- Email: [email protected]
Correlation of bone marrow colony growth in the myelodysplastic syndromes with the FAB classification and the bournemouth score
Leukemia Research Vol. 13, No. 9, pp. 833--839, 1989. Printed in Great Britain.
CORRELATION
0145-2126/89 $3.00 + .00 Pergamon Press plc
OF BONE
MARROW
MYELODYSPLASTIC CLASSIFICATION
AND
COLONY
SYNDROMES
GROWTH
IN THE
WITH THE FAB
THE BOURNEMOUTH
SCORE*
D. G. OSCIER, A. WORSLEY,S. DARLOW,A. FIGES, J. D. WILLIAMSt and T. J. HAMBLIN Department of Haematology, Royal Victoria Hospital, Bournemouth and tDepartment of Medical Statistics and Computing, University of Southampton, U.K.
(Received 14 February 1989. Accepted 28 May 1989) Abstract--Erythroid and myeloid colonies were grown from the bone marrow of 81 patients with myelodysplasia and the median number of colonies correlated with the FAB classification and Bournemouth score. CFU-GM were increased in CMML compared to RAEB and RAEBt. BFU-E were higher in RA than in the other FAB subgroups. Patients with a high Bournemouth score had poorer CFU-GM and BFU-E growth than those with a low score.
Key words: Bournemouth score, FAB classification, myelodysplasia.
INTRODUCTION
Attempts to correlate in vitro growth patterns with the FAB classification have met with mixed results. Rutuu et al. [4] found normal numbers of granulocyte-macrophage colonies (CFU-GM) in refractory anaemia with ring sideroblasts (RAS) and in refractory anaemia (RA) associated with partial deletion of the long arm of chromosome 5 ( 5 q syndrome), but reduced numbers of CFU-GM with other types of MDS, whereas May et al. [5] found no correlation between FAB subtypes and CFU-GM numbers. Erythroid colony formation has been much less studied than CFU-GM, but most workers have shown defective BFU-E colonies in the great majority of patients [4, 6]. No correlation with FAB subtype has been shown. Chronic myelomonocytic ieukaemia (CMML) has clinical features both of MDS and the myeloproliferative syndromes. Bone marrow colonies have been little studied in this syndrome, and none of the series have had sufficient numbers of this subtype of MDS to make a clear statement about its colony growth characteristics. We have studied a large enough number of patients to discriminiate reasonably between the different FAB subtypes of MDS. In this report we have attempted to correlate the Bournemouth score and the FAB subtype with CFU-GM and BFU-E colony formation characteristics. In addition we have included a large number of patients with RA whose features of MDS, though definite, were not gross, in an attempt to evaluate whether reduced BFU-E growth is a consistent feature of RA and therefore a
THE MYELODYSPLASTICsyndromes (MDS) are clonal disorders of the bone marrow in which the capacity of stem cells to differentiate is impaired so that characteristically a hypercellular bone marrow with distinctive morphological abnormalities is accompanied by peripheral cytopenias. The subclassification introduced by the FAB group [1] has prognostic value. Recently we have shown that the Bournemouth score, a simple scoring system based on the haemoglobin, neutrophil and platelet counts, and bone marrow blast percentage is a valuable guide to prognosis [2]. There have been a number of studies of bone marrow colony growth in MDS. Not surprisingly a "leukaemic-type" of growth with micro- or macrocluster formation and low colony formation is associated with a higher incidence of transformation to acute leukaemia and a shorter survival than a "nonleukaemic" growth pattern [3]. * We gratefully acknowledge the support of the Leukaemia Research Fund and the Wessex Regional Research Fund. We thank Mrs K. Avery for typing the manuscript. Abbreviations: MDS, myelodysplastic syndrome; RA, refractory anaemia; RAS, refractory anaemia with ring sideroblasts; RAEB, refractory anaemia with excess of blasts; RAEBt, refractory anaemia with excess of blasts in transformation; CMML, chronic myelomonocytic leukaemia; FAB, French-American-British; BS, Bournemouth score; CFU-GM, colony-forming unit-granulocytemacrophage; BFU-E, burst-forming unit-erythroid. Correspondence to: Dr D. G. Oscier, Haematology Department, Royal Victoria Hospital, Shelley Road, Bournemouth, U.K. 833
834
D . G . OSCIERet al. 500-
400.
300-
o 3
U m
o
x
\
0
200.
0 O0
O L ~J
o
8 o o
o ~ 8
0
O0 O0
o o o C~D
o
0
O0
O0
oo
8
8
o% 0 0
oo
0
o
o8
on
I
Control
RA
RAS
CMML
RAEB
n:B10
n~23
n:13
n=20
n=21
c~ RAEBt n= 4
FAB TYPE
FIG. 1. C F U - G M growth classified according to the FAB type. useful test for d i s t i n g u i s h i n g R A f r o m o t h e r c a u s e s o f d y s e r y t h r o p o i e s i s in w h i c h n o r m a l o r i n c r e a s e d BFU-E growth may be expected.
