Oral presentations

Oral presentations

Leukemia Research 25 Supplement Number 1 (2001) S1 – S22 www.elsevier.com/locate/leukres Or1 CLASSIFICATION OF MYELODYSPLASTIC SYNDROME: FAB VERSUS W...

241KB Sizes 0 Downloads 36 Views

Leukemia Research 25 Supplement Number 1 (2001) S1 – S22 www.elsevier.com/locate/leukres

Or1 CLASSIFICATION OF MYELODYSPLASTIC SYNDROME: FAB VERSUS WHO AND PEDIATRIC MDS 1

1

1

2

D Head , J Bennett , P Greenberg , T Hamblin , The Pediatric MDS Working Group (H Hasle, C Niemeyer, J Chessells, M Freedman, I Baumann, JB, DH)1 1 United States, 2United Kingdom Contact e-mail:[email protected] The FAB classification of MDS (1982) distinguishes refractory anemia (RA), RA with ring sideroblasts (RARS), RA with excess blasts (RAEB), RAEB in transformation (RAEB-T) and chronic myelomonocytic leukemia (CMML) using blast content, ring sideroblasts and peripheral monocytosis. It has gained widespread usage. A WHO-sponsored committee has recommended revision of this classification (preliminary publication, 1999). The WHO classification includes RA (with or without RS) and RAEB, subdivides RAEB using marrow blast percent, lowers the marrow blast percent for AML to 20%, adds 5q-syndrome and adds refractory cytopenia (RC) with multilineage dysplasia (RCMD). It separates CMML with juvenile myelomonocytic leukemia (JMML) and Ph-chronic myelogenous leukemia (Ph-CML) in a new category with mixed myelodysplastic and myeloproliferative features. It separates from MDS patients with a low marrow blast percent, but recurring translocations, e.g. t(8;21) or inv(16). However, some WHO changes have engendered debate: 1. CMML forms a spectrum of disease, from mild peripheral monocytosis with clinical findings similar to MDS, to a proliferative process with pronounced monocytosis. The spectrum shares genetic abnormalities and dysplastic morphology with MDS, with no evident means to subdivide the CMML group. Separation of hypoproliferative CMML from MDS (WHO) is problematic, but inclusion of proliferative CMML in MDS (FAB) is equally problematic. 2. Using a 20% blast threshold for AML (WHO) includes in AML cases with predominant ineffective hematopoiesis (i.e. MDS) rather than blast burden. A 30% threshold (FAB) is also imprecise. A better approach to diagnose AML in this context may be serial marrows to demonstrate rapidly increasing blasts. 3. RCMD is conceptually confusing to many observers, who require multilineage abnormalities to diagnose MDS. Disagreement centers on whether MDS should be considered a syndrome (with loose diagnostic criteria allowing inclusion of disparate entities) or a specific set of interrelated diseases (with more stringent diagnostic criteria). 4. Both WHO and FAB are of limited use in pediatric patients. A proposed Pediatric MDS Working Group Classification has categories of RC and RAEB and distinguishes AML from MDS by rapid progression of marrow blasts or a very high blast percent. It separates JMML from MDS, similar to WHO, placing it with CMML and h-CML (both uncommon in childhood). It segregates Down syndrome patients and removes disease with recurring translocations from MDS. It also requires notation of preceding events (e.g. congenital marrow failure syndromes, aplastic anemia, familial MDS). Discussants will debate the relative merits of each classification.

Or2 SUBCLASSIFICATION OF REFRACTORY ANEMIA: JAPANESE VERSUS WHO PROPOSAL Y Yoshida1, S O Oguma2 Department of Hematology, Takeda General Hospital, Kyoto, Japan, 2Department of Medicine, Kiboukan Hospital, Takasaki, Japan Contact e-mail:[email protected]

1

Refractory anemia (RA) is the most heterogeneous disorders among the MDS. Some cases show atypical features, such as marrow hypocellularity, lack of megakaryocytes and minimal dysplasia, while other cases have predominantly low white cell and/or platelet counts with minimal or normal erythroid counts. We have previously proposed subclassification of RA into typical RA (123 cases), RA with acquired a megakaryocytic thrombocytopenia (AATP, 27 cases), minimal dysplasia RA (MD-RA, 19 cases), marrow hypocellular RA (hypo-RA, 35 cases) and refractory cytopenia (RC, 47cases). The criteria has been reported elsewhere (IJH 1999;70:65– 67) and the definition of cytopenia was similar to that of IPSS, except for RC in which levels of cytopenia were more strict. Comparison of this classification with the new WHO proposal revealed that, all but MD-RA showed multilineage dysplasia, qualifying the WHO criteria of RA with multilineage dysplasia. Of note was \90% of typical RA falling into RA with multilineage dysplasia. Cases with multilineage dysplasia had significantly higher marrow blast percentage, marrow monocyte percentage, serum LDH and lysosome, while their hemoglobin and white cell count were lower. Those cases showed a significantly higher rate of leukemic transformation and shorter overall survival. The results reinforce the splitting of RA into those with and without multilineage dysplasia (For the Japanese Collaborative Study Group).

Or3 MDS/AML: MODELS OF GENETIC PROGRESSION AND CLUES TO ETIOLOGY C Willman University of New Mexico, Albuquerque, USA Contact e-mail:[email protected] MDS and AML are a heterogeneous group of clonal neoplastic disorders with variable biologic behavior and clinical severity; their common biologic and genetic features suggest that they are part of the same continuous disease spectrum. Current genetic models for MDS hypothesize the existence of cytogenetically undetectable initiating lesions in clonal hematopoietic stem cells that promote the acquisition of the ‘secondary’ cytogenetic abnormalities more traditionally associated with MDS (such as 3p-, 3q-, 5q-, 7q-, 12p-, − 17, − 18, 20q11– 12, + 8). Detailed cytogenetic and molecular mapping studies reveal that rearrangements and deletions involving these chromosomes are very complex and that multiple regions may contribute to the MDS phenotype or disease progression; two distinct regions are implicated on chromosome 7q (7q22 and 7q32– 34) and four on chromosome 5q (5q11, 5q13– q21, 5q31 and distal 5q33– 35). Genetic progression in MDS is also accompanied by mutations of genes such as p53, FLT3 or RAS; methylation of specific gene promoters; and in some cases by the reciprocal translocations and inversions more frequently associated with AML. While recent studies have demon-

0145-2126/01/$ - see front matter © 2001 Published by Elsevier Science Ltd. PII: S0145-2126(01)00057-1

S2

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

strated that mutations in specific genes (AML1, NF1 or genes mediating DNA repair) may predispose to the acquisition of secondary cytogenetic abnormalities and MDS, it is likely that such inherited genetic mutations will account for only a minority of MDS cases. Epidemiologic case-control and population-based studies suggest that the far more common ‘sporadic’ cases of MDS may arise as a result of cumulative environmental exposures in genetically predisposed individuals; in this case, the genetic predisposition is a multi-locus complex genetic trait such as a constellation of risk-promoting and risk-modifying alleles in genes mediating DNA repair or carcinogen metabolism (such as variants in NAD(P)H:quinone oxidoreductase (NQO1) or the glutathione S-transferases (GSTT1; GSTTM)). Although the hypothesis that MDS arises due to cumulative environmental exposures in genetically predisposed individuals is indeed attractive, these studies of natural human genetic variation and disease association are only in their infancy and will require large population-based cohorts.

Or4 MUTATION ANALYSIS OF THE DNA-DAMAGE CHECKPOINT GENE CHK2 IN MYELODYSPLASTIC SYNDROMES AND ACUTE MYELOID LEUKEMIAS W K Hofmann1, S Takeuchi1, C W Miller1, D F Hoelzer2, H P Koeffler1 1 UCLA School of Medicine, Hematology/Oncology, CSMC, Los Angeles, USA, 2University Hospital, Hematology and Oncology, Frankfurt/Main, Germany Contact e-mail:[email protected] Background: Checkpoint genes code for a family of proteins which sense DNA damage in eukaryotic cells. They play an important role in the control of the cell cycle. The human CHK2 is a homolog of the yeast G2 checkpoint kinases, known as CDS1 and RAD53. The CHK2 may be a tumor suppressor gene because it was found mutated in some individuals with the Li –Fraumeni syndrome. These cases had a normal, non-mutated p53 gene. Methods: We performed a mutational analysis of the 14 exons of the CHK2 gene using polymerase chain reaction-single strand conformation polymorphism (PCRSSCP) in genomic DNA extracted from 41 bone marrow samples from individuals with myelodysplastic syndrome (MDS) and 41 samples of acute myeloid leukemias (AML). DNA from aberrantly migrating bands was sequenced. Results: We found a novel G to C transversion resulting in a change from Ala to Gly at codon 507 of CHK2 in an MDS sample. Normal lymphocytes from the peripheral blood did not have the abnormality of the CHK2 gene and the p53 gene was not altered in this case. In addition, we demonstrated a polymorphism at codon 84 (A to G at nucleotide 252) of exon 1 of CHK2 in three of 41 MDS and three of 41 AML patients. Conclusion: The identification of a mutation at a potentially critical region of the CHK2 gene in MDS highlights the importance of cell cycle checkpoints in early leukemic alterations of hematopoiesis.

Or5 INVOLVEMENT OF THE LYMPHO-MYELOID STEM CELLS IN DIFFERENT CYTOGENETIC SUBGROUPS OF MYELODYSPLASTIC SYNDROMES

L Nilsson1, I A, strand-Grundstro¨ m1, I Arvidsson2, B Johansson3, P Hokland4, L Kjeldsen5, E Hellstro¨ m-Lindberg6, R Hast7, S E W Jacobsen1 1 Department of Stem Cell Biology, Institute of Laboratory Medicine, Lund, Sweden, 2Department of Pathology, Karolinska Institute,

Stockholm, Sweden, 3Department of Clinical Genetics, Lund, Sweden, 4Department of Hematology, Aarhuus, Denmark, 5Department of Hematology, Rigshospitalet, Copenhagen, Denmark, 6Department of Medicine, Division of Hematology, Huddinge, Sweden, 7 Department of Medicine, Karolinska Institute, Stockholm, Sweden Contact e-mail:[email protected] Purpose: The cell of origin in myelodysplastic syndromes (MDS) has been controversial. A recent study of MDS patients with trisomi 8 (+ 8) demonstrated that the hematopoietic stem cell (HSC) pool (CD34+Thy-1 +) was devoid of +8 cells, suggesting that the HSC in these patients might not be involved in the MDS clone (Blood 1998;89:2804). In contrast, we recently demonstrated that a lymphomyeloid (CD34 + CD38-) HSC is the origin of transformation in MDS patients with deletions of the long arm of chromosome 5 (5q-) (Blood 2000;96:2012). To address these apparent discrepancies, we have investigated the HSC pool in +8 patients with regard to phenotypic, cytogenetic and functional characteristics. Methods: CD34+ CD38-HSC were purified using advanced fluorescence activated cell sorting (FACS) and investigated for clonality with fluorescence in situ hybridization (FISH). In addition, functional in vitro and in vivo stem cell assays were used to verify the stem cell activity of purified CD34 + CD38- cells. Results: In six patients with + 8, a variable (range: 0 – 69%) fraction of CD34 +CD38-HSC were found to be involved in the MDS clone. In addition, one patient with both 5q- and +8 was investigated with dual FISH and although almost all HSC were 5q deleted, only a fraction were +8. Furthermore, no HSC with + 8 without 5q- were found, suggesting that acquisition of + 8 is secondary to the 5q deletion. Functional stem cell assays revealed a severe deficiency in the HSC pool for all investigated patients, also in +8 patients lacking + 8 in the HSC pool. As for 5qpatients, in MDS with 20q deletions, \90% of cells in the HSC pool carried the 20q deletion. Conclusions: A very high frequency of the lympho-myeloid stem cells in MDS patients with 5q and 20q deletions are involved in the MDS clone. Although in many cases of trisomi 8 patients, no + 8 cells can be detected in the HSC pool, functional and cytogenetic data support that the HSC are the primary targets for transformation also in this group of MDS patients. Thus, in most cases MDS appear to be a disease of HSC origin.

Or6 RESOLVING COMPLEX ABERRANT KARYOTYPES IN MDS AND AML EVOLVING FROM MDS WITH 24-COLOR-FISH: A STUDY OF 40 CASES C Schoch1, S Schnittger2, S Bursch3, W Hiddemann4, T Haferlach1 1 Internal Medicine III, Munich, Germany, 2Internal Medicine, Munich, Germany, 3Internal Medicine III, Munich, Germany, 4Internal Medicine III, Munich, Germany Contact e-mail:[email protected] A total of 554 patients with MDS (n =444) or AML evolving from MDS (n =110) were analyzed by chromosome banding analysis. According to the IPSS score, 337 (61%) had a good prognosis karyotype (normal: 291, −Y alone: six, del(5q) alone: 30, del(20q) alone: ten), 106 an intermediate (19%) ( + 8 alone: 34, +8 and one additional aberration: seven, del(5q) and one additional aberration: eight, other: 57) and 111 a poor prognosis karyotype (20%) (complex: 89, abnormalities of chromosome 7: 22), respectively. In order to characterize chromosome abnormalities within complex aberrant karyotypes in more detail, 40 cases were studied with 24-color-FISH and loci specific probes for 5p15, 5q31, 7q31 and for the p53-gene. The number of chromosomes involved per case varied from three to 18 with a median of ten abnormalities per case. Altogether, 340 abnormalities were observed in 40 cases. A gain of a whole chromo-

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22 some was observed 24 times with + 8 (n= 5), + 22 (n= 3) and +9 (n= 3) being the most frequently gained chromosomes. A loss of a whole chromosome occurred 103x. The chromosomes most often lost were chromosomes 7 (n = 7), 18 (n= 4) and Y (n= 4). Structural aberrations, leading to a gain of chromosomal material, were detected only 33x, while a loss of chromosomal material was found 138x. Balanced aberrations, including only two chromosomes, were found only 5x. Losses due to structural abnormalities involved most frequently 5q (n =34), 17p (n =14), 7q (n=12) and 12p (n =9), while gains were observed of 21q (n =5), 11q (n= 5) and 1q (n =3). Using locus specific probes, a deletion of the EGR1-locus (5q31) was observed in 34 cases, of 7q31 in 17 and of the p53-gene in 14 patients, respectively. While a loss of 5q occurred due to deletion or due to unbalanced rearrangements in 17 patients each, a loss of a complete chromosome 5 was not observed. In contrast, a loss of 7q31 was found due to a loss of a whole chromosome 7 in seven cases, due to unbalanced rearrangements in six and due to simple deletion in four cases. Comparing data of this cohort with 89 patients with de novo AML studied the same way, a similar pattern of chromosomal gain and loss was observed. 24-color-FISH identified unbalanced translocations as a major mechanism for loss of chromosomal material that seemed to be simple deletions according to G-banding analysis. In conclusion, loss of chromosomal material rather than gain was more frequently observed. Therefore, it can be hypothesized that in MDS and AML with complex karyotype, a loss of tumor-suppressor genes is a more important mechanism of leukemogenesis than activation of oncogenes. As the pattern of chromosomal losses is non-random, molecular studies focussing on deletion hot spots must clarify the pathogenesis of this very poor prognostic subgroup of MDS/AML.

Or7 PHYSICAL AND HEMATOPOIETIC TRANSCRIPT MAP OF A 5q31 ‘CRITICAL SUBREGION’ ASSOCIATED WITH THE 5q-SYNDROME V Konstantinopoulou1, S Kamakari1, I Ragoussis2, N P Anagnou1 1 Institute of Molecular Biology and Biotechnology and University of Crete School of Medicine, Heraklion, Greece, 2United Dental and Medical Schools of Guy’s and St. Thomas’s Hospital, London, UK Contact e-mail:[email protected] The 5q-syndrome is a distinct type of MDS characterized by refractory anemia, abnormalities of megakaryocytes and an interstitial deletion (5q23 – 31), as the sole cytogenetic abnormality. Independent studies have identified four critical subregions that seem to be involved in leukemogenesis. As a first step to delineate the role of one of these subregions, we constructed a YAC contig encompassing the region between the GM-CSF/IL-3 and TCF-1 genes. Extensive PCR screening of the CEPH and ICI YAC libraries, resulted in the isolation of 12 YACs positive for either GM-CSF/IL-3 genes and/or the TCF-1 gene and of four YACs positive for seven STSs. STS mapping resulted in the construction of the first complete YAC contig of this subregion and documented that the GM-CSF/IL-3 and TCF-1 genes are linked. In order to develop a hematopoietic transcript map of the subregion, 26 chromosome 5-specific ESTs were selected on the basis of their origin and assignment. PCR analysis and in silico searches, resulted in sublocalization of nine of these ESTs to the identified YACs and documented their expression in a bone marrow cDNA library. Furthermore, by exploring the recently available genomic sequence information derived from nine PAC clones of the region (http://www.jgi.doe.gov/prod/RESEQ.html), we employed a comprehensive set of sequence analysis tools to electronically assess the presence and the location of transcription units. Our analysis has identified several novel putative transcript units, which were further investigated, utilizing RT-PCR and RACE techniques along with a novel method for efficient cloning of full-length cDNAs, using circu-

S3

larized first strand cDNA and two gene specific primers to amplify both the 5% and 3% cDNA ends in one reaction. The data from these approaches are expected to provide a systematic physical and transcript hematopoietic map of this ‘critical subregion’ involved in the early stages of leukemogenesis.

Or8 PROGNOSTIC FACTORS IN AUTOLOGOUS STEM CELL TRANSPLANTATION (ASCT) FOR PATIENTS WITH MDS AND AML FOLLOWING MDS T De Witte UMC St Radboud, Hematology, Nijmegen, Netherlands Contact e-mail:[email protected] Chemotherapy followed by ASCT is being increasingly applied as an alternative therapy for young patients with MDS or sAML lacking a suitable donor. We analyzed the results of 266 patients with MDS/ sAML transplanted with autologous stem cells. The 3-year DFS of the 185 patients transplanted in first CR was 32%. Treatment failure was mainly due to a high relapse risk of 59%. Nonrelapse mortality was 22%. The results of the 81 ASCTs performed beyond CR-1 were significantly worse: the DFS was 14%, the relapse rate: 72% and TRM: 50%. Age had a significant impact on treatment outcome. The DFS of patients transplanted in CR-1 was 50% when they were B 20 years. The DFS was significantly (P =0.02) lower at older age: 36% between 20 and 40 years and 22% \ 40 years. This difference is due to a higher TRM in the older age group: 33% in the group \ 40 years and 12% in the group B 40 years. In addition, a trend to a higher relapse rate in patients \ 40 years contributed to the poorer DFS in this age group: 43% when B20 years, 67% when \40 years. Recently, peripheral stem cells have replaced bone marrow (BM) as the source of stem cells for autologous transplantation in MDS. Some 112 patients in CR-1 received BM and 58 patients PSC. The DFS relapse incidence and TRM were no different between the two groups. The DFS was 36% when the interval between diagnosis and CR-1 was B 3 months and 15% when the interval was \3 months. The stage of disease at start of chemotherapy (RAEB(t) versus sAML) did not influence the outcome. Forty-two patients had primary MDS and 29 patients had post-cytotoxic MDS. The outcome after ASCT was completely identical for both groups. The results of ASCT performed in patients with MDS are gratifying, especially when the transplant is performed for a young patient in first CR and when treatment is initiated within 3 months of diagnosis.

