British joirrnul of Huemutologg. 1993, 85, 63-66

Serum stem cell factor concentration in patients with myelodysplastic syndromes

DAVID BOWEN, SUE YANCIK,* LARRY BENNETT,* DOMINIC CULLIGAN AND KENNETH R E S S E R ~ Department of Haematology, University of Wales College of Medicine, Cardiff, U.K., *AMGEN Boulder Inc., Boulder, Colorado, U.S.A., and TAMGEN Center, Thousand Oaks, Calijornia, U.S.A.

Received 22 February 1993; accepted for publication 23 March 1993

Summary. Stem cell factor (SCF) is characterized by its capacity to synergize dramatically with other haemopoietic growth factors in in vitro erythroid, myeloid, and lymphoid progenitor culture systems. We have measured serum SCF concentrations by enzyme immunoassay in 8 5 patients with myelodysplasia (MDS). Serum samples were taken in 1988- 89 and in 1991-92 and stored at - 20°C. Mean serum SCF concentration in the MDS patients was 2.81 ng/ml (range 0.6-8.0). This was significantly lower (P=O.OOOl) than the values for 234 normal subjects: mean 3 . 3 0 ng/ml (range

1.3-8.0). No significant relationship between SCF concen- tration and peripheral blood counts, bone marrow para- meters, red cell transfusion status, survival or FAB subtype was found, although a trend of decreasing SCF concentration from refractory anaemia through sideroblastic anaemia and chronic myelomonocytic leukaemia to refractory anaemia with excess blasts was seen. The reduced SCF serum concen- tration in some patients with myelodysplasia suggests a rationale for therapy with recombinant SCF in these patients.

The recently identified haemopoietic growth factor (HGF), stem cell factor (SCF), is characterized by its capacity to synergize with other HGFs in in vitro erythroid, myeloid and lymphoid progenitor culture systems (McNiece et al, 1991). The receptor for SCF is the product of the KIT proto-oncogene, the cellular homologue of the oncogene v-kit. KIT encodes a tyrosine kinase receptor protein with sequence homology to the CSF-1 receptor (CSFR, FMS) and the platelet-derived growth factor receptors (PDGFRA and B). In the mouse, mutations or deletions at the S1 (Steel) and W (Dominant white spotting) loci produce a macrocytic anaemia. SI mutants result in a defect affecting the bone marrow microenvironment, while W mutants produce an abnor- mality intrinsic to the stem cell. A number of groups have confirmed that the W locus is allellic to KIT (Chabot et a!. 1988). and theSllocus toencodeforSCF(Zseboetal, 199Ua).

The myelodysplastic syndromes (MDS) are a heteroge- neous group of clonal preleukaemic disorders characterized by a macrocytic anaemia, with or without neutropenia and thrombocytopenia. Mutations of the KIT related gene, FMS, have already been described in patients with myelodysplasia (Ridge c t al. 1990), but their functional significance is as yet unknown. The potential parallel with macrocytic anaemia in

Correspondence: Dr David Bowen. Department of Haematology. University of Wales College of Medicine, Heath Park, Cardiff CF4 4XN.

W/Sl mutant mice, and the frequent FMS mutations suggests that abnormalities of the KITISCF interaction may play a role in the pathogenesis of MDS.

PATIENTS AND METHODS

Eighty-five patients with MDS were studied. Mean age was 65 years, with a range from 20 to 91. Forty-one patients were female and 44 were male. FAB subtypes were: 35 refractory anaemia (RA), 17 sideroblastic anaemia (SA). 14 chronic myelomonocytic leukaemia (CMML), 10 refractory anaemia with excess blasts (RAEB), four RAEB in transformation (RAEB-t), and five acute myeloid leukaemia transformed from MDS (t-MDS). Blood samples were taken to coincide with routine bone marrow aspirates, or at clinic visits within 3 months of a bone marrow aspirate. Three patients declined bone marrow aspiration. All samples were taken during the morning, and all patients were well at the time of sampling. Blood was also sampled from 10 elderly haematologically normal subjects, also invariably during the morning. Blood samples were taken in accordance with the Guidelines of the South Glamorgan Joint Ethics Committee. Serum was separ- ated and stored at -2OOC. Samples were taken in 1988-89 (nos. 1-59), andin 199 1-92 (nos. 60-85). No patients were receiving courses of cytotoxic therapy or treatment with cytokines at the time of sampling. Red cell transfusion history

6 3

64 David Bowen et a1 was known for 8 1 patients and of those 40 were receiving red cell transfusions at least six weekly.