PATIENTS AND METHODS We have studied 81 patients with MDS either at presentation or at a time when they were not receiving chemotherapy, and 10 normal controls. The patients were classified according to the F A B criteria. There were 22 with R A , 13 with RAS, 19 with CMML, 17 with refractory anaemia with an excess of blasts ( R A E B ) and four with refractory anaemia with excess of blasts in transformation (RAEBt). Patients were also categorized according to the Bournemouth score in which one point is allocated for each of the following: Hb < 10 g/dl, neutrophils <2.5 x 109/1,
TABLE 1.
NUMBER OF PATIENTS GROUPED ACCORDING TO BOURNEMOUTH
Bournemouth score No. of patients
0 12
SCORE
1 37
2 16
3 7
4 9
platelets <100 x 109/1, and bone marrow blasts > 5 % . The distribution of patients according to the Bournemouth score is shown in Table 1. B o n e m a r r o w culture
Bone marrow (2--4 ml) was aspirated and diluted in Hank's balanced salt solution. Mononuclear cells were isolated by Ficoll-Hypaque density separation. The mononuclear cell suspension was washed twice, counted in a Coulter ZB1 counter and the cell viability estimated by Trypan Blue staining.
Bone marrow colonies in MDS
835
TABLE 2. MEAN NUMBER OF COLONIES/2 × 105 CELLS PLATED IN CONTROLS AND PATIENTS CATEGORIZED BY FAB CATEGORY AND BOURNEMOUTH SCORE
CFU-E
BFU-E
Clusters
CFU-GM colonies
No. of patients
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Normal
10
133
66--390
85
41-138
198
126-427
96
45-126
FAB type RA RAS CMML RAEB RAEBt
22 13 19 17 4
65 37 21 22 4
0-94 0-95 0-122 0-34 0-13
50 15 14 8 8
0-139 0-38 0-43 0-82 0-27
158 283 305 122 193
73-796 42-952 68-901 40-215 7-558
79 75 147 51 14
10-136 0-187 0--463 0-282 0-53
Bournemouth score 0 12 1 37 2 16 3 7 4 9
66 34 23 6 2
0-95 1-122 0-78 2-34 0-13
34 30 5 3 1
0-102 0-139 0-17 0-11 0--4
249 261 163 135 86
42-952 68-901 57-370 40-215 7-558
114 107 59 33 13
0-216 0-463 0-282 1-92 0-53
Erythroid progenitors
Erythroid colonies were cultured in 75 patients using a method described by Iscove et al. [7] and modified by Goldman et al. [8]. The culture medium consisted of 0.8% methyl cellulose, 30% fetal calf serum, 10% bovine serum albumin, 10-4M mercaptoethanol and alpha medium. Mononuclear cells were plated at 2 x 105 cells/ml in the presence of 3 units of erythropoietin (Step I. Terry Fox Laboratories). Granulocyte-macrophage progenitors
CFU-GM were assayed in all 81 patients using a method described by Hibbin et al. [9]. Culture conditions were identical to those used for erythroid colony growth except that 10% phytohaemagglutinin-stimulated lymphocyte conditioned medium was included instead of erythropoietin. Cultures were plated in triplicate and incubated for 14 days in a fully humidified incubator at 37°C in an atmosphere of 5% CO2 and air. Plates were scored at 7 days for CFU-E (colony-forming units, erythroid) and clusters (aggregates of 3-40 cells), and at 14 days for BFU-E and CFU-GM. Statistical methods
A Kruskal-Wallis non-parametric analysis of variance was used initially to reveal the existence of an overall significant difference in colony growth between patients of different FAB type and Bournemouth score. MannWhitney tests were then used to identify specific pairs of groups with significantly different median growths. RESULTS The m e a n n u m b e r of B F U - E , C F U - E , C F U - G M and clusters for patients in each F A B group and for each B o u r n e m o u t h score is shown in Table 2. The numbers of C F U - G M and B F U - E grown in each patient classified according to the F A B classification and the B o u r n e m o u t h score is shown in Figs 1-4.
A Kruskal-Wallis analysis revealed significant differences between mean erythroid and myeloid colony growth both for different F A B types and Bournemouth scores (Table 3). The results of applying M a n n - W h i t n e y tests to pairs of F A B types and B o u r n e m o u t h scores for both erythroid and myeloid colony growth are shown below: (1) C F U - G M vs F A B . The median C F U - G M growth of patients with C M M L was significantly higher than for patients with R A E B ( p - - 0 . 0 0 6 ) and R A E B t , but not for patients with R A or RAS. Nine out of 20 patients with C M M L had increased C F U - G M growth c o m p a r e d to that found in the controls. (2) C F U - G M vs B o u r n e m o u t h score. The median colony growth of groups 3 and 4 was significantly lower than that of groups 0 and 1 (control vs group 3, p -- 0.0112). In addition the median colony growth of patients in group 1 was significantly higher than that in group 2 (p = 0.049). (3) B F U - E vs F A B . Median B F U - E growth of patients with R A was significantly lower than that of the control group (p = 0.028) but significantly higher than that of the other FAB groups. (4) B F U - E vs B o u r n e m o u t h score. The median B F U - E growth of the control group was higher than for all the patient groups (control vs group 0, p = 0.01). G r o u p 0 patients had significantly higher growth than patients in group 2 (p = 0.01) and group 4 (p = 0.023) but not group 1 or 3. There was no significant difference in median B F U - E growth between patients in group 1 and those in groups 2, 3 or 4. B F U - E g r o w t h in R A
Patients with R A were divided into two groups
836
D.G. OSCIERet al.
