Diels–Alder “click” chemistry for the cross-linking of furfuryl-gelatin-polyetheramine hydrogels

Abstract

This paper deals with the design of biopolymeric hydrogels after the chemical cross-linking of gelatin through the Diels–Alder (DA) reaction. Furan modified gelatin (Gel-FGE) was prepared by the reaction of furfuryl glycidyl ether with the free amino groups present in the gelatin. A Jeffamine®-based bismaleimide was employed as a cross-linking agent and the influence of the amount of cross-linker used on the final hydrogel properties was studied. The DA and retro-DA reactions were followed by ultraviolet spectroscopy and the final properties of the hydrogel assessed. Scanning Electron Microscopy was used to analyze the structure of the final material and rheology studies confirmed the formation of a chemically cross-linked network. The swelling behavior in response to external stimuli such as pH and salt concentration was also studied. By virtue of the DA “click” reaction, biopolymer-based cross-linked hydrogels, with promising properties for biomedical applications, were obtained in a simple one step procedure free of catalysts, additives or coupling agents.

---

Introduction

Biopolymer-based materials are receiving increasing interest for their potential applicability in the biomedical field. Biopolymers are natural polymers which are known to mimic the extracellular matrix and promote tissue growth.¹ One of the advantages in the use of these materials is their similarity to the components of the tissues they replace or contact, leading to high biocompatible as well as biodegradable systems.² Furthermore, they are composed of a wide variety of functional groups, compared with synthetic polymers, opening the way to diverse coupling chemistries.

Gelatin is derived from the partial acidic or alkaline hydrolysis of collagen, commonly used for biomedical applications.³ Gelatin is biodegradable, biocompatible and relatively inexpensive, important features for the design and development of new functional materials. Due to its solubility at physiological pH and its capability to form triple helices with a sol–gel transition between 25 and 30 °C, gelatin appears to be an ideal precursor for hydrogel formation.^3,4 Hydrogels are three-dimensional polymeric networks which are able to swell and retain huge amounts of water.⁵ The biocompatibility of these materials is promoted by their high water content making them similar to hydrated body tissues. Since hydrogels are able to show responsive properties to a variety of external stimuli, these materials have attracted much attention regarding their use in biomedicine.⁶

However, the applications of gelatin-based biomaterials are still hindered by some drawbacks such as poor mechanical properties and water sensitivity. Chemical cross-linking has been widely studied with the aim of enhancing the thermal and mechanical properties of gelatin-based physical hydrogels. The chemical composition of the gelatin, composed of a large variety of amino acids (glycine, alanine, lysine, and hydroxylysine among others) offers the possibility of achieving a wide number of chemical modifications via its amino, carboxyl or hydroxyl groups. Traditionally, bifunctional cross-linkers have been employed such as glutaraldehyde, genipin or carbodiimides.^7–9 Covalently cross-linked gelatin hydrogels were also obtained through thiol–ene reaction between thiolated gelatin and poly(ethylene glycol) diacrylate.¹⁰ Vlierberghe et al. prepared disulfide-cross-linked gelatin hydrogels from thiolated gelatin derivatives.¹¹

Among the various strategies which can be followed for the formation of cross-linked gelatin hydrogels, “click” reactions emerge as suitable candidates since they usually take place under mild reaction conditions and in absence of side events. The Diels–Alder (DA) reaction is a “click” type reaction which consists in a [4 + 2] thermo-reversible cycloaddition between a diene and a dienophile.¹² More specifically, the reaction is carried out in the absence of catalysts or initiators, which could affect the biocompatibility of the final material.¹³ This reaction has been used previously to build up polymeric hydrogels,^13–16 mainly starting from (meth) acrylic polymers or polysaccharides. The use of modified gelatin was recently reported by Yu et al. as a component of a DA cross-linked interpenetrating hydrogel, whit hyaluronic acid and chondroitin sulphate.¹⁷ In this work, promising hydrogels were developed entirely composed of furan modified gelatin and the as-prepared bismaleimide cross-linker. Furan groups were attached to gelatin by the epoxy-amine reaction with furfuryl glycidyl ether, and then cross-linked with Jeffamine®-based bismaleimides.

Jeffamines®, aliphatic polyether diamines composed of polypropylene oxide and polyethylene glycol units, have been used before for the cross-linking of polysaccharides, diglycidyl ether of bisphenol A or glycidyl methacrylate copolymer.^18–20 Recently, Jeffamines® have been modified with maleic anhydride for the design of a suitable dienophile for the formation of DA networks based on epoxy-amine type oligomers bearing furan groups.²¹ In this study, due to its water solubility, along with its potential biocompatibility,²² Jeffamine® ED 900 was employed for the synthesis of a water soluble bismaleimide. Indeed, Jeffamines® have been used in biomedicine in combination with other polymers for the preparation of gene delivery systems²³ and for the design of latexes and nanocapsules for drug delivery applications.²²

In the present work, gelatin was modified by the reaction of the free ε-amino groups (mainly in lysine and hydroxylysine residues) of the main chain with furfuryl glycidyl ether. For hydrogel formation, the Diels–Alder reaction was carried out between furan-modified gelatin and a Jeffamine®-based bismaleimide. The as-designed hydrogels were studied in terms of their microstructure and their rheological and swelling behavior. Finally, the thermal reversibility of the DA reaction was analyzed by means of spectroscopic techniques.

