Gelation and Cryogelation of Chitosan: Origin of Low Efficiency of Diglycidyl Ethers as Cross-Linkers in Acetic Acid Solutions

Abstract

Although diglycidyl ethers of glycols (DEs)—FDA-approved reagents for biomedical applications—were considered unsuitable for the fabrication of chitosan (CH) hydrogels and cryogels, we have recently shown that CH cross-linking with DEs is possible, but its efficiency depends on the nature of the acid used to dissolve chitosan and pH. To elucidate the origin of the low efficiency of chitosan interactions with DEs in acetic acid solutions, we have put forward two hypotheses: (i) DEs are consumed in a side reaction with acetic acid; (ii) DE chain length strongly affects the probability of cross-linking. We then verified them using FT-IR spectroscopy, rheological measurements, and uniaxial compression tests. The formation of esters in acetic acid solutions was confirmed for ethylene glycol diglycidyl ether (EGDE) and poly(ethylene glycol) diglycidyl ether (PEGDE). By the 7th day of gelation at pH 5.5, the G’_HCl/G’_HAc ratio was 5.1 and 1.5 for EGDE and PEGDE, respectively, indicating that the loss of cross-linking efficiency in acetic acid solution was less pronounced for the long-chain cross-linker. Under conditions of cryotropic gelation, only weak cryogels were obtained from acetic acid solutions at a DE:CH molar ratio of 1:1, while stable cryogels were fabricated at a molar ratio of 1:20 from HCl solutions.

1. Introduction

Chitosan-based materials of different shapes and morphologies, including beads [1], films [1], fibres [2], aerogels [3], hydrogels, and cryogels [4,5], are applied in drug delivery, tissue engineering, wound healing, environment remediation, and other fields. Finding an optimal balance between mechanical properties, chemical stability in a broad pH range, and density of accessible amino groups is a major challenge that can be overcome using an appropriate chitosan (CH) cross-linking method. In general, glutaraldehyde (GA) is used for the covalent cross-linking of CH [6,7,8,9]. Cross-linking agents, which form covalent bonds with amino groups, improve the stability of materials in acidic media at the expense of functional activity [10,11]. In addition, high GA contents are responsible for a loss of chitosan biocompatibility [6,7].

All the above stimulates the search for alternative cross-linking methods using natural cross-linkers, such as genipin [12], or conducting reactions in partially frozen media, where lower amounts of cross-linkers can be sufficient due to the cryoconcentration effect [13]. The latter approach results in the formation of cryogels, which demonstrate advantages over hydrogel analogues in toughness, elasticity, responsivity against external stimuli [14,15], and flow permeability [16]. However, due to its low reactivity, genipin is not efficient at subzero temperatures, which limits its applicability to the cross-linking of lyophilized and neutralized CH scaffolds yielding so-called macroporous cryostructurates [12]. A similar multi-step approach was applied to the fabrication of porous CH materials via solvent replacement in a frozen solution and subsequent structure stabilization using diglycidyl ether of polyethyleneglycol (PEGDE) in alkaline medium at 60 °C [17].

For a long time, it was assumed that diglycidyl ethers (DEs), FDA-approved cross-linkers for biomaterials [18,19,20], were not efficient for CH cross-linking in acidic media and, thus, for the fabrication of CH hydrogels and cryogels. Limited examples of DE reactions with CH in acidic media include the curing of diglycidyl ether of bisphenol A (DGEBA) coatings by CH at 160–200 °C [21]; the cross-linking of hydrogel films from CH solutions in 0.4% and 2% acetic acid using PEGDE at 80 °C [22]; and the fabrication of CH/PVA hydrogel nanofiber mats using ethylene glycol diglycidyl ether (EGDE) as a cross-linker in formic acid solution at 60 °C [23]. We have shown that DEs with different chain lengths can serve as efficient cross-linkers for CH at room and even subzero temperatures, if the reaction proceeds in CH solution in hydrochloric acid at pH > 4.5 [24]. Using solid-state ¹³C CP/MAS NMR spectroscopy, we also demonstrated that unreacted epoxy groups were still present in the reaction mixture of CH and EGDE at pH < 7, when the reaction in alkaline medium was already completed [24].

