Improved chitosan-mediated gene delivery based on easily dissociated chitosan polyplexes of highly defined chitosan oligomers.

Nonviral gene delivery systems based on conventional high-molecular-weight chitosans are efficient after lung administration in vivo, but have poor physical properties such as aggregated shapes, low solubility at neutral pH, high viscosity at concentrations used for in vivo delivery and a slow dissociation and release of plasmid DNA, resulting in a slow onset of action. We therefore developed highly effective nonviral gene delivery systems with improved physical properties from a series of chitosan oligomers, ranging in molecular weight from 1.2 to 10 kDa. First, we established structure-property relationships with regard to polyplex formation and in vivo efficiency after lung administration to mice. In a second step, we isolated chitosan oligomers from a preferred oligomer fraction to obtain fractions, ranging from 10 to 50-mers, of more homogeneous size distributions with polydispersities ranging from 1.01 to 1.09. Polyplexes based on chitosan oligomers dissociated more easily than those of a high-molecular-weight ultrapure chitosan (UPC, approximately a 1000-mer), and released pDNA in the presence of anionic heparin. The more easily dissociated polyplexes mediated a faster onset of action and gave a higher gene expression both in 293 cells in vitro and after lung administration in vivo as compared to the more stable UPC polyplexes. Already 24 h after intratracheal administration, a 120- to 260-fold higher luciferase gene expression was observed compared to UPC in the mouse lung in vivo. The gene expression in the lung was comparable to that of PEI (respective AUCs of 2756+/-710 and 3320+/-871 pg luciferase x days/mg of total lung protein). In conclusion, a major improvement of chitosan-mediated nonviral gene delivery to the lung was obtained by using polyplexes of well-defined chitosan oligomers. Polyplexes of oligomer fractions also had superior physicochemical properties to commonly used high-molecular-weight UPC.

Relationship between the physical shape and the efficiency of oligomeric chitosan as a gene delivery system in vitro and in vivo.

BACKGROUND: Chitosans of high molecular weights have emerged as efficient nonviral gene delivery systems, but the properties and efficiency of well-defined low molecular weight chitosans (<5 kDa) have not been studied. We therefore characterized DNA complexes of such low molecular weight chitosans and related their physical shape and stability to their efficiency as gene delivery systems in vitro and in vivo. METHODS: Individual complexes between six different chitosan oligomers (6-, 8-, 10-, 12-, 14- and 24-mers) and fluorescence-labeled T4 DNA were visualized and classified into six physical shapes using video-enhanced fluorescence microscopy. The effects of chitosan chain length, charge ratio (+/-) and solvent properties (pH and ionic strength) on the stability and structure of the complexes were studied. Gene expression in vitro and in vivo were studied using a luciferase reporter gene. RESULTS: Free DNA appeared as extended coils. Chitosan complexes had a variety of physical shapes depending on the experimental conditions. In general, the fraction of complexes that had nonaggregated, globular structures increased with increasing chain length of the chitosan oligomer, increasing charge ratio and reduction of pH (from 6.5 to 3.5). A further increase in charge ratio for globular complexes or a further reduction in pH (to 2.5) increased the fraction of aggregates, indicating a window where pharmaceutically desirable globules are obtained. Gene transfection efficiencies in vitro and in vivo were related to the physical shape and stability of the complexes. Only the 24-mer formed stable complexes that gave a high level of gene expression comparable to that of high molecular weight ultrapure chitosan (UPC) in vitro and in vivo. CONCLUSIONS: Chitosan oligomers form complexes with DNA in a structure-dependent manner. We conclude that the 24-mer, which has more desirable physical properties than UPC, is more attractive as a gene delivery system than the conventional high molecular weight chitosans.

Preparation and characterisation of fluorescent chitosans using 9-anthraldehyde as fluorophore.

Chitosans with chemical composition ranging from a fraction of N-acetylated units (F(A)) of 0.01 to 0.61 were used to prepare fluorescence labelled chitosans by reductive amination with 9-anthraldehyde. Fluorescent chitosans with a low theoretical degree of substitution (DS, 0.001-1%) were prepared, and the actual DS of the products were determined by UV and (1)H NMR spectroscopy. The fluorescence excitation and emission spectra of the chitosan with F(A) of 0.09 and DS 1% showed an excitation maximum at 254 nm and an emission maximum at 413 nm. The intrinsic viscosities ([eta]) of the fluorescent chitosans were compared to those of the original chitosans, showing that the derivatisation procedure lead only to a negligible decrease in [eta]. The conformation of these fluorescent chitosans with very low DS-values is not altered and they can conveniently be directly quantified by UV or fluorescence spectroscopy.