Materials and methods

Materials

Gelatin (from porcine skin type A ≈ 300 bloom), furfuryl glycidyl ether (FGE, 96%), sodium bicarbonate (99.7–100%), Jeffamine® ED 900 (O,O-bis (2-aminopropyl) polypropylene glycol-b-polyethylene glycol-b-polypropylene glycol), sodium acetate trihydrate (≥99.0%) and 2,4,6-trinitrobenzenesulfonic acid solution (TNBS, 5% w/v in H₂O) were purchased from Sigma-Aldrich. Acetic anhydride, triethylamine and hydrochloric acid solution (0.1 N) were purchased from Panreac. Acetone (99.5%) was purchased from Oppac S.A. Chloroform was purchased from Lab Scan Analytical Sciences. Sodium hydroxide solution (0.5 N) was purchased from Scharlau. Deuterium oxide (deuteration degree 99.96%) was purchased from Merck. Deionized water was employed as solvent. All reagents and solvents were employed as received.

Methods

Fourier Transform Infrared Spectroscopy (FTIR). FTIR spectra were obtained using a Nicolet Nexus spectrophotometer at room temperature. KBr pellets were used in the range from 4000 to 600 cm⁻¹, with 4 cm⁻¹ resolution and 40 scans. For bismaleimide and hydrogel samples spectra were run with a Specac MKII Golden Gate accessory equipped with a diamond crystal at a nominal incident angle of 45° and ZnSe lens.

Nuclear Magnetic Resonance Spectroscopy (NMR). Proton and carbon NMR (¹H and ¹³C NMR) spectra were recorded with an Avance Bruker equipped with BBO z-gradient probe. Experimental conditions were as follows: (a) for ¹³C NMR: 125.75 MHz, number of scans 14 000, spectral window 25 000 Hz, and recovery delay 2 s; (b) for ¹H NMR: 500 MHz, number of scans 64, spectral window 5000 Hz, and recovery delay 1 s. The solvent employed in all cases was D₂O.

Determination of residual amino groups in gelatin-based polymers. The extent of gelatin modification was assessed by a spectrophotometric method using TNBS to determine the free amino groups in modified gelatin samples.²⁴ The reaction of TNBS with the free ε-amino groups remaining in the lysine and hydroxylysine residues produces its trinitrophenyl derivative, TNP-lys, which absorbs in the UV region. Typically, 11 mg of modified gelatin, 1 mL of a 4% w/v NaHCO₃ solution and 1 mL of 0.5% w/v TNBS were incubated at 40 °C for 4 h. 3 mL of HCl 6 N were added and the mixture was autoclaved for 1 hour at 120 °C. The hydrolysate was then diluted with water and extracted with diethyl ether. 5 mL of the aqueous phase were heated for 15 min in order to eliminate the residual ether, cooled to room temperature and diluted with 15 mL of water. The absorbance was measured at 346 nm in a UV-vis spectrometer with a 1 cm optical path quartz cuvette. Blanks were prepared by the same procedure except that the HCl was added before TNBS to avoid any reaction of amino groups of the gelatin with TNBS. Three replicates were used in each determination. The same procedure was used for neat gelatin samples which were used as controls. The relationship between the absorbance and the moles of ε-amino groups per gram of gelatin was obtained using eqn (1):where 1.46 × 10⁴ L mol⁻¹ cm⁻¹ is the molar absorptivity of TNP-lys, b is the cell path length in cm, and x is the sample weight in grams.

Differential Scanning Calorimetry (DSC). Calorimetric measurements were performed using a DSC (Mettler DSC822). Thermograms were obtained with a heating rate of 10° min⁻¹ in the temperature range of 25–250 °C and under inert atmosphere (10 mL min⁻¹).

Scanning Electron Microscopy (SEM). Scanning Electron Microscopy experiments were performed by a JEOL JSM-6400 with a wolframium filament operating at an accelerated voltage of 20 kV and at a working distance of 15 mm. Freeze-dried samples were coated with approximately 20 nm of chromium using a Quorum Q150 TES metallizer.

UV-vis spectroscopy (UV-vis). The DA cycloaddition of the polymers was followed by UV-vis Spectroscopy using a UV-3600/3100 from Shimadzu equipped with a thermoelectric cell holder at 65 °C, operating in a scan range of 200 to 500 nm with a 0.1 mm optical path quartz cuvette. UV spectra were taken every 30 min for 24 h. For the retro-DA reaction the temperature was raised to 90 °C and the same procedure was followed.