Or9 CLONALITY DURING INTENSIVE CHEMOTHERAPY AND AUTOLOGOUS TRANSPLANTATION IN MYELODYSPLASTIC SYNDROMES M Delforge University Hospital Leuven, Hematology, Leuven, Belgium Contact e-mail:[email protected] The natural disease course of MDS patients with increased blast counts and/or unfavorable cytogenetics is infaust. Allogeneic stem cell transplantation is the treatment of choice, but is limited to a small subpopulation of young patients with histocompatible donors. With intensive AML-like chemotherapy regimens, remission rates in MDS vary between 15 and 65%, but relapses are more common compared to primary AML. Autologous stem cell transplantation (ASCT) has been proven to be a valuable consolidation treatment in MDS with a disease-free survival of 33% at 3 years (de Witte, 2000). Since MDS is a monoclonal (MC) disease of an early hematopoietic

S4

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

progenitor cell, the intriguing question regarding the clonal nature of the hematopoietic progenitors in remission marrow, blood and autografts of MDS patients is raised. Conventional karyotyping, FISH and X-chromosome inactivation assays (XCI) are most frequently used to assess clonality in MDS. In a single center study (Delforge, 1998, 1999) we have performed clonality studies with cytogenetics and XCI in nine high-risk female MDS patients before and after intensive chemotherapy. In contrast to T-lymphocytes, monoclonality was detected with XCI in mature myeloid cells and immature and committed marrow progenitors from all patients. Five of nine patients obtained a polyclonal (PC) and cytogenetic remission. Clonality analysis on stem cell harvests revealed similar results as in remission marrow. This study also showed the feasibility of ASCT with PC stem cells. These and other preliminary data have proven that a polyclonal and cytogenetic remission in MDS can be achieved, but have also raised several other questions: what is the percentage of PC and MC remissions in MDS? do patients with PC remissions have higher stem cell yields? what is the relapse rate in PC versus MC remission after consolidation chemotherapy and ASCT? These questions are currently addressed by a multicentric randomized phase III trial (CRIANT, coordinator: Professor Dr T. de Witte) comparing autologous PBPCT with a second intensive consolidation course in patients with bad prognosis MDS and sAML. Currently, cytogenetic and XCI data of the CRIANT trial are being analyzed and interpreted. On behalf of the CRIANT coworkers, available clonality data will be presented and discussed.

Or10 AUTOLOGOUS STEM CELL TRANSPLANTATION (ASCT) AFTER INTENSIVE CHEMOTHERAPY (IC) IN MYELODYSPLASTIC SYNDROMES (MDS). UPDATED RESULTS OF A PROSPECTIVE STUDY. GROUPE FRANCAIS DES MYELODYSPLASIES (GFM) AND GROUPE GOELAMS S Ducastelle1, L Ade`s2, C Gardin2, H Dombre2, T Pre´ bet1, A Merlat2, E Solary3, D Caillot3, B Mahe´ 4, J P Jouet5, F Maloisel6, F Lefrere2, A Guerci7, A Stamatoullas8, P Casassus9, A M Stoppa10, B Desablens11, D Guyotat12, P Lepelley5, C Dumontet1, F Dreyfus2, P Fenaux5, E Wattel1 1 Lyon, France, 2Paris, France, 3Dijon, France, 4Nantes, France, 5 Lille, France, 6Strasbourg, France, 7Nancy, France, 8Rouen, France, 9 Bobigny, France, 10Marseille, France, 11Amiens, France, 12SaintEtienne, France Contact e-mail:[email protected] Background: IC in MDS and AML following MDS gives shorter CR duration than in de novo AML and a very low proportion of long-term survivors (10% at 4 years in our experience) is observed. IC followed by ASCT provides an alternative therapy for MDS patients who lack a histocompatible donor. Retrospective studies revealed an ASCT-related mortality ranging between 10 and 28% and a 3-year disease free survival (DFS) of 33% (Br J Haematol 2000;110:620 – 30). Little remains know about the prognostic factors of ASCT. In the EBMT experience, patients B 40 years had a significantly better DFS than older patients (Blood 1997;90:3853 –7). In a recent prospective study, we reported that ASCT could be performed in \ 60% of patients who achieve CR with IC and that peripheral blood stem cells (PBSC) collection yielded higher numbers of stem cells than marrow collection (Leukemia 1999;13:524 –9). With a median follow-up of 9 months, gender was the only prognostic factor of the outcome of autograft. Methods: An updated analysis of this prospective study was performed at the reference date of January 15, 2001. Results: Between January 1993 and April 2000, 43 patients with high risk MDS in first CR after IC received autologous bone marrow transplantation (ABMT, 16 patients) or autologous peripheral blood stem

cell transplantation (APSCT) (27 patients). The median age was 52 years (range: 18 – 66). There were 26 MDS-AML, ten RAEB-T and seven RAEB. Bone marrow and PSC harvest were performed 0.5 –5 months (median: 2.3 months) and ASCT 1 – 7 months (median: 3.7) after achievement of CR, respectively. The conditioning regimen combined cyclophosphamide and busulfan in 29 patients and cyclophosphamide, busulfan and etoposide in 14 patients. Four patients (9%) died from the procedure. Hematological reconstitution occurred in the remaining 39 patients without significant differences between the two conditioning regimens. As previously described, reconstitution was faster after APSCT than after ABMT. Twenty-four patients (56%) relapsed after 1 – 36 months and 15 (35%) were still in CR after 2 – 92 months. The Kaplan – Meier (KM) estimate of overall survival was 8695% at 6 months, 74 9 7% at 12 months, 44 9 8% at 24 months and 3598% at 36 months. The KM estimate of DFS was 72 97% at 6 months, 53 98% at 12 months, 40 9 8% at 24 months and 31 9 8% at 36 months. Median KM DFS and survival were 16 and 22 months after autograft, respectively. In the 34 successfully karyotyped patients, the estimated 2-year survival rates were 50, 42 and 33% for patients in good (n =15), intermediate (n = 13) and poor risk (n = 6) IPSS cytogenetic subgroups, though the difference was not significant (P =0.273). In Cox univariate analysis, gender was the only prognostic factor of the outcome of autograft: males had both shorter DFS (median not reached in females versus 9 months in males, P =0.0290) and shorter overall survival (median 42 months in females versus 15 months in males, P= 0.0096) after the autograft. Other factors such as age, FAB classification, time between diagnosis and treatment, interval between CR and ASCT, WBC count, hemoglobin level, platelet count, absolute number of circulating blasts, percentage of bone marrow blasts and cytogenetic abnormalities (normal versus abnormal) at the time of induction chemotherapy did not significantly influence the overall outcome. Of the 23 patients autografted before 1997, with sufficient follow-up, seven (30%) have survived \4 years after the autograft. Six of these were alive in first CR after 60 + to 92 + months and probably cured, one relapsed 28 months after ASCT and achieved a second prolonged CR (23 + months) after IC. All six long-term survivors were female (versus four of the 17 patients who survived B4 years, P =0.0012) and were characterized by a higher incidence of normal karyotype (P=0.036). Conclusion: The present results confirm that gender is a strong prognostic factor in autografted MDS patients and suggest that ASCT could increase CR duration in some patients, by comparison with consolidation chemotherapy.

Or11 DO PATIENTS WITH LOW-GRADE MYELODYSPLASTIC SYNDROME MOBILISE NORMAL HEMOPOIETIC PROGENITORS? A Mijovic1, M Delforge2, M Sekhavat1, B Czepulkowski1, G J Mufti1 1 King’s College Hospital, Hematology, London, UK, 2University Hospital, Hematology, Leuven, Belgium Contact e-mail:[email protected] Methods: We have studied haemopoietic progenitor cell (HPC) mobilization in 16 patients with low-grade myelodysplastic syndrome (MDS) (IPSS score 0 – 1). Eleven patients had refractory cytopenia (two with hypocellular BM), three had 5q-syndrome and two had sideroblastic anemia. None had prior cytotoxic treatment. HPC were mobilized with G-CSF 10 mcg/kg for 5 days. Successful mobilization was defined as the collection of \2 × 106/kg CD34+ cells. Cytogenetics and in vitro cell culture were carried out on all HPC collections; clonality of individual CFU-GM and BFU-E colonies grown from blood progenitor cell (BPC) harvests was tested in females informative at the HUMARA locus of the X-chromosome. Results: In eight of 16 (50%) patients \2 ×106 CD34+ cells/kg could be harvested; in two additional patients between 1.5 and 2.0 × 106

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22 CD34+ cells/kg were collected. Both patients with hypocellular MDS failed mobilization. Of six patients with an abnormal bone marrow karyotype, four had the abnormality in 50 – 87% of harvested cells; this included all three patients with 5q-syndrome. Two of six patients had only cytogenetically normal cells in BPC harvests. However, in these patients the abnormal clone was present in a minority of mitoses (12 and 25%) at diagnosis. Of 12 evaluable patients, five displayed an abnormal in vitro growth pattern, i.e. low CFU-GM ‘output’ (0.15 –1.70%, normal 7.07 9 4.04% of CD34 + cells), skewed colony size distribution and low/absent BFU-E growth. In total, only three of 16 patients (19%) mobilized sufficient, qualitatively adequate HPC. An additional two patients nearly attained desired CD34 + cell counts in two aphereses. Of five females tested, two were informative at the HUMARA locus. 29 and 33%, respectively, of the CFU-GM and 24% BFU-E colonies grown from their BPC harvests expressed the allele not expressed by the dominant clone, suggesting that these progenitors were derived from residual non-clonal stem cells. Conclusions: Our results indicate that in a majority of low-grade MDS patients, G-CSF mobilizes insufficient HPC; furthermore, mobilized HPC are often cytogenetically and functionally abnormal. However, the demonstration of residual normal cells in the BPC harvests could prompt novel approaches to enhance mobilization or expand normal HPC.

Or12 APOPTOTIC MECHANISMS PLASTIC SYNDROMES

IN

MYELODYS-

S Mundle Rush Cancer Institute, Chicago, USA Contact e-mail:[email protected] Excessive spontaneous intramedullary apoptotic death in the bone marrow (BM) has been well established as a salient pathobiologic feature underlying inefficient hematopoiesis in the majority of patients with myelodysplastic syndromes (MDS). Since stem cell disorder is suspected in MDS, the propensity of CD34 positive progenitor cells to undergo apoptosis is studied widely, albeit with discrepant conclusion. The conflicting results may likely indicate an existence of a spectrum of susceptibility to apoptosis in the subclones of the original MDS stem cell constituting hematopoiesis in these patients. The studies using separated CD34 positive and negative fractions of BM aspirate however, suggest that at least in some MDS patients the apoptosis inducing activity may reside in the CD34 negative (CD34neg) maturing/matured cells. The BM microenvironment in MDS patients shows increased numbers of different immune cells such as macrophages and mast cells, NK cells, subsets of T cells, B cell aggregates, etc. and their inflammatory cytokines, such as TNFa, IL1, IL6, TGFb, etc. Although the exact nature of the apoptosis inducer still awaits confirmation, the final pathway involves TNFa triggered, Fas-mediated signal that is transduced intracellularly as a result of either rapid turnover or a lack of the negative regulator; Fas associated phosphatase-1 (Fap-1) and via caspase cascade. Our group has demonstrated that an additional abnormality in a transcription factor; E2F1 shifts the otherwise cell-cycle independent apoptosis to S-phase (signal antonymy). Differential susceptibility to apoptosis of the MDS subclones may be related to varying stoichiometry of the antiapoptotic proteins like Fap-1, bcl-2, etc. The relatively non-susceptible cells seem to differentiate and contribute to the peripheral blood cell pool. It remains to be seen if the leukemic transformation would occur due to a secondary event blocking differentiation on the background of non-susceptibility to apoptosis. The real therapeutically relevant challenge in the arena of apoptosis in MDS thus lies in unraveling the origin and basis for TNFa/Fas mediated cytopathic effects and also in finding the determinants of differential susceptibility to apoptosis in MDS subclones.

S5

Or13 POTENTIAL ROLE OF APOPTOSIS-MODIFYING TREATMENT FOR MYELODYSPLASTIC SYNDROMES P L Greenberg Stanford Medical Center, Hematology, Stanford, CA, USA Contact e-mail:[email protected] Increased apoptosis, a major pathogenetic mechanism underlying myelodysplastic syndromes (MDS), contributes to the ineffective hemopoiesis and peripheral cytopenias present at early stages of MDS. Conversely, excessive survival/decreased apoptosis of progressing abnormal clonal cells occurs upon evolution of MDS into AML. Mechanisms underlying this changing cellular survival profile relate to both intrinsic (i.e. processes occurring within the stem cells) and extrinsic (within the marrow stroma), generally paracrine, signaling abnormalities. The intracellular apoptotic-related phenomena are associated with switches within the hemopoietic precursors from an increased ratio of pro-apoptotic:anti-apoptotic oncoprotein/gene expression into a reversal of this ratio. The net balance between stimulatory and inhibitory cytokines present within the marrow stromal microenvironment contributes to modulating hemopoietic cell survival. Such abnormalities enhance the intracellular activation (or protection against) apoptosis-generating proteases, which underlie much of the pathogenesis of this disorder. As microenvironmental imbalance relates to suboptimal levels of certain hemopoietic stimulatory factors (HSFs), treatment has included their provision (e.g. G-CSF, IL3, erythropoietin), with variable degrees of benefit related to the patients’ clinical features (marrow status, transfusion need and endogenous HSF production). To diminish the apoptotic effects of increased levels of certain inhibitory factors (tumor necrosis factor a/TNF, vascular endothelial growth factor/VEGF), treatment with TNF-receptor-fusion protein and angiogenesis inhibitors (thalidomide, Hu McAb-VEGF) have been used. Immune-mediated apoptosis in a portion of MDS patients (often hypocellular MDS with evidence of a PNH clone) has been treated with immune modulators (ATG, cyclosporin). Encouraging Phase I/II clinical trials with the above agents, alone or combined, have been performed, predominantly for early stage disease. For more advanced MDS stages, treatment aimed at enhancing apoptosis of the neoplastic clone has been employed, including anti-ras agents (e.g. farnesyl transferase inhibitors) and chemotherapy. Data from these investigations will be reviewed. Ongoing risk-stratified Phase III trials should provide much useful data to determine the subsets of patients who may benefit clinically from such biologically specific therapeutic approaches. These studies will also aid our understanding of the relative contribution of specific biologic derangements to the disordered hemopoiesis of MDS.

Or14 PROAPOPTOTIC ACTIVITY OF T-CELLS IN MYELODYSPLASTIC SYNDROMES (MDS) R Neuwirtova1, J Karban1, H Krejcova2, J Bartunkova3, M Plasilova2, A Jonasova4, J Cermak2 1 Department of Clinical Haematology, First School of Medicine, Charles University Hospital, Prague, Czech Republic, 2Institute of Hematology and Blood Transfusion, Prague, Czech Republic, 3 Department of Clinical Immunology, First School of Medicine, Prague, Czech Republic, 4Dana Farber Institute, Boston, USA Contact e-mail:[email protected] Certain MDS patients benefit from the therapy either with ATG or cyclosporin A. Both drugs suppress T lymphocytes. Premature apoptosis of hemopoietic cells in MDS is now generally accepted. We try to answer two questions: (1) do T-cells play a role in apoptosis

S6

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

induction in MDS? and (2) is the fact that T-cells are frequently polyclonal and myeloid cells monoclonal of any significance for apoptosis inducing effect of T lymphocytes? 1. Published data are bringing evidence for proapoptotic activity of T cells in MDS: an increased colony formation in MDS bone marrow (BM) cultures depleted by T-cells or due to an addition of anti-TNFa antibody, an elevation of cytokines produced by activated T-cells as TNFa and Fas-L in MDS BM. We found an increase of the percent of CD8 + cells in the blood of 59 RA + RARS patients in comparison to controls (P= 0.01) with the exception of RARS women. Cocultivation of BM of 11 RA with TRAIL (TNF family member), which mediates T-cells cytotoxicity, reduces number of CFU-GM colonies (P= 0.003). The analysis of our 16 LGL leukemia patients demonstrates that cytopenia in LGL leukemia might serve as a model for the role of T-cells in MDS. 2. We have examined the clonality of myeloid cells and T lymphocytes in 25 women with MDS (using ioduronate-2suphatase polymorphism gene method). Among 17 women with totally monoclonal hemopoiesis only three patients had monoclonal T-cells. With passage of time, polyclonal T-cells are replaced by monoclonal cells: four of nine patients, whose T-cells were originally polyclonal had, after 12 – 22 months, monoclonal T-cells. Polyclonal T-cells in MDS either derived from remaining normal stem cells or they are long living T naive and memory cells. We try now to demonstrate with relevant antibodies (CD4+CD45RO and CD62Lhi, CD4+ 45RA and CD62Llow and CD8+ HLADR+CD11c) what imunophenotype polyclonal T-cells have and to find out whether these cells are activated by the contact with myeloid dysplastic cells, recognizing them as pathologic. The behaviour of monoclonal T-cells is a further line of inquiry. We conclude that polyclonal activated cytotoxic T-cells in MDS are probably accomplishing their physiologic function, which is to eliminate potentially malignant cells. If premature apoptosis is an event inherent to MDS, then T-cells induced apoptosis is a secondary event, which further worsens ineffective hemopoiesis in MDS. The implications for an immunosuppressive therapy of MDS are not straightforward.

Or15 REDUCED EXPRESSION BUT FUNCTIONAL INTEGRITY OF a4b1 AND a5b1 INTEGRINS ON HEMATOPOIETIC PROGENITORS OF MDS PATIENTS WITH HIGH LEVELS OF INTRAMEDULLARY APOPTOSIS M Delforge1, V Raets2, V Van Duppen2, G Verhoef1, M Boogaerts1 1 University Hospital Leuven, Hematology, Leuven, Belgium, 2 Catholic University Leuven, Hematology, Leuven, Belgium Contact e-mail:[email protected] Increased intramedullary apoptosis is frequently observed in patients with myelodysplastic syndromes (MDS). Currently, it is not known if a decreased expression and/or reduced function of b1 integrins contributes to apoptosis and impaired proliferation of MDS progenitors. Since adhesion-mediated signal transduction by a4b1 and a5b1 integrins is important in the control of the cell cycle of normal hematopoietic progenitors, we have studied the expression levels and anti-apoptotic function of these integrins on MDS progenitors. Expression levels of a4, a5 and b1 were measured by FACS analysis on freshly isolated and purified CD34 + marrow progenitors of 22 MDS patients (six RA, four RARS, five RAEB, four RAEB-t and three sAML). Viable, apoptotic and death progenitors were discriminated based on their levels of Annexin V and 7AAD binding. A significantly reduced expression of b1 (P =0.02), but not a4 or a5, was found when

comparing non-fractionated CD34 + cells from all patients with a group of healthy control subjects. Additional analyses on viable (Annexin-7AAD-) and apoptotic (Annexin +7AAD-) cells revealed that high levels (at least 25%) of intramedullary apoptosis were also associated with a significant decreased expression of a4 (P =0.04) and a5 (P =0.01) on the remaining viable progenitor cells. Using standardized functional adhesion assays, this reduced integrin expression did not result in altered adhesion to intact stroma or fibronectin. Finally, the influence of reduced b1 integrin expression on progenitor survival was assessed in ten patients by growing the CD34 + progenitors in serum-free medium for 72 h in contact with 33/66 kDa (a4b1 ligand) and RGDS (a5b1 ligand) fibronectin fragments, or separated from these ligands by a microporous membrane. These experiments revealed that, both in patients with low and high levels of intramedullary apoptosis, engagement of b1 integrins on CD34 + progenitors by fixed ligands resulted in apoptosis inhibition. We conclude from these experiments that CD34 + progenitor cells from MDS patients with high levels of intramedullary apoptosis have decreased expression of a4b1and a5b1 integrins but a normal adhesive and survival-promoting function of their residual b1 integrins.

