Endotheliopathy, Innate-Immune Activation and Hemostatic Misbalance Underlying CAR-T Cell Toxicities: Tools for Early and Differential Diagnosis

Background: The CAR-T cell-based immunotherapy constitutes one of the most revolutionary and feasible advances for the treatment of relapsed/refractory (R/R) hematologic malignancies nowadays. Nevertheless, the toxicities related to the treatment, such as cytokine release syndrome (CRS) and the immune effector cell-associated neurotoxicity syndrome (ICANS) can be life-threatening and early diagnosis is cornerstone. Emerging evidence points to endothelial damage as a crucial pathophysiological substrate. Furthermore, differential diagnosis among these toxicities and septic syndromes is usually challenging. Since their treatments are different, a laboratory tool for the differential diagnosis would be of great interest.

Methods: We prospectively included 62 patients admitted to our center for the treatment with CAR-T cell immunotherapy (76% anti-CD19, 24% anti-BCMA). Plasma samples were collected: (A) before the CAR-T cell infusion; (B), 24-48h after; (C), at the suspicion of the onset of any toxicity (fever, hypotension, hypoxia and/or neurotoxicity); and (D) 24-48h after immunomodulatory treatment. Biomarkers of endothelial dysfunction (sVCAM-1, sTNFR-1, ST-2, ANG-2), innate immunity activation (NETs, sC5b-9) and hemostasis/fibrinolysis balance (VWF:Ag, ADAMTS-13, alpha-2-antiplasmin, PAI-1) were evaluated by ELISA, fluorimetry or immunoturbidimetry. Levels of these biomarkers were also compared with those in septic patients and healthy donors.

Results: Patients who developed CAR-T cell-related toxicities (n=19) presented a significant increase in the levels sVCAM-1, sTNFR-1 and ST-2 at the clinical onset of the toxicity (point C) vs. the values at the post-infusional point (point B). Values expressed as median ± interquartile range (IQR) were of: VCAM-1 of 359 ng/mL ± 455 ng/mL vs. 223 ng/mL ± 499 ng/mL, p=0.028; sTNFR-1 of 4252 pg/mL ± 4733 vs. 3559 pg/mL ± 2259 pg/mL , p=0.023; and ST-2 of 191 ng/mL ± 130 ng/mL vs. 124 ng/mL ± 130 ng/mL, p=0.031, respectively. ST-2, already at point B, exhibited a good sensitivity for the early diagnosis of any CAR-T cell-related toxicity (AUC 0.7, P=0.020) (Panel A). Moreover, levels of ST-2, ANG-2, NETs and sC5b-9 at point B were also significantly higher in patients who required ICU admission due to CAR-T cell-related toxicity (n=17) than in those with milder clinical scenarios (ANG-2 of 2841 ng/mL ± 2959 pg/mL vs. 1729 ng/mL ± 1354 pg/mL, p= 0.043; ST-2 of 76 ng/mL ± 176 ng/mL vs. 28 ± 29, p=0.000; NETs of 9 μg/mL ± 6 μg/mL vs. 6 μg/mL ± 2 μg/mL, p=0.008; and sC5b-9 of 763 ng/mL ± 659 ng/mL vs. 440 ng/mL ± 351 ng/mL, p=0.05, respectively). Patients with sepsis presented superior levels of NETs, sC5b-9, VWF:Ag and PAI-1 Ag than those with CAR-T cell toxicity at their onset (point C). When considering only severe patients who required ICU admission, in both groups, the ROC curve from the regression model created by the combination of the four biomarkers showed an excellent discrimination for the diagnosis of severe sepsis or septic shock (AUC 0.96, p=0.000) (Panel B).

Conclusions: Biomarkers based on endotheliopathy, innate-immunity activation and hemostatic misbalance pathways have a potential role in the laboratory confirmation of the CAR-T cell-related toxicities, for the prediction of their clinical severity and, finally, for the differential diagnosis between CAR-T cell toxicities and septic syndromes.

Ortiz-Maldonado:Kite: Honoraria; Novartis: Honoraria; BMS: Honoraria; Janssen: Honoraria. Fernández De Larrea:Janssen: Consultancy, Honoraria, Research Funding; BMS: Consultancy, Honoraria, Research Funding; Amgen: Consultancy, Honoraria, Research Funding; Takeda: Honoraria, Research Funding; GSK: Honoraria; Sanofi: Consultancy; Beigene: Consultancy, Honoraria; Pfizer: Honoraria. Diaz-Ricart:Cellphire Therapeutics Inc: Research Funding.