ALBERT

Description: Introduction: Myeloid-derived suppressor cells (MDSC) are immunosuppressive immature cells of the myeloid lineage that play a role in cancer induction, progression and immune evasion. Two major subtypes of MDSC can be distinguished: the polymorphonuclear (PMN-MDSC) and monocytic (M-MDSC). Immune reconstitution after autologous hematopoietic stem cell transplantation (ASCT) may play a role in the therapeutic effect of ASCT in multiple myeloma (MM) and lymphoma patients. Since MDSC are immunosuppressive cells, we postulate they might play a role in the immune reconstitution following ASCT. Here, we studied the effect of G-CSF mobilization on MDSC accumulation and the kinetics of MDSC during G-CSF mobilization and ASCT in MM and lymphoma patients. Methods: Flow cytometry was used to measure MDSC in the stem cell graft (SCG) and peripheral blood mononuclear cells (PBMC) of 9 MM and 4 aggressive lymphoma patients. PMN-MDSC were defined as CD11b+ CD33dim HLA-DR- CD15+ cells and M-MDSC as CD11b+ CD33+ HLA-DRlow/- CD14+ cells. MDSC were measured at G-CSF start, time of stem cell (SC) collection and following ASCT at predefined time points (day 0, day +14, day +28, day +100, day +360). Values are reported as mean percentage of MDSC of living cells. For SC mobilization, MM patients received high-dose cyclophosphamide and G-CSF, lymphoma patients received R-DHAP and G-CSF. 3/9 MM patients and 4/4 lymphoma patients had a complete response at the time of SC collection. As conditioning regimen MM patients received melphalan, lymphoma patients received BEAM. Results: In PBMC of both MM and lymphoma patients, G-CSF mobilization increased PMN-MDSC levels: 10.06% at start G-CSF vs 46.44% at SC collection in MM patients (p=0.0006); 19.65% at start G-CSF vs 50.74% at SC collection in lymphoma patients (p=0.1143). After ASCT, PMN-MDSC levels tended to decrease: 28.21% at day 0 vs 7.94% at day +180 in MM patients (p=0.1061); 40.35% at day 0 vs 2.32% at day +180 (p=0.2000) in lymphoma patients. This in contrast to M-MDSC levels, which were not influenced by the use of G-CSF: 5.77% at start G-CSF vs 3.67% at SC collection in MM patients (p=0.3704); 6.52% at start G-CSF vs 3.23% at SC collection in lymphoma patients (p=0.3429). M-MDSC levels remained rather stable during the transplant period, but tended to increase after ASCT: 3.80% at day 0 vs 8.81% at day +180 in MM patients (p=0.0417); 1.24% at day 0 vs 16.70% at day +180 in lymphoma patients (p=0.3333). When comparing MDSC levels in PBMC and the SCG, we observed a decrease in the PMN-MDSC/M-MDSC ratio in the SCG in MM patients (p=0.0078) and in lymphoma patients (p=0.1250). In the SCG, fewer PMN-MDSC and more M-MDSC were present as compared to PBMC. Interestingly, the expression of IL-4Rα on PMN-MDSC (p=0.0078) and M-MDSC (p=0.0078) in PBMC was increased after G-CSF, at time of SC collection. Conclusion: The kinetics of both MDSC subtypes were comparable between MM and lymphoma patients. G-CSF mobilization resulted in the accumulation of PMN-MDSC, but not of M-MDSC. In addition, the expression of IL-4Rα increased on both MDSC subtypes after G-CSF, indicating a potential higher immunosuppressive capacity of MDSC. In both MM and lymphoma patients, PMN-MDSC levels decreased after ASCT, whereas M-MDSC levels increased after ASCT. This might be due to the natural development of these cells. However, further research, including the effect of MDSC on immune reconstitution and the effect of G-CSF on the immunosuppressive capacity of MDSC, is warranted. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 4102 Donor leucocyte infusion (DLI) after alloHSCT can induce strong graft-versus-leukemia (GvL) effects, inspiring investigators to examine this approach for solid tumors resistant to conventional therapies (Grivas et al. Curr Clin Pharmacol. 2011). However, DLI produces graft-versus-host disease (GvHD). Recipient-type leucocyte infusion (RLI) is currently being explored clinically as a means to induce GvL without risk of GvHD (clinicaltrials.gov; Rubio et al. Blood 2003; De Somer et al. Haematologica. 