Halina Miller-Podraza

N‐Glycolylneuraminic Acid Xenoantigen Contamination of Human Embryonic and Mesenchymal Stem Cells Is Substantially Reversible

Human embryonic and mesenchymal stem cell therapies may offer significant benefit to a large number of patients. Recently, however, human embryonic stem cell lines cultured on mouse feeder cells were reported to be contaminated by the xeno‐carbohydrate N‐glycolylneuraminic acid (Neu5Gc) and considered potentially unfit for human therapy. To determine the extent of the problem of Neu5Gc contamination for the development of stem cell therapies, we investigated whether it also occurs in cells cultured on human feeder cells and in mesenchymal stem cells, what are the sources of contamination, and whether the contamination is reversible. We found that N‐glycolylneuraminic acid was present in embryonic stem cells cultured on human feeder cells, correlating with the presence of Neu5Gc in components of the commercial serum replacement culture medium. Similar contamination occurred in mesenchymal stem cells cultured in the presence of fetal bovine serum. The results suggest that the Neu5Gc is present in both glycoprotein and lipid‐linked glycans, as detected by mass spectrometric analysis and monoclonal antibody staining, respectively. Significantly, the contamination was largely reversible in the progeny of both cell types, suggesting that decontaminated cells may be derived from existing stem cell lines. Although major complications have not been reported in the clinical trials with mesenchymal stem cells exposed to fetal bovine serum, the immunogenic contamination may potentially be reflected in the viability and efficacy of the transplanted cells and thus bias the published results. Definition of safe culture conditions for stem cells is essential for future development of cellular therapies.

Glycomics of bone marrow-derived mesenchymal stem cells can be used to evaluate their cellular differentiation stage

Human mesenchymal stem cells (MSCs) are adult multipotent progenitor cells. They hold an enormous therapeutic potential, but at the moment there is little information on the properties of MSCs, including their surface structures. In the present study, we analyzed the mesenchymal stem cell glycome by using mass spectrometric profiling as well as a panel of glycan binding proteins. Structural verifications were obtained by nuclear magnetic resonance spectroscopy, mass spectrometric fragmentation, and glycosidase digestions. The MSC glycome was compared to the glycome of corresponding osteogenically differentiated cells. More than one hundred glycan signals were detected in mesenchymal stem cells and osteoblasts differentiated from them. The glycan profiles of MSCs and osteoblasts were consistently different in biological replicates, indicating that stem cells and osteoblasts have characteristic glycosylation features. Glycosylation features associated with MSCs rather than differentiated cells included high-mannose type N-glycans, linear poly-N-acetyllactosamine chains and α2-3-sialylation. Mesenchymal stem cells expressed SSEA-4 and sialyl Lewis x epitopes. Characteristic glycosylation features that appeared in differentiated osteoblasts included abundant sulfate ester modifications. The results show that glycosylation analysis can be used to evaluate MSC differentiation state.

Recognition of glycoconjugates by Helicobacter pylori: an apparently high-affinity binding of human polyglycosylceramides, a second sialic acid-based specificity*

Helicobacter pylori has been reported to agglutinate erythrocytes and to bind to various other cells in a sialic acid-dependent way. The binding was inhibited by sialyllactose or fetuin and other sialylated glycoproteins. The specificity apparently requires bacterial growth on agar, since we found that it was lost after growth in the nutrient mixture Hams F12. Instead, the bacteria bound with high affinity and in a sialic acid-dependent way to polyglycosylceramides of human erythrocytes, a still incompletely characterized group of complex glycolipids.Bacteria grown in F12 medium were metabolically labelled with35S-methionine and analysed for binding to glycolipids on thin-layer chromatograms and to glycoproteins on blots after electrophoresis, with human erythrocyte glycoconjugates in focus. There was no binding to simpler gangliosides including GM3 or sialylparagloboside, or to a mixture of brain gangliosides. In contrast, polyglycosylceramides of human erythrocyte membranes bound at a pmol level. The activity was eliminated by mild acid treatment, mild periodate oxidation or sialidase hydrolysis. Erythrocyte proteins as well as a range of reference glycoproteins did not bind, except band 3, which was weakly active. However, this activity was resistant to periodate oxidation.These results indicate a second and novel sialic acid-recognizing specificity which is expressed independently of the previously described specificity.

