Azacitidine and Donor Lymphocyte Infusions As First Salvage Therapy For

Home , Azacitidine

ORIGINAL ARTICLE Azacitidine and donor lymphocyte infusions as ﬁrst salvage therapy for relapse of AML or MDS after allogeneic stem cell transplantation

T Schroeder1,8, A Czibere1,8, U Platzbecker2, G Bug3, L Uharek4, T Luft5, A Giagounidis6, F Zohren1, I Bruns1, C Wolschke7, K Rieger4, R Fenk1, U Germing1, R Haas1, N Kro¨ ger7 and G Kobbe1

The combination of azacitidine and donor lymphocyte infusions (DLI) as ﬁrst salvage therapy for relapse after allogeneic transplantation (allo-HSCT) was studied in 30 patients with acute myeloid leukemia (AML; n ¼ 28) or myelodysplastic syndromes (MDS; n ¼ 2) within a prospective single-arm multicenter phase-II trial. Treatment schedule contained up to eight cycles azacitidine (100 mg/m2/day, days 1–5, every 28 days) followed by DLI (from 1–5 106 to 1–5 108 CD3 þ cells/kg) after every second azacitidine cycle. A median of three courses azacitidine (range 1–8) were administered, and 22 patients (73%) received DLI. Overall response rate was 30%, including seven complete remissions (CRs, 23%) and two partial remissions (7%). Five patients remain in CR for a median of 777 days (range 461–888). Patients with MDS or AML with myelodysplasia-related changes were more likely to respond (P ¼ 0.011), and a lower blast count (P ¼ 0.039) as well as high-risk cytogenetics (P ¼ 0.035) correlated with the likelihood to achieve CR. Incidence of acute and chronic graft-versus-host disease was 37% and 17%, respectively. Neutropenia and thrombocytopenia grade III/IV occurred during 65% and 63% of treatment cycles, while infections were the most common grade III/IV non-hematological toxicity. Azacitidine and DLI as salvage therapy is safe, induces long-term remissions and may become an alternative for patients with AML or MDS relapsing after allo-HSCT.

Leukemia (2013) 27, 1229–1235; doi:10.1038/leu.2013.7 Keywords: AML; MDS; transplantation; relapse; DLI; azacitidine

INTRODUCTION The DNA-methyltransferase inhibitor 5-azacitidine (Aza) is Among patients with acute myeloid leukemia (AML), allogeneic approved for the treatment of patients with high-risk MDS and hematopoietic stem cell transplantation (allo-HSCT) used as post- AML with up to 30% blasts and might provide these properties: remission therapy for those with high-risk molecular profiles or as beyond its direct cytotoxic effects and ability to induce remissions salvage therapy for those with resistant disease offers the highest even in patients refractory to chemotherapy, Aza is considered to potential for long-term survival.1 In addition, allo-HSCT is the only increase the immunogenicity of AML blasts by re-expression of 5,10–14 curative option for patients with myelodysplastic syndromes (MDS).2 important antigens. Furthermore, animal models suggest an Still, relapse after allo-HSCT remains the major cause of treatment immunomodulatory role of Aza that might attenuate GvHD after 15,16 failure and is associated with a dismal prognosis.3,4 Treatment DLI. options for those patients are mainly restricted to chemotherapy, Considering this, we and others have treated small numbers of 17–21 donor lymphocyte infusions (DLI) or second allo-HSCT.5,6 patients with Aza for relapse after allo-HSCT. We here report In patients receiving DLI or second transplantation, remission at the results of a prospective phase-II trial, aiming to study efficacy the time of DLI or second transplantation is a major predictor for and safety of Aza in combination with DLI as first salvage therapy survival.7–9 However, remission induction by chemotherapy is for hematological relapse of AML or MDS after allo-HSCT. challenging because of low response rates and high toxicity.5 Thus, there is need for novel strategies for patients with AML or MDS relapsing after allo-HSCT. Such treatment approach should PATIENTS AND METHODS ideally offer efficient antileukemic activity, acceptable toxicity Study design and mechanisms to direct the donor immune system towards The AZAcytidine and donor lymphocyte infusions for treatment of RELApse an enhanced graft-versus-leukemia effect without provoking of AML or MDS after allo-HSCT (AZARELA) trial was a prospective, EBMT graft-versus-host disease (GvHD). (European Group for Blood and Marrow Transplantation)-labeled,