Kiuchi et al., who observed CH gelation with PEGDE in 0.4% but not in 2% solutions of acetic acid, suggested that the drop in nucleophilicity due to the almost complete protonation of CH amino groups at high concentrations of acetic acid was responsible for low cross-linking efficiency [22]. This hypothesis does not contradict earlier published data on the decrease of gelation time along with the increase in pH in the presence of DEs [24] and other cross-linkers [25] interacting with CH via amino groups. However, it does not explain why the efficiency of CH cross-linking with DEs in hydrochloric acid is much higher than in acetic acid at the same pH.

Since potential applications of epoxide reactions with CH under homogeneous conditions in acidic solutions are not limited to cross-linking, we consider it important to understand the origin of the low efficiency of DE-CH interactions in acetic acid solutions, which are the most widely used media to dissolve chitosan.

2. Materials and Methods

2.1. Materials

Chitosan (CH) with a degree of deacetylation of 0.9 and a molecular weight of 30 kDa was purchased from BioLog Heppe GmbH (Landsberg, Germany). Poly(ethylene glycol) diglycidyl ether, average Mn 500 (PEGDE), and 1,4-butanediol diglycidyl ether (BDDE) were supplied by Sigma-Aldrich (St. Louis, MO, USA), Ethylene glycol diglycidyl ether (EGDE) was supplied by J&K Scientific (Shanghai, China). Glacial acetic acid, hydrochloric acid, and sodium hydroxide of analytical grade were purchased from JSC Vekton (Saint Petersburg, Russia).

2.2. Methods

2.2.1. Gelation Time and Hydrogel Fabrication

First, 3% CH solutions were prepared by dissolution of the CH in hydrochloric or acetic acid at a 1:1 molar ratio to NH₂ groups, and the pH of the solutions was adjusted to 5.0 or 5.5. The pH was varied to elucidate the effect of amino group nucleophilicity on gelation. EGDE, BDDE, or PEGDE was added to the CH solutions at a DE:CH molar ratio of 1:1 while stirring; after mixing for 15 min, the reaction mixtures, labelled as CH:EGDE, CH:BDDE, and CH:PEGDE, were left to allow gelation at 25 °C, and the gelation time was determined. Then, the CH:EGDE and CH:PEGDE mixtures were kept sealed at 25 °C for 7 days to investigate evolution of their mechanical properties. The frequency sweep curves for the reaction mixtures were recorded at 25 °C using a Physica MCR 301 rheometer (Anton Paar GmbH, Graz, Austria), and gelation time was determined as described previously [26]. After gelation for 7 days, the reaction mixtures (hydrogels) CH:EGDE and CH:PEGDE were lyophilized, one half of each sample was immersed in water for 8 h, and then in a water/ethanol mixture (20:80 v/v) for 8 h to remove unreacted chemicals. Washed samples were labelled as CH:EGDE (EtOH) and CH:PEGDE (EtOH), respectively.

2.2.2. Cryogels Fabrication

EGDE and PEGDE were added under constant stirring to the 3% CH solutions in acetic or hydrochloric acid (pH 5.5) at molar ratios of DE to NH₂ groups of 1:1, 1:4, and 1:20. After mixing for 5 min at 25 °C, solutions were put into plastic syringes with an inner diameter of ~1 cm and kept at −10 °C for 12 days. After thawing, the cryogels were washed with distilled water (Scheme 1).

