Specific Lectin Binding Research Articles

Differential expression of cell surface sialoglycoconjugates on wild-type and cultured Ehrlich tumor cells as revealed by quantitative lectin-gold ultrastructural cytochemistry.

Three variants of the classical Ehrlich ascites tumor (EAT) cell have been studied by quantitative, sialic acid-specific, lectin-gold ultrastructural cytochemistry. Electron microscopic examination revealed pronounced differences in the surface morphology of the three cell variants. The wild-type Ehrlich cells (EAT-wt), grown in the peritoneal cavity of mice, exhibited a smooth surface profile. A variant form selected for growth as monolayer on basement membrane (EAT-c) showed a complex surface profile with numerous microvilli. The third variant (EAT-c/m), the cultured cells reinoculated into mice and passaged 20-25 times as ascites, presented a smooth surface profile similar to the EAT-wt cells. Quantitative single as well as double lectin-gold labeling revealed significant differences in the nature of cell surface sialoglycoproteins. The most significant finding was the presence of cell surface Neu5Ac alpha 2-6Gal residues as detected with the Sambucus nigra lectin on EAT-c and EAT-c/m cells, whereas EAT-wt cells contained little or none of such carbohydrate sequences. On the contrary, labeling by Maackia amurensis lectin, which recognizes the Neu5Ac alpha 2-3Gal beta 1-4GlcNAc sequence, was intense on all three Ehrlich cell variants; it was 20-60 times greater than alpha-2,6-linked sialic acid-containing glycoconjugates. Specific cell surface lectin binding combined with morphologic study appears to have identified a small subpopulation of cells within the ascites tumor that are capable of attaching to and growing on a basement membrane.

Bauhinia Purpurea (BPA) Binding to Normal and Neoplastic Thyroid Glands

Biotinylated bauhinia purpurea agglutinin (BPA) was studied, using the ABC method, in different thyroid gland conditions (26 follicular adenomas, 39 papillary carcinomas, 15 follicular carcinomas, and 10 normal thyroids), to determine whether specific lectin binding patterns were developed during malignant transformation that could enable the distinction of carcinoma from adenoma. In normal thyroids, BPA was very rarely and faintly reactive with follicular cells. In neoplastic conditions, BPA binding profiles for follicular adenoma were essentially identical to those of normal follicles, whereas BPA unequivocally reacted with follicular carcinoma and papillary carcinoma. BPA reacted more strongly with cells of papillary structures than with those forming solid nests and follicles. In papillary carcinoma, BPA binding was observed mostly in the apical surface and cytoplasm of carcinoma cells, whereas a diffuse cytoplasmic binding pattern was predominant in follicular carcinoma. Neuraminidase treatment had little or no effect on either normal or adenomatous follicular epithelium, whereas in follicular carcinoma, the number of positive cells and the staining intensity was increased. These findings suggest that BPA would be useful for the differential diagnosis of papillary and follicular carcinomas, and less consistently so, for differentiating follicular carcinoma from follicular adenoma. Peanut agglutinin with similar sugar specificity was not reactive with follicular cells either in normal or neoplastic glands.

The potential use of tomato lectin for oral drug delivery. 1. Lectin binding to rat small intestine in vitro

Tomato lectin, an apparently non-toxic dietary glycoprotein with a molecular mass of 71 kDa, was isolated from ripe tomato fruits. Its potential as an intestinal bioadhesive was examined using an in vitro organ culture system and Western blotting. The lectin was shown to bind avidly to rings of small intestine and this binding could be inhibited with specific competing sugars. After electrophoretic separation and transfer onto nitrocellulose, several of the proteins located on the small intestinal brush border membranes were found to bind tomato lectin, particularly those in the 160-116 kDa molecular mass range. These correspond to intestinal brush border digestive enzymes, which are known to be complex glycoproteins. The specific binding of tomato lectin to the enterocyte cell surface may give it potential for increasing the gastrointestinal residence time as a component of an oral drug formulation.

Native and sialic acid masked Lewis(a) antigen reactivity in medullary thyroid carcinoma. Distinct tumour-associated and prognostic relevant antigens.