1Medical Faculty, Department of Hematology, Oncology and Clinical Immunology, University of Duesseldorf, Medical Faculty, Duesseldorf, Germany; 2University Hospital Carl Gustav Carus, Medizinische Klinik I, Dresden, Germany; 3Goethe University, Medizinische Klinik II, Frankfurt, Germany; 4Charite´ Campus Virchow Klinikum, Department of Haematology, Oncology and Tumor Immunology, Berlin, Germany; 5University of Heidelberg, Medizinische Klinik V, Heidelberg, Germany; 6St. Johannes Hospital, Department of Hematology and Oncology, Duisburg, Germany and 7University Hospital Hamburg-Eppendorf, Clinic for Stem Cell Transplantation, Hamburg, Germany. Correspondence: Dr T Schroeder, Department of Hematology, Oncology and Clinical Immunology, University of Duesseldorf, Medical Faculty, Moorenstr. 5, Duesseldorf, Nord Rhine Westphalia, 40225, Germany. E-mail: [email protected] Parts of this study have been presented at the 52nd American Society of Hematology (ASH) Annual Meeting, Orlando, FL, 4–7 December 2010, the BMT Tandem Meeting 2011, Honolulu, HI, 17–21 February 2011, the 37th Annual Meeting of the European Group for Bone and Marrow Transplantation (EBMT), Paris, France, 3–7 April 2011 and at the 53rd American Society of Hematology (ASH) Annual Meeting, San Diego, CA, 10–13 December 2011. 8These authors contributed equally to this work. Received 16 October 2012; revised 16 December 2012; accepted 3 January 2013; accepted article preview online 14 January 2013; advance online publication, 8 February 2013 Aza and DLI for AML or MDS relapse after allo-HSCT T Schroeder et al 1230 multicenter, open-label, single-arm, phase-II study. It was opened in Although hematotoxicity is difficult to determine owing to disease- January 2009 at the University of Duesseldorf and conducted with five related cytopenias in the majority of these patients, myelosuppression other transplant centers (Hamburg, Dresden, Frankfurt, Berlin, Heidelberg). is a notable side effect of Aza. This concern was incorporated into a The last patient was enrolled in May 2010. The trial received local dose-modification plan based on baseline and nadir leukocyte counts institutional review board’s approval from all participating sites and was (Supplementary Figure 1). conducted in accordance with the Declaration of Helsinki. Sponsor of the In case of proliferative disease with an increase of white blood cell study was the University of Duesseldorf (http://clinicaltrials.gov (NCT- counts 410 000/ml in the interval between two Aza cycles, the subsequent 00795548)). cycle could be administered early. Sequential cytoreductive therapy with low-dose cytarabine was allowed until normalization of white blood cells, but co-administration with Aza was prohibited. Patients Patients aged X18 years with hematological relapse of AML (according to World Health Organization classification) or MDS following allo-HSCT were Study end points eligible. Relapse was defined as bone marrow blasts X5%, reappearance The primary end point was the best response within the first 6 months of of blasts in peripheral blood and/or extramedullary disease.22 Other major treatment. Responses were reported as defined by an international indispensable prerequisites were availability of DLI and that no other consensus panel for AML and by the 2006 International Working Group antileukemic or cellular treatment for relapse (including pre-emptive DLI) criteria for MDS.22,23 In patients achieving complete remission (CR), a had been given. Major exclusion criteria were an Eastern Cooperative cytomorphological review of slides obtained at diagnosis and CR was Oncology Group performance status 42, uncontrolled infection, renal performed by an independent external physician (AG). insufficiency (glomerular filtration rate o50 ml/min) and hepatic Secondary end points were overall response, duration of remissions and impairment, including bilirubine elevation 41.5-fold above the upper restoration of complete donor chimerism. Adverse events (AEs) were normal range or malignant liver tumors. All patients gave written informed recorded and graded using National Cancer Institute Common Toxicity consent. Criteria (NCI CTC) version 3.0. Incidence and severity of acute and chronic GvHD (aGvHD and cGvHD, respectively) was reported according to established criteria.24,25 Patients were followed until death, data lock Treatment (December 2011) or second transplantation. Patients were scheduled to receive six cycles Aza (Vidaza, Celgene Corporation, Summit, NJ, USA) 100 mg/m2/day subcutaneously on days 1–5 repeated every 28 days. In case of any type of response, two additional Statistics cycles of Aza could be given according to individual investigator’s Continuous variables were summarized using median (range), while judgment. frequency tables were used for categorical variables. Time-to-event curves DLIs were given on the sixth day of every second Aza cycle thereby were estimated using the Kaplan–Meier method, and the log-rank test was immediately after the completion of the 5 days of Aza administration used for comparison of subgroups. For comparison of biological variables’ (corresponding to day 34, 90 and 146) with increasing dosages of CD3 þ cross tabulation, Fisher’s exact test as well as Mann–Whitney test were used. cells, if no GvHD was observed. Additional DLI were permitted according to the treating physician’s decision (Figure 1). The 5-day dosing scheme of Aza was chosen to facilitate an outpatient RESULTS setting while still delivering a dose almost equal to the approved dosage of Patient characteristics Aza. Taking into account the results of Schmid et al.7 and our own experience,19 the timing for the first DLI was scheduled after cycle 2 in Thirty patients were enrolled into this study (Table 1). The majority order to reduce the leukemic burden before DLI and to ensure the of patients suffered from AML (n ¼ 28, 93%), including 15 patients availability of DLI. The dose-escalating strategy was intended to manage with AML with MDS-related changes, while 2 patients (both the balance between the GvL effect and the risk of GvHD following DLI. with bone marrow blast count X10% at relapse) had MDS or

Cycle 1 30 pts † Infection (n = 1) n = 5 Progress (n = 4)

Cycle 2 25 pts † Infection (n = 1) n = 9 † Bleeding (n = 2) st 1 DLI Progress (n = 5) Patient‘s wish (n = 1) Cycle 3 16 pts n = 3 Progress (n = 3)

Cycle 4 13 pts n = 3 Progress (n = 3) 2nd DLI Cycle 5 10 pts n = 2 † Infection (n = 1) Progress (n = 1) Cycle 6 8 pts n = 3 Progress (n = 1) rd 3 DLI CR (Relapse d396) (n = 1) Cycle 7 5 pts aGvHD (CR ongoing) (n = 1)

CR ongoing (n = 4) Cycle 8 5 pts † aGvHD/MOF (n = 1) (CR ongoing) d203

Figure 1. Treatment schedule and clinical course of all 30 patients during the study. Patients were scheduled to receive up to 6 cycles of Aza (100 mg/m2/day, days 1–5) on a 28-day schedule. Two additional cycles of Aza could be given in case of response. Three DLI with increasing CD3 þ cell dosages ranging from 1 106/kg bodyweight to 5 108/kg bodyweight were planned following every second Aza cycle. Reasons for drop-out and causes of deaths (w) are indicated per cycle. A total of seven patients (pts) reached a CR, of which ﬁve are still in CR for a median of 777 days (d, range 461–888 d). Of the two remaining patients, one died due to multiorgan failure (MOF) associated with aGvHD retaining CR for 203 days. The other patient stayed in CR for 396 days, then relapsed again and ﬁnally died of progressive disease.

