GD2 ‐ CAR T Cell Therapy for H3K27M + Diffuse Intrinsic Pontine Glioma: A Phase I Clinical Trial and Mechanistic Insights

Gliomas are the most prevalent malignant tumours of the central nervous system, and among them, diffuse midline gliomas (DMGs) are particularly challenging. These tumours are difficult to surgically remove due to their deep location in the brainstem or spinal cord, and they exhibit poor responses to chemoradiotherapy. The H3K27M mutation is the most common genetic alteration in DMGs and is closely associated with tumour aggressiveness and drug resistance. Over the past decade, chimeric antigen receptor T (CAR T) cell immunotherapy has shown remarkable efficacy in patients with chemotherapy-refractory haematologic malignancies, including lymphoma, leukaemia and multiple myeloma [1]. However, its application to solid tumours, particularly brain tumours, still faces significant hurdles. Autologous CAR T cells have achieved impressive results in refractory B cell and plasma cell malignancies but have yet to demonstrate sustained antitumour effects in solid tumours, including brain cancers. The disparity in activity between liquid cancers and solid/brain tumours may stem from a lack of safe targets with high homogeneous expression, insufficient T cell trafficking and tumour microenvironment-induced T cell dysfunction [2, 3]. A recent clinical trial involving GD2-CAR T cells for neuroblastoma highlighted the challenges of applying CAR T therapy to solid tumours but also suggested potential pathways for success. In 2022, early results from the first patients in Group A of the NCT04196413 trial were published, indicating that GD2 is highly expressed on H3K27M-mutated glioma cells. This discovery provided a basis for the preclinical success of GD2-CAR T cells and laid the groundwork for the first-in-human Phase I clinical trial (NCT04196413) [4]. Monje and colleagues conducted an open-label, nonrandomised Phase I clinical trial (NCT04196413) to evaluate the safety, tolerability and maximum tolerated dose of GD2-CAR T cell therapy in patients with H3K27M+ DMGs (Figures 1 and 2). The study included two dose groups: DL1 (1 × 106 kg−1) and DL2 (3 × 106 kg−1), each receiving different doses of GD2-CAR T cells. The primary objectives were to assess manufacturing feasibility, tolerability and identify the maximum tolerated intravenous (IV) dose, while secondary objectives included evaluating early efficacy through radiological response on MRI and clinical improvements in neurological function. Thirteen patients were enrolled, with 11 receiving GD2-CAR T cell therapy. All patients had completed primary radiotherapy at least 4 weeks prior to enrollment, with no additional antitumour therapies (including targeted or immunotherapies) allowed during the study. GD2-CAR T cells were successfully produced for all patients. No dose-limiting toxicity was observed in the DL1 group, but three patients in the DL2 group developed dose-limiting cytokine release syndrome (CRS), thus DL1 was established as the maximum tolerated IV dose. Nine patients received intracerebroventricular (ICV) injections, with no dose-limiting toxicity observed. While all patients showed neurotoxicity related to tumour inflammation, these symptoms were effectively managed with intensive monitoring and care. Four patients experienced major reductions in tumour volume (52%, 54%, 91% and 100%), and three others showed smaller reductions. One patient achieved a complete response, with the tumour fully disappearing on imaging; this patient has maintained the complete response for over 30 months since enrollment. Additionally, nine patients exhibited neurological benefits, as measured by protocol-guided clinical improvement scores (CIS) (CIS is simple tool to quantify changes in the neurological exam in which each item tested in a comprehensive neurological exam is assigned one point if improved, and −1 point if worse than preinfusion baseline for each infusion). Compared to other H3K27M-targeted therapies such as ONC201, which report an overall survival (OS) of 11–15 months in H3K27M+ DMGs [5], this study demonstrated superior outcomes, with an OS of 17.6 months in DIPG patients and 31.96 months in spinal DMG (sDMG) patients. Despite notable tumour reductions (52%, 54%, 91% and 100%) and neurological improvements, only two DIPG patients achieved long-term survival, underscoring the need to clarify mechanisms limiting durable benefit. Treatment toxicity emerged as a major constraint: all patients developed CRS after IV infusion, with Grade 4 CRS at DL2 halting treatment in three cases. Tumour inflammation-associated neurotoxicity (TIAN) occurred in 91% post-IV and 100% postinitial ICV infusion. CRS led to treatment discontinuation and early mortality, while TIAN caused neurological deficits—including mass-effect edema and CSF flow obstruction—that disrupted immune signalling. The immunosuppressive TME further compounded these issues, with elevated CSF TGF-β postinfusion correlating with disease progression. Collectively, these factors offset the benefits of tumour reduction by accelerating deterioration and mortality. Although optimised dosing and administration mitigated some risks, future strategies will require predictive CRS and TIAN biomarkers to enable timely prophylactic interventions. This trial represents one of the first successful applications of CAR T cell therapy in solid tumours. Key factors underlying these promising results include the