Differences in the distribution of CD34 epitopes ... - Wiley Online Library

British Journal of Haematology, 1996, 94, 597–605

Differences in the distribution of CD34 epitopes on normal haemopoietic progenitor cells and leukaemic blast cells R I TA S TEEN , 1 G EIR E. T J ØNNFJ OR D , 1,2 G USTAV G AU D ER NACK , 1 L ORENTZ B R I NCH 2 AND T OR STEIN E G ELAND 1 Institute of Transplantation Immunology and 2Medical Department A, The National Hospital, University of Oslo, Norway

1

Received 1 April 1996; accepted for publication 6 June 1996

Summary. The CD34 molecule expressed on haemopoietic progenitor cells contains a large number of epitopes whose expression may be related to the maturation or function of the cells. Monoclonal antibodies specific for different epitopes have been reported to detect different numbers of CD34 leukaemic blast cells. We wanted to confirm this observation and study whether parallel findings could be observed for normal haemopoietic progenitor cells. The cells were immunophenotyped by flow cytometry with a series of monoclonal antibodies reactive with different CD34 epitopes. Class III epitopes (resistant to enzymatic cleavage with neuraminidase, chymopapain and a glycoprotease from Pasteurella haemolytica) showed a broader distribution on normal haemopoietic progenitor cells and leukaemic blast cells than class I epitopes (sensitive to cleavage with all three enzymes) and class II epitopes (sensitive to degradation with glycoprotease and chymopapain, and resistant to neuraminidase). The subpopulation of normal progenitor cells which exclusively expressed class III epitopes had flow cytometric characteristics compatible with mature myeloid progenitor

The CD34 molecule is a transmembrane glycoprotein which is expressed on haemopoietic progenitor cells (Civin et al, 1984; Tindle et al, 1984; Andrews et al, 1986; Watt et al, 1987), vascular endothelium, high endothelial venules (Watt et al, 1987; Fina et al, 1990; Baumhueter et al, 1993) and some fibroblasts (Greaves et al, 1995; Parravicini et al, 1995). The molecule has been biochemically characterized as a mucin-like sialoglycoprotein which contains a large number of O-linked and nine N-linked glycosylation sites (Sutherland et al, 1988; Girard et al, 1995). A high number of epitopes can be detected on the CD34 molecule (Watt et al, 1987; Sutherland & Keating, 1992; Sutherland et al, 1992), and a classification of these epitopes by grouping them into three classes according to their differential sensitivity to enzymatic Correspondence: Dr Rita Steen, Institute of Transplantation Immunology, Rikshospitalet, The National Hospital, N-0027 Oslo 1, Norway.

# 1996 Blackwell Science Ltd

cells, whereas class I, II and III epitopes were equally expressed on cells enriched for immature subsets. No discordant epitope expression could be observed for the more immature leukaemias (AML-M0/1) and a higher percentage of the more mature leukaemic blast cells (AMLM3 and AML-M4/5) expressed class III epitopes compared to the percentage expressing class I and II epitopes. These data indicate that CD34 class III epitopes are more broadly distributed on normal haemopoietic progenitor cells and leukaemic blast cells than class I and II epitopes, and that class I and II epitopes may be down-regulated prior to class III epitopes during normal haemopoietic progenitor cell differentiation. These findings should be considered when selecting CD34 mabs for quantification and positive selection of haemopoietic progenitor cells for research and clinical purposes. Keywords: CD34 epitopes, haemopoietic progenitor cells, CD34 acute leukaemias, CD34 monoclonal antibodies.

cleavage has been suggested (Sutherland et al, 1992). Class I epitopes are sensitive to cleavage with neuraminidase, chymopapain, and a glycoprotease from Pasteurella haemolytica; class II epitopes are sensitive to enzymatic cleavage with glycoprotease and chymopapain, and resistant to cleavage with neuraminidase; and class III epitopes are resistant to cleavage with all three enzymes. This nomenclature has been adapted by the Leukocyte Typing V Workshop (Greaves et al, 1995). Little is known about the function of the CD34 molecule on haemopoietic progenitor cells. However, recent reports suggest that CD34 may play a role in progenitor cell localization/adhesionin the bone marrow (BM) (Majdic et al, 1994; Healy et al, 1995). The CD34 molecule on endothelial cells may provide an adhesive binding site recognized by Lselectin on circulating lymphocytes (Baumhueter et al, 1993). Interestingly, the isoform of the CD34 molecule found on high endothelial venules expresses class II and III epitopes,