Leukemia (2013) 1229 – 1235 & 2013 Macmillan Publishers Limited Aza and DLI for AML or MDS relapse after allo-HSCT T Schroeder et al 1231

Table 1. Patient demographics and clinical characteristics. Table 2. Relapse characteristics

No. % No. %

No. 30 Median time to relapse, days 175 Median age, years 55 Range 19–1688 Range 29–71 Median time from allo-HSCT to 179 start of Aza-treatment, days Sex Range 30–1700 Male 13 43 Perfomance status Female 17 57 Karnofsky Index AML 28 93 Median, % 90 WHO Range 60–100 With recurrent genetic abnormalities 2 7 Sorror Score MDS-related changes 15 50 Low 12 40 NOS 11 36 Intermediate 9 30 MDS 27High 7 23 RAEB-I 1 3.5 Median WBC, 109/l 3.5 CMML-I 1 3.5 Range 1.05–8.7 Median blasts PB, % 1 Karyotype Range 0–51 Normal 13 45 Median blasts BM, % 34 Aberrant 16 55 Range 5–100 Complex 9 31 Median Hb, g/dl 10.7 Monosomala 414 Range 5–16.4 Missing 1 3 Median Plts, 109/l 47 Range 4–360 Molecular/genetic riskb Median LDH, U/l 223 Favorable 2 7 Range 137–628 Intermediate 13 43 Median BM chimerism, % 67 Adverse 14 47 Range 1–99 Missing 1 3 Extramedullary relapse Disease status at allo-HSCT Cutaneous 1 3 Remission 14 47 GvHD before relapse No remission 16 53 Acute 13 43 Grade I 4 13 CR1 12 40 Grade II 6 20 CR2 2 7 Grade III 2 7 Relapse 1 6 20 Grade IV 1 3 Untreated 3 10 Chronic 4 13 Limited 3 10 Conditioning Extensive 1 3 Standard-dose 4 13 Dose-reduced 26 87 Immunosuppression at study entry Cy/TBI 7 23 Yes 6 20 Flu/Mel 7 23 No 24 80 Flu/Bu 6 20 Bu/Cy 5 17 Taper/stop 5 17 Flu/TBI 3 10 GvHD flare 1 3 Flu/Treo 1 3 Abbreviations: BM, bone marrow; GvHD, graft-versus-host disease; Hb, Flu/BCNU/Mel 1 3 hemoglobin; LDH, lactate dehydrogenase; PB, peripheral blood; Plts, Donor/HLA-match platelets; WBC, white blood cells. Related 10 33 Unrelated 20 67 myelodysplastic/myeloproliferative neoplasm. Of 29 patients who had karyotype analysis performed at diagnosis, 16 patients (55%) 10/10 23 77 had an abnormal karyotype, including 13 with complex (31%) or 9/10 4 13 monosomal (14%) karyotype (Supplementary Table 1). Based 8/10 3 10 on molecular data, 2 patients belonged to a favorable (7%), 13 to Graft source an intermediate (43%) and 14 to an adverse (47%) risk group PBSC 29 97 according to ELN (European LeukemiaNet) consensus classification BM 1 3 (for the 28 AML patients) and International Prognostic Scoring System (for the 2 MDS patients).22,26 At transplant, 16 patients Abbreviations: AML, acute myeloid leukemia; BCNU, carmustine; (53%) had active disease, including 6 patients with induction failure, Bu, busulfan; BM, bone marrow; CMML, chronic myelomonocytic leukemia; 7 patients with first or second relapse and 3 previously untreated CR, complete remission; Cy, cyclophosphamide; Flu, fludarabine; HLA, human leukocyte antigen; MDS, myelodysplastic syndromes; Mel, melpha- patients. A dose-reduced conditioning regimen was used in 87%. lan; NOS, not otherwise specified; PBSC, peripheral blood stem cells; RAEB, All 30 patients suffered from hematological relapse mirrored by refractory anemia with excess blasts; TBI, total body irradiation; Treo, a median bone marrow blast count of 34% (range 5–100%) and a treosulfan; WHO, World Health Organization. aDefined as in Breems et al.41 median donor chimerism of 67% (range 1–100%). Median time to bAccording to Dohner et al.22 and Greenberg et al.26 relapse was 175 days (range 19–1688 days). Table 2 displays detailed relapse characteristics.

