WO2016101068A1 - Conjugated chitosans and method of use and of making same - Google Patents

Abstract

There is provided a chitosan conjugate of formula (I): wherein CS represents a chitosan residue; X comprises a Zl group, the Zl being a linear thioacetal or forming an n-membered thioacetal ring, n being from 4 to 9, the Zl optionally being ramified and/or substituted; and POLY represents a ligand, or a pharmaceutically acceptable salt thereof. There is also provided uses of such conjugate as well methods of making same.

Description

CONJUGATED CHITOSANS AND METHOD OF USE AND OF MAKING SAME

CROSS-REFERENCE TO RELATED APPLICATION

[01] The present application claims the benefit of U.S. provisional patent application serial number 62/095,435 filed on December 22, 2014 by Pickenhahn et al. The contents of the above- referenced document are incorporated herein by reference in their entirety.

TECHNICAL FIELD

[02] This application generally relates to the field of conjugated chitosans, methods of use, and methods of making same.

BACKGROUND

[03] Chitosan, a linear and cationic polysaccharide is derived from chitin by deacetylation. This non-toxic cationic polysaccharide holds great interest due to its biocompatibility, biodegradability and mucoadhesive properties. Rinaudo, Progress in Polymer Science 2006, 31 (7), 603-632. Chitosan and its derivatives have been proposed for applications including gene and drug delivery, tissue repair, water purification and cosmetics. Two general approaches have been explored to chemically modify chitosan - lateral "graft" and "block" modifications. The former involves conjugation to chitosan lateral functional groups (amine or hydroxyl) while the latter relies on conjugation to chitosan end groups.

[04] Chitosan lateral grafting, however, can potentially compromise the ability of chitosan to bind nucleic acid and thus limit the stability and efficiency of chitosan/nucleic acid complexes for gene delivery applications. Indeed, lateral grafting can impede the ability of chitosan to electrostatically bind to negatively charged species by reducing its effective charge density and by potentially creating steric hindrance with bulky moieties. Casettari et al., Progress in Polymer Science 2012, 57 (5), 659-685.

[05] Chitosan block modification strategies have been recently proposed as a means to modify the chitosan properties without compromising its ability to bind oppositely charged macro-ions such as nucleic acids. Two different chitosan attachment sites have been explored to date: the first site is formed after chitosan depolymerization by nitrous acid (HONO) where a 2,5-anhydro-D- mannose unit (M-Unit) is formed at the reducing end of the cleaved polymer (Allan et al., Carbohydrate research 1995, 277 (2), 257-272), while the second site is available on the open- chain form, present in trace amounts, of the chitosan reducing extremity (either GlcNH₂ or GlcNHAc units) and allows mutarotation between the alpha and beta anomers. These two chitosan block modification strategies thus rely on the reaction of the aldehyde moiety with nucleophilic species. However, in both cases, the aldehyde moiety appears to be mostly present in its hydrated and unreactive form, also referred to as the geminal or gem-& o\ form, under acidic aqueous conditions. Los et al., Journal of the American Chemical Society 1956, 78 (8), 1564-1568; T0mmeraas et al., Carbohydrate research 2001, 333 (2), 137-144; and Azevedo et al., Carbohydrate Polymers 2012, 87 (3), 1925-1932.

[06] The amines of chitosan in their neutral form are strong nucleophiles that can react with the aldehyde of chitosan' s reducing end as depicted in the following scheme 1 at reaction 3:

R

MONO

CHITOSAN _*■

Chitosan

Depolymerization

Scheme 1

[07] Therefore block conjugation to the chitosan end group requires that the proportion of chitosan amines in their reactive form be minimized, for example by performing reactions at pH significantly lower than the chitosan pK_a (-6.5). However, chitosan pK_a varies with both ionic strength and chitosan charge density and can reach values as low as about 5.5 at high charge density and in the absence of added salt. Filion et al., Biomacromolecules 2007, 8 (10), 3224-34. To date, all chitosan end-group conjugation reactions known to the inventors to have been implemented rely on oxime-click chemistry. [08] For example, U.S. 2012/0238735 describes polyethylene glycol (PEG)-chitosan conjugates prepared by an oxime-click chemistry reaction of a single PEG moiety with an aldehyde on a single chitosan backbone. While this reaction affords a control over the ratio PEG/chitosan, the presence of the aldehyde on the chitosan depends on a mutarotation equilibrium in which the amount of non-reactive cyclic form at equilibrium is very high compared to the reactive opened- ring intermediate one (e.g., only 0.0051% of aldehyde opened form for D-Glucose at 52°C, pH 5). Maple et ai, Journal of the American Chemical Society 1987, 109 (10), 3168-3169. Although the reported yield of the reaction is high, it is worth mentioning that the chitosan-PEG adducts conjugation efficiency was mostly evaluated by gel permeation chromatography (GPC), where the free chitosan and the block-copolymer peaks are hardly distinguishable from each other. Indeed, similar elution volumes are obtained for both the block-copolymer and its corresponding glycan precursors, which precludes their precise quantification and thus the determination of conjugation efficiency (Novoa-Carballal, R.; Silva, C; Moller, S.; Schnabelrauch, M.; Reis, R. L.; Pashkuleva, I., Tunable nano-carriers from clicked glycosaminoglycan block copolymers. Journal of Materials Chemistry B 2014, 2 (26), 4177-4184).

[09] In light of the above, there is a need to provide chitosan conjugates as well as a method of making same that alleviate at least in part deficiencies of the existing conjugates and methods of making same.

SUMMARY

[10] In one non-limiting broad aspect, the present disclosure relates to a chitosan conjugate of formula (I):

wherein CS represents a chitosan residue; X comprises a Zl group, the Zl being a linear thioacetal or forming an n-membered thioacetal ring, n being from 4 to 9, the Zl optionally being ramified and/or substituted; and POLY represents a ligand, or a pharmaceutically acceptable salt thereof.

[11] In another non-limiting broad aspect, the present disclosure relates to a compound of formula (II):

wherein CS represents a chitosan residue, X comprises a thioacetal forming an n-membered ring, n being from 4 to 9, the ring optionally being ramified and/or substituents, and the ring further comprising a free SH, or a pharmaceutically acceptable salt thereof.

[12] In another non-limiting broad aspect, the present disclosure relates to a process for manufacturing depolymerized chitosan residues, comprising: depolymerization of chitosan with nitrous acid (HONO) to obtain a depolymerized chitosan residue salt, and drying the depolymerized chitosan residue salt under acidic conditions, the acidic conditions comprising a pH < 4, preferably pH < 3.5, more preferably pH < 3.

[13] In another non-limiting broad aspect, the present disclosure relates to a process for manufacturing a conjugate of chitosan residues, comprising: depolymerization of chitosan with nitrous acid (HONO) to obtain a depolymerized chitosan residue salt, drying the depolymerized chitosan residue salt under acidic conditions, the acidic conditions comprising a pH < 4, preferably pH < 3.5, more preferably pH < 3, rehydration of the salt in an aqueous solvent to obtain a solution, and incorporating a thiol molecule in the solution thus producing a reaction medium for obtaining said chitosan conjugate, said reaction medium being at a pH of about 1.

[14] In one non-limiting embodiment, the above drying step comprises thermal vacuum drying. [15] In another non-limiting embodiment, the above drying step comprises freeze-drying.

[16] In another non-limiting broad aspect, the present disclosure relates to a process for manufacturing a conjugate of chitosan residues, comprising: providing a nitrous acid (HONO) depolymerized chitosan residue salt previously dried under acidic conditions, said acidic conditions comprising a pH < 4, preferably pH < 3.5, more preferably pH < 3, rehydration of the salt in an aqueous solvent to obtain a solution, and incorporating a thiol molecule in the solution thus producing a reaction medium for obtaining said chitosan conjugate, said reaction medium being at a pH of about 1.

[17] In one non-limiting practical embodiment, the herein described reaction for obtaining said chitosan conjugate can be performed at a temperature selected from the range of about 5 °C to about 90 °C. In a particular practical embodiment, the temperature is selected from the range of about 20 °C and 50 °C.

[18] In one non-limiting practical embodiment, the herein described reaction for obtaining said chitosan conjugate can be performed substantially instantly (e.g., in case where there is a direct freeze-drying of the reaction medium) or can be performed for a given amount of time, for example up to several weeks. In a particular practical embodiment, the reaction can be performed for a time period ranging from about 24h to about 72h.

[19] In another non-limiting broad aspect, the present disclosure relates to particles comprising the chitosan conjugate of formula (I) as described previously, wherein the ligand comprise a polyethylene glycol (PEG) molecule, and wherein the particles have a substantially spherical form and have a reduced zeta potential than a comparative particle being prepared in the same conditions to said particle except for comprising chitosan residues instead of chitosan-PEG conjugates.

[20] All features of embodiments which are described in this disclosure and are not mutually exclusive can be combined with one another. Elements of one embodiment can be utilized in the other embodiments without further mention. Other aspects and features of the present disclosure will become apparent to those ordinarily skilled in the art upon review of the following description of specific embodiments in conjunction with the accompanying Figures.

BRIEF DESCRIPTION OF THE FIGURES

[21] A detailed description of specific embodiments is provided herein below with reference to the accompanying drawings in which:

[22] Figure 1 is a non-limiting schematic representation of a chitosan conjugate and a reaction scheme to make same in accordance with a non-limiting embodiment of the present disclosure.

[23] Figure 2 is a non-limiting schematic representation of another chitosan conjugate and a reaction scheme to make same in accordance with a non-limiting embodiment of the present disclosure.

[24] Figure 3 is a non-limiting schematic representation of other chitosan conjugates and a reaction scheme to make same in accordance with a non-limiting embodiment of the present disclosure.

[25] Figure 4 (top) is a non-limiting schematic representation of a prior art grafted chitosan residues on a surface ("standard pathway") and of a grafted conjugate on a surface in accordance with a non-limiting embodiment of the present disclosure ("BCL pathway"), as well as (bottom) a non-limiting schematic representation of a reaction scheme to obtain the grafted conjugate on a surface in accordance with the non-limiting embodiment of the present disclosure ("BCL pathway") shown in (top).

[26] Figure 5 is a non-limiting schematic representation of another chitosan conjugate and a reaction scheme to make same in accordance with a non-limiting embodiment of the present disclosure.

[27] Figure 6 is a non-limiting schematic representation of another chitosan conjugate and a reaction scheme to make same in accordance with a non-limiting embodiment of the present disclosure in which the ligand includes a siRNA. [28] Figure 7 is a non-limiting flowchart of an experimental design for (A) mechanistic studies and (B) reactivity studies performed in the present disclosure.

[29] Figure 8 is a non-limiting schematic representation of chitosan conjugates in accordance with a non-limiting embodiment of the present disclosure for quantitation of chitosan residues derivatization efficiency: functionalization degree (F) calculations.

[30] Figure 9 is a non-limiting graph of 1H NMR spectrum of a CS 92-1 in depolymerization medium [D₂0/DC1 (50 mM), T=70°C, HOD peak was presaturated, number of scans (ns) = 2000, relaxation period (dl) = 6s, Acquisition time=2s, Exponential apodization = 1 Hz]. No aldehyde proton peak was observed around 8-10 ppm, the hydrated gem-dio\ form remaining predominant.

[31] Figure 10 is a non-limiting schematic representation of potential reaction scheme occurring during conjugation of 2,5-anhydro-D-mannose (M-Unit) and 2 thiol-bearing molecules (3- mercaptopropionic acid and β-mercaptoethanol, MPA and BME respectively) giving the following expected products: Product A is the hemithioacetal intermediate that is in equilibrium with its corresponding oxonium, whereas products B and C correspond to the oxathiolane (for BME reactions only) and thioacetal, respectively. Molecule D represents the α,β-unsaturated sulfide.

[32] Figure 11 is a non-limiting graph of ^!H NMR spectrum of the CS-BME product after workup II [D₂0/DC1, T=70°C, HOD peak was presaturated, number of scans (ns) = 64, relaxation period (dl) = 6s, Acquisition time=2s, Exponential apodization = 1 Hz]. Integration of BME (-CH₂-S-) protons peaks was used to calculate the functionalization degree (F=70% in this particular case, according to Equation 2).

[33] Figure 12 is a non-limiting graph of ^]H NMR spectrum of the CS-MPA product after workup II (D₂0/DC1, T=70°C, HOD peak was presaturated, number of scans (ns) = 64, relaxation period (dl) = 6s, Acquisition time =2s, Exponential apodization = 1 Hz). Integration of MPA protons peaks was used to calculate the functionalization degree (F=52% in this particular case, according to Equation 3).

[34] Figure 13 is a non-limiting Diffusion ordered spectroscopy experiments (DOSY) spectrum of the CS HC1 salt M-Unit conjugated to MPA. 32 gradients between 1 1.2 and 358.4 gauss.cm-1 with a gradient pulse (δ) of 1 ms, a diffusion time (Δ) of 60 ms. Both CS and MPA have the same translational diffusion coefficient at 25°C in 2% DCl in D₂0 indicating that they are conjugated given the large difference between their molar masses.

[35] Figure 14 is a non-limiting schematic representation of potential reaction scheme for thiol addition to the aldehyde group of the M-Unit CS HCl salt under acidic aqueous conditions.

[36] Figure 15 is a non-limiting graph of 1H NMR spectrum of a CS-b-PEG₂ block-copolymer conjugate in accordance with a non-limiting embodiment of the present disclosure after workup II (D₂0, T=70°C, HOD peak was presaturated, number of scans (ns) = 64, relaxation period (dl) = 6s, Acquisition time=2s, Exponential apodization = 1 Hz). Integration of Gem-dio\ proton peak was used to calculate the functionalization degree (in this particular case, F=51% according to Equation 5).

[37] Figure 16 represents non-limiting Environmental Scanning Electron Microscopy (ESEM) pictures (High vacuum mode, accelerating voltage = 20.0 kV; spot size = 3 and working distance = 5mm) of polyplexes formed with pDNA and unmodified CS or CS-b-PEG₂ block-copolymer conjugates in accordance with a non-limiting embodiment of the present disclosure (amine to phosphate ratio = 3.7, N/P=3.7). A & B (x80000 and xl60000, respectively): polyplexes formed with CS 92-10 are heterogeneous in size and present various morphologies (globular, rod-like and toroidal). Pictures C and D (x80000 and xl60000, respectively): polyplexes formed with CS-b- PEG₂ (CS 92-10 and mPEG-SH 2kDa), are substantially uniformly spherical.

[38] Figure 17 is a non-limiting schematic representation of proposed model of reaction scheme for thiol addition to the aldehyde group of the M-Unit CS HCl salt under acidic aqueous conditions in accordance with a non-limiting embodiment of the present disclosure.

[39] Figure 18 is a non-limiting graph of ¹³C solid state NMR (CP-MAS) of an extra-dried CS 99-1 salt. The sample preparation as well as the analysis was performed under inert atmosphere (N₂) to avoid contact with water content in air. (T=25°C, t=6h, v=12kHz). [40] Figure 19 is a non-limiting graphical representation of M-Unit chitosan 92-2 (deacetylation degree of 92%; 2-3 kDa) / Thiol-hook models (Ethanedithiol and propanedithiol, EDT and PDT, respectively) conjugation efficiencies determined by ^!H NMR (N>3 +SD).

[41] Figure 20 is a non-limiting schematic representation of proposed model of a reaction scheme for a triskelion linker synthesis that corresponds to a two-steps process taking place in organic conditions in accordance with a non-limiting embodiment of the present disclosure.

[42] Figure 21 is a non-limiting schematic representation of a proposed model of reaction scheme for a 2-step process for the conjugation of the acetyl-protected triskelion to the M-Unit aldehyde in accordance with a non-limiting embodiment of the present disclosure.