Forty-six medullary thyroid carcinomas (MTC) were subjected to a qualitative and quantitative characterization of native and sialic acid masked Lewis(a) (Le(a)) antigens. Immunohistochemical investigations included monoclonal antibodies (MABs) directed against alpha(2,3)-sialyl-Le(a), i.e. CA19-9 (MAB 19-9), native Le(a) (MAB anti Le(a)) and alpha(2,3)sialyl type 1 structure, i.e. CA 50 (MAB C50). To detect sialic acid masked Le(a) reactivity, MAB anti-Le(a) was also applied to native and enzymatically desialylated tissue sections with and without masking of sialic acid residues by sialic acid and sequence specific lectins. Only 7 MTC (15%) displayed a weak expression of CA19-9, while 16 (33%) showed moderate positive staining for native Le(a). Twenty-seven tumours exhibited a strong staining by the N'ase MAB anti Le(a) staining sequence. The latter could most effectively be inhibited by the simultaneous masking of alpha(2,3)-and alpha-(2,6)-linked sialic acid residues due to the competitive binding of sialic acid and sequence specific lectins: Maackia amurensis agglutinin (specific alpha(2,3)-linked sialic acid) and Sambucus nigra agglutinin (specific alpha(2,6)-linked sialic acid). Thus, in MTC the major portion of sialic acid masked Le(a) antigen reactivity is different from that detected by the MAB 19-9. The antigen reactivity is probably due to Le(a) structures containing both alpha(2,3) and alpha(2,6)-linked sialic acid residues. A highly significant correlation between the expression of CA50 and that detected by the N'ase MAB anti-Le(a) staining sequence indicates that the alpha(2,3)-sialyl type 1 chain represents a common intermediate structure within the pathway of the biosynthesis of sialylated Le(a) antigens, excluding the formation of CA19-9 via the formation of the disialyl type 1 structure. This is subsequently fucosylated to the corresponding sialic acid masked Le(a). Preliminary clinicopathological studies indicate that the sialic acid masked Le(a) antigens detected by the N'ase MAB anti-Le(a) staining sequence are related to biologically aggressive MTC.

Effect of Hypoxia on Endothelial Cell Surface Glycoprotein Expression: Modulation of Glycoprotein IIIa and Other Specific Surface Glycoproteins

Exposure to hypoxia alters many aspects of endothelial cell metabolism and function; however, changes in surface glycoconjugates under these conditions have not been extensively evaluated. In the current studies, we examined surface glycoproteins of cultured bovine aortic (BAEC) and pulmonary arterial (BPAEC) endothelial cells under standard culture conditions (21% oxygen) and following exposure to hypoxia (0% oxygen) for varying time periods (30 min to 18 h) using a system of biotinylation, lectin binding (concanavalin A, Con A; Griffonia simplicifolia , GSA; Arachis hypogaea, PNA; Ricinus communis, RCA; or Triticum vulgaris, WGA), subsequent strep-avidin binding, and staining. Using these methods, we identified differences in lectin binding between the two cell types cultured in 21% oxygen with all lectins except PNA. With exposure to 0% oxygen, there was no change in lectin binding to most surface glycoproteins. Several surface glycoproteins, including glycoprotein IIIa on both cell types, demonstrated a time-dependent decrease in lectin binding; in addition, there was an increase in lectin binding to a few specific surface glycoproteins on each cell type within 30-60 min of exposure to 0% oxygen. These changes in specific surface glycoproteins were confirmed in both cell types by 125I labeling. Increased lectin binding was observed for Con A binding BAEC glycoproteins at molecular weight (MW) 116, 130, and 205 kDa, GSA binding BAEC glycoproteins at MW 120 and 205 kDa, and RCA binding BPAEC glycoproteins at MW 140 and 205 kDa. Increased binding of WGA or PNA was not observed during exposure to hypoxia. The specificity of lectin binding was further confirmed by competitive inhibition with the appropriate sugar. These studies demonstrate that there are baseline differences between BAEC and BPAEC cell surface glycoproteins and that exposure to hypoxia is associated with little change in lectin binding to most surface glycoproteins. There is, however, increased surface expression of a few glycoproteins that differ depending of the origin of the endothelial cell. Although the mechanism of this increase in lectin binding is not yet clear, subsequent studies suggested that it is due to increased availability of select carbohydrate moieties. The time course of these alterations suggests a possible role in the endothelial cell response to decreases in ambient oxygen tension.