Rheological measurements. The dynamic rheological behavior of the hydrogels was measured with a Rheometric Scientific Advanced Rheometric Expansion System (ARES), using parallel plate geometry (25 mm diameter). Frequency sweep measurements were performed at 37 °C from 0.1 to 500 rad s⁻¹ at a fixed strain in the linear viscoelastic region, previously assessed by strain sweep experiments.

Swelling. The swelling capacity of freeze-dried hydrogels was studied by a general gravimetric method. Samples (n = 3) were incubated in deionized water, physiological solution (NaCl 0.9% w/v) and simulated gastric fluids (HCl 0.1 M). At selected time intervals of 0.5, 1, 1.5, 2, 24 and 48 h, the swollen hydrogels were removed, the excess of water absorbed with a filter paper, and weighed. Hydrogels were then replenished with fresh solution. The swelling ratio (SR) was calculated using eqn (2):where W_s and W_d are the weight of the swollen and dried hydrogel samples respectively. The equilibrium swelling was considered to be achieved when the weight of the hydrogels no longer increased.

Synthesis of furfuryl-gelatin (Gel-FGE). Gelatin (1 g) was dissolved in 200 mL of water at 40 °C. The pH was adjusted to 11 using an aqueous NaOH solution 0.5 N. After cooling the solution in an ice bath, FGE (401 μL) was added and the solution was stirred at 55 °C for 24 h in the dark under a nitrogen atmosphere (Scheme 1). The solution was then neutralized to pH = 7 using an aqueous 0.1 N HCl solution and purified by dialysis against deionized water for 48 h using a Spectra/Por Dialysis Membrane, MWCO 12–14 000 (Spectrum Laboratories). Gel-FGE was finally obtained as a yellowish solid after freeze-drying.

Scheme 1 Modification of gelatin with furan groups using FGE.

Synthesis of bismaleimide (BMI). Water soluble bisdienophiles were prepared by the modification of Jeffamine® ED with maleic anhydride following a procedure reported elsewhere (Scheme 2).²⁵ First, a solution of maleic anhydride (51.0 mmol) in 50 mL of chloroform was kept under nitrogen at 10 °C for 3 h. The Jeffamine® ED (25.9 mmol) was added dropwise to it under stirring and the mixture allowed to warm to room temperature for 2 h. The ensuing bismaleamic acid (BMA) was dried under vacuum as a yellow viscous liquid. Then, the BMA (4.56 mmol), triethylamine (2.96 mmol) and sodium acetate trihydrate (3.47 mmol) were stirred in 10 mL of acetone under nitrogen atmosphere. Acetic anhydride (28.96 mmol) was added and the temperature and the resulting mixture was heated at reflux (70 °C) for 2.5 h. About 5 mL of acetone were rotatory evaporated and the viscous solution (BMI) was dried under vacuum at a temperature below 60 °C.

Scheme 2 Bismaleimide synthesis from Jeffamine® ED.

Hydrogel formation. For hydrogel preparation Gel-FGE was dissolved in deionized water (60 wt%) and mixed with different amounts of BMI (Scheme 3). Different furan-to-maleimide ratios were used in order to study the influence of employing an excess of cross-linking agent on the final hydrogel properties. Typically, Gel-FGE to BMI weight ratios for hydrogel formation were 1 : 1.0, 1 : 1.5 and 1 : 2.0. The mixture was allowed to gel at 65 °C for 5 h. Hydrogels are referred hereafter as shown in Table 1.

Scheme 3 Hydrogel formation through DA reaction between Gel-FGE and BMI.

Table 1 Composition of the different hydrogel formulations

Hydrogel	Gel-FGE (wt%)	Gel-FGE : BMI (w/w)
HGEL1-1.0	60	1 : 1.0
HGEL1-1.5	60	1 : 1.5
HGEL1-2.0	60	1 : 2.0

---

Results and discussion

Synthesis and characterization of furfuryl-gelatin

Furan modified gelatin was prepared by the modification of the non protonated ε-amino groups of lysine and hydroxylysine and the terminal amino acid residues by reaction with furfuryl glycidyl ether, following a procedure reported in the literature with some modifications.²⁶ According to the literature, the isoelectric point of gelatin from porcine skin is close to 8.5 and the pK_a value for the ε-amino group of lysine is 10.53.²⁷ For this reason, the reaction was carried out at strongly basic pH in order to ensure the availability of the non-protonated amino groups for the modification reaction. The chemical structure and the degree of substitution (DS) were confirmed by ¹H NMR (Fig. 1). The peaks at 6.38, 6.41 and 7.46 were attributed to the presence of the furan ring.²⁸ From the peak area of the furan signal it was demonstrated that Gel-FGE contained 2.20 mol% of furan groups. The content of lysine residues in gelatin is known to be 2.69 mol%.²⁷ Considering this value, the DS calculated using ¹H NMR spectroscopy was 82%, with respect to the total content of available ε-amino groups.

Fig. 1 ¹H NMR spectra of (a) gelatin and (b) Gel-FGE in D₂O.

