Human B-lymphocyte antigens expressed by lymphocytic and myelocytic leukemia cells. I. Detection by rabbit antisera

A previously uncharacterized human B-lymphocyte antigen has been detected by rabbit antisera raised to papain digests of spleen cell membranes. The unabsorbed sera reacted in both cytotoxicity and immunofluorescent tests with normal B lymphocytes and cultured B-cell lines but not with normal T lymphocytes or cultured T-cell lines. The cytotoxicity titers against B cells were as high as 1:32,000, whereas the same sera undiluted were negative against T cells. By immunofluorescent staining 6-14% of unfractionated normal lymphocytes and 48-85% of B-rich lymphocyte preparations were positive. Normal peripheral blood granulocytes, platelets, erythrocytes, and phytohemagglutinin blasts were negative. The antisera reacted with the same high titers against leukemia cells from approximately 70% of the patients with acute lymphocytic leukemia, acute myelocytic leukemia, chronic myelocytic leukemia, and seven of eight cases of chronic lymphocytic leukemia. From absorption studies it appeared that the same antigen was being expressed by leukemia cells and normal B lymphocytes. Using immunofluorescent staining the anti-B-cell antisera were able to detect positive leukemia cells in the bone marrow of patients with advanced leukemia and to monitor the elimination of these cells after chemotherapy. Soluble B-cell antigen was found in the serum of some leukemia and lymphoma patients do but not in normal serum.     

Poor Mixed Lymphocyte Reaction Stimulatory Capacity of Leukemic B Cells of Chronic Lymphocytic Leukemia Patients Despite the Presence of Ia Antigens

The human Ia-like antigens, selectively expressed on B lymphocytes, are now recognized to be closely associated with, or identical to, the gene products of the major histocompatibility complex responsible for stimulation in the mixed lymphocyte reaction. The leukemic B lymphocytes of patients with chronic lymphocytic leukemia express these antigens very well. In the present study they were readily detected by several techniques utilizing both allo- and heteroantisera. However, the leukemic B cells from most patients were found to be extremely poor stimulating cells in the mixed lymphocyte reaction. This was particularly apparent when comparisons were made on a B-cell basis with isolated normal B lymphocytes.Leukemic cell death, abnormal kinetics of leukemic cell-mediated stimulation, and serum or cellular suppressor factors do not appear to explain these findings. Studies comparing cells from a leukemic patient with those of her HLA identical sibling and results of mixed lymphocyte reactions between normal and leukemic subjects discordant for D-region-associated Ia antigens ruled out genetic explanations for the differences observed. Experiments with normal peripheral blood mononuclear cells depleted of T cells and monocytes exclude the quantitative deficiency of monocytes which is found in the peripheral blood of most leukemic patients as an explanation.The present results with chronic lymphocytic leukemia cells indicate that the mere expression of the Ia-like antigens by cell populations does not render them effective stimulators. The accumulated evidence obtained indicate that abnormalities, particularly of membrane function and metabolism, known to occur in chronic lymphocytic leukemia lymphocytes may be involved in the poor stimulatory capacity of the leukemic B cells.     

Combined Studies of Complement Receptor and Surface Immunoglobulin-Bearing Cells and Sheep Erythrocyte Rosette-Forming Cells in Normal and Leukemic Human Lymphocytes

Human lymphocytes from normal peripheral blood, thymus, spleen, thoracic duct, and peripheral lymphocytes from patients with chronic lymphatic leukemia were studied for complement receptor sites (CRL), surface immunoglobulin (SIg), and for the ability to form rosettes with sheep erythrocytes (TRFC). The two B cell markers (CRL and SIg) were found to be in overlapping, but not totally identical populations, whereas cells that were able to form rosettes were found in a totally unrelated population of lymphocytes; TRFC is therefore probably a reliable marker for T cells. In peripheral blood 24% of lymphocytes had SIg, but only half of these were also CRL. Almost all of the non-SIg peripheral blood lymphocytes were TRFC. In the spleen and thoracic duct only a few lymphocytes were observed that had SIg and were not CRL. On the other hand, in two of three spleens studied 10-20% of cells were CRL that did not have SIg. In the thoracic duct all non-CRL that did not have SIg. In the thoracic duct all non-CRL, non-SIg cells were TRFC. In chronic lymphatic leukemia three findings were made: (a) The presence or absence of CRL was independent of the presence or absence of SIg so that in individuals whose cells were non-SIg. CRL were usually plentiful. (b) Leukemic cells were essentially negative for TRFC. (c) Leukemic cells reacted poorly with human C3 compared to mouse C3, EACmo detecting up to 20-fold more CRL than EAChu. This latter finding was in sharp contrast to normal CRL that reacted somewhat preferentially with EAChu. These data suggest that altered surface Ig receptors and complement receptors are present in chronic lymphatic leukemic cells. Since the cells obtained from all leukemic patients tested in this study had either the complement receptor or surface immunoglobulin in a high percentage of their cells and were essentially negative for TRFC, it is strongly suggested that leukemic lymphocytes are of B cell origin. The finding of lymphocytes with only one of the two B cell markers suggests that these markers are not uniformly present on all B cells and that depending on the source, one or the other may be deficient.     