597

598

Rita Steen et al

whereas class I epitopes are absent (Baumhueter et al, 1993; Greaves et al, 1995). Reports on the expression of different CD34 epitopes on normal haemopoietic progenitor cells and leukaemic blast cells are inconsistent (Lanza et al, 1994; Egeland & Gaudernack, 1994a; Gaudernack & Egeland, 1995; Greaves et al, 1995; Titley et al, 1995; Traore et al, 1995). Furthermore, the biological significance of the distribution of class I, II and III epitopes on different CD34 cell subpopulations is not known. The present work was designed to study how the CD34 epitope distribution varies on normal haemopoietic progenitor cells and acute leukaemic blast cells. In particular, we wanted to determine whether certain distributionpatterns correlate with given subpopulations of these cells. Normal haemopoietic progenitor cells from human BM, umbilical cord blood (UCB), and from peripheral blood (PB) during short-term administration of granulocyte-colony stimulating factor (G-CSF), and leukaemic blast cells were studied by flow cytometry employing a series of CD34 epitope class I, II and III reactive monoclonal antibodies (mabs). We found differences in the distribution of CD34 epitope classes on both leukaemic blast cells and normal haemopoietic progenitor cell subsets, which might reflect their stage of maturation. MATERIALS AND METHODS Cell preparation. Heparinized samples from BM, UCB, and from PB during G-CSF administration, were obtained following informed consent under regulations provided by the regional ethical committee. BM was obtained by aspiration from the posterior iliac crest. UCB samples were collected during vaginal or caesarean full-term deliveries after normal pregnancies. PB samples were obtained from healthy adults on day 4 or 5 of G-CSF administration, 10 mg/kg/d s.c. (Tjønnfjord et al, 1994). In addition, cryo-preserved light density BM cells from 20 patients with acute myeloid leukaemia (AML) and eight patients with acute lymphoblastic leukaemia (ALL) were included in the study. These leukaemias had all been immunophenotyped as CD34 by the institute’s routine screening method (Gaudernack &

Lundin, 1989), using the CD34-specific class I mab BI3C5 (Tindle et al, 1984) and the class III mab 561 (Gaudernack & Egeland, 1995). The leukaemias were classi-fied according to the French–American–British (FAB) classi-fication of acute leukaemias (Bennett et al, 1976, 1985, 1991). Mabs used for flow cytometry. Table I shows the CD34specific mabs that were used in flow cytometry: unconjugated, fluorescein isothiocyanate (FITC)- or phycoerythrin (PE)-labelled mab 8G12 (anti-HPCA-2) and unconjugated mab My-10 (anti HPCA-1) [Becton Dickinson Immunocytometry Systems (BDIS), San Jose, Calif., U.S.A.], unconjugated, FITC- and PE-labelled mab QBEND-10 (Serotec, Oxford, U.K.) and ICH3 (Caltag, San Francisco, Calif., U.S.A.), and unconjugated mab 547 (Institute of Transplantation Immunology, Rikshospitalet, Oslo, Norway). In addition, we used the CD34 mabs 561, 566 and 581, which were developed in our laboratory (Gaudernack & Egeland, 1995), and directly FITC- and PE-labelled either by ourselves or BDIS. All the mabs except 561 and 566 have been clustered as CD34 specific at previous workshops [Leukocyte Typing Workshop IV and V (Civin et al, 1989; Greaves et al, 1995)]. All mabs were screened against a series of CD34 cell lines including the myeloid leukaemic cell lines KG1A expressing CD34 at high levels and HL-60 transfected with the CD34 gene (a kind gift from Adrian P. Gee), and against CD34-negative cell lines including the untransfected cell line HL-60. The specificity of the mabs was also confirmed by blocking experiments (Gaudernack & Egeland, 1995). Isotype-matched, irrelevant mabs served as negative controls. Simultest Con-trol 1/ 2 (FITC-IgG1 PE-IgG2a) and 1 (PE-IgG1) were obtained from BDIS, monoclonal mouse-IgG1 and -IgG2a were obtained from PharMingen (San Diego, Calif., U.S.A.), and FITC- and PE-goat anti-mouse IgG1 and IgG2a polyclonal antibodies were purchased from Southern Biotechnology Associates, Inc. (Birmingham, Ala., U.S.A.). Staining and flow cytometry. Staining for flow cytometry was performed on unfractionated PB, UCB and BM, and on light density gradient separated cells (Lymphoprep; 1.077 g/ml; Nycomed Pharma, Oslo, Norway) from acute

Table I. CD34-specific monoclonal antibodies (mabs) used in the present study.

Mabs

Epitope class

My10 (HPCA-1) 547

I I

QBEND10 ICH3

II II

8G12 (HPCA-2) 561 566 581

III III III III

Neuraminidase sensitivity

Glycoprotease/ chymopapain sensitivity

ÿ ÿ ÿ ÿ ÿ ÿ

ÿ ÿ ÿ ÿ

Isotype (Ig)

Immunofluorescence

IgG1 IgG1

Indirect* Indirect*

IgG1 IgG2a

Direct/indirect Direct/indirect

IgG1 IgG2a IgG1 IgG1

Direct/indirect Direct/indirect Direct/indirect Direct/indirect

The mabs are listed according to the classification of their corresponding epitopes. * Since these mabs are unable to bind to CD34 after fluorochrome conjugation, they can only be used in indirect immunofluorescence.