& 2013 Macmillan Publishers Limited Leukemia (2013) 1229 – 1235 Aza and DLI for AML or MDS relapse after allo-HSCT T Schroeder et al 1232 (14.5% versus 45.5%, P ¼ 0.039, Supplementary Figure 2). Inter- 1.0 p < 0.001 estingly, the presence of high-risk cytogenetics was also associated with a greater probability to achieve CR (P ¼ 0.035), while white blood cells, peripheral blood blast count, chimerism, GvHD before relapse, donor type and interval from allo-HSCT to 0.8 CR relapse had no predictive value. In line with this hypoproliferative phenotype bearing a high frequency of karyotype abnormalities, the overall response rate in patients with MDS or AML with MDS- 0.6 related changes was significantly higher than in those patients with other AML subtypes (65% versus 15%, P ¼ 0.011) By December 2011, median follow-up for surviving patients was 817 days (range 732–974 days) and 5 patients (17%) were alive 0.4 and free of disease. Twelve patients died from PD, while 7 patients Cum Survival died during or after treatment due to infections (n ¼ 4), bleeding (n ¼ 2) or GvHD (n ¼ 1) (Figure 1). Five non-responding patients received second allo-HSCT and subsequently died. The remaining All pts 0.2 patient got lost to follow-up. Median overall survival of the entire group was 117 days (range 25–974 days), not reached (range 253–974 days) for patients achieving CR and 83 days (range 25–446 days) for patients without CR (Po0.001, Figure 2). 0.0 No CR GvHD 0.00 200.00 400.00 600.00 800.00 1000.00 Thirteen patients (43%) had suffered from aGvHD and four Days following start of Aza treatment (OS) patients (13%) from cGvHD before relapse. Six patients still Figure 2. Overall survival (OS) following Aza treatment. received immunosuppressive therapy at study entry. Immunosup- Kaplan–Meier plots are shown for all patients (pts, n ¼ 30, gray pression could be tapered in five patients, while one patient curve), patients with complete remission (CR, n ¼ 7, green curve) suffered from GvHD flare (Table 2). and patients without CR (non-CR, n ¼ 23, blue curve). Log-rank: Eleven patients (37%) developed aGvHD while on study: six Po0.001. patients overall grade I, 2 patients grade II and 3 patients grade III. Eight of these patients (73%) had received DLI before aGvHD onset. Two of the remaining patients did not receive DLI due to Treatment response and survival active aGvHD and PD, respectively, while the third patient developed aGvHD 5 days before the first DLI. Median time from The combination of Aza and DLI was the first treatment for relapse first DLI to aGvHD onset was 97 days (range 20–251 days) after allo-HSCT in all patients reflected by a median time between corresponding to a median of 1 DLI (range 1–3 DLI) and 5 Aza diagnosis of relapse and onset of treatment of 9 days (range 0–60 cycles (range 2–8 cycles). days). This corresponds to a median time of 179 days (range 30– Eight patients required systemic immunosuppressive therapy. 1700 days) between allo-HSCT and the start of Aza therapy. No Six patients responded to therapy, while one patient died due to patient had received any antileukemic or cellular therapy before aGvHD and the other due to PD 9 days following GvHD onset. Five inclusion to the study. Patients received a median of three cycles patients showed symptoms of cGvHD, which was limited in all but of Aza (range 1–8 cycles). DLIs were administered to 22 patients one case (Supplementary Table 3). (73%). The median number of DLI per patient was 1 (range 1–4 DLI). Twelve patients received 1 DLI, 3 patients 2 DLI, 5 patients 3 DLI and 2 patients 4 DLI, respectively. By this, the median number Safety and toxicity of CD3 þ cells per patient was 5.0 106/ kg (range 1–165 106). At baseline, 50% of the patients already had neutropenia grade III/ Reasons for not giving DLI in the remaining patients were IV, 53% thrombopenia grade III/IV, while no patient had anemia progressive disease (PD) or death before DLI (n ¼ 6), infection grade III/IV. Only six patients did not exhibit at least one cytopenia (n ¼ 1) and GvHD flare before DLI (n ¼ 1) (Figure 1). Xgrade II at baseline. During study, a total of 112 cycles of Aza Overall response rate was 30% with seven patients achieving CR were administered to the patients. As shown in Table 3, grade III/IV (23%) and two patients a partial remission (7%). In addition, five neutropenia, thrombopenia and anemia were observed during 65, patients achieved a stable disease (17%). Median time and 63 and 33% of treatment cycles. Liver and renal dysfunctions were number of Aza cycles to best response were 79 days (range 28– uncommon following Aza treatment (Table 3). Of 606 AEs 299 days) and 3 cycles (range 1–8 cycles), respectively (Figure 1). reported, 163 were considered to be drug-related and 97 of Detailed characteristics of patients achieving CR are displayed in these were non-hematological AEs. The most common AEs were Supplementary Table 2. Median time to CR was 84 days (range 33– local skin reactions and gastrointestinal symptoms, which were 299 days) corresponding to 3 cycles of Aza (range 2–8 cycles). All mostly mild. Infections were more severe with 10 of 14 (71%) patients with CR had received DLI (median 3, range 3–4 DLI), while classified as grade III/IV. One thrombocytopenic patient suffered three patients were already in CR before first DLI. In all patients from hemianopsia, which was caused by intracerebral bleeding with CR, we observed restoration of complete donor chimerism in (Table 4). median 90 days (range 33–299 days) following start of Aza therapy. Five of the 7 patients continue in CR for a median duration of 777 days (range 461–888 days) without any additional DISCUSSION antileukemic treatment. One patient relapsed after 396 days after In this first prospective study, we investigated the combination of achieving CR and died due to PD, and one patient died after 203 Aza and DLI as first salvage therapy for hematological relapse of days in CR from aGvHD. AML or MDS following allo-SCT. We observed a CR rate of 23%, When looking for pretreatment variables associated with the with 5 of 7 patients remaining in remission for 42 years despite likelihood of reaching CR, these patients had a lower bone their unfavorable risk profile mirrored by a median age of 55 years marrow blast count at relapse compared with patients without CR and adverse cytogenetics in 90% of patients. Moreover, these