At this point, there is no apparent evidence that the secondary amine generated by the epoxy-primary amine reaction cannot react further with an epoxy moiety leading to complete amine substitution. However, considering the fact that the second reaction usually shows a significantly lower rate constant and the steric hindrance of the second substitution, it is likely that the furfuryl glycidyl ether was predominantly added to the primary amines of gelatin leading to secondary amines.^29,30

In order to confirm that furan ring was bonded to amine groups and not merely present in the polymer TNBS titration experiments were performed. The reaction of TNBS with ε-amino groups could be used to determine the number of unreacted lysine residues after the modification, using a UV assay for their quantification. Thereby, the number of modified groups is equal to the resulting difference between the chemically determined number of free ε-amino groups before and after the modification reaction.²⁴ The assay value calculated from the moles of free ε-amino groups per gram of gelatin resulted in a DS of 85%, which was consistent with the aforementioned DS calculated by ¹H NMR spectroscopy. According to the coincident values resulting from both techniques, it can be concluded that not only the furan unit was bonded to the primary amines of gelatin but also that there was a unique furan ring per amine. Otherwise, if a second substitution would have taken place to a significant extent a clear difference between the substitution degree calculated by ¹H NMR and the one calculated using TNBS titration should have been observed.

Gel-FGE was also characterized by FTIR spectroscopy (Fig. 2). The characteristic bands of gelatin for amide I and amide II could be observed at 1630 cm⁻¹ at 1540 cm⁻¹, respectively (referred as I and II, respectively, in Fig. 2). The amide I band corresponded to the C=O stretching vibration, whereas the amide II band is associated with the N–H bending vibration and the C–N stretching vibration.²⁷ As the total furan content of the gelatin samples was low, only slight differences were detected between gelatin and Gel-FGE spectra. However, in the case of the Gel-FGE, at 920 cm⁻¹ a band corresponding to =C–H deformation vibration of the furan ring was observed.³¹

Fig. 2 FTIR spectra of (a) gelatin and (b) Gel-FGE.

DSC measurements were conducted in order to analyze the thermal behavior of the new gelatin furanic derivate. First and second scans for gelatin and Gel-FGE samples are shown in Fig. 3. A dramatically different thermal behavior for Gel-FGE samples compared with that of the non-modified gelatin was observed. In the case of gelatin samples, two endothermic peaks associated to absorbed water evaporation and the helix–coil transition near 100 °C and 220 °C, respectively, were detected in the first scan (Fig. 3a). Both transitions disappeared upon re-heating after rapid cooling, as expected. On the contrary, in the case of the Gel-FGE samples, the endothermic peak corresponding to the amount of triple helix in gelatin did not appear in the first scan,³² indicating that the introduction of the furan ring into the biopolymer structure hindered the rearrangement of the triple helix by disturbing the formation of hydrogen bonds between gelatin chains.³² Besides, the glass transition temperature (T_g) (calculated at the onset of the T_g region) obtained from the second scan decreased significantly from 190 °C for the neat gelatin to 170 °C for the Gel-FGE derivative (Fig. 3b). The disruption of the interaction between gelatin chains provoked by the presence of the furan ring could facilitate the chain mobility and, consequently, the T_g decreased.

Fig. 3 DSC thermograms of gelatin and Gel-FGE. (a) First scan and (b) second scan.

Synthesis of bismaleimide (BMI)

The synthesis of the bisdienophile started from the modification of the polyetheramine Jeffamine® ED 900 with maleic anhydride, resulting in the bismaleamic acid (BMA), followed by a cyclization reaction of the BMA to a bismaleimide (BMI). The FTIR spectra of Jeffamine® ED 900, BMA and BMI are shown in Fig. 4. The BMA spectrum showed two peaks at 1718 and 1632 cm⁻¹ corresponding to the carbonyl stretching vibration of the –COOH and –CONH, respectively (Fig. 4b). In the case of BMI, two peaks at 1752 and 1705 cm⁻¹ were observed, associated to the stretching vibration of the carbonyl groups of the maleimide indicating that the cyclization had been achieved successfully (Fig. 4c).²¹

Fig. 4 FTIR spectra of (a) Jeffamine® ED 900, (b) BMA and (c) BMI.

Fig. 5 and 6 show the ¹H and ¹³C NMR spectra of Jeffamine® ED 900 and BMI, respectively. The presence of the maleimide ring was confirmed by the appearance of a peak at 6.21 ppm (Fig. 5b). Nevertheless, according to the results, BMI seemed not to be completely cyclized because the peaks corresponding to the protons of the double bond of the open maleamic acid were observed at 6.18 and 6.58 ppm.²¹ In the ¹³C NMR spectra (Fig. 6), the peak corresponding to the C=C of the maleimide ring appeared near 135 ppm and that at 171 ppm corresponded to the C=O. However, the presence of open maleamic acid moieties was again confirmed by the presence of more than one peak near 135 and 171 ppm.

Fig. 5 ¹H NMR spectra of (a) Jeffamine® ED 900 and (b) BMI in D₂O (* refers to solvent impurities).

Fig. 6 ¹³C spectra of (a) Jeffamine® ED 900 and (b) BMI in D₂O (* refers to solvent impurities).

