Agglutinability of Red Cells by Anti-i in Patients with Thalassæmia Major and Other Hæmatological Disorders

Characterization and statistical modeling of glycosylation changes in sickle cell disease

Sickle cell disease is an inherited genetic disorder that causes anemia, pain crises, organ infarction, and infections in 13 million people worldwide. Previous studies have revealed changes in sialic acid levels associated with red blood cell sickling and showed that stressed red blood cells bare surface-exposed clustered terminal mannose structures mediating hemolysis, but detailed glycan structures and anti-glycan antibodies in sickle cell disease remain understudied. Here, we compiled results obtained through lectin arrays, glycan arrays, and mass spectrometry to interrogate red blood cell glycoproteins and glycan-binding proteins found in the plasma of healthy individuals and patients with sickle cell disease and sickle cell trait. Lectin arrays and mass spectrometry revealed an increase in α2,6 sialylation and a decrease in α2,3 sialylation and blood group antigens displayed on red blood cells. Increased binding of proteins to immunogenic asialo and sialyl core 1, Lewis A, and Lewis Y structures was observed in plasma from patients with sickle cell disease, suggesting a heightened anti-glycan immune response. Data modeling affirmed glycan expression and plasma protein binding changes in sickle cell disease but additionally revealed further changes in ABO blood group expression. Our data provide detailed insights into glycan changes associated with sickle cell disease and refer glycans as potential therapeutic targets.

Branched I antigens on leukemia cells enhanced sensitivity against natural killer-cell cytotoxicity through affecting the target-effector interaction

BACKGROUND

The aberrant glycosylation on proteins and lipids has been implicated in malignant transformations for promoting the tumorigenesis, metastasis, and evasion from the host immunity. The I-branching β-1,6-N-acetylglucosaminyltransferase, converting the straight i to branched I histo-blood group antigens, reportedly could influence the migration, invasion, and metastasis of solid tumors.

STUDY DESIGN AND METHODS

We first chose the highly cytotoxic natural killer (NK)-92MI cells as effector against leukemia for this cell line has been used in several clinical trials. Fluorescence-activated cell sorting and nonradioactive cytotoxicity assay were performed to reexamine the role of NK-activating receptors, their corresponding ligands, and the tumor-associated carbohydrate antigens in this NK-92MI-leukemia in vitro system. The I role on cytotoxic mechanism was further studied especially on the effector-target interactions by cytotoxic analysis and conjugate formation assay.

RESULTS

We showed that expression levels of leukemia surface ligands for NK-activating receptors did not positively reflect susceptibility to NK-92MI. Instead, the expression of I antigen on the leukemia cells was found important in mediating the susceptibility to NK targeting by affecting the interaction with effector cells. Furthermore, susceptibility was shown to dramatically increase while overexpressing branched I antigens on the I- cells. By both conjugate and cytotoxicity assay, we revealed that the presence of I antigen on leukemia cells enhanced the interaction with NK-92MI cells, increasing susceptibility to cell-mediated lysis.

CONCLUSION

In our system, branched I antigens on the leukemia were involved in the immunosurveillance mediated by NK cells specifically through affecting the effector-target interaction.

Phosphorylation status of transcription factor C/EBPalpha determines cell-surface poly-LacNAc branching (I antigen) formation in erythropoiesis and granulopoiesis

The cell-surface straight and branched repeats of N-acetyllactosamine (LacNAc) units, called poly-LacNAc chains, characterize the histo-blood group i and I antigens, respectively. The transition of straight to branched poly-LacNAc chain (i to I) is determined by the I locus, which expresses 3 IGnT transcripts, IGnTA, IGnTB, and IGnTC. Our previous investigation demonstrated that the i-to-I transition in erythroid differentiation is regulated by the transcription factor CCAAT/enhancer binding protein alpha (C/EBPalpha). In the present investigation, the K-562 cell line was used as a model to show that the i-to-I transition is determined by the phosphorylation status of the C/EBPalpha Ser-21 residue, with dephosphorylated C/EBPalpha Ser-21 stimulating the transcription of the IGnTC gene, consequently resulting in I branching. Results from studies using adult erythropoietic and granulopoietic progenitor cells agreed with those derived using the K-562 cell model, with lentiviral expression of C/EBPalpha in CD34(+) hematopoietic cells demonstrating that the dephosphorylated form of C/EBPalpha Ser-21 induced the expression of I antigen, granulocytic CD15, and also erythroid CD71 antigens. Taken together, these results demonstrate that the regulation of poly-LacNAc branching (I antigen) formation in erythropoiesis and granulopoiesis share a common mechanism, with dephosphorylation of the Ser-21 residue on C/EBPalpha playing the critical role.