Haemoglobin concentration, haematocrit, mean cell volume, mean cell haemoglobin, total white cell count and differential and platelet counts were analysed with a Techni- con H1 counter, and bone marrow smears stained with Jenner-Giemsa, to obtain a 400 cell differential count. Per cent myeloblasts and erythroblasts were counted as a percentage of total nucleated cells, and per cent of erythro- blasts showing dyserythropoiesis also noted. Perls stain was used to count per cent ring sideroblasts. Peripheral blood BFU-E were measured with a modification of the method of lscove et a1 (1974) in 64/85 patients.

Serum SCF concentration was measured by enzyme immunoassay (Langley et aJ, 1993). Briefly, test microtitre wells (Dynatech Laboratories, Inc., Alexandria. Va.) were coated with affinity purified polyclonal anti-SCF antibody raised against CHO-derived recombinant human SCF. Serum was added to the test wells, and a standard curve also generated using known concentrations of CHO-derived recombinant human SCF. Bound SCF was measured using the 7H6 monoclonal antihuman SCF antibody (Amgen Inc., Thousand Oaks, Calif.) conjugated to horseradish peroxidase. The assay range for human serum SCF concentration is 0.5- 10 ng/ml. The interassay coefficient of variation (COV) is 8% across the range of the assay. The MDS patient specimens were analysed on the same day with the same reagents on three test plates.

Both log transformed and rank analyses were performed. The Mann-Whitney U test was used to compare the distribu- tion of SCF values between MDS patients and normals. One- way analysis of variance was used to compare SCF concen- trations between FAB subtypes. Spearman Rank Correlation coefficients were computed to define the association between SCF values in MDS patients and other variables.

RESULTS

Mean serum SCF concentration in the MDS patients was 2.8 1 ng/ml (median 2.54. standard deviation= 1.34), with a range from 0.63 to 7.96 ng/ml. This was significantly lower than serum concentrations of samples from 2 34 apparently healthy volunteer donors from IJCLA, Los Angeles, Calif.; mean 3.30 ng/ml (median 3.12, SD 1.1) with a range of 1.3-8.0 ng/ml (L. Bennett, unpublished observations) (P=O.OOOl) , and from a group of 10 age-matched control subjects: mean 3.47 ng/ml (median 3.15, SD 0.87) with a range of 2.38-5.14 ng/ml (mean age=66 years, range 30- 87) from Cardiff ( P = 0.05) (Fig 1). No significant relationship was found between SCF concentration and peripheral blood haemoglobin concentration, haematocrit, mean cell volume, mean cell haemoglobin, total white cell count and differential and platelet counts, BFU-E or per cent bone marrow erythro- blasts, dyserythropoiesis, myeloblasts, ring sideroblasts or karyotype in the MDS patients. However, patients with higher white cell counts (> 14 x lO'/l) all had serum SCF concentrations less than 2.5 ng/ml (Fig 2). There was no statistically significant difference in SCF concentration between FAB subtypes, although inspection of the means

i

m

rn

m I I

NORMALS NORMALS MDS (UCLA) (CARDIFF)

Pig 1. Serum SCF concentration in two cohorts of normal subjects (IJCLA and Cardiff), and in patients with MDS. Mean +standard deviation indicated by bars.

showed a progressive decrease in serum SCF concentration from FAB subtypes RA, through SA. CMML to RAEB (Table I). The mean values for RAEB-t and t-MDS patients did not continue this trend, but the small numbers and wide range of values in these groups make interpretation difficult. No relationship between serum SCF and patient sex, stage of disease (months from diagnosis), patient survival, or red cell transfusion dependence (versus non-transfused) was observed. No significant difference was seen in the values obtained from samples taken in 1988-89 compared with those in 1991-92, suggesting that SCF remains to be detected in serum stored at - 2OoC for at least 4 years. SCF

Table I. Serum SCF concentration by FAB subtype.

