Systemic IL-15 promotes allogeneic cell rejection in patients treated with natural killer cell adoptive therapy

Abstract

Natural killer (NK) cells are a promising alternative to T cells for cancer immunotherapy. Adoptive therapies with allogeneic, cytokine-activated NK cells are being investigated in clinical trials. However, the optimal cytokine support after adoptive transfer to promote NK cell expansion, and persistence remains unclear. Correlative studies from 2 independent clinical trial cohorts treated with major histocompatibility complex-haploidentical NK cell therapy for relapsed/refractory acute myeloid leukemia revealed that cytokine support by systemic interleukin-15 (IL-15; N-803) resulted in reduced clinical activity, compared with IL-2. We hypothesized that the mechanism responsible was IL-15/N-803 promoting recipient CD8 T-cell activation that in turn accelerated donor NK cell rejection. This idea was supported by increased proliferating CD8+ T-cell numbers in patients treated with IL-15/N-803, compared with IL-2. Moreover, mixed lymphocyte reactions showed that IL-15/N-803 enhanced responder CD8 T-cell activation and proliferation, compared with IL-2 alone. Additionally, IL-15/N-803 accelerated the ability of responding T cells to kill stimulator-derived memory-like NK cells, demonstrating that additional IL-15 can hasten donor NK cell elimination. Thus, systemic IL-15 used to support allogeneic cell therapy may paradoxically limit their therapeutic window of opportunity and clinical activity. This study indicates that stimulating patient CD8 T-cell allo-rejection responses may critically limit allogeneic cellular therapy supported with IL-15. This trial was registered at www.clinicaltrials.gov as #NCT03050216 and #NCT01898793.

© 2022 by The American Society of Hematology.

Figures

Graphical abstract

Figure 1.

N-803 promotes donor NK and recipient CD8 T-cell expansion in vivo. (A) WUSTL and UMN trial schemas. Briefly, relapsed/refractory AML patients were lymphodepleted with fludarabine (25 mg/m3 × 5) on study days -6 to -2 and cyclophosphamide (60 mg/kg × 2) on study days -5 and -4. On study day -1, related, haploidentical donors were apheresed, NK cells were purified (WUSTL) or enriched (UMN) and activated with IL-12, IL-15, and IL-18 (WUSTL), IL-2 (UMN cohort 1), or N-803 (UMN cohort 2). Products were washed and infused into patients on study day 0 (NK purity for each cohort indicated as a percentage). Infused products were supported with IL-2 (WUSTL/UMN cohort 1) or N-803 (WUSTL/UMN cohort 2). (B) N-803 concentration in the PB from patients at the indicated times. (C) Donor NK cell expansion over time, as determined by flow cytometry between IL-2- (blue) and N-803 (purple) supported WUSTL patients (IL-2 n = 6; N-803 n = 7). (D) Fold reduction in cells from day 21 compared with maximal measure NK cells, typically days 8 through 14. (E) Representative overlay viSNE plot of purified donor NK cells (baseline, BL), infusion product (activated, ACT), and in vivo differentiated donor ML NK cells assessed by mass cytometry. (F-J) WUSTL patient CD8 T cells from PBMC and bone marrow (BM) were assessed by mass cytometry at the indicated days after NK cell infusion. (F-G) Summary data depicting recipient CD8 T-cell frequency (of CD45+ lymphocytes) in the (F) PBMC (IL-2; day 7 n = 8, day 14 n = 3; N-802 n = 6) and (G) BM (day 7 n = 4, day 14 n = 3). (H) Summary data showing percent Ki-67+ CD8 T cells in recipient BM at 7 and 14 days, after NK cell infusion (day 7 n = 4, day 14 n = 3). (I-J) Summary data showing absolute CD8+ (I) and Ki67+ CD8 (J) T-cell numbers in the PBMC at the indicated days after infusion (IL-2 day 7 n = 8, day 14 n = 3; N-802 n = 6). (K-N) Patients treated on UMN trials using IL-2 activated NK cells supported in vivo with IL-2 (gold) or N-803 (red) were also assessed by mass cytometry (see schema, supplemental Figure 2). Summary data depicting percent CD8 T cells from the (K) PBMC before (day 0) and the indicated days after NK cell infusion (IL-2 day 0 n = 6, day 7 n = 3, day 14 n = 6; N-802 day 0 n = 3, day 7 n = 2, day 14 n = 3), (L) BM (IL-2 n = 6; N-802 day 0 n = 3, day 14 n = 5). Summary data depicting percent CD4 T cells from the (M) PBMC and (N) BM. Summary data were analyzed using 2-way analysis of variance. Mean is depicted with error represented as standard error of the mean. P values are indicated within the graphs; no significant differences were detected in panels K-N.

Figure 1.