4~-
0 0 0
o
0
m
x O
?
O 0
200-
0
Ilk U
0
0 0 0
o
0
o O O0 0 O0
8
~
o
0 0
Oo c~
0
°
o
0
0
0
0
8
oo
0
Control
0
nmlO
n-ll
1
2
n-38
n,,16
3
4
n-7
nm9
80URNEMOUTH SCORE
FIG. 2. CFU-GM growth classified according to the Bournemouth score. depending on whether their haemoglobin was greater or less than 10 g/dl at the time of colony assay. The results of BFU-E assay in these patients are shown in Table 4. Normal BFU-E numbers were found in nine of the 23 patients studied and there was no significant difference in the pattern of BFU-E growth between the two groups. These results would suggest that BFU-E growth is not a useful diagnostic test for RA. DISCUSSION Many studies have shown that the in v i t r o growth of myeloid [4], erythroid [4, 6], and megakaryocytic [10] precursors is abnormal in MDS. Often there is
no colony growth, but where colonies are formed karyotypic analysis of individual colonies has shown that at least some are derived from the neoplastic clone and do not simply represent residual normal haemopoiesis [11]. The FAB classification of MDS is based on morphological criteria and is of prognostic significance. However, some patients present with features which would allow them to be classified into more than one FAB group, and not infrequently patients evolve from one FAB group into another [12]. Karyotypic abnormalities are common in MDS, but unlike in A M L there is no particular abnormality associated with a single FAB type. For these reasons it is unclear whether the FAB types represent distinct disease
B o n e m a r r o w c o l o n i e s in M D S
837
150-
II
u
m o
2
x 100
o o
ell
\
o
w i
o
L
o
m o o o
50
o o
o
o o
8 8o
8o
o o
o
o o No
Control n,,lO
RA n=22
RAS n:13
CMML n=19
RAEB
RAEBI
n=17
n= 4
FAB TYPE
FIG. 3. BFU-E growth classified according to the FAB type. entities or whether they reflect differing presdiagnosis included MDS. All 21 patients in whom entations of a single disease. It is therefore of interest MDS was confirmed had abnormal erythroid growth, to examine whether particular in vitro colony growth while six of the remaining patients had normal BFUpatterns are related to the FAB type. This is E numbers. They concluded that normal colony especially the case in CMML, which has many of the growth excluded MDS. features of a myeloproliferative disorder. When colony growth was correlated with the Our data show that CFU-GM numbers are sigBournemouth score we found that both erythroid nificantly higher in CMML than in RAEB and and myeloid growth was uniformly impaired in RAEBt, in accord with an earlier study of a smaller patients with Bournemouth score 3 and 4. In addition erythroid growth was reduced or absent in all patients number of patients [13]. Geissler et al. [14] have recently demonstrated increased circulating CFUwith Bournemouth score 2, and in many patients with GM in two patients with CMML, and found that a score of 1 or 0. The majority of patients with colony growth occurred even in the absence of Bournemouth score 3 and 4 have RAEB or RAEBt, and there is therefore a clear association between exogenous colony-stimulating factors. Both propimpaired or absent colony formation and an erties are features of colony growth in myeloproliferative states. BFU-E growth was low or absent increased percentage of blasts in the bone marrow. in the majority of patients with RAS, CMML, RAEB The reason for absent colony growth and reduced and RAEBt, but it was normal in 10 of the 22 patients growth in long-term cultures in patients who have with RA. The diagnosis of RA was based on morhypercellular marrows is still unclear. A recent study phological criteria, and care was taken to exclude of colony growth in four patients with RAEB, using other disorders associated with dysplastic morrecombinant growth factors rather than conditioned phological features. There was no significant cormedia to stimulate colony growth, found that granurelation between erythroid colony growth and the locytic colonies were formed in all cases, but one severity of anaemia in this group of patients. At varipatient [16] was unable to produce macrophage colonies and in two there was no eosinophil or erythroid ance with these findings is a recent report from Shihab-EI-Deen et al. [15], who studied in vitro colcolony formation. The authors concluded that there were intrinsic abnormalities of colony-forming units ony growth in 34 patients whose initial differential
838
D.G. OSCIERet al. 150-
0 0
0
io
0 0
U m 0
0 0
× 100o
o
O
I,I.