I branching formation in erythroid differentiation is regulated by transcription factor C/EBPα

The histo-blood group i and I antigens have been characterized as straight and branched repeats of N-acetyllactosamine, respectively, and the conversion of the straight-chain i to the branched-chain I structure on red cells is regulated to occur after birth. It has been demonstrated that the human I locus expresses 3 IGnT transcripts, IGnTA, IGnTB, and IGnTC, and that the last of these is responsible for the I branching formation on red cells. In the present investigation, the K-562 cell line was used as a model to show that the i-to-I transition in erythroid differentiation is determined by the transcription factor CCAAT/enhancer binding protein α (C/EBPα), which enhances transcription of the IGnTC gene, consequently leading to formation of the I antigen. Further investigation suggested that C/EBPα IGnTC-activation activity is modulated at a posttranslational level, and that the phosphorylation status of C/EBPα may have a crucial effect. Results from studies using adult and cord erythropoietic cells agreed with those derived using the K-562 cell model, with lentiviral expression of C/EBPα in CD34+ hemopoietic cells demonstrating the determining role of C/EBPα in the induction of the IGnTC gene as well as in I antigen expression.

Back to the Beginnings: An Autobiography

After receiving BS and MS degrees from the University of Washington in Seattle, I entered its new medical school in 1947, receiving an MD degree in 1951. After internship and residency, I obtained a 2-year postdoctoral fellowship in hematology under the guidance of Dr Clement Finch. The last 6 months of the fellowship were spent in London, England, at Dr Patrick Mollison's Blood Transfusion Research Unit. There I met and worked with Marie Cutbush (later Crookston) who has been a long-term friend. On returning to Seattle, I joined the faculty of the medical school and became the associate director of the Puget Sound Blood Center. There, I supervised the blood typing and cross-matching laboratory, introducing methods I had learned in London and measuring the effectiveness of various cross-matching procedures. My own research was largely directed toward human genetic polymorphism, and I wrote a textbook published in 1969, describing the biochemical structure, function, inheritance, and geographic distribution of the genetic markers. Subsequently, I discovered that 2 forms of inherited immunodeficiency disease were due to deficiencies of the enzymes adenosine deaminase and purine nucleoside phosphorylase. In 1979, I became the director of the blood center and was shortly afterwards elected to the National Academy of Sciences. I retired in 1987 and have spent most of the intervening years relearning to play the violin and exploring the wonders of chamber music.

Traveling for the glycosphingolipid path

Our studies on glycosphingolipids (GSLs) were initiated through isolation and structural characterization of lacto-series type 1 and 2 GSLs, and globo-series GSLs. Lacto-series structures included histo-blood group ABH and I/i antigens. Our subsequent studies were focused on GSL changes associated with: (i) ontogenic development and differentiation; (ii) oncogenic transformation and tumor progression. Various novel types of GSLs such as extended globo-series, sialyl-Le(x) (SLe(x)), sialyl-dimeric-Le(x) (SLe(x)-Le(x)), dimeric-Le(x) (Le(x)-Le(x)), Le(y)-on-Le(x), dimeric-Le(a) (Le(a)-Le(a)), Le(b)-on-Le(a), etc. were identified as tumor-associated antigens. These studies provide an essential basis for up- or down-regulation of key glycosyltransferase genes controlling development, differentiation, and oncogenesis. GSL structures established in our laboratory are summarized in Table 1, and structural changes of GSLs associated with ontogenesis and oncogenesis are summarized in Sections 2 and 3. Based on these results, we endeavored to find out the cell biological significance of GSL changes, focused on (i) cell adhesion, e.g., the compaction process of preimplantation embryo in which Le(x)-to-Le(x), Gb4-to-GalGb4 or -nLc4 play major roles; and (ii) modulation of signal transduction through interaction of growth factor receptor tyrosine kinase with ganglioside, e.g., EGF receptor tyrosine kinase with GM3. Recent trends of studies on i and ii lead to the concept that GSL clusters (microdomains) are organized with various signal transducer molecules to form 'glycosignaling domains' (GSD). GSL-dependent adhesion occurs through clustered GSLs, and is coupled with activation of signal transducers (cSrc, Src family kinase, Rho A, etc.). Clustered GSLs involved in cell adhesion are recognized by GSLs on counterpart cells (carbohydrate-to-carbohydrate interaction), or by lectins (e.g., siglecs, selectins). Our major effort in utilization of GSLs in medical science has been for: (i) cancer diagnosis and treatment (vaccine development) based on tumor-associated GSLs and glycoepitopes; (ii) genetically defined phenotype for susceptibility to E. coli infection; (iii) clear identification of physiological E-selectin epitope (myeloglycan) expressed on neutrophils and myelocytes; (iv) characterization of sialyl poly-LacNAc epitopes recognized as male-specific antigens. Utilization of these GSLs or glycoepitopes in development of anti-adhesion approach to prevent tumor metastasis, infection, inflammation, or fertilization (i.e., contraceptive) is discussed. For each approach, development of mimetics of key GSLs or glycoepitopes is an important subject of future study.