N-803 promotes donor NK and recipient CD8 T-cell expansion in vivo. (A) WUSTL and UMN trial schemas. Briefly, relapsed/refractory AML patients were lymphodepleted with fludarabine (25 mg/m3 × 5) on study days -6 to -2 and cyclophosphamide (60 mg/kg × 2) on study days -5 and -4. On study day -1, related, haploidentical donors were apheresed, NK cells were purified (WUSTL) or enriched (UMN) and activated with IL-12, IL-15, and IL-18 (WUSTL), IL-2 (UMN cohort 1), or N-803 (UMN cohort 2). Products were washed and infused into patients on study day 0 (NK purity for each cohort indicated as a percentage). Infused products were supported with IL-2 (WUSTL/UMN cohort 1) or N-803 (WUSTL/UMN cohort 2). (B) N-803 concentration in the PB from patients at the indicated times. (C) Donor NK cell expansion over time, as determined by flow cytometry between IL-2- (blue) and N-803 (purple) supported WUSTL patients (IL-2 n = 6; N-803 n = 7). (D) Fold reduction in cells from day 21 compared with maximal measure NK cells, typically days 8 through 14. (E) Representative overlay viSNE plot of purified donor NK cells (baseline, BL), infusion product (activated, ACT), and in vivo differentiated donor ML NK cells assessed by mass cytometry. (F-J) WUSTL patient CD8 T cells from PBMC and bone marrow (BM) were assessed by mass cytometry at the indicated days after NK cell infusion. (F-G) Summary data depicting recipient CD8 T-cell frequency (of CD45+ lymphocytes) in the (F) PBMC (IL-2; day 7 n = 8, day 14 n = 3; N-802 n = 6) and (G) BM (day 7 n = 4, day 14 n = 3). (H) Summary data showing percent Ki-67+ CD8 T cells in recipient BM at 7 and 14 days, after NK cell infusion (day 7 n = 4, day 14 n = 3). (I-J) Summary data showing absolute CD8+ (I) and Ki67+ CD8 (J) T-cell numbers in the PBMC at the indicated days after infusion (IL-2 day 7 n = 8, day 14 n = 3; N-802 n = 6). (K-N) Patients treated on UMN trials using IL-2 activated NK cells supported in vivo with IL-2 (gold) or N-803 (red) were also assessed by mass cytometry (see schema, supplemental Figure 2). Summary data depicting percent CD8 T cells from the (K) PBMC before (day 0) and the indicated days after NK cell infusion (IL-2 day 0 n = 6, day 7 n = 3, day 14 n = 6; N-802 day 0 n = 3, day 7 n = 2, day 14 n = 3), (L) BM (IL-2 n = 6; N-802 day 0 n = 3, day 14 n = 5). Summary data depicting percent CD4 T cells from the (M) PBMC and (N) BM. Summary data were analyzed using 2-way analysis of variance. Mean is depicted with error represented as standard error of the mean. P values are indicated within the graphs; no significant differences were detected in panels K-N.

Figure 2.

N-803 hastens T-cell activation and allogenic rejection in mixed lymphocyte reactions. (A) MLR experimental design. Briefly, PBMCs were carboxyfluorescein diacetate succinimidyl ester (CFSE) labeled and incubated with unmatched, irradiated PBMCs with or without IL-2 and increasing concentrations of N-803. NK cells from stimulator donor were 12/15/18-activated and allowed to differentiate into ML-NK in parallel. (B-C) CD8 T cells were examined by flow cytometry from days 4 through 11 after incubation with allogenic stimulators. (B) Representative histograms depicting proliferation (CFSE dilution) and activation markers CD25 and CD38. (C) Summary data from panel B. (D) Summary data from CD4 T cells as examined in panels B and C. (E-I) 51CR-release killing assays using MLR-stimulated PBMCs (R; responder) as effectors against allogeneic in vitro differentiated ML NK cells as targets. (E) Killing assay schema. PBMCs were harvested from MLR on day 10 and cocultured with 51Cr-pulsed ML NK cells (matched to original allogenic stimulator [S] cells; S-ML NK) and killing assessed. (F) Summary data from panel E. (G-I) PBMCs incubated with IL-2 and 35 ng/mL N-803 were incubated with anti-MHC-I (G), anti-FAS-L (H), or anti-TRAIL (I) before addition of labeled ML-NK targets and allogenic killing assessed in the presence of blocking antibodies. N = 6 normal donor responders, 4 normal donor stimulator/targets in 2 independent experiments. Data were analyzed using 2-way analysis of variance, *P < .05, **P < .01, ***P < .001. Unless indicated (purple asterisk), statistics are for each condition, compared with IL-2 only condition. Purple asterisk indicates significance for 35 ng/mL N-803 + IL-2 compared with IL-2 only. Mean is depicted with error represented as standard error of the mean.