Mean SCF FAB N concentration Standard

deviation subtype values (ng/ml)

RA 3 5 2.99 1.1 3 SA 17 2.77 1.76 CMML 14 2.56 1.38 RAEB 10 2 . 3 5 0.92 RAEB-t 4 3 . 1 5 2.26 t-MDS 5 2.97 2.97

Serum Stem Cell Factor Concentrations in MDS 65

100 :

5 Q)' 10: s & I

-

0 z ~~

cn -I - 0 1 :

0.1

Fig 2.

rn

. .

. . % . w. . . . . . .-. . . .9 .

.9 . - .. . 9 .- . 9 .. 8f..

. ' m . . .+ . . 0 -

.

I I I I I I 1

MDS.

concentration was positively correlated with patient age in the MDS group ( r , = 0 . 3 2 ; P<0.01). although examination of the data suggests that this is unlikely to be of clinical relevance. No such age relationship was seen in either group of normal subiects.

DISCUSSION

SCF concentration in serum is high compared with other HGFs such as granulocyte colony-stimulating factor ( 10- 100-fold greater), granulocyte-macrophage colony-stimulat- ing factor (lOO-fold), or erythropoietin (10-100-fold). In those haemopoietic abnormalities studied thus far, namely red cell aplasia and Diamond-Blackfan anaemia (DBA), serum SCF concentrations have been normal except for high values in a patient in haematological remission from DBA (Abkowitz et al, 1991). In myelodysplasia, however, mean SCF concentration is low compared with those of haematolo- gically normal subjects. We cannot be certain that a single measurement reflects the patient's disease state, but the large numbers in each group and the statistical difference between values for the MDS and two groups of normal subjects would suggest that this is so. In addition, the trend to reducing serum SCF concentration from patients with RA, through SA and CMML to KAEB would provide further evidence for the validity of our considering a single measurement in our patients. The serum SCF concentration of the majority of MDS patients was within the range of values for normal subjects, although clustered towards the lower end, with 7/ 85 patients having values below the lower limit of this range. Whether this is an appropriate or inappropriate value for this particular pathological state is unknown. Serum concentra- tions of some other HGFs have been measured in MDS patients. Serum erythropoietin (EPO) concentration is vari-

able (Jacobs e t al, 1989), but some patients have inapprop- riately low levels for the degree of anaemia. It is as yet unknown whether this is due to decreased production or increased utilization of EPO. There is, however, accumulating evidence that these patients will respond better to recombi- nant human EPO therapy than those with higher serum concentrations (Bowen et al, 1991; Erslev, 1991). These low levels are in contrast to the concentration of colony- stimulating factor-1 (CSF-1) and TNF-c( in patients with MDS, which are high (Janowska-Wieczorek et al, 1991; Verhoef et al, 1992).

In mice, SCF exists in membrane-bound and soluble forms, produced by alternative mRNA splicing (Flanagan ef al, 1991). The function of the membrane-bound isoform is unknown, but it is reasonable to speculate that this isoform, found on bone marrow fibroblasts, may bind the KIT protein receptor on haemopoietic stem/progenitor cells, thus facili- tating the interaction of other adhesion molecules and their receptors, and holding cells within a concentration gradient of HGFs in the bone marrow microenvironment. The precise production sites of SCF are still uncertain. SCF was purified and cloned from cultured mouse Balb/3T3 fibroblasts (Flana- gan & Leder, 1990) and Buffalo rat liver cells simultaneously (Zsebo ef al, 1990b). The nature of the SCF-producing cell in liver is unknown. It may indeed be that cells responsible for SCF production are clonally abnormal in some MDS patients. The involvement of bone marrow fibroblasts in the clonal process of myelodysplasia has been little studied. In Dexter long-term bone marrow culture (LTBMC). the stromal layer derived from MDS patient marrow is morphologically abnor- mal, although functionally adequate to support the growth of adherent-cell depleted marrow from normal subjects (Cou- tinho et al, 1990). Fibroblast colony forming cells (CFU-F) derived from MDS stroma in LTBMC occur at a normal

66 David Bowen et a1 frequency, although following continuous subculture, diminished or absent growth was observed (Zipori et al, 1985).