# 1996 Blackwell Science Ltd, British Journal of Haematology 94: 597–605

CD34 Epitope Distribution leukaemias. Unfractionated samples were lysed by FACS Lysing Solution (BDIS) after staining, and all cell samples were fixed in 1.0% paraformaldehyde prior to flow cytometric analyses. Flow cytometric analyses were performed on a FacScan (BDIS) or a FacSort (BDIS) equipped with an argon-ion laser tuned at 488 nm. The Lysis II or the CellQuest software were used for data acquisition and analysis. A gate for viable cells was set according to forward and side light scatter (data not shown). For the purpose of enumeration of CD34 cells and in order to obtain a sufficient number of CD34 cells for epitope analyses on normal haemopoietic progenitor cells, an acquisition gate was set according to low side light scatter and a fluorescence intensity range comprising the cells with positive CD34 fluorescence signals (Fig 1, B and C). A total of 0.25 2 106 to 1.0 2 106 events were run through the cytometer per test and 2000–7500 events within the acquisition gate were stored in list mode data files. To minimize background fluorescence signals, only events which fell within a CD34 gate in forward versus side light scatter

599

cytogram (Fig 1, D and E, region R1 and R2) were accepted as CD34 cells (Smeland et al, 1992; Steen et al, 1994). The number of CD34 cells was related to the total number of viable cells. These steps were performed to optimize CD34 cell quantification to allow analysis of low concentrations of certain epitopes on some cell subsets. Statistics. The data were shown to conform to a normal distribution by calculation of standardized skewness and kurtosis. Multiple group comparisons were performed using analysis of variance (ANOVA), and group versus group comparisons were based on a post hoc t-test with the Bonferroni method for correction of P values (Altman, 1994). A P value < 0.05 was considered significant. RESULTS Distribution of CD34 epitopes on leukaemic blast cells from acute lymphoblastic and myeloid leukaemias Table II summarizes the analyses of the acute leukaemias. To evaluate inter-epitope class variability, the reactivity of

Fig 1. (A-C) Representative cytograms [side scatter (cell granularity) versus fluorescence intensity] of light-densitycells obtained from normal bone marrow. (A) Irrelevant subclass matched negative control. (B) Reactivity with the class III mab 8G12 (HPCA-2). (C) Reactivity with the class II mab QBEND10. The illustrations indicate how an acquisition gate was set according to low side scatter and fluorescence intensity. (D and E) Forward scatter (cell size) versus side scatter cytograms of gated events in B and C, respectively. Only events which fell within a CD34 gate (region R1 and R2) were accepted as CD34 cells for further analyses and calculation.


600

Rita Steen et al Table II. Summary of the expression of CD34 epitopes on leukaemic blast cells from eight acute lymphoblastic leukaemias (ALL) and 20 acute myeloid leukaemias (AML).

FAB classification

No. of leukaemias

Class I

Class II

Class III

ALL-L1 ALL-L2 AML-M0 AML-M1 AML-M2 AML-M3 AML-M4/5

3 5 5 1 6

81 (73–95) 58 (16–88) 94 (83–100) 39 (38–40) 46 (4–83)

79 (61–94) 56 (22–91) 95 (87–99) 38 (38–38) 45 (13–78)

84 (77–95) 67 (35–96) 96 (88–100) 38 (35–39) 53 (22–89)

1

32 (25–40)

25 (21–30)

54 (46–60)

7

38 (12–77)

29 (17–55)

60 (31–93)

(> 98) > 85 (< 1)

(> 97) > 80 (< 1)

(> 99) > 85 (< 1)

KG1A HL-60tranf. HL-60

g

*, †

The percentage of all light-density cells stained with the CD34-specific mabs are given as the mean value and range in parentheses. The data are simplified by pooling the reactivity of individual mabs within the same epitope class for every subgroup of the acute leukaemias. Positive controls included KG1A and HL-60 transfected with the CD34 gene (HL-60transf.). Statistical analyses of the difference between the number of cells stained with the class I, II and III reactive mabs: * P 0.0005 and †P < 0.0001 for the I versus III and II versus III mabs, respectively. No statistical differences were found for the other leukaemia subgroups.

individual mabs within the same epitope class was pooled for every subgroup of the leukaemias. The mean percentage of the light-density cell fraction that was stained is given for each FAB group. Leukaemic blast cells of the ALL-L1, ALL-L2 and the AML-M0, AML-M1 and AML-M2 subtypes did not show differences in expression of different CD34 epitope classes. However, a higher number of leukaemic blast cells of the AML-M3 and the AML-M4/5 expressed the class III epitopes than class I and II epitopes. Fig 2 provides an example of the difference in reactivity of the CD34-specific mabs with blast cells of a leukaemic cell line (KG1A) (A), two CD34 AMLs (B and C) and a CD34-negative AML (D). As can be seen in the AML-M4/5 example, the number of positive cells and the fluorescence intensity of these cells, i.e. epitope density, are reduced for class I and II mabs compared to class III mabs. AML-M0 and AML-M1 represent less differentiated leukaemic blast cells than blast cells of the AML-M3 and AMLM4/5 subtypes (Bennett et al, 1976). To determine whether discordant epitope expression was a feature of the maturation stage of the acute leukaemic blast cells, we compared the epitope distribution on blast cells of the AML-M0 and AMLM1 subtypes with blast cells of the AML-M3 and AML-M4/5 subtypes (Table II). The class I and II mabs bound a lower number of AML-M3 and AML-M4/5 blast cells than did the class III mabs (P 0.0005 and P < 0.0001 for class III versus class I mabs, and class III versus class II mabs, respectively). No statistical difference between the number of cells stained with the class I and II mabs was observed. For the AML-M0 and AML-M1 blast cells, the class I, II and III mabs reacted with the same proportion of cells. Furthermore, despite variations in fluorescence intensity between blast cells of the same FAB subgroup, the density of CD34 epitopes tended to be