precise selection of GD2 as a target antigen, given its high and uniform expression on H3K27M-mutant glioma cells (Table 1) [4]. This expression pattern is closely associated with the tumour's aggressive biology, as the H3K27M mutation drives progression through aberrant epigenetic regulation, while GD2 overexpression may further enhance invasiveness and immune evasion. Importantly, GD2 expression in normal CNS tissue is minimal, significantly reducing the risk of off-target toxicity and allowing CAR T cells to precisely target tumour cells while sparing healthy tissue—a major advantage in CAR T cell therapy [8]. Moreover, GD2 maintains its high expression within the restrictive microenvironment of the blood–brain barrier (BBB), reinforcing its status as an optimal target for CAR T cell therapy [9]. Monje et al. also optimised the dosing strategy, determining DL1 as the maximum tolerated dose through a 3 + 3 dose escalation design, avoiding the toxicity associated with higher doses. Furthermore, the infusion method was adjusted, initially using IV injection and later transitioning to direct injection into the cerebrospinal fluid based on patient response, which helped reduce adverse reactions and improve therapeutic efficacy. This trial represents a significant advance in CAR T cell therapy for solid tumours, particularly CNS malignancies, by addressing key challenges including target antigen heterogeneity, the immunosuppressive TME, and fibrotic barrier [10]. Unlike previous GBM therapies targeting EGFRvIII or IL13Rα2, which were limited by antigen escape and immune suppression, this trial leveraged GD2's highly uniform expression on H3K27M-mutant cells. This specificity reduces on-target, off-tumour toxicity while enhancing T cell infiltration, tumour clearance and sustained activation in patient-derived organoids [10]. Additionally, direct intracranial administration bypassed the BBB, improving tumour access and overcoming a major barrier to effective treatment. Collectively, these findings establish a new paradigm for CAR T cell therapy in refractory solid tumours and underscore the feasibility of refining this approach for clinical application. However, this study has some limitations. The patients recruited had relatively good baseline health, making it challenging to compare survival data with historical data for DMG patients. Future Phase II trials are needed to assess whether GD2-CAR T therapy improves patient survival. Additionally, despite clinical and imaging benefits in nine patients, only two DIPG patients survived. Notably, deceased patients achieved a median OS of 17.6 months from diagnosis, markedly surpassing the historical DIPG benchmark of 11 months. Surviving patients maintained responses for 33 months (Patient 007) and 30 months (Patient 010) at data cutoff, with one patient sustaining complete remission beyond 30 months, indicating meaningful therapeutic benefit. Additionally, the TME may also negatively impact the penetration, survival and function of CAR T cells, as tumours can evade CAR T attacks by expressing immunosuppressive molecules or recruiting immunosuppressive cells [11]. Furthermore, this study underscores persistent challenges for CAR T cell therapy in brain tumours such as DMGs. A major barrier is the immunosuppressive TME, where regulatory T cells (Tregs), myeloid-derived suppressor cells (MDSCs) and tumour-associated macrophages (TAMs) inhibit infused immune cells while promoting tumour progression through angiogenesis and anti-inflammatory factor secretion, collectively reducing CAR T cell efficacy. Additional challenges include tumour heterogeneity enabling antigen escape and relapse, as well as delivery limitations imposed by the BBB and deep tumour locations that hinder CAR T cell trafficking [12, 13]. In summary, this study demonstrated the potential of GD2-CAR T cell therapy for treating H3K27M+ DMGs, with some patients showing tumour shrinkage and improved neurological function. However, treatment was accompanied by significant toxic reactions. Further research is needed to optimise the treatment regimen. To further enhance CAR T cell efficacy, emerging strategies focus on remodelling the tumour microenvironment by blocking immunosuppressive cytokines and related pathways. Approaches include engineering CAR T cells to secrete immunostimulatory cytokines or resist immunosuppressive signals. Additionally, multitargeted CAR T designs and adapter molecule strategies are being developed to address tumour heterogeneity, while specialised delivery vehicles and engineering modifications aim to improve CAR T cell trafficking and penetration into solid tumours [12, 13]. This study provides important insights into the application of CAR T cell therapy in solid tumours, particularly central nervous system tumours, and is expected to drive further advancements in the field. Congfa Jiang and Hening Xu analysed the case and wrote the manuscript. Congfa Jiang, Hening Xu and Jianqiao Shentu drafted the figure. Hening Xu and Jianqiao Shentu conceived the idea. Shiwei Duan reviewed and revised the manuscript. All authors have read and approved the final manuscript. The authors would like to thank PubMed for the valuable information. Figure 1 was created by BioRender (biorender.com). The authors have nothing to report. All authors have read and agreed to the published version of the manuscript. The authors declare no conflicts of interest. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.