Professor G. J.Mufti Department of Haematology, King's College Hospital and School of Medicine and Dentistry, Bessemer Road, London SE5 9RS.
We performed flow cytometric analysis of CD34+ cell apoptosis in 59 patients with myelodysplastic syndrome (MDS) or acute myeloid leukaemia (AML) secondary to MDS (MDS-AML) using annexin V-FITC, which binds to exposed phosphatidylserine on apoptotic cells. Apoptosis was significantly increased in FAB subtypes RA, RARS and RAEB (<10% blasts) (56.5% (15.1–86.5%)) compared to normal controls (18.5% (3.4–33.4%), P < 0.0001) and RAEB-t/MDS-AML (16% (2.1–43.2%), P < 0.0001). There was no correlation between % apoptosis, Full blood count or cytogenetics in any disease category. Two-colour cytometric analysis of permeabilized CD34+ cells stained with antibodies to Bcl-2, Bcl-X (anti-apoptotic), Bax and Bad (pro-apoptotic), demonstrated significantly higher ratios of pro- v anti-apoptotic proteins in early MDS (2.47 (1.19–9.42) compared to advanced disease (1.14 (0.06–3.32), P = 0.0001). Moreover, using repeated measures of variants (ANOVA), we found that variations between individual Bcl-2-related proteins differed significantly according to disease subtype (P < 0.0005). Our results confirm that CD34+ cell apoptosis was significantly increased in MDS subtypes RA and RARS and fell with disease progression. Early MDS was also associated with a significantly higher CD34+ cell pro- v anti-apoptotic Bcl-2-family-protein ratio than advanced disease. Furthermore, patterns of expression of individual Bcl-2 related proteins differed significantly between different disease categories. However, no correlation between pro- v anti-apoptotic Bcl-2-family-protein ratios and the degree of apoptosis was observed.
The myelodysplastic syndromes (MDS) are clonal disorders characterized by dysplastic haemopoiesis and a propensity towards leukaemic transformation ( Mufti & Galton, 1986). Despite a normo- or hypercellular marrow in the majority of cases, most patients have peripheral blood cytopenias. This paradox has recently been ascribed to excessive haemopoietic cell apoptosis ( Yoshida, 1993; Raza et al, 1995 ). In turn, leukaemic progression could result from acquisition of genetic lesions which inhibit programmed cell death (PCD), thereby allowing illegitimate survival of the neoplastic clone.
PCD can be triggered by a wide variety of extracellular stimuli and intracellular signals. Major regulators of these apoptotic pathways are members of the Bcl-2 family. These proteins function as homo- or heterodimers and it is the ratio of pro- v anti-apoptotic molecules that determines a cell's susceptibility to death signals ( Yang & Korsmeyer, 1996). Deregulation of Bcl-2 family members has been demonstrated in a variety of human cancers and increased expression of the anti-apoptotic proteins Bcl-2 and Bcl-X has been linked to the transition from a pre-malignant to a frankly neoplastic clone in some studies ( Ter Harmsel et al, 1996 ; Krajewska et al, 1996 ). It is feasible, therefore, that altered expression of these oncoproteins is responsible for increased apoptosis and ineffective haemopoiesis in early MDS and/or the leukaemic progression witnessed in some patients. To this end, Rajapaksa et al (1996 ) have recently demonstrated reduced Bcl-2 expression in CD34+ cells of patients with early MDS subtypes compared to those with advanced disease and normal controls.
In the present study the role of apoptosis in the pathogenesis of MDS was evaluated using two-colour flow cytometry and fluorescein isothiocyanate (FITC) conjugated annexin V, a protein which binds specifically to exposed phosphatidylserine on the surface of apoptotic cells ( Koopman et al, 1994 ). Following cell permeabilization, we also measured intracellular expression of two pro- and two anti-apoptotic Bcl-2 family members, Bax and Bad, Bcl-2 and Bcl-X respectively, within haemopoietic progenitor cell (CD34+) populations in order to evaluate whether deregulated apoptosis could be attributed to altered expression of the Bcl-2-related proteins.
Table 1. Table I. Comparison of clinical and laboratory parameters in early MDS (RA, RARS, RAEB <10% blasts), late MDS (RAEB >10% blasts, RAEB-t, MDS-AML) and de novo AML.
