Bioscience Reports, Vol. 19, No. 2, 1999

Glycotargeting: Influence of the Sugar Moiety on Both theUptake and the Intracellular Trafficking of Nucleic AcidCarried by Glycosylated Polymers

Michel Monsigny,1,2 Patrick Midoux,1 Roger Mayer,1 and Annie-Claude Roche1

Received April 7, 1999

Nucleic acids (plasmids as well as oligonucleotides) used to specifically express or modulatethe expression of a gene, must reach the cytosol and/or the nucleus. Several systems havebeen developed to increase their uptake and their efficiency. Glycosylated polylysines havebeen shown to specifically help nucleic acids to be taken up in cells expressing a given cellsurface membrane lectin. However, it appeared that the efficiency of the imported nucleicacid was not directly related to the extent of the uptake. Indeed, some glycosylated poly-lysines bearing sugar moities which are poor ligands of the cell surface lectins of a givencell were found to be more efficient than those bearing better sugar ligands. The interpret-ation of this paradoxal result is discussed with regards to the nature of the compartmentallowing the nucleic acid to cross the membrane and to be delivered in the cytosol on theone hand, and to the presence of intracellular lectins on the other hand.

KEY WORDS: Oligonucleotides; gene transfer; routing; membrane lectins;glycoconjugates.

INTRODUCTION

Nucleic acids (oligonucleotides and genes) could be used to modulate selectively theexpression of intracellular proteins and/or to introduce exogenous genes producingnew proteins in cells. Several synthetic systems have been developed with the purposeto efficiently introduce nucleic acids in specific cells. Amongst these synthetic sys-tems, cationic polymers have shown to attract a real interest (see for reviews, Mon-signy et al., 1994a; Wagner et al., 1994; Frese et al., 1994). Polylysine has beenconjugated to various proteins such as asialoorosomucoid (Wu and Wu, 1987) ortransferrin (Wagner et al., 1990), both proteins are known to be specifically recog-nized by membrane receptors, they actively carry their ligands into intracellularendosomes. This approach has been extended to various polylysine derivativesincluding polylysine substituted with simple or complex sugars suitable to efficientlytransfect hepatoma cells, (Plank et al., 1992; Midoux et al., 1993; Perales et al., 1994;

1Vectorologie et trafic intracellulaire, Glycobiologie, CBM-CNRS, Rue Charles Sadron, 45071 Orleanscedex 02, France.

2To whom correspondence should be addressed.125

0144-8463/99/0400-0125$16.00/0 c 1999 Plenum Publishing Corporation

126 Monsigny, Midoux, Mayer, and Roche

Wadhwa and Rice, 1995; Wadhwa et al., 1995), macrophages (Erbacher et al., 1996;Ferkol et al., 1996) and then many other cell types.

By using glycosylated polylysine to deliver oligonucleotides to lung tumor cells(Stewart et al., 1996) and genes to airway epithelial cells, (Kollen et al., 1996; Fajacet al., 1999), as well as smooth muscle cells (Boutin et al., 1999), it clearly appearedthat the role of sugar moieties was not limited to the binding and uptake of thecomplexes but the nature of the sugar had a marked effect on the activity elicitedby the delivered nucleic acids, most probably related to the intracellular traffic ofthe complexes.

Glycosylated Polymers

Natural glycoproteins and synthetic glycoproteins, called neoglycoproteins,have been used to evidence cell surface endogenous lectins and to study their sugar-dependent uptake in various cells (see reviews by Monsigny et al., 1984, 1994a, b;Lee and Lee, 1994a, b, c; Gabius and Gabius, 1997). Neoglycoproteins have alsobeen used to deliver drugs and oligonucleotides into cells in a sugar dependent man-ner (see for reviews Monsigny et al., 1994a, b; Molema and Meijer, 1994; Wadhwaand Rice, 1995). In all cases the drugs or the oligonucleotides were linked througha covalent bond. In order to obtain protein-free carriers, polylysines were eitherglycosylated by adding sugar derivatives such as glycosidophenylisothiocyanateand/or gluconoylated by reaction with gluconolactone on polylysine made of tensto hundreds of amino acid residues (Derrien et al., 1989; see for reviews Monsignyet al., 1994a, b).

