Poly(2-ethyl-2-oxazoline) grafted gellan gum for potential application in transmucosal drug delivery

Gellan gum (GG) has been used to prepare polymeric carriers with prolonged retention on the eye surface for topical ocular drug delivery. In this work, GG was chemically modified with short poly(2-ethyl-2-oxazoline) (PEtOx) chains that were expected to have minimal adhesion to mucosal tissues (mucoadhesion). The choice of synthetic procedure, solvents, and reagents has been dictated by biocompatibility of the materials and possible application in drug delivery. The grafts were synthesized via cationic ring-opening polymerization and their living chains were attached onto deprotonated gellan backbone. The derivatives with three degrees of grafting were prepared by varying the in-feed mass ratio of PEtOx grafts over GG. NMR and FT-IR spectroscopies, thermogravimetric analysis, and SEC evidenced that the grafting had actually taken place. However, a greater diffusion coefficient determined for the copolymer, using diffusion-ordered spectroscopy (NMR), in relation to the diffusion of the unmodified GG, suggested either partial degradation of the backbone or a more compact structure of the copolymer. GG and its graft copolymers (GG-g-PEtOx) were found to be highly biocompatible with cells cultured under their induction at concentration of 1, 0.1 and 0.01 mg/mL demonstrated a physiological morphology, as well as an increase in viability and proliferation.


| INTRODUCTION
Ocular drug delivery is a challenging field of pharmaceutics owing to specific conditions on the eye surface. Many drug carriers often have disadvantages such as poor retention on ocular surfaces; additionally their ocular administration may cause irritation and blurred vision.
Topically administered drugs are easily diluted by tears or removed by blinking, which results in subtherapeutic drug levels. 1 Ongoing demands exist for innovative drug carriers capable of prolonged retention on the ocular surface and enhanced effectiveness of therapy. Gellan gum (GG) is an attractive material that has already found applications in the design of dosage forms for ocular drug delivery due to its mucoadhesive properties and ability to form gels in situ. 2 GG is an anionic tetra-saccharide with a repeating sequence comprising of D-glucuronate, L-rhamnose and two D-glucose units. 3 GG is typically produced in a fermentation process by Pseudomonas bacteria isolated from Elodea plant. 4 The native form of GG contains on average one glyceryl and 0.5 acetyl groups per glucose unit that can be removed by hot alkaline treatment giving the deacetylated form, so called low-acyl GG. 5 Both native and low-acyl GG form gels. Gels formed by the native GG are typically elastic and weak, whereas more compactly packed low acyl GG produces hard and brittle gels. Gelation of GG occurs upon cooling and the gel turns back to fluid if heated again. 6 Gelation is also promoted by minute amounts of mono and divalent metal cations, which is not always the case with all gelling polysaccharides. [7][8] The advantages of GG over the other gelling polysaccharides are the thermal and enzymatic stability and high strength of the gel.
GG has been approved as a gelling agent in foods (E418, FDA code 21CFR172.665). 9 Its low toxicity, biodegradability, commercial availability, and low cost render GG suitable for applications in drug delivery and tissue engineering. [10][11][12] Tear fluid contains over four times excess of cations required to promote gelling of a 1 wt% GG solution. 13 Therefore, effective ophthalmic drug delivery can be thought in two phases: liquid formulation facilitates the administration of the drug onto the eye surface, whereas subsequent gelation promoted by tear fluid resists ocular drainage and prolongs the drug retention time. Thus, significant increase in the contact time due to the fast ion-activated in situ gelling has been demonstrated for the low acyl GG. 14 Administration of solutions on human and rabbit eyes showed minimal irritation, though some blurred vision was observed. 15 Above-mentioned applications employ mucoadhesive properties of GG. It has recently been demonstrated that modification of silica nanoparticles surface with poly(2-methyl-2-oxazoline) and poly (2-ethyl-2-oxazoline) (PEtOx) decreases mucoadhesion and facilitates wthe diffusion of the particles in gastric mucin, [16][17] whereas more hydrophobic poly(2-n-propyl-2-oxazoline) shows no significant increase in the particles mobility compared to the pristine particles. 17 As a development of this research, we propose the modification of GG with short PEtOx chains.
PEtOx is chemically stable, non-toxic, and has structural similarities with polypeptides owing to the amide group in its repeating units.
It has shown stealth properties comparable with poly(ethylene glycol), that is, low recognition by the immune system. [18][19][20][21][22] These features have risen interest in PEtOx for biomedical applications. [23][24][25][26][27] Homogeneous and controlled modification of GG with PEtOx is not an easy task owing to the gelling of GG and the limited number of common solvents appropriate for PEtOx synthesis and modification. Additionally, the chemicals and synthetic reactions used ought to be biologically safe, which rules out click-reactions for well-defined grafting.
Post-functionalization of intentionally terminated end-groups of PEtOx has to be also avoided.
Therefore, in this research, GG was modified by grafting with short PEtOx chains to prepare graft copolymers of various degrees of grafting. The length of the PEtOx blocks was selected to be approximately 5000 g/mol, which has been shown to have mucus penetrating properties. 16,17 PEtOx has been synthesized via cationic ring-opening polymerization (CROP). 28 The method suggested for poly(2-isopropyl-2-oxazoline-co-2-butyl-2-oxazoline) grafting to β-glucan 29 and to κ-carrageenan 30 was adjusted to GG-g-PEtOx. A key factor for successive PEtOx synthesis followed by its grafting to deprotonated GG is the living nature of polymerization with a minimum number of chain transfer and termination reactions. The living chain ends can be terminated with nucleophilic reagents. Therefore, special care has been taken of purity and dryness of chemicals. Three GG-g-PEtOx samples with different degrees of grafting were synthesized.