higher on the AML-M0 and AML-M1 than on the AML-M3 and AML-M4/5 blast cells. Moreover, the number of cells which stained with either of the CD34-specific mabs was significantly higher in the light-density cell fraction of the AML-M0 and AML-M1 patients than in the light-density cell fraction of the AML-M3 and AML-M4/5 patients (P < 0.0001). Distribution of CD34 epitopes on normal haemopoietic progenitor cells Since leukaemic blast cells may be regarded as clonally expanded haemopoietic progenitor cells at a ‘fixed’ stage of maturation, it was of interest to investigate whether some of the differences in CD34 epitope distribution on leukaemic blast cells could be observed on normal haemopoietic progenitor cells. Flow cytometric analyses showed differences in epitope class expression. The number of cells which stained with the class III mabs was significantly higher than those observed for the class I or II mabs (P 0.05 and P 0.02, respectively) (Table III), but there were no significant differences between the number of cells stained with the class I and II mabs. To exclude the possibility that the difference in CD34 epitope class expression was primarily a feature of circulating haemopoietic progenitor cells (i.e. progenitor cells in UCB and in PB during G-CSF administration), BM cells from healthy individuals in steadystate haemopoiesis were analysed separately. We found the same pattern of CD34 epitope distribution on BM progenitor cells, i.e. the number of cells stained with the class III mabs was significantly higher than the number of cells stained with the class II mabs (P 0.018). This was also the trend when comparing the class III with the class I


CD34 Epitope Distribution

601

Fig 2. Reactivity of the CD34-specific monoclonal antibodies (mabs), listed according to the classification of their corresponding epitopes, with leukaemic blast cells in one CD34-highly positive myeloid leukaemic cell line KG1A (A), one immature AML of the M0 subtype (B), and a more mature AML of the M4/5 subtype (C). The reactivity with a mature CD34-negative acute leukaemia of the subtype AML-M4/5 is shown for comparison (D). Since the IgG subclass-matched negative controls were overlapping, only one irrelevant negative control (Irr. mab) is shown. The relative cell number is shown on the vertical axis. The horizontal axis shows the log fluorescence intensity on a scale from 100 to 104.

Table III. Summary of the expression of CD34 epitopes on normal haemopoietic progenitor cells from seven bone marrow (BM) samples, four umbilical cord blood (UCB) samples, and four samples of peripheral blood (PB) obtained during G-CSF administration. The CD34 epitope expression on BMCD34 cells is also shown separately. Percentage of CD34 cells Pooled samples from BM, UCB and PB (n 15)

BM samples (n 7) 0.41 (0.34–0.53) 0.42 (0.33–0.49) 0.30 (0.23–0.42) 0.33 (0.30–0.38) 0.61 (0.47–0.96) 0.50 (0.34–0.78) 0.53 (0.43–0.72) 0.51 (0.41–0.73)

Mabs

Epitope class

My10 (HPCA-1) 547

I I

QBEND10 ICH3

II II

0.31 (0.02–0.53) 0.29 (0.07–0.49) 0.23 (0.04–0.42) 0.23 (0.02–0.38)

8G12 (HPCA-2) 561 566 581

III III III III

0.36 (0.07–0.96) 0.44 (0.04–0.78) 0.33 (0.04–0.72) 0.31 (0.05–0.73)

The percentage of all cells stained with the CD34-specific mabs is given as the mean value and range in parentheses. n: number of samples.


602

Rita Steen et al

Fig 3. (A and B) Forward scatter (cell size) versus side scatter (cell granularity) cytograms of normal BM haemopoietic progenitor cells after staining with the class III mab 8G12 (HPCA-2) (A) and the class II mab QBEND10 (B). According to cell size and granularity, the two regions labelled R1 and R2 were drawn. The region R2 was further divided vertically into two equal areas; R2left and R2right representing the left and right part of the region R2, respectively. (C) Histogram showing cell size (x-axis) and number of cells ( y-axis). The filled curve represents cells which were stained with the class II reactive mab QBEND10 and the open curve represents cells stained with the class III mab 8G12 (HPCA-2). Note that more cells are confined to the region R2right when the samples were stained with the class III mab compared to staining with the class II mabs.