Leukemia (2013) 1229 – 1235 & 2013 Macmillan Publishers Limited Aza and DLI for AML or MDS relapse after allo-HSCT T Schroeder et al 1233 patients had a median time to relapse of only 175 days, including strategies is difficult due to different patient characteristics and nine patients relapsing within the first 100 days. Our prospective the fact that the retrospective studies on DLI or second transplant data now confirm the results observed in retrospective analyses are biased as they do only include patients who survived until the using varying schedules of Aza alone or in combination with DLI intervention. Therefore, randomized trials are definitely needed to for treatment of AML or MDS relapse.17–21 In contrast to them, as a compare different treatment strategies for relapse after allo-HSCT. particular strength, this combined pharmacological and cell-based In order to identify those patients who benefit most from our approach was the first intervention for relapse in our study. approach, a lower blast count at relapse and presence of high-risk Considering other treatment options such as chemotherapy, DLI cytogenetics appeared as significant predictors for CR. Surpris- or second transplantation5,6 our CR rate of 23% seems comparable ingly, in contrast to earlier studies, the interval between allo-SCT as CR rates following chemotherapy vary from 17–30%, are usually and relapse had no impact. short lived and consequently survival after 2 years is o10%.5 Our finding with regard to blast counts is in line with large Following DLI, a CR rate of 34% has been reported resulting in a registry studies, demonstrating that lower disease burden is 2-year survival of 21%,7 while a 5-year leukemia-free and overall associated with better outcome.27 In addition, blast counts may survival of 28% was observed following second transplantation in also reflect relapse kinetics, thereby suggesting that patients with a selected group of young patients.8 Finally, our 2-year overall slow relapse kinetics might be those who benefit. survival rate of 17% compares at least well with the 2-year overall These issues were also addressed by two other study groups. survival rate of 13.9% in a cohort of 776 patients with AML relapse One group is currently investigating an Aza-based maintenance after allo-HSCT receiving chemotherapy, DLI and/or second therapy in patients with high-risk leukemia in a randomized trial.28 transplantation.27 However, a direct comparison between these Other investigators studied Aza without cellular therapy guided by donor chimerism in circulating CD34 þ cells.29 This approach is promising, as it incorporates the idea that it is probably more Table 3. Hematotoxicity and laboratory findings according to the NCI successful to treat patients with lower disease burden, but may CTC version 3.0 during the study also have some limitations. The methods to detect imminent relapse in circulating CD34 þ cells are technically sophisticated Grade Grade and may not be available in all transplant centers. In addition, despite improvement of chimerism in 50% (10 of 20) of patients ANC Plts Hb Bilirubine Creatinine hematological relapse occurred in the majority of patients and continuous remission was only achieved in 3 (30%) of the No. % No. % No. % No. % No. % responders. This is in contrast to our results where relapse occurred in only 1 out of 7 patients (15%, median follow-up I 5 4161426238 7 2724 777 days), probably related to the fact that DLI are required to II 9 8 10 9 36 32 3 3 3 3 achieve long-term disease control. Supporting this idea, Schmid III 13 12 14 12 36 32 3 3 2 1 et al. as well as a consortium of the French transplant centers IV 59 53 57 51 1 1 1 1 0 0 No toxicity 26 23 15 14 13 12 97 86 80 84 (SFGM-TC) have recently demonstrated that reinduction of CR alone is not sufficient for long-term survival, but donor-cell-based Abbreviations: ANC, absolute neutrophile count; Hb, hemoglobin; NCI CTC, consolidation is required.27,30 National Cancer Institute Common Toxicity Criteria; Plts, platelets. In all, 112 The finding, that patients with high-risk cytogenetics had a cycles were administered to 30 patients. Hematotoxicity as well as higher probability to achieve CR, is of interest, as MDS patients creatinine and bilirubine were graded in every cycle. The maximal toxicity with high-risk cytogenetics in non-transplant setting exhibited per cycle is given as absolute number and relative proportion of cycles. higher response rates following treatment with decitabine.31

Table 4. Drug-related non-hematological adverse events according to the NCI CTC version 3.0 during the study

Events Grade

I ll III IV

No. No. % No. % No. % No. %

Skin reactions 26 24 92 2 8 0 0 0 0

Nausea 15 12 80 2 13 1 7 0 0 Infections 14 1 7 3 22 8 57 2 14 Diarrhea 9 9 100 0 0 0 0 0 0 Vomiting 8 6 75 2 25 0 0 0 0 Impairment of performance status 6 4 67 2 33 0 0 0 0 Pain 4 2 50 1 25 0 0 1 25 Dyspnea 3 3 100 0 0 0 0 0 0 Laboratory changes 3 1 33 1 33 1 33 0 0 Meteorism 2 2 100 0 0 0 0 0 0 Fever 2 1 50 1 50 0 0 0 0 Loss of apetite 1 1 100 0 0 0 0 0 0 Joint pain 1 1 100 0 0 0 0 0 0 Dizziness 1 1 100 0 0 0 0 0 0 Vision impairment 1 0 0 0 0 1 100 0 0 Hemianopsia 1 0 0 0 0 0 0 1 100 Abbreviation: NCI CTC, National Cancer Institute Common Toxicity Criteria.

