Tolerability and Clinical Activity of PostTransplantation Azacitidine in Patients Allografted for Acute Myeloid Leukemia Treated on the RICAZA Trial

the majority of patients. Thirty-one patients completed 3 or more cycles of AZA. Sixteen patients relapsed at a median time of 8 months after transplantation. No patient developed extensive chronic graft-versus-host disease. The induction of a post-transplantation CD8 þ T cell response to 1 or more tumor-speci ﬁ c peptides was studied in 28 patients. Induction of a CD8 þ Tcell response was associated with a reduced risk of disease relapse (hazard ratio [HR], .30; 95% con ﬁ dence interval [CI], .10 to .85; P ¼ .02) and improved relapse-free survival (HR, .29; 95% CI, .10 to .83; P ¼ .02) taking into account death as a competing risk. In conclusion, AZA is well tolerated after transplantation and appears to have the capacity to reduce the relapse risk in patients who demonstrate a CD8 þ T cell response to tumor antigens. These observations require con ﬁ rmation in a prospective clinical trial. The maximum age of patients receiving allogeneic hematopoietic stem cell transplantation (alloHCT) has been moving up over time. However, the availability of a suitable HLA-matched sibling donor may limit access of this patient population to alloHCT. We retrospectively investigated the outcomes of umbilical cord blood transplantation (UCBT) after reduced-intensity conditioning regimens in patients aged (cid:4) 70 years with myelodysplastic syndrome (MDS) and acute myelogenous leukemia (AML) between 2010 and 2014. During this period 70 patients with AML/MDS were referred to our center for alloHCT consideration. Twenty-two patients (33%) received alloHCT: 10 UCBT, 9 HLA full-matched sibling donor transplantation, 2 hap- loidentical alloHCT, and 1 unrelated donor alloHCT. In UCBT, cumulative incidences of nonrelapse mortality and relapse were 20% and 30% at 2 years, respectively. The cumulative incidence of acute graft-versus-host disease (GVHD) at day þ 100 and chronic GVHD at 2 years was 10%. Seven patients had viral reactivation/ infections. Rates of overall survival and disease-free survival were 60% and 50% at 2 years, respectively. Moreover, these outcomes seemed to be similar to that of patients aged 60 to 69 years receiving UCBT (n ¼ 60) and patients aged (cid:4) 70 years receiving HLA full-matched sibling donor transplantation (n ¼ 9). These results suggest that UCBT is feasible in selected AML/MDS patients aged (cid:4) 70 years. In fact, UCBT shortens the required time for an unrelated donor search and thus increases the chance of proceeding with alloHCT, which might contribute to higher rates of alloHCT in the referral group. Outcomes of UCBT are promising; however, larger studies with a longer follow-up are needed. (cid:1) 2016 American Society for Blood and Marrow Transplantation.

Allogeneic stem cell transplantation (SCT) represents an increasingly important curative option in adults with highrisk acute myeloid leukemia (AML) [1]. To a large degree, this reflects the improved tolerability of reduced-intensity conditioning (RIC) regimens, which have permitted the extension of a potentially curative graft-versus-leukemia (GVL) effect to patients up to their eighth decade [2][3][4]. Although the increased availability of allogeneic transplantation represents a major advance in the treatment of older adults with AML, its curative potential remains limited by both disease relapse and graft-versus-host disease (GVHD), which now represent the 2 major causes of treatment failure [5]. Thirty percent to 80% of patients allografted for AML will relapse; the great majority in the first 12 months after transplantation, and the outcome for this patient population remains extremely poor [6]. The administration of cellular or pharmacological therapies with the ability to either augment a GVL effect or deliver direct antitumor activity after transplantation represents a promising strategy to reduce the risk of disease relapse. However, because the preponderance of patients allografted for AML relapse within the first year after transplantation, any intervention aimed at reducing the risk of relapse must be delivered early. Although donor lymphocyte infusions (DLI) have the potential to augment a GVL effect, they are associated with a significant risk of severe GVHD, particularly when administered early after transplantation, which complicates their routine utilization [7,8]. In contrast, pharmacological agents can, in principle, be administered early after transplantation, and consequently are the subject of increasing scrutiny as a strategy to reduce the risk of disease relapse after transplantation [9][10][11].
Azacitidine (AZA) is a DNA methyltransferase inhibitor that demonstrates significant clinical activity in AML and myelodysplasia [12,13] and can also be an effective salvage therapy in patients who relapse after allogeneic transplantation [14,15]. It has been postulated that the antileukemic activity of AZA after transplantation may, in part at least, be a consequence of the upregulation of epigenetically silenced minor histocompatibility and tumor antigens on leukemic blasts, resulting in an augmented GVL response [16][17][18]. This is supported by the observation that AZA has the capacity to induce a CD8 þ T cell response to a range of tumor antigens in patients with AML both before and after transplantation [16,17]. Separately, AZA has been shown to augment the reconstitution of T regulatory cells (Tregs) in the immediate post-transplantation period and, in mouse models, its administration reduces the risk of severe GVHD [17,[19][20][21]. Taken together, these data raise the possibility that post-transplantation AZA may either deliver a direct antitumor effect or epigenetically manipulate the allo-immune response, augmenting a GVL effect without increasing the risk of GVHD [22]. To date, however, there has been no systematic study of the clinical activity of adjunctive post-transplantation AZA when administered as a maintenance strategy. Two small studies have identified the maximum tolerated dose of AZA after transplantation. None to date have studied the impact of sustained administration of hypomethylating agents on clinical outcome [10,11]. We, therefore, now report clinical outcomes in a cohort of patients with AML who underwent transplantation using a standardized RIC regimen incorporating post-transplantation AZA in the RICAZA study.