Antigen heterogeneity of human B and T lymphocytes.

Rhesus monkeys were immunized with normal human lymphoid cells, cultured lymphoid cells, and chronic leukemic lymphocytes. Antisera were analyzed by cytotoxicity and immunofluorescence techniques to study the antigenic characteristics of human lymphocytes. In an attempt to obtain a reagent specifically reactive with T (thymus-derived) lymphocytes, an antispleen antiserum was absorbed with cellf from five B- (bone marrow-derived) cell lines. After absorption, the antiserum killed 60-75% of peripheral blood lymphocytes and 40-50% of tonsil cells, so that there was a relationship between the percentage of killed cells and the proportion of T lymphocytes. However, when cells after cytotoxic treatment were assayed for rosette formation with sheep erythrocytes (a T-cell marker) 5-20% of viable rosette-forming lymphocytes were found. Therefore, this antiserum was cytotoxic for only 75-90% of T cells. From studies performed with antisera prepared against spleen and B-cell lines, we conclude that lymphoblastoid cells are antigenically different and deficient in comparison to normal B lymphocytes. In addition, cultured B-cell lines appear to be antigenically heterogenous, as shown by the cytotoxic activity remaining in antispleen and anti-B-cell lines sera after absorption with various numbers and types of lymphoid cell lines. After absorption with normal lymphocytes, an antiserum produced against chronic lymphatic leukemia cells had specific activity associated with 12 chronic lymphatic leukemia cells tested. Absorption of the same antiserum with leukemic cells from two patients showed that a certain degree of antigenic heterogeneity also exists among chronic leukemic lymphocytes.     

Multicomponent analysis of amino acid transport in human lymphocytes. Diminished L-system transport in chronic leukemic B lymphocytes

We have examined the amino acid transport in B cell chronic lymphocytic leukemia and compared it with the amino acid transport in isolated B lymphocytes from human blood and tonsils. L-system transport was measured with 2-amino-2-carboxy-bicyclo (2,2,1)-heptane, which is a synthetic amino acid whose transport is limited to the L-system. Amino acid uptake was subjected to a multicomponent analysis that partitioned the total uptake into the saturable carrier-mediated transport system and the uptake by diffusion. The maximal velocity of L-system transport in chronic lymphocytic leukemia cells, 81 mumol/1 cell water per min, was less than 10% that of blood B lymphocytes, which was 1,029 mumol/1 cell water per min. The uptake of 2-amino-2-carboxy-bicyclo (2,2,1)-heptane by tonsillar B cells, by a B lymphocyte cell line, and by blood T-lymphocytes was also 10-fold greater than that observed in chronic lymphocytic leukemic cells. Similarly, the L-system uptake of leucine and phenylalanine, which are naturally occurring amino acids usually transported primarily by the L-system, was reduced in chronic lymphocytic leukemic B cells to 15 and 10% of normal B cells, respectively. Total leucine uptake by chronic lymphocytic leukemic cells, however, was sustained at 30% of that expected because of transport via an alternative transport system. The A- or ASC-systems, the other major amino acid transport pathways, were not defective in chronic lymphocytic leukemic cells. These data indicate that there is a specific, profound decrease in L-system carrier-mediated amino acid transport in chronic lymphocytic leukemic B cells, as judged by the system-specific synthetic amino acid, 2-amino-2-carboxy-bicyclo (2,2,1)-heptane. This defect was confirmed by studies with two naturally occurring L-system amino acids, leucine and phenylalanine. This specific abnormality of membrane transport by chronic lymphocytic leukemic B lymphocytes is not shared by other B lymphocyte types, and thus appears to be related to the neoplastic nature of the leukemic B cells rather than to their immunologic subtype.     

