Multi-Antigen T Cell Therapy Shows Safety and Durable Responses in Pediatric Brain Tumors: Phase 1 ReMIND Trial
Key Insights
The ReMIND phase 1 trial demonstrated that multi-antigen-targeting TAA-T cell therapy is feasible and well tolerated in pediatric patients with DIPG and relapsed CNS tumors, with 88% of adverse events being low grade.
Three proteins — WT1 (search), PRAME (search), and survivin (search) — were targeted simultaneously, and the maximum tolerated dose was established at 8 × 10⁷ cells/m² per dose for future phase 2 studies.
Among relapsed/refractory patients, median overall survival post-infusion was 12.7 months, with one patient remaining in complete response over 3 years and others disease-free beyond 4 years.
A first-in-human phase 1 trial led by Children's National Hospital has demonstrated that a novel multi-antigen-targeting T cell therapy can be safely administered intravenously to children and young adults with some of the deadliest brain tumors, including diffuse intrinsic pontine glioma (search) (DIPG) and relapsed central nervous system (CNS) malignancies. The findings, published in Nature Medicine, establish key milestones — feasibility of manufacturing, a maximum tolerated dose (MTD), and an early safety profile — needed to advance the therapy into future phase 2 studies.
The ReMIND trial (Research on Multi-antigen T cell Infusion Against Neuro-oncologic Disease), which enrolled 33 patients between September 2018 and July 2024, evaluated an autologous T cell product targeting three tumor-associated antigens (TAAs): WT1 (search), PRAME (search), and survivin (search). These proteins are commonly expressed in pediatric brain tumors, and the multi-target design may help address tumor heterogeneity — a major barrier to successful treatment of aggressive childhood cancers.
"We were excited to see that we could preserve safety and quality of life while generating anti-tumor responses by attacking three targets at once," said Eugene Hwang, MD, chief of Oncology at Children's National and co-senior author of the study.
Study Design and Patient Population
The trial employed an adaptive dose-finding design across three arms. Arm A enrolled 16 patients with newly diagnosed radiographic DIPG, of whom 11 received infusions at a median of 3.5 months after radiotherapy. Arms B and C enrolled patients with relapsed, refractory, or recurrent non-brainstem high-grade CNS malignancies; 18 of 28 patients in arm B and 4 of 7 in arm C were infused. Arm C included fludarabine/cyclophosphamide lymphodepletion before the first infusion, while arms A and B did not.
Three dose levels (DLs) were evaluated: 2 × 10⁷ cells/m² (DL1), 4 × 10⁷ cells/m² (DL2), and 8 × 10⁷ cells/m² (DL3). Both arms A and B independently dose-escalated to DL3 as the model-estimated MTD and recommended phase 2 dose (RP2D).
TAA-T cells were manufactured by culturing antigen-presenting cells with peptide libraries spanning WT1 (search), PRAME (search), and survivin (search), then co-culturing autologous T cells with the peptide-loaded APCs. The final products were primarily CD3+ T cells (median 96.9%).
Manufacturing Feasibility
Of 48 patients who underwent blood procurement, 41 (85.4%) had TAA-T products manufactured at DL1 or higher. Following optimization of procurement and manufacturing protocols — including increasing the maximum blood collection volume and implementing an absolute lymphocyte count eligibility criterion — 100% of subsequently procured patients met DL1 or above for at least one infusion. The median number of available doses per manufacturing run increased significantly from 3.8 to 11.0 (P < 0.01).
Safety Profile
Treatment was generally well tolerated across all arms, with 88% (293 of 332) of all adverse events (AEs) being low grade. The most common treatment-emergent AEs included fatigue, headache, and vomiting. Grade ≥3 AEs at least possibly related to TAA-T therapy occurred in 9% of patients (n = 3).
One dose-limiting toxicity (DLT) occurred in arm A at DL2: a patient with DIPG developed hydrocephalus, tumor edema, and respiratory distress 10 days post-infusion, which was fatal. No additional DLTs occurred in any arm. Two serious adverse events (SAEs) were reported, both in patients with the largest tumor areas in their respective arms.
Notably, only one participant met criteria for grade 1 cytokine release syndrome (CRS), with a temperature of 38.1°C on day 6 post-infusion and normal blood pressure and oxygen saturation. This contrasts favorably with CAR-T cell trials in similar populations, where inflammatory toxicities including CRS were more common, particularly following intravenous administration.
"This study represents an important step toward developing safer and more effective T-cell therapies for children with devastating brain cancers," said Catherine Bollard, MBChB, MD, senior vice president and chief research officer at Children's National and co-senior author. "Even in this early-stage trial focused on safety, we were encouraged to see lasting clinical benefit in several patients who otherwise had very few options."
Clinical Responses
In arm A (DIPG), median progression-free survival (PFS) from diagnosis was 10.5 months, and median overall survival (OS) from diagnosis was 13.7 months. Among ten evaluable patients, best response was stable disease (SD) in 60% and progressive disease (PD) in 40%.
In arms B/C, median PFS from first infusion was 5.0 months, and median OS was 12.7 months. Among ten patients evaluable for objective radiographic response, best response was SD in 50% and PD in 50%. The median percent change in tumor area from baseline to best response was 10% (range, −93% to +91%).
Several patients achieved remarkable outcomes. One participant (P41), an 8-year-old with EWSR1-BEND2 rearranged astroblastoma, achieved a complete response (CR) one year after completing three infusions and remains disease-free more than 3 years later. Two additional patients — one with multiply relapsed medulloblastoma (P35) and another with recurrent pediatric glioblastoma (P36) — remain disease-free more than 2.6 and 4.3 years post-infusion, respectively.
Immunobiological Correlates
TCR profiling demonstrated that product-derived clonotypes were detectable post-infusion, with dominant clones persisting in blood up until the time point most proximate to confirmed disease progression. Plasma cytokine analyses revealed significant increases in IL-6 and IL-8 log2 fold changes at weeks 1 through 4 post-infusion (P < 0.05), along with a transient increase in CCL2 at week 2 in arms B/C. Notably, CXCL10 — an interferon-associated chemokine linked to CAR-T toxicity — showed no significant changes, suggesting a differentiated toxicity profile.
Elevated IL-6 and IL-8 levels were associated with moderate (grade 2) but not severe (grade ≥3) AEs. Patients with maximum IL-6 or IL-8 concentrations ≥500 pg/ml experienced more grade ≥2 AEs than those below this threshold (P < 0.05).
Limitations and Next Steps
The trial was not powered to evaluate efficacy, and the authors caution that conclusions about clinical benefit must await more definitive studies. Additional limitations include heterogeneity in patient tumor phenotype and clinical characteristics, lack of direct measures of T cell migration into tumors, and the confounding effect of post-study therapies on survival interpretation.
Two subsequent phase 1 trials are now underway: LIFT, which will examine low-frequency ultrasound to promote tumor penetration of TAA-T cells, and IMPACT (NCT06193759), which uses the same TAA-T platform against individualized neoantigens in children with embryonal tumors and recurrent ependymoma.
"Pediatric tumors are one of the greatest challenges in cancer research, with children still facing extremely limited treatment options, and existing treatments often causing severe side effects," said David Scott, director of Cancer Grand Challenges, which supports the NexTGen team co-led by Dr. Bollard.