Patient Inclusion
Adult patients with AML whose outcome with conventional chemotherapy was predicted to be poor, in whom a matched related donor or matched unrelated donor had been identified, were eligible for this trial. The clinical trial protocol was approved by the research ethics committee and all patients gave informed consent in accordance with the Declaration of Helsinki. The trial was registered at http://isrctn.org as #ISRCTN36825171. The primary endpoint of the study was to assess the safety and tolerability of AZA in patients after RIC allogeneic transplantation for AML. The impact of AZA administration on biological parameters in the first 27 patients treated on the RICAZA study have been reported in a previous publication, but with a median follow-up of 7 months, it was not possible to comment on the impact of AZA administration on clinical outcomes [17]. In this paper, we report an updated analysis of the 37 patients treated on the RICAZA study with a median follow-up of 24 months (range, 6 to 28 months).

Transplantation Regimen
Patients who underwent allogeneic SCT using either a conditioning regimen consisting of fludarabine (30 mg/m 2 intravenously for 5 days), melphalan (140 mg/m 2 intravenously), and alemtuzumab (10 mg intravenously for 5 days) (n ¼ 34) or a FLAMSA regimen (n ¼ 3) were eligible for this trial [23]. GVHD prophylaxis consisted of cyclosporine commencing day À1 at an adjusted dose to achieve therapeutic levels between 100 mg/L and 200 mg/L after transplantation, with the aim of tapering immunosuppression in patients with no evidence of active GVHD between day 60 and day 90 after transplantation. Patients were not eligible to receive DLI in the first year after transplantation.

AZA Schedule
Patients with stable engraftment (neutrophil count: >1 Â 10 9 /L; platelet count: >50 Â 10 9 /L) commenced treatment with AZA on day þ42 after transplantation at a dose of 36 mg/m 2 subcutaneously for 5 days. This was administered every 28 days up to 12 months after transplantation. Toxicity was graded against the National Cancer Institute's Common Terminology Criteria for Adverse Events. The dose of AZA was reduced to 24 mg/m 2 in patients experiencing grade 3 or 4 hematologic toxicity that persisted for >2 weeks. Initial data concerning the tolerability of post-transplantation AZA in a cohort of 27 patients, of whom only 17 had received more than 6 courses of post-transplantation AZA, were presented in a previous publication [17] and we now report the final toxicity and clinical outcome data in the complete trial cohort of 37 patients. The presence of active acute GVHD or a history of GVHD were not exclusion criteria to trial entry. Clinical responses and transplantation outcome were assessed every 3 months by sequential bone marrow aspirates and peripheral blood lineage specific chimerism analyses, as previously described, until 12 months after transplantation.

Quantification of Circulating Tumor-specific CD8 þ T Cells
The number of circulating tumor-specific cytotoxic T lymphocytes was measured in peripheral blood mononuclear cells prepared from 50 mL of fresh peripheral blood using a CD137 expression and enrichment assay (Miltenyi Biotec, Bergisch Gladbach, Germany). The procedure was conducted as previously described with some minor adjustments [17,24,25]. The HLA type of each patient was known and peptides that matched the HLA type were chosen from the list of previously described tumor-associated antigens [17]. The frequency of CD137 þ antigen-specific T cells was calculated as a percentage of the total CD8 þ T cell pool from the pre-enrichment estimates. The postenrichment analysis was used to validate the results obtained in pre-enrichment samples.