2.2.3. FT-IR Spectroscopy

Fourier transform infrared (FT-IR) spectra were recorded using an IR Affinity-1 spectrometer with a QATR 10 single-reflection ATR accessory (Shimadzu, Kyoto, Japan). The FT-IR spectra were recorded for (i) cross-linkers (EGDE, PEGDE); (ii) lyophilized CH solution (CH); (iii) lyophilized reaction mixtures obtained as described in Section 2.2.1 after 1 h; 1, 3, and 7 days of gelation (samples were labelled as CH:EGDE/CH:PEGDE_1h/1d/3d/7d); (iv) lyophilized and washed with EtOH reaction mixtures after 7 days of gelation (samples were labelled as CH:EGDE (EtOH) and CH:PEGDE (EtOH)); (v) cryogels obtained as described in Section 2.2 at a DE:CH molar ratio of 1:1 and labelled as CH:EGDE cryogel and CH:PEGDE cryogel.

2.2.4. Swelling and Morphology

Swelling (total swelling), swelling due to the free-flowing water in macropores (pores), and swelling of the polymer phase (walls) were determined for the never-dried cryogels, as described in [26].

The morphology of the lyophilized cryogels was investigated using a scanning electron microscope (SIGMA VP, Carl Zeiss, Jena, Germany). Prior to the analysis, the samples were coated with chromium.

2.2.5. Uniaxial Compression Test

Uniaxial compression tests were performed at a constant speed of 0.01 mm/s using a Physica MCR 301 rheometer (Anton Paar GmbH, Graz, Austria). Strain–stress curves were calculated from the normal force (F_N) and gaps recorded in loading-unloading cycle for the compression of the swollen cylindrically shaped cryogels with a diameter of 10–15 mm and a height of 8–10 mm.

3. Results

3.1. Chemical Gelation and Cryogelation in Chitosan Solutions with Diglycidylethers of Glycols (DEs)

The efficiency of CH chemical cross-linking in acetic and hydrochloric acid solutions was investigated using DEs with different chain lengths at a fixed pH value (pH 5.5) and DE:CH molar ratio (1:1). Figure 1A shows that the gelation time non-linearly depended on the DE chain length and increased in solutions of both acids in the order PEGDE < EGDE < BDDE. The gelation time was 3–3.7-fold longer in acetic acid, when short-chain cross-linkers, EGDE and BDDE, were used, while only a minor difference was observed for PEGDE. The evolution of the rheological properties of the CH solution in the presence of short-chain (EGDE) and long-chain (PEGDE) cross-linkers was monitored over a 7-day period using the oscillatory sweep test (Figure 1B–D).

Figure 1B,C show that the storage moduli of the CH hydrogel cross-linked with EGDE and PEGDE at pH 5.5 increased gradually in all solutions over 7 days and were higher for CH solutions in hydrochloric acid at all time points. The G’_HCl/G’_HAc ratio after 24 h was close to 10 for hydrogels cross-linked with EGDE; the weakness of the polymer network formed in an acetic acid solution was also evidenced by the drop in the storage modulus at a high frequency (65 Hz)—Figure 1B. In the case of a long-chain cross-linker (PEGDE), the difference between storage moduli in acetic and hydrochloric acid solutions after 24 h was negligible (Figure 1C) and correlated to the same gelation time in both solutions (Figure 1A).

By the 7th day of gelation, the G’_HCl/G’_HAc ratio was equal to 5.1 and 1.5 for EGDE and PEGDE, respectively, indicating that the loss of cross-linking efficiency in acetic acid solution was less pronounced for PEGDE. This effect cannot be attributed to the difference in amino group nucleophilicity, as suggested in [22]. However, deprotonation of the CH amino group undoubtedly has a positive effect on the cross-linking efficiency, since the hydrogel cross-linked with PEGDE at pH 5.0 (Figure 1D) was significantly weaker than that formed at pH 5.5 (Figure 1C).