mabs; however, the differences did not reach statistical significance (Table III). The cells which stained with CD34-specific mabs could be divided into two regions labelled R1 and R2 according to cell size and granularity (Fig 3, A and B) (Smeland et al, 1992). The region R1 contained small agranular cells and the region R2 contained larger, slightly granular cells. The small agranular cells in the region R1 are known to represent mainly B-cell precursors (Smeland et al, 1992; Tjønnfjordet al, 1994). As previously described (Steen et al, 1994), nearly all UCB- and PB-CD34 cells were confined within the region R2 (96.8 6 1.4% and 95.1 6 2.3% respectively), whereas the region R2 contained only 81.5 6 5.4% of all BM-CD34 cells. However, independent of the source of CD34 cells, an equal proportion of the CD34 cells in the region R1 expressed the class I, II and III epitopes. The region R2 contains a heterogenous mixture of CD34 cells. Since immature CD34 cell subsets are enriched in the left part of the region R2 (R2left) (Terstappen et al, 1991), whereas the right part (R2right) contains mainly the more differentiated myeloid progenitor cell subsets (Andrews et al, 1989; Terstappen et al, 1991; Tjønnfjord et al, 1995), the region R2 was divided vertically into two equal areas (Fig 3, A and B). A typical finding is illustrated in Fig 3, i.e. more cells were confined to the region R2right when the samples were stained with the class III mabs (A) compared to staining with the class II mabs (B). The same observation is also illustrated in the histogram (C). Fig 4 summarizes a number of experiments, analysing CD34 epitope expression in the region R2. Irrespective of the mabs used, the number of CD34 cells in the region R2left was comparable within each experiment. However, when analysing cells within the region R2right, a significantly higher number of cells was detected when class III mabs were used for staining compared to class I and II mabs (P 0.03 and

Fig 4. The percentage of CD34 cells within the region R2right of the CD34 cells in the region R2 for BM, UCB, and PB after staining with the CD34-mabs ( h, class I mabs; s, class II mabs; g, class III mabs). The data are simplified for presentation by pooling the reactivity of individual mabs within the same epitope class for every single sample, each column represents the mean value, the bars indicate standard deviation. Each group of three columns represents one experiment. 3, Epitope class II expression not determined.


CD34 Epitope Distribution 0.0005, respectively). Taken together, the findings show that there are normal haemopoietic progenitor cells which express CD34 class III, but not class I and II epitopes, and that these cells are confined to the region R2right. DISCUSSION The CD34 molecule, first identified with the mab My10 (Civin et al, 1984), is a transmembrane glycoprotein rich in O-linked carbohydrates and sialic acid (Watt et al, 1987; Sutherland et al, 1988). The heavily glycosylated cell surface molecule gives rise to a number of different epitopes (Watt et al, 1987; Sutherland & Keating, 1992; Sutherland et al, 1992). The expression of these epitopes may be related to function and/or maturation (Egeland & Gaudernack, 1994b; Majdic et al, 1994; Greaves et al, 1995; Krause et al, 1996). The presence or absence of different glycosylation-related epitopes is therefore of interest for the study of haemopoiesis and for immunophenotyping and positive selection of CD34 cells. Ultimately, this information may provide some clues to the functional significance of the CD34 molecule on haemopoietic progenitor cells. Since experiments with CD34 endothelial cells have shown that high endothelial venules selectively lack expression of class I epitopes, while class II and III epitopes are expressed (Baumhueter et al, 1993), it was of interest to study potential differences in CD34 epitope expression on haemopoietic cells. The present findings on leukaemic blast cells show that differential expression of CD34 epitopes is not restricted to endothelial cells. We demonstrate that all the CD34-specific mabs studied reacted with CD34 leukaemic blast cells in acute leukaemias, but the reactivity differed between the morphological subgroups of the leukaemias. Thus, for the AML-M3 and AML-M4/5, i.e. leukaemic blast cells with a relatively mature phenotype, the number of cells which stained with the class III mabs was significantly higher than the number of cells which stained with either of the class I or II mabs. No differences in the reactivity between different epitope class specific mabs were observed for the AML-M0 and AML-M1. These findings are in accordance with previous observations (Egeland & Gaudernack, 1994a). No consistent pattern in epitope class expression could be observed on leukaemic blast cells from the ALL and the AML-M2 patients. One possible explanation for the variability in epitope class expression on these subgroups of leukaemias might be that they are heterogenous with respect to the stage of maturation. The present data indicate that there is a selective reduction of expression of class I and II epitopes by the mature CD34 leukaemic blast cells compared to the more immature CD34 blast cells. Based on this observation, we hypothesized that similar differences in CD34 epitope expression might also be found during maturation of normal haemopoietic progenitor cells. In accordance with the findings on leukaemic blast cells, we observed that CD34 class III epitopes were more frequently expressed on normal haemopoietic progenitor cells than class I and II epitopes. The cells which stained with class III mabs, but not class I or II mabs, were relatively large and slightly