Expression of blood group i antigen and fetal hemoglobin in paroxysmal nocturnal hemoglobinuria

BACKGROUND

The mechanism by which the paroxysmal nocturnal hemoglobinuria (PNH) clone progressively takes over normal hematopoietic cells remains unknown. The respective in vivo differentiation of normal and PNH erythroid progenitors was investigated through the expression of two fetal erythroid markers (i antigen and fetal hemoglobin [HbF]) whose expression in adult red cells is associated with altered erythropoiesis.

STUDY DESIGN AND METHODS

Murine monoclonal antibodies directed against HbF and i and CD59 antigens were used to phenotype red cells of 10 PNH patients. A multiparametric flow cytometry assay of red cells and reticulocytes was designed to assess a possible association of i and HbF with PNH or normal red cells.

RESULTS

Most patients exhibited greater expression of i and HbF than did normal controls. In each case, the percentages of i-positive or HbF-positive cells within CD59-deficient and CD59-positive red cells were very close, clearly showing a lack of preferential association of these markers with normal or PNH cells.

CONCLUSION

In PNH patients, normal and PNH erythroid progenitors have the same ability to promote HbF and i antigen expression, which suggests that normal and PNH erythroid progenitors (burst-forming units-erythroid, colony-forming units-erythroid, erythroblasts) behave similarly in response to bone marrow stress.

Congenital dyserythropoietic anaemia type II (HEMPAS) and its molecular basis

Congenital dyserythropoietic anaemia type II (CDA II) is a rare genetic anaemia in humans, inherited in an autosomally recessive mode. CDA II is also called HEMPAS as this disease is characterized by hereditary erythroblastic multinuclearity with positive acidified serum lysis test. Analyses of CDA II erythrocyte membranes showed that the band 3 glycoprotein is underglycosylated. An aberrant glycosylation pattern is seen in the polylactosamine carbohydrates which are normally attached to the band 3 and band 4.5 glycoproteins. The polylactosamines are, however, accumulated in the form of glycolipids. Therefore a genetic factor in CDA II appears to block the glycosylation of protein acceptors and shift these carbohydrates to the lipid acceptors. Structural analysis of CDA II band 3 carbohydrates identified truncated hybrid-type oligosaccharides and suggests that the Golgi glycosylation enzyme(s), alpha-mannosidase II or N-acetylglycosaminyltransferase II is defective in CDA II. By using a cDNA probe for alpha-mannosidase II, one CDA II case has been identified as being defective in the gene encoding alpha-mannosidase II. At present, it is not clear whether CDA II is a genetically heterogenous collection of glycosylation deficiencies, or genetically homogenous but apparently heterogenous in phenotype expression. Freeze-fracture electron microscopy and immunoelectron microscopy revealed that the band 3 glycoproteins are clustered in CDA II erythrocyte membranes. The abnormal distribution of band 3 might cause an unstable membrane organization. In CDA II erythroblasts, the membrane proteins might also be underglycosylated and abnormally distributed. When normal erythroblasts were cultured in vitro in the presence of swainsonine (alpha-mannosidase inhibitor) the erythroblasts became multinucleared. It is, therefore, quite possible that the enzymic defect of alpha-mannosidase II could cause various morphological anomalies including multinuclearity. Because the genes encoding glycosylation enzymes are housekeeping genes, the enzyme defect of CDA II is not restricted to erythroid cells and there is also an abnormal glycosylation of hepatocyte glycoproteins. On the other hand, there are many types of cells and tissues which appear not to be affected by the CDA II defect. A mechanism for the erythroid-specific manifestation of CDA II and its tissue specificity are also discussed.