In both cases the global charge of the substituted polylysine was reduced,depending on the degree of e-amino group substitution. Such glycosylated poly-lysines were found to be suitable to prepare polyplexes by mixing glycosylated poly-lysine solutions and DNA solutions (Midoux et al., 1993). Some of thoseglycosylated polylysines were also found to be suitable to carry and deliver oligonu-cleotides into cells (Stewart et al., 1996).

The glycosylated polylysines we used were made by allowing glycosides bearingan activated aglycone to react on some of the amino groups of a polylysine dissolvedin an organic solvent such as dimethylformamide or in a buffer. For instance, lacto-sylated polylysine was prepared by adding galactopyranosido B-4 glucopyranosidoB phenylisothiocyanate to a polylysine solution, leading to a phenylthioureido conju-gate (Fig. 1).

We prepared a panel of glycosylated polymers containing alpha or beta ano-meric derivatives of common sugars: galactose, glucose, mannose, N-acetylglucos-amine, N-acetylgalactosamine, L-fucose, as well as of several disaccharides.Glycosylated polylysines containing about 485 residues were selected to carry oligo-nucleotides and glycosylated polylysines containing 190 residues to carry plasmids.

Sugar Dependent Oligonucleotide Targeting

Oligonucleotides have been developed as a new generation of putative thera-peutic agents (Helene and Toulme, 1990; Stein and Cheng, 1993). Oligonucleotides

Glycotargeting 127

Fig. 1. Structure of glycosylated polylysine. R is either a lactosyl-B-phe-nylthio-ureido moiety corresponding to the side chains substituted witha sugar derivative, or NH3+, corresponding to the unsubstituted lysylresidue.

are anionic oligomers, which are not able to cross a lipid bilayer. Therefore, whenthey reach the surface of a cell, they may be taken up by fluid or by adsorptiveendocytosis and they have then to leave these compartments to reach the cytosoland finally freely diffuse through nuclear pore to the nucleus.

Free oligonucleotides are usually taken up very slowly by cultured cells; theiruptake may be enhanced when they are associated with lipocations, liposomes orpolymeric carriers. In an attempt to make use of cell surface lectins to increase theuptake of oligonucleotides, we prepared oligonucleotide-neoglycoprotein conjugates(Bonfils et al., 1992; Sdiqui et al., 1995a) while Hangeland et al. (1995) used neogly-copeptides and Wu and Wu (1992) used asialoorosomucoid-polylysine conjugates.

128 Monsigny, Midoux, Mayer, and Roche

However, because of the amount of work required to prepare such conjugates, wethen started to use oligonucleotide-glycosylated polylysine complexes (Stewart et al.,1996). We selected antisense oligonucleotides complementary to the 3' non-codingregion of intercellular adhesion molecule-1 (ICAM-1) in A549 human non-smalllung carcinoma cells (according to Chiang et al., 1991). Various polylysines wereused, those with a mean degree of polymerization of 455 gave the best results inboth their ability to form complexes with oligonucleotides and their ability to inhibitthe expression of the related ICAM-1 protein. The striking point was that there wasno direct relation between uptake and efficiency. Indeed complexes prepared withpolylysine (dp 455) free or substituted with 100 sugar or gluconoyl residues led tosimilar uptake of the oligonucleotides bearing a fluorescent or a radioactive tag. Theinhibitory activity was not detected up to 2 mM when oligonucleotides were carriedby unsubstituted polylysine or by gluconoylated polylysine, but was quite efficient(more than 50% inhibition at 500 nM) when oligonucleotides were carried by a fuco-sylated polylysine. Furthermore, while the uptake was independent of the numberof fucose units linked to polylysine, the antisense efficiency increased when the num-ber of fucose units increased up to 100. The biological activity of the carried oligo-nucleotides was sequence related in agreement with the data of Chiang et al. (1991)who were using lipocations (Bennet et al., 1992), to help oligonucleotides to be takenup by the same cells.