Origin of human T-lymphotrophic virus I-positive T cell lines in adult T cell leukemia. Analysis of T cell receptor gene rearrangement

Using the clone-specific rearrangement of the T cell receptor gene as the genetic marker of the clonotype, we analyzed the clonal origin of the interleukin 2 (IL-2)-dependent human T-lymphotrophic virus I (HTLV-I)-positive T cell lines established from various adult T cell leukemia (ATL) patients. From a patient with chronic ATL, whose leukemic cells proliferated in vitro in response to IL-2, we repeatedly established leukemic T cell clones having the same rearrangement profile of the T beta chain gene as the leukemic cells. By contrast, established cell lines from acute ATL patients had different beta chain gene rearrangements from those of the leukemic cells. These HTLV-I+ T cell lines might not be the direct progeny of the leukemic cells, but that of T cells infected either in vivo or in vitro. These IL-2-reactive nonleukemic T cells might have been selected in vitro, because their leukemic cells failed to respond to IL-2, despite the expression of IL-2 receptor. The analysis of the T cell receptor gene rearrangement may give a new approach for the elucidation of the mechanism of leukemogenesis and the origin of the HTLV-I+ T cell lines in ATL.     

Chronic myelogenous leukemia: in vitro studies of hematopoietic regulation in a patient undergoing intensive chemotherapy.

A patient heterozygous for the X-linked enzyme glucose-6-phosphate dehydrogenase and with Philadelphia chromosome-positive chronic myelogenous leukemia (CML) was treated with combination chemotherapy and had a partial loss of Philadelphia chromosome accompanied by partial restoration of nonclonal hematopoiesis as determined by glucose-6-phosphate dehydrogenase. Studies of in vitro hematopoiesis were performed after chemotherapy to evaluate the influences of neoplastic stem cells on normal cells and to determine whether there were physical and cell kinetic differences between leukemic stem cells and their normal counterparts. The data revealed the following: (a) The frequencies of normal committed granulocytic stem cells (CFU-C) and erythroid stem cells (BFU-E) in blood did not differ from the frequencies in marrow. (b) Normal late erythroid progenitors (CFU-E) were found at a significantly lower frequency that the more primitive BFU-E. Calculations indicated that not only was there a decrease in CFU-E production by normal BFU-E, but there was also abnormal clonal expansion of CML BFU-E (CFU-E:BFU-E ratio for normal progenitors was 1.1, whereas for the CML clone it was 11.5). (c) No increase in frequency of normal CFU-C was found after marrow cells were exposed to high specific activity tritiated thymidine. (d) Normal CFU-C and those from the CML clone were not separable on the basis of density. (e) The frequency of normal BFU-E was consistently greater than that of CFU-C, suggesting that regulatory differences influence the commitment of normal progenitors to the two pathways.     

Phorbol ester restores L-system amino acid transport of B lymphocytes in chronic lymphocytic leukemia.

L (leucine-favoring)-system amino acid transport is uniquely and selectively diminished in chronic lymphocytic leukemia B lymphocytes: the maximal velocity of transport is 10% of normal B lymphocytes. We examined L-system transport in chronic leukemic B lymphocytes after incubation with tetradecanoyl phorbol acetate to determine if the transport abnormality can be corrected by the apparent cell maturation induced by this agent. Amino acid uptake was measured using 2-amino-2-carboxy-bicycloheptane, an L-system specific synthetic amino acid. Marked enhancement of L-system transport occurred in each of 12 leukemic cell populations; the initial velocity of transport in phorbol ester-treated cells increased 8-fold and 14-fold at 16 and 40 h, respectively, compared with untreated cells. The Vmax of the L-system in phorbol ester-treated leukemic cells was similar to that of phorbol ester-treated normal B lymphocytes. The L-system enhancement of the leukemic cells paralleled the development of plasmacytoid features at 40 h. Uptake of leucine, a naturally occurring L-system amino acid, was also increased by tetradecanoyl phorbol acetate. Cycloheximide, 100 micrograms/ml, which inhibited over 90% of protein synthesis in phorbol ester-treated chronic leukemic cells, blocked completely the phorbol ester-induced L-system enhancement. Phorbol ester treatment restores the selective L-system transport defect in chronic lymphocytic leukemia B lymphocytes, and this process coincides with in vitro maturation of the leukemic cells.     