Screening of chitosans and conditions for bacterial flocculation.

Chitosans with different chemical compositions and molecular weights have been evaluated as flocculants of Escherichia coli suspensions. The flocculation performance of chitosans at different conditions (pH, ionic strength) was followed by residual turbidity measurements. For precise comparison, the chitosan concentrations corresponding to 75% flocculated bacteria (x(75)) were calculated from a mathematical function fitted to the measured data. At all conditions, an increase in the fraction of acetylated units (F(A)) resulted in lower x(75) and thereby better flocculation efficiency. Especially the most acetylated chitosans (F(A) 0.49 and F(A) 0.62) were excellent flocculants. An increase in F(A) from 0.002 to 0.6 caused a 10-fold reduction in necessary concentrations, at both pH 5 and 6.8. pH was a rather insignificant factor in the range 4-7.4, further pH increase led to either increase of necessary doses at low F(A) or sudden ceasing of flocculation at high F(A). The chitosans flocculated in a broad range of molecular weights, although an increase in molecular weight was a favorable factor. Increase in ionic strength caused a severalfold reduction in x(75) for all chitosans and considerable broadening of flocculation intervals.

Binding of a highly de-N-acetylated chitosan to Japanese pheasan lysozyme as measured by 1H-NMR spectroscopy.

Binding of a highly de-N-acetylated chitosan to Japanese pheasant lysozyme (JPL), which differs from hen egg white lysozyme (HEWL) by nine amino acid substitutions (including Arg114-->His), was investigated by 1H-NMR spectroscopy. The profile of the one-dimensional spectrum of JPL is essentially identical to that of HEWL. Using two-dimensional spectra of JPL and HEWL, several aromatic and aliphatic proton resonances of JPL were assigned by comparison. When a highly de-N-acetylated chitosan (number-average degree of polymerization, about 18; degree of acetylation, 0.04), where the N-acetylated units are predominantly surrounded by de-N-acetylated units (a monoacetylated chitosan), was added to the JPL solution, the NMR signals were clearly affected in Trp28 C5H and Ile98 gammaCH, as in the case of binding to HEWL. The dissociation constant of the monoacetylated chitosan evaluated from the NMR signal responses was calculated to be 0.23+/-0.05 mm (-31.5 kJ/mol), which is similar to that of HEWL (0.11+/-0.02 mm, -33.3 kJ/mol). Thus, the Arg-->His substitution of the 114th amino acid, which participates in sugar residue binding at the right-sided subsite F, did not significantly affect the chitosan binding. In addition, the C2H signal of His114 of JPL was not affected by the chitosan binding. These results suggest that the monoacetylated chitosan binds to subsites E and F through the left-sided binding mode.

Chitosan as a nonviral gene delivery system. Structure-property relationships and characteristics compared with polyethylenimine in vitro and after lung administration in vivo.

Chitosan is a natural cationic linear polymer that has recently emerged as an alternative nonviral gene delivery system. We have established the relationships between the structure and the properties of chitosan-pDNA polyplexes in vitro. Further, we have compared polyplexes of ultrapure chitosan (UPC) of preferred molecular structure with those of optimised polyethylenimine (PEI) polyplexes in vitro and after intratracheal administration to mice in vivo. Chitosans in which over two out of three monomer units carried a primary amino group formed stable colloidal polyplexes with pDNA. Optimized UPC and PEI polyplexes protected the pDNA from serum degradation to approximately the same degree, and they gave a comparable maximal transgene expression in 293 cells. In contrast to PEI, UPC was non toxic at escalating doses. After intratracheal administration, both polyplexes distributed to the mid-airways, where transgene expression was observed in virtually every epithelial cell, using a sensitive pLacZ reporter containing a translational enhancer element. However, the kinetics of gene expression differed - PEI polyplexes induced a more rapid onset of gene expression than UPC. This was attributed to a more rapid endosomal escape of the PEI polyplexes. Although this resulted in a more efficient gene expression with PEI polyplexes, UPC had an efficiency comparable to that of commonly used cationic lipids. In conclusion, this study provides insights into the use of chitosan as a gene delivery system. It emphasises that chitosan is a nontoxic alternative to other cationic polymers and it forms a platform for further studies of chitosan-based gene delivery systems.