O
m
O O O o o
50'
o
o
0
8
0
8
0
0
(30 0
0
g~ Control
0
n : 10
n = 12
1
2
n : 37
n:12
O
3 n=5
4 n:9
BOURNEMOUTH SCORE
FIG 4. BFU-E growth classified according to the Bournemouth score. TABLE 3. KRUSKAL WALLIS ANALYSIS OF VARIANCE FOR
CFU-GM ANDBFU-E GROWTHACCORDINGTO FAB TYPE AND BOURNEMOUTH SCORE ( l S )
3. CFU-GM CFU-GM BFU-E BFU-E
vs FAB vs BS vs FAB vs BS
Chi 2
p
12.46 22.20 34.53 28.23
0.02 <0.001 <0.001 <0.001
TABLE 4. BFU-E GROWTHIN PATIENTSWITHRA
4.
5.
BFU-E Hb
No. of patients
Normal
Decreased
>10 g <10g
14 9
7 2
7 7
6.
7. in MDS, and our data would suggest that these become more p r o n o u n c e d with advanced disease. REFERENCES 1. Bennett J. M., Catovsky D., Daniel M. T., Flandrin G., Galton D. A. G., Gralnick H. R. & Sultan C. (1982) Proposals for the classification of the myelodysplastic syndromes. Br. J. Haemat. 51, 189. 2. Mufti G. J., Stevens J. R., Oscier D. G., Hamblin T.
8.
9.
J. & Machin D. (1985) Myelodysplastic syndromes: a scoring system with prognostic significance. Br. J. Haemat. 59, 425. Greenberg P. L. (1986) In vitro culture techniques defining biological abnormalities in the myelodysplastic syndromes and myeloproliferative disorders. Clin. Haemat. 15, 973. Ruutu T., Partanen S., Lintula R., Teerenhovi L. & Knuutila S. (1984) Erythroid and granulocyte-macrophage colony formation in myelodysplastic syndromes. Scand. J. Haemat. 32, 395. May S. J., Smith S. A., Jacobs A., Williams A. & Bailey-Wood R. (1985) The myelodysplastic syndrome: analysis of laboratory characteristics in relation to the FAB classification. Br. J. Haemat. 59, 311. Amato D. & Khan N. R. (1983) Erythroid burst formation in cultures of bone marrow and peripheral blood from patients with refractory anemia. Acta Haemat. 70, 1. Iscove N. N., Sieber F. & Winterhalter K. H. (1974) Erythroid colony formation in cultures of mouse and human bone marrow: analysis of the requirements of erythropoietin by gel filtration and affinity chromatography on aga-rose-concanavalin A . J. Cell Phys. 83, 309. Goldman J. M., Hihbin J., Kearney L., Orchard K., & Th'ng K. H. (1982) HLA-DR monoclonal antibodies inhibit the proliferation of normal and chronic granulocytic leukaemia myeloid progenitor cells. Br. J. Haemat. 52, 411. Hibbin J. A., Njoku O. S., Matutes E., Lewis S. M.
Bone marrow colonies in MDS & Goldman J. M. (1984) Myeloid progenitor cells in the circulation of patients with myelofibrosis and other myeloproliferative disorders. Br. J. Haemat. 57, 495. 10. Juvonen E., Partanen S., Knuutila S. & Ruutu T. (1986) Megakaryocyte colony formation by bone marrow progenitors in myelodysplastic syndromes. Br. J. Haemat. 63, 331. 11. Amenomori T., Tomonaga M., Yoshida Y., Kuriyama K., Matsuo T., Jinnai I., Ichimaru M., Omiya A. & Tsuji T. (1986) Cytogenetic evidence for partially committed myeloid progenitor cell origin of chronic myelomonocytic leukaemia and juvenile chronic myeloid leukaemia: both granuiocyte-macrophage precursors and erythroid precursors carry identical marker chromosomes. Br. J. Haemat. 64, 539. 12. Vallespi T., Torrabadella M., Julia A., Irriguible D., Jaen A., Acebo G. & Triginer J. (1985) Myelodysplastic syndromes: a study of 101 cases according to the FAB classification. Br. J. Haemat. 61, 83.
839
13. Milner G. R., Testa N. G., Geary C. G., Dexter T. M., Muldai S., Maclver J. E. & Laitha L. G. (1977) Bone marrow culture studies in refractory cytopenia and 'smouldering leukaemia'. Br. J. Haemat. 35, 251. 14. Geissler K., Hinterberger W., Bettelheim P., Haas O. & Lechner K. (1988) Colony growth characteristics in chronic myelomonocytic leukemia. Leuk. Res. 12, 373. 15. Shihab-E1-Deen A., Guevara C. & Prchal J. F. (1987) Bone marrow cultures in dysmyelopoietic syndrome: diagnostic and prognostic evaluation. Acta Haemat. 78, 17. 16. Schouten H. C., Delwel R., Bot F. J., Haegmeijer A., Touw I. P. & Lowenberg B. (1989) Characterization of clonogenic ceils in refractory anemia with excess of blasts (RAEB-CFU): response to recombinant hematopoietic growth factors and maturation phenotypes. Leukemia Res. 13, 245.