Statistical Analysis
Safety and tolerability outcomes were defined in terms of hematological and nonhematological toxicities. Studied outcomes included rates of complete remission (CR), relapse-free survival (RFS), and overall survival (OS). RFS was defined as the time from transplantation to relapse or death, censoring alive patients at date last seen. OS was defined as time from transplantation to death, censoring alive patients at date last seen. The sample size was calculated using A'Herns single stage design and was based on the primary outcome measure of tolerability. A tolerability rate of 50% or less was deemed to be unacceptable and the probability of obtaining a false positive result was set at 5%. A tolerability rate of 70% was deemed to be an acceptable figure and the probability of a false negative result (ie, incorrectly rejecting for further study a treatment with a true tolerability rate of >70%) was set at 10%. The analysis reported is based on the per-protocol population, including all patients who received the protocol-defined RIC regimen and commenced AZA after transplantation. Statistical analyses were performed using STATA 12 and R version 3.1.

Patient Demographics
Fifty-one patients were registered for treatment on the RICAZA trial and underwent allogeneic transplantation. Fourteen patients did not commence AZA therapy because of post-transplantation complications, including infection (n ¼ 8), patient withdrawal of consent or ineligibility (n ¼ 5), or acute GVHD (n ¼ 1). Thirty-seven patients commenced monthly courses of AZA at a median time of 54 days after transplantation (range, 40 to 194 days) and are the subject of this report. The median follow-up for alive patients was 24 months (range, 6 to 28 months). The median age of the 37 patients who commenced AZA was 60 years (range, 40 to 71 years) ( Table 1). Twenty-four patients (65%) were in CR1, 8 patients (22%) were in CR2, 3 patients (8%) were in first relapse, and 2 patients (5%) had primary refractory disease (Table 1). Thirteen (35%) patients underwent transplantation using a matched related donor and 24 (65%) had an adult volunteer unrelated donor. Thirty-four patients received granulocyte colonyestimulating factoremobilized peripheral blood stem cells and 3 had bone marrow as the stem cell source. All patients engrafted with a median time to neutrophil engraftment of 13 days (range, 1 to 22 days) and a median time to platelet engraftment of 13 days (range, 10 to 33 days).

Tolerability of Post-transplantation AZA
AZA was well tolerated in the majority of patients. Hematological and nonhematological toxicities experienced by 10% or more of patients are described in Table 2. Four patients experienced treatment delays due to neutropenia or thrombocytopenia. The most common nonhematological toxicities observed were abnormalities of liver function, injection site reaction, nausea, and infection. Thirty-one patients completed at least 3 cycles of AZA and 16 patients completed 10 cycles. Twenty patients discontinued AZA before 12 months after transplantation because of disease relapse (n ¼ 10), infection or hematological toxicity (n ¼ 6), or miscellaneous reasons (eg, physician decision to administer DLI, withdrawal of consent, and protocol deviation) (n ¼ 4).

Chimerism, GVHD, Relapse, and Outcome
At day þ90 after transplantation, 22 (59%) patients demonstrated full donor chimerism in whole blood, of whom 7 (19%) demonstrated full donor chimerism in the T cell fraction. Serial chimerism studies are available on 14 patients who received AZA after transplantation, which demonstrate broad stability of T cell chimerism with no significant changes observed over time. Grade 1 or 2 acute GVHD was documented in 17 patients. No patient developed severe (grades 3 or 4) acute GVHD during the period of treatment on the RICAZA trial, although 1 patient developed severe GVHD after being withdrawn and receiving DLI for the treatment of disease relapse. Ten patients developed limited chronic GVHD but no patient developed extensive chronic GVHD.

Correlation of Peripheral Blood Antitumor CD8-specific T Cell Response with Clinical Outcome
Twenty-eight patients treated with AZA were screened for antitumor CD8 þ T cell responses to a range of tumor antigens. AZA induced a CD8 þ T cell response to 1 or more tumor-specific peptides in 16 patients. A summary of their characteristics by CD8 T cell response is displayed in Table 3. As previously reported, no CD8 þ T cell response was detected in 7 patients allografted using a similar conditioning regimen who did not receive AZA after transplantation [17]. The frequency of CD8 þ T cells directed against the studied tumor antigens ranged from .01% to 1.6% (mean, .45%) of circulating CD8 þ T cells, which is comparable to the frequency of cytomegalovirus-specific T cell responses detectable in cytomegalovirus-seropositive patients. Induction of a CD8 þ anti-tumor response was associated with a reduction in relapse risk taking into account death as a competing risk (hazard ratio [HR], .30; 95% CI, .10 to .85; P ¼ .02) and  improved RFS (HR, .29; 95% CI, .10 to .83; P ¼ .02) (Figure 1). The beneficial impact of detection of a CD8 þ T cell response to candidate tumor antigens on relapse risk and RFS is still observed, with adjustment for time to start of AZA therapy (HR, .30; 95% CI, .1 to .94; P ¼ .039) or if restricted to only 24 patients who commenced AZA before day þ90 after transplantation (HR, .32; 95% CI, .10 to .95; P ¼ .041) in the competing risk analysis. The impact of detection of a CD8 þ T cell response on relapse risk remains strong with adjustment of donor type (HR, .26; 95% CI, .09 to .70; P ¼ .008) and with additional adjustment of start time to AZA therapy after transplantation (HR, .27; 95% CI, .09 to .82; P ¼ .022). There was no correlation between the timing or magnitude of the CD8 þ tumor response and relapse risk.

