- Top of page
- MATERIALS AND METHODS
Urokinase-type plasminogen activator receptor (UPA-R-CD87) is a GPI-anchored membrane protein which promotes the generation of plasmin on the surface of many cell types, probably facilitating cellular extravasation and tissue invasion. A flow cytometric quantitative analysis of expression levels for UPA-R was performed on fresh blast cells from patients with acute myeloid leukaemia (AML, n = 74), acute lymphoblastic leukaemia (ALL, n = 24), and biphenotypic leukaemia (BAL, n = 3) using two CD87 monoclonal antibodies (McAbs) (3B10 and VIM5). Peripheral blood and bone marrow (BM) cells from 15 healthy adults served as controls. Using 3B10 McAb, UPA-R was expressed (>99%) by blood monocytes, neutrophils, and BM myelomonocytic precursors in controls, whereas resting T and B lymphocytes, and CD34+ cells were UPA-R negative. We also attempted to clarify whether UPA-R has a role in mediating neutrophil functions. Oriented locomotion induced by different chemotaxins and lysozyme release by granules stimulated with fMLP or PMA were significantly decreased when UPA-R was neutralized by CD87 McAb. In contrast, the anti-UPA-R McAb had no effect on superoxide anion generation of normal neutrophils. Blasts from AML showed a heterogenous pattern of expression for the UPA-R McAbs, with reactivity strictly dependent on FAB subtype. The highest UPA-R expression was seen in the M5 group: all patients tested (n = 20) showed strong positivity for the UPA-R McAb whereas only 12% (3/24) of ALL patients were CD87 positive, and 2/3 of BAL patients showed a dim expression for CD87. The number of receptors expressed by blast cells in 6/74 (8.1%) AML patients was higher than those of normal samples; in addition, since co-expression of UPA-R and CD34 was not found in normal haemopoietic cells, it may be postulated that CD87 can be used alone (when overexpressed) or in combination with CD34 for the detection of minimal residual disease. Results also indicated that patients with UPA-receptors >12 × 103 ABC/cell, irrespective of FAB subtype, had a greater tendency for cutaneous and tissue infiltration and a higher frequency of chromosome abnormalities, thus suggesting the concept that cellular UPA-R content positively correlates with the invasive potential of AML cells. The combination of higher UPA-R positivity, abnormalities of chromosome 11, and M5 FAB morphology may identify a peculiar subset of AML, characterized by a more aggressive clinical course.
Urokinase is a specific serine protease which catalyses plasminogen conversion to plasmin principally within the extravascular compartments. Plasmin is a broad-spectrum protease which is able to degrade protein constituents of the extracellular matrix and adhesion molecules in the pericellular space. As a consequence, these enzymes are thought to contribute to the invasive properties and metastasis formation of tumour cells ( Bu et al, 1994 ; Stephens et al, 1996; Lijnen, 1996). Urokinase plasminogen activator (UPA)-mediated cell surface plasminogen activation occurs on a specific membrane receptor that can be found in many cell types including monocytes, granulocytes, a subset of activated natural killer cells, endothelial cells, fibroblasts, smooth muscle cells, keratocytes, hepatocytes and placental trophoblasts, as well as in several tumour cells such as fibrosarcoma, rhabdomyosarcoma, melanoma, breast, prostate and colon carcinoma, and the U937 leukaemia cell line ( Pyke et al, 1991 ; Todd et al, 1995 ; Lanza et al, 1994 ; Sitrin et al, 1994 ). The UPA-receptor (UPA-R) is a specific and saturable receptor, composed of a single-chain, highly glycosylated protein with a molecular weight of 55–60 kD. CD87 serves as a receptor for pro-UPA and UPA with high affinity (Kd 10−9–10−10 molar) ( Todd et al, 1995 ). The protein is linked to the plasma membrane via a glycosylphosphatidylinositol (GPI) anchor and thus has no intracytoplasmic domains; as a consequence, signal transduction by UPA-R may require an adaptor protein connected to the cell surface. Since UPA-R is spatially and temporally associated with cellular structures that regulate cell adhesion and migration, it has been hypothesized that formation of complexes between UPA-R and integrins such as CD11b (Mac-1) might provide an integrin mediated link between UPA-R and the cytoskeleton, thus promoting adhesion ( Gyetko et al, 1994 , 1995, 1996; Xue et al, 1994 ; Plesner et al, 1994 ; Cao et al, 1995 ; Kindezelskii et al, 1996 ; Simon et al, 1996 ). When UPA is complexed to its specific inhibitors PAI-1 (plasminogen activation inhibitor-1) and PN-1, it is internalized and degraded through a mechanism requiring both UPA-R and the α2 macroglobulin receptor. CD87 may also form a functional linkage with β1, β2, β3 integrins or with other proteins that have tyrosine kinase activity ( Wei et al, 1996 ).