Biochemical activities of lysosomal acid hydrolases in leukemic cells

The biochemical activities of 8 lysosomal acid hydrolases in leukemic cells from 48 patients were examined. Characteristic alterations were found in α-mannosidase, β-galactosidase and N-acetyl-β-glucosaminidase activities of leukemic cells. The level of α-mannosidase activity was much higher in myelo(mono)genous leukemias (AML, AMoL, AMMoL, CML and CMMoL) than in lymphogenous ones (ALL, T-cell leukemia, hairy cell leukemia and CLL) without exception. The β-galactosidase activity also differed as a result of α-mannosidase, except in T-cell leukemia. In T-cell leukemia it was within the range of normal lymphocytes, but in the other lymphogenous leukemias it was significantly below normal. N-acetyl-β-glucosaminidase activity in myelo(mono)genous leukemic cells was above the range of normal granulocytes. The changes in these enzyme levels were consistent. The lymphocytic or myelocytic nature of three cases of acute undifferentiated leukemia could be determined by enzyme studies. In two cases it was lymphocytic and in one it was myelocytic. The enzymatic abnormalities were also found in morphologically mature neutrophils from patients with not only chronic types (CML, CMMoL) but also acute types (AMoL, AMMoL) of leukemias, and were similar to those of their respective leukemic cells. Analysis of lysosomal enzymes (at least three of those mentioned above), can elucidate one of the biochemical properties of leukemic cells and may be valuable in the differentiation of leukemias.     

Flow cytometric detection of recombinant human granulocyte-colony stimulating factor binding to leukemic cells

To examine binding of recombinant human granulocyte-colony stimulating factor to myeloid cells, the factor was labeled with fluorescein isothiocyanate, and incubated with blood specimens, which were then analyzed by flow cytometry. Neutrophils demonstrated an increased fluorescence, while lymphocytes were negative. These cell fractions were used as controls for cytometric binding assays of leukemic cells. Six patients with lymphocytic leukemia were negative in this assay. Ten of 15 patients with myelocytic leukemia were positive. All patients (n = 5) in myeloblastic crisis of chronic myelogenous leukemia were also positive. The flow cytometry results correlated well with the results of colony formation in response to granulocyte-colony stimulating factor. The results indicate that our method is useful in predicting the susceptibility of leukemic cells to recombinant growth factors. © 1993 Wiley-Liss, Inc.     

Humoral regulation of human leukemic progenitor cell cycle: the case for turning off the cell cycle

Colony-forming cells of chronic myelocytic leukemia as well as acute nonlymphocytic leukemia in the resting state can be turned on and mobilized into cell cycle with exposure to leukocyte conditioned medium. In addition, evidence indicates that the leukemic colony-forming cells in cell cycle can be turned off and mobilized out of cell cycle by withholding the leukocyte conditioned medium. The system used in this study will be invaluable in biologic studies of leukemic cell cycling and its manipulation. With the concept of leukemic cell-cycle manipulation established, new approaches to treatment of leukemia based on growth manipulation should be explored.     

SHEEP RED CELL BINDING TO HUMAN LYMPHOCYTES TREATED WITH NEURAMINIDASE; ENHANCEMENT OF T CELL BINDING AND IDENTIFICATION OF A SUBPOPULATION OF B CELLS

Neuraminidase treatment of normal human lymphocytes enhances their capacity to form SRBC rosettes; more red cells are bound and the rosettes are more stable. Under these conditions approximately 90% of peripheral blood lymphocytes form rosettes. In addition to the effects on T cells, another population of lymphocytes which possess surface immunoglobulin and have the Fc receptor acquire the rosette-forming property after neuraminidase. This subpopulation of ‘B’ cells represents approximately half of the lymphocytes with surface immunoglobulin but is not found among the leukemic lymphocytes of patients with chronic lymphocytic leukemia. Electron microscope observations indicate close approximation and intimate association of the red cell and lymphocyte membranes after neuraminidase.     

Immunoglobulin mu-chain gene rearrangement in a patient with T cell acute lymphoblastic leukemia

A 6-yr-old girl with T cell acute lymphoblastic leukemia (ALL) is described. She had a mediastinal mass and her leukemic cells expressed T cell-associated antigens (Leu 1+, OKT3+, OKT9+, and OKT10+). When we examined genomic DNA from the leukemic cells, we detected Ig mu-chain gene rearrangement with germ-line configuration of light chain genes. As reported recently, detecting Ig gene rearrangement has become an important procedure for further classifying B cell precursor cells. This case, however, suggests that there is also heterogeneity among patients with T cell ALL, not only at the level of cell surface phenotypes, but also at the level of the Ig gene. These findings have major implications when we consider both the ontogenesis of these leukemic cells and the normal differentiation of human lymphocytes.     

