Immunomonitoring of Stage IV Relapsed Neuroblastoma Patients Undergoing Haploidentical Hematopoietic Stem Cell Transplantation and Subsequent GD2 (ch14.18/CHO) Antibody Treatment

Abstract

Haploidentical stem cell transplantation (haplo SCT) in Stage IV neuroblastoma relapsed patients has been proven efficacious, while immunotherapy utilizing the anti-GD2 antibody dinutuximab beta has become a standard treatment for neuroblastoma. The combinatorial therapy of haplo SCT and dinutuximab may potentiate the efficacy of the immunotherapy. To gain further understanding of the synergistic effects, functional immunomonitoring was assessed during the clinical trial CH14.18 1021 Antibody and IL2 After haplo SCT in Children with Relapsed Neuroblastoma (NCT02258815). Rapid immune reconstitution of the lymphoid compartment was confirmed, with clinically relevant dinutuximab serum levels found in all patients over the course of treatment. Only one patient developed human anti-chimeric antibodies (HACAs). In-patient monitoring revealed highly functional NK cell posttransplant capable of antibody-dependent cellular cytotoxicity (ADCC). Degranulation of NK cell subsets revealed a significant response increased by dinutuximab. This was irrespective of the KIR receptor-ligand constellation within the NK subsets, defined by the major KIR receptors CD158a, CD158b, and CD158e. Moreover, complement-dependent cytotoxicity (CDC) was shown to be an extremely potent effector-cell independent mechanism of tumor cell lysis, with a clear positive correlation to GD2 expression on the cancer cells as well as to the dinutuximab concentrations. The ex vivo testing of patient-derived effector cells and the sera collected during dinutuximab therapy demonstrated both high functionality of the newly established lymphoid immune compartment and provided confidence that the antibody dosing regimen was sufficient over the duration of the dinutuximab therapy (up to nine cycles in a 9-month period). During the course of the dinutuximab therapy, proinflammatory cytokines and markers (sIL2R, TNFa, IL6, and C reactive protein) were significantly elevated indicating a strong anti-GD2 immune response. No impact of FcGR polymorphism on event-free and overall survival was found. Collectively, this study has shown that in-patient functional immunomonitoring is feasible and valuable in contributing to the understanding of anti-cancer combinatorial treatments such as haplo SCT and antibody immunotherapy.

Keywords: GD2 antibody therapy; antibody-dependent cellular cytotoxicity; complement-dependent cytotoxicity; haploidentical allogeneic stem cell transplantation; immunomonitoring immunotherapy; neuroblastoma.

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2021 Seitz, Flaadt, Mezger, Lang, Michaelis, Katz, Syring, Joechner, Rabsteyn, Siebert, Troschke-Meurer, Zumpe, Lode, Yang, Atar, Mast, Scheuermann, Heubach, Handgretinger, Lang and Schlegel.

Figures

Figure 1

Lymphoid immune reconstitution posttransplant, ch14.18/CHO patient serum levels as well as serum levels of human anti-chimeric antibodies (HACA). (A) Posttransplant lymphoid immune reconstitution of CD3+ T cells (CD4+ and CD8+), CD19+ B cells, and CD56+CD16+ NK cells was assessed by flow cytometry at indicated time points. A total of n = 36 eligible patients were included in the analysis. Absolute cell counts per microliter blood (mean ± SEM) were calculated from flow cytometric frequencies (%) and total lymphoid cells derived from the whole blood counts of patients. In the right panel the mean frequencies [%] of CD3+ and CD56+CD16+ are shown. (B) ch14.18/CHO patient serum levels at indicated time points during cycles 1, 3, 4, and 6 were measured by triple-ELISA as previously described (15). A minimum of one and a maximum of n = 15 in cycle 1, n = 14 in cycle 3, n = 12 in cycle 4, and n = 13 in cycle 6 serum of patients were measured. Basically in all ch14.18/CHO treatment cycles, a continuous relevant serum concentration above 1 µg/ml was sustained, facilitating a strong immunologic anti-tumor effect and consolidation by complement-dependent and cellular-dependent cytotoxicity throughout the course of antibody therapy (>6 months). (C) Neutralizing HACA antibodies were measured by ELISA as previously described (19). Only in one patient (red dots), HACAs were detected at relevant levels, indicated by the red horizontal line, in cycle 3 and cycle 4. No further data for this patient is available since no additional testing was performed in this patient afterwards. In cycle 1 n = 32, cycle 3 n = 37, cycle 4 n = 36, cycle 6 n = 28, cycle 9 n = 14 patients were evaluated. In order to display the 0 values in the logarithmic scale, 0 was substituted by 0.002 below the detection threshold (indicated by the blue horizontal line).