| Selection of solvent
Preparation of GG-g-PEtOx proceeds in two main phases: polymerization of PEtOx via CROP and then attachment of living PEtOx chains to deprotonated GG at 70 C. An appropriate solvent, which is common for GG and PEtOx and also suitable for CROP is a prerequisite for this synthesis. Commercial low acyl GG is poorly soluble in other solvent than water, whereas PEtOx synthesis cannot be conducted in water. Acetonitrile is a typical solvent for CROP, 32 however GG is not soluble in it. Water-free DMSO do not cause significant gelling of gellan 33 and expected to be good candidate for this synthesis. DMSO and sulfolane are aprotic dipolar solvents that can be used for CROP. 34 Sulfolane is less hydroscopic than DMSO. Sulfolane is slightly more toxic than DMSO, while chain transfer is more probable in DMSO. 32 Commercial GG is not completely soluble neither in sulfolane nor in DMSO in the "as received" form, though it swells in sulfolane. Therefore, GG was purified as described below. After purification, GG becomes soluble in DMSO, which was selected as the solvent for grafting. The polymerization of EtOx was performed in sulfolane and then the solution of living PEtOx chains was added to the DMSO solution of GG.

| Preparation of GG for modification
Commercial low acyl GG is in a powder form and per se cannot be homogeneously modified. A 1 w/v % GG solution is not uniform even in water; it forms a fraction of soluble chains and a dispersed fraction of swollen gellan particles. GG may contain Ca 2+ , Mg 2+ , Na + , and K + , from which divalent cations impair the solubility the most and act as intermolecular crosslinks. 35  The product was freeze-dried and collected as dry fluffy white fibers.
The yield of purification was 45%-50% of the original weight of GG. FT-IR and NMR spectroscopy did not reveal any significant differences between purified and pristine GG. Thus, there was no indication that GG was incompletely deacetylated and that a fraction was removed during purification. This conclusion is also supported by the elemental analysis, which reveals equal carbon and hydrogen contents in both samples. Therefore, the most likely reason for the low yield of purification is divalent cations. Small amount of these metal cations binds gellan chains into particles, which are removed by centrifugation.
Purified GG homogeneously dissolves in water and becomes soluble in DMSO. However, it is not soluble in sulfolane but remains soluble in DMSO/sulfolane mixtures used in this report. This determined the choice of DMSO as the solvent for grafting.
Freeze-dried GG is highly hydroscopic. Azeotropic distillation with tetrahydrofuran and toluene was performed prior to grafting to remove absorbed moisture. Two azeotropic distillations were tested for the GG drying. In the first, GG was dried during distillation of THF and was stored in a vacuum desiccator over silica gel. Thermogravimetric analysis and NMR spectroscopy estimate the water content of dried GG to be about 6%. The other method was shown to be more effective. GG (200 mg, 1.14 mmol that is an averaging of the four sugar units) was first dissolved in 20 mL of anhydrous DMSO.
The solution was dried by addition of anhydrous toluene (3 × 5 mL) and subsequent evaporation under reduced pressure. If several reactions were performed simultaneously, excess GG solution was prepared, which was at this stage split into separate flasks.