Recently, UPA-R cDNA has been cloned and the primary structure of the protein defined; it consists of 313 amino acid residues organized in three repeating sequences of approximately 90 amino acids ( Todd et al, 1995 ). The human CD87 gene has been mapped to chromosome 19q1.3.
At the Fifth International Workshop on Leukocyte Differentiation Antigens, held in Boston, U.S.A., on 3–7 November 1993, six monoclonal antibodies (McAbs) were verified as recognizing the UPA-R, and were referred to as CD87 (CD stands for cluster of differentiation) ( Todd et al, 1995 ; Lanza et al, 1994 ). Four further CD87 McAbs were assigned in 1996 at the Sixth International Workshop on Leukocyte Differentiation Antigens.
The aim of this study was to investigate the cellular expression and functional role of UPA-R in haemopoietic cells obtained from healthy adult subjects; in addition, UPA-R expression was evaluated by flow cytometry in acute leukaemic blasts, and the results were tentatively correlated with clinical and biological features of the patients examined.
- Top of page
- MATERIALS AND METHODS
The cellular expression and clinical significance of the UPA receptor was investigated in acute leukaemic blasts and normal haemopoietic cells through the use of a flow cytometer and appropriate standards, which enabled a reliable measurement of antibody binding capacity and MESF values in the various cell types.
In previous studies it has been shown that UPA-R may be involved in cellular extravasion and in metastasis formation ( Cao et al, 1995 ; Todd & Petty, 1997). In leukaemia, overexpression of UPA-R and/or other components of this system was found to consume plasma inhibitors, producing a tendency to haemorrhage ( Todd et al, 1995 ; Lijnen, 1996; Bennett et al, 1997 ). The evidence that UPA-R and its ligand UPA are involved in the metastatic potential of tumour cells derives from a wide range of observations, including (i) a close relationship between oncogenic transformation and UPA synthesis, and between UPA and cellular invasion as well as metastasis in several model systems (melanoma, neuroblastoma, breast, colon and prostate carcinoma, etc.); (ii) the inhibition of these processes by anticatalytic McAbs to UPA; and (iii) the histochemical localization of UPA at the invasive fronts of tumours ( Bu et al, 1994 ; Lijnen, 1996; Todd et al, 1995 ). However, little is known about the cellular expression and clinical significance of the UPA-receptor in leukaemic cells from patients with acute leukaemia ( Knapp et al, 1994 ).
In this study we have demonstrated that both neutrophil granulocytes and monocytes taken from the peripheral blood of healthy subjects had detectable levels of UPA-R, whereas resting B and T lymphocytes lacked CD87 in all the cases examined. In healthy subjects, reactivity for UPA-R McAb was also detected on bone marrow CD34-negative myelomonocytic precursors (promyelocytes, myelocytes, metamyelocytes and promonocytes), whereas BM and mobilized peripheral blood CD34+ progenitors were UPA-R negative, supporting the concept that UPA-R expression is a stage-specific feature of late myelomonocytic cells. The lack of UPA-R in a significant number of bone marrow neutrophil granulocytes may be explained by the peculiar homing characteristics of this cell subset, possibly correlated to an unfinished degranulation of surface proteins.