Or16 CD95 EXPRESSION ON CD34 + MYELODYSPLASTIC SYNDROMES

CELLS

IN

L Tartuferi1, L Depaoli 1, M C Sacchi1, S Riva1, M Ciriello2, T Callegari2, F Salvi1, C Isidoro3, A Levis1 1 Department of Haematology, SS Antonio Biagio Hospital, Alessandria, Italy, 2SS Antonio Biagio Hospital, Clinical Pathology, Alessandria, Italy, 3Amedeo Avogadro University, Immunology, Novara, Italy Contact e-mail:[email protected] Background: Ineffective hemopoiesis, typical of MDS, is supposed to result from an excessive cellular death rate charged to hemopoietic progenitors, thus avoiding their differentiation and entering the bloodstream. An increased and early apoptosis would delete, in MDS, the effect of the high proliferative activity and would justify the common pattern of peripheral pancitopenia. Patients: We evaluated the expression of the cell-surface receptor Fas (CD95) in different marrow populations of 30 patients affected by MDS, in comparison to the marrows from five normal controls in consideration of its fundamental role in the activation of the apoptotic pathway. Results: Flow cytometric analysis, executed on bone marrow mononuclear cells (BMMNC), immediately after separation on density gradient, firstly allowed us to point out that MDS are characterized by a constitutive increased expression of CD95 in both all BMMNC (25% in controls versus 44% in MDS; P value B 0.01) and in CD34 + population (1% in controls versus 5% in MDS; P value =0.01). Moreover the CD95 evaluation, performed after 18 h (t18) of incubation in RPMI 1640 added with 1% glutamine, 1% PES and 10% FCS, allowed us to point out that normal marrows are characterized by a slight non-significant spontaneous increase in CD95 expression in both all BMMNC (t0 25% versus t18 28%; P value =n.s.) and CD34 + population (t0 1% versus t18 2%; P value =n.s.). On the contrary, in MDS patients, even if the difference of CD95 expression in BMMNC was not evident from t0 to t18 (t0 44% versus t18 45%; P value = n.s.), a significant increase was shown when the analysis was limited to the CD34 + population (t0 5% versus t18 7%; P value= 0.01). Conclusion: Our data lead to suppose that the marked proneness to enter the apoptotic pathway, shown by MDS cells and particularly by CD34 + population, may be due to the fact that MDS costitutively express CD95 values higher than normal marrows. Since CD95 expression spontaneously increases only in MDS CD34+ population, we are performing further studies based on treatment of MDS and normal mononuclear cells with cytokines as IFN-g and TNF-a, whose seriological levels are increased in MDS patients, intended to a better comprehension of the exact role of CD95 in CD34+ apoptotic pathway and its modulation.

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

Or17 IN VITRO AMPLIFICATION AND DIFFERENTIATION OF MYELODYSPLASTIC SYNDROMES DERIVING ERYTHROID PROGENITORS Y E Claessens1, D Bouscary1, J M Dupont2, J Melle3, F Picard3, M Guesnu3, S Gisselbrecht1, C Lacombe1, F Dreyfus1, P Mayeux1, M Fontenay-Roupie1 1 ICGM INSERM U363 Hoˆ pital Cochin, Hematology, Paris, France, 2 Hoˆ pital Cochin, Cytogenetics, Paris, France, 3Hoˆ pital Cochin, Hematology, Paris, France Contact e-mail:[email protected] Background: Normal erythropoiesis results from the proliferation and differentiation of pluripotent stem cells in erythroid progenitors, i.e. burst/colony-forming unit-erythroid (B/CFU-E). Erythropoietin (Epo) permits survival and growth of late BFU-Es and their terminal differentiation in erythrocytes. Myelodysplastic syndromes (MDS) commonly exhibit a dyserythropoiesis which pathophysiology is still unclear. Methods: MDS, (RARS = six, RA =seven, RAEB B10% blasts= seven) versus normal marrow (n=5)-deriving CD34+ cells were allowed to amplify and differentiate along the erythroid lineage in liquid culture in the presence of serum, SCF and Epo. In a second stage, terminal differentiation was obtained in the presence of Epo alone. Results: In patients and controls, \80% cells were CD34− 71+36+ at d7 and CD71+36+GPA+ at day 18. Enucleated erythrocytes represented 1/3 of the cells at the end of culture. At day 18, the amplification factor was slightly lower in RA/RAEB (× 750) than in RARS and normal marrows (× 1000). However, the number of BFU-E at day 7 was significantly lower in RARS (500/105) and RA/RAEB (350/105) compared to controls (1000/105). Detection of chromosomal abnormalities by FISH analysis of interphase nuclei confirmed that abnormal clones were amplified:

RARSc 1 (−Y) RAEBc 2 (−Y) RAEBc 3 del(4p)

Day 7–8 (%)

Day 17–18 (%)

24 75 35

80 95 75

Epo receptor signalling (STAT5 DNA binding, Akt and ERK1/2 phosphorylation) was normal at day 12. Apoptosis markers (Dym, phosphatidylserine exposure) was equivalent in MDS and normal CD34+ cells at day 0 and significantly increased ( × 2– 3; PB 0.05) in MDS from day 12. Marked Fas/CD95 expression was detected from day 3 in MDS but not in normal controls. Conclusions: These results demonstrate that MDS CD34+ cells proliferated with a growth advantage for abnormal clones. Despite Epo-dependent transduction of anti-apoptotic signals was normal, erythroid progenitors exhibited increased features of apoptosis. This suggests that impaired BFU-E growth and also intramedullary abortion of the erythroid lineage could be related to an upregulation of pro-apoptotic signallings.

Or18 ERYTHROPOIETIN: BASIC ASPECTS-STRUCTURE, REGULATION, MECHANISM OF ACTION AND BIOLOGICAL EFFECTS M Mittelman Rabin Medical Center-Hasharon Hospital, Petah-Tikva, Sackler School of Medicine, Tel-Aviv, Israel Contact e-mail:[email protected] Erythropoietin (Epo) is a 30 –34 kDa-glycoprotein hormone, produced in the kidneys (90%, by a minor subset of cortical interstitial

S7

cells) and liver (10%) circulates to the bone marrow, to bind receptors (EpoR) on BFU-E and CFU-E. The Epo gene was cloned and mapped to ch 7 (7pter-22q), as a single copy gene. The cDNA contains 5 exons, coding for a 193 amino acid (AA) peptide. A 27 AA leader sequence is cleaved of the N-terminus, leaving the known 166 AA structure of Epo. Epo contains four cysteinyl residues with two disulfide bridges and four oligosacharides, which are biologically important. Epo is heavily glycosylated (35 – 40%). The C-terminus contains the biological activity. Using the genetic sequence, the secondary structure, with a-helix and a-pleated could be predicted. The tertiary structure consists of four helices bundled together with two overhand loops. The Epo gene is expressed in fetal liver, whereas after birth the kidney becomes the predominant site of synthesis. The physiological stimulus for Epo production is tissue hypoxia. A hypoxia-inducible enhancer in the Epo 3% flanking region, contains binding sites for several transcription factors, including HIF-1, a factor whose DNA-binding activity is induced by hypoxia. Binding of HIF-1 is required for Epo transcriptional activation. EpoR gene is located in ch 19 and belongs to the cytokine receptor superfamily. It encodes a 508 AA peptide (72 – 78 kDa), with a homodimer structure. The 8-tyrosine residues are phosphorylated by JAK2 after Epo stimulation and act as docking sites for signaling molecules. Signaling pathways include the JAK/STAT and ras/MAP kinase. JAK2 phosphorylates the STATs, which dimerize and move to the nucleus where they stimulate gene transcription. The ras/MAP kinase pathway is activated by a number of intermediates, which link up to SOS, which activates ras. Active GTP-ras localizes raf to the plasma membrane, where it becomes phosphorylated and activates MEK, which in turn activates the MAP kinases. Dephosphorylation and inactivation of EpoR and JAK2 is achieved via the phosphatase SHP-1, which binds to the negative regulatory domain of EpoR. Another Epo-induced negative regulatory protein is CIS, which binds to EpoR at a major STAT5 binding site and acts as a negative modulator of STAT5 activation. Epo serves as a survival factor for erythroid cells in the relatively later stages of differentiation, and thus, has been shown to be effective in the treatment of several types of anemia. However, Epo appears to have additional biological effects. Epo has been proposed for treatment of congestive heart failure. Recently, Epo has been shown to reduce the injury following cerebral ischemia by 50 – 75% in rodents and was shown to have additional neuroprotective effects. Several of our multiple myeloma patients, treated with rHuEpo for anemia, lived longer than expected. Thus, we hypothesized that Epo might have another effect, unrelated to the erythroid effect. In mouse models, we showed that Epo resulted in a regression of :50% of the myeloma and long-term survival ( \ 15 months) compared with control mice, who all progressed to death within 40 – 50 days. The mechanism of this unrecognized hitherto anti neoplastic effect of Epo appears to be immune mediated.

Or19 ERYTHROPOIETIN TREATMENT IN CANCER ANAEMIA G Birgega˚ rd Department of Hematology, University Hospital, Uppsala, Sweden Contact e-mail:[email protected] Anemia is one of the most common consequences of chronic disease, especially inflammatory and malignant disease. Recent studies show that doctors tend to underestimate the importance of symptoms of anemia in these patients, probably because there has been no effective treatment for the anemia other than successful treatment of the underlying disease and, in severe anemia, transfusion. There is no consensus concerning the trigger level for transfusions, but few doctors aim for a hemoglobin level higher than 100 g/l other than with concomitant cardiac disease. During the last 8 years, a large number

S8

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

of studies have shown that treatment with rHuEpo (epoetin) can increase the Hb level in patients with anemia of chronic disease (ACD). With a response criterion of a Hb increase with 20 g/l the frequency of response varies from : 60% in cancer anemia to : 80% in rheumatoid arthritis and inflammatory bowel disease. Generally speaking, patients with an intense inflammation are less likely to respond as are patients with progressive malignant disease. The influence of the endogenous S-Epo level is weak. Previous attempts to use the ratio between expected and observed S-Epo level to predict response have not been found to be clinically useful, partially because the basis for what is an ‘expected’ S-Epo level is weak, partly because these patients rarely have S-Epo levels above 200 U/l. Once it was shown that the Hb level can be increased by epoetin treatment of ACD, the next step was to study whether this meant any functional or subjective improvement for the patients. A number of non-randomized studies have included quality of life (QOL)-measurements and improvement in physical activity, mental well-being and overall QOL has been reported. In 1977, a new QOL-instrument (Fact-An) was published by Cella et al., which was specifically designed to measure QOL life changes in anemia. Three large community-based, non-randomized studies have confirmed the effect on QOL by epoetin treatment of cancer anemia (Glaspy et al., Demetri et al. and Gabrilove et al.). The influence of disease status has been investigated and it has been shown that the magnitude of QOL changes is dependent on the magnitude of the Hb rise, regardless of whether the cancer disease is improving, stable or progressive. Lately, two large randomized trials have confirmed the improvement i QOL (O8 sterborg et al. and Littlewood et al.). One may ask whether it will be possible — for ethical reasons — to perform further randomized QOL studies in cancer patients. Possibly this can be done in order to further clarify what the treatment goal, in terms of Hb levels, should be. An interesting finding has emerged from several of the recent studies: patients seem to benefit with QOL improvement even when a mild anemia is corrected and when severe and moderate anemia is corrected, QOL improvements do not stop at a Hb level of 100 g/l but continue into the normal range. This has also been found in patients with renal anemia. These results need confirmation in randomized trials. A different treatment strategy has been adopted in patients who undergo cytostatic treatment with many repeated cycles of chemotherapy, especially in patients with small cell lung cancer: prophylactic epoetin treatment is started at an early stage when the Hb level starts falling but before anemia has developed, in order to avoid anemia and transfusions. Risk models for severe anemia after cytotoxic therapy are being developed (Ray-Coquard et al.).

Or20 TREATMENT OF ANEMIA IN MYELODYSPLASTIC SYNDROMES WITH ERYTHROPOIETIN M Cazzola Department of Hematology, University of Pavia School of Medicine, IRCCS Policlinico S. Matteo, Pavia, Italy Contact e-mail:[email protected] Anemia is a major clinical problem in MDSs with many patients being adversely affected by transfusion-dependency and secondary hemochromatosis. Several phase I –II clinical trials have been conducted to test the usefulness of recombinant human erythropoietin (rHuEpo) in the treatment of anemia in MDS patients. In these studies, a variety of criteria were adopted to define response. Most frequently, however, this latter was defined as abolishment of transfusion requirement in transfusion-dependent patients or as an increase in Hb] 1.5 g/dl in patients without transfusion requirement. Overall, 15 –20% of patients responded to rHuEpo treatment, but the vast majority of responders were not transfusion-dependent and the doses required to achieve response were \450 IU/kg per week. Several

reasons may account for this limited efficacy of erythropoietin in myelodysplastic syndromes: the most likely explanation, however, is that in the majority of MDS patients endogenous erythropoietin production is adequate for the degree of anemia. Most studies, in fact, showed that the major factor predicting response to rHuEpo in MDS patients is a serum erythropoietin levels B100 mU/ml. Consequently, these reports point to the use of a serum erythropoietin threshold of 5100 – 200 mU/ml to predict response to rHuEpo. In the study by Stenke et al., all responders but one had serum erythropoietin levels lower than 100 mU/ml. Rose et al. treated 100 MDS patients, 28% of whom responded to rHuEpo. Most responders (24/28) had baseline erythropoietin levels 5100 mU/ml. Hellstro¨ mLindberg analyzed 205 MDS patients from 17 published reports; her meta-analysis confirms that baseline serum erythropoietin is crucial for response to rHuEpo, since no patient with serum levels \ 200 mU/ml benefited from rHuEpo. This was later confirmed in the Italian Cooperative Study, where serum levels \ 200 mU/ml predicted for a non-response. The Italian study also showed that, in addition to low serum erythropoietin, low-risk MDS subtype and no transfusions prior to rHuEpo treatment were associated with higher probability of response. MDSs are typical stem cell disorders, so that the typical anemic patient is expected to have a high serum Epo level, i.e. an appropriately increased endogenous Epo production. It is therefore unclear why some individuals show inappropriately low Epo levels. We have recently shown that serum Epo reflects a balance between renal production and erythroid consumption, suggesting that factors other than renal production may affect the erythroid hormone levels. In spite of the low proportion of responses, recognizing potential responders to rHuEpo can be extremely important in individual cases of MDS. We favor a patient-oriented approach to the use of rHuEpo where the physician carefully evaluates the individual patient’s needs and likelihood of response: such approach can be applied also to MDS patients.

Or21 TREATING THE ANEMIA OF MDS WITH ERYTHROPOIETIN + G-CSF E Hellstro¨ m-Lindberg Department of Hematology, Karolinska Institutet, Huddinge University Hospital, Stockholm, Sweden Contact e-mail:[email protected] The majority of patients with MDS suffer from anemia with or without a need for transfusions. Besides the complications and costs associated with long-standing transfusion need, anemia reduces quality of life and may probably also aggravate co-morbid conditions. Erythropoietin (epo) is essential for erythroid colony growth in vitro, but the effect may be enhanced by the addition of early-acting or myeloid growth factors. Several clinical trials have studied the combination of epo and G-CSF, GM-CSF or IL 3 for the cytopenia of MDS. The response rate to epo +G-CSF is : 40% in MDS patients with B10% bone marrow blasts, i.e. twice as much as for epo alone. A large American phase II study, the trials within the Scandinavian MDS group and more recently a Spanish study have demonstrated evidence of a clear synergistic effect between the drugs in vivo. The American study showed that 50% of the patients with a response lost it when G-CSF was withdrawn and regained it when G-CSF was reintroduced, and the Scandinavian study showed that addition of G-CSF could induce erythroid responses in epo-resistant patients. The synergistic effect was most pronounced in patients with RARS, since this subgroup respond poorly to epo alone and showed the best response to the combination ( :50%). These data are supported by recent laboratory findings showing an independent effect of G-CSF on erythropoiesis A recent German study showed that some patients develop late responses to treatment, appearing after several months.

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22 A study of predictive factors for a response to treatment identified S-epo and the level of pre-treatment transfusion need as significant variables and these were combined in a predictive model for treatment with G-CSF +epo. The value of this model was confirmed in an independent patient material from the Spanish study and in a third prospective study within the Scandinavian MDS Group. This study included 53 evaluable patients and clearly verified the value of the predictive model (P =0.004). Moreover, quality of life (QLQ 30) improved in patients with a response to treatment (P B 0.05 for fatigue and global QOL). The median duration of response was 23 months for all responders and 29 months for those with a complete erythroid response. A simplified and clinically useful decision model was developed based on these results. Thus, patients with S-epo B 500 U/l and a transfusion need B2 U per month have a predicted response rate to treatment with epo 70000 u/w+GCSF dosed to give ANC counts of 6 to 10 × 109/l of 74%. In conclusion, the combination of G-CSF + epo has a definite role in the treatment arsenal for MDS with anemia and in particular for RARS.