| Synthesis of PEtOx
Glassware used in the synthesis and distillations was kept overnight in an oven at 90 C prior to use. Calcium hydride was used for drying EtOx and sulfolane. EtOx placed in a round-bottom flask, and CaH 2 was added until the mixture turned milky. Drying tube filled with calcium chloride granules was connected to the flask and the mixture was stirred until the next day. For sulfolane, drying was identical, but the flask was placed in an oil bath set to 32 C, which was above the freezing point of sulfolane (28.4 C). 36 Next day EtOx was distilled under a nitrogen atmosphere and sulfolane under reduced pressure.
Polymer synthesis described in Loukotová et al. 29 was adapted. The monomer-to-initiator ratio was 50:1 in the PEtOx polymerizations targeting the molar mass of 5000 g/mol. The trial polymerizations were performed using methyl p-toluenesulfonate (MeOTs) as an initiator.
Although these polymerizations proceeded successfully, the solid form of the initiator at room temperature made handling challenging and for that reason MeOTs was replaced with methyl trifluoromethanesulfonate (MeOTf). MeOTf was extremely reactive as initiator and could not be added directly to the hot reaction mixture: loss of the initiator was observed due to the evaporation at elevated temperature and subsequent condensation on the reaction flask walls. The addition of MeOTf to monomer mixtures has been reported at 0 C 37 and at room temperature 18 followed by heating. Because sulfolane is solid under 28 C, the handling of the initiator at lower temperatures was not possible.

| Synthesis of GG-g-PEtOx
Procedure for the GG-g-PEtOx synthesis is shown in Figure S5. Gellan chains were grafted in two parallel processes starting with polymerization of EtOx. The GG backbone was activated with NaH, producing alkoxide ions. However, grafting can also take place at carboxylic groups of D-glucuronate. In the course of the grafting, a new covalent ether bond is formed between a PEtOx chain and GG.
Grafting was tested at three temperatures. Grafting at 70 C resulted in extensive degradation of gellan gum. Synthesis at and below 40 C promotes gelling that is not favorable. About 50 C was selected for the reported polymers.
Sodium hydride (340 mg, 8.4 mmol) was added to the 10 mg/mL solution of GG in DMSO azeotropically dried using toluene (200 mg, 1.14 mmol that is an averaging of the four sugar units was dissolved in 20 mL). The mixture was stirred for 3 h at 50 C. A needle was placed through the rubber septum to release formed hydrogen gas. Next, the pre-determined amount of living PEtOx chains in sulfolane (sample P5) was added with a needle to the gellan/DMSO mixture, and the mixture was stirred overnight. Polymerization of the remaining PEtOx chains was terminated by the addition of 1 M methanolic NaOH and this mixture was stirred at room temperature for 24 h to confirm complete termination. Finally, the PEtOx mixture was dialyzed (cut off 3.5 kDa) for a week, freeze-dried, and characterized.
Fifteen milliliters of water was added to the resulted mixture and the mixture was then poured into the 250 mL separatory funnel. The mineral oil was removed by washing the mixture with 50 mL of diethyl ether and separating the aqueous phase. This was done at least twice.
If gelling occurred during this washing, warm water was added to the separatory funnel and the mixture was shaken to dissolve the gel.
Finally, the aqueous mixture of GG-g-PEtOx with unreacted PEtOx was dialyzed (cut off 12-14 kDa) for a week to remove PEtOx and other low molar mass impurities. The product was isolated as white fibers after freeze-drying.