A further explanation for reduced serum SCF concentra- tion is increased utilization of SCF. The routes of elimination of SCF from the serum are unknown. The detectable expression of mRNA and protein of the receptor for SCF, c-kit, on leukaemic blasts in 20/25 patients with acute myeloid leukaemia (in contrast to undetectable expression in normal bone marrow), including five of six patients transformed from MDS (Ikeda et ul, 1991), may provide a mechanism for increased removal of SCF from the circulation.

Finally, there is preliminary evidence from in vitro clono- genic assays that haemopoietic progenitors from MDS patients can be stimulated by the combination of recombi- nant human SCF (rHuSCF) and EPO alone (erythroid col- onies) or rHuSCF and GM-CSF and/or interleukin-3 (myeloid colonies) in vitro (Backx et al, 1992; Goselink et al, 1992). Soluble SCF is now available as a recombinant protein and preliminary studies have demonstrated an in vivo stimulatory effect on haemopoiesis in primates (Andrews et ul, 1991). These combined observations provide support for a thera- peutic trial of rHuSCF to treat the cytopenias in MDS patients, although more data is required on the stimulatory effects of SCF on leukaemic blasts to safely proceed with this.

ACKNOWLEDGMENTS

This study was supported by a grant from the Leukaemia Research Fund of Great Britain. We are grateful to Professor Allan Jacobs whose patients we studied.

REFERENCES

Abkowitz, J.L.. Sabo, K.M., Nakamoto. B., Blau, C.A., Martin, F.H., Zsebo, K.M. & Papayannopoulou. T. (1991) Diamond-Blackfan anemia: in vitro response of erythroid progenitors to the ligand for c-kit. Blood, 9, 2198-2202.

Andrews, R.G., Knitter, G.H., Bartelmez. S.H.. Langley, K.E., Farrar. D., Hendren, R.W., Appelbaum. F.R.. Bernstein, I.D. & Zsebo, K.M. (1991) Recombinant human stem cell factor. a c-kit ligand, stimulates hematopoiesis in primates. Blood, 78, 1975-1980.

Backx, B.. Broeders, L. & Lowenberg. B. (1992) Kit ligand improves in vitro erythropoiesis in myelodysplastic syndromes. Blood. 80,

Bowen, D.T., Culligan. D. & Jacobs, A. (1991) The treatment of anaemia in the myelodysplastic syndromes with human recombi- nant erythropoietin. British Journal of Haematology, 77,419-423.

Chabot. B.. Stephenson, D.A., Chapman, V.M.. Besmer, P. & Bernstein. A. (1988) The proto-oncogene c-kit encoding a trans- membrane tyrosine kinase receptor maps to the mouse W locus. Nature. 335, 88-89.

Coutinho, L.H.. Geary, C.G., Chang. J., Harrison, C. & Testa, N.G. (1 990) Functional studies of bone marrow haemopoietic and stromal cells in the myelodysplastic syndrome (MDS). British ]ourrial of Haematology. 75, 16-25.

121 3-121 7.

Erslev, A.J. (1991) Erythropoietin. New England Journal of Medicine,

Flanagan, J.G., Chan. D.C. & Leder, P. (1991) Transmembrane form of the kit ligand growth factor is determined by alternative splicing and is missing in the SI ‘d’ mutant. Cell, 64, 1025-1035.

Flanagan, J.G. & Leder, P. (1990) The kit ligand: a cell surface molecule altered in Steel mutant fibroblasts. Cell. 63, 185-194.