RECOGNITION OF GLYCOCONJUGATES BY HELICOBACTER PYLORI. COMPARISON OF TWO SIALIC ACID-DEPENDENT SPECIFICITIES BASED ON HAEMAGGLUTINATION AND BINDING TO HUMAN ERYTHROCYTE GLYCOCONJUGATES

Helicobacter pylori expresses separate binding characteristics depending on growth conditions, as documented by binding to human erythrocyte glycoconjugates. Cells grown in Hams F12 liquid medium exhibited a selective sialic acid-dependent binding to polyglycosylceramides, PGCs (Miller-Podraza et al. (1996) Glycoconjugate J 13:453–60). There was no binding to traditional sialylated glycoconjugates like shorter-chain gangliosides, glycophorin or fetuin. However, cells grown on Brucella agar bound both to PGCs and other sialylated glycoconjugates. Fetuin was an effective inhibitor of haemagglutination caused by agar-grown cells, but had no or a very weak inhibitory effect on haemagglutination by F12-grown bacteria. PGCs were strong inhibitors in both cases, while asialofetuin was completely ineffective. The results indicate that H. pylori is able to express two separate sialic acid-dependent specificities, one represented by binding to fetuin, as described before, and another represented by a selective binding to PGCs. Abbreviations: PGCs, polyglycosylceramides; TLC, thin-layer chromatography; SDS PAGE, sodium dodecylsulfate polyacrylamide gel electrophoresis; BSA, bovine serum albumin; C, chloroform; M, methanol. The carbohydrate and glycosphingolipid nomenclatures are according to recommendations of IUPAC-IUB Commission on Biochemical Nomenclature (Lipids (1977) 12:455–68; J Biol Chem (1982) 257:3347–51 and J Biol Chem (1987) 262:13–18).

Are globoseries glycosphingolipids SSEA-3 and -4 markers for stem cells derived from human umbilical cord blood?

Umbilical cord blood (UCB) is an efficient and valuable source of hematopoietic stem cells (HSCs) for transplantation. In addition to HSCs it harbours low amounts of mesenchymal stem cells (MSCs). No single marker to identify cord blood-derived stem cells, or to indicate their multipotent phenotype, has been characterized so far. SSEA-3 and -4 are cell surface globoseries glycosphingolipid epitopes that are commonly used as markers for human embryonic stem cells, where SSEA-3 rapidly disappears when the cells start to differentiate. Lately SSEA-3 and -4 have also been observed in MSCs. As there is an ongoing discussion and variation of stem-cell markers between laboratories, we have now comprehensively characterized the expression of these epitopes in both the multipotent stem-cell types derived from UCB. We have performed complementary analysis using gene expression analysis, mass spectrometry and immunochemical methods, including both flow cytometry and immunofluoresence microscopy. SSEA-4, but not SSEA-3, was expressed on MSCs but absent from HSCs. Our findings indicate that SSEA-3 and/or -4 may not be optimal markers for multipotency in the case of stem cells derived from cord blood, as their expression may be altered by cell-culture conditions.

Structural basis for differential receptor binding of cholera and Escherichia coli heat‐labile toxins: influence of heterologous amino acid substitutions in the cholera B‐subunit

The closely related B‐subunits of cholera toxin (CTB) and Escherichia coli heat‐labile enterotoxin (LTB) both bind strongly to GM1 ganglioside receptors but LTB can also bind to additional glycolipids and glycoproteins. A number of mutant CT B‐subunits were generated by substituting CTB amino acids with those at the corresponding positions in LTB. These were used to investigate the influence of specific residues on receptor‐binding specificity. A mutated CTB protein containing the first 25 residues of LTB in combination with LTB residues at positions 94 and 95, bound to the same extent as native LTB to both delipidized rabbit intestinal cell membranes, complex glycosphingolipids (polyglycosylceramides) and neolactotetraosylceramide, but not to non‐GM1 intestinal glycosphingolipids. In contrast, when LTB amino acid substitutions in the 1–25 region were combined with those in the 75–83 region, a binding as strong as that of LTB to intestinal glycosphingolipids was observed. In addition, a mutant LTB with a single Gly‐33→Asp substitution that completely lacked affinity for both GM1 and non‐GM1 glycosphingolipids could still bind to receptors in the intestinal cell membranes and to polyglycosylceramides. We conclude that the extra, non‐GM1 receptors for LTB consist of both sialylated and non‐sialylated glycoconjugates, and that the binding to either class of receptors is influenced by different amino acid residues within the protein.

New method for the isolation of polyglycosylceramides from human erythrocyte membranes

A new procedure was developed for the isolation of long-chain, highly polar glycosphingolipids from human erythrocytes. The membrane material left after extraction of membrane lipids with organic solvents was peracetylated in a mixture of formamide, pyridine and acetic anhydride, and the acetylated products were then extracted with chloroform. The material was fractionated and purified by means of Sephadex LH-20, Sephadex LH-60 and silica-gel chromatography. The final preparations were mixtures of highly polar glycosphingolipids containing from 7 to 31 sugar residues relative to sphingosine. GC-MS analysis of the sugar part of the isolated fractions showed the presence of branched polyglycosyl chains of N-acetyllactosaminyl type. Endo-beta-galactosidase (Bacteroides fragilis) liberated from the deacetylated material two glycosphingolipids, which were identified by fast atom bombardment-mass spectrometry as Hex-Cer and HexNAc-Hex-Hex-Cer with sphingosine and mainly 24 and 22 carbon fatty acids. Endoglycoceramidase (Rhodococcus) degraded polyglycosylceramides to free ceramides and free polysaccharides. The released sugars were fractionated by high-pH ion-exchange chromatography into fractions differing in sialic acid content. The procedure presented in this paper can be used for large and small scale preparations of complex glycosphingolipids. It proved to be especially suitable for screening for polyglycosylceramides in different tissues.