The average functionality (f) of the samples was determined by comparing the integration of the proton signals associated to the maleimide and the remaining maleamic acid as described in the literature eqn (3):²¹

where A_h is the area of the protons of the double bond of the maleimide ring (designed as h in Fig. 5) and A_h′ is the area of the protons of the double bond of the remaining open maleamic acid (designed as h′ in Fig. 5). The obtained f value was 1.42, indicating that the compound was not a pure bismaleimide but a mixture of mono and bismaleimide. However, the fact that the bismaleimide prepared for this purpose contained some monomaleimide homologue did not prevent the achievement of cross-linked gelatin-based gels using the furan-maleimide Diels–Alder reaction.

DA for hydrogel formation

DA hydrogels were obtained after the reaction of Gel-FGE with BMI for 5 h at 65 °C. A 60 wt% solution of Gel-FGE was mixed with different amounts of bismaleimide in order to study the influence of the furan to maleimide ratio employed on the final properties of the hydrogels. As mentioned previously, even if the BMI contained monomaleimide, that could react with the appended furan rings giving rise to short branching, the cross-linking took place efficiently. The amount of bismaleimide was amply sufficient to induce gelation, as shown by the formation of the expected gels. The chemical structure of the final hydrogel was analyzed by FTIR spectroscopy in order to confirm the DA reaction (Fig. 7). The adduct formation was confirmed by the bands appearing at 1014 and 1375 cm⁻¹ due to the C–O–C stretching vibration and the C=C stretching vibration, respectively.^13,31 However, it has to be considered that since some pending primary amino groups remained unmodified, their nucleophilic Michael addition to the double bond of the bismaleimides could compete with the DA cycloaddition. In order to discard the existence of this interference in the cross-linking mechanism for hydrogel formation a blank experiment was performed. Unmodified gelatin, with all the primary ε-amine groups available, was made to react with BMI under the same conditions as for the as-prepared DA hydrogels. No reaction was detected after 5 h at 65 °C, no gel formation ensued and the viscosity of the medium did not increase during that time. In the light of these results, it can be concluded that the cross-linking of Gel-FGE by BMI arose indeed from the intermolecular Diels–Alder reaction.

Fig. 7 FTIR spectra of gelatin-based hydrogel, HGEL1-1.0.

The DA reaction was also followed by UV-vis spectroscopy and, the formation of hydrogel HGEL1-1.5 was chosen as the model (Fig. 8). 0.08 g of Gel-FGE were mixed with 0.11 g of BMI in 50 μL of deionized water and the solution was transferred to a 0.1 mm cell. Measurements were conducted at 65 °C. The loss of the conjugation between the two carbonyl groups (nπ* (C=O)) and the double bond (ππ* (C=C)) of the maleimide ring of the BMI, which involves the formation of the DA adduct, could be monitored by the progressive decrease in the intensity of the absorption band at ∼300 nm.³³ During the formation of the chemical hydrogel, as the cross-linking reaction proceeded, the absorption band associated to the maleimide decreased and no absorption was observed in that region after 24 h, indicating that the reaction had been completed.

Fig. 8 UV-vis spectra following the progress of DA reaction of HGEL1-1.5 in H₂O at 65 °C.

It is well known that hydrogels require an interconnected porous structure when acting as drug carriers or scaffolds in tissue-engineering applications.¹⁶ In order to examine the microstructure of the hydrogels freeze-dried samples were analyzed by SEM. Fig. 9 shows the SEM images of the three different hydrogel compositions where, prior to lyophilization, the pores were occupied by water. HGEL1-1.5 presented a continuous and regularly distributed porous microstructure with a pore size varying from 9 to 80 μm. On the contrary, in the case of HGEL1-1.0 and HGEL1-2.0, pits and grooves were observed with sizes ranging from 8 to 50 μm and from 16 to 120 μm, respectively. Therefore, it could be concluded that the amount of cross-linker played an important role in the final microstructure of the hydrogels. The influence of the cross-linker concentration on the morphology of gelatin-based hydrogels was previously analyzed by other authors who deduced that the use of lower amounts of cross-linker increased the surface morphology of the final material.³⁴ In the present work, the highest porous microstructure was obtained when using an intermediate amount of cross-linking agent. However, HGEL1-1.0 and HGEL1-1.5, with lower cross-linker concentration than HGEL1-2.0, showed increasing surface morphology. On the other hand, the difference between the morphology of HGEL1-1.0 and HGEL1-1.5 could be attributed to the fact that higher cross-linker amounts would lead to more cross-linked networks. Thus, the number of junction points would increase, influencing the pore size and its distribution.³⁵ In the case of HGEL1-2.0, the use of a higher excess of cross-linking agent did not result in the formation of a more porous structure, indicating that the degree of cross-linking was not increased by adding further excess of BMI.

Fig. 9 SEM images of (a) HGEL1-1.0, (b) HGEL1-1.5 and (c) HGEL1-2.0.