Earlier, we showed [26] that a long gelation time at room temperature (>40 h at a DE:CH molar ratio of 1:20) does not exclude the possibility to fabricate mechanically stable CH cryogels. This becomes possible thanks to the effect of cryoconcentration, which consists of the enhancement of cross-link formation in a partially frozen system due to the concentration of the reagents in the unfrozen microphase [27]. However, despite the notably shorter than 40 h gelation time (Figure 1A), CH cryogels obtained from acetic acid solutions (Figure 2A) had a swelling degree 6.5–6.7-fold higher than those obtained from HCl solutions (Figure 2B), a looser porous structure (Figure 2C), and much weaker mechanical properties even in comparison with the cryogels formed at lower concentrations of the cross-linkers in HCl solutions (Figure 2D). A minimal difference in cryogel swelling (<50%) depending on the type of acid used to dissolve CH was observed when PEGDE was used as a cross-linker (Figure 2B). The PEGDE-cross-linked cryogel formed from the acetic acid solution was significantly stronger than EGDE-cross-linked one, although in the HCl solution, EGDE yielded cryogel with higher compressive strength (Figure 2D). It should be mentioned that CH cryogels, which are stable enough to retain the shape of a mould, were not obtained from acetic acid solutions at DE:CH molar ratios of 1:4–1:20 (Figure 2A). More detailed information on how variation in the DE chain length and DE:CH molar ratio affects mechanical properties, swelling, and susceptibility to the enzymatic hydrolysis of CH cryogels obtained from HCl solutions is reported in [26].

3.2. Investigation of Cross-Linking Mechanism Using FT-IR Spectroscopy

Such a strong effect of the DE chain length on the cross-linking efficiency at the same pH value in acetic and hydrochloric acids cannot be explained by the drop in nucleophilicity due to the protonation of CH amino groups, as suggested in [22]. Thus, we have hypothesized that it can originate from competition between target reactions (cross-linking) and side reactions of epoxy groups with water or carboxylic groups of acetic acid yielding esters, both of which are known for DEs [28] (Scheme 2).

Interactions between CH, DEs, and acetic acid were investigated by means of FT-IR spectroscopy, focusing on the analysis of the following spectral regions:

(i)

Between 1730 and 1750 cm⁻¹ to follow ester formation in the side reaction between DE and acetic acid;

(ii)

Between 1200 and 1000 cm⁻¹ to follow changes in the position of partially overlapping bands corresponding to C-O-C bridges in DEs (1092–1099 cm⁻¹), and 1,4-glycosidic bonds C-O-C (1152 cm⁻¹) and C-O stretching (1064 cm⁻¹ and 1024 cm⁻¹) in CH [29,30] depending on the cross-linking conditions;

(iii)

Around 910 cm⁻¹ to follow the kinetics of the epoxy ring opening and to identify unreacted epoxide groups in the final products, where they can be distinguished from symmetric vibrations of 1,4-glycosidic bonds C-O-C at 945 and 895 cm⁻¹ in CH [31].

Figure 3A,B show that the band at 1735 cm⁻¹ appeared only in reaction mixtures of CH with EGDE and PEGDE in acetic acid, which proves the formation of esters at least in one of the possible side reactions (Scheme 2, reactions 2 or 5). The relative intensity of the band at 1735 cm⁻¹ increased over time from 1 h to 7 days (Figure 3C). This shows that PEGDE was consumed in both target and side reactions, since the storage moduli of the reaction mixture (Figure 1C) also increased during this period. The absence of this band in the spectra of hydrogels washed with water and ethanol (Figure 3A,B) suggests that at room temperature, the side reaction proceeds without CH participation (Scheme 2, reaction 2). At the same time, a small shoulder in this region for CH:EGDE (EtOH) hydrogel does not exclude esterification of the terminal epoxy group in the EGDE graft completely (Scheme 2, reaction 5). After reaction at subzero temperatures, the band at 1735 cm⁻¹ in the spectrum of EGDE-cross-linked cryogel was identified more undoubtedly, confirming EGDE graft esterification. Since esterification blocks further participation of the second epoxy group of EGDE in cross-linking reactions in acetic acid solution, the mechanical properties of EGDE-cross-linked hydrogel (Figure 1B) and cryogel (Figure 2D) were much more weaker in comparison with those of PEGDE-cross-linked ones (Figure 1C and Figure 2D).