Glycolipid composition of blood group P erythrocytes.

The P blood-group system of human erythrocytes is made up of five phenotypes P1, P2, P, pk, and pk, which can be differentiated with three antibodies: anti-P, anti-P1 and anti-p k (see [1]). These antibodies react with the corresponding antigens at the erythrocyte surface, but the specificity of anti-p k and anti-P~ is sometimes overlapping. The two common P1 and P2 phenotypes are defined by the presence on erythrocytes of the combination of P + Pt antigen and of the P antigen alone, respectively. The three remaining phenotypes are very rare. In p phenotype all three P antigens are missing. Erythrocytes of the pk blood group contain pk and P~ antigens, whereas only the former antigen is present in pk cells. The probable chemistry of P antigens has been elucidated only recently. Cory et al. [2] isolated the determinant of P1 -active glycoprotein from hydatid cyst fluid and characterized it as a-D-galactopyranosyl(1-~4)-D-galactose. Naiki and Marcus [3] found that ceramide trihexoside and globoside, the common cell-membrane glycolipids, displayed significant pk and P blood-group activity, respectively. More recently, Marcus et al. [4] isolated from erythrocyte membranes a P1 -active ceramide pentasaccharide (for structures see table 1). On the basis of these findings Naiki and Marcus [3] predicted that individuals with p phenotype should be unable to synthesize ceramide

Bacterium-host protein-carbohydrate interactions

Publisher Summary This chapter investigates the bacterium–host protein–carbohydrate interactions, and to illustrate this, briefly discusses two cases: recognition of globo glycolipids by uropathogenic Escherichia coli ( E. coli ) and recognition of glycoconjugates by the gastric colonizer Helicobacter pylori ( H. pylori ). Binding of a radiolabeled clinical isolate of E. coli to a long list of globo series and other glycolipids separated on thin-layer chromatography (TLC) plates revealed a binding to all species carrying galabiose in the terminal or internal position. H. pylori appears to have several carbohydrate-binding specificities. In the case of E. coli , FimH on type 1 pili and recognizing Man oligosaccharide are required for the establishment of bladder infection, whereas PapG on P pili and binding the galabiose epitope are requirements for the more serious pyelonephritis to occur. So, these interactions represent two separate niches of urinary tract infection. In the case of H. pylori , a similar map is far from clear.

Unexpected carbohydrate cross-binding by Escherichia coli heat-labile enterotoxin. Recognition of human and rabbit target cell glycoconjugates in comparison with cholera toxin

The bacterial protein enterotoxins, cholera toxin (CT) of Vibrio cholerae and heat-labile toxin (LT) of Escherichia coli, induce diarrhea by enhancing the secretory activity of the small intestine of man and rabbit (animal model). This physiological effect is mediated by toxin binding to a glycolipid receptor, the ganglioside GM1, Gal beta 3GalNAc beta 4(NeuAc alpha 3)GAl beta 4Glc beta 1Cer. However, LT, but not CT, was recently shown by us to bind also to paragloboside, Gal beta 4GlcNAc beta 3Gal beta 4Glc beta 1Cer, identified in the target cells. By molecular modeling of this tetrasaccharide in the known binding site of LT, the saccharide-peptide interaction was shown to be limited to the terminal disaccharide (N-acetyllactosamine). This sequence is expressed in many glycoconjugates, and we have therefore assayed glycolipids and glycoproteins prepared from the target tissues. In addition to paragloboside, receptor activity for LT was detected in glycoproteins of human origin and in polyglycosylceramides of rabbit. However, CT bound only to GM1. Two variants of LT with slightly different sequences, human (hLT) and porcine (pLT), were identical in their binding to target glycoproteins and polyglycosylceramides, but different regarding paragloboside, which was positive for pLT but negative for hLT. This difference is discussed on basis of modeling, taking in view the difference at position 13, with Arg in pLT and His in hLT. Although N-acetyllactosamine is differently recognized in form of paragloboside by the two toxin variants, we speculate that this sequence in human glycoproteins and rabbit polyglycosylceramides is the basis for the common binding. Much work remains, however, to clear up up this unexpected sophistication in target recognition.