A two step model aimed at delivering antisense oligonucleotides in targeted cells J. Toth, a,b I. Boszormenyi, a,b Z.S. Majer, b I. Laczko, c C. Malvy, a M. Hollosi, b and J.-R. Bertrand a, * a CNRS UMR 1582, Institut Gustave Roussy, 39 rue Camille Desmoulins, 94805 Villejuif cedex, France b Department of Organic Chemistry, E otv os University, Post Box 32, H-1518, Budapest 112, Hungary c Biophysical Institute, Biological Research Center, Post Box 521, H-6701, Szeged, Hungary Received 8 March 2002 Abstract To be ecient in vivo antisense oligonucleotides must reach the targeted cells and then cross the cellular membrane. We propose a two step system where the oligonucleotide is ®rst electrostatically bound to a peptide coupled to a ligand of a cellular receptor. A complex is formed which allows the oligonucleotide to be bound to the membrane of the targeted cells. These oligonucleotides are then delivered inside the cells by the subsequent use of a transfection agent. As a reductionist model of peptide coupled to a ligand we have used a lipopeptide and characterized by a ®lter elution assay the stoichiometry between the peptide and the oligonucleotide in the complexes. Using HeLa cultured cells we have shown that addition of these complexes to the cells triggers the oligonucleotide binding to the cell membrane. The subsequent addition of dendrimers allows these antisense oligonucleotides to inhibit a reporter gene inside the cells. Ó 2002 Elsevier Science USA). All rights reserved. Keywords: Antisense oligonucleotide; Lipopeptide; Cell targeting; Delivery The use of antisense oligonucleotides ODNs) as therapeutic agents has been greatly developed in the past few years. Their eciency is related to their ability to recognize their mRNA target into cells and then to block gene expression. The general mode of activity of antisense ODN is mainly dependent on RNase H hy- drolysis of the RNA part of the duplex formed between the antisense ODN and the RNA target. However, one of the major limiting aspects of this strategy is the poor cell penetration for large anionic molecules such as an- tisense ODN. To enhance intracellular delivery of anti- sense ODN, dierent strategies using adsorption to cationic lipids [1] and polymers dendrimer, polylysine), polyethyleneimine PEI), ) [1±3] or direct coupling to peptides to be able to go through the cellular mem- branes [4] or KDEL signal peptide [5] have been devel- oped. Another approach with peptides has been obtained by the adsorption of the antisense ODN with a peptide which was designed to contain two domains, one in charge of the cell penetration and the other re- sponsible for ODN binding [6]. These strategies constitute a major advance, however, one important drawback is the lack of cellular speci®c- ity. This type of delivery is not specially designed to recognize one particular type of cell in an organism. It has to be noticed that, for instance, in cancer most of the studied targets in the antisense ®eld with the exception of the fusion oncogenes) are also present in normal cells. In this study, we propose a two step model for ad- vancing towards a more speci®c cellular approach. Our aim is to use in the ®rst step a peptide to bind antisense ODN to bring them to the cellular membrane of the targeted cells. In the second step an agent able to pierce through the cellular membrane with no speci®city is then used to deliver the antisense ODN into the cell. This can be done in the ®rst step with speci®city using a ligand of a cellular receptor. However to show the fea- sibility of such an approach we have used here in a re- ductionist approach a cationic peptide derived from SV40 T antigen nuclear signal peptide associated to a Biochemical and Biophysical Research Communications 293 2002) 18±22 www.academicpress.com * Corresponding author. Fax: +33-1-42-11-52-45. E-mail address: jrb@igr.fr J.-R. Bertrand). 0006-291X/02/$ - see front matter Ó 2002 Elsevier Science USA). All rights reserved. PII:S0006-291X02)00177-8