T-Cell Attributes of Axi-cel Linked to Outcomes in LBCL

Jason Westin, MD

The T-cell attributes of axicabtagene ciloleucel (axi-cel; Yescarta) impacted tumor burden, efficacy outcomes, peak levels of proinflammatory cytokines, and toxicities such as neurologic events (NEs) and cytokine release syndrome (CRS) in patients with relapsed/refractory large B-cell lymphoma (LBCL), according to an analysis from the phase 3 ZUMA-7 trial (NCT03391466) presented during the 2022 AACR Annual Meeting.¹

Specifically, CCR7-positive/CD45RA-positive T cells that expressed CD27 and CD28 were associated with all efficacy metrics, including durability or response. Products with central memory T cells, defined as CCR7 positive and CD45RA negative, were linked with higher levels of proinflammatory cytokines and CRS of grade 2 or greater. Finally, CCR7-negative/CD45RA- negative, or differentiated, T cells were negatively associated with efficacy and were further linked to higher peak levels of additional proinflammatory serum molecules and high levels of grade 3 or greater neurologic events.

“The quality of the T cells appears to matter more than the quantity in the ZUMA-7 clinical trial,” Jason Westin, MD, director of the Lymphoma Clinical Research and section chief of Aggressive Lymphoma in the Department of Lymphoma/Myeloma, Division of Cancer Medicine of The University of Texas MD Anderson Cancer Center in Houston, said in a presentation of the data. “We found that optimizing the product composition towards the juvenile T-cell phenotype to find a CCR7-positive, CD45RA-positive, CD27-positive, and CD28-positive [product] may improve the axi-cel therapeutic index.”

Axi-cel is an autologous anti-CD19 CAR T-cell that was investigated in the phase 1/2 ZUMA-1 trial (NCT02348216), which examined axi-cel as therapy in adults with aggressive refractory non-Hodgkin lymphoma. In the trial, CAR T-cell peak expansion was associated with overall response rate, durability of response, and NEs. These findings also revealed associations of CCR7/CD45RA positivity with efficacy and toxicity outcomes, prompting investigators to further investigate these attributes in ZUMA-7.

Outcomes of ZUMA-7 showed event-free survival (EFS) improvement in 60% of patients with relapsed/refractory LBCL who were being treated with axi-cel as second-line treatment vs standard of care (HR, 0.398; 95% CI, 0.308-0.514; P <.0001)2 and led to recent approval of the therapy in this setting. Notably, the rates of toxicities in ZUMA-7 were much lower compared with ZUMA-1, specifically NEs of grade 3 or greater (21% vs 32%, respectively) and CRS of grade 3 or greater (6% vs 11%, respectively). Samples from a total of 170 patients were used for biomarker analyses and pharmacokinetic (PK), pharmacodynamic (PD), and T-cell composition studies were performed.

The PK profile between the 2 studies were consistent, with peak CAR T-cell expansion occurring around 7 days following infusion with a rapid decline thereafter. Peak levels of CAR T-cells were slightly higher in ZUMA-1 at 38.3 cells/μl (IQR, 14.7-83.0) vs 25.84 cells/μl (IQR, 8.15-57.93) in ZUMA-7, but time to peak was 8 days in both studies.

Notably, tumor regression was associated with peak CAR T-cell expansion, but this did not necessarily translate to durability or response in the second-line setting. “What that means is patients who had any response vs no response had a much higher peak CAR T-cell level. However, [when comparing] those with a durable or ongoing response vs those who have progression after initial response, we unfortunately do not see any difference in the CAR T-cell levels. Therefore, the peak levels do not predict for durability of response,” Westin explained.

Looking at toxicities associated with CAR T-cell therapy, grade 3 or greater CRS was not statistically significantly associated with higher levels of peak CAR T-cell expansion, although Westin pointed out that these results could be unreliable due to a small patient sample. Conversely, grade 3 or greater NEs were significantly associate with higher peak levels of CAR T cells.

Serum PD analytes of chemokines and cytokines within the blood revealed associations with CRS and NEs in both ZUMA-1 and ZUMA-7, namely that more differentiated cells tended to have a positive association with more toxicities. CCR7-positive/CD45RA-negative, or central memory T cells, were positively associated with the toxicities with the higher Spearman R. Additionally, CCR7-negative T-effector memory cells and T-effector cells were also correlated with increased toxicity.

Products enriched with CCR7-positive/CD45RA-positive naïve T cell were significantly associated with durable response (P = .0408). Patients with a higher proportion of naïve T cells also had improved EFS (P = .0264), although Westin pointed out that axi-cel was favorable vs chemotherapy regardless of this biomarker (P <.0001).

Poor responses could be characterized by T-cell products exhibiting pronounced effector phenotypes with exhaustion markers. Lower levels of CD27/CD28-negative, PD-1/TIM-3/CD8-positive, or exhausted T cells, were linked with better EFS (P = .0156). On the other hand, high levels of CD27/CD28/PD-1/TIM-3/CD8-positive T cells, or juvenile phenotypes, were correlated with better EFS (P = .0127).

“When we look at the T cells within the axi-cel product, we find that CAR-positive and CAR-negative naïve T cells were associated with improved outcomes and more differentiated and exhausted T cells were associated with worse outcomes,” Westin concluded.

References

1. Filosto S, Vardhanabhuti S, Canales M, et al. Product attributes of axicabtagene ciloleucel (axi-cel) that associate differentially with efficacy and toxicity in second-line large B-cell lymphoma. Presented at: American Association for Cancer Research (AACR) 2022 Annual Meeting; April 8-13, 2022; Abstract CT004.
2. Locke F, Miklos DB, Jacobson CA, et al. Primary analysis of ZUMA‑7: a phase 3 randomized trial of axicabtagene ciloleucel (axi-cel) versus standard‑of‑care therapy in patients with relapsed/refractory large B-cell lymphoma. Blood. 2021;138(suppl 1):2. doi:10.1182/blood-2021-148039