A unique cell surface antigen identifying lymphoid malignancies of B cell origin.

A monoclonal antibody (anti-B1) specific for a unique B cell surface differentiation antigen was used to characterize the malignant cells from patients with leukemias or lymphomas. All tumor cells from patients with lymphomas or chronic lymphocytic leukemias, bearing either monoclonal kappa lambda light chain, expressed the B1 antigen. In contrast, tumor cells from T cell leukemias and lymphomas or acute myeloblastic leukemia were unreactive. Approximately 50% of acute lymphoblastic leukemias (ALL) of non-T origin and 50% of chronic myelocytic leukemia in blast crisis were also anti-B1 reactive. moreover, 21 of 28 patients with the common ALL antigen (CALLA) positive form of ALL were anti-B1 positive, whereas 0 of 13 patients with CALLA negative ALL were reactive. These observations demonstrate that an antigen present on normal B cells is expressed on the vast majority of B cell lymphomas and on approximately 75% of CALLA positive ALL, suggesting that these tumors may share a common B cell lineage.     

Studies on N5-methyltetrahydrofolate-homocystein methyltransferase in normal and leukemia leukocytes.

A cobalamin-dependent N5-methyltetra-hydrofolate-homocysteine methyltransferase (methyl-transferase) was demonstrated in unfractioned extracts of human normal and leukemia leukocytes. Activity was substantially reduced in the absence of an added cobalamin derivative. Presumably, this residual activity reflects the endogeneous level of holoenzyme. Enzyme activity was notably higher in lymphoid cells than in myeloid cells. Thus, mean specific activities (+/-SD) were: chronic lymphocytic leukemia lymphocytes, 2.15+/-1.16; normal lymphocytes, 0.91+/-0.59; normal mature granulocytes, 0.15+/-0.10; chronic myelocytic leukemia granulocytes, barely detectable activity. Properties of leukocytes enzymes resembled those of methyltransferases previously studied in bacteria and other animal cells. Granulocytes and chronic myelocytic leukemia cells contain a factor or factors that inhibits Escherichia coli enzyme. The data suggest that the prominence of this cobalamin-dependent enzyme in lymphocytes and other mononuclear cell types may be related to their potential for cell division.     

Chronic lymphocytic leukemia B cells contain anomalous Lyn tyrosine kinase, a putative contribution to defective apoptosis

B cell chronic lymphocytic leukemia (B-CLL) is a neoplastic disorder characterized by accumulation of B lymphocytes due to uncontrolled growth and resistance to apoptosis. Analysis of B cells freshly isolated from 40 patients with chronic lymphocytic leukemia demonstrated that the Src kinase Lyn, the switch molecule that couples the B cell receptor to downstream signaling, displays anomalous properties. Lyn is remarkably overexpressed at the protein level in leukemic cells as compared with normal B lymphocytes, with a substantial aliquot of the kinase anomalously present in the cytosol. Whereas in normal B lymphocytes Lyn activation is dependent on B cell-receptor stimulation, in resting malignant cells, the constitutive activity of the kinase accounts for high basal protein tyrosine phosphorylation and low responsiveness to IgM ligation. Addition of the Lyn inhibitors PP2 and SU6656 to leukemic cell cultures restores cell apoptosis, and treatment of malignant cells with drugs that induce cell apoptosis decreases both activity and amount of the tyrosine kinase. These findings suggest a direct correlation between high basal Lyn activity and defects in the induction of apoptosis in leukemic cells. They also support a critical role for Lyn in B-CLL pathogenesis and identify this tyrosine kinase as a potential therapeutic target.     

T cells and probably B cells arise from the malignant clone in chronic myelogenous leukemia.