[43] Figure 22 is a non-limiting schematic (left) and graphical representation (right) of the relative proportion of the M-Unit aldehyde / Triskelion linker conjugation products determined by liquid chromatography-mass spectrometry (LCMS). Products A & B correspond to the desired products obtained by intramolecular cyclization (A: M-Unit-triskelion conjugate; B: M-Unit- triskelion conjugates linked by disulfide bond through the third remaining thiol moiety). Products C & D correspond to the side-products obtained by intermolecular thioacetylation (D: M-Unit- (triskelion)₂ conjugate; E: Oxidized M-Unit-(triskelion)₂ conjugates. (N>3 + SD).

[44] Figure 23 is a non-limiting graphical representation of the M-Unit chitosan (deacetylation degree of 92%; 2-3 kDa and lOkDa*) / Triskelion linker conjugation efficiencies determined by Ή NMR (N>3 ±SD).

[45] In the figures, non-limiting embodiments are illustrated by way of example. It is to be expressly understood that the description and drawings are only for the purpose of illustrating certain embodiments and are an aid for understanding. The scope of the claims should not be limited by the embodiments set forth in the present disclosure, but should be given the broadest interpretation consistent with the description as a whole.

DETAILED DESCRIPTION OF EMBODIMENTS [46] Specific examples of chitosan conjugates, methods of use and method of making same will now be described to illustrate the manner in which the principles of the present disclosure may be put into practice.

[47] In the present disclosure, the term "chitosan residue" (also referred to as "CS") generally refers to a chitosan residue having a deacetylation degree (%DDA) from about 50% to about 100% and/or a molecular weight (Mn) of from about 0.2 to about 200 kDa. The term "chitosan residue" may also generally refer to any modified chitosan residue where the modification(s) is either on the chitosan lateral amines and/or on the chitosan hydroxyl groups. The person of skill will readily envision the types of modifications which can be suitable for this purpose.

[48] In the present disclosure, the term "M-Unit" generally refers to a 2,5-anhydro-D-mannose linked at the reducing end of a chitosan residue via a glycosidic bond.

[49] In the present disclosure, the term "Triskelion" generally refers to a mercapto compound having a structure where at least two S atoms are on one end and at least one S atom is on an opposite end of the structure, for example as per the following formula:

Z1-L1 *-Z2*-L2*-Z1 ' where Zl, LI, Z2, and L2 are as defined later in this text, the asterisk * represents an optional element and Zl includes at least two free S atoms and Ζ includes at least one free S atom. The Triskelion molecule may include a compound as described for example, but without being limited thereto, in EP0528590 (e.g., Zl includes two free S atoms, LI is -S-CH₂-CH₂- and ΖΓ includes one free S atom).

[50] In the present disclosure, the term "linear alkyl" generally refers to an alkyl chain having any length within CJ-CJOO, for example but without being limited thereto, Cj-Cgo, Q-Cso, C2-C90, C₂- C₈o, Ci-C₅₀₎ C]-C_40> C!-C₃o_; C1 -C25, C!-C₂o, C!-C_15; C!-C₁₀ and the like, optionally ramified and/or substituted.

[51] In the present disclosure, the term "thioacetal" generally refers to the chemical structure RCH(SR')₂ where R is not H, and where R' is an alkyl chain C1 -C100 if a linear thioacetal. [52] In the present disclosure, the term "ligand" (also referred herein as "POLY") generally refers to any desired moiety and/or surface, for example but without being limited thereto, a nucleic acid molecule (e.g., linear DNA, mRNA, shRNA, or siRNA), a polypeptide, a non-peptidic polymer (e.g., a poly(alkylene oxide) such as a PEG), another chitosan conjugate (of identical structure or not), a planar or particulate surface, and the like. In one non-limiting embodiment, the ligand may further include one or more functional group(s) in protected or unprotected form, optionally attached via a linker group. The functional group can be, for example, one or more of a passive or active targeting moiety (fusogenic peptide, folate TAG, Galactose, and the like), dyes/fluorophores, polymers, and the like.

[53] In one non-limiting embodiment of the present disclosure, the chitosan conjugate can be represented according to the following formula:

CS - M - Zl - LI * - Z2* - L2* - Z3* - POLY - Z4* - Y*

According to the following table:

ELEMENT SIGNIFICATION DETAILS

DDA = about 50% to about 100 % and/or Mn =

CS Chitosan residue about 0.2 to about 200 kDa, optionally

modified

M-Unit (2,5-anhydro-D-

M Linked to Chitosan residue via glycosidic bond mannose)

Comprises an S atom, e.g., thioacetal forming an n-membered ring, n being from 4 to 9, or linear

Zl Chemical transformation

thioacetal, optionally ramified and/or

substituted, thiazolidine, and the like

alkyl chain, C C^o, and/or alkyl n-membered ring, n being from 4 to 9, optionally ramified

LI * Optional linker and/or substituted; optionally including at least one S atom such as -S-CH₂-CH₂-; Cleavable or uncleavable covalent linkage; and the like

Peptide bond, thioester, thioether, Schiff base

Z2 * Chemical transformation

and the like

alkyl chain, Ci-C_]0o, and/or alkyl n-membered ring, n being from 4 to 9, optionally ramified

L2 * Optional linker or spacer and/or substituted; optionally including at least one S atom such as -S-CH₂-CH₂-; Cleavable or uncleavable covalent linkage; and the like Z3 * Chemical transformation cleavable or uncleavable covalent linkage

polypeptide, nucleic acid, planar or particulate

POLY ligand

surface, polymers and the like

Z4 * Chemical transformation Between POLY and Y

Active/passive targeting (fusogenic peptide,

Y * Optional ligand folate TAG, Galactose, and the like) and/or

dyes/fluorophores and/or polymers, and the like

* Denotes that the presence of this group is optional

[54] In a non-limiting embodiment of the present disclosure, there is provided a method of making such chitosan conjugates which makes use of a thiol-based chemistry. The proposed thiol- based chemistry may advantageously overcome at least some of the limitations of the hereinbefore discussed oxime-click pathway.

[55] Without being bond to any particular theory or scheme, it is believed that thiol moieties are highly reactive towards double bonds as well as towards carbonyl groups in aqueous conditions at pH as low as 1 where CS amines are present only in the ionized and non-reactive form (Lienhard, G. E.; Jencks, W. P., Thiol Addition to the Carbonyl Group. Equilibria and Kinetics 1. Journal of the American Chemical Society 1966, 88 (17), 3982-3995). Moreover many equilibrium measurements have demonstrated the ability of thiols to add to the carbonyl group more efficiently than other nucleophiles (e.g., hydroxyls or amines) in both acid- and base-catalyzed pathways (Sander, E. G.; Jencks, W. P., Equilibria for additions to the carbonyl group. Journal of the American Chemical Society 1968, 90 (22), 6154-6162).

[56] Thiols react with either aldehydes or ketones producing hemithioacetal products through a double equilibrium, as depicted in Scheme 2.

Scheme 2 [57] Acid catalyzed hemithioacetal formation takes place optimally below pH 3 (Lienhard, G. E.; Jencks, W. P., Thiol Addition to the Carbonyl Group. Equilibria and Kineticsl . Journal of the American Chemical Society 1966, 88 (17), 3982-3995) and the final product is unstable under alkaline conditions, since the attack of hydroxide ions readily reverts the product to the starting reactants (Barnett, R. E.; Jencks, W. P., Base-catalyzed hemithioacetal decomposition at a diffusion-controlled rate. Journal of the American Chemical Society 1967, 89 (23), 5963-5964; Caraballo, R.; Dong, H.; Ribeiro, J. P.; Jimenez-Barbero, J.; Ramstrom, O., Direct STD NMR Identification of β-Galactosidase Inhibitors from a Virtual Dynamic Hemithioacetal System. Angewandte Chemie 2010, 122 (3), 599-603).

[58] As mentioned previously, the reactive species is the dehydrated carbonyl compound so that dehydration and hemithioacetal formation represent the rate limiting steps of this pH dependent process (Lienhard, G. E.; Jencks, W. P., Thiol Addition to the Carbonyl Group. Equilibria and Kineticsl . Journal of the American Chemical Society 1966, 88 (17), 3982-3995; Schubert, M. P., Compounds Of Thiol Acids With Aldehydes. Journal of Biological Chemistry 1936, 114 (1), 341- 350).

[59] Hemithioacetals can be stabilized by thioacetal formation via a second thiol nucleophilic attack (intra- or inter-molecular) associated with the release of water (Campaigne, E., Chapter 14 - Addition Of Thiols Or Hydrogen Sulfide To Carbonyl Compounds. In Organic Sulfur Compounds, Kharasch, N., Ed. Pergamon: 1961 ; pp 134-145; Fournier, L.; Lamaty, G.; Natat, A.; Roque, J. P., Addition des thiols sur les cetones-III: Reinvestigation du mecanisme de I'addition du mercapto-2/ethanol. Tetrahedron 1975, 31 (8), 1025-1029), as depicted in Scheme 3.

Scheme 3

[60] This chemical process is widely used in organic synthesis as a carbonyl group protection strategy but is more conveniently performed in anhydrous organic solvent (Peter G. M. Wuts, T. W. G., Protection of the Carbonyl Group - Dithioacetals and Ketals. In Protective groups in organic synthesis, Wiley, Ed. 2006; pp 477-500). To the best of the inventors' knowledge, such a strategy has not been implemented yet in aqueous conditions for polymer derivatization.

[61] Advantageously, the process for manufacturing a conjugate of chitosan residues described in the present disclosure can have at least one of the following benefits over the oxime click method developed previously: it can be used for CS derivatization without interfering with amine groups that are fully protonated and thus unreactive; it is efficient in aqueous media; and there is no need for an additional chemical treatment to stabilize the products. In the present disclosure, despite the presence in trace amounts of the reactive aldehyde moiety on the M-Unit, CS-thiol adducts were unexpectedly produced with at least 50% coupling rates within acidic aqueous solvent, for example, using linear thioacetal the inventors unexpectedly obtained CS-thiol adducts with about 55%) to about 70% coupling rates (intermolecular pathway in aqueous conditions) and using thiol hook/triskelion the inventors unexpectedly obtained CS-thiol adducts with about 70%» to about 90%) conjugation efficiencies (intramolecular pathway in aqueous conditions with a co-solvent addition to solubilize the linkers).

[62] In one non-limiting practical embodiment, the herein described reaction medium can also include a co-solvent such as any suitable polar protic and/or aprotic co-solvent. For example, the suitable polar protic and/or aprotic co-solvent can be selected from, but without being limited thereto, methanol, ethanol, 2-propanol, butanol, isobutanol, tert-butanol, tetrahydrofuran, dioxane, dichloromethane, and any combination thereof. The person of skill will readily be able to select a suitable co-solvent without undue effort.

[63] In one non-limiting practical embodiment, the above co-solvent is present in the reaction medium at a proportion selected from the range of > 0 v/v % to about 95 v/v %>. Preferably, the person of skill may wish to use a lower proportion in water. The person of skill will readily be able to select a suitable proportion of co-solvent depending on the particular properties of the co- solvent without undue effort. [64] The presence of such co-solvent in the reaction medium can be advantageous for example for solubilisation of the herein described triskelion linker prior to conjugation with M-Unit and/or M-Unit chitosans.

[65] Figure 1 represents a non-limiting schematic representation of a chitosan conjugate and a reaction scheme to make same in accordance with a non-limiting embodiment of the present disclosure. This chitosan conjugate includes a chitosan residue having a single M-Unit obtained via depolymerization by nitrous acid (HONO), the M-Unit being linked to a thioacetal (Zl) and each S atom in the thioacetal being linked to a respective ligand, which in this case is represented with a polyethylene glycol (PEG) molecule. Without being limited to a particular theory, and as discussed elsewhere in this text, it is believed that in this reaction scheme, the ligand bearing a thiol moiety (e.g., PEG-SH) will attack the chitosan residue M-Unit aldehyde in aqueous conditions to form an unstable hemithioacetal intermediate, which will be stabilized by a second attack of another ligand bearing a thiol moiety (e.g., PEG-SH) to form the linear thioacetal structure (Zl). The reaction scheme of Figure 1 provides a conjugate with two ligands per conjugate molecule.

[66] Figure 2 is a non-limiting schematic representation of another chitosan conjugate and a reaction scheme to make same in accordance with a non-limiting embodiment of the present disclosure. This chitosan conjugate includes a chitosan residue having a single M-Unit, the M-Unit being linked to a thioacetal (Zl) and the S atoms in the thioacetal forming an n-membered ring, where n can be from 4 to 9, and in this particular case n = 6, and where a peptide bond (Z2) links the ring and the ligand, which in this case is represented with a PEG molecule. Without being limited to a particular theory, and as explained elsewhere in this text, it is believed that in some cases, in the reaction scheme of Figure 1, the structure stabilization by the second POLY-SH attack may be sterically hindered by the presence of the first POLY-SH attached to the M-Unit and that a reaction according to the scheme depicted in Figure 2, where the ligand includes a "thiol hook" ((SH-CH₂)₂CHR) may provide higher coupling rates. To implement the reaction scheme of Figure 2, the ligand is functionalized with the thiol hook prior to conjugation with the chitosan residue. In the particular case depicted in Figure 2, the functionalized ligand is represented with a functionalized PEG (PEG-NH-thiol hook). The reaction scheme of Figure 2 provides a conjugate with a single ligand per conjugate molecule.

[67] Figure 3 is a non-limiting schematic representation of other chitosan conjugates and a reaction scheme to make same in accordance with a non-limiting embodiment of the present disclosure. These chitosan conjugates include a chitosan residue having a single M-Unit, the M- Unit being linked to a thioacetal and the S atoms in the thioacetal forming an n-membered ring, where n can be from 4 to 9, and in this particular case n = 6, and where a peptide bond links the ring to the ligand (in this case represented with PEG) via either a cleavable linkage represented here with -S-S- (to form in this case a CS-b-SS-PEG conjugate) or with an uncleavable covalent linkage represented here with -S-CH₂-S0₂- (to form in this case a CS-b-PEG).

[68] Advantageously, a cleavable linkage (e.g., a disulphide linkage) may be cleaved upon exposure to a chemical trigger (e.g., exposure to a reductive environment, such as upon entry in an endosome), whereas an uncleavable covalent linkage (e.g., S-CH₂-S0₂-) may be cleaved upon exposure to a physical trigger (e.g., exposure to light, such as a laser) thus affording some control over the chitosan conjugate's stability and/or release of transported molecular cargo (e.g., nucleic acid molecules). Peter et al. Protective groups in organic synthesis, Wiley, Ed. 2006; pp 477-500.

[69] In Figure 3, the cleavable or uncleavable covalent linkage links the peptide bond to the ligand, which in this particular case is represented with a PEG molecule. In Figure 4, the cleavable or uncleavable covalent linkage links the peptide bond to the ligand, which in this case is represented with a functionalized surface, for example a nanoparticle surface or a planar surface. In Figure 5, the cleavable linkage links the peptide bond to a second chitosan residue which may have identical characteristics as those of the first chitosan residue. In Figure 6, the cleavable linkage (a disulphide linkage) links the peptide bond to a siR A. While these non-limiting embodiments all include a peptide bond, the person of skill will readily understand that this bond is optional. The person of skill will also be able to envision other chitosan conjugates having alternate structures based on the herein teachings.

[70] Figures 3-6 are discussed in further details in the following paragraphs. [71] Figure 3 also represents a schematic representation of the reaction scheme for making block- copolymers in accordance with a non-limiting embodiment of the present disclosure. This reaction scheme will be referred hereinafter as the "Triskelion" reaction scheme. Advantageously, this reaction scheme has an increased coupling rate relatively to the reaction scheme of Figure 1 , since it involves in a first step coupling a small linker (small linkers are more reactive than long ligand chains and they can be used at much higher concentration), and coupling the ligand (POLY) in a second step having smoother conditions than the first step, where for example the first step includes a reaction at a pH value of about 1. As discussed previously, this reaction scheme also allows the introduction of a cleavable or an uncleavable covalent linkage in the chitosan conjugate.

