Dendritic cell vaccination as postremission treatment to prevent or delay relapse in acute myeloid leukemia

Acute myeloid leukemia (AML) still has a dismal prognosis.^1,2  According to the latest data from the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute,¹  5-year overall survival (OS) of patients with AML is only ∼25%. One of the major reasons for this is that a majority of patients relapse even after complete remission (CR) is achieved with standard chemotherapy.³  Relapse is usually caused by the persistence of a small population of residual leukemic cells, a condition designated as minimal residual disease (MRD).⁴  Allogeneic hematopoietic stem-cell transplantation (allo-HSCT), the best established postremission treatment to eradicate MRD, decrease the risk of relapse, and increase survival after chemotherapy, is still beset by substantial morbidity and mortality.⁵  As a consequence, allo-HSCT is generally not considered as a therapeutic option in the large group of older patients with AML. For these patients and for younger patients without compatible donors, there is currently no standard adjuvant treatment to prevent postchemotherapy relapse.⁴ 

The beneficial effect of allo-HSCT against leukemia is mediated in large part by T cells that are capable of recognizing antigens expressed on the leukemia cells and of subsequently mediating AML-cell killing.⁶  Stimulation of autologous T cells by in vivo immunization with leukemia-associated antigens is an innovative strategy to combat relapse in AML,^7-11  acting via the reduction or eradication of MRD. Several antigens have been identified to serve as T-cell targets in AML, including the Wilms’ tumor protein 1 (WT1), which is highly overexpressed in AML and is also involved in leukemogenesis.¹²  In view of their role as the most potent antigen-presenting cells of the immune system, dendritic cells (DCs) are eminently equipped to stimulate antigen-specific T-cell immunity.¹³  This explains the strong interest in the use of these cells for cancer vaccination strategies.¹⁴ 

The aim of this phase 2 study was to determine the clinical efficacy of DC vaccine therapy in AML and, more specifically, whether this form of immunotherapy can be applied in the adjuvant setting to decrease the risk of relapse after chemotherapy and improve survival. To this end, we vaccinated 30 patients with AML in remission but at very high risk of relapse with autologous DCs loaded with the WT1 antigen by means of messenger RNA (mRNA) electroporation, a technique that allows for human leukocyte antigen (HLA) haplotype-independent, multiple-epitope antigen presentation to T cells.^15,16 

Thirty patients with AML were enrolled in this phase 2 study, whereby the first 10 patients were also included in a preceding feasibility, safety, and immunogenicity study (registered at www.clinicaltrials.gov as #NCT00834002).¹⁷  This study was approved by the ethics committee of the Antwerp University Hospital, Edegem, Antwerp, Belgium. Enrollment criteria for the phase 2 study were adult patients with AML (except acute promyelocytic leukemia), diagnosed according to World Health Organization criteria, in remission after ≥1 course of polychemotherapy and at high risk of relapse as defined by: age >60 years or if <60 years without matched sibling donor for allo-HSCT, poor-risk cytogenetic or molecular markers, hyperleukocytosis at presentation, and/or previous relapse.

Clinical grade WT1 mRNA-electroporated DC (WT1/DC) vaccines were prepared and administered intradermally as described previously.^16,17  Three different WT1 constructs were used to generate mRNA by in vitro transcription: construct 1 (WT1), encoding full-length WT1^16,17 ; construct 2 (WT1-DC-LAMP), incorporating a Sig-DC-LAMP major histocompatibility complex (MHC) class 2–skewing signal with deletion of the WT1 nuclear localization signal; and construct 3 (WT1-DC-LAMP-OPT), a codon-optimized version of construct 2 (Figure 1; supplemental Table 1, available on the Blood Web site).¹⁸ 