In a second set of experiments we attempted to confirm whether UPA-R had a role in mediating neutrophil functions ( Gyetko et al, 1994 , 1995). Oriented locomotion induced by different chemotaxins, such as fMLP and denatured casein, was significatively diminished when UPA-R was neutralized by McAbs. Similarly, lysozyme release by granules stimulated with fMLP or PMA was markedly blocked. The inhibitory effect does not depend on either the presence of membrane agonist receptor or the intracellular pathway activated by chemotaxins. In fact, although fMLP possesses at least two receptor isoforms ( Ye & Boulay, 1997) (each one able to induce a specific biological response), denatured peptides or proteins, such as casein, could recognize neutrophil surface by aspecific hydrophobic interactions. These results, however, further support those of Gyetko et al (1994 ), indicating that UPA-R has a structural and functional role in the cytoskeleton reorganization required for cell activation.
The observation that UPA-R, in our hands, seemed not to be involved in NADPH oxidase activation was not in contrast with the findings of Cao et al (1995 ) who showed the priming effect of UPA of fMLP-triggered superoxide generation, since in our assay conditions, cytochalasin B was added to potentiate the ability of fMLP to produce superoxide anion, and probably the priming effect mediated by UPA-R ligands was dimmed. Moreover, the amount of O−2 produced was determined at 5 min, and not at 30 min, when stimulated with fMLP. This observation, together with the lack of UPA-R involvement in NADPH activation, may exclude a receptor–receptor association able to block the second messenger pathway, but may support the hypothesis that UPA-R is involved in the cytoskeletal reorganization required for cell movements, namely locomotion and degranulation ( Gyetko et al, 1994 ).
As far as the expression of UPA-R in blast cells is concerned, we found that 81% of AML patients showed reactivity for the anti-UPA-R 3B10 McAb, but only 67% for VIM5 McAb. The number of positive cases and the level of expression was significantly higher in AML patients belonging to the FAB M5 subgroup, thus supporting the concept that a close association between monocytic commitment and expression of UPA-R may exist. Experiments of two- and three-colour fluorescence further showed that FAB M1–M3 AML were characterized by low-intermediate expression for UPA-R and lysozyme, and bright-intermediate expression for MPO, whereas the M5 FAB subvariety had high-intermediate levels of UPA-R and lysozyme, and little or no expression of MPO. On the basis of these data, it can be postulated that the combined use of UPA-R, MPO and lysozyme McAbs may provide useful information for the diagnosis of acute leukaemia, and for the distinction of myeloid from monocyte leukaemias ( Knapp et al, 1994 ; Lanza et al, 1997b ).
Interestingly, since the co-expression of UPA-R and CD34 was not found in normal haemopoietic cells, this marker combination could be used for detecting leukaemic cells in remission bone marrows and in peripheral blood leukaphersis products. Furthermore, the number of UPA-R expressed by leukaemic cells from 8.1% of AML patients was above the highest values seen in normal samples, making it a feasible marker, even alone, for monitoring the minimal residual disease (MRD) in a subset of AML. Concerning the expression of these markers in AML with minimal evidence of myeloid differentiation, we found that four out of seven cases of FAB M0 AML were positive for the UPA-R McAb. Only one of them showed reactivity for the anti-MPO McAb, and none for lysozyme, thus indicating that CD87 may have a role in the recognition of FAB M0 AML.
We have also shown that the great majority of leukaemic blasts from AML patients had intracellular stores for UPA-R, as demonstrated by the comparison of surface expression evaluated on unfixed cells and total antigen content in permeabilized cells. This finding is in agreement with the observation that, in resting neutrophils, UPA-R are stored in three distinct intracellular compartments (secretory vesicles, specific granules and gelatinase granules) ( Plesner et al, 1994 ; Borregaard & Cowland, 1997). Functional studies have also shown that, after activation, the intracellular pools of UPA-R may translocate to the cell surface ( Plesner et al, 1994 ; Borregaard & Cowland, 1997). Our own studies provide evidence that both neutrophils and monocytes contain an intracellular pool of UPA-R detectable by flow cytometry on permeabilized cells.