Comparison of FT-IR spectra of the reaction mixtures before and after washing with EtOH in the wavenumber range of 1200–800 cm⁻¹ showed (Figure 3D,E) that after washing, the maxima of the overlapped bands corresponding to C-O-C in DEs and C-O bonds in CH shifted to lower wavenumbers. In addition, the shift was more pronounced for the hydrogel formed from the CH solution in acetic acid. These changes can be attributed to the increase in the CH content in the washed hydrogel (EtOH) over the reaction mixtures after removal of the unreacted chemicals and soluble products of the side-reactions. A similar effect was observed for PEGDE-cross-linked cryogel. In this case, the higher content of CH in cryogel formed from acetic acid solution was also evident from the appearance of bands at 1152 and 895 cm⁻¹, corresponding to C-O-C bonds in CH. In the case of hydrogel and cryogel cross-linked with EGDE, changes in the spectral range of 1026–1094 cm⁻¹, where the bands of C-O bonds in CH and C-O-C bonds in the cross-linker overlap, were less evident due to the lower relative weight % of EGDE in the equimolar reaction mixtures compared to PEGDE. However, more intensive bands at 1152 and 895 cm⁻¹ in the spectra of the washed hydrogel (EtOH) and cryogel formed from acetic acid solution confirm a lower cross-linking degree in comparison with those formed from HCl solutions. Despite the lower reactivity of DEs at subzero temperatures, the unreacted epoxy groups were not identified in the FT-IR spectra of cryogels.

Detailed analysis of the reaction mixtures’ spectra in the region near 910 cm⁻¹ over time from 1 h to 7 days confirmed the faster opening of EGDE epoxy rings in acetic acid solution (Figure 4A); in the HCl solution, unreacted epoxy groups were identified after 3 days of gelation, while in acetic acid, corresponding bands partially overlapped with C-O-C symmetric vibrations in CH, shifted to lower wavenumbers after 1 day of gelation, and disappeared by the 3rd day. This correlates to a drastic drop in the strength of EGDE-cross-linked hydrogel formed from acetic acid solution (Figure 1B). The reactivity of PEGDE in acetic acid was lower, and unopened epoxy rings were identified in the reaction mixture at least up to the 3rd day of gelation, thus making cross-linking more efficient in comparison with EGDE (Figure 4B).

4. Discussion

To explain the low efficiency of CH cross-linking with DEs in acetic acid solutions, we have considered the following factors: (i) decrease in amino group nucleophilicity due to protonation, when CH was dissolved at a stoichiometric excess of acid, as suggested in [22]; (ii) consumption of DEs in side reaction with acetic acid; (iii) dependence of epoxy ring-opening reactions kinetics and probability of target (cross-linking) reaction on DE chain length.

We have previously shown [24] that the gelation time in CH solutions in the presence of EGDE sharply decreased at pH > 4.5, which correlates to the deprotonation of CH amino groups, in a way similar to that observed for other cross-linkers [25]. Indeed, increasing the pH from 5.0 to 5.5 in the CH-PEGDE solution in acetic acid resulted in a notable increase in hydrogel strength (Figure 1C,D), confirming the positive effect of amino group deprotonation on cross-linking efficiency. However, at pH 5.5, we still observed significantly longer gelation times (Figure 1A) and the formation of weaker CH hydrogels (Figure 1B,C) in acetic acid solutions. In addition, the difference in DEs’ performance as cross-linkers in solutions of these two acids was dependent on the DE chain length: the loss of efficiency in acetic acid solutions was the lowest in the case of the long-chain cross-linker, PEGDE. PEGDE-cross-linked cryogel formed from the acetic acid solution had the highest mechanical strength and the lowest swelling in the row of DEs (Figure 2), although in hydrochloric acid solution, PEGDE yielded the softest and most swellable cryogel (Figure 2B,D) [26]. This led us to the conclusion that a decrease in the nucleophilicity of CH amino groups negatively affects CH cross-linking with DEs, but it is, most likely, not the main factor responsible for the inefficiency of DEs in acetic acid solutions at room and subzero temperature.