Longitudinal monitoring of WT1 transcripts was performed as described previously using an in-house assay,^17,19  the Ipsogen WT1 ProfileQuant Kit (Qiagen)²⁰  or the WT1 mRNA OneStep Assay (Otsuka Pharmaceutical Co).¹⁹ WT1 mRNA levels above background (respectively >1 and 25 copies of WT1 mRNA per 1000 ABL copies in blood and marrow in the in-house assay; according to manufacturers’ instructions in the commercial kits) are indicative of MRD and herald full relapse.^21-25  Responders were characterized by MR or SD. MR was defined by normalization of WT1 mRNA levels in blood and/or bone marrow during WT1/DC vaccination, while reaching or maintaining hematological CR. SD was defined by stable WT1 blood transcript levels above background, with stable blood values without blasts. The minimum duration of SD was 2 months, after which WT1 mRNA levels did not increase by more than a factor of 0.5 log₁₀ and ≥1 basic blood value (hemoglobin, thrombocyte count, and/or absolute neutrophil count) was normal. Patients who relapsed without achieving MR or SD status were categorized as nonresponders. The classical definitions of CR, partial remission (PR), relapse, and survival were used.²⁶  Long-term survivors were defined as surviving for ≥3 years^26,27  and long-term responders as patients with an MR or SD response remaining in CR for ≥3 years after the first dose of WT1/DCs.

Detection and subtyping of anti-WT1 antibodies in pre- and postvaccination plasma samples were performed as described previously.^17,28  The increase in WT1 immunoglobulin G antibody levels after vaccination was determined by subtraction of the corresponding prevaccination values.

Cytokine plasma levels were determined using the T-helper 1 (Th1)/Th2 multiplex immunoassay (Bender MedSystems). Ex vivo flow cytometric analysis of lymphocyte subsets was performed using directly conjugated monoclonal antibodies (BD Biosciences).

Circulating WT1-specific CD8⁺ T cells obtained before vaccination and after the fourth dose of WT1/DCs were stained with peptide-HLA-A*0201 tetramers and quantified as described previously.^17,29  Whenever cells were available, they were also analyzed at different time points afterward until relapse and/or progression.

Two weeks after the fourth WT1/DC vaccination, delayed-type hypersensitivity (DTH) was tested against the complete vaccine (ie, keyhole limpet hemocyanin [KLH]–exposed WT1/DCs, except in patients UPN11, 12, and 13, in whom non–KLH-exposed WT1/DCs were used). This was performed by intradermal injection of 0.5 × 10⁶ WT1/DCs on the back of the patient. Forty-eight hours later, erythema and induration at the injection site were measured and skin punch biopsies were taken for culture of DTH-infiltrating lymphocytes (DILs). These DILs were allowed to expand for 2 to 3 weeks in medium with interleukin-2 (100 IU/mL). Expanded DILs were then harvested and tested for antigen specificity as described previously.¹⁷ 

Statistical calculations and data graphing were carried out using Prism version 5.01 (GraphPad). A P value <.05 was considered statistically significant.

WT1/DC vaccination as a postremission treatment was evaluated in 30 patients with high-risk AML in remission after chemotherapy. Patient characteristics with baseline evaluation and initial treatments are listed in supplemental Table 1. There were 15 men and 15 women, with a median age at diagnosis of 65 years. Before WT1/DC vaccination, 27 patients had achieved CR after chemotherapy, whereas 3 had PR. Six patients had a preceding hematological disorder: myelodysplastic syndrome (n = 5) or myeloproliferative neoplasia (n = 1). The cytogenetic risk group³⁰  was adverse in 1, intermediate in 23, and favorable in 6 patients. All patients were at increased risk of relapse, as indicated by the unfavorable prognostic features listed in supplemental Table 1. In particular, WT1 transcript levels above background postinduction and/or postconsolidation chemotherapy were predictive of relapse^21-25  in 5 of 6, 18 of 23, and 1 of 1 patients at respectively favorable, intermediate, and adverse cytogenetic risk. Of the remaining 6 patients without increased WT1 transcript levels after chemotherapy, only 1 with erythroleukemia in the intermediate-risk group (UPN22) did not relapse, and his response was categorized as undefinable (Figure 1).