Preparation and characterisation of oligosaccharides produced by nitrous acid depolymerisation of chitosans.

Two chitosans with widely different chemical composition (fraction of N-acetylated units (F(A))<0.001 and F(A)=0.59), were degraded by nitrous acid, to obtain the reactive 2,5-anhydro-D-mannose- (M-) unit at the new reducing end. The fully N-acetylated and fully N-deacetylated oligomers were separated by size-exclusion chromatography. Both the chemical structure and purity were studied by one- and two-dimensional 1H and 13C NMR methods. The fully N-acetylated oligomers were found to be stable, whereas the N-deacetylated oligomers reacted intermolecularly by a Schiff base reaction between the 2-amino group on the N-deacetylated units and the M-units, facilitating the cleavage of the glycosidic bond next to the M-unit and the formation of 5-hydroxymethylfurfural (HMF).

Electrophoretic light scattering studies of chitosans with different degrees of N-acetylation.

Electrostatic properties of three chitosans with fractions of N-acetylated units (F(A)) of 0.01, 0.13 and 0.49 were examined by electrophoretic light-scattering technique (ELS) and (1)H NMR spectroscopy. From the dependency of mean electrophoretic mobilities on pH, the pK(a) values were calculated. Despite their large differences in chemical composition, all chitosans had similar pK(a) values of 6.5-6.6. All chitosans also showed the same polyelectrolyte behavior when apparent pK(a) values were calculated according to Katchalsky and plotted as a function of the degree of ionization alpha. The intrinsic pK(a) values (pK(0)) extrapolated to zero charge were about 9. The results derived from an independent (1)H NMR study of the same chitosan samples showed no effect of F(A) on titration behavior of chitosan, confirming the results obtained by ELS.

Chitosans as absorption enhancers of poorly absorbable drugs. 3: Influence of mucus on absorption enhancement.

Chitosans are potent nontoxic absorption enhancers after nasal administration but their effects on the intestinal epithelium in vivo has not been studied in detail. In this study, the effects of chitosans with varying molecular weights and degrees of acetylation on the absorption of a poorly absorbed model drug (atenolol) were studied in intestinal epithelial cell layers with or without a mucus layer and in an in situ perfusion model of rat ileum. The effects of the chitosans on epithelial morphology and release of lactate dehydrogenase (LDH) into the perfusate were investigated in the in situ model. The chitosans had pronounced effects on the permeability of mucus-free Caco-2 layers and enhanced the permeation of atenolol 10- to 15-fold, with different absorption kinetics for different chitosans, in accordance with previous results. In contrast, enhancement of atenolol absorption through rat ileum was modest. LDH release from the tissues perfused with chitosans did not increase, indicating that the chitosans were used at nontoxic concentrations. Morphological examination of the perfused ileal tissues revealed more mucus discharge from the tissues exposed to chitosans than from controls, which suggested that the discharged mucus may inhibit the binding of chitosan to the epithelial surface and hence decrease the absorption-enhancing effect. This hypothesis was supported by studies with intestinal epithelial HT29-H goblet cells covered with a mucus layer. The binding of chitosan to the epithelial cell surface and subsequent absorption-enhancing effects were significantly reduced in mucus-covered HT29-H cultures. When the mucus layer was removed prior to the addition of chitosan, the cell surface binding and absorption-enhancing effects of the chitosans were increased. We conclude that the modest absorption-enhancing effects of unformulated chitosan solutions in the perfused rat ileum are a result of the mucus barrier in this tissue. This effect may be overcome by increasing the local concentrations of both chitosan and drug, i.e,. through formulation of the chitosan into a particulate dosage form.

Quantitative studies of the non-productive binding of lysozyme to partially N-acetylated chitosans. Binding of large ligands to a one-dimensional binary lattice studied by a modified McGhee and von Hippel model.

