Receptor-mediated endocytosis of phosphodiester oligonucleotides in the HepG2 cell line: evidence for non-conventional intracellular trafficking.

Having identified an oligonucleotide (ON) receptor in the HepG2 cell line, we have re-examined here the kinetics of ON uptake, subcellular distribution and intracellular localisation in these cells, at concentrations relevant for the study of a receptor-dependent process. Kinetic parameters of ON endocytosis were comparable with those of the receptor-mediated endocytosis tracer, transferrin (uptake equilibrium, saturation with concentration, specific competition and rapid efflux) and were clearly distinct from those of fluid-phase endocytosis. By analytical subcellular fractionation, particulate ON showed a bimodal distribution after 2 h of uptake, with a low-density peak superimposed on the distribution of endosomes, and a high-density peak overlapping lysosomes. After an overnight chase, only the high-density peak remained, but it could be dissociated from lysosomes, based on its refractoriness to displacement upon chloroquine-induced swelling. After 2 h of uptake at 300 nM ON-Alexa, a punctate pattern was resolved, by confocal microscopy, from those of transferrin, of a fluid-phase tracer, and of vital staining of lysosomes by LysoTracker. At 3 microM ON-Alexa, its pattern largely overlapped with the fluid-phase tracer and LysoTracker. Taken together, these data suggest that ON may be internalised at low concentrations by receptor-mediated endocytosis into unique endosomes, then to dense structures that are distinct from lysosomes. The nature of these two compartments and their significance for ON effect deserve further investigation.

Glyco-dependent nuclear import of glycoproteins, glycoplexes and glycosylated plasmids.

This short review deals with some properties of nuclear sugar-binding proteins also called nuclear lectins, the sugar-dependent nuclear import of neoglycoproteins and the attempts of using this pathway to enhance the nuclear import of plasmids in order to hopefully increase the expression of transferred genes.

Transcription of plasmid DNA: influence of plasmid DNA/polyethylenimine complex formation.

Polyethylenimine (PEI) is one of the most potent non-viral vectors. We have developed a lactosylated PEI (Lac-PEI) to enhance cell-specific transfection and have shown that Lac-PEI is more efficient than unsubstituted PEI for gene transfer into immortalized cystic fibrosis airway epithelial SigmaCFTE29o-cells. As both intact PEI/plasmid and Lac-PEI/plasmid complexes are found in the cell nucleus, we have investigated the transcription efficiency of the plasmid complexed with PEI or Lac-PEI, according to the polymer nitrogen/DNA phosphate (N/P) ratio (from 0 to 20). The initiation of transgene transcription was analyzed in an acellular nuclease S1 transcription assay. For both PEI and Lac-PEI complexes, transcription efficiency varied with the N/P ratio of the complexes. Transcription inhibition was observed when plasmid DNA was either loosely (N/P<5) or tightly condensed (N/P>15). For an N/P ratio of 5 and up to 15, transcription of the complexed plasmid was as efficient as that of the free plasmid. Similar results were observed when gene expression was studied after nuclear microinjection of the complexes into SigmaCFTE29o-cells. Our study shows that condensation of DNA influences the accessibility of the plasmid to the transcription machinery. Interestingly, the charge ratios that allow the most efficient transcription are those usually known to be the most efficient for gene transfer in vitro and in vivo.

Carbohydrate-lectin interactions assessed by surface plasmon resonance.

The specificity, the strength, the kinetics and some thermodynamic parameters of sugar-protein interactions are easily assessed by surface plasmon resonance (SPR). This paper intends to present both theoretical and practical considerations. This includes: the principle of SPR, the analysis according to Langmuir and Scatchard, the problems linked either to mass transport limitation, to the heterogeneity of the immobilized ligand density or to the non-linearity due to cluster effects. The non-linearity may be taken into account by either one of two ways: the fractal or the Sips approaches that have been developed with the aim of linearizing the data. In addition, selected data obtained by using either immobilized carbohydrates or immobilized lectins are summarized. The SPR has also been found useful to collect information concerning oligosaccharide structure as well as lectin-sugar specificity and to develop new tools with medical applications. Finally, a series of practical considerations are gathered in the hope of avoiding some of the common pitfalls arising in sugar-lectin interaction studies based on the use of SPR.

Chromatographic separation of Lewis a and Lewis x oligosaccharides as glycosynthons.