An alternative hypothesis that the drop in the cross-linking efficiency in acetic acid solutions originates from competition between the target reaction (cross-linking) and side reactions—deactivation of DE epoxy groups in side reactions with water and acetate (Scheme 2)—was verified using FT-IR spectroscopy. The formation of an ester between DEs and acetate (Scheme 2, reactions 2 and 5) was confirmed by the appearance of a new band at 1735 cm⁻¹ in reaction mixtures of CH with EGDE and PEGDE in acetic acid solution at room temperature (Figure 3A,B) and an increase in its intensity with time (Figure 3C). Despite the similarity of the spectral changes in EGDE- and PEGDE-cross-linked materials, there was an obvious difference in the kinetics of epoxy ring-opening reactions (Figure 4) and the structure of hydrogels and cryogels after the removal of insoluble products (Figure 3D,E).

The absence of the ester band in the spectrum of PEGDE-cross-linked cryogel and hydrogel washed with the EtOH/water mixture (Figure 3E) suggests that the reaction between PEGDE and acetate proceeded without the participation of CH at both room and subzero temperatures (Scheme 2, reaction 2). The lower cross-linking density in acetic acid solutions results from PEGDE consumption in side reactions, but the structure of the cross-linked network remains similar to that formed in the hydrochloric acid solution (Scheme 2, reaction 1). In contrast, ester groups were detected in EGDE-cross-linked hydrogel and cryogel formed from acetic acid solutions (Figure 3D). This can be attributed to the formation of the EGDE grafts (Scheme 2, reaction 3), in which the esterification of the terminal epoxy group (Scheme 2, reaction 5) with acetate blocks further possibilities of cross-linking. Since the kinetics of EGDE epoxy ring-opening reactions is faster than that of PEGDE (Figure 4) and the probability of interchain cross-linking increases with an increase in the DE chain length, deactivation of EGDE in side reactions (Scheme 2, reactions 4 and 5) proceeds faster in comparison with PEGDE. At subzero temperatures, due to the lower mobility of polymer chains, the consumption of DE in side reactions in acetic acid solution proceeds faster than cross-linking, resulting in the formation of extremely weak EGDE-cross-linked cryogel (Figure 2D).

5. Conclusions

For a long time, it was considered that FDA-approved diglycidyl ethers (DEs) are not efficient for CH cross-linking in acidic media and the fabrication of CH hydrogels and cryogels suitable for biomedical applications. Only a few known attempts to cross-link CH with DE under homogeneous reaction conditions were performed in acetic acid solutions and resulted in the formation of very weak hydrogels, despite the high amounts of cross-linkers used. Recently, we have shown that mechanically stable CH hydrogels and cryogels can be obtained using DEs, if the reaction proceeds in hydrochloric acid solutions at pH > 4.5, but the origin of the low efficiency of DE-CH interactions in acetic acid solutions has not been elucidated.

Here, we have investigated the gelation and cryogelation of CH with DEs in acetic and hydrochloric acid solutions at the same pH value and shown that the gelation time was longer and the mechanical strength of hydrogels and cryogels was lower if the reaction proceeded in acetic acid solutions. Since the observed effects could not be explained by the difference in amino group nucleophilicity, two alternative hypotheses were put forward and proved using a combination of FT-IR spectroscopy, rheological measurements, and uniaxial compression tests: (i) the consumption of DEs in the side reaction with acetate is responsible for the low cross-linking efficiency of DEs in acetic acid solutions; (ii) the reaction mechanism and probability of side reactions is affected by the DE chain length.

The data obtained allowed us to conclude that epoxy–amine reactions in CH solutions should be conducted in the presence of hydrochloric acid at a pH value corresponding to the minimal degree of CH protonation, which is still sufficient to assure polymer solubility.