Or22 MDS IN CHILDHOOD AND ADOLESCENCE: WHEN TO TREAT AND HOW TO TREAT? C Niemeyer University Children’s Hospital, Freiburg, Germany Contact e-mail:[email protected] MDS in children and adolescents is a rare clonal hematopoietic disorder characterized by multilineage dysplasia and defined genetic abnormalities. It may develop in patients with congenital or acquired bone marrow failure disorders, previous chemo- or radiation therapy, or recognized syndromes (secondary MDS). About half of the cases arise in children without known predisposing conditions (primary MDS). Although there is a consensus that stem cell transplantation (SCT) is the only curative therapy for all patients with MDS, a number of critical questions remain unanswered. For patients with primary advanced MDS the role of AML-like induction therapy prior to SCT is controversial. We recently presented data that imply that AML-like induction therapy prior to SCT may not be warranted for the majority of children with advanced stage primary MDS. For patients with primary low grade MDS close surveillance is often the most appropriate form of therapy. Once transfusion dependency is reached, patients who have a matched family donor (MFD) should receive allogeneic SCT. For patients without a MFD, the question of a trial of immunosuppressive therapy prior to SCT may be raised. The optimal preparative regimen for patients with primary low grade MDS is still open to discussion; it can be less intense than that used for advanced disease. For all children with MDS secondary to chemo- or radiation therapy and for children with MDS after a diagnosis of aplastic anemia allogeneic SCT without prior intensive chemotherapy is the choice of therapy. Options for children with MDS after congenital bone marrow failure disorder are more diverse. In Kostmann’s syndrome it has been shown that G-CSF receptor mutations, ras mutations and monosomy 7 herald the development of MDS. Therefore, it may be advisable to perform SCT shortly the demonstration of these acquired molecular or cytogenetic abnormalities. In patients with Fanconi anemia and MDS options include varies forms of SCT with conditioning regimens tailored to accommodate the increased chromosomal breakage of somatic cells. Although SCT can cure a relevant number of children with primary or secondary MDS cooperative multi-national studies will be required to answer the above mentioned questions in a scientific manner.

S9

Or23 CLASSIFICATION, EPIDEMIOLOGY AND PROGNOSTIC FACTORS IN CHILDHOOD MDS H Hasle Department of Pediatrics, Skejby Hospital, Aarhus, Denmark Contact e-mail:[email protected] While the FAB classification has been generally accepted for adults, the classification of MDS in childhood has been the subject of some controversy. Recently, international consensus has been reached on some of the controversial issues. (1) Only clonal disorders with dysplastic morphological features should be considered MDS. (2) Juvenile myelomonocytic leukemia (JMML) is a distinct disorder. (3) Monosomy 7 is the most common acquired abnormality in children with MDS and JMML. The concept of monosomy 7 as a distinct syndrome has been abandoned. (4) Myeloid leukemia in children with Down syndrome (DS) is distinct from the disease in non-DS children. The group of myelodysplastic disorders in children may accordingly be divided into three main groups; JMML, myeloid leukemia of DS and MDS (de novo and secondary). The annual incidence of JMML, myeloid leukemia of DS and MDS is 1.2, 0.9 and 1.8 per million, respectively. Neurofibromatosis is known in 15% of JMML cases and has helped understanding the pathogenesis involving the RAS signaling pathway. Myeloid leukemia of Down syndrome has unique clinical and biological features and a very favorable prognosis when treated on AML protocols. MDS may occur ‘de novo’ or develop in a child with a known predisposing condition-‘secondary’ (after chemo- or radiation therapy, in inherited bone marrow failure disorders, following acquired aplastic anemia or as familial MDS). It is to be recognized, however, that children with so-called ‘primary’ MDS may have an underlying yet unknown genetic defect predisposing them to MDS at young age. The International Prognostic Scoring System (IPSS) does not provide prognostic information in JMML. Low platelet count, elevated HbF and age of 2 years or more are associated with a poor prognosis in JMML. According to IPSS, more children than adults are high-risk patients. Preliminary data on childhood MDS showed that 2 – 3 lineage cytopenia and BM blasts \5% correlated with poor survival, whereas cytogenetics did not provide prognostic information.

Or24 PROBLEMATIC ASPECTS IN BONE MARROW BIOPSIES OF CHILDHOOD MDS D Anagnostou1, S Polychronopoulou2, A Tasidou1, M Fameli1, I P Panagiotou2, S A Haidas2 1 Hematopathology Department, ‘Evangelismos’ General Hospital, Athens, Greece, 2Department of Pediatric Hematology/Oncology, ‘Aghia Sophia’, Athens, Greece The establishment of the EWOG-MDS has decisively contributed to international collaborative studies, opening thus new horizons in the diagnosis, management and registration of childhood MDS. However, in the series that have been reported BM biopsy (BMB) does not represent an integrated procedure of the bone marrow examination in pediatric MDS. Thus, the histological findings and especially the evaluation of cellularity and subtyping of childhood MDS have not been well established. Aim of this study: (i) a detailed histological and immunohistochemical analysis of BMB in childhood MDS at diagnosis and during follow up; and (ii) their correlation to: (a) clinical presentation; (b) BM and PB smears; and (c) adult MDS. Material and methods: 32 samples of formalin fixed, EDTA decalcified, BMBs sized 0.5 – 3.0 cm (median value = 1.05) from 19 pediatric patients with primary MDS (RA =12, RAEB/RAEB-t= 7) and four patients with JMML. Patients’ age ranged from 2 to 170 months (median value =54.5) and their follow up from 3 to 140 months

S10

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

(median value = 19.5). The histological parameters included overall cellularity and status of hematopoietic series (quantitative differences, dysplasia, ALIP, distortion of the architecture) using three degrees: mild, moderate and severe. In 12 of 23 patients, associated abnormalities were identified. Results: The BMB in: (1) RA presents a number of histological differences from the corresponding group in adult MDS. More specifically: hypocellularity of significant degree (moderate three of ten and severe six of ten) was a constant and predominant finding; hematopoietic series especially myeloid (ten of 12) and megakaryocytic (eight of 12) were decreased, moderately and severely, respectively; dysplasia was pronounced mainly in megakaryocytic (nine of 12) and erythroid (ten of 12) series; distortion of the architecture was not a constant finding. (2) RAEB/ RAEB-t and JMML did not present substantial differences from corresponding adult MDS-subtypes. In conclusion, it seems that according to BMB findings childhood MDS deserve reconsideration in regard to: (i) determination of RA; (ii) definition of hypocellular pediatric MDS; and (iii) investigation of predisposing significance of associated abnormalities in BM failure syndromes (six of 12 belong to RA).

Or25 PAEDIATRIC MYELODYSPLASTIC MES IN BRITAIN 1990– 1999

SYNDRO-

S J Passmore1, J M Chessells2, C A Stiller3, I M Hann2 1 Hospital for Sick Children Great Ormond Street, Oncology, London, UK, 2Hospital for Sick Children Great Ormond Street, Haematology, London, UK, 3Childhood Cancer Research Group, Oxford, UK Contact e-mail:[email protected] Aims: To establish incidence and outcome for pediatric myelodysplasia (MDS) in the population of Great Britain. Methods: Members of the UK Children’s Cancer Study Group and the Pediatric Haematology Forum of the British Society of Haematologists were asked to notify cases of possible MDS. Eligibility for the MDS register was decided on the basis of central slide review and data from a standard questionnaire. Follow-up information was collected annually. Results: 136 children had ‘Primary MDS’. Exclusions were: ten Down syndrome (DS) MDS, eight transient abnormal myelopoiesis (six with DS), two refractory Cytopenia (RC) in patients with Shwachman’s syndrome, two MDS in patients with Kostman’s syndrome, four MDS following treatment for severe aplastic anemia, two MDS following chemotherapy, one acute lymphoblastic leukemia and one ovarian teratoma. The crude annual incidence of primary MDS was 1.25 per 106 (ASR 1.47 per 106) Of the 136 children, five had neurofibromatosis type 1, six had cardiac anomalies, 16 had other abnormalities; eight had a family history of acute myeloid leukemia or platelet storage pool disorder. The subtypes of MDS were: juvenile myelomonocytic leukemia/chronic myelomonocytic leukemia (JMML/CMML) n =61, RC n= 36, refractory anemia with excess blasts (RAEB) n = 21, RAEB in transformation (RAEBt) n = 17 and refractory anemia with ringed sideroblasts n= 1; 47 children fulfilled the internationally agreed criteria for JMML. Cytogenetics was available for 129 patients: 64 showed no clonal abnormality at diagnosis, 23 had monosomy 7, a further six had loss of chromosome 7 with additional changes, eight had trisomy 8 and 28 had other abnormalities. Five-year survival (95% CI) for the whole group was 49% (40–58), for JMML 44% (30 –59). Prognostic factors affecting survival were platelet count ( B40× 109/l is worse) P= 0.01and fetal hemoglobin (HbF) for JMML (\ 10% is worse) P= 0.004. Conclusions: The outlook for children with MDS is improving. The prognosis still depends on platelet count and HbF at diagnosis.

Or26 GRANULOCYTIC SURFACE MARKER ABNORMALITIES ARE DETECTED ONLY IN PATIENTS WITH MYELODYSPLASTIC SYNDROME AND PATIENTS WITH SHWACHMAN DIAMOND SYNDROME BUT NOT OTHER BONE MARROW FAILURE SYNDROMES M T Elghetany1, J Martinez2, J Patel2, K Sackey2, B P Alter3 1 Department of Pathology, University of Texas Medical Branch, Galveston, Texas, USA, 2University of Texas Medical Branch, Galveston, Texas, USA, 3National Cancer Institute, Bethesda, Maryland, USA Contact e-mail:[email protected] Background: Patients with bone marrow failure syndromes may evolve into myelodysplastic syndrome (MDS), in which an abnormal pattern of granulocytic surface marker maturation may be identified using flow cytometry (FC). Methods: To test this hypothesis, granulocytes from bone marrow (BM) (and from some peripheral bloods (PB)) from 52 patients with BM failure were examined. The patients had the following diagnoses: Shwachman Diamond syndrome (SDS) 19, Fanconi’s anemia (FA) nine, Diamond-Blackfan anemia (DBA) six, MDS (all refractory anemia, RA) seven, acquired aplastic anemia (AA) six, dyskeratosis congenita (DC) four and thrombocytopenia with absent radii (TAR) one. The patients included 30 females and 22 males. The median age was 8 years (range: 9 months to 76 years). FC was performed using lysed whole blood technique and a panel of antibodies to the following granulocytic maturation stages (GMS): early GMS, CD13, CD33, CD15, CD43; intermediate GMS CD11b, CD24; late GMS, CD35, CD16; and very late GMS, CD87 and/or CD10. CD55 and CD59 were also added. FC results were correlated with appropriate age-matched control specimens and with morphologic examination of BM aspirates. Results: Surface marker abnormalities were detected in BM granulocytes in three of 19 patients with SDS (15%), one of whom had morphologic MDS and in six of seven patients with RA (86%). No surface marker abnormalities were seen in patients with FA, DBA, AA, DC and TAR and none of these was in MDS phase. The prevalence and distribution of granulocytic abnormalities in SDS and RA are summarized in table:

SDS RA

CD33 (%)

CD35 (%)

CD16 (%) CD87 (%) CD10 (%)

1/16 (6) 1/2 (50)

1/19 (5) 6/7 (86)

1/19 (5) 2/7 (29)

1/19 (5) 1/6 (17)

1/12 (8) 2/2 (100)

Four patients with RA and two patients with SDS (including the patient in the MDS phase) showed multiple abnormalities. Fifteen of 17 abnormalities (88%) were in the late and very late appearing antigens: CD35, CD16, CD87 and CD10. Conclusion: We conclude that an abnormal granulocytic surface marker maturation pattern is a common finding in RA and may provide an additional tool to support the diagnosis. This is the first report of granulocytic surface marker defects in SDS, which may be a part of the natural course of the disease or may be an early indication of the evolution to MDS.

Or27 PEDIATRIC MDS IN BRAZIL: THE BRAZILIAN WORKING GROUP ON MDS IN CHILDHOOD, FISRT RESULTS AND ITS EDUCATIONAL IMPACT L F Lopes1, L NieroMelo2, I LorandMetze3, L G Tone4, E Velloso5, M R D Latorre6

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22 1

H. Caˆ ncer, Ped. Oncology, Sa˜ o Paulo, Brazil, 2UNESP, Hematology, Botucatu, Brazil, 3UNICAMP, Hematology, Campinas, Brazil, 4 USP-Ribeira˜ o, Ped. Oncology, R. Preto, Brazil, 5USP-Sao Paulo, Hemocentro, Sa˜ o Paulo, Brazil, 6Fac. S. Publica-USP, Epidemiology, Sa˜ o Paulo, Brazil Contact e-mail:[email protected] The Grupo Cooperativo Brasileiro de Sindrome Mielodisplasica em Pediatria (GCB-SMD-PED) is a working group of hematologists, pediatric oncologists, pediatric hematologists, molecular biologists and others constituted since January 1997 with the aim to study pediatric (age B 18 years) MDS. It is composed of six Subcommittees with members from different Universities: cytology, histopathology, epidemiology, clinical, cytogenetics and molecular biology. Twentyone centers from seven different states of our country have registered cases for diagnostic review by sending clinical data, blood and bone marrow cytology and histology (when available). Among 114 cases already submitted, MDS was confirmed in 64. Using the WHO proposals, (MDS = bone marrow blasts B20%) 21 cases had acute leukemia (18 AML, three ALL), one had CML, 16 patients received other diagnoses. In 12 patients, the material sent was insufficient for a conclusive diagnosis. FAB types in the confirmed primary MDS cases were: six RA, 37 RAEB, 12 LMMC/J, one hypocellular-MDS, two not classifiable by FAB. Six patients had secondary MDS. Cytogenetic as well as molecular analyses for every submitted case could only be introduced in 2000. The GCB-SMD-PED functions as a consultant party for every pediatrician that needs diagnostic confirmation. But it is also a research group to analyze epidemiology, clinical and biological features of pediatric MDS in Brazil.

Or28 QUALITY OF LIFE IN MYELODYSPLASTIC SYNDROMES: MEASUREMENT ISSUES IN RESEARCH AND IN CLINICAL PRACTICE M Thomas Veterans Affairs Palo Alto Health Care System, Palo Alto, USA Contact e-mail:[email protected] Quality of life (QOL) is becoming increasingly viewed as an important concept when treating individuals who have myelodysplastic syndromes (MDS). Frequently used definitions of QOL emphasize the subjective, multi-dimensional nature of this concept. Yet an in-depth understanding of QOL in the MDS patient population remains elusive. A review of the MDS literature published within the past 5 years shows that, of those that identify QOL, the majority of the studies still made assumptions regarding QOL rather than actually performing measurements of this important concept. For example, many studies cite outcome variables such as length of stay while hospitalized, toxicity of therapy, complications from disease and treatment, and financial cost and conclude that patients’ QOL was therefore improved or worsened based on these variables alone. Studies of the MDS patient population have recently begun using quantitative instruments designed to measure health-related QOL in the setting of clinical trials. These instruments are designed to measure four key domains of health-related QOL: physical, functional, social and emotional well-being. Data from one such study will be briefly presented to illustrate the advantages and problems with using such instruments. These data will be contrasted with those obtained from a pilot study of MDS patients outside the clinical trial setting. Here, a structured interview format was used to elicit key factors patients themselves considered to be important vis-a-vis QOL and how their QOL was impacted by their experience with MDS. These data will serve as a vehicle for critically evaluating appropriate methods of QOL assessment in both the research and clinical practice setting. With improved assessment, interventions can be better

S11

targeted so that optimal QOL can be achieved and maintained for those living with MDS.

Or29 IMMUNE MEDIATED BONE MARROW FAILURE IN APLASTIC ANEMIA (AA) AND MYELODYSPLASIA (MDS) N S Young, J M Maciejewski, W Zeng, H Kook, G Chen, A Risitano, E S Sloand, J Melenhorst, A J Barrett Hematology Branch, National Heart, Lung and Blood Institute, Bethesda, USA Contact e-mail:[email protected] Blood count depression is frequent and important in AA and MDS. Despite a variety of possible etiologies, including toxic and infectious environmental insults and somatically acquired genetic changes, these distinct diseases share clinical features and indeed appear to overlap or merge and pathophysiologic mechanisms common to both may produce hematopoietic cell destruction. From the success of immunosuppressive therapy and supporting laboratory experiments, AA is now regarded as T cell mediated, organ-specific destruction of marrow. Indirect in vitro evidence, from early coculture inhibition to measurements of lymphocyte subsets and lymphokine expression and in vitro model systems, has implicated TH1/TC1 effectors causing apoptotic cell death via the Fas pathway. Molecular analysis of the T cell response is now feasible: using immortalized T cell clones, functionally inhibitory of autologous hematopoiesis, AA patients of similar HLA-DR show an oligoclonal immune response which is modulated by therapy and, remarkably, sharing of complementarity determining region sequence, suggesting reactivity to a common antigen. A similar approach can be used to assess the immune response in MDS. Other advanced methods applied to the study of both AA and MDS include T cell receptor Vb skewing and measurement of intracellular cytokines. While g-interferon has been most readily implicated in AA, tumor necrosis factor has figured more prominently in MDS. The frequent occurrence of paroxysmal nocturnal hemoglobonuria clones in both AA and MDSB at time of presentation B and the link between PNH and an immune pathophysiology is an important but not fully understood clue. Adaptation of the acute graft-versus-host system has provided an animal model of swift and severe stem cell destruction mediated by limited numbers of T cells that could be adapted to examination of different inciting antigens. Current AA treatments aim at ‘tolerizing’ patients, as for example using mycophenolate mofetil and delaying cyclosporine, as in our new protocol; high dose cyclophosphamide has prove, too toxic and no more effective than standard therapies. For MDS, adaptation of immunosuppressive therapy requires identification of appropriate disease subtypes and other positive risk factors, as well as regimens suited for an immune response perhaps primarily directed at genetically derived neoantigens.

Or30 CD 59 AND CD 55 DEFICIENCY IN PATIENTS WITH PRIMARY MYELODYSPLASTIC SYNDROME. AN ANALYSIS OF 35 PATIENTS J Cermak, Y Marinov, J Jelinek Institute of Hematology, Prague, Czech Republic Contact e-mail:[email protected] Deficiency of CD 55 and CD 59 antigens caused by mutation of PIG-A gene leads to abnormal sensitivity to complement-mediated lysis and to clinical symptoms of paroxysmal nocturnal hemoglobonuria (PNH). Cases of ‘PNH like’ disease with Ham test positivity have also been reported in patients with myelodysplastic

S12

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

syndrome (MDS). However, the relationship between PNH and MDS and the significance of CD 55 and CD 59 deficiency for clinical course and prognosis of MDS patients still remains unclear. We have studied by flow cytometry the incidence of CD 55 and CD 59 deficiency in isolated peripheral blood cell subpopulations from 33 patients with refractory anemia (RA) and two patients with refractory anemia with ringed sideroblasts (RARS). A statistical analysis of the relationship between the number of CD 55 and CD 59 negative (PNH III) cells and laboratory findings, clinical manifestation of MDS and prognosis included unpaired t-test, Mantel – Cox and Fischer tests. At the time of diagnosis, \ 3% (range: 4 – 28%) of CD 55- and CD59-red blood cells (RBC) were present in 14 patients (40%). Laboratory parameters suggesting increased hemolysis were found in only four patients (12%) with \ 20% of CD 59-RBC and none of the patients had clinically manifest hemolysis with hemoglobonuria. Three patients (9%) developed CD 55 and CD 59 deficiency 16 – 49 months after the diagnosis of MDS, the amount of CD 59-RBC was significantly higher than in the patients who were CD 59 deficient at the time of diagnosis (range: 31 –38%, PB0.001) and two out of these three patients had recurrent episodes of intravascular hemolysis with hemoglobonuria. No thrombotic events were observed in patients with increased amount of CD 55- and CD 59-cells. None of the patients with \ 20% of CD 59-RBC had hypoplastic bone marrow, all these patients exhibited low to normal in vitro growth of granulocyte macrophage progenitors (CFU-GM). No differences in karyotype abnormalities, IPSS score and overall survival (25 vs. 22 months, NS) were observed between patients with more and less than 20% of CD 59-RBC. Interestingly, no leukemic progression was present in patients with \ 20% of CD 59-RBC in contrast to three patients with B 20% of CD 59-RBC who subsequently developed MDS with excess of blasts. Univariate analysis of different parameters revealed increased number of reticulocytes ( \ 3%) as the only variable predictive for increased number of CD 59-RBC. Despite frequent CD 55 and CD 59 deficiency at the time of diagnosis in RA and RARS patients, hemolytic episodes were observed only in a minority of patients (6%) who developed a marked CD 55 and CD 59 deficiency in the course of the disease. This finding might reflect a genetic instability of preexisting MDS clone with tendency to multiple mutations of PIG-A gene. A potential relationship between a non-leukemic growth of CFU-GM in CD 55- and CD 59-patients and disease progression will be further investigated. This study was supported by scientific grant IGA NK 5937-3.