Figure 2

Elevation of proinflammatory cytokines and activation of NK cells by ch14.18/CHO antibody infusion therapy. The infusion of ch14.18/CHO induced robust immune response measured by means of secretion of proinflammatory cytokines (A) IL2, (B) TNFα, (C) IL6, and (D) CrP in serum of patients or lithium heparin plasma as well as by the percentage increase in number of activated NK cells (E, F). Cytokines and NK cells were measured on day 1 prior to the start of antibody infusion and on day 5 of cycles 1 to 9. The maximum level of CrP per cycle is shown for the cycles 1–6 in (D). Every single dot represents an independent single value per cycle and patient used from cycles 1–6 (A–D) but as indicated in (E) cycles 1–3 and (F) cycles 4–6. For the comparison of (A) IL2 n = 232, (B) TNFα n = 137, (C) IL6 n = 207 was available in pair values and for (D) CrP n = 198 single values were used. NK cell immunophenotype was assessed by flow cytometry and was defined as the CD56+CD16+ and CD3− subset of lymphoid cells. The early activation marker CD69 was used to distinguish resting (CD69−) from activated (CD69+) NK cells ncylces 1–3 = 47, ncylces 4–6 = 36. Statistical analysis was done by two-tailed paired t-test. P-values below 0.05 were defined significant. **** = <0.0001.

Figure 3

Comparison of cytokine secretion and activation of NK cells during cycles 1–3 without IL2 administration and cycles 4–6 with low dose subcutaneous IL2 administration. In contrast to cycles 1–3 starting as early as 60 days post haploidentical HSCT, in cycles 4–6 at days 6, 8, and 10 subcutaneous IL2 is administered at 1E06 IU/m2/day. Consequently, we performed a systematic analysis and comparison of cytokine secretion and the activation of NK cells between the cycles 1–3 versus the cycles 4–6. We compared the levels of the proinflammatory cytokines (A) IL2, (B) TNFα, (C) IL6 and (D) CrP in serum of patients or lithium heparin plasma and the percentage increase in number of activated NK cells (E, F). Every single dot represents an independent single value per cycle and patient used from cycles 1–6 (A–D) but as indicated in (E) cycles 1–3 and (F) cycles 4–6. For the comparison of (A) IL2 day 1 ncycles 1–3 = 62/ncycles 4–6 = 46, day 5 ncycles 1–3 = 62/ncycles 4–6 = 41, (B) TNFα day 1 ncycles 1–3 = 67/ncycles 4–6 = 55, day 5 ncycles 1–3 = 63/n cycles 4–6 = 47, (C) IL6 day 1 ncycles 1–3 = 129/ncycles 4–6 = 97, day 5 ncycles 1–3 = 125/ncycles 4–6 = 86 was available in pair values and for (D) CrP n = 99 single values were used. NK cell immunophenotype was assessed by flow cytometry and was defined as the CD56+CD16+ and CD3− subset of lymphoid cells. The early activation marker CD69 was used to distinguish resting (CD69−) from activated (CD69+) NK cells, day 1 and 5 ncylces 1–3 = 47, ncylces 4–6 = 36, day 1 p = 0.78, day 5 p = 0.58. Statistical analysis was done by two-tailed unpaired t-test. P-values below 0.05 were defined significant. ** = <0.01, **** = <0.0001. ns, not significant.

