Leukemia (2001) 15, 377–384  2001 Nature Publishing Group All rights reserved 0887-6924/01 $15.00 www.nature.com/leu Cytogenetic subgroups in acute myeloid leukemia differ in proliferative activity and response to GM-CSF G Jahns-Streubel 1 , J Braess 1 , C Schoch 1 , C Fonatsch 2 , D Haase 3 , C Binder 3 , B Wo ¨ rmann 3 , T Bu ¨ chner 4 and W Hiddemann 1 1 Department of Medicine III, University Hospital Großhadern, Ludwig Maximilians University, Munich; 3 Department of Hematology and Oncology, Georg-August-University, Go ¨ ttingen; 4 Department of Hematology and Oncology, Westfa ¨lische Wilhelms University, Mu ¨ nster, Germany; and 2 Department of Medical Biology, University of Vienna, Austria The current study was undertaken to search for differences in the biology of cytogenetic subgroups in patients with de novo acute myeloid leukemia (AML). In addition, factors influencing the metabolism of cytosine arabinoside (araC) as the key agent of antileukemic activity were assessed. Bone marrow aspirates from 91 patients with newly diagnosed AML in whom kary- otypes were successfully obtained were analyzed: (1) for spon- taneous proliferative activity by 3 H-thymidine ( 3 H-TdR) incor- poration; (2) proliferative response to GM-CSF by in vitro incubation of blasts for 48 h with or without GM-CSF (100 U/ml) followed by an additional 4-h exposure to 3 H-TdR (0.5 Ci/ml); and (3) parameters of araC metabolism comprising 3 H-araC uptake in vitro and the activities of polymerase alpha (poly ), deoxycytidine kinase (DCK) and deoxycytidine deaminase (DCD). According to the results of chromosome analyses four cytogenetic subgroups were discriminated: (I) normal kary- otypes (n = 38); (II) favorable karyotypes [t8;21), t(15;17), inv(16)] (n = 16); (III) unfavorable karyotypes [inv (3), -5, 5q-, t(6;9), + 8, t (9;11), complex abnormalities] (n = 20); (IV) kary- otypes of unknown prognostic significance (n = 17). Proliferat- ive activity of leukemic blasts was significantly higher in favor- able karyotypes (group II) as compared to cases with unfavorable cytogenetics (group III) with median values and range for 3 H-TdR uptake in group II of 2.48 pmol/10 5 cells (0.28– 25.8) and in group III of 0.51 pmol/10 5 cells (0.04–7.6) (P = 0.0096). The respective values in group I and group IV were 0.7 pmol/10 5 cells (0.0–6.7) and 0.98 pmol/10 5 cells (0.0–4.0), respectively. Inversely, response to GM-CSF, as defined by an increase in 3 H-TdR incorporation 1.5- fold over control values after 48 h of GM-CSF exposure, was significantly lower for patients with a favorable karyotype (group II) as compared to group I (P = 0.04) and group III (P = 0.013). No significant differ- ences between karyotype groups I, II, III and IV were found for 3 H-araC incorporation, nor for the activities of poly , DCK and DCD. These data demonstrate differences in the biology of cytogenetic subgroups in AML which may partly explain the well established differences in clinical outcome. Leukemia (2001) 15, 377–384. Keywords: cytogenetics; acute myeloid leukemia; proliferative activity; growth factors Introduction With the development of intensive induction chemotherapy and the improvement of supportive care, complete remissions (CR) are nowadays achieved in 50% to 80% of adult patients with acute myeloid leukemia (AML), but only 20% to 30% of cases experience long-term disease-free survival and potential cure. 1–4 Increasing insights into the biology of AML indicate that distinct cytogenetic and molecular aberrations can be identified that are closely related to treatment outcome. Cyto- genetic analyses thus allow discrimination of subgroups of patients with favorable, unfavorable and intermediate prog- Correspondence: W Hiddemann, Department of Medicine III, Univer- sity Hospital Grohadern, Ludwig Maximilians University, Marchioni- nistr 15, 81377 Mu ¨ nchen, Germany; Fax: 89 7095 8875 Received 3 April 2000; accepted 29 September 2000 nosis and provide the means to develop risk-adapted treat- ment strategies. 5–9 New therapeutic approaches that have been investigated recently include the application of cytosine arabinoside (araC) at high doses 10,11 and preliminary data indicate substantial dif- ferences in the efficacy of this treatment between distinct cyto- genetic subgroups. 12–17 High expectations have also been associated with the priming of leukemic blasts by hematopo- ietic growth factors to enhance their sensitivity for subsequent chemotherapy. Although this approach was supported by promising in vitro data, 18–22 the available data from prospec- tive clinical studies are mostly disappointing. 23–30 Such a glo- bal conclusion, however, appears too simplistic and neglects potential differences in the efficacy of priming concepts for distinct subtypes of AML. In particular, no data are yet avail- able to relate the use of hematopoietic growth factors to pre- therapeutic determinants of prognostic or biologic relevance such as cytogenetics. Overall, more information about the biology of AML, the determinants of response to cytostatic drugs, such as araC, and the mechanisms underlying the prognostic significance of cytogenetic subgroups is needed to develop appropriate regi- mens for patients facing different long-term expectations. A variety of biologic characteristics of leukemic blasts have been investigated with regard to their predictive value for treatment response. Analyses about the major determinants of araC metabolism in leukemic cells led to different conclusions about their relevance to treatment outcome. While some stud- ies found DCK activity in AML blasts to be predictive for thera- peutic response 31,32 other groups suggested that some still unidentified mechanism of metabolic control rather than the amount of active enzyme determines the level of araC incor- poration and cytotoxicity. 33,34 The same controversy also relates to DCD activity. A recent investigation by our group 35 confirmed earlier studies and found a high DCD activity to be predictive for resistance to therapy 31,36 while this association was rejected by other authors. 37,38 Conflicting data have similarly been reported about the association between the proliferative activity of leukemic blasts and clinical outcome. 39–41 Lo ¨ wenberg et al 42 described an inverse relation between proliferative activity of leukemic blasts and long-term prognosis while our group could demon- strate that patients with a high proliferative activity of AML blasts in vitro have a higher probability to experience an adequate response to induction therapy as compared to cases with a low spontaneous proliferation rate. 43 Sensitivity of blast cells to GM-CSF in vitro appeared to be inversely related with early treatment response which may be explained by the autocrine production of hematopoietic growth factors in AML blasts. 44–47 These results were confirmed by Tsuzuki and coworkers, 48 showing that patients whose leukemic cells had a positive proliferative response to growth factors had a poorer outcome. In all these studies the only generally accepted determinant