| Characterization
Elemental analysis was used to determine the composition of the commercial GG as received in comparison to purified GG. Elemental composition was determined by the HANAU Elementar Analysensysteme GmbH, Germany.vario MICRO cube using sulfanilamide as a standard.
Three samples (~2 mg of dry polymer per measurement) were studied and the average of those values is presented as a result.
NMR spectroscopy: Zg20 pulse sequence was applied to record 1 H NMR spectra with a Bruker Avance III 500 spectrometer. In diffusion- The optical transmittance of the graft copolymers aqueous solution (0.7 w/v %) was measured at 650 nm using a Lambda 25 Perkin-Elmer UV/vis spectrophotometer as the solution temperature was raised from 20 to 80 C.

| Characterization of GG-g-PEtOx
GG-g-PEtOx copolymers of three different grafting densities were prepared by adjusting the living PEtOx content in the reaction feed (Table 1). This is given as weight equivalent in feed, that is, the mass ratio of PEtOx grafts over GG used in the grafting reaction. The mass content of PEtOx is estimated from the assumption that all the initiator reacts and forms a propagating chain. This may deviate from the actual molar amount as M n of PEtOx was higher than the theoretical molecular weight. Also, the molar mass of the attached PEtOx grafts may slightly differ from the M n value measured for the corresponding PEtOx sample P5 obtained by polymerization termination. The grafting density is given as the number of repeating sequences (i.e., four sugar units constituting GG) per one PEtOx graft.
Hence, every 12th repeating sequences (i.e., 12 × 4 sugar units) of the G1 copolymer has in average one PEtOx grafted chain. In the case of G2, the grafting density is 3, and for G3, that is 2.
The peak intensity from the end-group is low and the peak from rhamnose becomes evident at the higher GG contents and overlaps with another peak at 0.96 (methyl protons of the side group of PEtOx), the accuracy of the grafting density calculation is not perfect. However, the difference in grafting densities between the copolymers is evident. Examination of the yields of the gellan copolymers indicates increasing graft content relative to the weight equivalent. Thermograms of pure GG, PEtOx, and the GG-g-PEtOx copolymers are presented in Figure 3. GG shows a two-step degradation process. Six percentage weight loss of absorbed water is observed in the temperature range of 25-100 C. Decomposition of the GG backbone takes place between 200 and 300 C (about 50% of the weight).
The fastest weight loss occurs at 247 C ( Figure S8). At 800 C, the total weight loss was 76%. Degradation of PEtOx happens between T A B L E 1 Experimental conditions for the GG-g-PEtOx copolymer synthesis 260 and 440 C. The fastest degradation appears at 420 C ( Figure S8).
The PEtOx decomposition proceeds to almost 100% weight loss and results in negligible char yield.
All thermograms of the GG-g-PEtOx copolymers show weight loss within 25 to 100 C, indicating the presence of water. The higher moisture content corresponds to the copolymer with the higher GG F I G U R E 1 1 H NMR spectra of pure gellan gum and the GG-g-PEtOx copolymers. Assignments of the peaks to corresponding chemical structures are given is Figures S2 and S3 F I G U R E 2 FT-IR spectra of the GG-g-PEtOx copolymers. FT-IR spectra of gellan gum and of the PEtOx homopolymer, sample P5 and the bands assignments are given in Figures S6 and S7 correspondingly F I G U R E 3 Thermograms of pristine gellan gum, PEtOx homopolymer and the graft copolymers content. Degradation of the GG backbone starts at~200 C and results in a 35% weight loss for G1, 20% for G2, and 10% for G3. Degradation of the PEtOx grafts happens within 300 to 400 C. The char yield of the samples follows the GG content with the highest yield found for G1. The derived weight losses of the graft copolymers ( Figure S9) show that the fastest losses happen at somewhat lower temperatures than those for the pure polymers. Also, the lower temperature process (GG backbone) splits into two with increasing grafting density. Very similar results were obtained for corresponding physical mixtures of GG and PEtOx, see Figure S10. In the derived thermogram, the splitting of the first process (GG degradation) is present too and therefore it does not originate from grafting. Therefore, more solid proofs of successful grafting are needed.
The grafting of PEtOx to the GG backbone was confirmed using G2 sample by means of diffusion-ordered spectroscopy (DOSY), see Figure S11. Interestingly that dilute aqueous solutions of the G1 copolymer are somewhat opaque or cloudy, which is not observed for solutions of G2 and G3. The optical transmittance of the graft copolymers does not change significantly upon heating in the range of 20-80 C ( Figure S12). Opacity may be a sign of particle formation and suggests the existence of specific interactions between GG and PEtOx (e.g., via H-bonding). This is partially confirmed by the DLS results ( Figure 4). Pristine GG shows a bimodal size distribution, which is typical for charged, highly swollen and interacting coils. The distributions of the copolymers are broad but monomodal with the peak, which corresponds to the size of the individual GG molecules. Taking into account the stronger scattering from the G1 copolymer solutions, the particles formed by G1 are expected to be denser than the coils of GG and the particles formed by G2 and G3. From the above, it is evident that the degree of grafting affects the solution properties of the synthesized copolymers.
In order to visualize the apparent difference between the GG-g-PEtOx copolymers, SEM images of the G1, G2, and G3 samples were taken. Figure 5 clearly demonstrates that the packing density of dried G1 is significantly higher than that for G2 and G3 and decreases with increasing grafting. Moreover, spherical particles are always present in the images of G1.
The nature of the interaction between GG and PEtOx and possible partial degradation of the GG backbone in the course of the F I G U R E 6 Cytotoxicity assessment of gellan gum (GG) and the GG-g-PEtOx copolymers (G1, G2, and G3). Cell viability and proliferation (A) and percentage viability compared to the TCPS control (B) after 2, 4, and 7 days in culture under the induction of 0.01, 0.1, and 1 mg/mL of the polymers grafting reaction will be investigated further and the synthesis will be optimized and improved.