The interpretation of the above mentioned results was based on the followingconsiderations:

—The fucosylated polylysine had been selected on the basis of preliminaryresults working with fluorescein labeled neoglycoproteins. A549 cells take upmore effectively neoglycoproteins bearing either a-L-fucose, a-mannose ora-6-phosphomannose units than sugar-free bovine serum albumin or neogly-coproteins bearing a-glucose. Therefore, in the case of neoglycoproteins, thesugar was involved in the uptake.

—Conversely, oligonucleotides, which were poorly taken up (10pmol/mg pro-tein) when they were free, were efficiently taken up (about 300 pmol/mg pro-tein) when they were associated with either sugar-free polylysine orfucosylated polylysine. Therefore, the cell uptake of the labeled oligonucleot-ides is not related to the presence of the sugar residues.

In conclusion, we postulate that oligonucleotides could be routed to differentintracellular compartments depending on the nature of the carrier, i.e., glycosylatedor non-glycosylated.

Sugar-Dependent Plasmid Targeting

Cationic polymers interact with polyanions to give either aggregates or com-plexes. Under appropriate conditions with regard to the concentration of the poly-mers, to the ionic strength and to the osmolarity, the mixing of a solution ofpolylysine and of a solution containing a plasmid led to the formation of complexescalled polyplexes. These polyplexes appeared to be compact particles with a diameteras low as 50 nm when the plasmid size was about 5000 base pairs and polylysine

Glycotargeting 129

contained about 200 aminoacids. Acylation of the amino groups of polylysine byglycosylation and/or gluconoylation led to a polylysine derivative which still inter-acts with a plasmid and gives compact complexes (Plank et al., 1992; Midoux et al.,1993; Wadhwa et al., 1995; Erbacher et al., 1995, 1996, 1997) called glycoplexes.These glycoplexes are usually more active than the classical polyplexes because theycan dissociate under conditions less drastic than polyplexes, specially in the presenceof polyamines. The enhanced activity of glycoplexes is also linked to the nature ofthe sugar born by the glycosylated polylysine called glycofectin. Indeed, a glycofectincontaining b-galactosyl residues was found to be 100 times more efficient than aglycofectin containing a-mannosyl residues in transfecting HepG2 cells, a humanhepatoma cell line known to express on its surface the asialoglycoprotein receptoralso known as Ashwell and Morell lectin (Ashwell and Harford, 1982); this lectin,which recognizes components with b-galactose units in a terminal non-reducing pos-ition but does not recognize components with a-mannose units, efficiently mediatesthe endocytosis of its ligands. However, the enhanced efficiency of glycoplexes incomparison with the efficiency of polyplexes is not limited at the step of internaliz-ation, it seems also related to the intracellular traffic.

Indeed, Boutin et al. (1999) clearly showed that the uptake of glycoplexes byrabbit vascular smooth muscle cells was limited when the glycofectin contained N-acetyl-b-galactosamine but very efficient when the glycofectin contained lactose, theefficiency ratio between those two glycoplexes was close to one tenth. Conversely,the expression of the luciferase gene present in the glycoplex plasmid was roughly100 times higher when glycoplexes were prepared with glycofectin containing N-acetyl-b-galactosamine than with glycoplexes prepared with the glycofectin contain-ing lactose. The absence of direct relation between uptake and expression efficiencywas also found with other sugars, for instance glycofectins containing either b-galac-tose or N-acetyl-a-galactosamine residues were taken up with an equal efficiencywhile that containing N-acetyl-a-galactosamine led to expression ten times higherthan that containing b-galactose units.

This observation is not limited to the rabbit vascular smooth muscle cells;indeed, differences have also been observed with transfection of other cells withglycoplexes. Recently, Fajac et al. (1999) showed that human airway epithelial cellswere efficiently transfected with glycoplexes prepared with a glycofectin containingfor instance N-acetyl-b-glucosamine residues, while these glycoplexes were verypoorly taken up by these cells. Conversely, glycoplexes made with a glycofectincontaining a-mannose were very efficiently taken up by these cells but led to almostno expression at all of the transfected luciferase gene.