Rheological analysis

Dynamic rheological analysis was performed with the aim of studying the viscoelastic properties of the different hydrogel compositions. Rheological measurements were carried out at 37 °C using a parallel plate geometry. All samples were previously submitted to strain sweep tests in order to establish the linear viscoelastic region where both the storage (G′) and the loss modulus (G′′) were independent of the applied strain. In all cases, G′ was constant and always higher than G′′ indicating that the synthesized hydrogels were chemically cross-linked (Fig. 10).³⁶ The mean G′ values are reported in Table 2. As can be observed, the G′ value of HGEL1-1.5 was slightly higher than that of HGEL1-1.0, which could be attributable to the fact that higher amounts of bismaleimide would lead to a more cross-linked network. On the contrary, for HGEL1-2.0 the G′ value was similar to that of HGEL1-1.0, suggesting that a wide excess of cross-linking agent did not improve the elastic properties of the hydrogels and could, instead, act as a plasticizer lowering the G′ value. This result is in accordance with the morphology observed in the SEM images where HGEL1-1.5 displayed the best interconnected porous microstructure, whereas the addition of more cross-linking agent, in the case of HGEL1-2.0, did not result in a more cross-linked network since the porosity was not improved. The elastic properties of these materials are crucial in terms of tissue engineering applications. The mean values of the storage modulus obtained for all the hydrogel compositions were in the range of those of liver, fat, relaxed muscle and breast gland tissue, namely 10³ to 10⁴ Pa.³⁷

Fig. 10 Frequency sweep of hydrogels. G′ (filled symbols) and G′′ (empty symbols). ■ HGEL1-1.0, ▲ HGEL1-1.5, ● HGEL1-2.0.

Table 2 Storage modulus (G′) values of hydrogels at 37 °C (average ± standard deviation, n = 3)

Hydrogel	G′ (Pa)
HGEL1-1.0	1709.0 ± 278.9
HGEL1-1.5	2040.2 ± 59.6
HGEL1-2.0	1705.7 ± 132.3

---

Swelling

The swelling properties of the DA cross-linked hydrogels were determined gravimetrically by measuring the water uptake over a period of time until equilibrium was reached, Fig. 11a shows the pictures of the three swollen hydrogels after 24 h of swelling in water. In order to study the swelling behavior under different environments the uptake was recorded in water, acidic solution simulating gastric fluids (HCl 0.1 M) and physiological solution (NaCl 0.9% w/v) at 37 °C. Fig. 11b shows the equilibrium swelling values of the different hydrogel compositions in the three different media. The equilibrium was reached after 2 h of incubation in the case of water and after 24 h of incubation in the case of the acidic and physiological solutions.

Fig. 11 (a) Hydrogels before and after swelling for 24 hours in water, (b) equilibrium swelling ratio of hydrogels at 37 °C in the different media employed; water, HCl (0.1 M) and NaCl (0.9% w/v).

Since the concentration of Gel-FGE in all hydrogel formulations was held constant, two parameters influence the swelling capacity of these hydrogels, viz. (i) the extent of cross-linking, since higher cross-linked networks restricts the solution to permeate inside the hydrogel, and (ii) the amount of BMI used in each case, due to the presence of hydrophilic groups in the cross-linker chains [17]. Hence, HGEL1-1.0 was expected to swell the most, since lower cross-linking densities favor the swelling capacity of hydrogels. However, HGEL1-1.5 swelled the most and HGEL1-2.0 swelled more than HGEL1-1.0 in water and physiological solution. HGEL1-1.5 displayed a continuous porous microstructure and contained an intermediate amount of cross-linking agent. In this case, both factors influenced the swelling properties and the solvent could permeate better into the network. Even if more cross-linked networks were supposed to show lower swelling ratios, higher concentrations of the hydrophilic cross-linking agent could imbibe more water, thus increasing the swelling capacity of the hydrogels.¹³ That was the case of HGEL1-2.0, where the amount of BMI played a more important role in the swelling properties than the cross-linking density. The pH of the medium and the type of counterions present in the solution during the swelling process were also important factors in the case of the swelling properties of these hydrogels due to the chemical composition of the gelatin. In the case of the acidic medium, swelling ratios were found to be higher than in water. At low pH, gelatin is positively charged and electrostatic repulsions between the chains produced an increase in swelling.³⁸ Furthermore, the higher swelling ratios observed for the as-prepared hydrogels could be attributed to the presence of ether groups in the BMI chain susceptible to protonation, enhancing even more the repulsion between the polymeric chains. While incubated in physiological solution, hydrogels showed lower swelling ratios. This fact could be explained by the presence of counterions which could induce a screening effect, reducing the repulsion between polymer chains and, consequently, the swelling. The swelling properties of hydrogels are decisive for their application in the biomedical field, where their environmental response is relevant, especially for drug delivery systems and biosensors. In this work, the as-prepared hydrogels showed responsive properties, varying the swelling as a function of such external stimuli as pH or salt concentration.