[72] The "Triskelion" strategy involves a two-step process: Firstly, a Triskelion molecule bearing 3 or more thiol groups (preferably 3 thiol moieties) is conjugated to the CS M-Unit aldehyde. This conjugation takes place at about pH 1, and it results in a functionalized CS chain (CS-b-Triskelion) that bears a highly reactive chemical moiety (the 3^rd unreacted thiol group) on its end group. Secondly, once the CS-b-Triskelion has been synthesized, further conjugations with POLY can be performed in milder conditions (pH 4-6).

[73] The Triskelion reaction scheme of Figure 3 includes providing as starting materials, a "Triskelion" molecule and a chitosan residue having a single M-Unit including a reactive aldehyde. In the specific example depicted in Figure 3, the Triskelion molecule includes a thiol hook linked to an -CH₂-CH₂-SH moiety (L2) via a peptidic bond (72). The person of skill will however be able to readily envision other suitable Triskelion structures without undue effort. The starting materials are then reacted, as further explained in more detail later in this text, via a step (a) to obtain an intermediate CS-b-Triskelion which includes, in this specific example, a chitosan residue where the M-Unit is linked to a thioacetal (Zl), where the thioacetal forms an n-membered ring, which in the particular case depicted in Figure 3 has an n = 6, and the ring is linked to the -CH₂-CH₂-SH moiety via the peptidic bond (Z2). The CS-b-Triskelion intermediate is then reacted, as further explained in more detail later in this text, via a step (b) either with a ligand functionalized to include a cleavable linkage (Z3) thus forming a CS-b-SS-ligand conjugate (in this case the ligand is represented by PEG) or via a step (c) with a ligand functionalized to include an uncleavable covalent linkage (Z3) thus forming a CS-b-ligand conjugate (in this case the ligand is also represented by PEG).

[74] Figure 4 (top) is a schematic representation of a prior art grafted chitosan residues on a surface, for example a planar surface, a nanoparticle surface, and the like. Conventional techniques for grafting chitosan residues to surfaces usually result in chitosan residues which are grafted, adsorbed or electrostatically bound in a mostly longitudinal configuration (parallel to the surface). In contrast, the Triskelion reaction scheme described here affords chitosan residues which are grafted in a mostly transverse configuration (perpendicular to the surface). This is better illustrated in Figure 4 (bottom) which is a schematic representation of a chitosan conjugate having a cleavable linkage S-S or an uncleavable covalent linkage S-CH₂-S0₂- which is grafted in a mostly transverse configuration onto a surface. The reaction scheme to obtain this conjugate includes reacting a CS-b-Triskelion intermediate with a functionalized surface, which includes a functional group comprising the cleavable or the uncleavable covalent linkage. Advantageously, a chitosan conjugate grafted in a mostly transverse configuration as described here may have its amines and other chitosan residue functional group(s) free to interact with their environment and may therefore provide novel properties in terms of bioactivity relatively to those chitosan conjugates of the prior art bound in a mostly longitudinal configuration.

[75] Figure 5 represents a non-limiting example of a double chitosan residue conjugate and a reaction scheme for making same. The conjugate includes a first chitosan residue where the M- Unit is linked to a thioacetal, where the thioacetal forms an n-membered ring, which in the particular case depicted in Figure 5 has an n = 6, and the ring is linked to a CH₂-CH₂-S-S-CH₂- CH₂ moiety (cleavable linkage) via a peptidic bond, and the cleavable linkage is linked to a second chitosan residue, which is this case is represented with an identical structure as the first chitosan residue. The reaction scheme to obtain this conjugate is based on the previously described Triskelion reaction scheme. In particular, two CS-b-Triskelion intermediates, each having a molecular weight of n kDa, are reacted with each other (oxidation) so as to obtain a chitosan conjugate having a molecular weight of 2n kDa, which includes two chitosan residues. A chitosan conjugate according to this embodiment can be particularly advantageous for cellular delivery of nucleic acid molecules since the the disulfide linkage between the two chitosan residues can be cleaved within the endosome (reductive environment), favoring the release of the nucleic acid molecules.

[76] Figure 6 represents a non-limiting example of a chitosan conjugate which includes a nucleic acid molecule, represented in this case with a siRNA molecule. The conjugate includes a chitosan residue where the M-Unit is linked to a thioacetal, where the thioacetal forms an n-membered ring, which in the particular case depicted in Figure 6 has an n = 6, and the ring is linked to a CH₂-CH₂- S-S-CH₂CH₂ moiety (cleavable linkage) via a peptidic bond, and the cleavable linkage is linked to the siRNA molecule. The reaction scheme to obtain this conjugate is based on the previously described Triskelion reaction scheme. In particular, the starting materials include the CS-b- Triskelion intermediate and a nucleic acid molecule modified to include a linker carrying a thiol group or another disulfide linkage (which in this case is represented with an siRNA having a 2- disulfanepyridine - C₅H₅NS₂). In this embodiment, the chitosan residue may have any molecular weight from 0.2 to 5 kDa so long as the chitosan residues is soluble in water when neutralized and can be attached this way without any electrostatic interactions between the chitosan residue and the nucleic acid. While an siRNA is shown here, the person of skill will understand that any other suitable nucleic acid molecule can be used, for example, an shRNA, mRNA, oligonucleotide, linear DNA, and the like. Such chitosan conjugate may be used, for example, for administration to a subject, such as by subcutaneous injection.

[77] Figure 7 represents a non-limiting of a flowchart of an experimental design for (A) mechanistic studies and (B) Chitosan M-Unit reactivity studies, which are further discussed later in the Examples:

(A) Mechanistic studies

[78] Glucosamine (GlcNH₂) was treated with nitrous acid (HONO) to form the 2,5-anhydro-D- mannose (M-Unit). The M-Unit was reacted with 2 thiol-bearing molecules (β-mercaptoethanol and 3-mercaptopropionic acid, BME and MP A, respectively). The reaction products were treated using one of 3 methods: a. Method I: Direct LC-MS analyses to determine to which extent thioacetal formation occurs in situ; b. Method II: Freeze-drying (FD) + LC-MS analyses to assess the effect of FD on the thioacetal proportion and to ascertain that no by-products appear post FD; and c. Method III: Acetate buffer pH 4 + FD + LC-MS analyses to determine the effect of an increase in pH prior to FD (this pH increase was included here to prevent any CS acid hydrolysis that could occur when Method II, i.e. FD at pH 1, would be transposed to the polymer).

(B) Chitosan M-Unit reactivity

[79] CS 92-200 was depolymerized with nitrous acid (HONO) to produce CS 92-2 HCl salt bearing the M-Unit at the cleaved end of the polymer. M-Unit CS 92-2 HCl salt were reacted with MP A and BME and the reaction products treated with one of 3 workups: a. Workup I: Dialysis vs. HCl ImM solution + FD to remove all thiol model excess and to determine the in situ thioacetal formation rate; b. Workup II: FD + Dialysis vs. HCl ImM solution + FD to determine the effect of FD on the functionalization rate; c. Workup III: Acetate buffer pH 4 + FD + Dialysis vs. HCl ImM solution + FD to determine the effect of an increase in pH prior to FD on the functionalization rate (this pH increase was included to prevent any CS acid hydrolysis that could occur during FD at pH 1 in Workup II).

[80] The degree of functionalization of the CS conjugates was determined by 1H NMR, whereas covalent conjugation was assessed by DOSY NMR experiments and Ellman assays in order to rule out the possibility of a simple physical mixture of reagents. Examples

1. Reagents, Materials

[81] Chitosan with a degree of deacetylation (%DDA) of 91.7%, Mn=193 kg.mol-1 (PDI=1.256) and 99.5%, Mn=0.8 kg.mol-1 (PDI=1.245) was provided by Marinard Biotech Inc. Deuterium oxide (Cat #151882), Deuterium chloride 35 wt. % in deuterium oxide (Cat #543047), Sodium nitrite (Cat #431605), Hydrochloric acid standard solution - 1.0 N in H₂0 (Cat #31,894-9), Hydrochloric acid 37% (Cat #320331), Sodium hydroxide solution l .OM (Cat #319511), Sodium acetate (Cat #241245), DTNB (5,5'-dithiobis-(2-nitrobenzoic acid)) (Cat #D8130), GlcNH₂ D-(+)- Glucosamine hydrochloride 99% (Cat #C-1276), MP A (3-Mercaptopropionic acid) >99% (Cat #63768), BME (β-Mercaptoethanol) (Cat #M6250), Sodium acetate trihydrate BioXtra (Cat #S7670), Dowex^® 50WX8-100 [H+] (Cat #217506), Dowex^® 1X8-50 [C1-] (Cat #217417), EDT (1,2-Ethanedithiol) >98.0% (Cat #02390), PDT (1,3-Propanedithiol) 99% (Cat #P50609), 2- Propanol anhydrous 99.5% (Cat #278475), 1,2,6-Hexanetriol 96% (Cat #T66206), Methanesulfonyl chloride >99.7% (Cat #471259), Potassium thioacetate 98% (Cat #241776), Dichloromethane anhydrous >99.8% (Cat #270997), Tetrahydrofuran anhydrous (Cat #186562), Triethylamine BioUltra >99.5% (Cat #90335), Sodium carbonate BioXtra >99.0% (Cat #S7795), Sodium chloride BioXtra >99.5% (Cat #S7653), Magnesium sulfate anhydrous ReagentPlus^® >99.5% (Cat #M7506), Ethyl acetate CHROMASOLV^® for HPLC >99.8% (Cat #439169), Cyclohexane for HPLC >99.9% (Cat #650455), N.N-Dimethylformamide anhydrous 99.8% (Cat #227056), Sodium methoxide 0.5 M in methanol (Cat #71751), Methanol for HPLC, >99.9% (Cat #34860) and Sodium azide (Cat #S2002) were purchased from Sigma-Aldrich. UltraPureTM TRIS (Cat #15504-020), Glacial acetic acid (Cat #351271-212) and Spectra/Por^®6 dialysis membrane (MWCO=1000 Da, Cat #132640) were purchased from Life Technologies, Fisher Scientific and Spectrum Labs respectively. mPEG-SH 2 kDa and the plasmid DNA (pDNA) pEGFPLuc were purchased from JenKem Technology USA and from Clontech Laboratories, respectively.

2. Characterization

2.1. Mass spectrometry [82] Liquid chromatography - mass spectrometry (LC-MS) data were acquired on an Agilent 6224 LC-TOF mass spectrometer in positive electrospray ion mode, coupled to an Agilent 1260 series liquid chromatography system (Agilent Technologies). Mass Hunter B.06 software (Agilent Technologies) was used to process data. Separations were carried out at 50 C on a XSELECT CSH™C18 column (4.6 x 100mm, 5μηι particles) from Waters. The auto-sampler was maintained at 15 C to avoid sample degradation. The eluents consisted of 0.1% formic acid in water (eluent A) and 0.1% formic acid in acetonitrile (eluent B). The initial mobile phase contained 1% eluent B and was held for 3 min. Eluent B content was increased from 1 % to 20 % from 3 to 5 min then from 20 % to 80 % from 5 to 7 min. The system returned to the initial conditions at 7.2 min and was held constant for up to 15 min to allow column equilibration. The injection volume was 1-3 μί. A needle wash solution containing methanol: water (60:40 v/v) was used after each injection to reduce carry-over. Mass spectra were acquired for m/z ranging from 50 to 1200.

[83] Liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) experiments were performed on a Thermo Scientific Quantum Ultra triple quadrupole mass spectrometer operated in positive electrospray ion mode, equipped with a Thermo Scientific Surveyor liquid chromatography system. Xcalibur software (Thermo Scientific) was used to process data.

Separations were carried out on a XSELECT CSH CI 8 column (4.6 x 100mm, 5μηι particles) from Waters operated under the same chromatographic gradients as those described above. MS/MS spectra were acquired on m/z values for protonated [M+H]⁺ and sodium adduct [M+Na]⁺ species of targeted compounds.

2.2. Chitosan characterization

[84] The deacetylation degree (%DDA) of chitosan was determined by 1H NMR spectroscopy as previously described using a Bruker Avance 500 spectrometer equipped with a Bruker 5 mm BBFO probe. Lavertu et al., Journal of Pharmaceutical and Biomedical Analysis, 2003, 32 (6), 1 149-1 158.

[85] Cross-polarization magic-angle spinning (CPMAS) and Bloch-decay (BD) ¹³C NMR spectra were collected on a Bruker Avance 600 instrument equipped with a Bruker 4mm BL4 CPMAS probe and samples were spun at the magic angle (54.7°) at a rate of 10-12 kHz. Diffusion ordered spectroscopy experiments (DOSY) were conducted on a Bruker II 400 equipped with a Bruker difBO probe, using 32 gradients between 1 1.2 and 358.4 gauss. cm^"1 with a gradient pulse (δ) of 1 ms, a diffusion time (Δ) of 60 ms.

[86] Molar mass of starting 92% DDA chitosan was determined by size-exclusion chromatography (SEC) as previously described. Lavertu et al., Carbohydrate Polymers, 2012, 87 (2), 1192-1198. Measurements were acquired on a gel permeation chromatography system equipped with LC-20AD isocratic pump, SIL-20AC HT autosampler, CTO-20AC oven (Shimadzu). This setup was coupled to the following detectors: Dawn HELEOS II multiangle laser light scattering, Viscostar II viscosimeter and Optilab rEX interferometric refractometer (Wyatt Technology Co.). The starting materials were eluted through two Shodex OHpak columns (SB- 806M HQ and SB-805 HQ) connected in series with a mobile phase composed of 0.15M acetic acid, 0.1M sodium acetate, 0.4mM sodium azide, 0.1M NaCl, pH 4.5. Nguyen et al., Carbohydrate Polymers, 2009, 75 (3), 528-533. A dn/dc value of 0.214 (DDA=92%) was used and the number and weight average molar masses (Mn and Mw) of the CS starting materials were found to be 193 kg.mor' and 242.5 kg.mol^"1 respectively.

[87] Modified CS (depolymerized CS and thiol coupled CSs) were analyzed in SEC using the same conditions but with columns SB-806M HQ and SB-803 HQ that are more suitable for the analysis of low molecular weight chitosans.

2.3. Nanoparticles characterization

[88] Average diameters (Z-Average) of chitosan/pDNA and CS-b-PEG₂/pDNA polyplexes were determined by dynamic light scattering (DLS) at an angle of 173° at 25°C, using a Malvern Zetasizer Nano ZS (Malvern, Worcestershire, UK). Samples (N = 2) were measured in triplicates using the viscosity of pure water in calculations.

[89] Environmental scanning electron microscopy (ESEM) imaging of the polyplexes were performed as previously described on an environmental scanning electron microscope, Quanta 200 FEG (FEI Company Hillsboro, OR) operated in high vacuum mode with accelerating voltage = 20.0 kV; spot size = 3 and working distance = 5mm. Niebel et al., Biomacromolecules 2014, 15 (3), 940-947.

3. Aldehyde availability

3.1. Chitosan depolymerization using deuterated species for direct Ή NMR measurements

[90] The depolymerization reaction was performed in deuterated solvent for direct M-Unit CS aldehyde detection by 1H NMR spectroscopy without further processing post-reaction. Chitosan with 92% DDA and Mn = 200 kg.mol^"1 (CS 92-200) was depolymerized using nitrous acid in deuterated solvent to achieve a specific number-average molar mass (Mn) target of 1 kg.mol^"1 (CS 92-1). These short CS chains were used to facilitate the detection and the quantification of aldehyde end groups. Chitosan (202.5 mg) was dissolved in 37.9 mL D₂0 and 170 μΐ_^ of DC1 35% (w/w) at 50°C. Then 2.435 mL of fresh sodium nitrite solution (10 mg.mL^"1 in D₂0) was added to the dissolved CS to reach 0.5% (w/v) chitosan concentration. These conditions correspond to a GlcNH₂:HONO molar ratio of 3. The mixture was stirred for 3h at 50°C. The pD (pD=pH+0.4) of the depolymerization medium was ca. 1.9 at the end of the reaction.