PERSPECTIVES

An efficient transfection using a non-viral approach is dependent on variousparameters. Amongst them, some are well identified, such as (i) an efficient protec-tion of the plasmid as long as it is not functional in the cell nucleus, (ii) an efficientand selective uptake by the relevant cells, (iii) an efficient protection of the complexagainst delivery to lysosomes, (iv) an efficient delivery to the cytosol allowing thecomplex/or the plasmid to cross the membrane from the lumen of the organelles

130 Monsigny, Midoux, Mayer, and Roche

where the complex or the plasmid finally reached, and (v) when cells are not dividing,an efficient transfer of the complex or of the plasmid to the nucleus, where theplasmid gene may be transcribed. The sugars borne by glycofectins may be quitehelpful to make very efficient and very selective the step of binding to the surface ofthe targeted cells and the step of the uptake by these cells. The use of histidylatedpolylysine may be helpful to cross the endosomal membrane (Midoux and Mon-signy, 1999).

New glycofectins bearing oligosaccharides are being developed, making use ofthe novel way of transforming a complex oligosaccharide into a glycosynthon, witha very high yield (almost quantitative) (Sdiqui et al., 1995b, Quetard et al., 1997,1998 see for review Monsigny et al., 1998); such glycosynthons with a very highbinding constant and selectivity with regards to cell surface lectins are easily linkedto polylysine, leading to glycofectins. In addition, as shown above, the sugar borneby glycofectins may be quite helpful to enhance the transfection efficiency of glyco-coplexes by guiding them intracellularly. Indeed, in addition to cell surface lectinssuch as the galactose specific lectin of parenchymal liver cells and as the 6-phospho-mannose specific lectin which recycle from the cell surface to endosomes and toprelysosomes respectively, a lectin (MR60/ERGIC53/P58) recycling from the endo-plasmic reticulum to the cis-Golgi on the one hand, and lectins recycling from cyto-sol to nucleus on the other hand have been evidenced (for a review see Roche andMonsigny, 1996).

The intracellular lectins could help glycoplexes to avoid their delivery to prely-sosomes and lysosomes and to reach another compartment from which their exit tocytosol could be facilitated. Then, soluble lectins present in the cytosol could helpglycoplexes to reach the nucleus by a mechanism independent of the classical pep-tidic nuclear localization signal as suggested by Duverger et al. (1995, 1996) in thecase of the entry of neoglycoproteins in the nucleus. Both mechanisms could explainthe enhancement of glycoplex efficiency related to that of polyplexes, but their directinvolvement has yet to be demonstrated.

ACKNOWLEDGEMENTS

This work was supported by grants from the Agence Nationale de Recherchessur le SIDA and from the Association de Recherche sur le Cancer (ARC 6132).M.M. is Professor at the Universite of Orleans, R.M. is Research Director CNRS,P.M. and A.C.R. are Research Directors INSERM.

REFERENCES

Ashwell, G. and Harford, J. (1982) Carbohydrate-specific receptor of the liver. Anna. Rev. Biochem.51:531–554.

Bennett, C. F., Chiang, M. Y., Chan, H., Shoemaker, J. E. S., and Mirabelli, C. K. (1992) Cationic lipidsenhance cellular uptake and activity of phosphorothioate antisense oligonucleotides. Mol. Pharma-col. 41:1023–1033.

Bonfils, E., Depierreux, C., Midoux, P., Thuong, N. T., Monsigny, M., and Roche, A. C. (1992) Drugtargeting: synthesis and endocytosis of oligonucleotide-neoglycoprotein conjugates. Nucleic AcidsRes. 20:4621–4629.

Glycotargeting 131

Boutin, V., Legrand, A., Mayer, R., Nachtigal, M., Monsigny, M., and Midoux, P. (1999) Glycofection:the ubiquitous roles of sugar bound on glycoplexes. Drug Delivery 6:45–50.

Chiang, M., Chan, H., Zounes, M. A., Freier, S., Lima, W. F., and Bennett, C. F. (1991) Antisenseoligonucleotides inhibit intercellular adhesion molecule-1 expression by two distinct mechanisms. J.Biol. Chem. 266:18162–18171.

Derrien, D., Midoux, P., Petit, C., Negre, E., Mayer, R., Monsigny, M., and Roche, A. C. (1989) Mura-myl dipeptide bound to poly-L-lysine substituted with mannose and gluconoyl residues as macroph-age activators. Glycoconjugate J. 6:241–255.