This paper considers the non-productive (inhibitory) binding of chitosans to lysozyme from chicken egg white. Chitosans are linear, binary heteropolysaccharides consisting of 2-acetamido-2-deoxy-beta-D-glucose (GlcNAc; A-unit) and 2-amino-2-deoxy-beta-D-glucose (GlcN, D-unit). The active site cleft of lysozyme can bind six consecutive sugar residues in subsites named A-F, and specific binding of chitosan sequences to lysozyme occurs with A-units in subsite C. Chitosans with different fractions of A-units (FA) induced nearly identical changes in the 1H NMR spectrum of lysozyme upon binding, and the concentration of bound lysozyme could be determined. The data were analysed using a modified version of the McGhee and von Hippel model for binding of large ligands to one-dimensional homogeneous lattices. The average value of the dissociation constant for different sequences that may bind to lysozyme (K(ave)D) was estimated, as well as the number of chitosan units covered by lysozyme upon binding. K(ave)D decreased with increasing FA-values at pH 3 and 4.5, while the opposite was true at pH 5.5. Contributions from different hexamer sequences to K(ave)D of the chitosans were considered, and the data revealed interesting features with respect to binding of lysozyme to partially N-acetylated chitosans. The relevance of the present data with respect to understanding lysozyme degradation kinetics of chitosans is discussed.

The interactions between highly de-N-acetylated chitosans and lysozyme from chicken egg white studied by 1H-NMR spectroscopy.

We have investigated the binding interactions between highly de-N-acetylated chitosans and lysozyme from chicken egg white by one-dimensional and two-dimensional 1H-NMR spectroscopy. A fully de-N-acetylated chitosan (fraction of N-acetylated units, F < 0.001) induced no observable changes in the 1H chemical shifts of lysozyme. However, a chitosan with F(A) = 0.04, where the N-acetylated units are predominantly surrounded by de-N-acetylated units (a monoacetylated sequence), induced significant shifts of several lysozyme resonances, demonstrating a specific interaction between lysozyme and de-N-acetylated units in the chitosan. The interaction between the two positively charged molecules increased with increasing ionic strength, as expected. The dissociation constant (Kd) between lysozyme and the monoacetylated sequence was strongly dependent on pH* (pH measured in D2O), with Kd = 0.02+/-0.01 mM at pH* 6.0, Kd = 0.11+/-0.02 mM at pH* 4.5, and Kd approximately 2 mM at pH* 3, suggesting that electrostatic forces contribute to the observed binding. The complex was strikingly stable, with bound lifetimes in the range of 10-25 ms at pH* 4.5 and 328-300 K. Most lysozyme resonances that were affected by the binding were assigned, and we suggest that the monoacetylated chitosan sequence binds to the active site cleft of lysozyme with the N-acetylated unit in subsite C. Assuming this binding mode, we have discussed the contributions in energetic terms from individual subsites of lysozyme towards binding of N-acetylated and de-N-acetylated units.

Chitosans as absorption enhancers for poorly absorbable drugs 2: mechanism of absorption enhancement.

PURPOSE: It has recently been shown that the absorption enhancing and toxic effects of chitosans are dependent on their chemical composition. In this study, the mechanisms underlying these effects were investigated at the cellular level. METHODS: The effects on epithelial cells of chitosans with different chemical composition, absorption enhancing properties and toxicities were studied in Caco-2 monolayers. Chitosan C( 1:31) has a low degree of acetylation (DA) (1%) and a low m.w. (31 kD), and displays dose-dependent absorption enhancement and cytotoxicity; chitosan C(35:170) has a higher DA (35%) and a higher m.w. (170 kD), is less dose-dependent in absorption enhancement, and is not cytotoxic. A third non-toxic chitosan C(49:22) with a high DA (49%), a low m.w. (22 kD), and no influence on epithelial permeability was used as control. RESULTS: C(1:31) and C(35:170) bound tightly to the epithelium. Cellular uptake of the chitosans was not observed. Both chitosans increased apical but not basolateral cell membrane permeability and induced a redistribution of cytoskeletal F-actin and the tight junction protein ZO-1. This resulted in increased paracellular permeability of hydrophilic marker molecules of different molecular weights. Addition of negatively charged heparin inhibited the cellular and the absorption enhancing effects of the chitosans, indicating that these effects are mediated via their positive charges. The onset of the effects of C(35:170) on apical membrane permeability and tight junction structure was much faster than that of C(1:31). C(49:22) did not influence any of the properties of the Caco-2 cell monolayers studied. CONCLUSIONS: The binding and absorption enhancing effects of chitosans on epithelial cells are mediated through their positive charges. The interaction of chitosans with the cell membrane results in a structural reorganisation of tight junction-associated proteins which is followed by enhanced transport through the paracellular pathway.