| In vitro biocompatibility assessment
In vitro testing was performed in order to assess the potential of the synthesized GG-g-PEtOx copolymers for ocular drug delivery.
The validation of the biocompatibility of GG and the GG-g-PEtOx Similar to our study, cell metabolic activity assessment on hydrophilic nanogels based on partially hydrolyzed poly(2-ethyl-2-oxazoline) evidenced that they are not cytotoxic when investigated in a concentration range from 0.1 to 400 μg/mL. 39

| CONCLUSIONS
In this research, we have synthesized biologically safe polymeric carriers for topical drug delivery based on graft copolymers of GG, an anionic polysaccharide that forms transparent gels in situ in water and in the presence of metal ions, and can bind drug molecules, and PEtOx chains, which enable the penetration of the copolymers in the human tissues.
We report the successful synthesis of GG-g-PEtOx copolymers of three different degrees of grafting. A feasible purification procedure for commercial low acyl GG was developed. GG was purified prior to modification, which removed gelling impurities and significantly improved its solubility in DMSO. The synthesis procedure reported for grafting of other polysaccharides was significantly altered and applied to GG. These results confirm that the proposed materials may further be developed for use in ocular drug delivery. At this stage, our concept is based on gellan as a natural drug delivery vehicle. 40 We expect that PEtOx grafts enable penetration of GG-g-PEtOx into the eye mucus whereas GG binds and carries the drug. Future studies will evaluate the effect of PEtOX grafts on the mucoadhesive properties of gellan gum. It is expected that these properties will be inhibited similarly to previously reported effect of PEtOX on mucoadhesion of thiolated silica nanoparticles 41 and this should lead to enhancement in mucuspenetrating properties. Ultimately, the performance of both types of materials will be compared for the ability to delivery drugs to the eye.