Calcium, cell shrinkage, and prolytic state of human red blood cells

The effects of intracellular calcium, or Cac, on the Na permeability of human red blood cells were examined during 3-h incubations with the Ca ionophore A23187 and varied external Ca, Cao. Above 3 microM Cao, Nac increased significantly as ATP decreased. Maintenance of normal ATP with vanadate did not prevent the gain of Nac. Similar amounts of Nac were gained in 3 h by ouabain-treated cells exposed to the K ionophore valinomycin or by cells osmotically shrunken. Cells shrunken with sucrose also exhibited partial loss of Kc. When the cells with increased Nac were subsequently transferred to Na-free, high-K medium, the Nac and Kc that had changed slowly over 3 h returned toward normal within 10 min. The development of irreversible high cation permeability in shrunken cells was not prevented by a variety of transport inhibitors. These observations and cell volume distributions suggest that prolonged shrinkage induces a subpopulation of cells to become highly cation permeable, or "prolytic". The major effect of Cac on Na permeability appears to be an indirect consequence of cell shrinkage due to KCl loss.

Calcium ions, drug action and the red cell membrane

Intracellular calcium regulates a number of membrane functions in the erythrocyte, including control of shape, membrane lipid composition and cation permeability. Measurement of total red cell calcium has yielded values between 5 and 15 nmol/ml cells, and these low values in part reflect the absence of Ca2+ -containing organelles. Most intracellular Ca2+ is bound and the low cell ionized Ca2+ concentration (approximately 0.2 microM) is maintained by a combination of low membrane permeability and a powerful Ca2+ -pump. This pump has been identified with a (Ca2+ + Mg2+)-stimulated ATPase, and both Ca2+ transport and ATP splitting are stimulated by calmodulin, a low molecular weight protein which binds Ca2+ avidly and activates many Ca2+ -dependent enzymes. Both high and low affinity kinetics for Ca2+ pumping have been demonstrated, depending on the extent of binding of calmodulin to the pump. A stoichiometry of either 1 or 2 Ca2+ ions pumped per ATP molecule split has been shown, and the value may vary with the level of intracellular Ca2+. Phenothiazines, such as chlorpromazine inhibit the Ca2+ -pump by antagonizing the increment in activity produced by calmodulin. The passive inward leak of Ca2+ into erythrocytes can be quantitated by 45Ca2+ uptake into red cells whose Ca2+ -pump has been inhibited. Estimates of the Ca2+ permeability, based on unidirectional influx, yield values many orders of magnitude lower than for nucleated cells. Influx of Ca2+ into human erythrocytes occurs by a facilitated diffusion process, which can be inhibited by phenothiazines and the cinchona alkaloids. Calcium affects many membrane functions including cation permeability, lipid composition and some cytoskeletal interactions which may determine cell shape. Any rise in intracellular Ca2+ activates a specific K+ channel which normally makes little contribution to K+ fluxes. Kinetic studies of this process demonstrate either high or low affinity Ca2+ -activation of K+ efflux, with low affinity of the channel to Ca2+ being the probable state in vivo. Propranolol is the best known activator of Ca2+ -stimulated K+ efflux, although the mechanism of stimulation is unclear. Like other tissues, red cells possess a Ca2+ -activated phosphoinositol phosphodiesterase. Although it has been suggested that the echinocytic shape change induced by Ca2+ is due to the hydrolysis of polyphosphoinositides, it seems more likely that this shape change results from an effect of Ca2+ on the macromolecular interactions of the cytoskeleton. Abnormal Ca2+ permeability may contribute to red cell destruction in a variety of diseases. For example, in sickle cell anemia a large Ca2+ influx occurs when cells are sickled under deoxy conditions, and moreover, the ability of the Ca2+ -pump to extrude the increment of cell Ca2+ is impaired. Thus, red cell Ca2+ is increased 3-7-fold above normal and this may contribute to the short survival of sickle red cells...