In vitro degradation rates of partially N-acetylated chitosans in human serum.

The initial degradation rates (r) in human serum of three chitosans with FA = 0.42, 0.51, and 0.60 were determined by measuring the decrease in viscosity as a function of time. A strong increase in r with increasing FA of the chitosans was observed, with r increasing proportionally to FA4.5. With increasing concentrations of lysozyme added to the reaction mixtures of chitosan and serum, the relative increase in degradation rate of chitosans with increasing FA was almost the same as that without lysozyme added. Addition of the chitinase inhibitor allosamidin (50 microM) did not inhibit the degradation rate of chitosan (FA = 0.60) by human serum. The results suggest that chitosans are actually mainly depolymerized by lysozyme in human serum, and not by other enzymes or other depolymerization mechanisms.

Chitosans as absorption enhancers for poorly absorbable drugs. 1: Influence of molecular weight and degree of acetylation on drug transport across human intestinal epithelial (Caco-2) cells.

PURPOSE: Chitosan has recently been demonstrated to effectively enhance the absorption of hydrophilic drugs such as peptides and proteins across nasal and intestinal epithelia (1-3). In this study, the effect of the chemical composition and molecular weight of chitosans on epithelial permeability and toxicity was investigated using monolayers of human intestinal epithelial Caco-2 cells as a model epithelium. METHODS: Eight chitosans varying in degree of acetylation (DA) and molecular weight were studied. The incompletely absorbed hydrophilic marker molecule 14C-mannitol was used as a model drug to assess absorption enhancement. Changes in intracellular dehydrogenase activity and cellular morphology were used to assess toxicity. RESULTS: Chitosans with a low DA (1 and 15%) were active as absorption enhancers at low and high molecular weights. However, these chitosans displayed a clear dose-dependent toxicity. Chitosans with DAs of 35 and 49% enhanced the transport of 14C-mannitol at high molecular weights only, with low toxicity. One chitosan (DA = 35%; MW = 170 kD) was found to have especially advantageous properties such as an early onset of action, very low toxicity, and a flat dose-absorption enhancement response relationship. CONCLUSIONS: The structural features of chitosans determining absorption enhancement are not correlated with those determining toxicity, which makes it possible to select chitosans with maximal effect on absorption and minimal toxicity.

Determination of enzymatic hydrolysis specificity of partially N-acetylated chitosans.

A new method for determining the specificity of hydrolysis of the linear binary heteropolysaccharide chitosan composed of (1-->4)-linked 2-acetamido-2-deoxy-beta-D-glucopyranose (GlcNAc; A-unit) and 2-amino-2-deoxy-beta-D-glucopyranose (GlcN; D-unit) residues is described. The method is based on the assignments of the 13C chemical shifts of the identity (A- or D-units) of the new reducing and non-reducing ends and the variation in their nearest neighbours, using low molecular weight chitosans with known random distribution of A- and D-units as substrate. A highly N-acetylated chitosan with fraction of acetylated units (FA) of 0.68 and a number-average degree of polymerization (DPn) of 30 was hydrolysed with hen egg-white lysozyme, showing that both the new reducing and non-reducing ends consisted exclusively of A-units, indicating a high specificity for A-units in subsites DL and EL on lysozyme. Our data suggests that the preceding unit of the reducing A-units, is invariable, and based on earlier studies, most probably an A-unit, while the unit following the non-reducing A-units can be either an A- or a D-unit. A more detailed study of the specificity of lysozyme at subsite DL was performed by hydrolyzing a more deacetylated chitosan (FA = 0.35 and DPn of 20) to a DPn of 9, showing that even for this chitosan more than 90% of the new reducing ends were acetylated units. Thus, lysozyme depolymerizes partially N-acetylated chitosans by preferentially hydrolyzing sequences of acetylated units bound to site CL, DL and EL of the active cleft, while there is no specificity between acetylated and deacetylated units to site FL. In addition, a moderately N-acetylated chitosan with fraction of acetylated units (FA) of 0.35 and a DPn of 20 was hydrolysed with Bacillus sp. No. 7-M chitosanase, showing that both the new reducing and non-reducing ends consisted exclusively of D-units. Our data suggests that the nearest neigbour to the D-unit at the reducing end is invariable, and based on earlier studies, most probably a D-unit, while the unit following the non-reducing D-units can be either an A- or a D-unit. We conclude that the Bacillus chitosanase hydrolyzes partially N-acetylated chitosan by preferentially attacking sequences of three consecutive deacetylated units, hypothetical subsites CC, DC and EC, where the cleavage occur between sugar units bound to subsites DC and EC. A hypothetical subsite FC on the chitosanase show no specificity with respect to A- and D-units. The new NMR method described herein offers a time and labour-saving alternative to the procedure of extensive hydrolysis of the binary heteropolysaccharide chitosan and subsequent isolation and characterization of the oligosaccharides.