CORRELATION
0145-2126/89 $3.00 + .00 Pergamon Press plc
OF BONE
MARROW
MYELODYSPLASTIC CLASSIFICATION
AND
COLONY
SYNDROMES
GROWTH
IN THE
WITH THE FAB
THE BOURNEMOUTH
SCORE*
D. G. OSCIER, A. WORSLEY,S. DARLOW,A. FIGES, J. D. WILLIAMSt and T. J. HAMBLIN Department of Haematology, Royal Victoria Hospital, Bournemouth and tDepartment of Medical Statistics and Computing, University of Southampton, U.K.
(Received 14 February 1989. Accepted 28 May 1989) Abstract--Erythroid and myeloid colonies were grown from the bone marrow of 81 patients with myelodysplasia and the median number of colonies correlated with the FAB classification and Bournemouth score. CFU-GM were increased in CMML compared to RAEB and RAEBt. BFU-E were higher in RA than in the other FAB subgroups. Patients with a high Bournemouth score had poorer CFU-GM and BFU-E growth than those with a low score.
Key words: Bournemouth score, FAB classification, myelodysplasia.
INTRODUCTION
Attempts to correlate in vitro growth patterns with the FAB classification have met with mixed results. Rutuu et al. [4] found normal numbers of granulocyte-macrophage colonies (CFU-GM) in refractory anaemia with ring sideroblasts (RAS) and in refractory anaemia (RA) associated with partial deletion of the long arm of chromosome 5 ( 5 q syndrome), but reduced numbers of CFU-GM with other types of MDS, whereas May et al. [5] found no correlation between FAB subtypes and CFU-GM numbers. Erythroid colony formation has been much less studied than CFU-GM, but most workers have shown defective BFU-E colonies in the great majority of patients [4, 6]. No correlation with FAB subtype has been shown. Chronic myelomonocytic ieukaemia (CMML) has clinical features both of MDS and the myeloproliferative syndromes. Bone marrow colonies have been little studied in this syndrome, and none of the series have had sufficient numbers of this subtype of MDS to make a clear statement about its colony growth characteristics. We have studied a large enough number of patients to discriminiate reasonably between the different FAB subtypes of MDS. In this report we have attempted to correlate the Bournemouth score and the FAB subtype with CFU-GM and BFU-E colony formation characteristics. In addition we have included a large number of patients with RA whose features of MDS, though definite, were not gross, in an attempt to evaluate whether reduced BFU-E growth is a consistent feature of RA and therefore a
THE MYELODYSPLASTICsyndromes (MDS) are clonal disorders of the bone marrow in which the capacity of stem cells to differentiate is impaired so that characteristically a hypercellular bone marrow with distinctive morphological abnormalities is accompanied by peripheral cytopenias. The subclassification introduced by the FAB group [1] has prognostic value. Recently we have shown that the Bournemouth score, a simple scoring system based on the haemoglobin, neutrophil and platelet counts, and bone marrow blast percentage is a valuable guide to prognosis [2]. There have been a number of studies of bone marrow colony growth in MDS. Not surprisingly a "leukaemic-type" of growth with micro- or macrocluster formation and low colony formation is associated with a higher incidence of transformation to acute leukaemia and a shorter survival than a "nonleukaemic" growth pattern [3]. * We gratefully acknowledge the support of the Leukaemia Research Fund and the Wessex Regional Research Fund. We thank Mrs K. Avery for typing the manuscript. Abbreviations: MDS, myelodysplastic syndrome; RA, refractory anaemia; RAS, refractory anaemia with ring sideroblasts; RAEB, refractory anaemia with excess of blasts; RAEBt, refractory anaemia with excess of blasts in transformation; CMML, chronic myelomonocytic leukaemia; FAB, French-American-British; BS, Bournemouth score; CFU-GM, colony-forming unit-granulocytemacrophage; BFU-E, burst-forming unit-erythroid. Correspondence to: Dr D. G. Oscier, Haematology Department, Royal Victoria Hospital, Shelley Road, Bournemouth, U.K. 833
834
D . G . OSCIERet al. 500-
400.
300-
o 3
U m
o
x
\
0
200.
0 O0
O L ~J
o
8 o o
o ~ 8
0
O0 O0
o o o C~D
o
0
O0
O0
oo
8
8
o% 0 0
oo
0
o
o8
on
I
Control
RA
RAS
CMML
RAEB
n:B10
n~23
n:13
n=20
n=21
c~ RAEBt n= 4
FAB TYPE
FIG. 1. C F U - G M growth classified according to the FAB type. useful test for d i s t i n g u i s h i n g R A f r o m o t h e r c a u s e s o f d y s e r y t h r o p o i e s i s in w h i c h n o r m a l o r i n c r e a s e d BFU-E growth may be expected.