Lewis a and Lewis x oligosaccharides Gal beta 3(Fuc alpha 4)GlcNAc beta 3Gal beta 4Glc and Gal beta 4(Fuc alpha 3)GlcNAc beta 3Gal beta 4Glc are easily isolated as a mixture from biological fluids, including human milk. However, because they behave almost identically in most chromatographic systems, it is difficult to have each of them as a pure compound. Incidentally, we found that they were easily separated by HPLC as glycosynthons [Gal beta 3(Fuc alpha 4)GlcNAc beta 3Gal beta 4Glc-Glp-beta Ala-OBzl and Gal beta 4(Fuc alpha 3)GlcNAc beta 3Gal beta 4Glc-Glp-beta Ala-OBzl] after substitution of the terminal reducing sugar by a short peptide (pyroglutamyl-beta alanyl-O-benzyl ester) in a one-pot two-step reaction (Carbohydr. Lett. 1 (1995) 269; Bioconjug. Chem. 9 (1998) 268). Such glycosynthons are easily either converted back to native Lewis a and Lewis x oligosaccharides upon hydrazinolysis or used to synthesize glycoconjugates, such as glycoclusters, glycopeptides, glycooligonucleotides, glycosylated polymers or glycosylated matrices for therapeutic or analytical purposes.

Carbohydrate-lectin interactions assayed by SPR.

Surface plasmon resonance is a valuable tool to determine the affinity between glycoconjugates and sugar-binding proteins such as plant and animal lectins. The main interest of using such an approach is that neither the lectins - which are proteins - nor their ligands - natural compounds such as glycoproteins, oligosaccharides, polysaccharides, or synthetic glycoconjugates such as glycoclusters or neoglycoproteins - require any tag. Because lectins bear several binding sites, they behave like immunoglobulin eliciting avidity phenomena. This peculiarity may lead to erroneous results if special conditions are not applied. We obtained best and reproducible results when the lectin was immobilized and its ligands were used as soluble analytes. With heterogeneous glycoconjugates such as neoglycoproteins (which are heterogeneous in terms of nature, number, and position of sugar residues) or a mixture of oligosaccharides, the data may be more accurately gathered by using the Sips approach, which has been used to determine mean binding constants of polyclonal antibodies. With small analytes such as oligosaccharides, we found it convenient to determine binding constants by using an inhibitory approach: a neoglycoprotein (M (r) = approximately 80,000) was allowed to bind to the immobilized lectin and small oligosaccharides were used as inhibitors. With larger glycoconjugates such as peptides substituted with glycoclusters, direct binding measurements gave accurate results. Because of the availability of low-cost simple sugars (mono- or disaccharides) it is very convenient to use large concentrations of such carbohydrates to clean the sensor chips instead of more drastic cleaning solutions such as acids or alkali, in such a way that the immobilized lectin is stable for many experiments.

In vivo gene delivery in the mouse lung with lactosylated polyethylenimine, questioning the relevance of in vitro experiments.

Polyethylenimine (PEI) is an efficient vector for in vitro and in vivo gene transfer into respiratory cells. Glycosylated PEIs were shown to enhance in vitro gene transfer by favoring the complex entry into the airway cells. The aim of our study was to evaluate the in vivo efficiency of gene transfer mediated by glycosylated PEIs in the mouse lung and to determine the transfected cell type and the intracellular trafficking of the complexes. Upon nasal instillation in mice of complexes made with various glycosylated PEIs, a high luciferase activity was observed while the green fluorescent protein (GFP) expression was similar for all the vectors tested with few cells expressing GFP. Complexes made with lactosylated PEI were then labeled and their localization studied by confocal microscopy. In the lungs, large numbers of complexes were taken up by epithelial cell which were mostly alveolar cells. In the airways, complex uptake varied greatly, depending on the area observed. Eight hours upon nasal instillation and in contrast with the in vitro situation, a dissociation between the plasmid DNA and the lactosylated PEI was usually observed, leading to the plasmid mostly localized in lysosomes and the Lac-PEI localized in the nucleus. These results emphasize the need to engineer a plasmid able by itself to overcome the nuclear barrier and to quickly move to in vivo experiments to select the best carrier.

Cytoskeletal involvement in the cellular trafficking of plasmid/PEI derivative complexes.