Bone marrow (BM) aspirate samples were obtained from 59 patients with MDS for determination of apoptosis and Bcl-2-related protein expression. BM from 25 patients with de novo AML and 20 normal donors were also evaluated. MDS patients were subdivided into early (refractory anaemia (RA) (n = 29)/RA with ringed sideroblasts (RARS) (n = 4)/RA with excess blasts (RAEB) (blasts <10% and good-risk cytogenetics) (n = 3)) and advanced (RAEB (blasts >10% and/or poor-risk cytogenetics) (n = 4)/RAEB in transformation (RAEB-t) (n = 1) and MDS transformed to AML (MDS-AML) (n = 18) subtypes according to the French–American–British (FAB) morphological guidelines ( Bennett et al, 1982 ) and a recent international prognostic scoring system (IPSS) ( Greenberg et al, 1997 ). Similarly, AML patients were subclassified (FAB M0–M7) according to morphology and immunophenotype ( Bennett et al, 1985 ). The median age of patients with ‘early’ and ‘late’ MDS was 64.5 years (range 19–92 years) and 66 years (range 17–81 years) respectively, compared to a median age of 51 years (range 27–96 years) in patients with de novo AML and 54.5 years (range 15 months to 76 years) for normal controls. Based on the IPSS ( Greenberg et al, 1997 ), 28/35 (80%) evaluable patients with early MDS had ‘good-risk’ cytogenetics (normal karyotype or isolated deletions of the long arm of chromosome 5 or 20 or −Y), compared to only 6/23 (26.1%) with advanced disease. Of this latter group, in contrast, 11/23 (47.8%) harboured a ‘poor-risk’ karyotype (more than two chromosomal abnormalities and/or abnormalities of chromosome 7) with 9/23 (39.1%) demonstrating abnormalities of chromosome 7. Patients with early MDS had higher haemoglobin, white cell and platelet counts than in patients with advanced disease and de novo AML. For quantification of apoptosis, only bone marrow samples <6 h old were used, whereas for measurement of intracellular Bcl-2-related proteins, samples were analysed up to 24 h after aspiration.
Bone marrow mononuclear cells (MNC) were isolated by density gradient centrifugation over Ficoll-Paque (Pharmacia Biotech) and washed twice in phosphate-buffered saline (PBS) (Sigma). 106 cells were used per sample. Cells were incubated with phycoerythrin (PE) conjugated CD34 monoclonal antibody (Mab) (Becton Dickinson) for 20 min at room temperature in the dark and then washed twice in PBS. Pelleted cells were resuspended in 390 μl binding buffer (10 m m Hepes/NaOH, pH 7.4, 140 m m NaCl, 2.5 m m CaCl2) (Bender Medsystems, Boehringer Ingelheim) and incubated with 10 μl FITC-conjugated annexin V (Bender Medsystems, Boehringer Ingelheim) for 10 min at room temperature in the dark. Cells were then washed once in PBS and resuspended in 400 μl binding buffer prior to flow cytometric analysis. Negative controls included cells incubated with neither CD34-PE Mab nor annexin V-FITC and cells incubated with CD34-PE Mab only.
Bcl-2-related protein analysis
For cell surface and intracellular protein staining, whole bone marrow (BM) containing 106 white blood cells (WBC) per sample were used. Cells were incubated with FITC conjugated CD34 Mab (Becton Dickinson) for 20 min at room temperature in the dark and then washed twice in PBS. Cells were then fixed in 100 μl ‘medium A’ containing formaldehyde (Fix & Perm, Caltag Laboratories, TCS Biologicals) for 15 min at room temperature in the dark and then washed in PBS. To prevent non-specific binding of monoclonal antibodies to Fc receptors, cells were pre-incubated with 10 μl rabbit anti-mouse (RAM) immunoglobulins (Dako) for 10 min. 100 μl permeabilization ‘medium B’ (Fix & Perm, Caltag Laboratories, TCS Biologicals) was subsequently added for a further 10 min prior to washing in ice-cold PBS to which 2.5% fetal calf serum (Sigma) and 0.1% sodium azide (Sigma) had been added (PBSF). To detect intracellular Bcl-2-related protein expression, cells were incubated with monoclonal antibodies to Bax (Immunotech, Coulter), Bcl-2 (Dako), Bad and Bcl-X (Transduction Laboratories, Affiniti Research Products) for 30 min on ice. After washing in ice-cold PBSF, cells were then incubated with PE-conjugated RAM (Dako) for 30 min on ice, washed again and resuspended in ice-cold PBSF prior to flow cytometric analysis. Negative controls were performed by incubating cells with isotype-specific antibodies (Sigma).
Flow cytometric analysis
Two-colour analysis of apoptosis and Bcl-2-related protein levels within CD34+ cell populations was performed using an EPICS XL flow cytometer (Coulter). Data was acquired in listmode, acquiring at least 2 × 105 events per sample. Analysis was based on gating of subpopulations of cells by forward scatter versus side scatter as well as side scatter versus fluorescence 2 (CD34-PE) for apoptosis detection and fluorescence 1 (CD34-FITC) for Bcl-2-related protein analysis. Percentage positivity was determined by comparison of the fluorescence distribution histogram of positively stained cells to that of cells to which no annexin V was added for apoptosis, and to isotype control stained cells for Bcl-2-related protein analysis. The mean fluorescence intensity (MFI) value for each Bcl-2-related protein was calculated by dividing the MFI value of the positively stained cells by that of cells stained with an isotype control antibody. To determine individual protein levels, we calculated a Protein Index (PI) as previously described ( Rajapaksa et al, 1996 ), whereby PI equalled the product of the percent positive cells and MFI.
