CAR natural killer T cells that co-express GD2 and interleukin-15 were found to be safe and to demonstrate evidence of in vivo expansion and localization to metastatic sites in patients with stage IV relapsed/refractory neuroblastoma.
CAR natural killer T (NKT) cells that co-express GD2 and interleukin (IL)-15 were found to be safe and to demonstrate evidence of in vivo expansion and localization to metastatic sites in patients with stage IV relapsed/refractory neuroblastoma, according to data from a first-in-human phase 1 trial (NCT03294954) presented during the 2021 American Society of Gene and Cell Therapy Annual Meeting.1
Of 11 patients treated, 2 achieved a partial response, 1 of whom is now in complete remission (CR) after receiving the second dose of CAR-NKT cell infusion, according to Andras Heczey, MD, lead study author and an assistant professor of Pediatrics-Oncology at Baylor College of Medicine. Additionally, 4 patients had stable disease, and 5 experienced disease progression.
“CAR NKTs can be effectively manufactured to clinical scale,” Heczey said. “These cells are safe and there’s evidence of in vivo expansion and localization to metastatic sites. Now, we are seeing tumor regression in 3 patients, including 1 durable CR.”
CAR T-cell therapy has demonstrated good clinical outcomes in multiple lymphoid malignancies, and this modality has changed the standard of care for many patients with these cancers. In terms of treatment for solid tumors, however, strategies with these therapies still need to be refined.
“Perhaps one way to improve [these strategies] is by looking at specific subsets of lymphocytes that may have better antitumor properties for solid tumors than general bulk expansion of αβ T cells,” Heczey said. “We believe that NKT cells are such a subset.”
T cells recognize peptides presented by the MHC class I and II molecules and they have a variety of T-cell receptors that allow them to do this. Unlike these T cells, NKT cells can recognize hydrophobic molecules, such as glycolipids, presented by the CD1d molecules. Additionally, they express an invariant T-cell receptor (iTCR) that incorporates the Vα24 chain preferentially paired with β11 chain, according to Heczey.
“All of us share this iTCR in our NKT cells, which is important, as they can regulate the same processes in all of us,” Heczey explained. “As the field is moving toward off-the-shelf approaches, these NKTs do not have potential for induced graft-vs-host-disease.”
In terms of their potential efficacy, NKTs traffic in a chemokine-dependent manner to solid tumors, specifically to neuroblastoma, in a CCL2/CCL20-dependent manner. More importantly, according to Heczey, is that once at the tumor site, the NKTs target tumor-associated macrophages (TAMs) through a CD1d/iTCR interaction; CD1d is expressed on TAMs. Through this indirect mechanism, NKTs infiltration in neuroblastoma tissues has been associated with improved outcomes.
Moreover, preclinical data have demonstrated that NKTs engineered with a GD2-expressing CAR can target tumor cells directly, and indirectly, by destroying tumor-supporting TAMs; this has been shown in neuroblastoma models.
These results provided the rationale to launch the first-in-human phase 1 study, which set out to examine genetically-engineered NKT cells expressing an optimized GD2-CAR and IL-15 to treat children with relapsed/refractory neuroblastoma in the autologous setting. “IL-15 was included because this cytokine can enhance the expansion persistence, and as a result, antitumor properties of NKTs in the preclinical setting,” Heczey noted.
To be eligible for enrollment, patients had to have confirmed relapsed/refractory high-risk neuroblastoma, a life expectancy of at least 12 weeks, and be between the ages of 1 and 21 years.2
The study utilized a 3+3 dose-escalation design, and patients were treated at 4 dose levels: dose level 1 of 3 x 106, dose level 2 of 1 x 107, dose level 3 of 3 x 107, and dose level 4 of 1 x 108. Additionally, a standard lymphodepletion regimen was given and this comprised cyclophosphamide given at a dose of 500 mg/m2 on days -4, -3, and -2, and intravenous fludarabine at a dose of 30 mg/m2 on days -4 and -3, prior to cell infusion.
On day 0, patients received the CAR-NKT cell infusion and were evaluated weekly in the outpatient setting. At week 2 following infusion, patients were biopsied, and at week 4, patients underwent imaging and response-to-therapy evaluation.
“The safety assessment focuses on the first 28 days post infusion,” Heczey added.
The primary end point of the study is safety, while other clinical end points include CAR-NKT cell persistence and trafficking, as well as antitumor responses.
Of the 11 patients enrolled and treated on the study thus far, the median age was 7 years (range, 2-12), and all had relapsed/refractory, high-risk, stage IV neuroblastoma. Moreover, all patients were able to receive the full prescribed lymphodepletion regimen and the full CAR-NKT cell infusion.
“NKTs are a small subset of peripheral lymphocytes…so there was a question of whether the appropriate amount of NKTs could be generated ex vivo,” Heczey said. “Thus far, we have been able to generate these products at high purity, up to 98% and sometimes higher, with nice CAR transduction efficiency, as well.”
With this, the absolute number of NKTs expanded is approaching, even in smaller children, about 1 billion cells, according to Heczey. In the beginning, investigators planned these expansions to account for approximately 21 days for manufacturing. “Now, we are down to 14 days or less and we are still generating a large amount of highly potent NKT cells,” Heczey said.
To measure the interpatient product heterogeneity, and differences in antitumor properties, NKT cells were repeatedly exposed to tumor cells in a co-culture assay. The CAR-NKT cells were replated with neuroblasts multiple times throughout this process. Investigators then evaluated their ability to expand and maintain their cytolytic ability and measured expression of exhaustion markers.
“Most CAR-NKT products can maintain their ability to kill, even after 5 rounds of co-culture,” Heczey noted. “However, there are some, where the ability to kill drops significantly by the fifth round. If you look at the fold expansion, there are obvious differences between these products. Perhaps the highest fold expansion is, for our product, at the beginning, as we would expect.”
In terms of exhaustion markers in the CAR-NKT products, differences were seen in proportion of TIM-3 and PD-13 dual-positive populations, which could suggest that those cells may be more exhausted than others, Heczey explained.
After infusion, CAR-NKT cells expanded in the peripheral blood of patients. Additionally, the absolute number of CAR-NKT cells significantly varied patient to patient and did not appear to be dependent on dose level. In all but 1 patient enrolled on the study, the frequency of CAR-NKT cells was higher than prior to infusion, according to Heczey. Moreover, in all patients, CAR-NKT cells were detected in the peripheral blood, with a peak expansion around week 2 and week 3 after infusion. Additionally, at all dose levels, CAR-NKT cells were found in the neuroblastoma tumor sites, as well as in bone marrow biopsies.
In terms of GD2-expression in the CAR-NKT cells, 2 patients were found to have downregulated expression. Investigators believe that this loss of expression could have been occurring prior to enrollment to study, but there are no biopsies to confirm this, according to Heczey.
In addition to CAR-NKT cell expansion, tumor burden, which was measured by Curie scores, was also found to be associated with antitumor activity in patients; those who had high area under the curve/Curie score values were the ones who experienced antitumor activity.
In terms of safety, no dose-limiting toxicities were observed in the first 28 days following CAR-NKT infusion. Most grade 3 or 4 adverse effects (AEs) reported were mostly hematologic, which is consistent with what has been seen in patients who have received a cyclophosphamide/fludarabine lymphodepletion regimen, according to Heczey. The most common grade 4 AEs included neutropenia (n = 10/11), lymphopenia (n = 6/11), leukopenia (n = 6/11), and thrombocytopenia (n = 2/11).