2011). High-risk neuroblastoma carries a bleak prognosis despite aggressive treatment with chemo-, radiotherapy and autologous HSCT. Clinical observations in such patients suggest that alloHSCT may produce a graft-vs-neuroblastoma effect (Kanold et al. Bone Marrow Transplantation. 2008). In mice, alloHSCT delays local neuroblastoma growth, and adoptive transfer of tumor-pulsed donor dendritic cells and donor leucocytes enhances this effect (Ash et al. Cancer Immunol Immunother. 2009, Br J Cancer 2010). In this study, we show that not only DLI by itself, but also RLI enhances the local anti-neuroblastoma effect of alloHSCT in mice. MHC-mismatched [C57BL/6 (H-2Kb) → A/J (H-2Kk)] bone marrow chimeras were given a subcutaneous inoculation with 1 × 106 Neuro2A cells (A/J neuroblastoma) on day 14 post HSCT. On day 21, we performed adoptive transfer of 10 × 106 donor splenocytes (DLI), 50 × 106 recipient splencoytes (RLI) and/or 1 × 106 recipient-type MACS-isolated DX5+ NK cells. We measured tumor volume twice weekly using a caliper (volume = width2 × length × 0,4). Validation of this model showed progressive tumor growth and mortality as a result of metastasis. Peripheral blood chimerism and tumor infiltrating lymphocytes were studied using flow cytometry. AlloHSCT chimeras developed mixed donor T cell chimerism by day 21 and full donor chimerism by day 76 post HSCT. DLI induced a conversion to full donor T cell chimerism and RLI induced a complete loss of donor T cells chimerism, both within 1 week. AlloHSCT chimeras showed reduced local growth of subcutaneous neuroblastoma tumors relative to synHSCT chimeras. This delay in tumor growth was enhanced not only by DLI, but also by RLI; DLI provoked lethal GvHD whereas mice treated with RLI remained healthy. Within tumor-infiltrating lymphocytes, T and NK cell chimerism mirrored the systemic chimerism changes seen after RLI and DLI, associated with an increased intratumoral CD8/CD4 ratio, CD8+ T-cell IFN-γ-expression and NK-cell Granzyme B-expression. This indicates a close relation between lymphohematopoietic alloreactivity and the anti-tumor effect, and suggests that the anti-tumor mechanism of DLI and RLI involves not only CD8+ T cells but also cytotoxic NK cells. The baseline antitumor effect seen in alloHSCT chimeras was also accompanied by increased Granzyme B expression by intratumoral NK-cells, supporting a role for NK cells also in this baseline antitumor effect. In vivo Neuro2A-inoculation experiments in (poly(I:C)-treated) C57BL/6, TCR−/− C57BL/6 and A/J mice, and in vitro NK cytotoxicity experiments showed that 1° donor T cells critically resist Neuro2A cells, 2° donor NK cells exhibit spontaneous cytotoxic reactivity to Neuro2A that is enhanced by NK activation, but also that 3° syngeneic NK cells may acquire reactivity to Neuro2A cells provided they are activated. Interestingly, when RLI was given, the intratumoral NK-cell frequency declined markedly, and adoptive transfer of additional NK cells obtained from naïve A/J mice enhanced the local antitumor effect of RLI. This supports the hypothesis that, in addition to donor NK cells, also syngeneic NK cells can play a critical role in mediating a growth-limiting effect on local neuroblastoma. Along this line, we observed that also in synHSCT mice, which showed a reduced local neuroblastoma growth relative to naïve mice, the intratumoral NK cells showed increased FasL-expression. These are the first experimental data showing that RLI after alloHSCT may induce immune-mediated anti-neuroblastoma effects, and that NK cells, in particular syngeneic NK cells, may play a critical role herein. Our data support the exploration of post alloHSCT adoptive cell therapy with recipient-type T and NK cells to produce enhanced immune antitumor effects for high-risk neuroblastoma, and potentially for other solid tumors that are unresponsive to traditional therapies, without the risk of GvHD. Disclosures: No relevant conflicts of interest to declare.