For comparison of Bax/Bad:Bcl-2/Bcl-X ratios and the degrees of apoptosis between patients within clinically different disease subgroups, the Mann-Whitney test was used, whilst correlations with clinical and laboratory parameters were calculated using Spearman Rank Test. In order to determine whether patterns of Bcl-2-related protein expression varied within the different diagnostic groups, we used repeated measures of variants (ANOVA), with Bcl-2-related protein as ‘within subjects factor’ and disease subgroup as ‘between subgroup factor’. A P value of <0.05 was considered significant.
Degree of apoptosis
As shown in the representative set of experiments (Fig 1), within the CD34+ cell compartment, apoptosis was significantly increased in MDS FAB subtypes RA, RARS and RAEB (<10% blasts) (median 56.5% (range 15.1–86.5%)) (n = 31) compared to normal controls (median 18.5% (range 3.4–33.4%), z = 4.84, P < 0.0001) (n = 15), patients with advanced MDS (median 16% (range 2.1–91.6%), z = 4.25, P < 0.0001) (n = 17) and de novo AML (median 15% (3.8–41.3%)) (n = 16). The highest levels of CD34+ cell apoptosis (91.6%) were recorded in a patient with RAEB (17% blasts) associated with poor-risk cytogenetics (45,XX,−7). 20/31 (64.5%) patients with early MDS had >50% CD34+ cells undergoing PCD. There was no correlation between patient age, full blood count or cytogenetics and the degree of apoptosis in either the early or late stages of the disease (data not shown).
Bcl-2-related protein expression
As shown in Fig 2, CD34+ cells of patients with advanced MDS expressed a significantly lower Bax & Bad v Bcl-2 & Bcl-X ratio (median ratio 1.14 (range 0.06–3.32), n = 16), than in patients with early disease (median ratio 2.47 (range 1.19–9.42), n = 16) (z = 3.8, P = 0.0001) and normal controls (median ratio 1.89 (range 0.65–4.12), n = 10) (z = 2.18, P = 0.029). In contrast, a higher ratio was observed in early MDS compared to normal controls although this did not reach statistical significance (z = 1.66, P = 0.097). 13/16 (81.25%) patients with early MDS had a ratio >2 compared to only 3/16 (19.75%) patients with advanced disease. The lowest ratios were observed in leukaemic blasts from patients with de novo AML (median 0.51 (0.03–1.29)) (n = 16), which were significantly lower than both normal controls (z = 4.04, P = 0.0001) and MDS RAEB-t/MDS-AML (z = 3.8, P = 0.0001). Low ratios in de novo AML appeared to be due to down-regulation of Bad and up-regulation of Bcl-2 in the majority of cases. In the different FAB subtypes of MDS, however, no consistent pattern of Bcl-2-related protein expression emerged. Using repeated measures of variants (ANOVA), nevertheless, we found a significant interaction between disease subtype and protein expression (f = 5.05, df = 9,161; P < 0.0005), indicating that variations between individual Bcl-2-related proteins differed according to disease subtype (Fig 3). There was, however, no correlation between Bax and Bad v Bcl-2 and Bcl-X ratio and the degree of apoptosis, full blood count or karyotype in any of the diagnostic categories.
This is the first study, to our knowledge, to address whether deregulated apoptosis in MDS can be attributed to altered expression of the Bcl-2-related proteins. It also utilizes new methods of apoptosis detection to clarify previous conflicting data as to whether excessive apoptosis occurs primarily in early or late disease ( Greenberg et al, 1994 ; Raza et al, 1995 ; Rajapaksa et al, 1996 ; Bogdanovic et al, 1997 ; Hellstrom-Lindberg et al, 1997 ). Indeed, our studies using flow cytometric detection of FITC-conjugated annexin V on the surface of apoptotic cells indicated that programmed cell death was significantly increased in bone marrow CD34+ cells of patients with early MDS FAB subtypes, RA, RARS and RAEB (<10% blasts) compared to normal controls and advanced disease and probably contributes to the ineffective haemopoiesis and peripheral cytopenias that are characteristic of the disease. The annexin V affinity assay used in this study offers advantages over older apoptosis assays in that it is rapid and detects cells in the early stages of programmed cell death ( Koopman et al, 1994 ; Martin et al, 1995 ; Vermes et al, 1995 ). Indeed, direct comparison with the TdT-mediated dUTP-nick end labelling (TUNEL) technique, which identifies DNA strand breaks generated during apoptosis, indicated that PS externalization could be measured prior to in vitro detection of DNA damage ( O'Brien et al, 1997 ). In this study, as has been previously demonstrated ( Rajapaksa et al, 1996 ; Hellstrom-Lindberg et al, 1997 ), levels of apoptosis in early MDS were highly variable (15.1–86.2%) and could not be explained by differences in patient age, peripheral blood counts or cytogenetics. In contrast, Bogdanovic et al (1997 ), using morphological techniques for apoptosis quantification, found that levels of PCD were inversely correlated with peripheral white cell count.