1H NMR (500 MHz, D₂0/DC1, 25°C, ns=2000, dl=6s, acquisition time=2s) δ 2.06 (s, 1.38H, NHAc), 3.13-3.19 (br, 4.5H, H2D), 3.49-3.51 (br, 1H, H2A), 3.73-3.95 (m, 27H, H3-H6), 4.12-4.13 (q, J=5.1Hz, 1H, H5M), 4.22 (t, J=3.9 Hz, 1H, H4M), 4.44 (t, J=3.9 Hz, 1H, H3M), 4.58 (br, 0.5H, HI A), 4.79-4.88 (m, 4.5H, HID), 5.09 (d, J=5.3Hz, 0.98H, HIM Gem-diol).

SEC-MALLS : Mn = 823 (± 41) g.mol^"1 ; Mw = 1024 (± 28) g.mof¹ ; PDI = 1.245 (± 0.027)

4. Thiol reactivity towards M-Unit CS aldehyde

4.1. Mechanistic evaluation of chitosan thioacetylation by mass spectrometry (Figure 7).

[91] The CS terminal end-group (2,5-anhydro-D-mannose) formed after HONO depolymerization was derivatized with thiol-bearing model molecules (BME and MP A). Since the expected products have similar structures, their sensitivity to ionization should be equivalent. These derivatized M- Unit products were analyzed in a semi-quantitative way by comparing the chromatogram integration peaks of specific m/z values corresponding to both proton ([M+H]+ and sodium adducts [M+Na]+) within the same run.

4.1.1. 2,5-anhydro-D-mannose (M-Unit) synthesis

[92] 2,5-anhydro-D-mannose was synthesized according to Claustre et al. Carbohydrate Research 1999, 315, 339-344. Briefly, GlcNH₂ HCl (5 mmol, lg) was dissolved in 25 mL degassed ddH₂0 and was allowed to stir overnight at room temperature. The colorless reaction medium was cooled down to 0°C and NaN0₂ (12.5 mmol, 862 mg) was added. Dowex^® 50WX8- 100 [H+] resin (42.5 mmol, 8.85 g dried, 25 mL) was added slowly under stirring and the heterogeneous mixture stirred for 4h at 0-5 °C. The H+ resin was removed by filtration and the filtrate was neutralized with Dowex^® 1X8-50 [C032-] resin (60 mmol, 17.14 g dried, 50 mL), flash-frozen and freeze-dried to give the expected yellowish solid with 85% yield.

1H NMR (500 MHz, D₂0, 25°C, ns=64, dl=6s, acquisition time=2s) δ 3.36-3.40 (m, 2H, H6), 3.91-3.95 (m, 2H, H2 & H5), 4.05-4.08 (t, J=5.6 Hz, 1H, H4), 4.18-4.21 (t, J=5.8 Hz, 1H, H3), 5.09-5.10 (d, J=5.4Hz, 0.88H, HI Gem-diol), 8.46 (s, 0.12H, HI Aldehyde).

MS (ESI+) : [M+H+] = 163.0625; [M+Na+] = 185.0460 (Expected: [M+H+] = 163.0601 ; [M+Na+] = 185.0420)

4.1.2. 2,5-anhydro-D-mannose (M-Unit) conjugation with thiol-bearing molecules

[93] The synthesized 2,5-anhydro-D-mannose M-Unit (0.1 mmol, 16.2 mg) was dissolved in 5 mL degassed ddH₂0. The pH of the solution was adjusted to 1 with 3M HCl solution prior to the addition of the thiol-bearing molecule (0.5 mmol, 53.2 L for MPA and 35.1 iL for BME). The pH was readjusted to 1 with 3M HCl solution. The reaction mixture was stirred for 72h at 25°C, under Ar atmosphere and covered with aluminum foil. The reaction mixture turned clear pink- orange after 72h and was split into 3 parts (Methods I, II and III): the first was dedicated to the direct LC-MS analysis of the reaction medium in order to determine the thioacetal proportion in resulting conjugates that formed in situ; the second one was directly flash-frozen and then freeze- dried prior to LC-MS analyses to assess the effect of FD on the thioacetal proportion in resulting conjugates and to ascertain that no by-products appear post FD, whereas the third was treated with 1M acetate buffer pH 4 before flash-freezing and freeze-drying in order to determine by LC-MS the effect of an increase in pH on the resulting conjugates. It is worth mentioning that Method III was included to prevent any CS acid hydrolysis that could occur when Method II, i.e. FD at pH 1, would be transposed to the polymer.

4.2. Chitosan end-group reactivity

[94] Chitosan with 92% DDA and Mn = 200 kg.mol-1 (CS 92-200) was depolymerized with nitrous acid (HONO) to a final molar mass of 2 kg.mol-1 (CS 92-2). The final product was kept in its hydrochloride salt form by dialysis vs. HCl ImM solution and freeze-drying to minimize CS amines reacting with the M-Unit. This low 2kDa molar mass was chosen in order to facilitate the elimination by dialysis of unreacted model thiols from the reaction mixture. The CS hydrochloride salt, carrying the 2,5-anhydro-D-mannose unit (M-Unit), was allowed to react at pH 1 with the two thiol models: 3-mercaptopropionic acid (MP A) and β-mercaptoethanol (BME). Each reaction was allowed to stir for 72h, at two different temperatures (25 and 50°C) under Ar atmosphere and were treated using three different workups: Workup I) Dialysis vs. HCl ImM solution followed by freeze-dry (FD) to remove all thiol-bearing molecule excess and to determine the in situ thioacetal functionalization degree); Workup II) Direct FD, dialysis vs. HCl ImM solution and a final FD to determine the effect of FD on the functionalization degree; Workup III) 1M Acetate buffer pH 4 addition to protect CS from acid hydrolysis during FD at pH 1, dialysis vs. HCl ImM solution and another FD to assess the effect of an increase in pH prior to FD on the functionalization rate. All conditions implemented are summarized in Figure 7. Each final product was characterized by ^]H NMR, Diffusion Ordered Spectroscopy (DOSY) (5 mg.mL-1 with 2% DC1 in D₂0), SEC-MALLS (1 mg.mL-1 in duplicates) and free thiol content was determined by Ellman assay (before and after Zn/HCl treatment to reduce any disulfide bond that would not have been detected by the Ellman method). Erlandsson et al., Int J Pept Res Ther 2005, 11 (4), 261-265.

[95] The following protocols describe non-limiting examples of CS preparation as well as non- limiting examples of conjugation reactions.

M-Unit CS 92-2 HCl salt synthesis [96] Chitosan was depolymerized using nitrous acid (HONO) to achieve specific number-average molar mass targets (Mn) of 2 kg.mol-1. For depolymerization, chitosan (1 g) was dissolved in 184.5 mL ddH₂0 and 9.54 mL HCl IN solution at 50°C. Then 5.975 mL of fresh sodium nitrite solution (10 mg.mL-1 in ddH₂0 obtained by solubilization of 76.5 mg NaN0₂ in 7.65 mL ddH₂0) were added to the completely dissolved CS to reach 0.5% (w/v) chitosan concentration. These conditions correspond to a GlcNH₂:HONO molar ratio of 6. The viscous colorless mixture was stirred for 3h at 50°C. The reaction medium was then dialyzed 5x against 4L of an aqueous solution of HCl at pH 3 (HCl ImM solution) over 2 days. The resulting colorless solution was flash-frozen with liquid nitrogen and freeze-dried over 3 days to give the desired white powder with 60-70% yield.

Ή NMR (500 MHz, D₂0, 70°C, ns=64, dl=6s, acquisition time=2s) δ 2.06 (s, 3.16H, NHAc), 3.14-3.21 (br, 13H, H₂D), 3.51-3.56 (br, 1H, H2A), 3.68-3.95 (m, 70H, H3-H6), 4.12 (br, 1H, H5M), 4.21-4.31 (br, 1H, H4M), 4.43 (br, 1H, H3M), 4.61 (br, 1H, HI A), 4.87-4.89 (m, 13H, HID), 5.08 (d, J=5.0 Hz, 1H, HIM Gew-diol).

SEC-MALLS: Mn = 2342 (± 11) g.mol-1 ; Mw = 3117 (± 4) g.mol-1 ; PDI = 1.332 (± 0.008)

4.2.2. M-Unit CS 92-2 HCl salt conjugation with thiol-bearing molecules

[97] CS 92-2 HCl salt (0.035 mmol, 70 mg) and thiol-bearing model molecules (0.175 mmol, 25.4 for MP A, 12.3 for BME) were solubilized in 1.73 mL degassed ddH₂0. The pH of the reaction medium was adjusted to 1 with 3M HCl. The reaction medium was stirred for 72h at either 25 or 50°C under Ar atmosphere. The resultant colorless liquid was directly flash-frozen with liquid nitrogen and then freeze-dried over 3 days. The freeze-dried white solid was solubilized in 5 mL ddH₂0 and dialyzed 5x against 2L HCl ImM solution to remove unreacted thiols. The colorless solution was flash frozen and freeze-dried to give the expected white solid with typically 70-80% yield.

Addition of BME (Figure 11): Ή NMR (500 MHz, D₂0/DC1, 70°C, ns=64, dl=6s, acquisition time=2s) δ 2.06 (s, 5.89H, NHAc), 2.91-2.95 (br, 2.78H, BME_CH₂S), 3.17-3.21 (br, 20H, H2D), 3.51-3.53 (br, 1H, H2A), 3.69-3.95 (m, 105H, H3-H6), 4.12-4.14 (br, 1H, H5M), 4.24-4.25 (br, 1H, H4M), 4.57-4.59 (br, 1H, H3M), 4.61 -4.62 (br, 1H, HI A), 4.91 - 4.92 (m, 20H, HID), 5.08-5.09 (d, J=5.0 Hz, 0.50H, HIM Gem-dio\).

SEC-MALLS: Mn = 3177 (± 57) g.mol^"1; Mw = 3680 (± 66) g.mol^"1; PDI = 1.160 (± 0.003).

Addition of MPA (Figure 12): 1H NMR (500 MHz, D₂0/DC1, 70°C, ns=64, dl=6s, acquisition time=2s) δ 2.06 (s, 4.18H, NHAc), 2.74-2.77 (t, J=7.1 Hz, 2.33H, MPA_CH₂- CO), 2.97-3.01 (q, J=6.8 Hz, 1.78H, MPA_CH₂S), 3.15-3.24 (br, 17H, H2D), 3.51-3.56 (br, 1H, H2A), 3.69-3.95 (m, 91H, H3-H6), 4.11 (br, 1H, H5M), 4.21-4.23 (br, 1H, H4M), 4.55 (br, 1H, H3M), 4.62 (br, JH, HI A), 4.87-4.92 (m, 17H, HID), 5.08 (d, J=5.0 Hz, 0.47H, HIM Gew-diol).

SEC-MALLS: Mn = 3053 (± 81) g.mol^"1; Mw = 3564 (± 48) g.mol^"1; PDI = 1.182 (± 0.016). 4.2.3. Ellman Assays

[98] Thiol-derivatized CSs were analyzed by the Ellman assay to assess the presence of free thiols within the products. Ellman stock solutions (50 mM sodium acetate, 2 mM DTNB) were prepared by dissolving 39.7 mg of Ellman reagent and 205.1 mg of sodium acetate in 50 mL double deionized water (ddH₂0). Tris 1M dilution buffer was prepared dissolving 6.1 g of Tris in 50 mL ddH₂0 and adjusting the pH to 8.0 using HC1 1.0 N standard solution.

[99] Thiol concentrations were measured in triplicate by mixing 50 of Ellman stock solution with 100 μΐ_^ of Tris dilution buffer and 10 \xL of sample solution. After 15 min the mixture was diluted by the addition of 840 μΐ, of ddH₂0 and the absorbance at 412 nm read using a microplate reader Tecan Infinite^® M200. Thiol concentrations were calculated from a standard curve prepared using either MPA or BME and measurements were performed in triplicates in a 96 well plate using 150 μΐ, sample volumes. The CS used as starting material was dissolved at the appropriate concentration for each sample and used as a blank. Both NaOH and Zn/HCl treatments of the CS adduct solutions were implemented on separate samples to determine the presence of hemithioacetal intermediates and any disulfide bond formation within the final product by the Ellman assay, respectively. Concentrated 1M NaOH and 1M HC1 solutions were used to minimize changes in CS concentration. After 45-60 min constant agitation of the reaction media, Ellman assays were performed using 10 of alkali sample solution for NaOH treatment. Zn/HCl treated samples were obtained by adding few of 1M HC1 (to reach pH 1) and 5 equivalents of Zn dust per CS; the supernatants were analyzed after centrifugation (lOOOg for 1 min).

4.2.4. Quantitation of CS derivatization efficiency: Functionalization degree (F) calculations

[100] The functionalization degrees (F) of each conjugation were calculated according to the following equations:

Equation 1

Where H_{mol peaks} refers to the well-defined proton peaks of the thiol-bearing molecule conjugated to CS and H_M__{Unil peaks} corresponds to the well-defined M-Unit characteristic proton peaks. Both integrations in Equation 1 are normalized to the number of protons used for the calculation, namely a and β for the thiol-bearing molecule and M-Unit, respectively.

[91] According to the mechanistic studies on the M-Unit model presented below, the hemithioacetal intermediate is fully stabilized into the corresponding thioacetal (as shown in Figure 8) after freeze-drying of the reaction mixture in acidic conditions, thus two thiol-bearing molecules per M-Unit CS salt were considered for the calculation of the functionalization degree (F). For MPA adducts, two well-defined peaks corresponding to -CH₂-S- and -CH₂-CO- protons (i.e. 8 protons) appear on the NMR spectra. However, for BME adducts, only the -CH₂-S- peak is visible on the spectra, in agreement with NMR spectrum simulation that predicts that -CH₂-CO- peak is hidden by the CS H3-H6 broad peaks. Banfi et al., Chimia 2008, 62 (4), 280-281 ; Lavertu et al., Journal of Pharmaceutical and Biomedical Analysis 2003, 32 (6), 1149-1 158.

[92] Thus, a values of 4 and 8 in Equation 1 where used for BME and MPA, respectively. For the M-Unit, the well-defined peaks corresponding to H₄M and H₅M protons were used for integration and a β value of 2 was thus used in Equation 1. From the above considerations, Equation 1 can be rewritten as Equation 2 and Equation 3 for BME and MPA conjugates, respectively:

Equation 2

(j //_{f V/ i}_, + J H_{( VJj}__{t YJ})

MPA x lOO

Equation 3

Where the protons used for integration are defined in Figure 8, for purified BME and MPA chitosan adducts.

[93] Similarly, CS PEGylation efficiency (F_PEG) was also calculated by adapting Equation 1 with the PEG characteristic peak integrations:

Equation 4

where H_PEG__0CH^ refers to the well-defined methyl protons (3 H) peaks located at the end of the PEG chain (a=6 as there are 2 PEG chains per CS).

4.3. CS-b-PEG₂ block-copolymer formation

[94] In order to reduce any mPEG-SS-PEGm disulfide bonds, mPEG-SH (2kDa, m = 200 mg, 0.1 mmol) was solubilized in 2 mL 1% TFA in ddH20. Zn (m = 9.8 mg, 0.15 mmol) was then added and the mPEG-SH solution stirred for lh. The clear colorless reduction medium was centrifuged at lOOOg for 2min prior to CS conjugation.