There was a demonstrable antileukemic effect of the WT1/DC vaccination only without any concomitant chemotherapy in 13 of 30 patients, corresponding to a clinical response rate of 43%. Of these 13 patients, 9 went into MR, as demonstrated by the normalization of WT1 transcript levels in blood and/or bone marrow (UPN01, 06, 08, 10, 11, 14, 15, 16, and 17). Of these 9 patients, 2 went from PR to CR (UPN08 and 16),¹⁷  and 5 were still in CR (UPN06, 08, 10, 14, and 15) with a median duration of 114.5 months and median follow-up after the first WT1/DC vaccination of 109.4 months. Four patients relapsed after reaching MR, with 3 ultimately dying as a result of AML (UPN01, 11, and 16) and 1 achieving long-term CR after undergoing allo-HSCT (UPN17). In all relapsing patients, recurrence of AML was preceded and accompanied by increased WT1 transcript levels. In the remaining 4 patients (UPN21, 33, 35, and 48), clinical response was characterized by SD, as demonstrated by elevated but stable WT1 transcript levels in blood and stable blood values without blasts. An example of SD is shown in Figure 2. The stable WT1 profile seen during SD contrasts with the steeply rising curve that is usually seen when patients with AML are relapsing.^19,23,25 

Only a minority of patients had a molecular MRD marker other than WT1 mRNA. Overall, there was a corresponding evolution between the WT1 transcript levels and other markers of MRD, such as the fusion transcripts RUNX1-RUNX1T1 [translocation t(8;21)] and CBFB-MYH11 [inversion inv(16)]. For instance, in clinical responder UPN15, the normalization of bone marrow WT1 expression after 4 WT1/DC vaccinations (from 638 to 4 copies per 1000 ABL copies) was paralleled by a decrease in RUNX1-RUNX1T1 transcript level from 1.43 to 0.0495 copies per 1000 ABL copies. In nonresponder UPN28, the increase in already-elevated blood WT1 transcript level (from 24.1 to 1195 copies per 1000 ABL copies) was mirrored by an increase in CBFB-MYH11 level from 10 to 469 copies per 1000 ABL copies.

The clinical response rate was 50% in patients with favorable cytogenetic risk (MR in 3 of 6 patients). Among the 22 treatment response–evaluable patients with intermediate cytogenetic risk, 6 experienced an MR and 4 achieved SD, corresponding to a clinical response rate of 45%. Response to WT1/DC vaccination, grouped according to the WT1 construct used, is indicated in Figure 1 and supplemental Table 1. Of the 29 patients with a definable response or nonresponse, there was an effect in 8 of 16 in the WT1 group, 3 of 5 in the WT1-DC-LAMP group, and 2 of 8 in the WT1-DC-LAMP-OPT group. Of the 3 patients receiving DCs prepared without KLH (UPN11, 12, and 13), 1 achieved MR (UPN11).

Overall, 6 of 30 patients (UPN06, 08, 10, 14, 15, and 22) had not yet relapsed and were still in first CR (CR1), with a median duration of 107.6 months and median follow-up of 101.8 months after the first dose of WT1/DCs. Of the remaining 24 of 30 patients, 1 did not reach CR1 (UPN20), 1 died presumably as a result of a lung adenocarcinoma without morphological evidence of AML relapse (UPN07), 3 had a second or third relapse (UPN02, 16, and 34), and 19 had a first relapse. Four of these 19 patients with AML in first relapse received supportive care (UPN01, 11, 21, and 38), whereas the remaining 15 were treated with salvage therapy (chemotherapy ± allo-HSCT; supplemental Table 1). The CR2 rate in this group was 73.3% (11 of 15 patients: UPN05, 12, 13, 17, 28, 30, 35, 36, 46, 47, and 48); the remaining 4 patients had progressive disease (UPN03, 09, 29, and 33).

Survival data are shown in Figures 1 and 3, Table 1, and supplemental Table 1. Five-year relative OS data compared favorably with those reported by the SEER program of the National Cancer Institute.¹  OS and RFS were significantly higher in responders as compared with nonresponders, irrespective of age category. OS data were not influenced significantly by allo-HSCT, which was carried out in some patients who relapsed during WT1/DC vaccination (data not shown).