Characterization of binding and TNF-alpha-inducing ability of chitosans on monocytes: the involvement of CD14.

Chitosans with different chemical composition were found to induce TNF-alpha production from human monocytes. Their ability to induce TNF-alpha was found to be highly dependent on neutral-solubility and molecular weight. Monoclonal antibodies against CD14 inhibited TNF-alpha production from monocytes stimulated with neutral-soluble chitosans. Binding studies indicated that lipopolysaccharides (LPS) and neutral-soluble chitosans share a binding site on monocytes which involves CD14. TNF-alpha production from monocytes stimulated with chitosans was dependent on serum. LPS-binding protein (LBP) enhanced the chitosan-induced TNF-alpha production only to a minor degree, suggesting that serum proteins other than LBP play an important role in the stimulatory effect.

Hydrodynamic characterization of chitosans varying in degree of acetylation.

The molecular weights and gross aqueous solution conformation of two chitosans of different degrees of acetylation, 'Sea Cure +210' (11% acetylated) and KN50 (58% acetylated) were characterized by viscometry, analytical ultracentrifugation and dynamic light scattering. The hydrodynamic parameters obtained were used to determine both the molecular weights and the gross solution conformation of the two chitosans. Using the Wales-Van Holde ratio of sedimentation coefficient concentration regression coefficient (ks) to intrinsic viscosity [eta], the Sea Cure +210 chitosan, which is much less acetylated than the KN50, is highly asymmetric in conformation. This, in conjunction with the charge on the molecule, would suggest a rod-like conformation in solution. The two largest KN50 chitosans have widely differing values for the Wales-Van Holde ratio, suggesting different solution conformation. However, when the series is examined as a whole using the Mark-Houwink-Kuhn-Sakurada relationships relating molecular weight to both intrinsic viscosity and translational diffusion coefficient, then a more spheroidal structure, approximating to a random coil, is predicted.

Chitosan cross-linked with Mo(VI) polyoxyanions: a new gelling system.

A procedure for preparing homogeneous chitosan gels by in situ molybdate cross-linking is described. The gels are obtained by dispersing solid MoO3 in a buffered chitosan solution and the polymer is cross-linked by formation of heavily negatively charged molybdate polyoxyanions. The resulting ionic gels are very transparent, thermoirreversible and can be made at low polymer concentrations. Depending on the ionic strength, these gels are able to swell several times their original size in aqueous solutions. Estimates of the degree of cross-linking reveal a very open pore structure which is confirmed by electron micrographs of the gel.

13C-n.m.r. studies of the acetylation sequences in partially N-deacetylated chitins (chitosans).

Chitosans obtained under homogeneous conditions of N-deacetylation, with degrees of N-deacetylation between 46% and 94%, were depolymerised and their 125-MHz 13C-n.m.r. spectra have been interpreted. The sequence of 2-acetamido-2-deoxy-beta-D-glucopyranose (GlcNAc) and 2-amino-2-deoxy-beta-D-glucopyranose (GlcN) residues influenced the chemical shifts, and the diad and triad frequencies have been calculated. Chitosans that were N-deacetylated under homogeneous and heterogeneous conditions gave values for the diad and triad frequencies that were consistent with a random arrangement of GlcN and GlcNAc residues.