[95] M-Unit CS 92-10 HC1 salt (m = 100 mg, 0.01 mmol, 5 mM aldehyde) was added to the reduced mPEG-SH solution and the pH of the reaction medium was adjusted to 1 with HC1 3N solution. The reaction medium was stirred for 72h at 70°C, under Ar atmosphere. At the end of the reaction, the reaction medium was flash-frozen and freeze-dried. Unreacted mPEG-SH was discarded by reprecipitation in 5 x 45 mL CH2CI2. The remaining white pellet was dried under reduced pressure overnight.

1H NMR (Figure 12) (500 MHz, D₂0, 70°C, ns=64, dl=6s, acquisition time=2s) δ 2.06 (s, 11H, NHAc), 3.14-3.22 (br, 46H, H2D), 3.37 (s, 3.67H PEG-OCH3), 3.51-3.56 (br, 3H, H2A), 3.69 (s, 181H, PEG Chain -0-CH₂-CH₂), 3.75-3.95 (m, 238H, H3-H6), 4.12-4.14 (br, 1H, H5M), 4.21-4.23 (br, 1H, H4M), 4.61 (br, 3H, HI A), 4.88-4.90 (m, 46H, HID), 5.08 (d, J=5.0 Hz, 0.49H, HIM Gem-diol).

4.4. CS-b-PEG₂ and CS/pDNA polyplexes formation

[96] Polyplexes were prepared as previously described. Lavertu et al., Biomaterials 2006, 27 (27), 4815-4824. Briefly: CS-b-PEG₂ and depolymerized unmodified chitosan (CS 10 kDa with 92.5% DDA) stock solutions were prepared by dissolution at 0.5% (w/v) in hydrochloric acid using a glucosamine :HC1 ratio of 1 :1. Polymer stock solutions were diluted with ddH₂0 to reach the amine to phosphate ratio of 3.7 (N/P=3.7) when equal volumes of chitosan and pDNA (100 μg.mL^"1) solutions would be mixed. Both CS-b-PEG₂/pDNA and CS/pDNA polyplexes were prepared at room temperature, by adding 100

of the diluted polymer solution to 100 of the pDNA solution followed by immediate mixing by pipetting up and down. The polyplexes were analyzed for their size and morphology by dynamic light scattering (DLS) and environmental scanning electron microscopy (ESEM) lh after their formation.

Results

1. Aldehyde availability

[97] Since hemithioacetal formation requires the dehydrated aldehyde as reactive species (referred to aldehyde in this manuscript), the CS aldehyde availability was assessed by NMR spectroscopy. 1.1. Chitosan 2,5-anhydro-D-mannose unit (M-Unit) - Gem-diol ubiquity

[98] Raw CS was depolymerized using HONO to a final molar mass of 1 kg.mol^"1 (CS 92-1). This low Mn was chosen to increase the concentration of aldehyde moieties, facilitating their detection by ^]H NMR spectroscopy. The use of deuterated solvent for the depolymerization reaction in this study allowed direct NMR analysis of the reaction mixture (Figure 9). In 1H NMR spectrum, no aldehyde group was observed either at 9-9.5 ppm (the expected aldehyde proton chemical shift), nor at 8.5 ppm (for the M-Unit model) despite the use of a large number of scans (2000). However its hydrated form, the gem-diol peak at 5.09 ppm was omnipresent within the reaction medium. It is worth mentioning that the absence of the dehydrated form in the NMR spectrum is not due to a fast exchange between hydrated and dehydrated forms since both forms were detected for the 2,5- anhydro-D-mannose (M-Unit).

1.2 Gem-diol stability

1.2.1. Equilibrium is strongly displaced towards the gem-diol form for the M-Unit CS

[99] The hydrated form of the aldehyde was the only form detected in each liquid NMR analysis, either at 25°C (data not shown) or 70°C (Figure 9). It is worth mentioning that these analyses were performed in D₂0 and/or D₂0/DC1, which are favorable conditions for the hydrated form or gem- diol formation. Bell, Adv. Phys. Org. Chem. 1966, 4, 1-29. Some authors also reported an increase in the acetaldehyde carbonyl hydration equilibrium constant (K_hyd - [Gem-diol]/[Aldehyde]) from 0.85 to 0.99 when experiments are performed in ddH₂0 and D₂0 respectively, showing that the equilibrium can be displaced towards the formation of Gem-diol in deuterated solvents. Lienhard et al., Journal of the American Chemical Society 1966, 88 (17), 3982-3995; Bell, Adv. Phys. Org. Chem. 1966, 4, 1-29.

[100] In order to eliminate the contribution of the aqueous solvent on this equilibrium and to favor a displacement towards the aldehyde or unhydrated form of CS end unit, 1 kg.mol^"1 M-Unit CS HC1 salt was analyzed by solid-state NMR (CP-MAS). Fully deacetylated CS (CS 99-1) was preferred to the CS 92-1 to avoid any confusion between the carbonyl chemical shift of the acetyl peak and the aldehyde peak. The same sample was analyzed at 2 different frequencies (10 kHz and 12 kHz) to detect the eventual presence of harmonics within the spectrum. All peaks corresponded to chemical moieties attributed according to Heux et al. Biomacromolecules 2000, 1 (4), 746-751 (data not shown). The CS salt did not form any Schiff base product, as expected (since protonated amines are not nucleophilic), however no aldehyde peak was detected in these spectra.

[101] It has been reported that hydration of an aldehyde in the gas-phase can be observed at relative humidity (RH%) level as low as 5%. Axson et al., Proceedings of the National Academy of Sciences 2010, 107 (15), 6687-6692. The relative humidity of the laboratory where the experiments were performed was between 20-50%, and it could be that all aldehyde groups were transformed into gem-diols during the sample transfer and preparation. To eliminate the exposure to air humidity that might favor this formation of the gem-diol, an inert atmosphere solid state NMR experiment was implemented on an extra-dried CS 99-1 salt (freeze-dried over 3 days and then dried using Speed- Vac Plus Centrifuge at 60°C, overnight under reduced pressure). Sample preparation was performed within an Ar glove box to verify if air humidity transforms the CS terminal aldehyde into its corresponding hydrate. The solid state NMR analysis was conducted under inert atmosphere as well (constant N₂ flow). Neither the aldehyde peak (expected around 190 ppm) [Saito et al., Macromolecules 1987, 20 (10), 2424-2430] nor the gem-diol peak (expected around 90 ppm) [Temmeraas et al., Carbohydrate research 2001, 333 (2), 137-144] were visible on the spectrum. It is worth mentioning that the expected chemical shift of gem-diol falls within the range of chemical shifts corresponding to C3-C5 peaks and the former is most probably hidden by the latter (Figure 18). In order to confirm that the absence of gem diol in the spectrum was not due to an unexpected side reaction occurring in the preparation of chitosan sample, the dried sample was subsequently dissolved in D₂0 and analyzed by standard ^!H NMR. This analysis revealed that the hydrated aldehyde form was present at the expected quantitative proportion, as established from CS Mn and DDA (data not shown).

1.2.2. H-bonding could stabilize the M-Unit CS gem-diol

[102] Although for most aldehydes and ketones the hydrates are generally less stable than their respective parent [Bell, Adv. Phys. Org. Chem. 1966, 4, 1-29.], their equilibrium can be displaced towards the gem-diol form by making the carbonyl more electropositive. Thus, the gem-diol form can predominate when the aldehyde is located close to a functional group allowing a negative inductive effect. For CS, some suitable electron-withdrawing substituents, such as hydroxyl and hemiacetal substituents might create a weak negative inductive effect, thereby increasing slightly the δ+ charge on the carbon of the carbonyl and favouring water nucleophilic attack. Since CS offers significantly more H-bond donors than 2,5-anhydro-D-mannose, intermolecular H-bonding may be responsible for the strong predominance of the gem-diol form [Schulman et al., Journal of the American Chemical Society, 1976, 98 (13), 3793-3799]. This hypothesis was confirmed with the ^!H NMR analysis of the synthesized 2,5-anhydro-D-mannose that presents a detectable proportion of the aldehyde in ¹H NMR spectroscopy (around 10% of the aldehyde form, data not shown). The NMR experiments described above suggest that the M-Unit CS aldehyde is only present in trace amounts since only the gem-diol form was detected. Nonetheless these trace amounts are reactive enough to be engaged with nucleophiles such as CS amines (Schiff base formation) or more particularly with thiol moieties (see, scheme 1).

2. Mechanisms of conjugation of 2,5-anhydro-D-mannose (M-Unit) and thiol-bearing molecules

[103] The reactivity of aldehydes toward thiols in aqueous conditions was assessed semi- quantitatively by LC-MS using the 2,5-anhydro-D-mannose as an aldehyde model.

2.1. Expected products of thiol conjugation to aldehydes include hemithioacetal, thioacetal, oxathiolane and α,β-unsaturated sulfide intermediate

[104] The expected products of all conjugations implemented with thiol-bearing molecules (MP A and BME) include hemithioacetal, thioacetal, oxathiolane and α,β-unsaturated sulfide intermediate (Figure 10). The first thiol attack on the aldehyde forms a hemithioacetal intermediate (A), which is in equilibrium with its corresponding protonated hemimercaptal form (oxonium) via a proton transfer. This structure may react in several ways: it could be stabilized with a second nucleophilic attack forming the corresponding thioacetal (Q after water removal. Another hypothetical pathway is the formation of ana,P-unsaturated sulfide intermediate (D) through an elimination process. The final possible product concerns the BME adducts that could form oxathiolane derivatized adducts (B), but this possibility is slight given their fast hydrolysis compared to the thioacetal. Satchell et al, Chemical Society Reviews, 1990, 19 (1), 55-81 ; Fournier et al., Tetrahedron, 1975, 31 (8), 1025-1029.

2.2. Low in situ stabilization of hemithioacetals

[105] Five equivalents of thiolated molecules (BME and MP A) per M-Unit aldehyde/gem-diol were reacted with a synthesized 2,5-anhydro-D-mannose (M-Unit model) for 72h at pH 1, under inert atmosphere. The relative proportions of the final expected compounds were calculated from LC-MS chromatogram integrations of specific m/z values corresponding to both proton and sodium adducts ([M+H]⁺ and [M+Na]⁺) within the same run (Table 1). This semi-quantitative evaluation was possible since the expected final products have similar structures and thus expected similar ionization behaviors. Direct LC-MS analyses (Table 1) of the reaction media (Method I, Figure 7) indicated that the hemithioacetal intermediate A corresponded to the major observed compound (75%), the minor product being the stable thioacetal C (25%), after 72h reaction. A highly similar 4:1 ratio of hemithioacetal to thioacetal was observed for all thiol models (BME and MP A) tested. Thus the stabilization to the thioacetal intermediate A seems to occur with a second thiol nucleophilic attack to form the corresponding thioacetal C with the release of water. However, our results suggest that this stabilization occurs only to a relatively low extent in aqueous medium.

2.3. Freeze-drying facilitates the hemithioacetal stabilization

[106] Water removal by freeze-drying (FD) is the key-step in Schiff base formation occurring between CS amines and CS terminal aldehyde. T0mmeraas et al., Carbohydrate research 2001, 333 (2), 137-144. A similar effect might be at play in the reaction with thiolated species. In order to assess whether or not FD could favor a second thiol nucleophilic attack to stabilize the structure, Method II (Direct FD of the reaction medium, Figure 7) was implemented. This strategy resulted in the synthesis of the thioacetal C without any detectable quantity of hemithioacetal A, as deduced from LC-MS analysis (Table 1). These trends were also observed using Method III (Increase in pH with 1M acetate buffer pH 4 followed by FD, Figure 7), initially proposed to prevent any CS acid hydrolysis that could occur when this method would be transposed to the polymer CS. Reaction mixtures that were treated this way resulted in a significant increase, when compared to Method I, of the relative proportion of stabilized thioacetal C vs. hemithioacetal A, corresponding to 96% and 82% thioacetal C for BME and MP A respectively (Table 1).

[107] It is worth mentioning that the LC-MS analyses only provide the relative proportion of observed species so that similar results obtained with both Methods II and III do not necessarily corresponds to equivalent absolute conversion rates. For instance, since the hemithioacetal formation equilibrium is pH sensitive [Barnett et ah, Journal of the American Chemical Society 1967, 89 (23), 5963-5964] (increase in pH is known to displace the equilibrium towards the starting materials), the increased relative proportion of thioacetal C observed with Method III vs. Method I could be due to a reduction of the absolute amount of hemithioacetal A in the reaction mixture. The conversion degrees or functionalization degrees, are calculated below by Ή NMR of the purified conjugated polymers.

[108] The oxathiolane B and α,β-unsaturated sulfide products D appeared as traces in both Methods II and III (Table 1). LC-MS chromatograms revealed the same elution time as for thioacetals C, suggesting an in-source decomposition of BIC into their respective D form. The hypothesis that the oxathiolane B was formed within the MS apparatus by the ionization of the thioacetal C was confirmed by LC-MS/MS analyses of C adduct obtained from the reaction of M- Unit and MPA: the fragmentation of C produced compound D (data not shown).

[109] These experiments suggest that the oxonium intermediate (which is in equilibrium with the hemithioacetal intermediate) is stable enough to favor the thioacetal formation notwithstanding the unsaturated compound D formation. The freeze-drying step apparently orients the reaction towards the stable thioacetal formation, more likely due to an increase in concentration by water removal to facilitate the second nucleophilic attack.

[110] In Table 1, percentages represent the relative proportion of expected final molecules resulting from each conjugation that were implemented in triplicates (N > 3 ± SD): A) Hemithioacetal intermediate, B) Oxathiolane (for β-mercaptoethanol only), C) Thioacetal, D) α,β- unsaturated sulfide. Calculations are based on chromatogram peak integrations of both proton and sodium adducts of a specific chemical formula, m/z given in parentheses represents the thioacetal in-source decomposition observations. Method I refers to direct LC-MS analysis of the reaction medium; Method II corresponds to the direct freeze-drying (FD) of the reaction medium before analysis; Method III corresponds to an increase in pH with acetate buffer pH 4 followed by FD. With both models, the hemithioacetal intermediate is stabilized by FD into the corresponding thioacetal. LC-MS/MS experiments rule out the possible formation (post-FD) of both oxathiolane and ,β-unsaturated sulfide (B and D forms in Figure 10, respectively).

TABLE 1 : Expected product (Figure 10) proportions as deduced from LC-MS analyses.

3. M-Unit Chitosan HC1 salt reactivity

3.1. Chitosan HC1 salt maintains the M-Unit integrity after freeze-drying

[111] The 2,5-anhydro-D-mannose unit (M-Unit) resulting from CS depolymerization using HONO is not stable after rehydration in aqueous acidic conditions. Indeed, when the reaction medium is neutralized, the reaction between CS amines and the M-Unit aldehyde moiety produces a reversible imino bond (Schiff base formation), which is accompanied with the release of water (scheme 1). It has been demonstrated that after FD, which is accompanied by Schiff base formation via equilibrium displacement, the solubilization of CS in acidic conditions (pH below 5) cleaves 2,5-anhydro-D-mannose unit from CS into hydroxymethyliurfural (HMF) [Temmeraas et al., Carbohydrate research 2001, 333 (2), 137-144] (scheme 1). In terms of reactivity, the M-Unit is available within the reaction medium after HONO treatment but its concentration is limited to that of the depolymerization medium (0.5% w/v in our case, corresponding to a concentration of reactive units of 2.5 mM for CS with Mn = 2 kg.mol^"1). Higher CS depolymerization concentrations are possible (typically up to 2% w/v for CS with Mn of a few hundreds of kg.mol^"1) but limited by the high viscosity of CS solutions, which may compromise stirring efficiency and homogeneity of the depolymerization medium. In order to maintain the M-Unit integrity and to work in a more concentrated regime, the depolymerized (i.e. less viscous) CS hydrochloride salt was freeze-dried, with all CS amines protonated, thus avoiding Schiff base formation and subsequent HMF formation upon rehydration. All the CSs that were prepared this way still carried their M-Unit after rehydration (M-Unit remaining > 80%), allowing higher CS concentration than the depolymerization medium (4% w/v vs. 0.5% w/v, respectively).