DISCUSSION
The ability of epigenetic therapies to upregulate tumor antigens represents a novel strategy by which a GVL response may be augmented. In this small prospective trial, we observed a reduced risk of disease relapse in patients treated with post-transplantation AZA who developed a CD8 þ T cell response to a range of candidate tumor antigens. It is important to emphasize the importance of repeating this observation in a larger randomized study, but these data are supportive of further investigation of the impact of other post-transplantation epigenetic therapies, such as DNA methyltransferase inhibitor and histone deacetylase inhibitors, on transplantation outcome. Of interest, the dose of AZA observed to induce a CD8 þ T cell response in this study is approximately one-half that utilized in the treatment of patients with de novo AML or myelodysplasia, consistent with the hypothesis that the observed reduction in relapse is consequent upon manipulation of the alloreactive response. Our study appears to refute the possibility that AZA might improve/outcome by simply postponing relapse, although a larger study will clearly be required to conclusively address this possibility. Because the great majority of patients destined to relapse after an allogeneic transplantation for AML will do so within the first year, a potential advantage of AZA administration, as opposed to DLI, is the ability to commence treatment early [6]. The majority of patients in this study were not only able to commence AZA within 3 months after transplantation but also complete the scheduled course of 1-year treatment. Nonetheless, approximately one third of the patients registered to this study did not receive AZA, emphasizing the potential limitations of posttransplantation interventions as a strategy to reduce the risk of disease relapse. Alternative approaches to selectively augment a GVL effect include vaccination to tumor antigens, such as WT1, and it would be of interest to combine such a strategy with AZA, as has been done for DLI [26,27]. Furthermore, although in this study we measured T cell responses to a broad range of candidate tumor antigens, future  studies should be aimed at correlating clinical outcome with the induction of immune responses to specific tumor antigens.
A number of studies have demonstrated the ability of AZA to accelerate the reconstitution of Tregs after allogeneic SCT [17,21], and in murine models this has been correlated with a reduction in the incidence of GVHD [19]. Although the observation that no patient who received posttransplantation AZA had extensive chronic GVHD is striking, particularly given the fact that most patients underwent transplantation using an unrelated donor, it is important to note that the RIC regimen utilized included alemtuzumab, which is known to reduce the incidence of both acute and chronic GVHD [28,29]. Nonetheless, these data support further examination of the impact of post-transplantation AZA on the incidence of acute or chronic GVHD. Reassuringly, our data suggest that the observed induction of Tregs and attendant absence of extensive chronic GVHD do not appear to be associated with an increased risk of disease relapse [22]. This is the first study to examine the impact of the posttransplantation administration of AZA on transplantation outcome. The reduced relapse rate in patients who demonstrate a CD8 þ T cell response directed against tumor antigens, coupled with the absence of chronic extensive GVHD, is consistent with the hypothesis that AZA has the capacity to epigenetically manipulate the alloreactive effect after transplantation. These observations require confirmation in a prospective randomized trial.

ACKNOWLEDGMENTS
Financial disclosure: This study was supported by Leukaemia and Lymphoma Research. Adjunctive biological studies and research nurse support were funded by the CRUK ECMC programme and Cure Leukaemia. Azacitidine was provided free of charge by Celgene Ltd. P.V. is funded by the Oxford Biomedical Research Centre under the NIHR Biomedical Research Centres funding scheme, the Medical Research Council (MRC) Disease Team Award and the MRC Molecular Haematology Unit. The support and time of participating patients and their families is gratefully acknowledged. Clinical trial registered at http://isrctn.org as #ISRCTN36825171.
Conflict of interest statement: C. Craddock and P.V. received honoraria and research support from Celgene. The remaining authors have nothing to report.
Authorship statement: C. Craddock designed the trial, analyzed the data, wrote the paper, and takes primary responsibility for the paper. N.Y.J., S.S., C.Y., J.N.K., J.W., P.F., P.H., and R.B. collected and analyzed data. S.N., P.V., E.T., C. Crawley, N.R., J.B., R.M., J.S., and M.D. recruited patients. O.G. performed the laboratory studies for this paper. All investigators contributed to the writing of the manuscript and have reviewed the submitted manuscript.