Or31 ABNORMAL T-CELL RECEPTOR V BETA REPERTOIRE IN PATIENTS WITH MYELODISPLASTIC SYNDROMES S Bonfigli1, M G Doro2, D Derudas1, C Fozza1, R M Nieddu1, M Longinotti1 1 Istituto di Ematologia, Universita` di Sassari, Sassari, Italy, 2Istituto CNR per lo studio della patologia del sangue e degli organi emopoietici, Area della ricerca di Sassari, Sassari, Italy Contact e-mail:[email protected] Autoimmune disorders are rather common in patients with myelodysplastic syndromes (MDS) who, however, frequently show asymptomatic immunologic abnormalities involving B and T lymphocytes as well as NK cells and monocytes. As the size of the human T-cell receptor V beta genes (BV TcR) repertoire is related to the immune competence status, the peripheral blood TcR repertoire was evaluated in the CD4 + and CD8+ T-cells of eleven MDS patients (six RA, one RAEB, one RAEB-T and three LMMoC) using monoclonal antibodies (MoAbs) to the BV 1, 2, 3, 5.1, 7, 8, 13.1, 14, 17, 20, 21.3 and 22 gene families. Twenty normal subjects

were studied as controls. As it is known, the BV TcR analysis by MoAbs allows to exactly quantify the BV usage and consequently to identify the presence of specific BV expansion or deletion. The mean expression values of the 12 BV genes in MDS and in controls did not vary in the CD4 + population, while in the CD8 + T cells BV 3 and BV 21.3 mean expression values resulted to be significantly reduced in the MDS patients. A skewed usage towards the CD4 + subpopulation was detected for the same BV families (BV 2, 13.1, 22) in the two groups: in addition the normal subjects showed the CD4 + skewing of the BV 5.1 and 17 as well. Having defined a BV expansion as any value of the BV family expression higher than the mean 9 2 S.D., no difference could be found between MDS and controls. Nevertheless, one patient showed an unusually high overexpression of the BV 5.1 that represented the 75% of the CD8 + T cells. These expansions will be investigated by the molecular ‘immunoscope’ analysis to define the clonality of the complementarity determining region 3 of each overexpressed BV family. In contrast to healthy controls, 50% of the MDS patients had a variable degree of TcR BV-specific T cell deletion and this phenomenon, ranging from one to five families, involved different BV families. These data demonstrate that the TcR BV repertoire expression is restricted in MDS affected patients.

Or32 CLONAL T-CELL AND IMMUNOGLOBULIN GENE REARRANGEMENTS IN MYELODYSPLASTIC SYNDROMES (MDS): A MORPHOLOGIC AND IMMUNOPEROXIDASE STUDY C A Hanson1, G A Dayharsh2, W G Morice1, J M Hodnefield1, D P Steensma3 1 Mayo Clinic, Hematopathology, Rochester, MN, USA, 2Mayo Clinic, Anatomic Pathology, Rochester, MN, USA, 3Mayo Clinic, Hematology, Rochester, MN, USA Contact e-mail:[email protected] Background: MDS and large granular lymphocyte (LGL) leukemia can both be associated with blood cytopenias. Clonal T cell gene rearrangement (GR) studies are commonly used as the diagnostic criteria for the diagnosis of LGL leukemia. However, GR can be non-lineage-specific and are known to occur in 5 – 20% of acute myeloid leukemias. Method: We analyzed 51 patients with MDS seen at our institution who had GR studies (15 women, 36 men; age range = 45 – 81 years; median age =65). Results: A diagnosis of MDS was made in all patients: RA/RCMD= 19, RARS= five, RAEB=17, CMML=ten. An MDS-associated karyotype was found in 25 of 51 (49%) of the patients; abnormalities of chromosome 5, 7, 8, or 20 accounted for 18 of these. All 51 patients presented with varying degrees of cytopenias; 48 were anemic, 34 were neutropenic, 36 were thrombocytopenic and 23 were pancytopenic. IPSS scores showed: Low =six, Int-1 =26, Int-2 =16 and High =three. GR were identified in 17 of the patients: three with clonal Ig GR and 14 with T-cell GR. The age distribution, hematologic findings, MDS classification, cytogenetic karyotype and IPSS scores of the GR + group were virtually identical to the overall group. All 17 GR + patients had cytopenias; 16 had thrombocytopenia, 11 were neutropenic and seven were pancytopenic. MDS classification included RA/RCMD= nine, RARS= one, RAEB= six, CMML=one. Ten of the 17 GR + cases had chromosomal abnormalities with chromosome 5, 7, 8, or 20 accounting for 70% of these. IPSS scores showed: Low =two, Int-1 = nine, Int-2 =five and High =one. Immunoperoxidase stains for cytotoxic T-cell markers (CD3, CD8, TIA-1, granzyme B) were performed in all cases. Two of the 17 cases with GR + demonstrated an abnormal staining pattern characteristic of T-LGL leukemia. None of the remaining cases with or without GR demonstrated any abnormal staining

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22 pattern. Three of the 17 GR + cases that progressed to acute myeloid leukemia had GR studies; all showed an identical GR pattern in the leukemic specimen as in the preceding MDS specimen. Cell sorting studies were done in four cases. In two cases, T-cell GR were confirmed in the myeloid cells. One case following sorting demonstrated a T-cell GR in only CD8-positive T-cells and a monosomy 7 karyotype by FISH in its myeloid cells. One case following sorting demonstrated clonal T-cell GR in both the CD8-positive T-cells and myeloid cells. Conclusion: These findings confirm that clonal rearrangements are non-lineage-specific and can occur in up to 30% of patients with MDS. Clonal rearrangements cannot be interpreted in isolation to make the diagnosis of T-LGL leukemia and do not exclude the possibility of an MDS. In addition, our data would indicate that GR in MDS do not identify a unique subset of disease.

Or33 RESPONSE TO ANTITHYMOCYTE OR ANTILYMPHOCYTE GLOBULIN IN LOW-RISK MDS PATIENTS WITH POLYCLONAL BLOOD AND BONE MARROW CELLS M Aivado1, A Rong1, M Stadler2, U Germing1, J Novotny3, K M Josten4, C Strupp1, S Knipp1, N Gattermann1, C Aul5, R Haas1, A Ganser2 1 University, Haematology, Du¨ sseldorf, Germany, 2University, Haematology, Hannover, Germany, 3University, Haematology, Essen, Germany, 4German Clinic of Diagnostics, Haematology, Wiesbaden, Germany, 5St. Johannes Hospital, Haematology, Duisburg, Germany Contact e-mail:[email protected] Background: Administration of antithymocyte or antilymphocyte globulin (ATG/ALG) is the treatment of choice for patients with aplastic anemia (AA). Recently, ATG/ALG was also shown to be effective in : 30% of patients with low-risk MDS. The distinction of MDS from AA can be very difficult and reliable predictive factors for successful treatment of low-risk MDS with ATG/ALG are lacking. X-inactivation studies have shown that AA is usually a polyclonal disorder, while in MDS-patients myeloid cells only rarely display polyclonal X-chromosome inactivation patterns (XCIP). We examined in female low-risk MDS-patients whether XCIP have an impact on the outcome of treatment with ATG/ALG. Patients and methods: Nine females (median age 58 years, range: 42 –65) were heterozygous and thus informative with regard to XCIP analysis. There were seven RA, one RAEB, one CMML. All patients received either ALG 15 or ATG 3.75 mg/kg per day i.v. for 5 days. We applied the humAR- and the PGK-1-assay on CD3 + T cells and granulocytes from peripheral blood (PMNC) as well as on bone marrow cells (BMC). Our data were obtained by automated DNA fragment analysis on an ALF express sequencer, which allows semiquantitative determination of the XCIP, yielding corrected ratios (CR). In accordance with other investigators, we defined a ‘clonal’ XCIP as a CR\ 3. Results: Six females experienced no improvement in their blood count. Pre-treatment, five of them had clonal XCIP and one had a skewed XCIP. Three patients had a response to ATG/ALG and all had polyclonal XCIP initially. No relevant alterations of XCIP were observed after treatment with ATG/ALG. Conclusion: polyclonal XCIP in PMNC or BMC may be a valuable predictor for a successful treatment of low-risk MDS with ATG/ALG. Recently, it has been speculated that there may be a common pathophysiology for some forms of AA and MDS. Further analyses on such patients are necessary in order to see whether polyclonality is a characteristic feature of such kind of AA/MDS.

S13

Or34 SECONDARY MDS AND AML, RELATIONSHIP BETWEEN TYPES OF PREVIOUS THERAPY AND SPECIFIC GENETICS AND CYTOGENETICS ABNORMALITIES J Pedersen-Bjaergaard Rikshospitalet, Copenhagen, Denmark Contact e-mail:[email protected] Based on characteristic chromosome aberrations, we have previously proposed different genetic pathways in therapy-related myelodysplasia (t-MDS) and acute myeloid leukemia (t-AML), (Blood 1995;86:3542 – 52). As already demonstrated in cases with balanced translocations to chromosome bands 11q23 and 21q22, cases of t-AML with t(15;17), inv(16) and translocations to band 11p15 have now also been related to therapy with topoisomerase II inhibitors. Furthermore, cases of t-MDS and t-AML following radiation have been shown characteristically to present deletions of various parts of 7q or monosomy 7. Mutations of p53 have been observed frequently in t-MDS and t-AML, significantly related to previous therapy with alkylating agents, to deletions or loss of 5q or monosomy 5 and to a complex karyotype. In some cases of t-AML with a normal karyotype possible representing cases of MDS and AML de novo, internal tandem duplications of the FLT3 and the MLL genes have been demonstrated. Finally, transgenic mice expressing the AML-specific chimeric rearrangements of the MLL and the RARA genes in bone marrow precursors, have been shown to develop overt leukemia, often at a young age, whereas mice expressing the rearranged core binding factor genes AML1 and CBFB presented only an insufficient hematopoiesis. Based on these new results at least eight different genetic pathways can now be proposed in t-MDS and t-AML. Most pathways are related to specific types of previous therapy and some pathways are associated with presentation of the disease as t-MDS, others to presentation as overt t-AML. Prognosis and response to intensive antileukemic chemotherapy differ between the pathways. The genetic pathways in t-MDS and t-AML seem to be very similar to pathways in MDS and AML de novo.

Or35 ATOMIC BOMB RADIATION ENHANCES THE RISK OF MDS A Kimura1, Y Takeuchi1, H Tanaka1, K Satoh2, M Ohtaki2, N Hayakawa3 1 Hematology/Oncology, Hiroshima University, Hiroshima, Japan, 2 Environmetrics/Biometrics, Hiroshima, Japan, 3Epidemiology, Hiroshima, Japan Contact e-mail:[email protected] To see the effect of radiation exposure on MDS development among Hiroshima A bomb survivors, patients with definite individual radiation dose were examined who were referred to our department between 1985 and 1999. The dose was estimated according to ABS 93 dosimetry system established in our institute. Among 735 patients without malignancy, 26 patients aged from 48 to 88 (median: 75) and ´ -ray had diagnosis of MDS, including exposed up to 429.8 cSv of f A 16 RA, two RARS, four RAEB and four RAEBt. We analyzed the relative risk in relation to exposed radiation dose using a multiple regression with Cox proportional hazard model. Highly significant dose-effect on the risk of MDS was found. The relative risk was 2.53 (95% CI: 1.49 – 4.30) at the dose of 100 cSv against 0 cSv (P = 0.0006). Age at the time of bombing showed significant effects on the risk (P= 0.0000); younger age at bombing has higher relative risk. The value of the relative risk was between that of leukemia and that of some solid tumors like lung, thyroid and breast cancer. This is the first time that the incidence of hematological disease other than leukemia was found to be radiation dose dependent.

S14

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

Or36 CLINICAL PRESENTATION AND OUTCOME OF MYELODYSPLASTIC SYNDROMES (MDS) WITH del (20q) M Montillo1, A Tedeschi1, K G Hayes2, E Morra1, E Estey3, M J Keating3 1 Department of Haematology, Niguarda Ca’Granda Hospital, Milano, Italy, 2Department of Cytogenetic, MD Anderson Cancer Center, Houston, TX, USA, 3Leukemia Unit, MD Anderson Cancer Center, Houston, TX, USA Contact e-mail:[email protected] The medical records of 59 MDS patients referred to the MDACC presenting the del(20q) as the sole cytogenetic abnormality or associated to other cytogenetic abnormalities were reviewed to identify clinical characteristics and outcome. The median age at diagnosis was 62.5 years (range: 34 – 77), 17 were female and 42 male. According to FAB classification 26 cases (44.1%) were RA, eight cases (13.6%) RARS, six cases (15.2%) RAEB, 15 cases (25.4%) RAEBt, one patient presented with CML. Seven (11.9%) patients had a secondary MDS: two RA developing after an autologous bone marrow transplantation procedure while one RARS, two RAEB and two RAEB-t were related to conventional dose chemotherapy. The del(20q) was present as the sole abnormality in 33 patients (56%); the association with a complex karyotype was higher in patients with RARS, RAEB and RAEBt compared to patients with RA (P B0.01). An association with the del(5q) or del(7q) and monosomy 5 or 7 was found in 23.7% of patients. A terminal deletion of the 20q was observed in 47 (79.7%) patients, the remaining patients presented an interstitial deletion q11q13. All but one patient were anemic at presentation (Hb B 12 g/dl), anemia was severe (Hb B 8 g/dl) in 22 cases (27.3%); severe neutropenia (PMN B 1 × 109/l) was observed in 19 patients; Plt were within the normal range in eight cases while 22 patients presented with a Plt count B40 × 109/l. The prognosis of the population presenting RA and RARS was generally poor as 15 of the 34 cases transformed to RAEBt or AL. No difference in survival was observed when the 23 patients with the isolated loss of the 20q were compared with patients with complex cytogenetic changes including del(20q), median survival 28 and 40 months. The median survival of patients with RAEB and RAEBt was 8 months; significantly different was the survival of these patients when compared to the RA and RARS population. No difference on survival was observed between the two groups when only patients with isolated del(20q) were observed; while complex karyotype conferred a shorter survival to patients with advanced MDS (P B0.004). The del(20q) in the setting of a complex karyotype was associated with unfavorable outcome in patients with RAEB and RAEBt in respect of the same setting of patients with the del(20q) alone.

Or37 DYSPLASTIC FEATURES IN 614 PATIENTS WITH DE NOVO ACUTE MYELOID LEUKEMIA (AML): FREQUENCY, CORRELATION TO CYTOGENETIC RESULTS AND PROGNOSTIC IMPACT T Haferlach1, C Schoch1, H Lo¨ ffler2, W Gassmann3, S Schnittger4, M C Sauerland5, W D Ludwig6, T Bu¨ chner7, W Hiddemann4 1 Department for Internal Medicine III, Munich, Germany, 2Medical Department II, Kiel, Germany, 3Hematology, Siegen, Germany, 4 Department for Internal Medicine III, Munich, Germany, 5 Biostatistics, Mu¨ nster, Germany, 6Hematology, Berlin, Germany, 7 Hematology, Mu¨ nster, Germany Contact e-mail:[email protected]

Between 1992 and 1999, we analyzed prospectively cytomorphological and cytogenetic features in 614 patients with de novo AML treated within the AMLCG-92 trial. Median age was 53 years (range: 16 – 82). Overall survival (OS) at 6 years was 22.6%. FAB types were: 3.3% M0, 19.1% M1, 32.1% M2, 6.7% M3(v), 16.9% M4, 6.2% M4eo, 5.9% M5a, 6.2% M5b, 3.3% M6, 0.3% M7. Dysgranulopoiesis = DysG, DysE or DysM as well as trilineage dysplasia (TLD) were analyzed according to criteria proposed by Goasguen and Bennett (Leukemia 1992;6:520). No dysplasia was seen in 45.1%, DysG in 9.8%, DysE in 5.2%, DysM in 15.1%, Dys GE in 2.3%, Dys GM in 4.4, Dys EM in 2.4, and TLD in 14.8%. No significant differences were found for overall or disease free survival (OS, DFS) when patients with 0, 1, 2 or 3 lineage dysplasia were compared. Following the WHO proposal, also in patients without any dysplasia or with one lineage involved in correlation to multi-lineage (i.e. two cell lines) or TLD no significant differences were found. Cytogenetic analysis using standard definitions for low risk (i.e. t(8;21); t(15;17) inv(16); 90 patients), intermediate (all others; 279 patients) and high risk ( − 5, 5q-, − 7, complex; 80 patients; overall 75% cytogenetic results available) demonstrated significant differences for OS and DFS: low vs. intermediate risk (P =0.0003); intermediate versus high risk (P = 0.0002) and low versus high risk (P =0.0001). We then correlated cytogenetic data to dysplastic features and detected an inverse correlation of the good cytogenetic risk group with trilineage dysplasia at one hand and a significant correlation of bad risk cytogenetics with trilineage dysplasia at the other hand (PB 0.001). However, in multivariate analysis dysplasia alone showed no major prognostic impact. In conclusion, cytomorphological features such as dysplasia demonstrated only very limited prognostic differences in 614 patients with de novo AML. In contrast, cytogenetic profiles reflected much better biological entities.