3.2. CS 92-2 HC1 salt M-Unit thioacetylation confirmed by NMR spectroscopy

[112] Five equivalents of thiolated molecules (BME and MP A) per M-Unit aldehyde/gem-diol were reacted with the M-Unit CS 92-2 HC1 salt for 72h at pH 1, in inert atmosphere. Resultant reaction media were treated according to 3 different workups: Workup I: Dialysis vs. HC1 ImM solution + FD; Workup II: FD + Dialysis vs. HC1 ImM solution + FD; Workup III: Increase in pH with acetate buffer pH 4 + FD + Dialysis vs. HC1 ImM solution + FD. The Workup I was proposed to evaluate the in situ thioacetal formation by removing all unreacted thiol moieties prior to FD, whereas the Workup II was implemented to assess the role of freeze-drying in hemithioacetal intermediate stabilization. Workup III was initially suggested to protect M-Unit CS HCl salt from acid hydrolysis during FD at pH 1 in Workup II, however SEC-MALLS analyses of Workup II polymers did not reveal any glycosidic linkage hydrolysis at pH 1. Other SEC-MALLS analyses involving different CS salts (Mn of 2, 4 and 10 kg.moi ¹) that were freeze-dried at low pH also did not reveal any alteration of the polymer (data not shown).

[113] The covalent nature of the conjugation of the CS HCl salt M-Unit to thiol-bearing molecules was confirmed by the Ellman assay where no free thiol moieties were detected after rehydration of the modified polymers. Note that free thiol moieties were not detected after Zn/HCl treatment that would have reduced any disulfide bond potentially formed in the course of the conjugation reaction and/or post-reaction workup. The absence of any hemithioacetal intermediate (base sensitive) was also confirmed by performing the Ellman assay on the product after exposure to 1M sodium hydroxide solution. Purified CS-thiol adducts were also analyzed by diffusion ordered spectroscopy (DOSY), a spectroscopic method that distinguishes compounds according to their respective translation diffusion coefficient. Figure 10 shows that both CS and thiol-bearing models have the same diffusion coefficient in D₂0 at 25°C, despite significant molar mass differences (2,300 g.mol^"1 vs. 106 g.mol^"1, for M-Unit CS HCl salt and MPA respectively). Altogether, the aforementioned controls confirmed the presence of the thioacetal linkage between the CS HCl salt M-Unit and both thiol-bearing model species. The results of the conjugation efficiencies between CS and BME or MPA were calculated using Equation 2 and Equation 3 respectively and are summarized in Table 2.

TABLE 2

[1 14] In Table 2, there is shown the efficiency of conjugation of the M-Unit CS HC1 salt to 5 equivalents of thiol-bearing molecules (3-mercaptopropionic acid and β-mercaptoethanol, MPA and BME respectively) per CS end unit for 72 hours at pH 1. Reaction media were treated according to the following workups: Workup I (Dialysis vs HC1 ImM solution + FD); Workup II (FD + Dialysis vs. HC1 ImM solution + FD); Workup III (Increase in pH with acetate buffer pH 4 + FD + Dialysis vs HC1 ImM solution + FD). F below corresponds to the functionalization degree, considering 2 thiol molecules per potential aldehyde and calculated using Equation 2 for BME and Equation 3 for MPA with N>3 (±SD). F was also calculated using Equation 5, considering only the relative proportion of the remaining gem-dio\ per M-Unit. (*) corresponds to the results of the conjugations implemented with 20 equivalents (instead of 5) of thiol-bearing molecule per end unit.

3.2.1. NMR and LC-MS analyses indicate that two thiol-bearing molecules regioselectively react with CS M-Unit aldehyde to form a thioacetal

[115] The regioselectivity of the CS M-Unit aldehyde conjugation to the thiol models was assessed by 2D NMR experiments (COSY and HMBC, data not shown) in order to detect long-range correlations between the M-Unit and the thiol characteristic peaks. However, such correlations were not visible in the NMR spectra, most probably because of the inherently low concentration of the end-group conjugated thiols within the synthesized structures and/or because the atoms to correlate are separated by a large number of bonds (3 and 4 for proton-carbon and proton-proton correlation, respectively - see Figure 8), especially for the COSY experiments. [Claridge, High- resolution NMR techniques in organic chemistry. Newnes: 2008; Vol. 27; Pinto et al., Recent Research Developments in Heterocyclic Chemistry, Pinho e Melo, Research Signpost, Kerala (India) 2007.] Moreover, the HMBC measurements were found to be insensitive, particularly with poorly resolved 'Η-'Η multiplets (Figure 11 and Figure 12). Bax et al., Journal of Magnetic Resonance, Series A, 1996, 119 (1), 134-138; Furrer, Chemical Communications 2010, 46 (19), 3396-3398.

[1 16] Despite the inability of these 2D NMR experiments to reveal the expected correlations, the combined NMR and LC-MS analysis indicated that two thiol-bearing molecules react regioselectively with the aldehyde of the terminal M-Unit of chitosan. As discussed above, the MS experiments performed with the mannose monomer indicated clearly that the stabilized form is the thioacetal form, so that, two thiols are expected to react similarly with the M-Unit of chitosan. This expected stoichiometry and regioselectivity for thiol-bearing molecules reacting on chitosan was validated by monitoring the relative proportion of gem-diol. Indeed, the gem-diol signal should decrease concomitantly with the conjugation of thiols onto the M-Unit of chitosan (one gem-diol consumed for two conjugated thiols). The calculated conjugation efficiencies obtained with either Equation 2 (BME) or 3 (MP A) and the following equation should therefore be the same if two thiols react regioselectively onto the terminal aldehyde function of chitosan:

Equation 5

[117] Where H₄ and H₅M are protons with well-defined NMR peaks from the M-Unit shown in Figure 8 (unchanged by the reaction of the aldehyde with thiol-bearing molecules) and G_em__Diol is the HI proton of the gem-diol form of CS M-Unit shown in Figure 9. It is worth mentioning that efficiency calculation using Equation 5 is independent of the reaction stoichiometry and relies only on the assumption that any thiol-bearing molecule will react selectively with the terminal unit of chitosan.

[118] For all conjugation reactions performed in this study, the conjugation efficiencies calculated with both approaches, namely with Equation 2 (BME) or Equation 3 (MP A), which both rely on the reaction stoichiometry, or Equation 5 that is independent of stoichiometry and relies only on the relative proportion of gem-diol vs. M-Unit, were found to be in very close agreement (Table 2). These results indicate that 1) thiol-bearing molecules react selectively with the terminal aldehyde functional group of chitosan and 2) the thioacetal is the only stable form of product observed.

3.2.2. The stabilization rate of the product from the hemithioacetal to thioacetal form within the reaction medium can be enhanced by FD [1 19] For reactions performed using 5 equivalents of thiol-bearing molecule per aldehyde, the first workup tested here (Workup I: Dialysis vs. HC1 ImM solution + FD), showed a limited conversion into the desired conjugates (F=2% and 10% as conversion degrees, for BME and MPA at 25°C respectively; Table 2). Similar results were obtained for Workup III (Increase in pH with acetate buffer pH 4 + FD + Dialysis vs. HC1 ImM solution + FD) with F=l 1% and 15% at 25°C, for BME and MPA, respectively (Table 2), while significantly higher functionalization degrees were obtained for Workup II (FD + Dialysis vs. HC1 ImM solution + FD) where F=18% at 25°C, for both BME and MPA (Table 2). Similar trends were observed for reactions performed at 50°C but with an overall increase in functionalization degrees. These results suggest that FD favors the second thiol nucleophilic attack to stabilize the hemithioacetal structure, possibly by concentrating the reaction medium. This FD effect is only seen in Workup II since in Workup I, all thiol-bearing molecules were removed by dialysis prior to FD, while in Workup III, most of the hemithioacetal intermediate was readily transformed into the starting reactants by an increase in pH. Thus, one of the reacting species is absent (or present in very low amount) during the last FD step in Workup I (thiol-bearing molecule removed with concomitant hemithioacetal formation equilibrium displacement towards the starting reactants, Figure 14) and Workup III (hemithioacetal intermediate amount reduced by pH increase) and the thioacetal form cannot be further increased by FD as compared to Workup II where both reacting species are present during FD. In fact, for Workup I and III, all observed thioacetals were mostly formed in situ, during the 72h reaction and results indicate that for reactions performed with 5 equivalents of thiol-bearing molecule per aldehyde, in situ stabilization into the thioacetal form is low.

3.2.3. Hemithioacetal-to-thioacetal conversion within the reaction medium is increased by large excess of thiol equivalents

[120] The conjugations implemented with 20 equivalents of thiol-bearing molecules per CS end unit revealed higher conversion rates (F = 55 - 70% at 50°C depending on the thiol-bearing molecules engaged) and were independent of the workup implemented (i.e. I and II, Table 2). These results also support the proposed reaction mechanism proposed in Figure 1 1. Indeed, at higher thiol concentrations, hemithioacetal intermediates and thioacetal are both favored within the reaction medium. However, in this case, FD had no significant impact on the conversion degree. Our results suggest that at high thiol concentration (20 equivalents per aldehyde) the amount of thiol-bearing molecules is sufficient to achieve significant hemithioacetal stabilization in situ. The fact that FD has no significant impact on the functionalization rate is unclear and would require additional investigations.

3.2.4. Temperature favors both hemithioacetal formation and stabilization to the thioacetal form

[121] The highest conversion degrees were obtained at 50°C, regardless of the workup implemented (Table 2). Indeed, an increase in temperature favors the hemithioacetal intermediate formation by increasing the probability of thiol-bearing molecules to react with the CS HC1 salt M-Unit aldehyde. Similarly, stabilization of the hemithioacetal intermediate occurred with an increase in temperature, favoring the second thiol model attack by increasing the probability of collisions between species. This mechanism is especially valid for the results corresponding to Workups I and III where no FD stabilization was reported. Indeed, the functionalization degree varied from 2% to 26% for BME and from 10% to 14% for MP A, for 25 and 50°C respectively. The proposed mechanism involving an equilibrium between the starting reactants and the hemithioacetal intermediate (Figure 11) is thus confirmed by this increase in conversion degree with temperature.

3.3. Effective CS PEGylation by thioacetylation of the CS M-Unit aldehyde 3.3.1. CS-b-PEG2 block-copolymer synthesis

[122] As a non-limiting implementation of the thioacetylation conjugation described herein, a 2 kDa mPEG-SH was reacted with a 10 kDa CS HC1 salt. Because of solubility limitations with these longer chains, the reaction was performed at 5 mM aldehyde instead of 20 mM that was used for the reactions between the 2 kDa CS and MPA or BME. In order to counterbalance the decrease in aldehyde concentration, the reaction was performed at 70°C (instead of 50°C) for 72h and ten thiol equivalents per aldehyde were used. After direct FD of the reaction medium and unreacted mPEG-SH removal by multiple precipitations, 1H NMR analysis of the final product (Figure 12) was performed and functionalization degree values (F) of 61% and 51% were found with Equation 4 and Equation 5 (where only the gem-diol peak integration decrease was considered), respectively.

[123] The slight discrepancy between these two values could possibly be the due to the presence of residual mPEG-SH post-purification. This hypothesis was confirmed by SEC analysis of the conjugates, where a small residual peak identified as mPEG-SS-PEGm was detected. Because PEG and CS molecular weights are close to each other, the DOSY NMR processing used to validate covalent conjugation of MPA and BME to CS was found to be inefficient for the block- copolymer (data not shown).

3.3.2. CS-b-PEG₂ block-copolymer/pDNA polyplexes are homogeneously spherical

[124] The CS-b-PEG₂ block-copolymer (CS 92-10 and 2 kDa mPEG-SH) synthesized above was used without further purification to form polyplexes with plasmid DNA (pEGFPLuc). Whereas ESEM imaging of polyplexes prepared with unmodified CS revealed various morphologies, namely toro'ids, spheres and rods, those prepared with CS-b-PEG₂ block-copolymer were substantially uniformly spherical (Figure 16). The structure modification of the polyplexes formed with PEGylated CS was also confirmed by Dynamic light scattering (DLS), where measured Z- average diameters decreased from 106 (±1) nm to 76 (±1) nm, for unmodified CS and CS-b-PEG₂ block-copolymer, respectively (Table 3).

TABLE 3

[125] Table 3 shows DLS measurements of unmodified CS and CS-b-PEG₂ polyplexes prepared with pDNA (pEGFPLuc, N/P=3.7). Samples were analyzed in triplicates (N=2, ±(max-min)/2). The size of CS-b-PEG₂ polyplexes is smaller as compared to native polyplexes.

[126] Since the PEGylated polyplexes are uniformly spherical and show a narrower size as compared to those prepared with corresponding homopolyions, these observations are consistent with the formation of micellar structures called "Block Ionomer Complexes" (BICs). Kabanov et al., Bioconjugate chemistry 1995, 6 (1), 7-20; Voets et al., Advances in Colloid and Interface Science 2009, 147-148 (0), 300-318; Pergushov et al., Chemical Society Reviews 2012, 41 (21), 6888-6901.

[127] Such particles having a size of less than a micron and which include chitosan conjugates as described herein where the ligand is a PEG molecule may be used as a cell delivery system. Advantageously, when used as a delivery system in a host blood vessels such particles have a prolongation of their circulation time within the host blood vessels relatively to known particles having a size of less than a micron, at least because these particles (i) have a net neutral charge, (ii) have less aggregation, (iii) have less protein interaction than the known particles, or (iv) any combinations thereof, which reduces recognition by the host immune system.

4. Mechanistic model for thiol-based end-group derivatization of chitosans

[128] Figure 17 illustrates a reaction scheme which summarizes a proposed mechanistic model for thiol-based end-group derivatization of chitosans described here. CS nitrous acid (HONO) depolymerization induces the formation of M-Unit that carries an aldehyde moiety at the end of the cleaved polymer (1). The equilibrium between the M-Unit aldehyde and its hydrated form (gew-diol) is strongly displaced towards the latter (2). If the CS depolymerization medium is freeze-dried at pH well below the CS pKa (i.e. pH -3-4 or below), all the CS amines are protonated and are therefore unable to react with any aldehyde group, maintaining the CS M-Unit integrity at the end of the cleaved polymer (3). Nevertheless, the equilibrium between the M-Unit aldehyde and the corresponding gem-άιοΧ is still displaced towards the hydrated form (4). Despite the undetectable aldehyde moieties, thiol molecules and the M-Unit CS aldehyde are engaged in a pH dependent equilibrium with the corresponding hemithioacetal intermediate (5). The stabilization of the latter into its thioacetal form (6) can occur at least either by increasing the amount of thiol-bearing reactants in the medium (in situ stabilization), or by freeze-drying the reaction medium when low amounts of thiol are engaged.

[129] While the present disclosure describes certain non-limiting conjugate embodiments, the person of skill will readily appreciate that the thiol-based reaction schemes set forth here can be applied to other polymers bearing aldehydes or ketones.

[130] The person of skill will also readily understand that if a PEG molecule is used according to the herein teachings, any suitable PEG entity may be used so long as it produces a conjugate having desired properties. For example, the person of skill may use a PEG molecule of formula H- [0-CH₂-CH₂]_n-OH, where n = 1 to 2500 or a ramified four-arms or 8-arms PEG molecule.