& 2013 Macmillan Publishers Limited Leukemia (2013) 1229 – 1235 Aza and DLI for AML or MDS relapse after allo-HSCT T Schroeder et al 1234 This finding as well as a hypoproliferative disease phenotype CONFLICT OF INTEREST suggests that patients suffering from relapse of MDS or AML with TS received financial travel support and honoraria from Celgene Corporation, MDS-related changes might benefit most from treatment with Aza Germany. UP received advisory fees, honoraria and research funding from Celgene and DLI. In line with this hypothesis, we observed a significant Corporation, Germany. GB received advisory fees, honoraria, travel grants and higher overall response rate in these patients in comparison to research funding from Celgene Corporation, Germany. FZ is employed as medical patients with other AML subtypes. Data on the use of DLI alone in director Millenium Pharmaceuticals, USA. RF received advisory fees, honoraria, travel patients with MDS and AML with MDS-related changes are limited grants and research funding from Celgene Corporation, Germany. UG received but indicate that long-term disease control is infrequent and even received honoraria and research funding from Celgene Corporation, Germany. NK received research funding from Celgene Corporation, Germany. The remaining more hampered by the fact that those mostly older patients are authors declare no conflict of interest. not eligible for intensive treatment before DLI.32 Compared with published data, our results achieved in this particular patient group argue in favor for the combination of Aza and DLI. Due to cumulative cytotoxic effects, regimen-related toxicity is a ACKNOWLEDGEMENTS relevant problem, especially for the treatment of early relapse This study was supported by research funding of Celgene Corporation, Germany to after allo-HSCT. In our study, the combination of Aza and DLI was GK. This work was also supported by the Leuka¨mie Lymphom Liga e. V. Duesseldorf, well-tolerated and given on an outpatient basis despite a rather Germany and the Forschungskommision of the Heinrich-Heine-University, Duessel- dorf, Germany. We would like to thank the staff of the transplant unit of the high patients’ age and a proportion of 53% having an Department of Hematology, Oncology and Clinical Immunology for excellent patient intermediate or high hematopoietic stem cell transplantation- care as well as Julia Fro¨bel, Ron-Patrick Cadeddu and Annemarie Koch for excellent comorbidity index. The great majority of patients in our study died laboratory work. of PD. In 7 patients (29%), PD was not the main cause of death, thereby comparing well with treatment-related mortality reported 8,27,33 with chemotherapy or second transplantation. Four of these AUTHOR CONTRIBUTIONS patients died due to infections. Infections were also the most common non-hematological AEs 4grade II. This is related to high Conception and design: GK and AC. Provision of study materials or patients: rates of grade III/IV neutropenia during our study. However, as NK, GB, UP, TL, LU, KR, CW, GK and TS. Collection and assembly of data: TS, nearly half of the patients had grade III/IV neutropenia at study GK, NK, GB, UP, TL, LU, KR and CW. Data analysis and interpretation: TS, GK, AC, entry, cytopenias and consecutive complications like infections RF, FZ and IB. Manuscript writing: TS, GK, RH and AC. Final approval of the and bleeding can be attributed, in part, not only to the Aza-related manuscript: all authors. myelosuppression but also to the underlying disease. As CR occurred after a median of 84 days, extensive microbiological surveillance and prophylaxis is recommended. Dose reduction of REFERENCES Aza may help to reduce hematotoxicity and associated 1 Schlenk RF, Dohner K, Mack S, Stoppel M, Kiraly F, Gotze K et al. Prospective complications. Jabbour et al.21 have reported a comparable CR evaluation of allogeneic hematopoietic stem-cell transplantation from matched rate of 33% with Aza doses ranging from 16–40 mg/m2 for 5 days. related and matched unrelated donors in younger adults with high-risk Another relevant problem following DLI is the induction of acute myeloid leukemia: German-Austrian trial AMLHD98A. J Clin Oncol 2010; 28: GvHD as indicated by an aGvHD rate of 40–50%, with 80% of 4642–4648. affected patients having aGvHD grade II-IV in the largest report.7 2 Saure C, Schroeder T, Zohren F, Groten A, Bruns I, Czibere A et al. Upfront allogeneic blood stem cell transplantation for patients with high-risk myelodys- Our protocol included DLI and 73% of the patients received at plastic syndrome or secondary acute myeloid leukemia using a FLAMSA-based least one DLI. The aGvHD incidence in our group was 37% being high-dose sequential conditioning regimen. Biol Blood Marrow Transplant 2011; mostly grade I/II disease, while cGvHD was observed in 17%. In 18: 466–472. line with our previous experience, aGvHD onset was rather late 3 Pavletic SZ, Kumar S, Mohty M, de LM, Foran JM, Pasquini M et al. NCI First with a median of 97 days following first DLI, and most patients International Workshop on the Biology, Prevention, and Treatment of Relapse requiring systemic immunosuppressive therapy responded.19 after Allogeneic Hematopoietic Stem Cell Transplantation: report from the Only one patient died due to GvHD. Furthermore in five of six Committee on the Epidemiology and Natural History of Relapse following Allo- patients entering the study while on immunosuppression, geneic Cell Transplantation. Biol Blood Marrow Transplant 2010; 16: 871–890. immunosuppressive therapy could be tapered. This rather low 4 van den Brink MR, Porter DL, Giralt S, Lu SX, Jenq RR, Hanash A et al. Relapse after allogeneic hematopoietic cell therapy. Biol Blood Marrow Transplant 2010; incidence and mild presentation of GvHD support recent results 16(1 Suppl): S138–S145. demonstrating a mitigation of GvHD probably related to an Aza- 5 Porter DL, Alyea EP, Antin JH, DeLima M, Estey E, Falkenburg JH et al. NCI First 15,34 induced expansion of regulatory T cells. Addressing this International Workshop on the Biology, Prevention, and Treatment of Relapse question, we found an 1.6-fold increase of circulating regulatory after Allogeneic Hematopoietic Stem Cell Transplantation: report from T cells (CD4 þ CD25 þ FoxP3 þ ) after four Aza cycles (baseline: the Committee on Treatment of Relapse after Allogeneic Hematopoietic Stem Cell 6.3 109/l; day 90: 10.3 109/l), albeit not reaching statistical Transplantation. Biol Blood Marrow Transplant 2010; 16: 1467–1503. significance due to the low number of only 7 patients being on 6 Kroger N. Approaches to relapse after allogeneic stem cell transplantation. study at that point. Curr Opin Oncol 2011; 23: 203–208. In conclusion, we demonstrate here that Aza and DLI is a safe 7 Schmid C, Labopin M, Nagler A, Bornhauser M, Finke J, Fassas A et al. Donor lymphocyte infusion in the treatment of first hematological relapse after and active approach for relapse after allo-SCT representing a allogeneic stem-cell transplantation in adults with acute myeloid leukemia: a valuable alternative to current treatment modalities, especially in retrospective risk factors analysis and comparison with other strategies by the patients with MDS or AML with MDS-related changes. Besides EBMT Acute Leukemia Working Party. J Clin Oncol 2007; 25: 4938–4945. identification of target patient groups, a better understanding of 8 Eapen M, Giralt SA, Horowitz MM, Klein JP, Wagner JE, Zhang MJ et al. Second these Aza-mediated mechanisms in the post-transplant situation transplant for acute and chronic leukemia relapsing after first HLA-identical will hopefully enable the development of combination strategies. sibling transplant. Bone Marrow Transplant 2004; 34: 721–727. In this context, lenalidomide appears as an interesting combina- 9 Bosi A, Laszlo D, Labopin M, Reffeirs J, Michallet M, Gluckman E et al. Second tion partner due to its promising synergistical effects with Aza in allogeneic bone marrow transplantation in acute leukemia: results of a survey patients with MDS and its immunomodulating potential.35–40 We by the European Cooperative Group for Blood and Marrow Transplantation. J Clin Oncol 2001; 19: 3675–3684. have therefore initiated a new trial that will investigate the 10 Kuendgen A, Graf T, Zohren F, Hildebrandt B, Hunerliturkoglu A, Gattermann N et combination of Aza, lenalidomide and DLI in patients with al. Induction of complete remission in a patient with acute myeloid leukemia relapse of MDS and AML with MDS-related changes following refractory to high-dose chemotherapy through treatment with 5-azacytidine. allo-HSCT. Leuk Res 2007; 31: 407–409.