Or38 THE ROLE OF AZACITIDINE IN THE TREATMENT OF THE MYELODYSPLASTIC SYNDROME L Silverman Mount Sinai Medical Center, Division of Medical Oncology, New York, USA Contact e-mail:[email protected] Although a number of therapeutic agents have been tested in MDS, to date no single agent or combination has emerged as a standard of treatment. Supportive care remains the mainstay of therapy. Azacitidine (Aza C), a ring analog of the pyrimidine nucleoside cytidine, has effects on DNA metabolism, cell differentiation and gene expression in part through effects on DNA methylation. The Cancer and Leukemia Group B (CALGB) reported results of a regimen of Aza C administered as a continuous IV infusion at 75 mg/m2 per day for 7 days q28d in patients with RAEB and RAEB-T. Responses occurred in 21 of 43 evaluable patients: 12% had a complete response (CR = complete normalization of blood and bone marrow); 25% had a partial response (PR = ]50% restitution in the deficit from normal in all three peripheral blood cell lines and the elimination of all transfusion requirements); 12% improved ( ]50% restitution in the deficit from normal in ]1 peripheral blood cell line and/or a ]50% reduction in transfusion requirements). A trilineage response (TLR = CR+PR) occurred in 37% of patients. The median survival and median duration of response was 13.3 and 14.7 months, respectively. In a second CALGB study, Aza C was administered via SC bolus in an ambulatory regimen at the same dose and schedule to 67 patients. Thirty-five patients (52%) responded: 12% CR; 12% PR; 28% improved with a TLR of 24%. In a phase III randomized controlled trial Aza C was compared to supportive care (SC). Responses occurred in 61% on the Aza C arm (9% CR, 15% PR, 38% improved) compared

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22 to 5% (improved) on SC (PB 0.001). The median time to leukemic transformation or death was 21 months for those on Aza C versus 13 months for SC (P= 0.004). Probability of transformation to AML as the first event was lower on Aza C (15%) than on SC (35%) (P =0.004). QOL assessment demonstrated significant major advantages in physical function, symptoms and psychological state for patients initially treated with Aza C and for those on SC after crossover to Aza C. Median survival for Aza C and SC (analyzed by intent to treat regardless of crossover) was 23 and 14 months, respectively (P =0.09). The probability of survival at 12 and 24 months is 73 and 43% for Aza C compared to 62 and 26% for SC, respectively. Aza C was well tolerated, with further exaggeration of pre-existing cytopenias the most significant side effect. Treatment-related mortality occurred in B 1% of patients. Treatment with Aza C results in significantly higher response rates, improved QOL, decreased rate of transformation to acute leukemia and a delayed time to leukemic transformation or death compared to SC. These results demonstrate that Aza C is the only agent other than allogeneic BMT to alter the natural history of MDS and is superior to supportive care.

Or39 LOW DOSE DECITABINE IN ELDERY PATIENTS WITH HIGH RISK MDS; WHO WILL RESPOND BEST? P W Wijermans1, M Luebbert2, G E G Verhoef3 1 Hospital Leyenburg, haematology, The Hague, Netherlands, 2 Albert-Ludwigs University, Haematology, Freiburg, Germany, 3 Universital Hospital Gasthuisberg, Haematology, Leuven, Belgium Contact e-mail:[email protected] Decitabine is a pyrimidine analogue that inhibits DNA methylation. The antiproliferative effect of Decitabine may result from DNA strand breaks or from reactivation by demethylation of genes involved in cellular proliferation and differentiation that are silenced by promotor hypermethylation (for instance P15 ink4B). RAEB, RAEBT and RA patients who were transfusion dependent were treated with low dose Decitabine in three different phase II studies. The data presented are based on the final results of the first two phase II studies and the first 43 evaluable patients of the third study. According the IPSS 38 Int.1 risk, 41 Int 2 risk and 59 high risk patients were treated. The median age was 70 years (38 –89 years). The response rate was 50% with a CR rate of 18% and a 2-year survival of 32%. A marked difference in response rate was seen for the different IPSS risk groups. Int 1 39%. Int 2 45% and high risk patients 61%. No difference was seen for sex or age. The response rate for the different FAB groups was as follows: RA 33%, RAEB 53%, RAEB-T 55% and CMML 50%. The response rate in patients without cytopenia was 33% whereas in patients with three cytopenic lineages the response rate was 59%. Patients with high risk cytogenetic abnormalities did not worse compared with those without abnormalities or low risk abnormalities. Several of the patients who relapsed after an initial response did show a response again after retreatment with Decitabine. The median and 2-year survival was 15 months and 32%, respectively for the whole group. The median and 2-year survival for the IPSS subgroups was Int 1: 19 months 41%. Int 2 14 months 26% and high risk also 14 months and 26%. For FAB subgroups it was as follows: RA 19 months, 30%. RAEB 15 months 34%, RAEB-T 12 months 18% CMML 18 months, 35%. No influence of sex was seen. The groups were too small for the analysis of the influence of age according the risk groups on survival but patients \ 75 years did worse than patients B65 years. Three lineages cytopenia was not a risk factor for survival. In 34% of the patients who could be followed during therapy a complete cytogenetic response was observed. This occurred even in patients with high risk abnormalities. Increase of the

S15

platelet count often already after the first therapy cycle was seen in 54% of the patients. Decitabine shows responses in patients with a MDS even in those with bad prognostic factors like high blast cell count and high risk cytogenetic abnormalities. A phase III study is needed to see if these responses translate into a better survival.

Or40 STEM CELL BIOLOGY OF MYELODYSPLASTIC SYNDROMES S E W Jacobsen Department of Stem Cell Biology, University Hospital of Lund, Lund, Sweden Contact e-mail:[email protected] Recent progress in identifying and purifying human lympho-myeloid stem cells as well as development of surrogate human stem cell assays, including in vivo xenograft models, have facilitated studies of stem cells in hematological malignancies. In this presentation, an overview of recent developments in normal stem cell biology will be reviewed, as a background for the recent attempts to identify, purify and characterize multipotent hematopoietic stem cells in MDS. In addition, the results from these studies will be summarized and discussed with regard to the involvement of lympho-myeloid stem cells in MDS and the biological and potential clinical implications of these findings.

Or41 N-RAS MUTATION AND FLT3 INTERNAL TANDEM DUPLICATION IN ACUTE MYELOID LEUKAEMIA WITH ABNORMAL KARYOTYPE; A STUDY OF ASSOCIATION WITH ALLELLIC VARIANTS IN CARCINOGEN-METABOLISING ENZYMES D T Bowen1, M E C Frew1, S Rollinson2, P Roddam2, A Dring2, M T Smith3, A K Burnett4, G J Morgan2 1 University of Dundee, Molecular and Cellular Pathology, Dundee, UK, 2University of Leeds, Pathology, Leeds, UK, 3University of California, Berkeley, Environmental Health Sciences, Berkeley, USA, 4 University of Wales College of Medicine, Haematology, Cardiff, UK Contact e-mail:[email protected] Background: Accumulating evidence for the leukaemogenicity of selected chemotherapeutic agents and for benzene is emerging but factors predisposing to the development of AML following exposure to these and other carcinogens are less clear. An association between the presence of an inactivating NQO1 polymorphism and the development of AML following both benzene exposure and prior chemotherapy suggest a potential paradigm for genetic predisposition. Methods: We have assayed DNA samples from 447 AML patients B55 years selected for abnormal karyotype from a cohort treated within the MRC AML ten and 12 clinical trials. Allelic variant frequencies were determined in genes encoding carcinogenmetabolizing enzymes GST T1, GST M1, SULT 1A1, CYP 1A1*3, CYP 2D6, CYP 2C19*2 and NQO1. GST T1/M1 null genotype status was assayed by multiplex PCR, CYP1A1 and 2D6 by PCRRFLP, and all other genotypes by allelic discrimination (ABI 7700 Taqman™). Flt3 ITD was assayed by previously published manual PCR. Samples were first screened for N-RAS mutation using the Transgenomic WAVE™ DNA Fragment Analysis System with separate assays for codons 12/13 and for codon 61. Abnormal patterns were confirmed as mutant by both direct sequencing (or cloning and sequencing) and mutant specific PCR. Results: Flt3 ITD frequency was 17% (77/447) and N-RAS mutation 12% (53/443) for the entire

S16

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

selected cohort. Both mutations were found together in only four patients (P = 0.05), confirming previous observations. No associations were found between enzyme allelic variant frequencies and the presence of Flt3 ITD for the entire cohort or cytogenetic subgroups. CYP 1A1*3 high-inducibility variant was more frequent in the entire cohort of patients with N-RAS mutation (8/53 mutation versus 26/374 non-mutation: P= 0.054, OR = 2.37, 95% CI 1.01-5.57), and also amongst the 5q/7q cytogenetic subgroup (4/7 mutation versus 4/79 non-mutation: P =0.001, OR = 25, 95% CI 4.11 – 151.8). Conclusion: Individuals with CYP 1A1*3 may metabolize leukaemogenic compounds more rapidly to genotoxic intermediates increasing the possibility of N-RAS mutation.

Or42 Lecam1/ICAM RATIO IDENTIFIES A DISTINCT PATTERN ON NORMAL AND MYELODYSPLASTIC SYNDROME (MDS) BONE MARROW CD34 + CELLS F Buccisano, A Venditti, G Del Poeta, L Maurillo, E Abruzzese, A Tamburini, S Amadori University Tor Vergata, Hematology, Rome, Italy Contact e-mail:[email protected] Background: interaction of adhesion molecules (AM) with bone marrow (BM) microenvironment regulates hematopoietic progenitors growth and survival. Consequently, the abnormal growth and proliferation observed in MDS/secondary acute myeloid leukemia (sAML) could be secondary to defective adhesive properties within the stem cell compartment. Methods: we compared, in a three-color flow cytometric assay, the expression of b1/b2-integrins, Lecam1, CD44 and ICAM1 on CD34+ progenitors from BM of 42 patients affected by MDS (19 RA, one RARS, 16 RAEB, six RAEBt) or sAML (nine) and 15 healthy donors. The expression of AM was measured as percentage, mean fluorescence intensity (MFI) and AM index (AMI=product of MFI and percent positive cells). Results: in the MDS/sAML group, a defective expression of Lecam1 in terms of percentage, MFI and AMI were observed, whereas percentage of positivity of ICAM1 was higher. Given this reciprocal expression we designed a Lecam1/ICAM1 AMI ratio to distinguish normal from MDS/sAML CD34+ cells. In fact, a lower ratio for MDS/sAML patients compared to normal controls was observed (mean 1.3 9 1.6 vs. 27.7 916.8, P = 0.000) with all value restricted below the threshold of 5. The ratio inversely correlated with BM blast infiltration (Spearman’s r −0.39, P= 0.005). Among MDS patients, a low ratio ( B1) identified a group with higher risk of leukemic evolution (38 vs. 4%, P= 0.006). Interestingly, a patient converted the AMI ratio from 0.33 to 12.12 after a chemotherapy-induced remission: when the ratio eventually decreased the patient experienced a clinical relapse. Conclusions: (1) Lecam1 is defective in the stem cell compartment of MDS, whereas ICAM1 is overexpressed; (2) a low ratio ( B 5) suggests a myelodysplastic pattern; (3) the value of the ratio may serve as a parameter to assess the risk of disease progression; and (4) after chemotherapy, the ratio could be useful to monitor minimal residual disease.

Or43 LENTIVIRUS-TRANSDUCTION OF HEMATOPOIETIC PRECURSORS FROM PATIENTS WITH SIDEROBLASTIC AND REFRACTORY ANEMIA T Matthes1, V Kindler1, F Leuba2, P Beris1, V Piguet2 1 Department of Internal Medicine, Division of Hematology, Geneva, Switzerland, 2Department of Dermatology, Division of Dermatology, Geneva, Switzerland

Contact e-mail:[email protected] Lentivectors derived from human immunodeficiency virus-1 (HIV-1) represent a novel investigational and therapeutic tool for the transduction of human hematopoietic stem and progenitor cells (HSCs). Using this approach, we developed an experimental protocol that will allow the study of genes implicated in the ineffective hematopoiesis and increased cell death occurring in low risk MDS (sideroblastic (SA) and refractory anemia (RA)). CD34 + cells were purified from the bone marrow of six adult patients with a normal marrow and of five patients with SA and two with RA. Cells were than cultured with erythropoietin (Epo), IL-3 and stem cell factor (SCF) and analyzed after 10 and 20 days of culture for proliferation, phenotype and morphology. Alternatively, cells were first cultured for 7 days with thrombopoietin, Flt-3 Ligand (F3L) and SCF before Epo/IL-3/SCF. CD34+ cells from normal adult marrow were amplified 96 920-fold (day 10) and 329 956-fold (day 20) after culture in Epo/IL-3/SCF. Of these, 88 9 9% became glycophorin +/CD45− cells and 10 –30% were CD71 − and displayed a morphology of mature RBCs. In two-step cultures amplification was similar but the percentage of glycophorin +/CD45− cells was only 50% at day 20. In cultures from SA and RA, proliferation was low (1- to 37-fold at day 10) and between day 10 and 20 of culture no further proliferation was observed in both culture conditions. The number of erythroid and mature RBCs obtained varied considerably in these patients from 0 to 60% (mean =24%). At various timepoints of culture (days 1, 3 and 7) cells were transduced with lentivectors containing either (I) the EF1a promoter or, (II) the PGK promoter or, (III) the PGK promoter and the central polypurine tract (cPPT) and analyzed at days 10 or 20 by flow cytometry for the presence of GFP-expressing glycophorin + cells. Transduction was optimal for CD34 + cells from normal bone marrows with the PGK-cPPT vector at day 3 of culture, reaching 68 96% GFP+/glycophorin + cells at a MOI of 10 (44% for EF1a vector; 38% for PGK-vector). SA- and RA-CD34 + cells transduced in the same conditions produced 27 910% in four of six patients. Thus, highly effective transduction of hematopoietic precursor cells from the bone marrow of adults and of some patients with SA and RA is achieved with this protocol. This approach will allow the study of the influence of various transducible genes (transcription factors, apoptosis genes,…) on normal erythroid differentiation, and eventually help to correct defects in dysplastic hematopoietic cells.

Or44 EXPRESSION OF A HYPERMETHYLATED AND SILENCED P15/INK4B GENE IN A SUBGROUP OF MDS PATIENTS IS RESTORED BY TREATMENT WITH THE METHYLATION INHIBITOR 5-AZA-2%-DEOXYCYTIDINE M Daskalakis1, T T Nguyen2, P Guldberg3, C Nguyen2, G Ko¨ hler4, P Wijermans5, P A Jones2, M Lu¨ bbert1 1 Division of Hematology/Oncology, Albert-Ludwigs-University, Freiburg, Germany, 2Norris Cancer Center, University of Southern California, Los Angeles, CA, USA, 3Institute of Cancer Biology, Danish Cancer Society, Copenhagen, Denmark, 4Department of Pathology, Albert-Ludwigs-University, Freiburg, Germany, 5 Department of Hematology, Leyenburg Hospital, The Hague, Netherlands Contact e-mail:[email protected] Background: Treatment of MDS patients with low doses of DNA methylation inhibitors (Azacitidine, Decitabine) results in 47 –64% overall response rates (Silverman et al., ASCO Proc. 1998;17:14a; Wijermans et al., J. Clin. Onc. 2000;18:956). In order to better understand the in vivo mechanisms of action of Decitabine (DAC), we examined methylation and expression of the p15/INK4B gene that

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22 is frequently hypermethylated in MDS and acute leukemia. Methods: p15 promoter/exon 1 methylation in bone marrow mononuclear cells (MNC) of MDS patients before and during DAC treatment was assayed by analysis of bisulfite-treated and PCR-amplified DNA using methylation-sensitive single nucleotide primer extension (MsSNuPE), denaturing gradient gel electrophoresis (DGGE) and sequencing. p15 protein was detected by immunohistochemical staining of bone marrow biopsies using a polyclonal antibody. Results: p15 was methylated in 15/24 patients (63%) and unmethylated in nine of 24 patients (37%). Seventeen patients were sequentially analyzed following initiation of DAC treatment. In nine of 11 patients, initial p15 hypermethylation was reduced by \25%. DGGE analyses and DNA sequencing were indicative of induction of hypomethylation at individual alleles. p15 protein levels in MDS bone marrow MNC were similar to normal bone marrow in nine of 20 patients (45%) and very low or absent in the remaining 11 patients (55%). Sequential results of p15 methylation and expression levels were available in ten patients. Normal p15 expression was observed in two of the three patients with unmethylated p15, and lack of p15 expression in four of the seven patients with hypermethylated p15. During DAC treatment, these four patients showed both p15 re-expression and a decrease in p15 methylation. In the remaining patients, pre-treatment expression levels were either stable (four patients) or increased (two patients). Conclusions: Simultaneous hypermethylation and silencing of p15 expression were reversed in a subgroup of MDS patients following DAC treatment. In other patients, either enhanced expression in the absence of initial hypermethylation, or lack of p15 expression despite unmethylated p15 were noted. The emergence of partially demethylated epigenotypes and re-establishment of normal p15 protein expression following the initial DAC courses implicate pharmacologic demethylation as a possible mechanism resulting in hematologic response in MDS. Analysis of a larger series of MDS patients is warranted to investigate this pharmacologic activity further.

Or45 RATIONALE FOR USING HIGH DOSE CHEMOTHERAPY WITH OR WITHOUT STEM CELL TRANSPLANTATION IN HIGH RISK MDS A Ganser Medizinische Hochschule Hannover, Abt. Ha¨ matologie und Onkologie, Hannover, Germany Contact e-mail:[email protected] Standard intensive chemotherapy leads to complete remission in about half of the patients with high-risk MDS. The various induction chemotherapy schedules do not differ significantly in the complete remission rate, however improved supportive care or changes in selection criteria apparently have resulted in a reduction of the early death rate over recent years. Intensified induction therapy with medium- to high-dose ara-C in conjunction with fludarabin or topotecan appear to increase the CR-rate, while the role of hematopoietic growth factors in combination with induction chemotherapy (priming) is still being studied. Despite the improvements in remission induction, the duration of remission usually is short with only a minority of patients remaining in long-term CR. Although consolidation therapy with chemotherapy or interleukin-2 has been proposed, its value for remission maintenance has either not been formally tested or is without value. Factors influencing remission induction and duration may vary between the reported trials but include age of the patients, presence of certain chromosomal abnormalities, stage of disease and duration of the disease prior to chemotherapy. Since CR is associated with improved clinical performance and probably also prolonged survival, intensive induction chemotherapy should be part of the treatment strategy in younger patients as well as in elderly patients with good performance status to reduce the tumor load.

S17

Post-remission therapy includes allogeneic stem cell transplantation, either conventionally in younger patients or with attenuated conditioning in the elderly or those with risk factors to exploit the graft-versus-leukemia effect. Patients not eligible for allogeneic stem cell transplantation should be entered into clinical trials studying the role of newer agents, e.g. farnesyl transferase inhibitors, drug –antibody conjugates, demethylating agents, thalidomide, arsenic trioxide and agents affecting the differentiation pathway.

Or46 CHEMOTHERAPY FOR HIGH-RISK MDS: MD ANDERSON APPROACHES E Estey, H Kantarjian, J Cortes, F Giles, M Beran MD Anderson Cancer Center, Leukemia Department, Houston, TX, USA Contact e-mail:[email protected] This presentation will review results of therapy of newly-diagnosed high-risk MDS(RAEB,RAEB-t) at MD Anderson. Much effort during the 1990s involved development of regimens containing topotecan +ara-C(TA) or fludarabine + ara-C(FA). However, results of a recent multivariate analysis indicate that neither FA nor TA regimens are superior to idarubicin +ara-C(IA) either in treatment of our patients with MDS or with AML. More recently, we have investigated gemtuzumab ozogamycin (GO, ‘mylotarg’) as a single agent in patients over age 65 or with poor performance status. CR rates are 0/9 in patients with abnormal cytogenetics; the only patient with a normal karyotype had a CR. These results are mirrored by the results considering patients with both MDS and AML in whom CR rates are 7/13 normal cytogenetics and 1/30 abnormal cytogenetics. Approximately 2/3 s of the failures is due to resistant disease. These results have led us to close the single agent trial in patients with abnormal cytogenetics. Given the low response rate, it appears unlikely that combinations of mylotarg with other agents will be useful in such patients (although we are currently investigating a mylotarg, fludarabin, ara-C regimen). In contrast, the remission rate in patients with a normal karyotype approaches that seen historically, suggesting that combinations of mylotarg with other drugs might be profitably investigated in patients with a normal karyotype. Currently, we are focusing on new agents in patients with abnormal karyotypes patients and are planning to randomize patients to troxacitabine + ara-C versus troxacitabine +idarubicin versus IA.