5. Thiol-hook models for end-group derivatization of chitosans

5.1. M-Unit / Thiol-hook models conjugation: Mechanistic studies by LCMS

[131] Protocol: 2,5-anhydro-D-mannose (M-Unit) conjugation with thiol-hook models. Thiol-hook models (Ethanedithiol, EDT and Propanedithiol, PDT) were used to assess the intramolecular thioacetylation process, where both thiol attacks occur simultaneously on the M-Unit aldehyde forming instantaneously the stable thioacetal conjugate. Briefly, the synthesized 2,5-anhydro-D- mannose M-Unit (0.1 mmol, 16.2 mg) was dissolved in 5mL degassed 30 or 40 %v/v 2-propanol in ddH₂0 for EDT or PDT coupling, respectively. The pH of the solution was adjusted to 1 with 3M HC1 solution prior to the addition of the thiol-bearing molecule (0.5 mmol, 41.9 μΐ_^ for EDT and 50.2 μΐ, for PDT). The reaction mixture was stirred for 72h at 50°C, under Ar atmosphere and covered with aluminum foil. The reaction mixture turned clear pink-orange after 72h and was split into 2 parts (Methods I and II): the first was dedicated to the direct LC-MS analysis of the reaction medium in order to determine the thioacetal proportion in resulting conjugates that formed in situ; whereas the second one was immediately flash-frozen and then freeze-dried prior to LC-MS analyses to assess the effect of FD on the thioacetal proportion in resulting conjugates and to ascertain that no by-products appear post FD.

[132] Results: Only one product was detected by LCMS, namely the corresponding stable thioacetal formed by intramolecular cyclization, independently of both the thiol-hook model engaged and the Method used post-reaction. No linear thioacetals were observed using the thiol hook models and no hemithioacetal intermediates were detected for all experiments performed (N>3 for each condition tested). MS-ESI: EDT conjugates [M+Na]⁺ = 261.0215 (261.0226 expected); PDT conjugates [M+Na]⁺ = 275.0374 (275.0382 expected).

5.2. M-Unit chitosan salts / Thiol-hook models conjugation

[133] Reaction 1 (30% 2-propanol. EDT 5 equivalents): CS 92-2 HCl salt (Mn = 2003 g/mol, 356.7 mg, 0.178 mmol aldehyde) was dissolved in a degassed mixture of 8.94 mL 30% v/v 2- propanol in ddH₂0 (c aldehyde ⁼ 20 mM). The pH of the homogeneous reaction medium was adjusted to 1 with degassed 3M HCl solution. This solution was divided into 3 equal volumes of v = 2980 μΐ, and each of those was treated with v = 24.9 ethanedithiol (EDT; 5 equivalents per aldehyde, n = 0.297 mmol). Reaction media were stirred for 72h at T=50°C, under inert atmosphere.

Π341 Reaction 2 (40% 2-propanoL PDT 5 equivalents): CS 92-2 HCl salt (Mn = 2003 g/mol, 354.4 mg, 0.177 mmol aldehyde) was dissolved in a degassed mixture of 8.93 mL 40% v/v 2- propanol in ddH₂0 (c aldehyde ⁼ 20 mM). The pH of the homogeneous reaction medium was adjusted to 1 with degassed 3M HCl solution. This solution was divided into 3 equal volumes of v = 2977 μΐ, and each of them was treated with v = 29.8 μΐ. propanedithiol (PDT; 5 equivalents per aldehyde, n = 0.295 mmol). Reaction media were stirred for 4h, 24h and 72h at T=50°C, under inert atmosphere.

[1351 Reaction (30% 2-propanul, EDT 20 equivalents): CS 92-2 HCl salt (Mn = 2003 g/mol, 70.0 mg, 0.035 mmol aldehyde) was dissolved in a degassed mixture of 1.75 mL 30% v/v 2- propanol in ddH₂0 (c aldehyde = 20 mM). The pH of the homogeneous reaction medium was adjusted to 1 with degassed 3M HCl solution. This homogeneous solution was treated with v = 58.7 μΐ, ethanedithiol (EDT; 20 equivalents per aldehyde, n = 0.700 mmol) and was allowed to stir for 72h at T=50°C, under inert atmosphere.

[1361 Reaction 4 (40% 2-propanol. PDT 20 equivalents): CS 92-2 HCl salt (Mn = 2003 g/mol, 70.0 mg, 0.035 mmol aldehyde) was dissolved in a degassed mixture of 1.75 mL 40% v/v 2- propanol in ddH₂0 (c aldehyde = 20 mM). The pH of the homogeneous reaction medium was adjusted to 1 with degassed 3M HC1 solution. This homogeneous solution was treated with v = 70.3 propanedithiol (PDT; 20 equivalents per aldehyde, n = 0.700 mmol) and was allowed to stir for 72h at T=50°C, under inert atmosphere.

[137] Workups and purification: All reaction media were divided into 2 equal volumes dedicated to Workup I (no freeze-drying) and Workup II (direct freeze-drying) and were treated as follows for comparison purposes. Workup I samples were directly engaged in the purification process, whereas Workup II samples were flash-frozen and freeze-dried prior to removal of unreacted thiol- bearing molecule. All reaction media were treated with IN sodium hydroxide solution (pH of the solutions was increased up to 9) in order to remove some potential hemithioacetal intermediates (even if they were not observed by LCMS in the conditions implemented therein), ensuring that only the stable thioacetals conjugates would be detected by 1H NMR. After acidification of the solutions for chitosan solubilization, unreacted thiol molecules (EDT and PDT) were discarded by 5 successive reprecipitations in pure 2-propanol. The remaining precipitates were dissolved in 5 mL ddH₂0 and these solutions were flash-frozen and freeze-dried. Conjugation efficiencies were determined by 1H NMR, using the herein described Equation 5 (normalization of the NMR spectra based on the M-Unit corresponding peaks and functionalization degree calculated via the integration of the gem-diol peak).

[138] Results corresponding to the conjugation efficiencies of M-Unit chitosan 92-2 (deacetylation degree of 92%; 2-3 kDa) / Thiol-hook models (Ethanedithiol and propanedithiol, EDT and PDT, respectively) determined by 1H NMR (N>3 +SD) are depicted in Figure 19. The labels below the bars correspond to reaction conditions, for example "EDT 72-5" refers to the ethanedithiol conjugation on the M-Unit chitosan for 72h with 5 equivalents of thiols per aldehyde. Workup I stands for non-freeze-dried samples, whereas Workup II refers to freeze-dried samples prior to purification. The freeze-drying step post-reaction may favor the conjugation as observed for PDT 4-5, PDT 24-5 and PDT 72-5. Nevertheless, with 72h reaction duration, the freeze-drying step seems to increase the conjugation to a lower extent than with shorter reaction durations. The use of 20 equivalents thiol per chitosan end-group aldehyde considerably increases the conjugation efficiency without any freeze-drying step, reaching about 70% conjugation efficiency. 6. Triskelion linker for end-group derivatization of chitosans

6.1. Triskelion linker synthesis

[139] The proposed triskelion linker synthesis is depicted in Figure 20 and corresponds to a two- steps process that takes place in organic conditions: 1) the triol (1,2,6-hexanetriol) was treated with mesylate chloride (methanesulfonyl chloride) to transform all triol hydroxyls into leaving groups. 2) The leaving groups were displaced by potassium thioacetate (CH₃COSK) to give the triskelion under its protected form.

[140] Step 1 : The triol starting material (m = 1.741 g) was dissolved in 36 mL anhydrous dichloromethane (DCM) + 24 mL anhydrous Tetrahydrofurane (THF) (COH = 648 mM). Mesylate chloride (3 equivalents per hydroxyl group; v = 9.24 mL) was added stepwise to the stirring reaction medium. While stirring, the clear and homogeneous reaction medium was cooled down to 0-5°C within an ice bath and v = 16.6 mL triethylamine (3 equivalents per hydroxyl group) was added. The reaction mixture was gently warmed up to room temperature and stirred for 24h under inert atmosphere. Heterogeneous and dark-orange reaction mixture was dissolved in v = 3 x 100 mL DCM and the organic bottom layer was successively extracted with 2 x 100 mL cold ddH₂0, 2 x 100 mL 10% v/v HC1 solution, 2 x 100 mL saturated Na₂C0₃ solution and v = 100 mL saturated NaCl solution. Remaining organic layer was dried over MgS0₄ and concentrated under reduced pressure to give an orange oil that was analyzed by TLC, 1H NMR and MS-ESI. This compound was engaged without further purification in the second step of the synthesis.

TLC: Rf = 0.3, Ethylacetate/Cyclohexane [7:3]

1H NMR (500 MHz, CDC1₃, 22°C, ns=32, dl=2s, acquisition time=2s): δ 1.58-1.81 (m, 6H, -(CH₂)₃-), 3.01-3.10 (3 s, 3x3H, -(S-CH₃)₃), 4.23-4.41 (m, 2x2H -CH₂-0-), 4.87-4.89 (m, IE, -CH-0-)

MS-ESI: [M+NH₄]⁺ = 386.06226 (386.06077 expected); [M+Na]⁺ = 391.01738 (391.01617 expected); [M+K]⁺ = 406.99055 (406.99010 expected).

[141] Step 2: The mesylate intermediate isolated in step 1) above was dissolved in v = 94 mL anhydrous DMF (CQMS ⁼ 467 mM). The clear orange solution was degassed with 3 cycles vacuum/N₂ and was cooled down to 0-5°C within an ice-bath for 15 min. Potassium thioacetate (m = 16.9 g, 5 equivalents per -OMs group) was added stepwise to the stirring and cold reaction medium under inert atmosphere. The first 8h of reaction were done at 0-5°C and the mixture stirred at room temperature for an additional 16h. The red-brown reaction medium was concentrated to dryness and was solubilized in v = 3 x 200 mL cyclohexane. The organic layer was extracted with 5 x 200 mL ddH₂0, dried over MgS0₄ and concentrated under reduced pressure. The crude oil was purified by flash-chromatography using Cyclohexane/Ethylacetate [95:5] as eluent, giving m = 2.25 g of acetylated triskelion linker with 75% yield over 2 steps.

TLC: Rf=0.3, Cyclohexane/Ethylacetate [9:1].

1H NMR (500 MHz, CDC1₃, 22°C, ns=32, dl=2s, acquisition time=2s): δ 1.36-1.73 (m, 6H, -(CH₂)₃-), 2.32-2.34 (3 s, 3x3H, -(CO-CH₃)₃), 2.82-2.86 (t, 2H, -CH2-CH2-S-), 3.09-3.25 (m, 2H, -CH-CH2-S), 3.59-3.65 (m, 1H, -CH-S-).

¹³C NMR (500 MHz, CDC1₃, 22°C, ns=320, dl=1.5s, acquisition time=2s): δ 25.80, 28.62, 28.96 (-CH₂-); 30.36, 30.46, 30.55 (-CH₃-); 32.51 (-CH-CH₂-S-); 33.84 (-CH₂-CH-); 43.91 (-CH-); 194.67, 194.79, 195.67 (CO).

MS-ESI: [M+H]⁺ = 309.06526 (309.06473 expected); [M+NH₄]⁺ = 326.09217 (326.09128 expected); [M+Na]⁺ = 331.04745 (331.04668 expected); [M+K]⁺ = 347.02041 (347.02062 expected).

6.2. M-Unit / Triskelion linker conjugation: Mechanistic studies by LCMS

[142] Protocol : The conjugation of the acetyl -protected triskelion linker to the 2,5-anhydro-D- mannose (M-Unit) is depicted in Figure 21. Briefly, the acetyl -protected triskelion was first deprotected using sodium methoxide (MeONa) treatment. The reaction medium was then quenched with HC1 in order to neutralize the excess of MeONa and the solvent was carefully removed under reduced pressure. In a second step, M-Unit that was synthesized by HONO treatment of glucosamine was dissolved in an appropriate mixture of solvents (30% THF v/v in ddH₂0 or 90% methanol v/v in ddH₂0 to reach 20 or 10 mM aldehyde concentration at pH 1, respectively). This solution was added to the deprotected triskelion and the reaction medium was allowed to stir for 72h under inert atmosphere at T=50°C. Reaction medium was treated according to the following methods prior to LCMS analysis: Method I where the reaction medium was concentrated by rotavap followed by extractions vs. Method II where extractions were performed without preliminary concentration of the medium by rotavap.

[143] Step 1 : Triskelion linker deprotection. The acetyl-protected triskelion linker (m = 171.2 mg, n = 5,55 10^"4 mol, 20 equivalents per potential M-Unit aldehyde) was dissolved in v = 10.7 mL of degassed 0.5M sodium methoxide solution (3.2 equivalents per acetyl group to cleave). The reaction medium stirred for 10 min at room temperature and under inert atmosphere. Reaction medium was quenched with v = 660 μΐ, of degassed HC1 37% and the heterogeneous solution was then extracted with v = 4 mL of degassed ddH₂0. The isolated organic layer was carefully concentrated under reduced pressure to give a clear yellowish oil confirmed to be the pure product (95% yield) by 1H NMR and was stored under inert atmosphere until conjugation reaction.

1H NMR (500 MHz, CDC1₃, 22°C, ns=32, dl=2s, acquisition time=2s): δ 1.35 (m, 1H, -SH-), 1.50-1.69 (br., 6H, -(CH₂)₃-), 1.79-1.86 (m, 2H, -SH-), 2.53-2.55 (m, 2H, S), 2.68-2.83 (m, 2H, -CH-CHz-SH), 2.88-2.90 (m, 1H, -CH-S-).

[144] Step 2 Reaction 1 : Triskelion linker conjugation (30% THF). The following description corresponds to the conjugation performed in a degassed mixture of 30% v/v THF in ddH₂O.The 2,5-anhydro-D-mannose (m = 5 mg, n = 2.78 10^"5 mol) was dissolved in 960 iL of degassed ddH₂0 and then v = 416 μL of degassed THF was added. The pH of the reaction medium was adjusted to 1 with v = 1 1 μί., of degassed HC1 37%, reaching a final aldehyde concentration of 20 mM. This solution was added to the deprotected triskelion prepared in step 1 above and the reaction medium was stirred for 72h at T=50°C under inert atmosphere.

[145] Step 2 Reaction 2: Triskelion linker conjugation (90% MeOH). The following description corresponds to the conjugation performed in a degassed mixture of 90% v/v Methanol in ddH₂0. The 2,5-anhydro-D-mannose (m = 5 mg, n = 2.78 10^"5 mol) was dissolved in 255 of degassed ddH₂0 and then v = 2498 μΐ. of degassed methanol was added. The pH of the reaction medium was adjusted to 1 with \ = 23 μL· of degassed HC1 37%, reaching a final aldehyde concentration of 10 n M. This solution was added to the deprotected triskelion prepared in step 1 and the reaction medium was stirred for 72h at T=50°C under inert atmosphere.

[146] Reaction media were treated according to the following methods prior to LCMS analysis: Method I where the reaction medium was concentrated by rotavap followed by extractions vs. Method II where extractions were performed without preliminary concentration of the medium by rotavap.

[147] Results: The relative proportion of the M-Unit aldehyde / Triskelion linker conjugation products determined by LCMS is depicted in Figure 22. Products A & B correspond to the desired products obtained by intramolecular cyclization (A: M-Unit-triskelion conjugate; B: M-Unit- triskelion conjugates linked by disulfide bond through the third remaining thiol moiety). Here the thioacetylation process between the M-Unit aldehyde and the triskelion thiol hook forms a 5- membered ring. Products C & D correspond to the side-products obtained by intermolecular thioacetylation (D: M-Unit-(triskelion)₂ conjugate; E: Oxidized M-Unit-(triskelion)2 conjugates. Here the thioacetylation process between the M-Unit aldehyde and the triskelion' s third thiol moiety forms a linear thioacetal. Independently of both the reaction conditions performed (N>3 ± SD), and the method post-reaction applied, the major compounds observed correspond to the intramolecular cyclization product (5-membered ring thioacetal), meaning that the intramolecular thioacetylation is favored vs. its intermolecular counterpart. These results are in agreement with those obtained with the thiol-hook models (Ethanedithiol and propanedithiol, EDT and PDT respectively). Moreover, the fact that the product B was observed among the major products underlines both the presence and the reactivity of the remaining thiol moiety which is not engaged into the thioacetal cyclization process.