Characterization of glycoproteins fromChlamydia trachomatisusing lectins

Glycoproteins in Chlamydia trachomatis, serotype L1, elementary bodies were studied by lectin blotting. A panel of 23 lectins representing a variety of sugar specificities was used. The pattern of lectin-binding specificities at a peptide band was studied in order to determine the type and structure of its glycoconjugate. To establish chlamydial origin of the glycopeptide bands in the blot, control samples from non-infected host cell membranes were run in parallel. Terminal mannosidic structures were demonstrated in a 72 kDa glycopeptide (gp72) by its selective binding of Galanthus nivalis lectin (GNA). Sialic acids were found in two chlamydial glycopeptides, gp40 and gp64, which appear to carry O-linked glycoconjugates as they bound the peanut agglutinin (PNA, both gp40 and gp64) and jackfruit lectin (Jac, only gp40). Such structures were also present in other chlamydial glycopeptides. Lectins with specificities for fucose in different links, galactose and N-acetyl glucosamine bound to several chlamydial peptides. On the basis of our results we suggest an alternative mechanism for uptake of chlamydial elementary bodies into host cells, namely phagocytosis mediated by eukaryotic cell surface lectins.

Characterization of the type of carbohydrate chains of the higher molecular weight (140 kDa) acid phosphatase of the frog liver

1.1. The higher molecular weight, (HMW, Mr 140 kDa) acid phosphatase (AcPase) of the frog liver (Rana esculenta) was separated into enzymatically active components by isoelectric focusing in an immobilized pH gradient and their carbohydrate chains were analyzed by specific lectin binding after native blotting.2.2. The lectin-binding patterns obtained with ConA, WGA, LcH and PNA as well as with WGA and PNA after desialylation indicate that the frog liver HMW AcPase contains predominantly N-linked complex and/or hybride type carbohydrate chains with terminal sialic acid and fucose residues; O-glycosylated enzyme components with free and sialic acid substituted Gal-GalNAc sequences were also detected.

Differentiative changes in fucosyltransferase activity in newborn rat epidermal cells

An enzymatic activity catalyzing the transfer of L-fucose from GDP-L-fucose to a glycoprotein that is associated with the surfaces of the basal cells has been found in the membranous fraction of the cutaneous epidermis from the newborn rat. This fucosyltransferase which is located in the differentiated cells alters the acceptor glycoprotein's lectin-binding specificity from the Isolectin I-B 4 of Griffonia simplicifolia (GS I-B 4) to the Agglutinin I of Ulex europeus (UEA) and could be responsible for the same change in lectin-binding specificity that occurs as the epidermal basal cell differentiates. Another membraneous fucosyltransferase that can use asialofetuin — but not the GS I-B 4-binding glycoprotein — as an acceptor, is also present in the membraneous fraction.

Effect of lectins on hepatic clearance and killing of Candida albicans by the isolated perfused mouse liver.

The isolated perfused mouse liver model was used to study the effects of various lectins on hepatic trapping and killing of Candida albicans. After mouse livers were washed with 20 to 30 ml of perfusion buffer, 10(6) C. albicans CFU were infused into the livers. At the time of recovery, 63% +/- 2% (mean +/- standard error of the mean) of the infused C. albicans CFU were recovered from the liver and 14% +/- 1% were recovered from the effluent for a total recovery of 77% +/- 2%. This indicated that 86% +/- 9% of the original inoculum was trapped by the liver and that 23% +/- 2% was killed within the liver. When included in both preperfusion and postperfusion buffers (0.2 mg of lectin per ml), Ulex europeaus lectin (binding specificity for fucose) decreased hepatic trapping of C. albicans by 37% and eluted trapped C. albicans from the liver only when included in postperfusion buffer. By comparison, treatment of C. albicans with U. europeaus lectin before infusion had no effect on the trapping or killing of yeast cells. When Lens culinaris lectin (binding specificity for mannose) was included in the perfusion buffers, hepatic killing of C. albicans increased by 16% with no significant effect on hepatic killing when yeast cells were treated with L. culinaris lectin before infusion. Forty to 55% of the infused C. albicans were killed when concanavalin A (binding specificities for mannose and glucose), Glycine max (binding specificity for N-acetylgalactosamine), or Arachis hypogea (binding specificity for galactose) lectin was included in the perfusion buffer or when yeast cells were treated with these lectins before their infusion. When C. albicans was treated with concanavalin A at a concentration of less than 0.02 mg/ml, hepatic killing of yeast cells was not significantly increased. The data suggest that a fucose-containing receptor on the surface of either sinusoidal endothelial cells or Kupffer cells is involved in the trapping of C. albicans by the perfused mouse liver. Moreover, lectins with binding specificities for mannose, N-acetylgalactosamine, and galactose increased hepatic killing of C. albicans.