Of the patients in CR1 (Table 1) who received WT1/DCs, survival from diagnosis compared favorably with that of patients from the SALR, a comprehensive population-based AML database.³¹  Five-year relative RFS of patients with AML in CR1 treated with WT1/DCs was similar to that of patients in the SALR. It should be noted that the risk of relapse in patients treated with WT1/DCs was most probably higher than that of patients in the SALR, some of whom (overall 32.1%) did not relapse after 5 years (Table 1). In contrast, all but 1 of the patients who were vaccinated with WT1/DCs were predicted to relapse based on increased prevaccination WT1 transcript levels and/or did in fact relapse. RFS of responding patients was markedly better than RFS of patients in the SALR. Of the 25 response-evaluable patients in CR1, 5 had not yet relapsed (UPN06, 08, 10, 14, and 15, all of whom were responders), compatible with a vaccine-induced relapse reduction rate of 25%. The median duration of CR1 in these 5 patients is indicated in “Clinical response.” There were 16 of 30 patients in the long-term survivor category (UPN01, 06, 08, 10, 11, 12, 14, 15, 17, 21, 22, 28, 29, 33, 34, and 35).

Long-term survival was observed in 11 of 13 responders and in 4 of 16 nonresponders; the difference was significant (Fisher’s exact test P = .0025), correlating long-term survival with response to WT1/DCs. Long-term survival was noted in 5 of 6 and 10 of 22 response-evaluable patients at respectively favorable and intermediate cytogenetic risk; the difference between the 2 groups was not significant (P = .20). There were no significant differences in OS outcome between the different WT1 constructs used (Figure 1; WT1 vs WT1-DC-LAMP, P = .48; WT1 vs WT1-DC-LAMP-OPT, P = .44; WT1-DC-LAMP vs WT1-DC-LAMP-OPT, P = .84; WT1 vs WT1-DC-LAMP + WT1-DC-LAMP-OPT, P = .34).

As shown in supplemental Table 1, as of 31 December 2016, 11 of 30 patients were alive in CR, with a median OS from diagnosis of 99.4 months (range, 72.6-125.5 months). Of these 11 patients, 6 were alive in continuing CR1 (UPN06, 08, 10, 14, 15, and 22). Five other surviving patients who relapsed after WT1/DC vaccination were brought back into CR by chemotherapy followed (UPN17 and 28) or not (UPN12, 34, and 35) by allo-HSCT. Remarkably, patients UPN12 and 34 were alive in continuing CR respectively >7 and 4 years after achieving CR2 and CR3 with chemotherapy alone.

Of the 19 patients with AML who were in first relapse after WT1/DC vaccination, 52.6% were alive at 3 years and 36.8% at 5 years from diagnosis. Of the latter patients, 4 achieved CR2 after chemotherapy (UPN12, 17, 28, and 35), and the other 3 had a remarkably long period of CR1 before relapsing (UPN01, 11, and 29; respectively 47, 51.6, and 59.8 months after the start of WT1/DC vaccination).

Immunomonitoring was performed on peripheral blood mononuclear cells, DILs, and plasma samples obtained before and after WT1/DC vaccination. There were no significant changes after vaccination with respect to: frequencies or absolute numbers of circulating lymphocyte subsets (CD4⁺ and CD8⁺ T cells, B cells, and natural killer cells); relative frequencies of naïve, terminally differentiated effector, effector memory, and central memory subsets within the CD4⁺ and CD8⁺ T-cell compartments; relative frequencies of regulatory T-cell subsets, or relative frequency of myeloid-derived suppressor cells. Th1/Th2 cytokine levels and anti-WT1 immunoglobulin G titers in plasma were also unchanged after WT1/DC vaccination.

WT1-specific T-cell responses were evaluated using pHLA-A*0201 tetramer staining and intracellular cytokine assays. Increased (>1.5-fold) frequency of WT1-specific tetramer⁺ CD8⁺ T cells was observed after vaccination in 6 of 12 evaluable (HLA-A*0201⁺) patients. A significant positive correlation (P = .018) was found between long-term clinical response (UPN01, 08, and 35) and increased circulating frequency of polyepitope WT1-specific tetramer⁺ CD8⁺ T cells (supplemental Table 2). Of the 9 HLA-A*0201⁺ nonresponder patients, only UPN30 had an increased number of polyepitope WT1-specific CD8⁺ T cells. The other non–long-term responders showed either no (6 of 9 patients) or monoepitope (2 of 9 patients) WT1-specific CD8⁺ T-cell response to WT1/DC vaccination. WT1-specific tetramer⁺ CD8⁺ T cells were also assessed in 7 HLA-A*0201⁺ patients (UPN16, 17, 30, 34, 35, 38, and 47) at relapse or disease progression; frequency was not lost in any patient upon relapse. In all but 1 patient (UPN16), we observed an increase in frequencies of ≥2 epitopes as compared with the postfourth WT1/DC vaccination sample and in all patients a maintenance and/or increase for the other epitope(s) examined (data not shown).