6.3. M-Unit chitosan salts / Triskelion linker conjugation

[148] Step 1 : Triskelion linker deprotection was performed as described above. The amount of triskelion used in the following examples corresponds to 20 equivalents per chitosan's M-Unit aldehyde. [149] Step 2 Reaction 1 ( 2 kDa CS. 30% TH K pH 1 , 72h, Workups Ι&ΙΓ): CS 92-2 HC1 salt (Mn = 2566 g/mol, 55 mg, 2.14 10^"5 mol aldehyde) was dissolved in v = 741 μΐ, degassed ddH₂0. The pH of the homogeneous reaction medium was adjusted to 1 with degassed 3M HC1 solution. Degassed THF (v = 322 μί) was added to the chitosan acidic solution giving a 30% v/v THF in ddH₂0 mixture and reaching a chitosan end-group concentration of 20 mM. The protected triskelion linker (20 equivalents per aldehyde, n = 4.29 10^"4 mol, m = 132.2 mg) was deprotected and purified according to the protocol described above. Chitosan solution was added to the triskelion oil and the reaction medium stirred for 72h at T=50°C, under inert atmosphere. The reaction medium was divided into 2 equal volumes dedicated to Workup I and Workup II (no drying and direct drying, respectively). Workup I samples were immediately engaged in the purification process, whereas Workup II samples were concentrated to dryness under reduced pressure prior unreacted thiol molecule models removal.

[150] Step 2 Reaction 2 (2 kDa CS. 85% MeOH. pH 1 , 24, 48 and 72h. Workup I: CS 92-2 HC1 salt (Mn = 3244 g/mol, 76.7 mg, 2.36 10^"5 mol aldehyde) was dissolved in v = 315 μΐ, degassed ddH₂0. The pH of the homogeneous reaction medium was adjusted to 1 with v = 39.4 μΐ. degassed HC1 37%. Degassed methanol (v = 2010 μϋ,) was added to the chitosan acidic solution giving an 85% v/v methanol in ddH₂0 mixture and reaching a chitosan end-group concentration of 10 mM. The protected triskelion linker (20 equivalents per aldehyde, n = 4.73 10^"4 mol, m = 145.9 mg) was deprotected and purified according to the protocol described above. Chitosan solution was added to the triskelion oil and the reaction medium stirred for 24h, 48h and 72h at T=50°C, under inert atmosphere. The reaction media corresponding to the time-points (24h, 48h and 72h) were treated with Workup II (concentration to dryness under reduced pressure prior unreacted thiol molecule models removal).

[151] Step 2 Reaction 3 (2 kDa CS, 90% MeOH, pH 1 . 72h, Workup I): CS 92-2 HC1 salt (Mn = 2432 g/mol, 32.6 mg, 1.34 10^"5 mol aldehyde) was suspended in v = 123 μΐ, degassed ddH₂0. The pH of the homogeneous reaction medium was adjusted to 1 with v = 1 1.1 μΐ, degassed HC1 37%. Degassed methanol (v = 1206 μΐ ) was added to the chitosan acidic solution giving an 90% v/v methanol in ddH₂0 mixture and reaching a chitosan end-group concentration of 10 mM. The protected triskelion linker (20 equivalents per aldehyde, n = 2.68 10^"4 mol, m = 82.7 mg) was deprotected and purified according to the protocol described above. Chitosan solution was added to the triskelion oil and the reaction medium stirred for 72h at T=50°C, under inert atmosphere. The reaction medium was treated with Workup I (no drying prior to purification).

[152] Slep 2 Reaction 4 (2 kDa CS. 85% MeOH. pH 4. 24, 48 and 72h. Workup II): CS 92-2 HC1 salt (Mn = 3244 g/mol, 66.8 mg, 2.06 10^'5 mol aldehyde) was dissolved in v = 299 uL degassed ddH₂0. The pH of the homogeneous reaction medium was adjusted to 4 with v = 10.3 μΐ, degassed HC1 37%. Degassed methanol (v = 1750 μΐ,) was added to the chitosan acidic solution giving an 85% v/v methanol in ddH₂0 mixture and reaching a chitosan end-group concentration of 10 raM. The protected triskelion linker (20 equivalents per potential aldehyde, n = 4.12 10^"4 mol, m = 127.0 mg) was deprotected and purified according to the protocol described above. Chitosan solution was added to the triskelion oil and the reaction medium stirred for 24h, 48h and 72h at T=50°C, under inert atmosphere. The reaction media corresponding to the time-points (24h, 48h and 72h) were treated with Workup II (concentration to dryness under reduced pressure prior unreacted thiol molecule models removal).

[153] Step 2 Reaction 5 f 1 0 kDa CS, 90% MeOH, pH 1 , 72h, Workups Ι&ΙΓ): CS 92-10 HC1 salt (Mn = 9020 g/mol, 26.2 mg, 2.91 10^"6 mol aldehyde) was suspended in v = 134 uL degassed ddH₂0. The pH of the homogeneous reaction medium was adjusted to 1 with v = 12 μΐ, degassed HC1 37%. Degassed methanol (v = 1310 uL) was added to the chitosan acidic solution giving an 90% v/v methanol in ddH₂0 mixture and reaching a chitosan end-group concentration of 2 mM. The protected triskelion linker (20 equivalents per potential aldehyde, n = 5.82 10^"5 mol, m = 18.0 mg) was deprotected and purified according to the protocol described above. Chitosan solution was added to the triskelion oil and the reaction medium stirred for 72h at T=50°C, under inert atmosphere. The reaction medium was divided into 2 equal volumes dedicated to Workup I and Workup II (no drying and direct drying, respectively).

[154] Step 3 Purificalion and analyses: All reaction media were treated with IN sodium hydroxide solution (pH of the solutions was increased up to 10) in order to remove some potential hemithioacetal intermediates (even if they were not observed by LCMS in the conditions implemented therein), ensuring that only the stables thioacetals conjugates will be detected by Ή NMR. After acidification of the solutions for chitosan solubilization, unreacted triskelion was discarded by 5 successive reprecipitations in fresh THF. The remaining precipitates were dissolved in 5 mL ddH₂0 and these solutions were flash-frozen and freeze-dried. Conjugation efficiencies were determined byΉ NMR, using both the herein described Equation 1 and Equation 5.

[155] Results: The M-Unit chitosan (deacetylation degree of 92%; 2-3 kDa and lOkDa*) / Triskelion linker conjugation efficiencies determined by 1H NMR (N>3 ±SD) are depicted in Figure 23. In this example, "THF 1-72" refers to the triskelion conjugation on the M-Unit chitosan aldehyde at pH 1 for 72h. Workup I stands for non-freeze-dried samples, whereas Workup II refers to freeze-dried samples prior purification. Note that MeOH 1-24, MeOH 1-48, MeOH 4-24, MeOH 4-48 and MeOH 4-72 were only treated with Workup II. As described in previous examples (with thiol-hook models, EDT &PDT), the freeze-drying step post-reaction may increase the conjugation degree as observed for THF 1-72, MeOH 1-72 and MeOH 1-72* samples. All reactions performed at pH 4 are less efficient than their pH 1 counterpart; this observation is in good agreement with previous results since the thioacetylation process is pH dependent.

[156] Other examples of implementations will become apparent to the reader in view of the teachings of the present description and as such, will not be further described here.

[157] Note that titles or subtitles may be used throughout the present disclosure for convenience of a reader, but in no way these should limit the scope of the invention. Moreover, certain theories may be proposed and disclosed herein; however, in no way they, whether they are right or wrong, should limit the scope of the invention so long as the invention is practiced according to the present disclosure without regard for any particular theory or scheme of action.

[158] All references cited throughout the specification are hereby incorporated by reference in their entirety for all purposes.

[159] It will be understood by those of skill in the art that throughout the present specification, the term "a" used before a term encompasses embodiments containing one or more to what the term refers. It will also be understood by those of skill in the art that throughout the present specification, the term "comprising", which is synonymous with "including," "containing," or "characterized by," is inclusive or open-ended and does not exclude additional, un-recited elements or method steps.

[160] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. In the case of conflict, the present document, including definitions will control.

[161] As used in the present disclosure, the terms "around", "about" or "approximately" shall generally mean within the error margin generally accepted in the art. Hence, numerical quantities given herein generally include such error margin such that the terms "around", "about" or "approximately" can be inferred if not expressly stated.

[162] Although the present invention has been described in considerable detail with reference to certain embodiments thereof, variations and refinements are possible and will become apparent to persons skilled in the art in light of the present description.

Claims

CLAIMS:

1. A chitosan conjugate of formula (I):

wherein:

CS represents a chitosan residue;

X comprises a Zl group, Zl being a linear thioacetal or a thioacetal forming an n-membered ring, n being from 4 to 9, and Zl optionally being ramified and/or substituted; and

POLY represents a ligand, or a pharmaceutically acceptable salt thereof.

2. A conjugate according to claim 1, wherein said Zl is a linear thioacetal, optionally ramified and/or substituted, wherein each S atom of said linear thioacetal is linked to a respective ligand.

3. A conjugate according to claim 1, wherein said Zl is a thioacetal forming an n-membered ring, n being from 4 to 9, the Zl optionally being ramified and/or substituted.

4. A conjugate according to claim 2 or 3, wherein said X comprises a Zl-Ll group, wherein said LI comprises a linear alky chain having a length within C]-C₁₀₀ and/or an alkyl n-membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂-; or a cleavable or an uncleavable covalent linker.

5. A conjugate according to claim 2 or 3, wherein said X comprises a Z1-L1-Z2 group, wherein said LI comprises a linear alky chain having a length within Ci-Cjoo and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-C¾-; or a cleavable or an uncleavable covalent linker; and said Z2 comprises a peptide bond, thioester, thioether, or Schiff base.

6. A conjugate according to claim 2 or 3, wherein said X comprises a Z1-L1-Z2-L2 group, wherein said LI comprises a linear alky chain having a length within Q-Cioo and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂-; or a cleavable or an uncleavable covalent linkage; said Z2 comprises a peptide bond, thioester, thioether, or Schiff base; and said L2 comprises a linear alky chain having a length within Q-Cioo and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂; or a cleavable or an uncleavable covalent linkage.

7. A conjugate according to claim 2 or 3, wherein said X comprises a Z1-L1-Z2-L2-Z3 group, wherein said LI comprises a linear alky chain having a length within Ci-C₁₀o and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂-; said Z2 comprises a peptide bond, thioester, thioether, or Schiff base; said L2 comprises a linear alky chain having a length within Ci-Cioo and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂-; and said Z3 comprises a cleavable or an uncleavable covalent linkage.

8. A conjugate according to claim 2 or 3, wherein said X comprises a Z1-L1-Z3 group, wherein said LI comprises a linear alky chain having a length within Cj-Cioo and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂-; and said Z3 comprises a cleavable or an uncleavable covalent linkage.

9. A conjugate according to claim 2 or 3, wherein said X comprises a Z1 -L1 -Z2-Z3 group, wherein said LI comprises a linear alky chain having a length within Ci-Cioo and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂-; said Z2 comprises a peptide bond, thioester, thioether, or Schiff base; and said Z3 comprises a cleavable or an uncleavable covalent linkage.

10. A conjugate according to claim 2 or 3, wherein said X comprises a Z1 -L1 -L2-Z3 group, wherein said LI comprises a linear alky chain having a length within Q-CHJ_O and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂; said L2 comprises a linear alky chain having a length within C Qoo and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂; and said Z3 comprises a cleavable or an uncleavable covalent linkage.

1 1. A conjugate according to claim 2 or 3, wherein said X comprises a Z1 -Z2-L2-Z3 group, wherein said Z2 comprises a peptide bond; said L2 comprises a linear alky chain having a length within Ci-C₁₀o and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂; and said Z3 comprises a cleavable or an uncleavable covalent linkage.

12. A conjugate according to claim 2 or 3, wherein said X comprises a Z1-L2-Z3 group, wherein said L2 comprises a linear alky chain having a length within Ci-Cioo and/or an alkyl n- membered ring, n being from 4 to 9, said linear alkyl or alkyl ring being optionally ramified and/or substituted; -S-CH₂-CH₂; and said Z3 comprises a cleavable or an uncleavable covalent linkage.

13. A conjugate according to claim 2 or 3, wherein said X comprises a Z1-Z3 group, wherein said Z3 comprises a cleavable or an uncleavable covalent linkage.

14. A conjugate according to any one of claims 1 to 13, wherein said POLY represents a nucleic acid molecule, a polypeptide, a non-peptidic polymer, a second chitosan residue, a planar or particulate surface, or any combinations thereof.

15. A conjugate according to claim 14, wherein said nucleic acid molecule is a linear DNA, mRNA, shRNA, or siR A.

16. A conjugate according to claim 14, wherein said non-peptidic polymer is a polyethylene glycol (PEG).

17. A process for manufacturing depolymerized chitosan residues, comprising:

- depolymerization of chitosan with nitrous acid (HONO) to obtain a depolymerized chitosan residue salt, and

- drying said depolymerized chitosan residue salt under acidic conditions, said conditions comprising a pH < 4.

18. A process according to claim 17, wherein said acidic conditions comprise a pH < 3.

19. A process for manufacturing a chitosan-ligand conjugate, comprising:

- providing a nitrous acid (HONO) depolymerized chitosan residue salt,

- drying said depolymerized chitosan residue salt under acidic conditions, said conditions comprising a pH < 4,

- rehydration of said salt in an aqueous solvent to obtain an aqueous solution, and

- incorporating a thiol molecule to said solution thus producing a reaction medium for obtaining said conjugate, said reaction medium having a pH of about 1.

20. A process for manufacturing a chitosan-ligand conjugate, comprising:

- providing a nitrous acid (HONO) depolymerized chitosan residue salt previously dried under acidic conditions, said conditions comprising a pH < 4,

- rehydration of said salt in an aqueous solvent to obtain an aqueous solution, and

- incorporating a thiol molecule to said solution thus producing a reaction medium for obtaining said conjugate, said reaction medium having a pH of about 1.

21. A process according to claim 19 or 20, wherein said acidic conditions comprise a pH < 3.

22. A process according to any one of claims 19 to 21, wherein said thiol molecule comprises said ligand.

23. A process according to any one of claims 19 to 22, further comprising freeze-drying said reaction medium.

24. A process according to any one of claims 19 to 22, wherein said thiol molecule is incorporated in an excess amount relatively to said chitosan residue.

25. A conjugate according to any one of claims 1 to 16, wherein said chitosan residues have an average molecular weight of from about 0.2 kDa to about 200 kDa.

26. A conjugate according to any one of claims 1 to 16, wherein said chitosan residues have a degree of deacetylation of from about 50% to about 100%.

27. A particle comprising the chitosan conjugate of any one of claims 1 to 12, wherein said ligand comprise a polyethylene glycol (PEG) molecule, and wherein said particles have a substantially spherical form and have a reduced zeta potential than a comparative particle being identical to said particle except for comprising chitosan residues instead of chitosan-PEG conjugates.

28. A particle according to claim 27, wherein said chitosan residues have an average molecular weight of from about 0.2 kDa to about 200 kDa.

29. A particle according to claim 27 or 28, wherein said chitosan residues have a degree of deacetylation of from about 50% to about 100%.

30. A particle according to any one of claims 27 to 29, further comprising plasmid DNA and having a Z-average diameter of < 200 nm.

31. A particle according to claim 30, wherein the Z-average diameter is of < 90 nm.

32. A particle according to any one of claims 27 to 29, further comprising a linear DNA, a siRNA, a shRNA, or a mRNA.

33. A process according to any one of claims 17 to 19, wherein said drying step includes freeze- drying.

34. A process according to any one of claims 17 to 19, wherein said drying step includes oven vacuum drying.

35. A process according to claim 20, wherein said nitrous acid (HONO) depolymerized chitosan residue salt is previously dried by freeze-drying.

36. A process according to claim 20, wherein said nitrous acid (HONO) depolymerized chitosan residue salt is previously dried by oven vacuum drying.