Affinity sorbent for galactose specific lectins consisting of deproteinized plant cell wall ghosts

Purified plant cell wall ghosts with high pectin content were investigated for affinity sorption of the galactose specific plant lectins from Ricinus communis L., Glycine max (L.) Merr. and Vviscum album L. The experiments were carried out with the previously described material Permselect prepared from suspension cultured cells of Chenopodium album L. which has been originally developed for permeation chromatography, and a similar material obtained from parenchyma of cabbagehead ( Brassica oleracea L.). After a pretreatment with 2% Na 2CO 3 at room temperature, both cell wall materials showed strongly increased capacity for affinity binding of galactose specific lectins from the crude extracts. Ricin (RCA 2) is the Ricinus lectin preferently bound to pretreated Permselect. Results are discussed with respect to cell wall composition and porosity as well as practical application of Permselect for purification of galactose specific lectins.

Bioadhesion by means of specific binding of tomato lectin.

The possibility of developing bioadhesive drug delivery systems on the basis of molecules which selectively bind to the small intestinal epithelium by specific, receptor-mediated mechanisms was investigated using a lectin isolated from tomato fruits (Lycopersicum esculentum). The tomato lectin (TL) was found to bind specifically onto both isolated, fixed pig enterocytes and monolayers of human Caco-2 cell cultures with a similar affinity. TL-coated polystyrene microspheres (0.98 micron) also showed specific binding to enterocytes in vitro. Lectin binding was found to be favored at neutral pH and to be reduced in an acidic environment. Crude pig gastric mucin, however showed a marked cross-reactivity in vitro, indicating that lectin binding to the cell surface in vivo might be inhibited by mucus.

Ascaris suum: Stage-specific differences in lectin binding to the larval cuticle

A glycosylated component with affinity for wheat germ agglutinin (specific binding to n-acetyl- d-glucosamine monomers and oligomers) and weak affinity for poke weed mitogen (specific binding to n-acetyl- d-glucosamine oligomers) was detected temporally on the surface of Ascaris suum larvae developing in vitro and on in vivo-derived larvae. The component was identified on the surface of in vitro-derived late second stage larvae, on all late third stage larvae (derived from pig lung), and all fourth stage larvae (obtained from pig intestines and from in vitro culture) of A. suum. None of the newly hatched first molt, L2, or early L3 bound any of the lectins tested. The component exhibited no affinity for concanavalin A (specific binding to α- d-mannosyl and α- d-glucosyl residues) or Dolichos biflorus lectin (specific binding to n-acetyl- d-galactosamine). A component with similar lectin binding specificities had previously been found on the cuticle of adult A. suum.

Lectin Staining of Peritoneal Mesothelial Cells in Vitro

A survey of lectin-binding specificities present on rodent and human mesothelial cells propagated and maintained in tissue culture was made using fluorescein isothiocynate conjugated (FITC) lectins. Rodent and human cells exhibited cell-associated fluorescence following exposure to the FITC-lectins from C. ensiformis, T. vulgaris, A. hypogaea, E. cristagalli and B. simplicifolia, but not with lectins from G. max and D. biflorus. Rodent cells were also positive for FITC-M. pomifera lectin binding. Human, but not rodent, cells were positive for FITC-T. purpureas lectin binding. Exposure of rabbit mesothelial cells in vitro to FITC-lectins that bound to the cell surface resulted in the appearance of discrete loci of putatively intracellular fluorescence. Exposure of cells to ferritin-labelled T. vulgaris lectin at 37 degrees C for as little as 7.5 minutes resulted in the appearance of ferritin-size particles in intracellular vesicles. These results demonstrate 1. the presence of lectin-binding sites in and on peritoneal mesothelial cells from rodents and humans and 2. a possible role of such sites in mediating the entry of lectin-like endogenous molecules into the vacuolar apparatus of these cells.