The swelling kinetics of hydrogels was determined using an equation reported in the literature:³⁹

where SR is swelling ratio at time t, SR_max is the maximum swelling and k_s is the swelling rate constant. Table 3 summarizes the swelling properties at 37 °C of the different hydrogels. The swelling process in all cases followed second order kinetics with correlation coefficients exceeding 0.99. Under acidic pH, the SR_max value was considerably higher than the SR_eq, whereas in water and in physiological solution both values were quite similar. When using acidic pH as the swelling medium, k_s was low, since the amount of solution absorbed was high and the equilibrium swelling slower, making the swelling rate quite low.⁴⁰ It can be observed that the swelling media had greater influence on the swelling behavior of the hydrogels than the extent of cross-linking.

Table 3 Swelling parameters of gelatin-based hydrogels at 37 °C

	SReq (%)	SRmax (%)	ks × 10^4 (ghydrogel gsolution^−1 s^−1)
H2O
HGEL1-1.0	442	468	2.29
HGEL1-1.5	600	632	2.48
HGEL1-2.0	508	530	1.85
HCl (0.1 M)
HGEL1-1.0	583	811	0.20
HGEL1-1.5	696	1000	0.12
HGEL1-2.0	490	576	0.65
NaCl (0.9% wt)
HGEL1-1.0	365	380	5.46
HGEL1-1.5	394	403	3.62
HGEL1-2.0	429	450	3.47

---

Disassembly by retro-DA reaction

The retro-DA depolymerisation of the hydrogel was followed at 90 °C using the same technique as for the DA reaction (Fig. 12). The time evolution of the UV spectra mimicked the features shown in Fig. 8, with the corresponding progressive increase in the absorption near 300 nm as the retro-DA reaction proceeded. As reported by other authors, water seemed to accelerate the DA reaction while impeding the retro-DA, since the polymeric diene and dienophile, released during hydrogel disassembly, could react again with each other quickly.⁴¹ However, in light of the results obtained using UV-vis spectroscopy it has been proved that hydrogel depolymerisation can be induced using water as solvent at high temperatures.

Fig. 12 UV-vis spectra following the progress of the retro-DA depolymerisation of HGEL1-1.5 in H₂O at 90 °C.

Conclusions

Diels–Alder “click” chemistry is an efficient method for the preparation of biopolymer-based chemically cross-linked hydrogels from furan-modified gelatin and maleimide functionalized polyetheramines. Hydrogels were cross-linked following a clean one step procedure in absence of any catalysts or initiators. Different hydrogel compositions were obtained by varying the furan to maleimide ratio and the influence of the amount of cross-linker used in the final properties of the hydrogels was studied. The DA and retro-DA for hydrogel formation and depolymerisation, respectively, followed by UV-vis spectroscopy confirmed the success of both reactions under the studied conditions. The as-obtained hydrogels had a storage modulus similar to that of liver, fat, relaxed muscle and breast gland tissue and showed pH-responsive properties. These biobased cross-linked hydrogels have promising structural, mechanical and swelling properties in view of future biomedical applications.

Acknowledgements

Financial support from the Basque Country Government in the frame of Saiotek S-PE12UN036 and Grupos Consolidados (IT-776-13) and from the University of the Basque Country (UPV/EHU) in the frame of EHUA12/19 is gratefully acknowledged. C. García-Astrain wishes to acknowledge the Universidad del País Vasco/Euskal Herriko Unibertsitatea (Ayudas para la Formación de Personal Investigador) for its PhD grant PIFUPV10/034. Moreover, technical and human support provided by SGIker (UPV/EHU, MINECO, GV/EJ, ERDF and ESF) is also gratefully acknowledged.