Description: A murine model of minor histocompatibility antigen (miHCag)–mismatched bone marrow transplantation (BMT) was used to study the development of immunoregulatory cells in the posttransplantation period and their possible involvement in the dissociated graft-versus-host (GVH) and graft-versus-leukemia (GVL) reactivity of posttransplantation donor lymphocyte infusions (DLIs). DLI, applied immediately after BMT, induced GVH disease (GVHD), but when DLI was delayed for 3 weeks, GVHD was avoided while a distinct GVL response was allowed to develop. A population of Mac1+Ly6-G+Ly6-C+ immature myeloid cells, found in small numbers in normal mice, strongly expanded in spleens of chimeras, reaching a maximum level at week 3 and returning to base level by week 12. Upon isolation, these cells exhibited interferon-γ (IFN-γ)–dependent, nitric oxide (NO)–mediated suppressor activity toward in vitro alloresponses, suggesting that, after in vivo DLI, they are activated by IFN-γ to produce NO and suppress GVH reactivity. Because not only alloactivated T-cell proliferation but also leukemia cell growth was found susceptible to inhibition by exogenous NO, in vivo activation of these cells after DLI may explain the occurrence of a GVL effect despite suppression of GVHD. This suggested sequence of events was supported by the finding that the ex vivo antihost proliferative response of spleen cells, recovered shortly after in vivo DLI, was characterized by strong mRNA production of the monokines interleukin-1 (IL-1), IL-6, and tumor necrosis factor-α (TNF-α) and of inducible nitric oxide synthase (iNOS). Our data suggest that transiently expanding Mac1+Ly6-G+Ly6-C+ immature myeloid cells (probably as a result of extramedullary myelopoiesis) may play a role in controlling GVH while promoting GVL reactivity of DLI after allogeneic BMT.

Description: A murine model of minor histocompatibility antigen-mismatched bone marrow transplantation (BMT) was used to study the role of timing of donor lymphocyte infusion (DLI) in eliciting graft-versus-host (GVH) and graft-versus-leukemia (GVL) reactivity. We gave DLI at weeks 3 and 12 after BMT and related its ability to induce a GVL effect with (1) evolution of T cell chimeric status and (2) the extent to which DLI could elicit lymphohematopoietic GVH (LHGVH) reactivity. All mice remained free of GVH disease, but only week 3 DLI chimeras exhibited a significant GVL response when challenged with host-type leukemia cells. In these week 3 DLI chimeras, host-reactive T cells were found to proliferate in vivo (5- [and-6]-carboxyfluorescein diacetate, succinimidyl esther [CFSE]-labeled DLI inocula, TCR-Vβ6+ T-cell frequency) and T-cell chimerism rapidly converted from mixed into complete donor type, indicating the occurrence of LHGVH reactivity. In week 12 chimeras, DLI elicited none of the activities noted at week 3. Yet, in both instances, splenocytes, recovered following DLI, generated an equally strong antihost proliferative response in a mixed lymphocyte reaction, thereby arguing against a decisive role of regulatory cells. The lack of in vivo LHGVH reactivity after week 12 DLI was associated with a substantially increased level of pre-existing host-type T-cell chimerism. We conclude that elicitation of a GVL effect may require LHGVH reactivity and that the reason why timing of DLI was critical for obtaining LHGVH reactivity and the desired GVL effect may lie in the evolution of chimeric status. A possible direct involvement of residual host-type antigen-presenting cells in eliciting LHGVH reactivity after DLI should be studied using models that allow chimerism analysis in non–T-cell lineages.

Description: Xenoantibody production directed at a wide variety of T lymphocyte–dependent and T lymphocyte–independent xenoantigens remains the major immunologic obstacle for successful xenotransplantation. The B lymphocyte subpopulations and their helper factors, involved in T-cell–independent xenoantibody production are only partially understood, and their identification will contribute to the clinical applicability of xenotransplantation. Here we show, using models involving T-cell–deficient athymic recipient mice, that rapidly induced, T-cell–independent xenoantibody production is mediated by marginal zone B lymphocytes and requires help from natural killer (NK) cells. This collaboration neither required NK-cell–mediated IFN-γ production, nor NK-cell–mediated cytolytic killing of xenogeneic target cells. The T-cell–independent IgM xenoantibody response could be partially suppressed by CD40L blockade.