PATIENTS AND METHODS We have studied 81 patients with MDS either at presentation or at a time when they were not receiving chemotherapy, and 10 normal controls. The patients were classified according to the F A B criteria. There were 22 with R A , 13 with RAS, 19 with CMML, 17 with refractory anaemia with an excess of blasts ( R A E B ) and four with refractory anaemia with excess of blasts in transformation (RAEBt). Patients were also categorized according to the Bournemouth score in which one point is allocated for each of the following: Hb < 10 g/dl, neutrophils <2.5 x 109/1,
TABLE 1.
NUMBER OF PATIENTS GROUPED ACCORDING TO BOURNEMOUTH
Bournemouth score No. of patients
0 12
SCORE
1 37
2 16
3 7
4 9
platelets <100 x 109/1, and bone marrow blasts > 5 % . The distribution of patients according to the Bournemouth score is shown in Table 1. B o n e m a r r o w culture
Bone marrow (2--4 ml) was aspirated and diluted in Hank's balanced salt solution. Mononuclear cells were isolated by Ficoll-Hypaque density separation. The mononuclear cell suspension was washed twice, counted in a Coulter ZB1 counter and the cell viability estimated by Trypan Blue staining.
Bone marrow colonies in MDS
835
TABLE 2. MEAN NUMBER OF COLONIES/2 × 105 CELLS PLATED IN CONTROLS AND PATIENTS CATEGORIZED BY FAB CATEGORY AND BOURNEMOUTH SCORE
CFU-E
BFU-E
Clusters
CFU-GM colonies
No. of patients
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Normal
10
133
66--390
85
41-138
198
126-427
96
45-126
FAB type RA RAS CMML RAEB RAEBt
22 13 19 17 4
65 37 21 22 4
0-94 0-95 0-122 0-34 0-13
50 15 14 8 8
0-139 0-38 0-43 0-82 0-27
158 283 305 122 193
73-796 42-952 68-901 40-215 7-558
79 75 147 51 14
10-136 0-187 0--463 0-282 0-53
Bournemouth score 0 12 1 37 2 16 3 7 4 9
66 34 23 6 2
0-95 1-122 0-78 2-34 0-13
34 30 5 3 1
0-102 0-139 0-17 0-11 0--4
249 261 163 135 86
42-952 68-901 57-370 40-215 7-558
114 107 59 33 13
0-216 0-463 0-282 1-92 0-53
Erythroid progenitors
Erythroid colonies were cultured in 75 patients using a method described by Iscove et al. [7] and modified by Goldman et al. [8]. The culture medium consisted of 0.8% methyl cellulose, 30% fetal calf serum, 10% bovine serum albumin, 10-4M mercaptoethanol and alpha medium. Mononuclear cells were plated at 2 x 105 cells/ml in the presence of 3 units of erythropoietin (Step I. Terry Fox Laboratories). Granulocyte-macrophage progenitors
CFU-GM were assayed in all 81 patients using a method described by Hibbin et al. [9]. Culture conditions were identical to those used for erythroid colony growth except that 10% phytohaemagglutinin-stimulated lymphocyte conditioned medium was included instead of erythropoietin. Cultures were plated in triplicate and incubated for 14 days in a fully humidified incubator at 37°C in an atmosphere of 5% CO2 and air. Plates were scored at 7 days for CFU-E (colony-forming units, erythroid) and clusters (aggregates of 3-40 cells), and at 14 days for BFU-E and CFU-GM. Statistical methods
A Kruskal-Wallis non-parametric analysis of variance was used initially to reveal the existence of an overall significant difference in colony growth between patients of different FAB type and Bournemouth score. MannWhitney tests were then used to identify specific pairs of groups with significantly different median growths. RESULTS The m e a n n u m b e r of B F U - E , C F U - E , C F U - G M and clusters for patients in each F A B group and for each B o u r n e m o u t h score is shown in Table 2. The numbers of C F U - G M and B F U - E grown in each patient classified according to the F A B classification and the B o u r n e m o u t h score is shown in Figs 1-4.
A Kruskal-Wallis analysis revealed significant differences between mean erythroid and myeloid colony growth both for different F A B types and Bournemouth scores (Table 3). The results of applying M a n n - W h i t n e y tests to pairs of F A B types and B o u r n e m o u t h scores for both erythroid and myeloid colony growth are shown below: (1) C F U - G M vs F A B . The median C F U - G M growth of patients with C M M L was significantly higher than for patients with R A E B ( p - - 0 . 0 0 6 ) and R A E B t , but not for patients with R A or RAS. Nine out of 20 patients with C M M L had increased C F U - G M growth c o m p a r e d to that found in the controls. (2) C F U - G M vs B o u r n e m o u t h score. The median colony growth of groups 3 and 4 was significantly lower than that of groups 0 and 1 (control vs group 3, p -- 0.0112). In addition the median colony growth of patients in group 1 was significantly higher than that in group 2 (p = 0.049). (3) B F U - E vs F A B . Median B F U - E growth of patients with R A was significantly lower than that of the control group (p = 0.028) but significantly higher than that of the other FAB groups. (4) B F U - E vs B o u r n e m o u t h score. The median B F U - E growth of the control group was higher than for all the patient groups (control vs group 0, p = 0.01). G r o u p 0 patients had significantly higher growth than patients in group 2 (p = 0.01) and group 4 (p = 0.023) but not group 1 or 3. There was no significant difference in median B F U - E growth between patients in group 1 and those in groups 2, 3 or 4. B F U - E g r o w t h in R A
Patients with R A were divided into two groups
836
D.G. OSCIERet al.