Because KLH was shown in our previous study¹⁷  to skew T cells toward a Th2 profile, which could be detrimental for a cytotoxic antitumoral response, it was omitted from the preparation of the DCs in 3 patients (UPN11, 12, and 13). In those patients, there was no local immunoreactivity at the site of DC injection, nor was there any DTH reactivity, and it was thus impossible to assess in vivo the quality of the DCs for their capacity to migrate to the lymph nodes or elicit T-cell response. For this reason, it was therefore decided to reincorporate KLH into the DC vaccine preparation in all other patients. All the KLH-exposed patients showed a DTH response; in 13 of these patients, DILs were obtained for immune response assessment. Functional analysis of DILs restimulated with autologous WT1-loaded DCs demonstrated WT1-specific CD8⁺ but not CD4⁺ T-cell responses after vaccination, as shown by significant increases in WT1-specific interferon-γ (IFN-γ) and tumor necrosis factor-α (TNF-α) but not interleukin-5 production. This vaccine-specific CD8⁺ T-cell response was present in the long-term but not in the non–long-term survivor group (Figure 4A-C). Notably, significant WT1-specific bifunctional TNF-α⁺/IFN-γ⁺ CD8⁺ T-cell responses were also detected in the long-term survivor group (Figure 4D). In some long-term responders (UPN14) or survivors (UPN17, 34, and 35), the proportion of IFN-γ⁺ and/or TNF-α⁺ WT1-specific CD8⁺ DILs was very high (range, 5%-50%). IFN-γ⁺ and TNF-α⁺ WT1-specific CD8⁺ but not CD4⁺ DIL responses were significantly higher in patients vaccinated with the DC-LAMP–containing WT1 constructs as compared with those vaccinated with the wild-type WT1 construct (data not shown).

In this phase 2 study, we demonstrated clinical activity of autologous WT1 mRNA-electroporated DC vaccination in patients with AML in remission and showed that this form of cancer vaccine therapy may offer OS benefit that is linked to the induction of WT1-specific CD8⁺ T-cell immunity. The clinical response rate that was obtained in this study (43%) is of considerable interest, with 30% of patients achieving MR and 13% SD. The high rate of patients achieving MR (9 of 30 patients) is an important finding, because these patients were otherwise likely to relapse based on their increased prevaccination WT1 transcript levels.^19,21-25  In the SD response group (4 of 30 patients), WT1/DC vaccination was effective in temporarily halting AML progression (as characterized by stabilization of WT1 transcript levels and stable blood values without blasts), thereby delaying the occurrence of florid relapse. SD is an unexpected response category in the context of AML, where relapses are normally characterized by exponentially increasing WT1 transcript levels at a constant doubling time.¹⁹  Similar observations of disease stabilization have been made in WT1 peptide vaccine trials in AML,^7,8,32  indicating that SD should be included as a separate category in the response assessment of immunotherapeutic interventions in hematological malignancies, as is now routinely the case in the field of solid tumor immunotherapy.³³  Altogether, our data indicate that WT1-targeted DC vaccination can be an effective strategy to prevent or delay relapse in AML, without the toxicity of allo-HSCT. This is an important finding in light of the growing number of elderly patients with AML who are generally not considered candidates for allo-HSCT because of toxicity considerations.⁵  On the basis of these results, WT1/DC vaccination could also have a role among younger patients with AML who do not proceed to allo-HSCT after standard chemotherapy because of refusal or lack of suitable donors.