603

granular, i.e. cells with characteristics of mature myeloid progenitor cells (Andrews et al, 1989; Terstappen et al, 1991; Tjønnfjord et al, 1995). Smaller, agranular cells, which are known to have a high content of immature, uncommitted subsets (Terstappen et al, 1991), expressed comparable numbers of class I, II and III epitopes. Multicolour immunophenotyping and functional studies are in progress to verify the hypothesis that reduction in expression of class I and II epitopes are confined to mature myeloid progenitor cells. Preliminary data indicate that CD34 cells exclusively stained with class III mabs display a mature myeloid progenitor cell phenotype and are enriched for colony forming unit-granulocyte macrophages (CFU-GM) (unpublished observation). Normal haemopoietic progenitor cells were obtained from three different compartments. There are several reports to support the view that there are differences both between neonatal and adult haemopoietic progenitor cells (Kinniburgh & Russell, 1993; Lansdorp et al, 1993; Traycoff et al, 1994) and between circulating and quiescent haemopoietic progenitor cells (Steen et al, 1994; To et al, 1994; Wickenhauser et al, 1995; Tarella et al, 1995). One may therefore speculate that the CD34 molecule expressed by circulating cells in UCB and in PB after G-CSF administration, display other epitopes than quiescent BM CD34 cells of adults. However, our data show that BM CD34 progenitor cell subsets also selectively express class III epitopes. Thus, lack of expression of class I and II epitopes is not a consequence or a prerequisite for being in a circulating state or being exposed to exogenous growth factors. A decreased intensity of CD34 staining per se has been associated with maturation of the progenitor cell population (Caux et al, 1989; Sutherland et al, 1989; Okumura et al, 1992). This has been explained by a decrease in the number of CD34 molecules on individual cells. An additional phenomenon related to maturation might be that the expression of class I and II epitopes is selectively reduced on individual CD34 molecules during progenitor cell differentiation. This view is supported by the present study. Several explanations may account for this; enzymatic degradation and/or changes in glycosyl transferase activity during maturation may selectively reduce the expression of class I and II epitopes, whereas the glycosylation independent class III epitopes are unaffected. A reduction in class I and II epitopes may also cause ‘hidden’ class III epitopes to become detectable (Sutherland & Keating, 1992; Gaudernack & Egeland, 1995). Thus, during maturation of the progenitor cells, class I and II specific immunofluorescence signals may go below detection level whereas exposure of ‘hidden’ epitopes causes a maintenance of detectable epitope class III specific immunofluorescence intensity. To conclude, our findings demonstrate that CD34 class III epitopes, compared to class I and II epitopes, are more broadly distributed on both normal haemopoietic progenitor cells and leukaemic blast cells, and that down-regulation of CD34 class I and II epitope expression might reflect maturation of the progenitor cells. This observation is important for selection of CD34-specific mabs for immunophenotyping and enumeration of CD34 cells, and must be considered in relation to


604

Rita Steen et al

positive selection of CD34 cells for research and clinical purposes. For haemopoietic stem cell transplantation, one implication may be that CD34 class III mabs should be preferred for positive selection of the cells to ensure a high yield of all subpopulations of CD34 cells. ACKNOWLEDGMENTS This work was supported by a research fellowship from the Norwegian Cancer Society. The authors gratefully acknowledge the excellent technical assistance of Lill Anny Grøseth and Eva Borreti (The National Hospital, Oslo, Norway). We thank Hans E. Johnsen (Herlev Hospital, University of Copenhagen, Herlev, Denmark) for providing some of the bone marrow samples from acute leukaemic patients, and Narve Moe (The National Hospital, Oslo, Norway) for making umbilical cord blood samples available for the study. We also thank Adrian P. Gee (University of South Carolina, South Carolina, U.S.A.) for providing the CD34-transfected cell line HL-60. We are endebted to Lars Mørkrid (The National Hospital, Oslo, Norway) for his helpful assistance with the statistical analysis and to Erik Thorsby (The National Hospital, Oslo, Norway) for helpful comments on the manuscript.

REFERENCES Altman, D.G. (1994) Multiple comparisons. Practical Statistics for Medical Research, pp. 210–211. Chapman & Hall, London. Andrews, R.G., Singer, J.W & Bernstein, I.D. (1986) Monoclonal antibody 12.8 recognizes a 115-kD molecule present on both unipotent and multipotent hematopoietic colony-forming cells and their precursors. Blood, 67, 842–845. Andrews, R.G., Singer, J.W. & Bernstein, I.D. (1989) Precursors of colony-forming cells can be distinguished from colony-forming cells by expression of the CD33 and CD34 antigens and light scatter properties. Journal of Experimental Medicine, 169, 1721–1731. Baumhueter, S., Singer, M.S., Henzel, W., Hammerich, S., Renz, M., Rosen, S.D. & Lasky, L.A. (1993) Binding of L-selectin to the vascular sialomucin CD34. Science, 262, 436–438. Bennett, J.M., Catovsky, D., Daniel, M.T., Flandrin, G., Galton, D.A.G., Gralnick, H.R. & Sultan, C. (1976) Proposals for the classification of the acute leukaemias. British Journal of Haematology, 33, 451–458. Bennett, J.M., Catovsky, D., Daniel, M.T., Flandrin, G., Galton, D.A.G., Gralnick, H.R. & Sultan, C. (1985) Proposed revised criteria for the classification of acute myeloid leukaemia. Annals of Internal Medicine, 103, 626–629. Bennett, J.M., Catovsky, D., Daniel, M.T., Flandrin, G., Galton, D.A.G., Gralnick, H.R. & Sultan, C. (1991) Proposals for the recognition of minimally differentiated acute myeloid leukaemia (AML-M0). British Journal of Haematology, 78, 325–329. Caux, C., Favre, C., Saeland, S., Duvert, V., Mannoni, P., Durand, I., Aubry, J.-P. & de Vries, J.E. (1989) Sequential loss of CD34 and class II MHC antigens on purified cord blood hematopoietic progenitors cultured with IL-3: characterization of CD34ÿ , HLA-DR cells. Blood, 74, 1287–1294. Civin, C.I., Strauss, L.C., Brovall, C., Fackler, M.J., Schwartz, J.F. & Shaper, J.H. (1984) Antigenic analysis of hematopoiesis. III. Hematopoietic progenitor cell surface antigen defined by a monoclonal antibody raised against KG-1a cells. Journal of Immunology, 133, 157–165.