Goselink, H.M., Williams, D.E., Fibbe. W.E.. Wessels. H., Beverstock, G.C., Willemze, R. & Falkenburg, J.H.F. (1992) Effect of mast cell growth factor (c-kit ligand) on clonogenic leukemic precursor cells. Blood, 80, 750-757.

Ikeda. H., Kanakura, Y., Tamaki. T.. Kuriu, A,, Kitayama, H., Ishikawa, J., Kanayama. Y.. Yonezawa, T., Tarui. S. & Griffin, J.D. (1 99 1) Expression and functional role of the proto-oncogene c-kit in acute myeloblastic leukemia cells. Blood, 78, 2962-2968.

Iscove. N.N., Sieber, F. & Winterhalter, K.H. (1 974) Erythroid colony formation in cultures of mouse and human bone marrow: analysis of the requirement for erythropoietin by gel filtration and affinity chromatography on agarose concanavalin A. Journal of Cell Physiology, 83, 309-320.

Jacobs, A., Janowska-Wieczorek, A., Caro. J., Bowen. D.T. &Lewis, T. (1 989) Circulating erythropoietin in patients with myelodysplas- tic syndromes. British ]ournal of Haematology, 73, 36-39.

Janowska-Wieczorek. A., Belch, A.R.. Jacobs, A., Bowen, D., Padua. R.-A.. Paietta. E. & Stanley, E.R. (1991) Increased circulating colony-stimulating factor-1 in patients with pre-leukemia. leuke- mia and lymphoid malignancies. Blood, 77, 1796-1803.

Langley, K.E.. Bennett, L.G., Wypych, J., Yancik. S.A.. Liu, X.-D., Westcott, K.R.. Chang. D.G., Smith. K.A. & Zsebo, K.M. (1993) Soluble stem cell factor in human serum. Blood, 81, 656-660.

McNiece, I.K., Langley, K.E. & Zsebo. K.M. (1991) Recombinant human stem cell factor synergises with GM-CSF. G-CSF, IL-3. and Epo to stimulate human progenitor cells of the myeloid and erythroid lineages. Experimental Haematology, 19, 226-23 1.

Ridge, S.A., Worwood, M.. Oscier, D., Jacobs, A. & Padua, R.A. (1990) Fms mutations in myelodysplastic, leukemic and normal subjects. Proceedings of the National Academy of Sciences of the United States of America, 87, 1377-1380.

Verhoef, G.E.G.. De Schouwer, P., Ceuppens, J.L.. Van Damme. J.. Goossens, W. & Boogaerts, M.A. (1992) Measurement of serum cytokine levels in patients with myelodysplastic syndromes. Leukaemia, 6, 1268-1272.

Zipori, D., Reichman, N., Arcavi, L., Shtalrid, M., Berrebi, A. & Resnitzky, P. (1 985) In vitro functions of stromal cells from human and mouse bone marrow. Experimental Haematology, 13, 603- 609.

Zsebo. K.M., Williams, D.A.. Geissler. E.N., Broudy. V.C., Martin, F.H., Atkins, H.L., Hsu. R.-Y., Birkett. N.C., Okino, K.H., Murdock, D.C., Jacobsen, F.W., Langley, K.E.. Smith, K.A.. Takeishi. T., Cattanch, B.M. & Suggs, S.V. (1990a) Stem cell factor is encoded at the S1 locus of the mouse and is the ligand for the c-kit tyrosine kinase receptor. Cell. 63, 213-224.

Zsebo, K.M., Wypych, J., McNiece. I.K., Lu. H.S.. Smith, K.A., Karkare, S.B., Sachdev, R.J., Yuschenkofl, V.N.. Birkett. N.C.. Williams, L.R., Satyagal. V.N., Tung. W., Bosselman, R.A.. Mendiaz, E.A. & Langley, K.E. (1 990b) Identification, purification and biological characterisation of haemopoietic stem cell factor from Buffalo-rat liver conditioned medium. Cell, 63, 195-201.

324,1339-1344.