Myeloid Derived Suppressor Cell Kinetics in Peripheral Blood Correlate with Efficacy and Toxicity of CAR T Cell Therapy in Relapsed or Refractory B Cell Lymphoma

Introduction. CAR T-cell therapy is potentially curative for patients with relapsed or refractory B-cell lymphoma (BCL). However, about 60% of patients fail CAR T-cell therapy. Immune-related adverse events (irAE) like cytokine release syndrome (CRS) and immune effector cell associated neurotoxicity syndrome (ICANS) occur in an inflammatory state. Myeloid-derived suppressor cells (MDSCs) expand during acute inflammation as steady-state hematopoiesis switches to emergency myelopoiesis. We hypothesized that the acute inflammatory conditions created by CAR T-cell therapy may alter MDSC kinetics and investigated MDSC expansion in the context of CAR T-cell efficacy and toxicity in BCL.

Methods. Patients with BCL receiving standard-of-care CAR T-cell therapy were enrolled in a prospective sample collection study. Blood samples were collected before lymphodepletion (baseline), prior to and at early intervals after CAR T-cell infusion. Viable white blood cell (WBC) and MDSC counts were evaluated by multicolor flow cytometry. Phenotypically, MDSCs were defined as immature myeloid cells (CD11b ⁺ CD33 ⁺ HLA-DR ^-/low IL-4R ⁺) and divided into monocytic (M-MDSC, CD14 ⁺) and polymorphonuclear (PMN-MDSC, CD15 ⁺) subsets. Clinical response was assessed by Lugano criteria at 1, 3 and 6 months. Incidence and severity of CRS and ICANS were recorded. Cell counts were analyzed by descriptive statistics and reported as mean ± standard error or fold-change from baseline. Kinetics of MDSCs in blood and correlation with clinical outcomes were evaluated by Wilcoxon matched pair signed rank t-test and Mann Whitney U-test.

Results. Forty-three BCL patients (41 large BCL, 2 mantle cell lymphoma) were enrolled in the study. Median age was 64 (range 27-83), 32.6% were female. Racial and ethnic minority groups represented 14% of enrolled subjects. One patient developed grade 5 ICANS and was not evaluable for response. Thirty-seven (86%) developed leukopenia and all received granulocyte colony stimulating factor (G-CSF) starting on day 5 post-CAR T infusion. WBC count fell sharply from baseline following lymphodepletion (7.7E ⁺⁰⁵±8.6E ⁺⁰⁴ baseline vs 2.9E ⁺⁰⁵±9.2E ⁺⁰⁴/mL, p<0.0001) and continued to decline 2-3 days post-CAR T infusion (baseline vs 1.4E ⁺⁰⁵±2.1E ⁺⁰⁴/mL, p<0.0001) before returning to baseline within 1 week post-CAR T (i.e. within 3 days of G-CSF). MDSCs followed similar kinetics early on but significantly increased from baseline levels by 1-week post-CAR T (3.4E ⁺⁰³±5.8E ⁺⁰² baseline vs 8.5E ⁺⁰³±1.7E ⁺⁰² /mL, p<0.0001). This effect was mainly driven by an exuberant expansion of M-MDSC observed in 33/43 patients (median 6.1-fold increase from baseline at 1 week, range 0.5-4301).

Seven patients (16.3%) had progressive disease (PD) at 1 month, 6 (13.9%) additional patients progressed at 3 months and 4 (9.3%) more progressed at 6 months after CAR T-cell infusion. M-MDSC expansion 1-week post-CAR T was only observed in 3/7 refractory patients compared with 28/35 responding patients (Fisher's exact test, p=0.0635). Patients who progressed at 3 months had poorer M-MDSC retention compared to patients with sustained response (1.7E ⁺⁰²±6.1E ⁺⁰¹ PD vs 1.4E ⁺⁰³±2.4E ⁺⁰¹/mL at 2-week post-infusion, p<0.1). Similarly, patients who progressed at 6 months had lower M-MDSC compared with sustained responders (2.0E ⁺⁰²±8.1E ⁺⁰¹ PD vs 4.5E ⁺⁰²±1.4E ⁺⁰² ml at 1-month post-infusion, p<0.05).

Thirteen patients (30.2%) developed > grade 2 CRS and had lower PMN-MDSC retention rate 2-3 days post-CAR-T compared with those with no/low-grade CRS (4.4E ⁺⁰²±2.6E ⁺⁰² vs 1.4E ⁺⁰³±3.9E ⁺⁰²/mL, p<0.05). Similarly, patients who had ICANS > grade 2 (n=14, 32.5%) had lower PMN-MDSC early retention compared to those with no or low-grade ICANS (5.0E ⁺⁰²±2.5E ⁺⁰² vs 1.5E ⁺⁰³±4.1E ⁺⁰²/mL, p<0.05). Further analyses are currently underway to bridge these findings with the kinetics of CAR T-cells and T-cell repertoire post-CAR T.

Conclusions. The proinflammatory milieu after CAR-T and the use of G-CSF can lead to significant expansion of M-MDSCs. Rather than an impediment to CAR T therapy, M-MDSC expansion and retention are associated with favorable clinical response and PMN-MDSC with a lower risk of developing high grades CRS or ICANS. Whether MDSCs play a mechanistic role remains to be elucidated, but they may have a role as a biomarker in predicting outcomes after CAR T-cell therapy in BCL.