We have studied the cytoskeletal involvement in the cellular trafficking of complexes made with plasmid/PEI or plasmid/lactosylated PEI in cystic fibrosis airway epithelial cells (SigmaCFTE29o- cells). Complexes were incubated in the presence of cytoskeletal inhibitors, and the number of transfected cells was determined by flow cytometry. Complexes were also generated with fluorescein-labeled PEI derivatives and the cell fluorescence intensity was determined by flow cytometry. In the presence of cytochalasin D to depolymerize actin filaments or nocodazole to disrupt microtubules, gene transfer efficiency with both PEI derivatives was decreased by 90%. The uptake of fluoresceinylated complexes studied by flow cytometry was decreased by 50% in the presence of cytochalasin D for both types of complexes (p<0.005) and unchanged in the presence of nocodazole. When cytoskeletal inhibitors were added to the cell culture after the complex uptake had occurred, gene transfer efficiency was decreased by 75% and 50% in the presence of nocodazole and cytochalasin D, respectively. Upon nocodazole-microtubule network disruption, the lysosomal localization of complexes was reduced, as assessed by confocal microscopy. Our results show a major cytoskeletal involvement in the cellular trafficking of complexes made with both PEI derivatives: actin filaments mainly in complex uptake, and microtubules in the trafficking of complexes towards the nucleus, probably through guided transport of complex-containing endosomal vesicles.

Synthesis of glycocluster-tumor antigenic peptide conjugates for dendritic cell targeting.

The use of dendritic cells (DC) for the development of therapeutic cancer vaccines is attractive because of their unique ability to present tumor epitopes via the MHC class I pathway to induce cytotoxic CD8+ T lymphocyte responses. C-Type membrane lectins, DC-SIGN and the mannose receptor (MR), present on the DC surface, recognize oligosaccharides containing mannose and/or fucose and mediate sugar-specific endocytosis of synthetic oligolysine-based glycoclusters. We therefore asked whether a glycotargeting approach could be used to induce uptake and presentation of tumor antigens by DC. To this end, we designed and synthesized glycocluster conjugates containing a CD8+ epitope of the Melan-A/Mart-1 melanoma antigen. These glycocluster-Melan-A conjugates were obtained by coupling glycosynthons: oligosaccharyl-pyroglutamyl-beta-alanine derivatives containing either disaccharides, a dimannoside (Manalpha-6Man) or lactoside, or a Lewis oligosaccharide, to Melan-A 16-40 peptide comprising the 26-35 HLA-A2 restricted T cell epitope, extended with an oligolysine stretch at the C-terminal end. We showed by confocal microscopy and flow cytometry that fluorescent-labeled Melan-A glycoclusters containing either dimannoside or Lewis oligosaccharide were taken up by DC and concentrated in acidic vesicles; conversely lactoside glycopeptides were not at all taken up. Furthermore, using surface plasmon resonance, we showed that dimannoside and Lewis-Melan-A conjugates bind MR and DC-SIGN with high affinity. DC loaded with these conjugates, but not with the lactose-Melan-A conjugate, led to an efficient presentation of the Melan-A epitope eliciting a CD8+ T-lymphocyte response. These data suggest that synthetically designed glycocluster-tumor antigen conjugates may induce antigen cross-presentation by DC and represent a promising tool for the development of tumor vaccines.

Which mechanism for nuclear import of plasmid DNA complexed with polyethylenimine derivatives?

BACKGROUND: To investigate the nuclear import mechanism of plasmid/polyethylenimine (PEI) derivative complexes and the putative nuclear targeting of therapeutic genes by the use of oligosaccharides, we have studied the nuclear import of plasmid DNA complexed either with PEI or with lactosylated PEI (Lac-PEI) in cystic fibrosis human airway epithelial cells ( summation operatorCFTE29o- cells). METHODS AND RESULTS: Cells were synchronized by a double-thymidine block protocol and gene transfer efficiency was evaluated: Lac-PEI- and PEI-mediated gene transfer was greatly increased when cells have undergone mitosis during the course of transfection. However, both types of complexes were able to transfect some growth-arrested cells. When the nuclear import of plasmid/Lac-PEI or plasmid/unsubstituted PEI complexes was studied in digitonin-permeabilized cells, the nuclear uptake of both types of complexes did not follow the classic pathway of nuclear localization sequence (NLS)-containing proteins and lactose residues did not act as a nuclear localization signal. CONCLUSIONS: Our results show that for complexes made with PEI derivatives, the major route for plasmid DNA nuclear entry is a passive nuclear importation during mitosis when the nuclear membrane temporarily breaks down. However, albeit to a lesser extent as that observed in dividing cells, a plasmid DNA importation also occurs in nondividing cells by a yet unknown mechanism.

Potocytosis and cellular exit of complexes as cellular pathways for gene delivery by polycations.