Increased erythrocyte cation permeability in thalassemia and conditions of marrow stress

Calcium and sodium permeability of erythrocytes from patients with untransfused alpha- or beta- thalassemia major has been studied and compared to mature erythrocytes or control cells with comparable reticulocytosis. Isotopic Na(+) influx was increased a mean fourfold greater than normals and threefold greater than reticulocyte rich control. Passive net leak of Na(+) into thalassemic cells incubated with ouabain was also increased corresponding to their greater (22)Na(+) influx. Erythrocyte Na(+) and K(+) concentrations and cell water content per unit volume of cells were normal. Quantitation of active cation pumps in the cell membrane by the technique of [(3)H]ouabain binding showed a 2.6- to 9.9-fold increase above normal. Inward Ca(2+) movement was studied in cells with absent Ca(2+) pumping produced by depletion of either ATP or Mg(2+)-ions. Calcium uptake by ATP depleted thalassemic cells was increased 12-fold above normals and 3.6-fold above reticulocyte-rich controls. The Ca(2+) uptake by Mg(2+)-depleted thalassemic cells was also increased above normal confirming that erythrocyte Ca(2+) permeability is increased in this disease. Osmotic fragility measurements show that the surface area to volume ratio of thalassemic erythrocytes was increased by 15 to 25% above mature erythrocytes. The increased passive cation permeability of thalassemic erythrocytes cannot be explained by either reticulocytosis or an increased surface area to volume ratio of these cells. Moreover, erythrocyte Na(+) and Ca(2+) influxes in congenital dyserythropoietic anemia (CDA type 2) were increased 2- and 14-fold, respectively, above normal. The increased cation fluxes and cation pump numbers in thalassemic and congenital dyserythropoietic anemia erythrocytes are consistent with the hypothesis of membrane immaturity arising from rapid marrow transit times, a concept previously advanced to explain the persistence of i-antigen on these cells.

Anomalous cell surface structure of sickle cell anemia erythrocytes as demonstrated by cell surface labeling and endo-beta-galactosidase treatment

Erythrocyte surface glycoproteins from patients with various types of sickle cell anemia have been analyzed and compared with those from normal individuals. By hemagglutination with various anti-carbohydrate antibodies, sickle cells showed profound increase of i antigens and moderate increase of GlcNAc beta 1 leads to 3Gal beta 1 leads to 3 Glc structure, whereas antigenicity toward globosidic structure was unchanged. In parallel to these findings, erythrocytes of sickle cell patients have additional sialylated lactosaminoglycan in Band 3. Thus, it can be concluded that erythrocytes of sickle cell patients are characterized by an altered cell surface structure which does not appear to be due to topographical changes of cell surface membrane. It is possible that the anemia or the "stress" hematopoiesis in these patients is responsible for these changes.

The blood group Ii system: a carbohydrate antigen system defined by naturally monoclonal or oligoclonal autoantibodies of man

The monoclonal anti-I and -i autoantibodies of cold agglutinin disease have drawn attention to a class of carbohydrate antigens which are of considerable biological interest. These antibodies have also served to illustrate the remarkable differences that can exist in the fine specificities of monoclonal antibodies directed against relatively simple antigens borne on hexa- to octasaccharide structures.

K562 human leukaemic cells express fetal type (i) antigen on different glycoproteins from circulating erythrocytes

During the ontogenic change from fetal to adult human erythrocytes, as well as fetal haemoglobin being replaced by adult haemoglobin, the cell-surface antigen i is converted to I (ref. 1). Recently it has been shown that this antigenic change is the conversion of the linear repeating Gal beta 1 leads to 4GlcNac beta 1 leads to 3Gal structure to branched Gal beta 1 leads to 4GlcNac beta 1 leads to 3(Gal beta 1 leads to 4GlcNac beta 1 leads to 6)Gal structure. We have shown that cell-surface labelling followed by endo-beta-galactosidase digestion can distinguish these two forms on the cell surface, and that band 3 and band 4.5 are the major carriers for these antigens on mature erythrocytes. Human leukaemic cell line K562, originally isolated from a patient at blast crisis of chronic myelocytic leukaemia, has recently been shown to synthesize glycophorin A, and to be capable of synthesizing haemoglobin upon induction. I demonstrate here that K562 cells express the fetal type (i) antigen on distinctly different glycoproteins from those of erythrocytes, by the use of cell-surface labelling followed by endo-beta-galactosidase digestion or followed by immunoprecipitation with specific antibodies.

