ST2 as a biomarker of cytokine release syndrome following CD19-targeted chimeric antigen receptor T-cell therapy

Abstract

Abstract Introduction Axicabtagene ciloleucel (axi-cel) is a CD19-targeted chimeric antigen receptor (CAR) T-cell therapy with curative potential in relapsed/refractory (R/R) B-cell lymphoma (BCL). However, it is associated with potentially severe, acute immune-mediated toxicities such as cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). Thus, strategies to identify patients at risk and prevent these adverse events are warranted. The suppression of tumorigenicity 2 (ST2) is involved in many inflammatory signaling pathways. We hypothesized that this cytokine could identify patients at an increased risk of high-grade CRS. Methods This was a retrospective, single-center study including adult patients with B-cell lymphoma (BCL) treated with axi-cel who had available cryopreserved plasma samples prior to lymphodepleting chemotherapy (LDC); and if they had paired samples on days -1 or 0 (day of CAR-T infusion), these were also analyzed. The primary endpoint was the incidence of high-grade CRS according to the levels of ST2. Secondary endpoints included the incidence of ICANS according to the levels of ST2. Serum levels of TNFa, IL-1a, IL-1b and IL-33 were also evaluated. To measure the cytokine plasma concentrations a Simple Plex kit (Biotechne) was used, the analysis was performed at the institution´s immunology laboratory. Also, a paired sample analysis was performed in order to identify the variability (delta) of ST2 between the pre-LDC and pre-infusion timepoints, when available. Logistic and linear regression were used for univariate and multivariate analyses. Wilcoxox test was applied for variability analysis. Cut-off values for continuous variables were defined using the maxstat method. All analyses were conducted using R 4.4.2. Results A total of 89 patients with R/R BCL who received axi-cel from 2019 to 2024 were included. Fludarabine and cyclophosphamide constituted the lymphodepleting therapy (LDC) in all patients. Pre-LDC samples were available in 88 patients, whereas pre-infusion samples were available in 22 patients. In terms of baseline characteristics, median age was 57 years (range, 23-78), most patients (57%) were male and had a DLBCL diagnosis (90%). Median number of previous lines was 2; 26% had received an autologous stem cell transplant. Thirty-one patients (37%) had a high CAR-HEMATOTOX. First, we carried out a paired sample analysis (pre-LDC and pre-infusion) of cytokine levels in samples available at both timepoints (n=22). This analysis showed a significant variability between these two timepoints, with a reduction of ST2, IL-1ra, and TNFa levels from LDC to pre-infusion (ΔST2= -21704 pg/mL, p=0.0124; ΔIL-1ra= -125 pg/dL, p=0.024; ΔTNFa= -1.95 pg/dL, p=0.015). No changes were found in IL-1B and IL-33 levels. Eighty-three (93%) patients presented any-grade CRS (41.6% grade ≥2 and 12.3% grade ≥3) and 45 (51%) developed any-grade ICANS (33% grade ≥2 and 15.9% grade ≥3). In univariate analysis (UA), factors associated with grade ≥2 CRS included LDH (OR 2.63, 95%CI, 1.11-6.48, p=0.031), male sex (OR 0.28, 95%CI, 0.11-0.67, p=0.005) and higher pre-infusion levels (≥29,261 pg/mL) of ST2 (n=22) (OR 2.91, 95%CI, 1.32-9.59, p=0.03). None of the other cytokines (IL-1ra, IL-1B, IL-33, TNFa) showed a correlation with this endpoint. In multivariable analysis (MVA), only the pre-infusion plasma level of ST2 (OR 3.41, 95%CI, 1.31-18.46, p=0.05) maintained its significance for grade ≥2 CRS. Focusing on grade ≥3 CRS, the only significant factor was a higher level of ST2 at pre-LDC time point (n=88) (≥37,565 pg/mL) in both UA (OR 8.14, 95%CI, 2.12-40.12, p=0.004) and MVA (OR 14.54, 95%CI, 2.64-128.99, p=0.005). Concerning ICANS, no cytokines were associated with any-grade or grade ≥3. Conclusion In this cohort of BCL patients treated with axi-cel, higher levels of ST2 at different time points before CAR T-cell infusion were predictive of the development of high-grade CRS. These results require external validation with larger, multicenter cohorts. We did not observe an association of pre CAR T-cell infusion levels of ST2 with ICANS.