Duverger, E., Pellerin-Mendes, C., Mayer, R., Roche, A. C., and Monsigny, M. (1995) Nuclear importof glycoconjugates is distinct from the classical NLS pathway. J. Cell Sci. 108:1325–1332.

Duverger, E., Roche, A. C., and Monsigny, M. (1996) N-acetylglucosamine-dependent nuclear import ofneoglycoproteins. Glycobiology 6:381–386.

Erbacher, P., Roche, A. C., Monsigny, M., and Midoux, P. (1995) Glycosylated polylysine/DNA com-plexes: gene transfer efficiency in relation with the size and the sugar substitution level of glycosylatedpolylysines and with the plasmid size. Bioconjugate Chem. 6, 401–410.

Erbacher, P., Bousser, M. T., Raimond, J., Monsigny, M., Midoux, P., and Roche, A. C. (1996) Genetransfer by DNA/glycosylated polylysine complexes into human blood-monocyte derived macro-phages. Hum. Gene Ther. 7:721–729.

Erbacher, P., Roche, A. C., Monsigny, M., and Midoux, P. (1997) The reduction of the positive chargesof polylysine by partial gluconoylation increases the transfection efficiency of DNA/polylysine com-plexes. Biochim. Biophys. Acta 1324:27–36.

Fajac, I., Briand, P., Monsigny, M., and Midoux, P. (1999) Sugar-mediated uptake of glycosylated poly-lysines and gene transfer into normal and cystic fibrosis airway epithelial cells. Human Gene Ther.10:395–406.

Ferkol, T., Perales, J. C., Mularo, F., and Hanso, R. W. (1996) Receptor-mediated gene transfer intomacrophages. Proc. Natl. Acad. Sci., USA 93:101–105.

Frese, J., Wu, C. H., and Wu, G. Y. (1994) Targeting of genes to the liver with glycoprotein carrier. Adv.Drug Delivery Rev. 14:137–152.

Gabius, H. J. and Gabius, S. (1997) Glycosciences. Status and Perspectives. Chapman and Hall,Weinheim, p. 631.

Hangeland, J. J., Levis, J. T., Lee, Y. C., Ts’o, O. P. (1995) Cell-type specific and ligand specific enhance-ment of cellular uptake of oligodeoxynucleoside methylphosphonates covalently linked with a neo-glycopeptide, YEE(ah-GalNAc)3. Bioconjugate Chem. 6:695–701.

Helene, C. and Toulme, J. J. (1990) Specific regulation of gene expression by antisense and anti-genenucleic acids. Biochem. Biophys. Acta 1049:99–125.

Kollen, W. J. W. et al. (1996) Gluconoylated and glycosylated polylysine as vector for gene transfer intohuman cystic fibrosis airway epithelial cells. Hum. Gene Ther. 7:1577–1586.

Lee, Y. C. and Lee, R. T. (1994a) Neoglycoconjugates. Part A. Synthesis. Meth. Enzymol. Vol. 242, p. 238.Lee, Y. C. and Lee, R. T. (1994b) Neoglycoconjugates. Part B. Biochemical application. Meth. Enzymol.

Vol. 247, p. 450.Lee, Y. C. and Lee, R. T. (1994c) Neoglycoconjugates: preparation and applications. Academic Press, San

Diego, p. 549.Midoux, P. et al. (1993) Specific gene transfer mediated by lactosylated poly-L-lysine into hepatoma cells.

Nucleic Acids Res. 21:871–878.Midoux, P. and Monsigny, M. (1999) Efficient gene transfer by histidylated polylysine/pDNA complexes.

Bioconjugate Chem. (in press.)Molema, G. and Meijer, D. K. (1994) Targeting of drugs to various blood cell types using (neo-)glyco-

proteins, antibodies and other protein carriers. Adv. Drug Delivery Rev. 14:25–50.Monsigny, M. et al. (1998) Glycotargeting: the preparation of glyco-amino acids and derivatives from

unprotected reducing sugars. Biochimie 80:99–108.Monsigny, M., Roche, A. C., and Midoux, P. (1984) Uptake of neoglycoproteins via membrane lectin(s)

of L 121O cells evidenced by quantitative flow cytofluorometry and drug targeting. Biol. Cell 51: 187–196.