Or47 CYTOGENETIC DATA PREDICT OUTCOME OF INTENSIVE CHEMOTHERAPY IN PATIENTS WITH MYELODYSPLASTIC SYNDROMES AGED OVER 60 YEARS U Germing1, A Ngomo1, C Strupp1, T Emde1, A A N Giagounidis2, C Aul2, N Gattermann1 1 University, Haematology, Du¨ sseldorf, Germany, 2Johanneshospital, Haematology, Duisburg, Germany Contact e-mail:[email protected] Background: Intensive chemotherapy in high-risk MDS is still discussed controversial, especially in patients aged over 60 years. Methods: Between 1991 and 2000, 38 patients over the age of 60 years (median age 66 years) with myelodysplastic syndromes underwent induction chemotherapy with an intensive regimen (Idarubicin/ARAC). Results: Initial diagnoses were RA in five cases, RARS in one case, RAEB in ten cases, RAEB-T in 19 cases and CMML in three cases. At the start of treatment, in 29 patients transformation to AML was documented. In 34 patients, cytogenetic data before the start of chemotherapy were available. Using the cytogenetic risk categories according to the IPSS 19 patients (56%) belonged to the

S18

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

low-risk group (A), three (9%) to the intermediate (B) and 12 (35%) to the high-risk group (C). Twenty-one patients (55%) entered complete remission, 14 patients failed to respond (37%), three (8%) patients succumbed to treatment-related complications in aplasia. CR rate in group A was 77%, in group B 67% and in group C 25% (P= 0.02). Six of the patients with CR in group A relapsed (40%), all patients of groups B and C relapsed. Median survival of the entire group was 10 months with a cumulative survival rate of 25% after 2 years and 10% after 5 years. There were no significant differences concerning therapy outcome and FAB subtype at the beginning of the chemotherapy. Patients who received CR with normal karyotype had a disease-free survival of 25 months. Patients with normal karyotype have a benefit in survival as compared to age and medullary blasts matched patients that did not receive intensive chemotherapy, but a very small group of patients achieve long-term remission. Patients with unfavorable karyotype have no benefit from intensive chemotherapy even if they enter CR, because they have early relapse. Conclusion: Intensive chemotherapy for high-risk MDS patients aged over 60 years with an unfavorable karyotype is not advisable.

Or48 TREATMENT OF MYELODYSPLASTIC DROMES WITH 5 AZACITIDINE

SYN-

R K Shadduck, Z R Zeigler, J M Raymond, J F Gryn, J Lister Western Pennsylvania Cancer Institute, Medicine, Pittsburgh, PA, USA Contact e-mail:[email protected] Fifty-one patients with myelodysplasia (MDS) were treated with one to 26 cycles of 5 azacitidine. Therapy was seven daily subcutaneous injections of 75 mg/m2 repeated every 28 days. The mean age of the patients was 69 years (range: 36 –88) and the time from diagnosis to start of therapy with 5 azacitidine was 14 months (range: 1 –83). Nine patients withdrew after one cycle of therapy, leaving 42 evaluable for response. Response is defined as either an increase in ANC to \1500/ml, platelet count to \10000/ml or sustained transfusion independence ( \ 8 weeks) from red cells and/or platelets. Responses were detected after one to four cycles of therapy. Six of the nine patients who withdrew developed neutropenia after one cycle; three patients declined further therapy. Of the 42 patients who received more than one cycle of therapy, there were 16 responders. Eight patients achieved complete remissions (CR) of which three had clearing of the cytogenetic abnormalities. The CRs were maintained on continued therapy for 2+ to 21 months. Eight patients achieved partial responses PRs that lasted on average, 10 months (range = 2 – 36 +). Nine of the responders died of disease progression, one of cerebral hemorrhage and one of cardiac failure. Five remain in remission after 2 + to 36+ months of therapy. Therapy with 5 azacitidine is well tolerated and provides effective palliation and/or remission for \ 30% of patients with myelodysplastic syndromes.

Or49 A PHASE II STUDY OF INTENSIVE CHEMOTHERAPY (IC) WITH FLUDARABINE, CYTARABINE AND MITOXANTRONE IN P GLYCOPROTEIN (PGP) NEGATIVE MYELODYSPLASTIC SYNDROMES (MDS). GROUPE FRANCAIS DES MYELODYSPLASIE (GFM) T Pre´ bet1, A Merlat2, A Stamatoullas3, E Deconinck4, C Fruchart5, B Mahe´ 6, N Ifrah7, N Gratecos8, P Casassus9, P Lepelley10, F Dreyfus2, P Fenaux10, E Wattel11

1

Lyon, France, 2Paris, France, 3Rouen, France, 4Besancon, France, Le Havre, France, 6Nantes, France, 7Angers, France, 8Nice, France, 9 Bobigny, France, 10Lille, France, 11He´ matologie, Hoˆ pital Edouard Herriot, Lyon, France Contact e-mail:[email protected] 5

Background: A previous study conducted by the GFM and GOELAMS groups demonstrated that quinine increases the complete remission (CR) rate and survival in PGP positive MDS cases treated with IC. The present trial was conducted in order to improve the CR rate in PGP negative MDS. Since leukemic cells primed by exposure to fludarabine exhibit enhanced accumulation of cytarabine triphosphate (the cytotoxic nucleotide of cytarabine), especially with continuous AraC infusion (Clin Cancer Res. 1998;4:45 – 52), a phase II trial was designed to explore the potential feasibility and efficacy of a combination chemotherapy consisting of fludarabine, mitoxantrone and cytarabine for high-risk MDS. Methods: Inclusion criteria included: aged 65 years; MDS with an excess of marrow blasts (RAEB, RAEB-T) or having progressed to AML (MDS-AML). Patients received mitoxantrone 12 mg/m2 per day, days 2 – 5 (over 30 min)+fludarabine, 30 mg/m2 per day (over 5 min) days 2 – 5+ AraC 1.5g m2 every 24 h (continuous infusion started 5 h prior to fludarabine) days 1 – 5 (FAM protocol). CR criteria were stringent: marrow blasts B5%, normalization of cytopenias and of karyotype, disappearance of MDS features. Patients B60 years with no HLA identical donor and who achieved CR were scheduled to receive autologous stem cell transplantation (ASCT) preceded by a moderate consolidation chemotherapy (CT) course. Older patients received several consolidation CT courses. The outcomes of the FAM-treated patients were compared with the outcomes of 32 MDR negative MDS patients who fulfilled the same inclusion and response criteria and who received mitoxantrone 12 mg/m2 per day, days 2 – 5 +AraC 1 g/m2 every 12 h, days 1 – 5 (MA protocol), between October 1992 and May 1996 (Br J Haematol 1998;102:1015 – 24). Results: Between March 1998 and January 2000, 29 patients were included (14 MDS-AML, eight RAEB-T and seven RAEB). Median age was 55 years (range: 32 – 70). Six patients (21%) died from the procedure, 16 (55%) achieved CR. In these patients, median duration of leukopenia, neutropenia, and thrombocytopenia were 29 (range: 22 – 45), 30 (range: 22 – 50), and 42 (range: 25 – 99), respectively. Of the 16 patients who achieved CR, nine received consolidation CT, five were autografted and two were allografted in first CR. ASCT was performed in five of the eight patients B 60 years with no HLA identical donor and who achieved CR. Reasons for not performing ASCT in the three remaining patients were: early relapse in two patients and poor marrow stem cell harvest in one patient. The Kaplan – Meier (KM) estimate of overall survival was 72 9 8% at 6 months, 32 9 9% at 12 months, 8 9 8% at 24 months. In the 26 successfully karyotyped patients, the estimated 1-year survival rates were 56, 25 and 0% for patients in good (n = 9), intermediate (n =8) and poor risk (n =9) IPSS cytogenetic subgroups, though the difference was not significant (P= 0.08). By Cox’s univariate analysis, abnormal karyotype was the only factors associated with poor survival (P =0.0088). There was no significant difference in the CR rate (55 vs. 47%), treatment-related mortality (21 vs. 25%), duration of leukopenia (mean: 26 vs. 24 days), neutropenia (mean: 27 vs. 25 days), ASCT feasibility (62 vs. 57%), relapse-free survival (median: 7.4 vs. 14 months, P=0.2) or overall survival (median: KM 11.4 vs. 14 months) between the FAM and MA protocols, respectively. Duration of thrombocytopenia was significantly longer in the FAM protocol (mean 33 vs. 25 days, P= 0.027). Conclusion: Present results suggest that addition of fludarabine to mitoxantrone and AraC had no effect on CR rate, survival, or DFS in high-risk MDS.

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

Or50 FLAG-IDA FOLLOWED BY INTENSIVE POSTREMISSION CHEMOTHERAPY WITH/WITHOUT AUTOLOGOUS HEMATOPOIETIC STEM CELL TRANSPLANTATION (HSCT) IN HIGH-RISK MDS AND AML SECONDARY TO MDS (sAML): PRELIMINARY RESULTS G F Sanz, J M Ribera, M C del Can˜ izo, L Palomera, R Garcia-Boyero, M J Sayas, C Rayon, M Tormo, R Andreu, M Batlle, J J Lahuerta, A Bendan˜ a, M Perez-Sanchez, M J Arilla, J J Moneva, L Benlloch, M A Sanz On behalf of, Spanish PETHEMA Cooperative Group, Valencia, Spain Contact e-mail:sanz – [email protected] Intensive AML-type chemotherapy (CT) with/without autologous HSCT could be curative in some patients with high-risk MDS or sAML. The main aim of this study was to evaluate in a large series of patients with high-risk MDS and sAML the results of a protocol including FLAG-IDA remission induction CT followed by intensive post-remission CT. Eligible patients were those with high-risk de novo MDS (IPSS score \ 1 and/or Spanish PSS score \ 2) or sAML. Induction regimen consisted in one cycle of FLAG-IDA (fludarabine 30 mg/m2 per day × 4 days, cytarabine (ARAC) 2 g/m2 per day × 4 days, idarubicin (IDA) 10 mg/m2 per day × 3 days and lenograstim (G-CSF) 300 mg/m2 per day, days − 1 to 5 and 263 mg per day from day 11). Patients in remission received consolidation CT with IDA +ARAC (3×7)+ G-CSF. Patients under 65 years of age in whom sufficient numbers of hematopoietic progenitor cells from mobilized peripheral blood or bone marrow were collected proceed to autologous HSCT; the remaining received intensification CT with carboplatin (300 mg/m2 per day × 4 days)+ G-CSF. Fifty-six patients (median age 62 years (range: 17 –79); MDS 30 and sAML 26; IPSS intermediate-1 in one, intermediate-2 in nine and high in 13; Spanish PSS intermediate in 11 and high in 19) were included. Thirty-two patients (57%) attained complete remission (CR), 11 (19%) had resistant disease (partial remission in eight) and 13 (23%) died in aplasia. Ten patients had relapsed, five died in CR and 13 remain alive and disease-free 1 + to 19+ months after CR. Median DFS was 10 months and DFS rate at 1 year was 44 911%. Ten patients had received an autologous HSCT and three a HSCT from an unrelated donor. Overall survival (OS) rate at 1 year was 44 98% in all patients, 58 911% in patients who achieved CR and 66 920% in those who attained partial remission. No factor clearly influenced CR, DFS, and OS rates. In summary, the CR rate with FLAG-IDA is acceptable. The protocol seems to eliminate the prognostic importance of recognized variables, though a longer follow-up and more patients are required to assess the overall efficacy of this strategy.

Or51 THE ROLE OF MITOCHONDRIA IN IRON TRANSPORT AND SIDEROBLAST FORMATION N Gattermann Heinrich-Heine-Universita¨ t Du¨ sseldorf, Du¨ sseldorf, Germany Contact e-mail:[email protected] Virtually all the iron that enters an erythroblast via the transferrin receptor goes into the mitochondria, where it is incorporated into protoporphyrin IX to make heme. Iron import into mitochondria is an energy-dependent process. It does not require ATP, but depends on the membrane potential of the inner mitochondrial membrane, which in turn is created by the activity of the mitochondrial respiratory chain. The transporter for mitochondrial iron import has not yet been identified. Iron appears to pass through the inner mitochondrial

S19

membrane in the reduced state (Fe2 + ). The reduced state of iron is also a prerequisite for its incorporation into protoporphyrin IX by ferrochelatase, an enzyme of the inner mitochondrial membrane. We have proposed that impairment of the respiratory chain, caused by mutations of mitochondrial DNA (mtDNA), disturbs mitochondrial iron metabolism in a way that hinders the reduction of iron or its maintenance in the reduced state. If iron is not kept in the right chemical form, it cannot be utilized by ferrochelatase and will therefore accumulate in the mitochondrial matrix. Our pathophysiological model is supported by the mtDNA mutations that we detect in the bone marrow of MDS patients, and also by a recent finding reported by Shirihai et al. (EMBO J 2000;19:2492 – 2502). These authors looked for genes that are selectively expressed in erythropoietic cells under the influence of the GATA-1 transcription factor. Among such genes was the gene for uncoupling protein 2 (Ucp2). Uncoupling proteins dissipate the strong proton gradient that is created by respiratory chain complexes pumping protons out of the mitochondrial matrix. The proton gradient forms the basis of mitochondrial energy production. Uncoupling proteins allow protons to slip back into the mitochondrial matrix, thereby stimulating the respiratory chain to run faster to restore the gradient. This produces heat and consumes oxygen. We propose that in erythropoietic cells, partial uncoupling of the respiratory chain by Ucp2 and the corresponding increase in oxygen consumption may be necessary to create a lowoxygen environment in the mitochondrial matrix in order to maintain iron in the reduced state until it reaches ferrochelatase. Respiratory chain defects, caused by mtDNA mutations, will impair mitochondrial oxygen consumption and may thereby promote reoxidation of iron in the mitochondrial matrix, thus contributing to sidero-blast formation. MDS with 50% ring sideroblasts was diagnosed in a 58-year-old female in 1992. Because of fluctuating peripheral monocytosis, the MDS was hard to classify (RARS/CMML). The karyotype was normal. We identified an acquired hetero-plasmic mutation of mitochondrial DNA (mtDNA) in a bone marrow aspirate from 1993. Heteroplasmy, i.e. coexistence of mutant and wildtype mtDNA in individual cells, is characteristic of mitochondrial DNA diseases. The mutation T6721C (Met- \Thr) affected the gene of subunit I of cytochrome-c-oxidase (Blood 1997;90:4961). The patient became transfusion dependent in 1996. In 4/1999, a slight increase in blast cells (6%) was noted in the marrow and erythropoiesis had become scarce ( B 10%). In two of 2000, a medullary blast count of 39% indicated transformation into AML (M4) (54% blasts 2 months later). The karyotype remained normal. Intensive chemotherapy did not achieve complete remission. After a second trial of CTx, the marrow did not recover from aplasia and the patient died from pneumonia and septicemia. On retrospective analysis of several bone marrow samples taken over the course of the disease, the proportion of mtDNA molecules carrying the mutation (the degree of heteroplasmy) showed little change from 1993 ( 50%) to 1999 (  40%), but decreased to B20% when AML developed. A point mutation of N-ras (exon XXX) was identified through DNA sequencing in biopsies from 1999 and 2000, but was not detectable in earlier samples. The N-ras mutation was thus associated with leukemic transformation. We assumed that this lesion had provided an additional proliferative advantage to a preleukemic subclone, which then became dominant. The decline of the mtDNA mutation might have been attributable to selective pressure against the mtDNA mutation under circumstances of accelerated cellular proliferation. However, to our surprise, X-chromosome inactivation analysis (HUMARA) revealed a different allelic pattern in the leukemic marrow versus the original MDS clone. The switch in clonality indicates that the leukemic clone originated from an unrelated hematopoietic stem cell. This finding suggests that in some cases of MDS, several independent stem cells may have accumulated genomic damage that predisposes them to leukemic transformation. We can only speculate whether stem cell damage in the present case may be related to gold therapy (oral and parenteral) that the patient received 1984 – 1986 for psoriatic arthritis. She never received methotrexate or alkylating agents.

S20

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

Or52 INDICATIONS AND CONTROVERSIES IN ALLOGENEIC STEM CELL TRANSPLANTATION FOR PATIENTS WITH MDS J E Anderson Department of Medicine/Division of Hematology, University of Texas Health Science Center at San Antonio, San Antonio, USA Contact e-mail:[email protected] The foremost indication for allogeneic stem cell transplantation (SCT) in MDS is simply the diagnosis of MDS and the desire for curative therapy. Although a very small subset of patients with newly diagnosed RAEB or RAEBT with normal karyotype may be cured with conventional chemotherapy and a small subset of patients with low or intermediate-1 risk MDS experience long-term survival with supportive care, long-term cure of MDS is only realistic with allogeneic SCT. The more difficult decision-making is choosing which MDS patients are candidates for SCT and when during the disease course SCT is appropriate. Non-randomized data suggest that longterm survival is better among patients B 60 years of age with IPS intermediate-1, 2, and high risk MDS who receive allogeneic SCT compared to no intensive therapy. Furthermore, the risk of relapse is B5% among patients with IPS intermediate-1 disease. Many studies suggest that longer disease duration is associated with greater non-relapse mortality and, in some studies, longer disease-free survival (DFS). However, some studies show no difference in DFS because of an inverse association with relapse and disease duration. Nonetheless, these data support a decision to perform transplant early in the disease course for all but IPS low risk patients, for whom transplant should be performed in only select situations. The appropriate upper age limit for allogeneic SCT is not defined, but preliminary singlecenter data have reported favorable outcome for patients up to 66 years with the use of busulfan-targeting and cyclophosphamide. Use of this preparative regimen has also resulted in a low 100-day mortality for recipients of HLA-matched unrelated donor marrow. The use of induction chemotherapy before the preparative regimen for patients with IPS intermediate-2 and high risk disease is controversial. It is likely that such treatment selects out the good risk patient who can achieve a CR, but does not result in a more favorable DFS compared to that obtained with upfront SCT. Improvement in the complex decision-making process for SCT in MDS patients will require prospective, randomized clinical trials.