Substitution of asparagine for serine-406 of the immunoglobulin μ heavy chain alters glycosylation at asparagine-402

Previous work suggested that the substitution of Asn for Ser at position 406 of the mu heavy chain of mouse IgM results in aberrant glycosylation at Asn402. In order to characterise the apparently abnormal glycosylation process more precisely, the mutant and wildtype mu chains were fragmented by cleavage with cyanogen bromide, and the resulting glycopeptides were analysed further. Measurements of lectin binding specificity as well as glycosidase sensitivity suggest that the oligosaccharide at Asn402 of wildtype mu is a hybrid type which does not contain terminal alpha(2-6) or alpha(2-3) linked sialic acid. By contrast, the corresponding oligosaccharide on Asn402 of mutant mu is complex and contains terminal sialic acid linked alpha(2-6) to galactose. The structural features for specifying the abnormal glycosylation are present in monomeric mutant IgM.

Lectin staining of neoplastic and normal background colorectal mucosa in nonpolyposis and polyposis patients

A lectin histochemistry approach was adopted for comparative assessment of a colon cancer risk. Binding of Ulex europaeus agglutinin-I (UEA-I), peanut agglutinin (PNA), Griffonia simplicifolia agglutinin-II (GSA-II), and Dolichos biflorus agglutinin (DBA) was investigated in tumor and background tissue from a total of 34 adenoma and 44 cancer patients and compared with reaction patterns in control and familial adenomatous polyposis (FAP) patients. Adenoma patients with UEA-I positive rectal mucosa were found to have a 33.3 percent familial history of large bowel cancer, which was significantly higher (P less than 0.05) than the respective 4.0 percent figure for patients with negative rectal mucosa. In the cancer patients, an even stronger correlation was noted, with a 63.2 percent UEA-I positive family history association being recorded, as opposed to 4.0 percent in the negative rectal mucosa patients (P less than 0.01). Thus, the results suggest that, apparently, normal rectal background mucosa of individuals genetically at high risk for colon and rectal cancer demonstrates a specific lectin binding ability similar to that of FAP patients and that the simple method using UEA-I staining of rectal biopsy specimens can be of practical use in identification of high-risk colorectal cancer.

Lectin Histochemistry of Trachea and Lung of Healthy Turkeys (Meleagris gallopavo) and Turkeys with Pneumonia

Thirteen lectins were used to characterize lectin-binding specificity of glycoconjugates on sections of formalin-fixed lung and trachea from seven normal turkeys, two turkeys with acute pneumonia, and two turkeys with chronic pneumonia. Neuraminidase was used to digest sialic acid residues. One N-acetylgalactosamine-binding lectin and two N-acetylgalactosamine/galactose-binding lectins stained the apical membrane and cytoplasm of multifocal cells that lined air atria and hyperplastic granular cells. Other lectins in these groups stained ciliated cells of the trachea and bronchi and air capillary epithelial cells. Sialic acid residues were on apical surfaces of ciliated and nonciliated tracheal and bronchial lining cells, air capillary epithelial cells, and vascular endothelial cells. Mannose/glucose-binding lectins stained reticular and elastic fibers in the lamina propria of trachea, primary and secondary bronchi, and the tunica adventitia of arteries and veins. By transmission electron microscopy, colloidal gold-Arachis hypogaea (peanut agglutinin) labeled microvilli on the apical surface of mature granular cells. The L-fucose-binding lectin, in addition to several other lectins, stained nonspecifically in both trachea and lung. These studies show that granular cells that line air atria can be identified with lectins of N-acetylglucosamine and N-acetylgalactosamine/galactose groups, and that apical surfaces of epithelial cells and endothelial cells in the trachea and lung express terminal sialic acid residues.

Oligosaccharide structures present on asparagine-289 of recombinant human plasminogen expressed in a Chinese hamster ovary cell line

The oligosaccharide structures linked to Asn289 of a recombinant (r) variant (R561S) human plasminogen (HPg) expressed in Chinese hamster ovary (CHO) cells, after transfection of these cells with a plasmid containing the cDNA coding for the variant HPg, have been determined. Employing high-performance anion-exchange liquid chromatography mapping of the oligosaccharide units cleaved from the protein by glycopeptidase F, compared with elution positions of standard oligosaccharides, coupled with monosaccharide compositional determinations and analyses of sequential exoglycosidase digestions and specific lectin binding, we find that considerable microheterogeneity in oligosaccharide structure exists at this sole potential N-linked glycosylation site on HPg. A variety of high-mannose structures, as well as bi-, tri-, and tetraantennary complex-type carbohydrate, has been found, in relative amounts of 1-25% of the total oligosaccharides. The complex-type structures contain variable amounts of sialic acid (Sia), ranging from 0 to 5 mol/mol of oligosaccharide in the different glycan structures. Neither hybrid-type molecules, N-acetylglucosamine bisecting oligosaccharides, nor N-acetyllactosaminyl-repeat structures were found to be present in the complex-type carbohydrate pool in observable amounts. Of interest, a significant portion of the Sia exists an outer arm structures in an (alpha 2,6) linkage to the penultimate galactose, a novel finding in CHO cell-directed glycosylation.(ABSTRACT TRUNCATED AT 250 WORDS)