A new red cell phenotype I-i-: red cells lacking both I and i antigens

When blood samples from 5,864 healthy Indian blood donors were tested with anti-I, seven samples failed to react or reacted only weakly. Further studies with a number of anti-I and anti-i sera suggested that these samples had the phenotype I-i- and reacted weakly with anti-IT. In one family, the I-i- phenotype was present in three generations.

Acquired haemoglobin H disease, complicating a myeloproliferative syndrome: a case report

A case of acquired haemoglobin H disease in association with a myeloproliferative disorder is described. Severe haemolysis with hypochromic microcytic anaemia was present. Haemoglobin H formed 18% of the circulating haemoglobin and 60% of the red cells showed multiple inclusions on incubation with brilliant cresyl blue. Blood film and absolute red cell values from a previous unrelated illness were normal, proving the acquired nature of the haemoglobin abnormality. Alpha/beta chain synthesis was measured in vitro and the degree of imbalance (alpha/beta ratio 0.39) was similar to that seen in the inborn thalassaemic disorder. A small proportion of red cells showed i-antigen reactivity but their haemoglobin H content was no different from the majority of cells which were l-antigen positive.

I and i antigens on normal and leukemic leukocytes

Human leukocyte I and i antigens were quantitated using 125I labelled purified antibodies. Binding of these antibodies to leukocytes was dependent on reduced temperature. No significant difference in antigen content was observed between normal and leukemic myeloid leukocytes. B lymphocytes bound much greater amounts of both I and i antibodies than did T lymphocytes. Neoplastic lymphoid cells bound widely divergent amounts of both antibodies with chronic lymphocytic leukemia and lymphosarcoma cell leukemia cells binding much decreased amounts compared to normal lymphocytes. Cells from patients with hairy cell leukemia bound very large quantities of these antibodies in a cold dependent fashion. These elevated levels of binding were not due to nonspecific binding of IgM.

[Enzyme deficiencies of blood cells in bone marrow insufficiency (author's transl]

Numerous enzyme defects-deficiency of pyruvate kinase, phosphofructo-kinase, glocosephosphate isomerase, adenylate kinase, 2,3-diphosphoglycerate mutase and glutathione reductase--in red blood cells have been described to be connected with dyserythropoietic or refractory anemias and panmyelopathies of different origin. These enzyme deficiencies also have been demonstrated in red cells of patients with acute leukemia. Most likely the enzyme deficiencies are acquired and are not important for the origin of anemia or bone marrow insufficiency. Partial derepression of fetal genes, qualitative and quantitative perturbations of genetic expression, and posttranslational variations of the enzyme protein by low molecular factors from plasma, erythrocytes or leukemic cells have been discussed as a reason of enzyme deficiency. The decrease of glutathione reductase deficiency is dependent of FAD deficiency.

Increased expression in erythrocytic Ii antigens in sickle cell disease and sickle cell trait

The reactivity of human erythrocytes with anti-I and anti-i sera was estimated in 36 normal adults, 18 individuals with sickle cell anemia and 27 individuals with sickle cell trait, by quantitative hemagglutination with a sensitive autoanalyzer system. SS erythrocytes showed significantly increased I and i reactivity when compared to normal erythrocytes. AS erythrocytes also demonstrated some increase in Ii reactivity. No correlation was found between erythrocytic HbF content and i reactivity.

Increase in strength of red cell Bga antigen following infectious mononucleosis

Following an attack of infectious mononucleosis, the red cells of HLA B7 patients may show a greatly increased reactivity with anti-Bga antibodies. This occurs from about the 3rd week of the disease, the reactivity then slowly decreasing over a period of months or years. Both specific and nonspecific reactions to other HLA antisera may also occur. Five patients were followed in detail and the reason for the increase in antigen strength was investigated without a conclusive result.