4~-
0 0 0
o
0
m
x O
?
O 0
200-
0
Ilk U
0
0 0 0
o
0
o O O0 0 O0
8
~
o
0 0
Oo c~
0
°
o
0
0
0
0
8
oo
0
Control
0
nmlO
n-ll
1
2
n-38
n,,16
3
4
n-7
nm9
80URNEMOUTH SCORE
FIG. 2. CFU-GM growth classified according to the Bournemouth score. depending on whether their haemoglobin was greater or less than 10 g/dl at the time of colony assay. The results of BFU-E assay in these patients are shown in Table 4. Normal BFU-E numbers were found in nine of the 23 patients studied and there was no significant difference in the pattern of BFU-E growth between the two groups. These results would suggest that BFU-E growth is not a useful diagnostic test for RA. DISCUSSION Many studies have shown that the in v i t r o growth of myeloid [4], erythroid [4, 6], and megakaryocytic [10] precursors is abnormal in MDS. Often there is
no colony growth, but where colonies are formed karyotypic analysis of individual colonies has shown that at least some are derived from the neoplastic clone and do not simply represent residual normal haemopoiesis [11]. The FAB classification of MDS is based on morphological criteria and is of prognostic significance. However, some patients present with features which would allow them to be classified into more than one FAB group, and not infrequently patients evolve from one FAB group into another [12]. Karyotypic abnormalities are common in MDS, but unlike in A M L there is no particular abnormality associated with a single FAB type. For these reasons it is unclear whether the FAB types represent distinct disease
B o n e m a r r o w c o l o n i e s in M D S
837
150-
II
u
m o
2
x 100
o o
ell
\
o
w i
o
L
o
m o o o
50
o o
o
o o
8 8o
8o
o o
o
o o No
Control n,,lO
RA n=22
RAS n:13
CMML n=19
RAEB
RAEBI
n=17
n= 4
FAB TYPE
FIG. 3. BFU-E growth classified according to the FAB type. entities or whether they reflect differing presdiagnosis included MDS. All 21 patients in whom entations of a single disease. It is therefore of interest MDS was confirmed had abnormal erythroid growth, to examine whether particular in vitro colony growth while six of the remaining patients had normal BFUpatterns are related to the FAB type. This is E numbers. They concluded that normal colony especially the case in CMML, which has many of the growth excluded MDS. features of a myeloproliferative disorder. When colony growth was correlated with the Our data show that CFU-GM numbers are sigBournemouth score we found that both erythroid nificantly higher in CMML than in RAEB and and myeloid growth was uniformly impaired in RAEBt, in accord with an earlier study of a smaller patients with Bournemouth score 3 and 4. In addition erythroid growth was reduced or absent in all patients number of patients [13]. Geissler et al. [14] have recently demonstrated increased circulating CFUwith Bournemouth score 2, and in many patients with GM in two patients with CMML, and found that a score of 1 or 0. The majority of patients with colony growth occurred even in the absence of Bournemouth score 3 and 4 have RAEB or RAEBt, and there is therefore a clear association between exogenous colony-stimulating factors. Both propimpaired or absent colony formation and an erties are features of colony growth in myeloproliferative states. BFU-E growth was low or absent increased percentage of blasts in the bone marrow. in the majority of patients with RAS, CMML, RAEB The reason for absent colony growth and reduced and RAEBt, but it was normal in 10 of the 22 patients growth in long-term cultures in patients who have with RA. The diagnosis of RA was based on morhypercellular marrows is still unclear. A recent study phological criteria, and care was taken to exclude of colony growth in four patients with RAEB, using other disorders associated with dysplastic morrecombinant growth factors rather than conditioned phological features. There was no significant cormedia to stimulate colony growth, found that granurelation between erythroid colony growth and the locytic colonies were formed in all cases, but one severity of anaemia in this group of patients. At varipatient [16] was unable to produce macrophage colonies and in two there was no eosinophil or erythroid ance with these findings is a recent report from Shihab-EI-Deen et al. [15], who studied in vitro colcolony formation. The authors concluded that there were intrinsic abnormalities of colony-forming units ony growth in 34 patients whose initial differential
838
D.G. OSCIERet al. 150-
0 0
0
io
0 0
U m 0
0 0
× 100o
o
O
I,I.