BACKGROUND: Although polycations are among the most efficient nonviral vectors for gene transfer, the gene expression they allow is still too low for in vivo applications. To engineer more potent polycationic vectors, the factors governing the intracellular trafficking of a plasmid complexed with current polycations need to be identified. METHODS AND RESULTS: The trafficking of plasmid DNA complexed to glycosylated polylysines or polyethylenimine (PEI) derivatives was studied by electron microscopy of human airway epithelial cells. The cellular processing of complexes varied with their size and the polycation derivative used: large complexes (> 200 nm) made with all polycationic vectors studied were internalized by macropinocytosis. In contrast, intermediate (100-200 nm) ligand-coupled polylysine and PEI complexes primarily entered through clathrin-coated pits. Complexes were then found in endosomal vesicles, accumulated in lysosomes or vesicles near the nucleus and their nuclear entry was limited. For the population of small complexes (< or = 100 nm) obtained with PEI derivatives, they were internalized through caveolae and pursued a traffic pattern of potocytosis to the endoplasmic reticulum where their fate remains unclear. Finally, some complexes exited the cells either by regurgitation when PEI derivatives were used or through an exosome-like pathway for glycosylated-polylysine complexes. CONCLUSIONS: The different pathways of complex trafficking observed in relation with complex size imply the development and study of vectors forming complexes with definite size. Moreover, the complex exit we describe may contribute to the well-established short-term efficiency of gene transfer based on synthetic vectors. It favors the engineering of vectors allowing repeated treatment.

Oligolysine-based saccharide clusters: synthesis and specificity.

In search of specific and highly selective sugar clusters for cell receptors, such as membrane lectins, various disaccharides were coupled to small peptide cores through an amide bond. In a first step, the reducing disaccharides, i.e. lactose and three different dimannoses, were converted into glycosyl-pyroglutamyl-beta-alanine derivatives. The free carboxylic group of these conjugates was then coupled to the alpha and epsilon amino groups of the core peptide (Lys( n )-Ala-Cys-NH2) with n =1 to 5, with complete substitution leading to homogeneous glycoclusters. The thiol group of the cysteine residue was used to tag the glycosylated oligolysines upon reaction with fluorescein iodoacetamide. The affinity of these glycoclusters towards two plant lectins was assessed by surface plasmon resonance. The selectivity of their cell uptake was investigated by flow cytometry using two types of cells: a human hepatoma cell line (HepG2 cells) expressing the plasma membrane galactose-specific lectin, and monocyte-derived dendritic cells expressing the plasma membrane mannose-specific lectin. The glycoclusters containing four or five disaccharides were shown to bind plant lectins and cell surface membrane lectins with a narrow selectivity and with a high affinity.

Sugar-dependent nuclear import of glycosylated proteins in living cells.

The nuclear import of proteins larger than Mr 40,000 depends on the presence of a nuclear localization signal (NLS) corresponding either to a short peptide sequence or to defined sugars. The sugar-dependent nuclear import was previously evidenced by using glycosylated proteins (neoglycoproteins) introduced into the cytosol of cells either by electroporation or on digitonin-permeabilization and was shown to be distinct from the peptide NLS-mediated pathway. In this work, we used a microinjection approach to compare the two nuclear import pathways in intact living cells. The intracellular localization of fluorescent NLS-BSA or Glc-BSA injected into the cytosol was analyzed by confocal microscopy. Novel differences between the two mechanisms were evidenced. First, Glc-BSA migrated less efficiently into the nucleus than NLS-BSA because of a cytosolic retention. Second, the import of neoglycoproteins was not affected by microinjection of antinuclear import factor importin/karyopherin beta antibodies, whereas the NLS-dependent transport was completely abolished. Third, the nuclear import activity of Glc-BSA was found to be cell cycle-dependent in thymidine and hydroxyurea-treated HeLa cells, with greatest efficiency during G1/S transition and S phases, whereas NLS-BSA was imported with the same efficiency during any stage of the cell cycle but the G2 phase. Fourth, we show that after mitosis, nonglycosylated BSA was excluded from the nucleus contrary to Glc-BSA. In both cases, the nuclear import signals (NLS or alpha-glucoside) were grafted onto BSA; such tools led to a clear-cut conclusion, which will reach a full physiological significance when they are confirmed in the case of endogenous (glyco)proteins.

Lactosylated polyethylenimine for gene transfer into airway epithelial cells: role of the sugar moiety in cell delivery and intracellular trafficking of the complexes.