OS data in our study compared favorably with current and new data from SEER and SALR, respectively. Importantly, our comparatively longer OS was observed not only in younger patients (≤65 years) but also in the poor-prognosis older-age category (>65 years). Long-term OS was seen in both favorable and intermediate cytogenetic risk groups. OS and RFS also correlated with clinical response to WT1/DCs. These observations further validate the use of WT1 transcript levels as a suitable marker for leukemic residual disease and for monitoring the effect of therapy in AML. Increased OS in AML was recently found to correlate with reduced WT1 mRNA levels and WT1-specific CD8⁺ T-cell responses in a cohort of patients with AML after chemotherapy and allo-HSCT.³⁴  The comparatively longer OS observed in patients with AML vaccinated with WT1/DCs is consistent with a meta-analysis indicating that DC vaccine therapy can offer OS benefit in patients with solid malignancies, including melanoma, prostate cancer, glioblastoma multiforme, and renal cell cancer.¹⁴ 

Intriguingly, 5-year relative OS but not RFS in patients vaccinated with WT1/DCs compared favorably with SALR data. Our data suggest that 1 major reason for the OS advantage was the unexpectedly high clinical response rate and long survival in patients who had relapsed after WT1/DC vaccination. For patients with AML in first relapse, the probability of achieving CR2 with salvage treatment (ie, chemotherapy and/or allo-HSCT) has been reported to be 46%,³⁵  whereas the CR2 rate in our study was 73.3%. Likewise, the 5-year OS rate of WT1/DC-vaccinated patients with AML after first relapse was 36.8% in our study, which compares favorably with the 11% 5-year OS rate for patients with AML in first relapse that has been described in the literature.³⁵  These data suggest that WT1/DC vaccination can potentiate the response to subsequent treatment, providing an explanation for how vaccination contributes to prolongation of survival. A similar scenario has been reported for solid tumors, where improved clinical outcomes have been documented in patients who received chemotherapy after the seeming failure of immunotherapy as compared with patients who received chemotherapy alone.^36-44  This outcome may reflect synergism between immunotherapy and chemotherapy, the latter having not only antiproliferative but also immunostimulatory effects.^6,13  The precise mechanisms underlying the anticancer synergy between tumor vaccines and chemotherapy are currently being investigated⁴¹ ; 1 potential mechanism involves the release of cytokines (eg, TNF-α) by vaccine-induced CD8⁺ T cells, which in turn enhances the ability of chemotherapy to induce apoptotic tumor-cell death.⁴⁵ 

In this study, clinical response and survival were found to be correlated with induction of WT1-reactive CD8⁺ T-cell immunity by DC vaccination, providing a mechanistic basis for the antileukemic activity of WT1/DCs. First, we found a correlation between long-term clinical response and increased circulating frequency of polyepitope WT1-specific tetramer⁺ CD8⁺ T cells. Maintenance of or increase in the frequency of WT1-specific CD8⁺ T cells at relapse or progression points toward antigen-driven immune activation associated with increasing exposure to the WT1 antigen at relapse⁴⁶  and with the continuation of the WT1/DC vaccination. This suggests that antigen-specific T-cell numbers may be necessary but not sufficient to ultimately control AML. Second, we found a correlation between WT1-specific IFN-γ⁺ and/or TNF-α⁺ DTH-infiltrating CD8⁺ T lymphocytes and long-term OS. This suggests that CD8⁺ T-cell function is needed for long-term control of AML, at least in the immunotherapy setting. Because DTH was not performed in this study at the time of relapse, we cannot exclude that despite maintenance of or increase in WT1-specific cell numbers, their function may be deficient at that time.

In line with preclinical data,¹⁸  3 different WT1 constructs were used in this study: a native full-length WT1 construct, a WT1 construct incorporating the lysosomal targeting signal of DC-LAMP, and a codon-optimized version of the latter construct. Theoretically, the DC-LAMP–containing constructs would facilitate MHC class 2 antigen presentation and subsequent CD4⁺ T-cell stimulation. Contrary to expectations, no statistical evidence for induction of WT1-specific CD4⁺ T-cell immunity was found in the patients in whom the DC-LAMP–containing constructs were used. This is in contrast with a recently published study in melanoma, which used DCs loaded with melanoma antigen-encoding mRNA linked to a similar DC-LAMP construct.⁴⁷  The apparent lack of CD4⁺ T-cell stimulation in our study may have resulted from the fact that the strongly immunogenic WT1_332-347 MHC class 2 epitope^48,49  is not encoded by the DC-LAMP–containing constructs because they lack the WT1 nuclear localization signal.¹⁸  Nevertheless, even without significantly increased stimulation of WT1-specific CD4⁺ T cells, the DC-LAMP–containing constructs did induce higher frequencies of WT1-specific CD8⁺ T cells as compared with the wild-type WT1 construct, presumably as a consequence of higher cytoplasmic WT1 expression and MHC class 1 epitope presentation.¹⁸  This, however, did not translate into superior clinical response rates or improved survival outcomes.