Light and electron microscopic studies of lectin binding on the glycocalyx of rat pancreatic cells

Lectin binding of the glycocalyx of pancreatic tissue sections as well as of isolated pancreatic acini and acinar cells was studied of healthy wistar rats by light and electron microscopy. For light microscopy, we used FITC (WGA, RCAI, LCA) or peroxidase marked (WGA, RCAI, PNA, PHA, LCA, UEAI, LPA) as well as unmarked lectins (Con A, VAA I). Gold marked lectins were used for electron microscopy (WGA, RCAI, LCA, HPA, PNA, VAAI-B). Intact acinar cells in pancreatic tissue sections and isolated acini showed a strong binding of WGA, RACI, and HPA on the apical cell surface, whereas VAAI, UEAI, LCA, and Con A reacted strongly with the basolateral glycocalyx, but not with the apical surface. The 2 main domains of the glycocalyx of pancreatic cells showed their specific lectin binding so long as the junctional complexes between the cells are intact. The polarity of the cell surface of pancreatic acinar cells is discussed in regard to the possible function of the 2 domains.

Glycoconjugate Heterogeneity among Five Strains of Mycoplasma gallisepticum

Five strains of Mycoplasma gallisepticum (MG) were examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis for the presence of carbohydrate-containing components. Staining with periodic acid-Schiff (PAS) demonstrated carbohydrate components in three of the five strains studied. The PAS-reactive bands counterstained for protein, indicating a possible glycoprotein nature. Western blot analysis using three biotinylated lectin probes demonstrated the presence of additional glycoconjugates in the blot profiles of each MG strain. The carbohydrate specificity of lectin binding was demonstrated by competition experiments using specific sugars. Differences in the number, electrophoretic mobility and the morphology of PAS and lectin reactive bands were reproducible among separate preparations of each MG strain. These findings indicate substantial phenotypic diversity among the five MG strains in their ability to produce or acquire glycoconjugates.

In Vitro Differentiation of Leukemic Cells to Eosinophils in the Presence of Interleukin-5 in Two Cases of Acute Myeloid Leukemia with the Translocation (8;21)(q22;q22)

We demonstrated the significant eosinophilic growth of leukemic cells in the presence of interleukin-5 (IL-5) in 2 of 15 cases of acute myeloid leukemia. These two cases were M2 (FAB classification) with the translocation (8;21)(q22; q22). Bone marrow examination revealed the rather high percentages (6% and 9%) of atypical eosinophils in the total nucleated bone marrow cells in these two cases. In the remaining 13 cases, eosinophils were less than 2% in the nucleated bone marrow cells. In the methylcellulose culture system, 142 +/- 18 or 54 +/- 2 colonies were formed by 5 x 10(4) mononuclear cells in the presence of IL-5 in these two cases. These colonies mainly comprised mature eosinophils. Eosinophils were confirmed by Biebrich scarlet staining and electron microscopic examination using a specific lectin binding assay. The eosinophilic differentiation and proliferation of leukemic cells were also observed in the liquid culture system. It was shown that eosinophils observed in both systems were derived from leukemic cells using the chromosomal marker of leukemic cells, t(8;21). Leukemic cells also differentiated to neutrophils or both neutrophils and eosinophils in response to granulocyte colony-stimulating factor or interleukin-3, respectively, but did not respond noticeably to granulocyte-macrophage colony-stimulating factor. Although IL-5 acts on normal eosinophil committed precursors as a lineage-specific growth factor, at least some leukemic cells reacted to IL-5 and could proliferate and differentiate along eosinophilic pathway. Our findings suggest that atypical eosinophils observed in the bone marrow were derived from the leukemic clone in two cases of AML.