O
m
O O O o o
50'
o
o
0
8
0
8
0
0
(30 0
0
g~ Control
0
n : 10
n = 12
1
2
n : 37
n:12
O
3 n=5
4 n:9
BOURNEMOUTH SCORE
FIG 4. BFU-E growth classified according to the Bournemouth score. TABLE 3. KRUSKAL WALLIS ANALYSIS OF VARIANCE FOR
CFU-GM ANDBFU-E GROWTHACCORDINGTO FAB TYPE AND BOURNEMOUTH SCORE ( l S )
3. CFU-GM CFU-GM BFU-E BFU-E
vs FAB vs BS vs FAB vs BS
Chi 2
p
12.46 22.20 34.53 28.23
0.02 <0.001 <0.001 <0.001
TABLE 4. BFU-E GROWTHIN PATIENTSWITHRA
4.
5.
BFU-E Hb
No. of patients
Normal
Decreased
>10 g <10g
14 9
7 2
7 7
6.
7. in MDS, and our data would suggest that these become more p r o n o u n c e d with advanced disease. REFERENCES 1. Bennett J. M., Catovsky D., Daniel M. T., Flandrin G., Galton D. A. G., Gralnick H. R. & Sultan C. (1982) Proposals for the classification of the myelodysplastic syndromes. Br. J. Haemat. 51, 189. 2. Mufti G. J., Stevens J. R., Oscier D. G., Hamblin T.
8.
9.
J. & Machin D. (1985) Myelodysplastic syndromes: a scoring system with prognostic significance. Br. J. Haemat. 59, 425. Greenberg P. L. (1986) In vitro culture techniques defining biological abnormalities in the myelodysplastic syndromes and myeloproliferative disorders. Clin. Haemat. 15, 973. Ruutu T., Partanen S., Lintula R., Teerenhovi L. & Knuutila S. (1984) Erythroid and granulocyte-macrophage colony formation in myelodysplastic syndromes. Scand. J. Haemat. 32, 395. May S. J., Smith S. A., Jacobs A., Williams A. & Bailey-Wood R. (1985) The myelodysplastic syndrome: analysis of laboratory characteristics in relation to the FAB classification. Br. J. Haemat. 59, 311. Amato D. & Khan N. R. (1983) Erythroid burst formation in cultures of bone marrow and peripheral blood from patients with refractory anemia. Acta Haemat. 70, 1. Iscove N. N., Sieber F. & Winterhalter K. H. (1974) Erythroid colony formation in cultures of mouse and human bone marrow: analysis of the requirements of erythropoietin by gel filtration and affinity chromatography on aga-rose-concanavalin A . J. Cell Phys. 83, 309. Goldman J. M., Hihbin J., Kearney L., Orchard K., & Th'ng K. H. (1982) HLA-DR monoclonal antibodies inhibit the proliferation of normal and chronic granulocytic leukaemia myeloid progenitor cells. Br. J. Haemat. 52, 411. Hibbin J. A., Njoku O. S., Matutes E., Lewis S. M.
Bone marrow colonies in MDS & Goldman J. M. (1984) Myeloid progenitor cells in the circulation of patients with myelofibrosis and other myeloproliferative disorders. Br. J. Haemat. 57, 495. 10. Juvonen E., Partanen S., Knuutila S. & Ruutu T. (1986) Megakaryocyte colony formation by bone marrow progenitors in myelodysplastic syndromes. Br. J. Haemat. 63, 331. 11. Amenomori T., Tomonaga M., Yoshida Y., Kuriyama K., Matsuo T., Jinnai I., Ichimaru M., Omiya A. & Tsuji T. (1986) Cytogenetic evidence for partially committed myeloid progenitor cell origin of chronic myelomonocytic leukaemia and juvenile chronic myeloid leukaemia: both granuiocyte-macrophage precursors and erythroid precursors carry identical marker chromosomes. Br. J. Haemat. 64, 539. 12. Vallespi T., Torrabadella M., Julia A., Irriguible D., Jaen A., Acebo G. & Triginer J. (1985) Myelodysplastic syndromes: a study of 101 cases according to the FAB classification. Br. J. Haemat. 61, 83.
839
13. Milner G. R., Testa N. G., Geary C. G., Dexter T. M., Muldai S., Maclver J. E. & Laitha L. G. (1977) Bone marrow culture studies in refractory cytopenia and 'smouldering leukaemia'. Br. J. Haemat. 35, 251. 14. Geissler K., Hinterberger W., Bettelheim P., Haas O. & Lechner K. (1988) Colony growth characteristics in chronic myelomonocytic leukemia. Leuk. Res. 12, 373. 15. Shihab-E1-Deen A., Guevara C. & Prchal J. F. (1987) Bone marrow cultures in dysmyelopoietic syndrome: diagnostic and prognostic evaluation. Acta Haemat. 78, 17. 16. Schouten H. C., Delwel R., Bot F. J., Haegmeijer A., Touw I. P. & Lowenberg B. (1989) Characterization of clonogenic ceils in refractory anemia with excess of blasts (RAEB-CFU): response to recombinant hematopoietic growth factors and maturation phenotypes. Leukemia Res. 13, 245.