Civin, C.I., Trishmann, T.M., Fackler, M.J., Bernstein, I.D., Buehring, H.J., Campos, L., Greaves, M.F., Kamoun, M., Katz, D.R., Lansdorp, P.M., Look, A.T., Seed, B., Sutherland, D.R., Tindle, R.W. & Uchanska-Ziegler, B. (1989) Report on the CD34 cluster workshop. Leukocyte Typing IV (ed. by W. Knapp et al), pp. 818– 825. Oxford University Press, New York. Egeland, T. & Gaudernack, G. (1994a) Immunomagnetic isolation of CD34 cells: methodology and monoclonal antibodies. Haematopoietic Stem Cells. The Mulhouse Manual (ed. by E. Wunder, H. Sovalat, P. R. Henon and S. Serke), pp. 141–148. AlphaMed Press, Ohio. Egeland, T. & Gaudernack, G. (1994b) CD34 the gateway to the study of lymphohematopoietic progenitor and leukemic cells. The Immunologist, 2/2, 65–70. Fina, L., Molgaard, H.V., Robertson, D., Bradley, N.J., Monoghan, P., Delia, D., Sutherland, D.R., Baker, M.A. & Greaves, M.F. (1990) Expression of the CD34 gene in vascular endothelial cells. Blood, 75, 2417–2426. Gaudernack, G. & Egeland, T. (1995) Epitope mapping of 33 CD34 mAbs, including the Fifth Workshop panel. Leukocyte Typing V (ed. by S. F. Schlossman et al), pp. 861–864. Oxford University Press. Gaudernack, G. & Lundin, K.E.A. (1989) Rapid immunomagnetic phenotyping of cells. Journal of Immunogenetics, 16, 169–175. Girard, J.P. & Springer, T.A. (1995) Expression of Sialomucin CD34 by high endothelial (HEV) in human tonsils. Leukocyte Typing V (ed. by S. Schlossman et al), pp. 1801–1803. Oxford University Press. Greaves, M.F., Titley, I., Colman, S.M., Buehring, H.-J., Campos, L., Castoldi, G.L., Garrido, F., Gaudernack, G., Girard, J.P., InglesEsteve, J., Invernizzi, R., Knapp, W., Lansdorp, P.M., Lanza, F., MerleBeral, H., Parravicini, C., Razak, K., Ruiz-Cabello, F., Springer, T.A., van der Schoot, C.E. & Sutherland, D.R. (1995) Leukocyte Typing V (ed. by S. Schlossman et al), pp. 840–846. Oxford University Press. Healy, L., May, G., Gale, K., Grosveld, F., Greaves, M. & Encer, T. (1995) The stem cell antigen CD34 functions as a regulator of hematopoietic cell adhesion. Proceedings of the National Academy of Sciences of the United States of America, 92, 12240–12244. Krause, D.S., Fackler, M.J., Civin, C.I. & May, W.S. (1996) CD34: structure, biology, and clinical utility. Blood, 87, 1–13. Kinniburgh, D. & Russell, N.H. (1993) Comparative study of CD34positive cells and subpopulations in human umbilical cord blood and bone marrow. Bone Marrow Transplantation, 12, 489–494. Lansdorp, P.M., Dragowska, W. & Mayani, H. (1993) Ontogenyrelated changes in proliferative potential of human haematopoietic cells. Journal of Experimental Medicine, 178, 787–791. Lanza, F., Moretti, S., Castagnari, B. & Castoldi, G. (1994) Complex pattern of reactivity of CD34+ antibodies in normal and leukaemic cells. (Abstract). European Stem Cell Club Meeting, Barcelona. Majdic, O., Stockl, J., Pickl, W.F., Bohuslav, J., Strobl, H., Scheinecker, C., Stockinger, H. & Knapp, W. (1994) Signalling and induction of enhanced cytoadhesivness via the hematopoietic progenitor cell surface molecule CD34. Blood, 83, 1226–1234. Okumura, N., Tsuji, K. & Nakahata, T. (1992) Changes in cell surface antigen expression during proliferation and differentiation of human erythroid progenitors. Blood, 80, 642–650. Parravicini, C., Capra, M., Soligo, D., Caneva, L., Bifulco, C., Berti, E., Cattoretti, G. & Delia, D. (1995) CD34 expression on fibroblastic cells. Leukocyte Typing V (ed. by S. Schlossman et al), pp. 857–858. Oxford University Press. Smeland, E.B., Funderud, S., Kvalheim, G., Gaudernack, G., Rasmussen, A.-M., Rusten, L., Wang, M.Y., Tindle, R.W., Blomhoff, H.K. & Egeland, T. (1992) Isolation and characterization of human hematopoietic progenitor cells: an effective method for positive selection of CD34 cells. Leukemia, 6, 845–852.