BACKGROUND: As we have previously shown that lactosylated polyethylenimine (PEI) is the most efficient glycosylated PEI for gene transfer into human airway epithelial cells in primary culture, we have studied here the role of the lactose residue in the enhancement of gene transfer efficiency observed with lactosylated PEI as compared with unsubstituted PEI in immortalized (Sigma CFTE29o- cells) and primary human airway epithelial cells. METHODS AND RESULTS: After three transfections of 1 h performed daily, 60% of Sigma CFTE29o- cells were transfected with lactosylated PEI, whereas 25% of cells were transfected with unsubstituted PEI (p < 0.05). Cell viability was 1.8-fold greater with lactosylated PEI as compared with unsubstituted PEI (p < 0.05). As assessed by flow cytometry, the cellular uptake of lactosylated complexes was greater than that of complexes made with unsubstituted PEI (p < 0.05) and involved mostly a receptor-mediated endocytosis. The study of the intracellular trafficking in airway epithelial cells of complexes showed an endosomal and lysosomal accumulation of lactosylated complexes. In the presence of a proton pump inhibitor, the level of lactosylated and unsubstituted PEI-mediated gene expression was reduced more than 20-fold, whereas the cell viability increased to almost 100%. For both complexes, a nuclear localization was observed for less than 5% of intracellular complexes. CONCLUSIONS: Our results show that the greater gene transfer efficiency observed for lactosylated complexes may be attributed to a higher amount of lactosylated complexes incorporated by airway epithelial cells and a lower cytotoxicity that might be related to reduced endosomolytic properties. However, the lactose residues substituting the PEI did not promote the entry of the plasmid into the nucleus.

Uptake of plasmid/glycosylated polymer complexes and gene transfer efficiency in differentiated airway epithelial cells.

BACKGROUND: We have studied gene transfer efficiency of glycosylated polylysines and glycosylated polyethylenimines as vectors in immortalized differentiated airway gland serous cells and primary cultures of human airway surface epithelial cells. METHODS AND RESULTS: In both cell types, lactosylated PEI was more efficient for gene transfer than unsubstituted PEI and lactosylated polylysine which requires the presence of endosomolytic agents. However, for all the vectors tested, gene transfer efficiency was lower in differentiated cells as compared with poorly differentiated cells. The presence of membrane lectins, i.e. cell surface sugar-specific receptors, was evaluated using fluorescein-conjugated neoglycoproteins and microscopy or flow cytometry. In differentiated airway surface epithelial cells, membrane lectins were not expressed and plasmid DNA/fluorescein-conjugated glycosylated polymer complexes were not incorporated. This accounted in part for the lack of gene transfer efficiency in these cells. In contrast, in differentiated airway gland serous cells, expression of lectins and their endocytotic properties appeared to be similar to that observed in undifferentiated cells, and plasmid DNA/fluorescein-conjugated glycosylated polymer complexes were incorporated in similar amounts by cells in both differentiated states CONCLUSIONS: Glycosylated PEI appears to be a promising gene delivery system since it is more efficient than the sugar-free polymer and does not require endosomolytic agents. However, in differentiated airway gland serous cells, a low gene transfer efficiency was observed that could not be attributed to low expression of membrane lectins or low uptake of glycosylated complexes. An impaired intracellular trafficking of glycosylated complexes in differentiated airway gland serous cells is suggested.

Oligolysine-based oligosaccharide clusters: selective recognition and endocytosis by the mannose receptor and dendritic cell-specific intercellular adhesion molecule 3 (ICAM-3)-grabbing nonintegrin.

Dendritic cells are potent antigen-presenting cells that express several membrane lectins, including the mannose receptor and DC-SIGN (dendritic cell-specific ICAM-3-grabbing nonintegrin). To identify highly specific ligands for these dendritic cell receptors, oligosaccharides were converted into glycosynthons (Os1) and were used to prepare oligolysine-based glycoclusters, Os-[Lys(Os)]n-Ala-Cys-NH2. Clusters containing two to six dimannosides as well as clusters containing four or five pentasaccharides (Lewisa or Lewisx) or hexasaccharides (Lewisb) were synthesized. The thiol group of the appended cysteine residue allows easy tagging by a fluorescent probe or convenient substitution with an antigen. Surface plasmon resonance was used to determine the affinity of the different glycoclusters for purified mannose receptor and DC-SIGN, whereas flow cytometry and confocal microscopy analysis allowed assessment of cell uptake of fluoresceinyl-labeled glycoclusters. Dimannoside clusters are recognized by the mannose receptor with an affinity constant close to 106 liter.mol-1 but have a very low affinity for DC-SIGN (less than 104 liter x mol-1). Conversely, Lewis clusters have a higher affinity toward DC-SIGN than toward the mannose receptor. Dimannoside clusters are efficiently taken up by human dendritic cells as well as by rat fibroblasts expressing the mannose receptor but not by HeLa cells or rat fibroblasts expressing DC-SIGN; DC-SIGN-expressing cells take up Lewis clusters. The results suggest that ligands containing dimannoside clusters can be used specifically to target the mannose receptor, whereas ligands containing Lewis clusters will be targeted to DC-SIGN.