Leukemia (2013) 1229 – 1235 & 2013 Macmillan Publishers Limited Aza and DLI for AML or MDS relapse after allo-HSCT T Schroeder et al 1235 11 Atanackovic D, Luetkens T, Kloth B, Fuchs G, Cao Y, Hildebrandt Y et al. 27 Schmid C, Labopin M, Nagler A, Niederwieser D, Castagna L, Tabrizi R et al. Cancer-testis antigen expression and its epigenetic modulation in acute myeloid Treatment, risk factors, and outcome of adults with relapsed AML after reduced leukemia. Am J Hematol 2011; 86: 918–922. intensity conditioning for allogeneic stem cell transplantation. Blood 2012; 119: 12 Pinto A, Maio M, Attadia V, Zappacosta S, Cimino R. Modulation of HLA-DR 1599–1606. antigens expression in human myeloid leukaemia cells by cytarabine and 28 de LM, Giralt S, Thall PF, de Padua SL, Jones RB, Komanduri K et al. Maintenance 5-aza-20-deoxycytidine. Lancet 1984; 2: 867–868. therapy with low-dose azacitidine after allogeneic hematopoietic stem cell 13 Pinto A, Attadia V, Fusco A, Ferrara F, Spada OA, Di Fiore PP. 5-Aza-20-deox- transplantation for recurrent acute myelogenous leukemia or myelodysplastic ycytidine induces terminal differentiation of leukemic blasts from patients with syndrome: a dose and schedule finding study. Cancer 2010; 116: 5420–5431. acute myeloid leukemias. Blood 1984; 64: 922–929. 29 Platzbecker U, Wermke M, Radke J, Oelschlaegel U, Seltmann F, Kiani A et al. 14 Goodyear O, Agathanggelou A, Novitzky-Basso I, Siddique S, McSkeane T, Ryan G Azacitidine for treatment of imminent relapse in MDS or AML patients after et al. Induction of a CD8 þ T-cell response to the MAGE cancer testis antigen by allogeneic HSCT: results of the RELAZA trial. Leukemia 2011; 26: 381–389. combined treatment with azacitidine and sodium valproate in patients with acute 30 Guieze R, Damaj G, Robin M, Mohty M, Michallet M, Tabrizi R et al. Treatment of myeloid leukemia and myelodysplasia. Blood 2010; 116: 1908–1918. relapse after allogeneic hematopoietic sem-cell transplantation for myelodys- 15 Sanchez-Abarca LI, Gutierrez-Cosio S, Santamaria C, Caballero-Velazquez T, plastic syndrome: a large-scale study on behalf of the Societe Francaise De Greffe Blanco B, Herrero-Sanchez C et al. Immunomodulatory effect of 5-azacytidine De Moelle Et De Therapie Cellulaire (SFGM-TC). ASH Annual Meeting Abstracts (5-azaC): potential role in the transplantation setting. Blood 2010; 115: 107–121. 2012; 120: 593. 16 Choi J, Ritchey J, Prior JL, Holt M, Shannon WD, Deych E et al. In vivo adminis- 31 Lubbert M, Wijermans P, Kunzmann R, Verhoef G, Bosly A, Ravoet C et al. Cyto- tration of hypomethylating agents mitigate graft-versus-host disease without genetic responses in high-risk myelodysplastic syndrome following low-dose sacrificing graft-versus-leukemia. Blood 2010; 116: 129–139. treatment with the DNA methylation inhibitor 5-aza-20-deoxycytidine. Br J Hae- 17 Graef T, Kuendgen A, Fenk R, Zohren F, Haas R, Kobbe G. Successful treatment of matol 2001; 114: 349–357. relapsed AML after allogeneic stem cell transplantation with azacitidine. Leuk Res 32 Campregher PV, Gooley T, Scott BL, Moravec C, Sandmaier B, Martin PJ et al. 2007; 31: 257–259. Results of donor lymphocyte infusions for relapsed myelodysplastic syndrome 18 Bolanos-Meade J, Smith BD, Gore SD, McDevitt MA, Luznik L, Fuchs EJ et al. after hematopoietic cell transplantation. Bone Marrow Transplant 2007; 40: 5-Azacytidine as salvage treatment in relapsed myeloid tumors after allogeneic 965–971. bone marrow transplantation. Biol Blood Marrow Transplant 2011; 17: 754–758. 