CD34 Epitope Distribution Steen, R., Tjønnfjord, G.E. & Egeland, T. (1994) Comparison of the phenotype and clonogenicity of normal CD34 cells from umbilical cord blood, granulocyte colony-stimulating factor-mobilized peripheral blood, and adult human bone marrow. Journal of Hematotherapy, 3, 253–261. Sutherland, D.R., Watt, S.M., Dowden, G., Karhi, K., Baker, M.A., Greaves, M.F. & Smart, J.E. (1988) Structural and partial amino acid sequence analysis of the human hematopoietic progenitor cell antigen CD34. Leukemia, 2, 793–803. Sutherland, D.R. & Keating, A. (1992) The CD34 antigen: structure, biology, and potential clinical applications. Journal of Haematotherapy, 1, 115–129. Sutherland, D.R., Marsh, J.C.W., Davidson, J., Baker, M.A., Keating, A. & Mellors, A. (1992) Differential sensitivity of CD34 epitopes to cleavage by Pasteurella haemolytica glycoprotease: implications for purification of CD34-positive progenitor cells. Experimental Hematology, 20, 590–599. Sutherland, H.J., Eaves, C.J., Eaves, A.C., Dragowska, W. & Lansdorp, P.M. (1989) Characterization and partial purification of human marrow cells capable of initiating long-term hematopoiesis in vitro. Blood, 74, 1563–1570. Tarella, C., Benedetti, G., Caracciolo, D., Castellino, C., Cherasco, C., Bondesan, V., Omede, P., Ruggieri, D., Gianni, A.M. & Pileri, A. (1995) Both early and committed haemopoietic progenitors are more frequent in peripheral blood than in bone marrow during mobilization induced by high-dose chemotherapy G-CSF. British Journal of Haematology, 91, 535–543. Terstappen, L.W.M.M., Huang, S., Safford, M., Lansdorp, P.M. & Loken, M.R. (1991) Sequential generations of hematopoietic colonies derived from single nonlineage-committed CD34 CD38ÿ progenitor cells. Blood, 77, 1218–1227. Tindle, R.W., Nichols, R.A.B., Chan, I.C., Compana, D., Catovsky, D. & Birnie, G. (1984) A novel monoclonal antibody BI.3C5 recognizes myeloblasts and non-B non-T lymphoblasts in acute leukemias and CGL blast crisis and reacts with immature cells in normal bone marrow. Leukemia Research, 9, 1–10.

605

Titley, I., Healy, L.E., Scott, M., Amos, T.A.S. & Gordon, M.Y. (1995) Extent of variability inherent in measurements of CD34-positive cells in different human hematopoietic tissues. Bone Marrow Transplantation, 16, 611–616. Tjønnfjord, G.E., Steen, R. & Egeland, T. (1995) Lineage commitment of human CD34+ lymphohematopoietic progenitor cells. Experimental Hematology, in press. Tjønnfjord, G.E., Steen, R., Evensen, S.A. & Egeland, T. (1994) Characterization of CD34 peripheral blood cells from healthy adults mobilized by recombinant human granulocyte colonystimulating factor. Blood, 84, 2795–2801. To, L.B., Haylock, D.N., Dowse, T., Simmons, P.J., Trimboli, S., Ashman, L.K. & Juttner, C.A. (1994) A comparative study of the phenotype and proliferative capacity of peripheral blood (PB) CD34 cells mobilized by four different protocols and those of steady-state-phase PB and bone marrow CD34 cells. Blood, 84, 2930–2939. Traore, Y., Fornelli, C., Komarover, H. & Hirn, J. (1995) A defined pool of CD34 mAb: potential for accurate immunophenotyping of acute leukaemias. Leukocyte Typing V (ed. by S. Schlossman et al), pp. 864–867. Oxford University Press. Traycoff, C.M., Abboud, M.R., Laver, J., Brandt, J.E., Hoffman, R., Law, P., Ishizawa, L. & Srour, E.F. (1994) Evaluation of the in vitro behaviour of phenotypically defined populations of umbilical cord blood haematopoietic progenitor cells. Experimental Hematol-ogy, 22, 215–222. Watt, S.M., Karhi, K., Gatter, K., Furley, A.J.W., Katz, F.E., Healy, L.E., Altass, L.J., Bradley, N.J., Sutherland, D.R., Levinsky, R. & Greaves, M.G. (1987) Distribution and epitope analysis of the cell membrane glycoprotein (HPCA-1) associated with human haematopoietic progenitor cells. Leukaemia, 1, 417–426. Wickenhauser, C., Thiele, J., Drebber, U., Kvasnicka, H.M., Thiel, A., Schmitz, B., Hansmann, M.-L. & Fischer, R. (1995) CD34 human hematopoietic progenitor cells of the bone marrow differ from those of the peripheral blood: an immunocytochemical and morphometric study. Acta Haematologica, 93, 83–90.
