HER2-Specific CAR T Cells Induce Early Efficacy Without Dose-Limiting Toxicities in Pediatric CNS Tumors

Repetitive locoregional infusion of HER2-specific CAR T cells did not induce any dose-limiting toxicities, according to findings from an interim analysis of the ongoing phase 1 BrainChild-01 trial (NCT03500991) that were published in Nature Medicine. Furthermore, the cells elicited clinical and correlative laboratory evidence of local central nervous system (CNS) immune activation in 3 pediatric patients with recurrent or refractory CNS tumors.

The clinical evidence included high concentrations of C-X-C motif chemokine ligand 10 (CXCL10) and C-C motif chemokine ligand 2 (CCL2) in the cerebrospinal fluid (CSF) and serum samples.

“This interim report supports the feasibility of generating HER2-specific CAR T cells for repeated dosing regimens and suggests that their repeated intra-CNS delivery might be well tolerated and activate a localized immune response in pediatric and young adult patients,” Nicholas Alexander Vitanza, MD, an assistant professor at the Ben Towne Center for Childhood Cancer Research, and a staff member of the Cancer and Blood Disorders Center, Brain Tumor Program, Apheresis, at Seattle Children’s, and coauthors, wrote in the study publication.

Although the integration of CAR T-cell therapy has provided a novel therapeutic modality to manage multiple hematologic malignancies, the utility of CAR T cells is not fully understood for pediatric patients with CNS tumors.

HER2 offers a valid target for CAR T-cell therapy in CNS tumors because it is widely expressed on a significant proportion of biologically diverse CNS tumors such as ependymoma, glioblastoma, and medulloblastoma, as well as CNS cancer stem cells. Moreover, HER2 is not expressed on normal CNS tissue.

Monoclonal antibodies, such as trastuzumab (Herceptin), are beneficial for patients with some HER2-expressing cancers but have limited activity in CNS tumors that require a therapy that crosses the blood-brain barrier. CNS tumors also harbor less HER2 expression compared with malignancies like breast cancer.

As such, directly administering HER2-directed therapy to the tumor site could be a lucrative strategy for patients with CNS tumors.

Preclinical data demonstrated that spacer length was correlated with improved activity of HER2-specific CAR T cells. Based on this, the single-institution BrainChild-01 trial used a medium-length spacer HER2CAR to evaluate repeated locoregional delivery of HER2-specific CAR T cells for pediatric patients with recurrent or refractory CNS tumors.

Following CAR T-cell manufacturing, patients were treated in the outpatient setting for up to 6 courses. Course 1 consisted of 3 weeks of a 1 x 107 dose of CAR T cells (DL1), followed by clinical evaluation in week 4. Course 2 consisted of 1 week of DL1 treatment, 2 weeks of a 2.5 x 107 dose of CAR T cells (DL2), followed by clinical and radiographic evaluation in week 4. Courses 3 through 6 retained the same dosing schedule at the highest tolerated dosing levels, which included 2 additional tiers: 5 x 107 [DL3] and 10 x 107 [DL4].

“The BrainChild-01 HER2CAR T-cell product was manufactured under a process designed to yield balanced numbers of CD4+ and CD8+ lentivirally transduced T cells exhibiting limited terminal differentiation with enrichment for the CAR+ population of cells mid-culture,” Vitanza and coauthors wrote.

The initial 3 patients were required to be from 15 to 26 years old. This age group is more capable of self-reporting neurologic changes compared with a younger patient population, so they were specifically used for the initial evaluation.

The first eligible 3 patients underwent apheresis and had CAR T-cell products that were in-line with release criteria. As such, the patients were assigned to the appropriate treatment arms: repeated locoregional CNS infusion into the CNS tumor or tumor cavity (arm A; n = 1) vs repeated locoregional CNS infusion into the ventricular system (arm B; n = 2).

All patients had undergone at least 3 prior tumor-directed surgical procedures, at least 1 prior irradiation, and at least 1 prior chemotherapy regimen. Additionally, all patients had presumed pediatric biology of their tumors.

A 19-year-old female patient enrolled on arm A was diagnosed with WHO grade III localized anaplastic astrocytoma. She had 1.95 x 109 total nucleated cells manufactured and 1.87 x 109 EGFRt+ CAR T cells manufactured. She received 6 doses of treatment.

Both patients enrolled on arm B were males with WHO grade III metastatic ependymoma. The first, a 16-year-old, had 3.2 x 109 total nucleated cells manufactured, 2.97 x 109 EGFRt+ CAR T cells manufactured, and received 9 doses of treatment. The second patient, aged 26, had 2.06 x 109 total nucleated cells manufactured, 1.87 x 109 EGFRt+ CAR T cells manufactured, and received 9 doses of treatment. The latter patient’s product in arm B had initial failure of viability screening, but with 2 additional manufacturing attempts, enough CAR T cells were generated to complete a minimum of 2 treatment courses.

The study was designed to primarily assess feasibility, safety, and tolerability, with assessment of CAR T-cell distribution and disease response as secondary objectives.

Patients experienced post-treatment symptoms. One patient who underwent imaging experienced radiographic evidence of treatment-mediated localized CNS immune activation.

Additional results showed that the most common adverse effects (AEs) observed in all patients were headache, pain at metastatic sites of spinal cord disease, and transient worsening of a baseline neurologic deficit. Additionally, the 2 patients on arm B experienced fever within 24 hours following infusion. These AEs were deemed possibly, probably, or definitely related to CAR T-cell therapy.

Systemic C-reactive protein elevation was also noted in all patients and overlapped with the timing of headaches and/or pain.

Regarding CSF cytokines and radiographic imaging, CAR T cells were not detected in any patient at any time point following infusion in CSF via flow cytometry or in peripheral blood via quantitative polymerase chain reaction. Non–CAR T cell populations of CD4+ and CD8+ T cells were detected in CSF after infusion.

Cytokines, including CXCL10, CCL2, granulocyte colony–stimulating factor, granulocyte-macrophage colony-stimulating factor, IFNα2, IL-10, IL12-p70, IL-15, IL1α, IL-6, IL-7, and tumor necrosis factor–α, were detected in the CSF following infusion. One patient also had elevated VEGF.

Additional studies are planned to evaluate the relationship between target antigen density and clinical toxicity and response.

With these findings, the trial is planned to enroll the broader age cohort of patients aged 1 to 26 years. Notably, the trial will include patients with diffuse midline glioma.

Two additional studies are also planned. BrainChild-02 (NCT03638167) will deliver EGFR-specific CAR T cells to pediatric patients with recurrent or refractory EGFR-positive CNS tumors. BrainChild-03 (NCT04185038) will deliver B7-H3–specific CAR T cells to pediatric patients with recurrent or refractory CNS tumors or diffuse intrinsic pontine glioma.

Gleaning the results of all 3 BrainChild studies, the investigators plan to use a multiplexed strategy to overcome tumor heterogeneity, which remains a challenge for drug development in this patient population, and antigen escape.

“Ultimately, the experience of the initial three patients treated on BrainChild-01 suggests that repeated locoregional HER2-specific CAR T-cell dosing might be feasible and that correlative CSF markers might be valuable in assessing on-target CAR T-cell activity in the CNS,” concluded Vitanza and coauthors.

Reference

1. 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. Published online July 12, 2021. Accessed July 20, 2021. doi:10.1038/s41591-021-0140