Monsigny, M., Roche, A. C., Midoux, P., and Mayer, R. (1994a) Glycoconjugates as carriers for specificdelivery of therapeutic drugs and genes. Adv. Drug Delivery Rev. 14:1–24.

132 Monsigny, Midoux, Mayer, and Roche

Monsigny, M., Roche, A. C., Midoux, P., and Mayer, R. (1994b) Sugar specific delivery of drugs, oligo-nucleotides and genes. In: Targeting of drugs, 4: Advances in system constructs. NATO AS1 SeriesA: Life Sciences, Vol. 273. G. Gregoriadis, B. McCormack, and G. Poste (eds.). Plenum Press, NewYork, pp. 31–50.

Perales, J. C., Ferkol, T., Beegen, H., Ratnoff, O. D., and Hanson, R. W. (1994) Gene transfer in vivo:sustained expression and regulation of genes introduced into liver by receptor-targeted uptake. Proc.Nat. Acad. Sci., USA 91:4086–4090.

Plank, C., Zatloukal, K., Cotten, M., Mechtler, K., and Wagner, E. (1992) Gene transfer into hepatocytesusing asialoglycoprotein receptor mediated endocytosis of DNA complexed with an artificial tetra-antennary galactose ligand. Bioconjugate Chem. 3:533–539.

Quetard, C. et al. (1998) Novel glycosynthons for glycoconjugate preparation: oligosaccharyl-pyrogluta-myl-anilide derivatives. Bioconjugate Chem. 9: 268–276.

Quetard, C., Normand-Sdiqui, N., Mayer, R., Strecker, G., Roche, A. C., and Monsigny, M. (1997)Simple synthesis of novel glycosynthons: oligosylpyroglutamyl derivatives. Carbohydr. Lett. 2:415–422.

Roche, A. C. and Monsigny, M. (1996) Trafficking of endogenous glycoproteins mediated by intracellularlectins: facts and hypotheses. Chemtracts Biochem. Mol. Biol. 6: 188–201.

Sdiqui, N., Arar, K., Midoux, P., Mayer, R., Monsigny, M., and Roche, A. C. (1995a) Inhibition ofHuman mammary cell line proliferation by membrane lectin-mediated uptake of Ha-ras antisenseoligodeoxynucleotide. Drug Delivery 2:63–72.

Sdiqui, N., Roche, A. C., Mayer, R., and Monsigny, M. (1995b) New synthesis of Glycopeptides. Car-bohydr. Lett. 1:269–275.

Stein, C. A. and Cheng, Y. C. (1993) Antisense oligonucleotides as therapeutic agents—is the bullet reallymagical? Sciences 261:1004–1012.

Stewart, A. J., Pichon, C., Meunier, L., Midoux, P., Monsigny, M., and Roche, A. C. (1996) Enhancedbiological activity of antisense oligonucleotides complexed with glycosylated poly-L-lysine. Mol.Pharmacol. 50:1487–1494.

Wadhwa, M. S., Knoell, D. L., Young, A. P., and Rice, K. G. (1995) Targeted gene delivery with a lowmolecular weight glycopeptide carrier. Bioconjugate Chem. 6:283–291.

Wadhwa, M. S. and Rice, K. G. (1995) Receptor mediated glycotargeting. J. Drug Target. 3:111–127.Wagner, E., Zenke, M., Cotten, M., Beug, H., and Birnstiel, M. L. (1990) Transferrin-polycation conju-

gates as carriers for DNA uptake into cells. Proc. Natl. Acad. Sci., USA 87:3410–3414.Wagner, E., Curiel, D., and Cotten, M. (1994) Delivery of drugs, proteins and genes into cells using

transferrin as a ligand for receptor-mediated endocytosis. Adv. Drug Delivery Rev. 14:113–135.Wu, G. Y. and Wu, C. H. (1987) Receptor-mediated in vitro gene transformation by a soluble DNA

carrier system. J. Biol. Chem. 262: 4429–4432.Wu, G. Y. and Wu, C. H. (1992) Specific inhibition of Hepatitis B viral gene expression in vitro by target

antisense oligonucleotides. J. Biol. Chem. 267:12436–12439.