Increased apoptosis in early MDS may represent a physiological mechanism whereby the haemopoietic system is able to eliminate potentially harmful clones. To this end, PCD might be triggered by the bone marrow microenvironment or arise from intrinsic defects within the stem cell itself such as abnormalities in cell cycling or DNA repair. Alternatively, an early ‘hit’ in the multistep pathogenesis of MDS may result in a higher proliferative rate of the neoplastic clone, creating an expanded population of CD34+ cells. Increased apoptosis may therefore merely represent a homeostatic process to control cell numbers. This hypothesis would explain why, despite >50% CD34+ cells underging PCD, stem cell exhaustion does not seem to occur, at least in the early stages of the disease. Recent studies using cell tracking of proliferating normal human CD34+ cells in short-term cell cultures have demonstrated that with increasing numbers of in vitro cellular divisions there is both a decline in the ability of primitive stem cells to sustain in vitro haemopoiesis and a concomitant increase in CD34+ cell apoptosis ( Traycoff et al, 1998 ). This work therefore provides a model to explain both the ineffective haemopoiesis and excessive apoptosis of haemopoietic progenitors witnessed in MDS.
We also sought to determine whether deregulated CD34+ cell apoptosis in ‘early’ and/or ‘late’ MDS could be attributed to altered expression of two pro- and two anti-apoptotic Bcl-2-related proteins Bax and Bad v Bcl-2 and Bcl-X respectively using two-colour flow cytometry. Certainly, Rajapaksa et al (1996 ) found low levels of Bcl-2 protein expression in early MDS compared to advanced disease. Moreover, in de novo AML, Bcl-2 overexpression is linked with CD34 positivity, resistance to chemotherapy and reduced survival ( Delia et al, 1992 ; Campos et al, 1993 ; Stoetzer et al, 1996 ; Bradbury et al, 1997 ), features commonly attributed to secondary disease ( De-Witte et al, 1989 ; Fenaux et al, 1991 ; Lima et al, 1995 ). In our group of patients we did find a higher although non-significant (P = 0.09) Bax and Bad v Bcl-2 and Bcl-X ratio in early MDS compared to normal controls. Furthermore, in advanced disease, ratios were significantly lower than both normal controls and early MDS. There was, however, no correlation between these ratios and the degree of apoptosis as determined by annexin V positivity. Nor did the Bcl-2-related protein ratio correlate with peripheral blood counts or karyotype. Studies measuring Bcl-2 expression in de novo AML have reached similar conclusions ( Banker et al, 1997 ) and suggest that mechanisms additional to Bcl-2-related protein expression are involved in the regulation of apoptosis in myeloid malignancies. Furthermore, although statistical analysis using repeated measures of variants (ANOVA) showed a significant difference in patterns of protein expression between the different disease categories, the failure of a consistent pattern of Bcl-2-related protein expression to emerge indicates that altered levels of these regulatory proteins are an effect and not a cause of deregulated apoptosis observed in MDS.
It is hoped that a clearer understanding of the mechanisms underlying programmed cell death and its role in the pathogenesis of MDS will lead to the design of effective therapeutic strategies in a disease in which no existing treatment, other than bone marrow transplantation, has convincingly prolonged survival.
We are indebted to Dr F. Al Refaie, Princess Alexandra Hospital, Harlow, Dr R. Carr, St Thomas' Hospital, London, Dr J. Duncan, Royal Sussex County Hospital, Brighton, Dr D. S. Gillett, Pembury Hospital, Dr J. P. L. A. Hayes, All Saint's Hospital, Chatham, Dr R. M. Ireland and Dr P. Black, Greenwich Hospital, Dr R. Jan-Mohamed, Dr R. Kaczmarski, The Hillingdon Hospital, Dr Mir, Lewisham Hospital, Dr S. M. B. Rassam, Queen Mary's Hospital, Sidcup, Dr V. Ratnayake, The William Harvey Hospital, Ashford, Dr I. R. Samaratunga, Farnborough Hospital, Dr B.Wells, The James Paget Hospital, Great Yarmouth, and Dr Y. F. Williams, The Kent and Canterbury Hospital, for donation of patient samples and to Dr R. Hooper for his help in statistical analysis.
This project was funded by a grant from the Leukaemia Research Fund, U.K.