Figure 4

Degranulation of patient NK cells and cytolysis mediated by ch14.18/CHO. Patient PBMCs acquired post haploidentical HSCT were used to test the capacity of degranulation, measuring CD107a (LAMP-1) expression as a strong indicator of NK cell activation and cytolysis targeting the neuroblastoma cell lines LAN-1 and LS in a 6 h flow cytometry based kill assay (A–D). Besides testing the overall degranulation capacity, NK cell subsets distinguished by the major inhibitory killer cell immunoglobulin-like receptors (KIR) CD158a/KIR2DL1, CD159b/KIR2DL2 and CD158e/KIR3DL1 were analyzed for interpopulation differences in CD107a positivity and were evaluated in the context of the KIR receptor ligand model. Additionally, a 2 h-DELFIA-EuTDA release cytotoxicity assay was used to evaluate the cytolysis of tumor cells by patient PBMCs via ADCC and CDC (E, F). The testing conditions (E) comprised I) PBMCs versus tumor, II) PBMCs versus tumor and patient serum (after infusion of ch14.18/CHO), and III) PBMCs versus tumor and 1 µg/ml ch14.18/CHO. Moreover, cytolysis of the neuroblastoma cell line LAN-1 by PBMCs from healthy volunteer donors (HVDs) compared to patient PBMCs was assessed. (F) The comparison included the conditions I) [HVD PBMCs plus HVD serum] versus tumor, II) [HVD PBMCs plus HVD serum plus 1 µg/ml ch14.18/CHO] versus tumor, III) [patient PBMCs plus heat inactivated patient serum] versus tumor, and IV) [patient PBMCs plus patient serum] versus tumor. Data shown in (A–D) represent mean of (n = 12) independent experiments and different donors in triplicates, respectively. Data shown in (E) represent mean of (nLAN-1 = 15; nLS = 13) and (F) represent single values of (nHVD = 5; npatients = 10) independent experiments and different donors in triplicates, respectively. Statistical significance was determined by one-way ANOVA and Tukey post-hoc test. P-values below 0.05 were defined significant. * = <0.05, ** = <0.01, *** = <0.0001, **** = <0.0001. ns, not significant.

Figure 5

GD2 expression and cytolysis of neuroblastoma cell lines via ADCC and CDC by PBMCs and serum of patients. (A) GD2 expression of the neuroblastoma cell lines LAN-1, LS, SK-N-AS, and SH-SY5Y was measured by flow cytometry using primary labeled GD2 PE (clone 14G2a) mAb and a mouse IgG2a, k PE mAb as isotype control. The median fluorescence intensity (MFI) was used to calculate the MFI ratio (MFIR) by the MFI GD2 PE mAb divided by MFI isotype control IgG2a, k PE mAb. Representative univariate histograms are shown. From left to right panels the relative GD2 expression declines MFIR LAN-1 = 112.9 > MFIR LS = 46.8 > MFIR SK-N-AS = 1.3 > MFIR SH-SY5Y = 1. (B) Patient PBMCs acquired post haploidentical HSCT were used to test the direct cellular cytotoxicity and ADCC at an effector to target ratio E:T 5:1 of PBMCs versus the indicated neuroblastoma cell line in a 24 h luciferase-based kill assay. Specific lysis was assessed in the conditions I) PBMCs without mAb versus tumor, II) PBMCs plus patient serum versus tumor, III) PBMCs plus 1 µg/ml ch14.18/CHO mAb versus tumor. (C) CDC was tested in a coincubation of human serum from healthy volunteer donors (HVDs) at different levels of concentrations of added ch14.18/CHO mAb or heat-inactivated serum of patients or serum of patients versus the indicated neuroblastoma cell lines in a 24 h luciferase-based kill assay without adding any effector cells. Data shown in (B) represent mean of triplicates of (n = 15) independent experiments; data shown in (C) represent (n = 3) independent experiments of (n = 15) different donors in (B, C), respectively. Statistical significance was determined by one-way ANOVA and Tukey post-hoc test. P-values below 0.05 were defined significant. ADCC, antibody-dependent cellular cytotoxicity; CDC, complement-dependent cytotoxicity; HVD, healthy volunteer donor. * = <0.05, ** = <0.01, **** = <0.0001. ns, not significant.