Or53 ALLOGENEIC HEMATOPOIETIC TRANSPLANTATION IN MDS. RESULTS OF THE SPANISH GROUP FOR HEMATOPOIETIC TRANSPLANTATION M C del Can˜ izo1, S Brunet2, E Conde3, C Martı´nez4, G Sanz5, C Vallejo 6, M V Mateos1 1 Hospital Universitario, Hematology, Salamanca, Spain, 2Hospital Santa Creu e Sant Pau, Hematology, Barcelona, Spain, 3Hospital Valdecilla, Hematology, Santander, Spain, 4Hospital Clinic, Hematology, Barcelona, Spain, 5Hospital La Fe, Hematology, Valencia, Spain, 6Hospital Provincial, Hematology, Murcia, Spain Contact e-mail:[email protected] In the present study, we have analyzed 68 patients with myelodysplastic syndrome (MDS) included in the Spanish Registry in whom an allogeneic hematopoietic stem cell transplantation have been performed between May 1985 and December 2000. The diagnosis according to Fab classification was as follows: Ten refractory anemia (RA), two refractory anemia sideroblastic (RARS), 28 refractory anemia with excess of blast (RAEB), 16 refractory anemia with excess

of blast in transformation (RAEBt), three chronic mielomonocityc leukemia (CMML) and three secondary MDS. The remaining seven cases corresponded to variants not included in the FAB, such as hypoplastic MDS and MDS with marrow fibrosis. The conditioning regimen was Cy +TBI in 32 cases and Bu + Cy in 29. In the remaining eight cases other different conditioning were used. GVHD prophylaxis was mainly based on CSP + MTX or CSP+ Prednisone. Peripheral blood (PB) stem cells were used in 27 cases while the other 41 patients received bone marrow (BM) mononuclear cells. Sixty five percent of patients showed acute GVHD ] and 28% of cases developed extensive chronic GVHD. The median overall survival from diagnosis was 39 months while from transplantation it was only 12 months. Upon analyzing the influence of FAB categories on survival from transplant it was observed that, although patients with favorable FAB categories (RA + RARS) had longer survival than patients with other prognostic groups (median time 37 vs. 7 months) differences were not statistically significant (P= 0.1). Moreover, neither the year of transplantation (before or after 1997) nor the time elapsed from diagnosis to transplantation (more or less of 6 months) had a significant influence (PB0.1 and \ 0.05). By contrast, interestingly, the source of stem cells used for transplantation showed a significant influence on survival (patients transplanted with PB 42 months versus BM 4 months; P= 0.004). Our results show that in MDS patients receiving allogeneic stem cell transplantation the use of PB in stead of BM as the source of precursor cells is associated with a significant prolongation on survival.

Or54 THE PROGNOSTIC IMPACT OF THE PRESENCE OF AN HLA-IDENTICAL DONOR ON OUTCOME OF PATIENTS WITH HIGH RISK MDS OR SECONDARY AML (SAML) TREATED WITH INTENSIVE CHEMOTHERAPY. A JOINT STUDY OF THE EORTC, EBMT, SAKK AND GIMEMA LEUKEMIA GROUPS M Oosterveld1, T De Witte2, S Suciu3, G Verhoef4, B Labar5, A Belhabri6, C Aul7, D Selleslag8, A Ferrant9, P Wijermans10, F Mandelli11, S Amadori12, U Jehn13, P Muus2, R Zittoun14, A Gratwohl15, H Zwierzina16, O Anak3, R Willemze17 1 University Medical Center, Nijmegen, Netherlands, 2University Medical Centre St. Radboud, Nijmegen, Netherlands, 3EORTC Data Center, Brussels, Belgium, 4University Hospital Gasthuisberg, Leuven, Belgium, 5Clinical Hospital Rebro, Zagreb, Croatia, 6Hopital Edouard Herriot, Lyon, France, 7Heinrich Heine University, Dusseldorf, Germany, 8St. Jan’s Hospital, Brugge, Belgium, 9Clinique Universitaire St. Luc, Brussels, Belgium, 10Leijenburg Hospital, The Hague, Netherlands, 11University La Sapienza, Roma, Italy, 12 Ospedale Eugenio, Roma, Italy, 13Ludwig-Maximilian University, Munich, Germany, 14Hopital Dieu, Paris, France, 15Kantonsspital Basel, Switzerland, 16Medizinische Klinik, Innsbruck, Austria, 17 Leiden University Medical Centre, Leiden, Netherlands Contact e-mail:[email protected]nl Between November 1992 and March 1997, 184 evaluable patients from 35 European centres participated in the EORTC 06921 study. Treatment consisted of remission-induction chemotherapy followed by one consolidation course. HLA family typing was to be initiated at the onset of therapy. All patients with an HLA-identical donor were candidates for allogeneic BMT either in CR-1 after the consolidation course or as salvage therapy if no CR was achieved. The upper age limit for allogeneic BMT varied from 45 to 60 years depending on the policy of the centres. Patients in CR without a donor were scheduled for autologous bone marrow transplantation or peripheral blood stem cell transplantation. The question is whether the presence of an

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22 HLA-identical sibling improves the outcome of MDS patients assuming that the majority of patients with an HLA-identical donor have been transplanted indeed. The actuarial survival rate at four years of the 184 evaluable patients was 26% (S.E. = 3.5%) with a median follow-up of 3.6 years. The disease-free survival rate at 4 years from CR was 29% (S.E. = 4.9%). Thirty-six of the 56 patients with an HLA-identical donor received an allograft (28 in CR-1). Sixty-eight patients did not have an HLA-identical donor. Autologous stem cell transplantation was performed in 32 patients (27 in CR-1), nine patients received an allograft with an alternative donor as salvage treatment. In 60 patients, the availability of a donor was unknown. The majority of these patients were older (72% \50 years). Only 32% reached CR after the remission-induction course and 32% died in hypoplasia. Nine patients underwent autologous stem cell transplantation in first CR. The actuarial survival at 4 years of the 56 patients with a donor was 37% (S.E. =6.3%) versus 31% (S.E. = 6.4%) for the 68 patients without a donor (P = 0.14). DFS from CR was 33% (S.E. = 7.7%) versus 31% (S.E. =7.9%). This analysis shows that intensive antileukemic treatment for young patients with poorrisk MDS results in a promising DFS. The outcome of patients without an HLA-identical donor may not be inferior compared to the outcome of patients with a donor, because the majority of them received either an autologous transplant or an allograft from other donor sources.

Or55 CHROMOSOMAL CHARACTERISTICS ARE AN INDEPENDENT RISK FACTOR FOR RELAPSE AFTER HLA-IDENTICAL SIBLING STEM CELL TRANSPLANTATION IN PATIENTS WITH MYELODYSPLASIA OR SECONDARY ACUTE MYELOID LEUKEMIA V Runde1, T de Witte2, R Brand3, A van Biezen3, A Bacigalupo4, T Ruutu5, R Willemze6, P Guardiola7, M Boogaerts8, C Cordonnier9, E Allesandrino10, A Gratwohl11, J Reiffers12, J Finke13, R Arnold14, D Niederwieser15 1 Department of Bone Marrow Transplantation, University Hospital Essen, Essen, Germany, 2Department of Hematology, University Hospital Nijmegen, Nijmegen, Netherlands, 3Department of Medical Statistics, University Hospital Leiden, Leiden, Netherlands, 4 Department of Hematology, Ospedale San Martino, Genova, Italy, 5 Department of Hematology, University Hospital Helsinki, Helsinki, Finland, 6Department of Hematology, University Hospital Leiden, Leiden, Netherlands, 7Department of Hematology, Hopital St. Louis, Paris, France, 8Department of Hematology, University Hospital Leuven, Leuven, Belgium, 9Department of Hematology, Hopital Henri Mondor, Creteil, France, 10Department of Hematology, Policlinico San Matteo, Pavia, Italy, 11Department of Hematology, University Hospital Basel, Basel, Switzerland, 12Department of Hematology, Hopital du Haut Leveque, Pessac, France, 13Department of Hematology, University Hospital Freiburg, Freiburg, Germany, 14Department of Hematology, University Hospital Charite´ , Berlin, Germany, 15 Department of Internal Medicine, University Hospital Leipzig, Leipzig, Germany Contact e-mail:[email protected] In 1073 patients with MDS or sAML after allogeneic stem cell transplantation (SCT) from an HLA-identical sibling with detailed information about stage and FAB classification at time of SCT analysis was performed. Some 557 patients were male, median age of the patients was 38 (range: 0 –62) years. In 163 cases, the origin of the disease was explained by a prior cytotoxic exposure. FAB diagnosis at SCT was RA/RARS (n =167), RAEB (n= 198), CMML (n =61), RAEB-t (n = 252) and sAML (n=395). A total of 35% of the patients underwent SCT without prior remission induction treatment

S21

(untreated), 31% in first CR (CR1) and 34% in more advanced stages after induction therapy (noCR1). Cytogenetic data were available in 566 cases. Nine cases with inv(16) and four cases with t(8;21) were excluded from further analysis. The remaining karyotypes were categorized as standard risk when they were normal (n =256), high risk when complex aberrations and/or abnormalities involving chromosome 7 were present (n = 136) and intermediate risk for other abnormalities (n =161). Median interval between diagnosis and SCT was 6 (range: 0 – 218) months. Source of stem cells was blood in 229 cases. T-cell depletion was reported in 266 cases. Multivariate analysis in a COX model included variables with significant influence in univariate analysis: stage of disease at SCT, FAB diagnosis at SCT, chromosomal classification, patient age, year of SCT, cytotoxic etiology, female donor and interval between diagnosis and SCT. Significant (positive) multivariate associations between improved EFS as well as survival were found for stage (CR1 \noCR1/untreated), FAB subtype (decreasing), cytotoxic etiology (absence \presence), age (younger \ older) and year of transplants (recent \earlier). An increased relapse risk was associated with stage (noCR1/untreated\ CR1), FAB subtype (decreasing), and chromosomal risk category (standard \ intermediate \high). TRM was associated with age (old \ young) and cytotoxic etiology (present \absent). Conclusion: In addition to well known risk factors, we demonstrate that cytogenetic characteristics independently influence the risk of relapse after HLA-identical sibling SCT in MDS/sAML and that a cytotoxic etiology of MDS/sAML increases the risk of TRM and decreases survival and EFS.

Or56 ALLOGENEIC STEM CELL TRANSPLANTATION (SCT) IN MDS: INTERIM RESULTS OF OUTCOME FOLLOWING NON-MYELOABLATIVE CONDITIONING COMPARED TO STANDARD PREPARATIVE REGIMENS J E Parker1, T Shafi1, A Mijovic1, S Devereux1, A Pagliuca1, M Garg2, J A Yin2, M Potter3, H G Prentice3, J Byrne4, N H Russell4, G J Mufti1 1 King’s College Hospital, Haematology, London, UK, 2Manchester Royal Infirmary, Haematology, Manchester, UK, 3Royal Free Hospital, Haematology, London, UK, 4Nottingham City Hospital, Haematology, Nottingham, UK Contact e-mail:j – e – [email protected] Outcome following allogeneic SCT was evaluated in 29 MDS patients conditioned with standard myeloablation compared to 23 receiving non-ablative regimens. Eighteen had RA; 15 RAEB; 15 RAEB-t/ CMML-t/MDS-AML; four t-MDS. Thirty received induction chemotherapy, with 15 (50%) achieving CR. SCs were infused from siblings (28), Alternative family donors (AFD) (two) or volunteer unrelated donors (VUD) (24). Recipient and donor were HLAmatched in 41 (77%). Eighteen sibling allografts were conditioned with busulphan (16 mg/kg) and cyclophosphamide (200 mg/kg) (Bu/ Cy) while AFD and nine VUD recipients received Bu (4 mg/kg), Cy (120 mg/kg), TBI (1440cGy) and Campath-1G (100 mg, iv). Ten siblings and 14 VUD recipients underwent non-ablative conditioning (fludarabine 150 mg/m2, busulphan 8 mg/kg, Campath-1H 100 mg, iv (22); fludarabine, idarubicin, cytarabine, G-CSF (2)) due to advanced age (47.5 vs. 37 years, P= 0.002) and/or co-morbidity which would normally preclude transplantation (21 vs. 3, P =0.003). Fifty-one received methotrexate/cyclosporin GvHD prophylaxis and two, cyclosporin only. Donor engraftment (100%) (VNTR/cytogenetic/FISH analysis) was achieved in 22 (92%) and 27 (93%) patients conditioned with non-ablative and ablative protocols respectively. Non-ablative conditioning was associated with significantly delayed onset and reduced duration of aplasia, less mucositis, fever, antibiotic, parenteral nutrition, acute and chronic GvHD and early (day 100)

S22

Abstracts / Leukemia Research 25 Suppl. No. 1 (2001) S1 – S22

procedure-related mortality (2 (8%) vs. 9 (31%)). Six patients relapsed (two ablative, four non-ablative) and two (non-ablative) experienced late graft failure. Twenty-nine (59%) patients are alive 1 – 77 months post-transplant (non-ablative, 16 (67%); ablative, 13 (45%)). Threeyear actuarial DFS is 49% in the non-ablative arm and 43% in the ablative group. Three-year non-relapse mortality is 31 and 51%, respectively. Predictors of DFS included B 10% blasts at diagnosis, good/intermediate-risk karyotype, low IPSS score and CR-attainment. Our data demonstrates that VUD or sibling allogeneic SCT following non-myeloablative conditioning is feasible in high-risk MDS patients considered unsuitable for standard myeloablation and is associated with comparable 3-year DFS to those receiving conventional preparative regimens.

throughout the infusion. Samples will be analyzed for evidence of re-expression of the silenced cell cycle regulatory gene p15. Conclusion: Sequential administration of ‘first generation’ MeTI AC and HDI PB appears feasible and without excessive toxicity. Correlation of molecular endpoints with clinical responses will allow assessment of the ‘molecular efficacy’ of this approach.

Or57 CLINICAL DEVELOPMENT OF HISTONE DEACETYLASE INHIBITORS FOR THE TREATMENT OF MDS AND RESISTANT MYELOID MALIGNANCIES S D Gore1, S B Baylin2, J G Herman2, C B Miller1 1 Johns Hopkins Oncology Center, Hematologic Malignancies, Baltimore, USA, 2Johns Hopkins Oncology Center, Cancer Biology, Baltimore, USA Contact e-mail:[email protected]

N Casadevall1, E Lepage2, P Durieux3, S Dubois1, F Dreyfus4, M C Quarre´ 4, P Fenaux5, G Damaj4, S Giraudier6, A Guerci7, G Laurent8, H Dombret4, A Vekhoff4, V Ribrag9, pour le Groupe Francais des Mye´ lodysplasies (GFM) 1 Hoˆ tel Dieu, Service d’He´ matologie Biologique, Paris, France, 2 Hoˆ pital Henri Mondor, Unite´ d’Information Me´ dicale, Cre´ teil, France, 3HEGP, Qualite´ et Gestion des Risques, Paris, France, 4 GFM, Paris, France, 5GFM, Lille, France, 6GFM, Cre´ teil, France, 7 GFM, Nancy, France, 8GFM, Toulouse, France, 9GFM, Villejuif, France, 10France Contact e-mail:[email protected]

Background: Recent elucidation of the important interaction between the acetylation status of histones and regulation of gene transcription has identified histone deacetylases (HDAC) as potentially important molecular targets for the treatment of malignancies. HDAC has been shown to be directly involved in transcriptional repression induced by leukemia-specific fusion proteins (AML1-ETO, PML-RAR). HDAC is also recruited by proteins such as MeCP2 which bind to methylated DNA; thus HDAC may play an important role in mediating methylation-associated transcriptional repression. A wide variety of genes are methylated and thus silenced in neoplastic cells, including AML and MDS. In addition, HDAC silences nuclear-hormone-dependent gene expression in absence of ligand; this may have an important impact on retinoid-dependent myeloid differentiation. A number of compounds with HDAC inhibitory activity are under clinical investigation. Of these, sodium phenylbutyrate (PB) has undergone the most extensive pharmacokinetic and early stage clinical evaluation. Plasma concentrations achieved during continuous infusions of PB induce acetylation of histones H3 and H4 in bone marrow mononuclear cells, leukemic cell lines, and primary leukemic bone marrow cells. Method: We have previously demonstrated that sequential treatment of cells with DNA methyltransferase inhibitors (MeTI) and histone deacetylase inhibitors (HDI) leads to synergistic re-expression of silenced tumor suppressor genes. To study the clinical feasibility of sequential MeTI/HDI in patients with MDS and AML, a Phase I study was initiated in which 5-azacytidine (AC) is given from 75 to 10 mg/m2 per day SQ for 5 days, followed by PB at 375 mg/kg per day IVCI days 5– 12. Dosing is repeated every 28 days. Results: No unexpected toxicities have been observed to date. Two of 12 patients have exhibited hematologic improvement. Significant inhibition of MeT has been observed in bone marrow mononuclear cells (42 – 85%). Administration of PB at 375 mg/kg per day leads to rapid (within 4 h) acetylation of histones H3 and H4 that is sustained

Context: In MDS, the use of rHuEpo alone is not really effective on anemia (10 – 20% of responses). But, rHuEpo plus G-CSF have a synergic effect demonstrated in vitro and in vivo. The ‘Groupe Francais des Mye´ lodysplasies’ (GFM) decided to confirm these results in a randomized trial that included the analysis of the quality of life and a cost study. Method and results: The study included 60 patients with RA (20), RARS (26) or RAEB (14) (marrow blasts B10%) and serum Epo level B500 UI/ml, randomized between 01.99 and 04.00 in two arms: arm A treated with rHuEpo (Eprex® (JanssenCilag), 20000 UI, 3 × per week by SC route) and G-CSF (Granocyte® 13 (Chugai), 105 mg, 3 × per week by SC route), arm B treated symptomatically. For each patient, treatment duration was 52 weeks. Both drugs were given over 12 weeks. At this time, either the patient was responder (R) and treatment was continued with rHuEpo alone or was no-responder (NR), and treatment was stopped. In case of relapse with rHuEpo alone, G-CSF was reintroduced. Response criteria at 12 weeks were: (a) hemoglobin (Hb) increase ]1.5 g/dl or absence of transfusions with a stable Hb level (partial response (PR)); (b) Hb level ] 11.5 g/dl (complete response (CR)). All 60 patients included in the trial have reached the 12th week and 24 patients have completed the study (ten arm A, 14 arm B). In arm A, 24 patients reached the 12th week (six stopped before week 12). Treatment was stopped in ten patients (NR). Fourteen patients were responders (one CR, 13 PR) and continued with rHuEpo alone. Nine of these patients relapsed and G-CSF was reintroduced. Response was restored in all of them. In arm B, there was no change in Hb level and transfusion needs during the study. Since all patients have not reached the end of the protocol, comparative data on the quality of life and costs are not yet available and will be presented at the meeting. Conclusions: These results confirm the synergy of rHuEpo and G-CSF in the treatment of anemia in myelodysplastic syndromes with a response rate of about 58% at the 12th week of treatment. Data on quality of life and costs are awaited.

Or58 ERYTHROPOIETIN (rHuEpo) PLUS G-CSF IN THE TREATMENT OF ANEMIA IN MYELODYSPLASTIC SYNDROMES: RESULTS OF A RANDOMIZED TRIAL WITH IMPACT ON QUALITY OF LIFE AND COSTS