Filgrastim enhances T-cell clearance by antithymocyte globulin exposure after unrelated cord blood transplantation

Abstract

Residual antithymocyte globulin (ATG; Thymoglobulin) exposure after allogeneic hematopoietic (stem) cell transplantation (HCT) delays CD4⁺ T-cell immune reconstitution (CD4⁺ IR), subsequently increasing morbidity and mortality. This effect seems particularly present after cord blood transplantation (CBT) compared to bone marrow transplantation (BMT). The reason for this is currently unknown. We investigated the effect of active-ATG exposure on CD4⁺ IR after BMT and CBT in 275 patients (CBT n = 155, BMT n = 120; median age, 7.8 years; range, 0.16-19.2 years) receiving their first allogeneic HCT between January 2008 and September 2016. Multivariate log-rank tests (with correction for covariates) revealed that CD4⁺ IR was faster after CBT than after BMT with <10 active-ATG × day/mL (P = .018) residual exposure. In contrast, >10 active-ATG × day/mL exposure severely impaired CD4⁺ IR after CBT (P < .001), but not after BMT (P = .74). To decipher these differences, we performed ATG-binding and ATG-cytotoxicity experiments using cord blood- and bone marrow graft-derived T-cell subsets, B cells, natural killer cells, and monocytes. No differences were observed. Nevertheless, a major covariate in our cohort was Filgrastim treatment (only given after CBT). We found that Filgrastim (granulocyte colony-stimulating factor [G-CSF]) exposure highly increased neutrophil-mediated ATG cytotoxicity (by 40-fold [0.5 vs 20%; P = .002]), which explained the enhanced T-cell clearance after CBT. These findings imply revision of the use (and/or timing) of G-CSF in patients with residual ATG exposure.

Graphical abstract

Figure 1.

The effect of residual ATG exposure on CD4⁺ IR after CBT and BMT. (A-B) Residual ATG exposure affects CD4⁺ T-cell reconstitution (defined as ≥50 × 10⁶ CD4⁺ T-cells/L in 2 consecutive measurements after HCT) more in CB recipients (A; n = 155, P < .001) than in BM/PB recipients (B; n = 120, P = .74). (C) When ATG exposure is low, CD4⁺ IR is faster after CBT compared with BMT (P = .018). P values are for comparisons among all 4 groups (multivariate log-rank test), with correction for covariates.

Figure 2.

Differential lytic effect of residual ATG exposure on T-cell IR in CB or BM recipients. The lytic effect of residual ATG exposure on T cells is depicted in patients who received CBT or BMT. ATG exposure is depicted as area under the curve after HCT with 95% confidence intervals (red line, gray area), with T-cell reconstitution evaluated as mean cell amounts over time with 95% confidence intervals (blue line, gray area).

Figure 3.

Higher ATG-mediated cytotoxicity by neutrophils after in vivo G-CSF treatment. Neutrophils were derived from healthy volunteers not receiving G-CSF (n = 5) or from donors who received 10 mg/kg G-CSF for at least 5 days (n = 5). (A) Percentage of pHRodo+CTV+ neutrophils as evaluated with flow cytometry; difference between neutrophils with and without G-CSF per ATG concentration tested: 0 μg/mL, P = .005; 1 μg/mL, P = .005; 10 μg/mL, P = .002; 20 μg/mL, P = .005; 100 μg/mL, P = .002. (B) The percentage of neutrophils that phagocytized a target cell as evaluated by confocal microscopy, no G-CSF vs G-CSF: 0 μg/mL 2-hour incubation, P = .99; 20 μg/mL 2-hour incubation, P < .001; 0 μg/mL overnight incubation, P = .90; 20 μg/mL overnight incubation, P = .001. Z-stack analyses were applied to ensure target cells were contained within neutrophils. (C) The geometric mean fluorescent intensity (MFI) of CD11b, CD62L (P = .03), CD66b, CD63 (P = .003), CD64 (P < .001), CD35, and CD16 (P = .03) on neutrophils, with (blue) and without (red) G-CSF treatment. Statistically significant differences (P < .05) are indicated with asterisks.

Figure 4.

Confocal microscopy confirms higher cytotoxicity through ATG-mediated phagocytosis by neutrophils after G-CSF treatment. Neutrophils were obtained from a healthy donor who received 10 mg/kg Filgrastim for 5 days and from a healthy donor who served as control. Neutrophils were incubated with CD14⁻ allogeneic PBMCs in the absence or presence of 20 μg/mL ATG for 2 hours or overnight. Neutrophils were stained with CD64⁻ APC (red), and target cells were CTV+ labeled (blue) for confocal microscopy visualization. Pictures were taken with same Z-resolution (in 1 slice). Original magnification ×65. (A,E) Control neutrophils and target cells without ATG, incubated for 2 hours (A) or overnight (E). (B,F) Neutrophils from G-CSF recipient and target cells without ATG, incubated for 2 hours (B) or overnight (F). (C,G) Control neutrophils and target cells with ATG, incubated for 2 hours (C) or overnight (G). (D,H) Neutrophils from G-CSF recipient and target cells with ATG, incubated for 2 hours (D) or overnight (H). For neutrophils from the G-CSF recipient, shown are target cells within neutrophils. Z-stack analyses were applied to ensure target cells were contained within neutrophils.

Figure 5.

ATG-binding and ATG-mediated cytotoxicity of CB and BM cells. ATG-binding and ATG-mediated cytotoxicity in cells from CB (n = 3) and BM grafts (n = 3). (A) Binding of ATG to graft cells after exposure to 0, 1, 10, 50, and 100 μg/mL rabbit ATG, as evaluated using a FITC-labeled donkey anti-rabbit antibody (total cells, T cells, B cells, NK cells, and monocytes). (B) CDC is evaluated as the percentage of 7-AAD+ immune cells after exposure to 0, 1, 10, 50, and 100 μg/mL rabbit ATG (total cells, T cells, B cells, NK cells, and monocytes). ATG-mediated ADCC is evaluated as the percentage of 7-AAD+ immune cells after exposure to 0, 1, 10, and 100 μg/mL rabbit ATG. (C) NK ADCC (total cells, T cells, B cells, and monocytes). (D) Neutrophil ADCC (total cells, T cells, B cells, and NK cells). (E) Monocyte/macrophage ADCP is evaluated as the amount of remaining target cells/1000 beads after exposure to 0, 1, 10, and 100 μg/mL rabbit ATG (total cells, T cells, B cells, and NK cells). No significant differences were observed. n.a., not applicable.