Another interesting finding, which may have clinical implications, is represented by the marked increase in the number of UPA-R receptors in relapsed disease after induction-consolidation chemotherapy, thus indicating that the quantitative analysis of this receptor may be of value for the monitoring of AML patients.
No correlation was found between number of UPA-R, kinetic status and DNA content, whereas the frequency of chromosome abnormalities was much higher in AML cells overexpressing UPA-R than in AML blasts having little or no expression of UPA-R, indicating that the expression of this protease may increase the aggressiveness of the disease.
In contrast, only 12% (3/24) of ALL patients were CD87 positive; however, the degree of positivity was considerably weaker in ALL blasts than in AML cells, allowing us to speculate that this molecule may help in the distinction of acute myeloid leukaemias from lymphoid malignancies. However, the specificity of this marker for the myeloid commitment of the blasts needs to be confirmed using larger numbers of acute leukaemia patients. Furthermore, two out of three byphenotypic leukaemias and one out of three cases of AML with minimal phenotypic deviation showed positivity for CD87. The diagnostic usefulness of UPA-R McAbs in the classifi-cation of byphenotypic and undifferentiated acute leukaemias deserves careful evaluation in prospective future studies.
A close analysis of our data further shows a certain variability in the reactivity of leukaemic cells with the two UPA-R McAbs. The significance of this discrepancy is uncertain; nevertheless, it is possible to speculate that this group of McAbs recognizes a distinct epitope of the molecule ( Todd et al, 1995 ). Competitive binding assays and extensive immunoprecipitation studies have confirmed the recognition of different epitopes in close proximity of, or within, the UPA binding domain of the urokinase receptor ( Garni-Wagner & Todd, 1995); the biological relevance of such a finding is still under evaluation.
A possible correlation between UPA-R expression and the clinical features of acute leukaemias was observed in patients having a number of UPA-receptors superior to 12 × 103ABC/cell who, irrespective of the FAB category, showed a greater tendency to cutaneous and tissue infiltration, together with a higher leucocyte count, as compared to AML patients whose blasts were UPA-R negative or dimly positive. Since proteolytic enzymes, such as UPA, play a key role in the dissolution of the extracellular matrix and in facilitating the cell egress from the bone marrow, it is conceivable that the expression of the UPA-R could contribute to the invasive properties and, possibly, the metastatic potential of leukaemic cells in a subset of AML. Further evidence in favour of the role of UPA-R in facilitating leukaemic cell extravasation from the bone marrow into the circulation derives from the demonstration that both the percentage of positivity and the surface antigen density of UPA-R was significantly higher in peripheral blood AML cells than in bone marrow blasts.
Furthermore, the bright expression of UPA-R on AML blasts was associated with a high incidence of chromosome abnormalities, most of which were complex. Interestingly, within M5 AML patients, who were cytogenetically characterized by either the t(9;11) or 11q− abnormalities, two groups could be recognized, i.e. those who showed a strong expression of UPA-R and those whose blast cells were dimly positive for UPA-R. However, since in the M5 FAB group, patients carrying 11q− abnormalities have a variable clinical outcome, It can be speculated that the combination of higher UPA-R positivity, abnormalities of chromosome 11 and M5 FAB morphology may identify a peculiar subset of AML characterized by distinct morpho-immunological and cytogenetic features, and a more aggressive clinical course. Moreover, although the incidence of complex karyotypes was significantly higher in the AML patient group with UPA-Rbright phenotype in comparison with that of UPA-Rneg/dim AML, no correlation was found between FAB category and the occurrence of complex chromosome aberrations.
In conclusion, cellular UPA-R expression was variable, ranged from undetectable (ALL and a minority of AML) to very high intensities in M5 AML, and was also documented in >50% M0 AML, thus supporting the concept that UPA-R may be considered a specific and rather sensitive marker to detect leukaemic cells committed to the myeloid lineage. As a consequence, the inclusion of this marker for the immunophenotype diagnosis of acute leukaemia is recommended, even if a closer analysis of its expression in a large prospective study is needed before drawing any definitive conclusion on this matter.