References

1. B. Balakrishnan and R. Banerjee, Chem. Rev., 2011, 111, 4453–4474.
2. G. Tronci, A. T. Neffe, B. F. Pierce and A. Lendlein, J. Mater. Chem., 2010, 20, 8875–8884.
3. K. M. Park, Y. Lee, J. Y. Son, D. H. Oh, J. S. Lee and K. D. Park, Biomacromolecules, 2012, 13, 604–611.
4. E. Hoch, C. Schuh, T. Hirth, G. E. M. Tovar and K. Borchers, J. Mater. Sci.: Mater. Med., 2012, 23, 2607–2617.
5. D. M. Dragusin, S. Van Vlierberghe, P. Dubruel, M. Dierick, L. Van Hoorebeke, H. A. Declercq, M. M. Cornelissen and I. C. Stancu, Soft Matter, 2012, 8, 9589–9602.
6. A. Pourjavadi and M. Kurdtabar, Eur. Polym. J., 2007, 43, 877–899.
7. J. A. Deiber, M. L. Ottone, M. V. Piaggio and M. B. Peirotti, Polymer, 2009, 50, 6065–6075.
8. D. M. Kirchmajer, C. A. Watson, M. Ranson and M. in het Panhuis, RSC Adv., 2013, 3, 1073–1081.
9. H. C. Liang, W. H. Chang, H. F. Liang, N. H. Lee and H. W. Sung, J. Appl. Polym. Sci., 2004, 91, 4017–4026.
10. Y. Fu, K. Xu, X. Zheng, A. J. Giacomin, A. W. Mix and W. J. Kao, Biomaterials, 2012, 33, 48–58.
11. S. V. Vlierberghe, E. Schacht and P. Dubruel, Eur. Polym. J., 2011, 47, 1039–1047.
12. A. Gandini, Prog. Polym. Sci., 2013, 38, 1–29.
13. C. M. Nimmo, S. C. Owen and M. S. Shoichet, Biomacromolecules, 2011, 12, 824–830.
14. H. Tan, J. P. Rubin and K. G. Marra, Macromol. Rapid Commun., 2011, 32, 905–911.
15. C. García-Astrain, A. Gandini, D. Coelho, I. Mondragon, A. Retegi, A. Eceiza, M. A. Corcuera and N. Gabilondo, Eur. Polym. J., 2013, 49, 3998–4007.
16. H. L. Wei, K. Yao, H. J. Chu, Z. C. Li, J. Zhu, Y. M. Shen, Z. X. Zhao and Y. L. Feng, J. Mater. Sci., 2012, 47, 332–340.
17. F. Yu, X. Cao, L. Zeng, Q. Zhang and X. Chen, Carbohydr. Polym., 2013, 97, 188–195.
18. G. Mocanu, D. Mihaï, V. Dulong, L. Picton and D. Le Cerf, Carbohydr. Polym., 2012, 87, 1440–1446.
19. L. Shan, K. N. E. Verghese, C. G. Robertson and K. L. Reifsnider, J. Polym. Sci., Part B: Polym. Phys., 1999, 37, 2815–2819.
20. D. Luo, Y. Li and M. Yang, J. Appl. Polym. Sci., 2011, 120, 2979–2984.
21. M. Marref, N. Mignard, C. Jegat, M. Taha, M. Belbachir and R. Megahbar, Polym. Int., 2013, 62, 87–98.
22. E. Marie, K. Landfester and M. Antonietti, Biomacromolecules, 2002, 3, 475–481.
23. L. Y. Peddada, N. K. Harris, D. I. Devore and C. M. Roth, J. Controlled Release, 2009, 140, 134–140.
24. C. M. Ofner 3rd and W. A. Bubnis, Pharm. Res., 1996, 13, 1821–1827.
25. US Pat., 4 876 658, 1989.
26. S. H. Park, S. Y. Seo, H. J. Lee, H. N. Na, J. W. Lee, H. D. Woo and T. I. Son, Macromol. Res., 2012, 20, 842–846.
27. S. Farris, J. Song and Q. Huang, J. Agric. Food Chem., 2010, 58, 998–1003.
28. T. Il Son, M. Sakuragi, S. Takahashi, S. Obuse, J. Kang, M. Fujishiro, H. Matsushita, J. Gong, S. Shimizu, Y. Tajima, Y. Yoshida, K. Suzuki, T. Yamamoto, M. Nakamura and Y. Ito, Acta Biomater., 2010, 6, 4005–4010.
29. E. Mounif, V. Bellenger and A. Tcharkhtchi, J. Appl. Polym. Sci., 2008, 108, 2908–2916.
30. H. Liu, A. Uhlherr, R. J. Varley and M. K. Bannister, J. Polym. Sci., Part A: Polym. Chem., 2004, 42, 3143–3156.
31. C. Goussé and A. Gandini, Polym. Int., 1999, 48, 723–731.
32. C. Peña, K. de la Caba, A. Eceiza, R. Ruseckaite and I. Mondragon, Bioresour. Technol., 2010, 101, 6836–6842.
33. A. Gandini, D. Coelho and A. J. D. Silvestre, Eur. Polym. J., 2008, 44, 4029–4036.
34. N. J. Einerson, K. R. Stevens and W. J. Kao, Biomaterials, 2002, 24, 509–523.
35. R. Dash, M. Foston and A. J. Ragauskas, Carbohydr. Polym., 2013, 91, 638–645.
36. G. Palui, A. Garai, J. Nanda, A. K. Nandi and A. Banerjee, J. Phys. Chem. B, 2010, 114, 1249–1256.
37. J. L. Vanderhooft, M. Alcoutlabi, J. J. Magda and G. D. Prestwich, Macromol. Biosci., 2009, 9, 20–28.
38. M. Amiji, R. Tailor, M. K. Ly and J. Goreham, Drug Dev. Ind. Pharm., 1997, 23, 575–582.
39. A. Altinisik and K. Yurdakoc, Polym. Bull., 2014, 71, 759–774.
40. I. Katime, N. Valderruten and J. R. Quintana, Polym. Int., 2001, 50, 869–874.
41. H. L. Wei, J. Yang, H. J. Chu, Z. Yang, C. C. Ma and K. Yao, J. Appl. Polym. Sci., 2011, 120, 974–980.

---

Footnote

† This paper is dedicated in memoriam of Professor Iñaki Mondragon founder of “Materials + Technologies” Group (GMT) in 1988, who passed away recently after his contribution to this work.

---