In summary, WT1-targeted DC vaccination can elicit antileukemia T-cell immunity in patients with AML at very high risk of relapse. The induction of functional WT1-specific CD8⁺ T cells is likely a mechanism to help eliminate residual leukemic cells, decrease the likelihood of AML relapse, and improve survival. Vaccination with WT1/DCs can therefore be considered as a nontoxic, postremission strategy to prevent or delay relapse of AML in the adjuvant setting.

The authors thank Charlotte Quisquater, Sandy Van den Eynde, Ho Wa Lau, Hans De Reu, Daphné Lathouwers, and Pascale De Graef for providing excellent technical assistance.

This study was supported by grants from the Antwerp University Hospital, Kom Op Tegen Kanker (formerly the Flemish League against Cancer), the Belgian Foundation against Cancer (SCIE216-2008), the King Baudouin Foundation (2013-J1810870-100687), the European Hematology Association (clinical research grant), the University of Antwerp (Concerted Research Action GOA-BOF and Methusalem grants), the Research Foundation-Flanders (G.0082.08), the Agency for Innovation by Science and Technology (IWT-TBM grant 80664), the National Cancer Plan (Action 29) of the Belgian government, the HEBA Fund, the Kaushik Bhansali Fund, the Suply Hope Fund for Cancer Immunotherapy, and the Antwerp International School Fund for Clinical Research in Hematology. E.L.S. and N.C. were postdoctoral fellows and S.A. was a research fellow of the Research Foundation–Flanders. S.A. and A.V.D. held a fellowship from the Stichting Emmanuel van der Schueren from Kom Op Tegen Kanker. D.A.P. is a Wellcome Trust Senior Investigator.

Contribution: S.A., A.L.V.d.V., E.L.S., V.F.V.T., G.J., N.C., E.L., A.V.D., L.M., P.J., H.G., Y. Oka, F.F., Y. Oji, H.S., and Z.N.B. designed research; S.A., A.L.V.d.V., E.L.S., V.F.V.T., G.J., N.C., G.N., B.S., A.V.D., K.V., M.-B.M., K.D., R.M., F.F., and Y. Oji performed research; G.J. contributed the data from the Swedish Acute Leukemia Registry; S.A., A.L.V.d.V., G.J., I.V., A.V., A.P.G., W.A.S., M.M.C., and Z.N.B. contributed and/or treated patients; K.V., M.-B.M., E.G., M.L., H.G., D.A.P., K.L., F.F., Y. Oji, and H.S. contributed new and/or critical reagents and/or analytic tools; S.A., A.L.V.d.V., V.F.V.T., G.J., N.C., K.V., D.A.P., K.L., Y. Oka, Y. Oji, H.S., and Z.N.B. analyzed data; S.A., A.L.V.d.V., G.J., V.F.V.T., and Z.N.B. prepared the tables and/or figures; S.A., A.L.V.d.V., E.L.S., S.A., V.F.V.T., G.J., K.V., Y. Oji, and Z.N.B. wrote the final version of the report; and all authors approved the report.

Conflict-of-interest disclosure: V.F.V.T. and Z.N.B. are coinventors of a patent covering the messenger RNA electroporation technique (WO/2003/000907; improved transfection of eukaryotic cells with linear polynucleotides by electroporation). Z.N.B. is a member of the Scientific Advisory Board of ExoCyte Therapeutics. The remaining authors declare no competing financial interests.

Correspondence: Zwi N. Berneman, Division of Hematology, Antwerp University Hospital, Wilrijkstraat 10, 2650 Edegem, Antwerp, Belgium; e-mail: zwi.berneman@uza.be.