Bone marrow cells from a patient with Ph' positive chronic myelogenous leukemia in chronic phase were cultured for multilineage hematopoietic colonies (CFU-GEMMT), erythroid bursts, and granulocytic colonies. With CFU-GEMMT colonies, T lymphocytes were identified by reaction with monoclonal antibodies Leu-5 and OKT-3; B cells were identified by reaction with B1. All CFU-GEMMT colonies examined contained the Ph' chromosome. Recloned secondary colonies of T cells reacted with Leu-5 and OKT-3 and were Ph' positive. This demonstrates that Ph' positive T lymphocytes were generated from the pluripotential stem cell of this patient. The presence of B cells in the mixed colonies indicates that these may also be derived from the neoplastic clone.     

Identification of a p69,71 complex expressed on human T cells sharing determinants with B-type chronic lymphatic leukemic cells

In the course of generating monoclonal antibodies to human thymus-dependent differentiation antigens, we were able to define specificities shared by T cells and by cells from patients with chronic lymphatic leukemia that were not detectable on normal B cells. In particular, one of these antibodies was reactive by indirect immunofluorescence with greater than 95% of the thymocytes and 80--95% of nonadherent sheep erythrocyte-rosetting peripheral blood lymphocytes (PBL), but was unreactive with normal B cells or cell lines derived from PBL by Epstein-Barr virus transformation. However, the leukemic cells from 11 of 14 patients with B-type chronic lymphatic leukemia were found to express detectable concentrations of this surface determinant. The target antigen recognized by this monoclonal antibody was shown by immunoprecipitation and sodium dodecyl sulfate-polyacrylamide gel electrophoresis to be a p69,71 complex. Our findings suggest a possible relationship between this antigen and the previously described GIX system in the mouse.     

Receptors for complement and immunoglobulin on human leukemic cells and human lymphoblastoid cell lines

The bone marrow-derived (B) lymphocyte can be identified by the presence of easily detectable surface immunoglobulin and a receptor for antigen-antibody-complement complexes (EAC'). Monocytes and macrophages also bear a receptor for EAC' and in addition possess a receptor for red cell-IgG complexes (EA). Thymus-derived (T) lymphocytes bear neither of these receptors. The cells of 15 patients with leukemia and 19 human lymphoblastoid cell lines were examined for the presence of the EAC' and EA receptors. Of the human leukemias studied, only the cells from the patients with chronic lymphatic leukemia (CLL) possess the EAC' receptor. The EA receptor could not be demonstrated on CLL cells; hence, CLL cells bear the lymphocyte EAC' receptor and by this criteria represent B lymphocytes. 12/19 of the cell lines studied could be classified as B lymphocytes by the presence of the EAC' receptor and absence of the EA receptor. 2/19 cell lines possessed both the EAC' and EA receptors and thus resemble the monocyte. 5/19 cell lines had no detectable receptor for EAC' or EA. The approach presented in this study for the classification of leukemias and cell lines as to their B lymphocyte, T lymphocyte, or monocyte origin may have useful diagnostic and therapeutic implications.     

Circulating levels of soluble CD23 reflect clinical and biological features of leukemic B-cell chronic lymphoproliferative disorders

Summary One hundred and twenty-four sera, from patients with various leukemic B-cell chronic lymphoproliferative diseases were investigated at diagnosis by ELISA for their soluble CD23 content. Immunophenotyping was carried out in all patients, and in a selected subset the mean number of membrane-bound CD23 molecules per cell was also investigated. Seventy-three patients had typical B chronic lymphocytic leukenia, 41 leukemic B-cell disorders with atypical morphological and/or immunophenotypic features, 5 had low-grade follicular cell lymphoma in the leukemic phase, and 5 had splenic lymphoma with villous lymphocytes Soluble CD23 levels were significantly higher than in normal sera (mean±SD: typical B chronic lymphocytic leukemia 3,650±4,654 U/ml, atypical B chronic lymphocytic leukemia 3,440±4,671 U/ml, follicular cell lymphoma 3,200±1,511 U/ml, splenic lymphoma with villous lymphocytes 8,236±7,294 U/ml, controls 137±128 U/ml;P<0.001). More advanced Rai's stages were related to higher soluble CD23 levels (P<0.01), both in typical and atypical B chronic lymphocytic leukemias, the highest levels and the best correlation with the absolute number of circulating CD19+ cells (r=0.50) being observed in the typical form. The number of membrane-bound CD23 molecules per cell was significantly higher in typical than in atypical B chronic lymphocytic leukemias (mean number 156,727±94,668 vs. 12,010±10,643,P<0.001). Our data suggest that soluble CD23 levels correlate with the clinical and biological features of leukemic B-cell lymphoproliferative disorders.