33 Arellano ML, Langston A, Winton E, Flowers CR, Waller EK. Treatment of relapsed 19 Czibere A, Bruns I, Kroger N, Platzbecker U, Lind J, Zohren F et al. 5-Azacytidine for acute leukemia after allogeneic transplantation: a single center experience. Biol the treatment of patients with acute myeloid leukemia or myelodysplastic Blood Marrow Transplant 2007; 13: 116–123. syndrome who relapse after allo-SCT: a retrospective analysis. Bone Marrow 34 Goodyear OC, Dennis M, Jilani NY, Loke J, Siddique S, Ryan G et al. Azacitidine Transplant 2010; 45: 872–876. augments expansion of regulatory T cells after allogeneic stem cell transplanta- 20 Lubbert M, Bertz H, Wasch R, Marks R, Ruter B, Claus R et al. Efficacy of a 3-day, tion in patients with acute myeloid leukemia. Blood 2012; 119: 3361–3369. low-dose treatment with 5-azacytidine followed by donor lymphocyte infusions 35 Blum W, Klisovic RB, Becker H, Yang X, Rozewski DM, Phelps MA et al. in older patients with acute myeloid leukemia or chronic myelomonocytic leu- Dose escalation of lenalidomide in relapsed or refractory acute leukemias. J Clin kemia relapsed after allografting. Bone Marrow Transplant 2010; 45: 627–632. Oncol 2010; 28: 4919–4925. 21 Jabbour E, Giralt S, Kantarjian H, Garcia-Manero G, Jagasia M, Kebriaei P et al. 36 Fehniger TA, Uy GL, Trinkaus K, Nelson AD, Demland J, Abboud CN et al. A phase 2 Low-dose azacitidine after allogeneic stem cell transplantation for acute leuke- study of high-dose lenalidomide as initial therapy for older patients with acute mia. Cancer 2009; 115: 1899–1905. myeloid leukemia. Blood 2011; 117: 1828–1833. 22 Dohner H, Estey EH, Amadori S, Appelbaum FR, Buchner T, Burnett AK et al. 37 Fenaux P, Giagounidis A, Selleslag D, Beyne-Rauzy O, Mufti G, Mittelman M et al. A Diagnosis and management of acute myeloid leukemia in adults: recommenda- randomized phase 3 study of lenalidomide versus placebo in RBC transfusion- tions from an international expert panel, on behalf of the European LeukemiaNet. dependent patients with Low-/Intermediate-1-risk myelodysplastic syndromes Blood 2010; 115: 453–474. with del5q. Blood 2011; 118: 3765–3776. 23 Cheson BD, Greenberg PL, Bennett JM, Lowenberg B, Wijermans PW, 38 Lioznov M, El-Cheikh Jr. J, Hoffmann F, Hildebrandt Y, Ayuk F, Wolschke C et al. Nimer SD et al. Clinical application and proposal for modification of the Inter- Lenalidomide as salvage therapy after allo-SCT for multiple myeloma is effective national Working Group (IWG) response criteria in myelodysplasia. Blood 2006; and leads to an increase of activated NK (NKp44( þ )) and T (HLA-DR( þ )) cells. 108: 419–425. Bone Marrow Transplant 2010; 45: 349–353. 24 Filipovich AH, Weisdorf D, Pavletic S, Socie G, Wingard JR, Lee SJ et al. National 39 Sekeres MA, List AF, Cuthbertson D, Paquette R, Ganetzky R, Latham D et al. Phase Institutes of Health consensus development project on criteria for clinical trials in I combination trial of lenalidomide and azacitidine in patients with higher-risk chronic graft-versus-host disease: I. Diagnosis and staging working group report. myelodysplastic syndromes. J Clin Oncol 2010; 28: 2253–2258. Biol Blood Marrow Transplant 2005; 11: 945–956. 40 Sekeres MA, O’Keefe C, List AF, Paulic K, Afable M, Englehaupt R et al. Demon- 25 Przepiorka D, Weisdorf D, Martin P, Klingemann HG, Beatty P, Hows J et al. stration of additional benefit in adding lenalidomide to azacitidine in patients 1994Consensus Conference on Acute GVHD Grading. Bone Marrow Transplant with higher-risk myelodysplastic syndromes. Am J Hematol 2011; 86: 102–103. 1995; 15: 825–828. 41 Breems DA, Van Putten WL, De Greef GE, Van Zelderen-Bhola SL, Gerssen-Schoorl 26 Greenberg P, Cox C, LeBeau MM, Fenaux P, Morel P, Sanz G et al. International KB, Mellink CH et al. Monosomal karyotype in acute myeloid leukemia: a better scoring system for evaluating prognosis in myelodysplastic syndromes. Blood indicator of poor prognosis than a complex karyotype. J Clin Oncol 2008; 26: 1997; 89: 2079–2088. 4791–4797.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

& 2013 Macmillan Publishers Limited Leukemia (2013) 1229 – 1235