Introduction

Chitosan, despite having strong biological activity, has limited biomedical applications due to its high molecular weight (MW) and lack of solubility in physiological pH conditions. It also forms highly viscous solutions because of its high MW, which makes it difficult to process in pharmaceutical or biomedical formulations. A way to circumvent these issues is to depolymerize chitosan to smaller chain chitooligosaccharides (COS), which have much lower MW that imparts greater solubility and higher biological activity. Usually, COS tend to have MW < 10 kDa and DP < 50–55 (Muanprasat and Chatsudthipong 2017). Some authors consider that chitosan oligomers should only have DP < 10 (Naqvi and Moerschbacher 2017), while others include MW up to 30 kDa as the upper limit (Liaqat and Eltem 2018). This is because true chitosan polymers are considered to have a DP > 100, which is roughly greater than 100 kDa (Naqvi and Moerschbacher 2017). Regardless of the demarcation, COS is characterized by a greater degree of solubility than its precursor, which extends its applicability. COS in place of chitosan has shown stronger anti-inflammatory, anti-tumor and anti-oxidant behavior (Chae et al. 2005; Harish Prashanth and Tharanathan 2007; Muanprasat and Chatsudthipong 2017; Tomida et al. 2009). Complete solubility across a wide range of pH is observed when MW is reduced to below 10 kDa (Mourya, Inamdar, and Choudhari 2011).

Similar to chitosan, COS exhibits various bioactive properties which include antihypertensive, anti-infectious, anti-tumor effects, etc. However, differences in DD, DP, or MW may cause differences in its application (Zou et al. 2016). In its antioxidant use, COS can be used to break up the oxidative sequence to prevent cardiovascular diseases. In fact, lower the molecular weight, stronger the anti-oxidative response against an array of radicals such as superoxide, –OH, 2,2-diphenyl-1-picrylhydrazyl (DPPH), etc., with MW < 5 kDa being more promising (Mendis et al. 2007; Mengíbar et al. 2013; Park et al. 2003). Additionally, COS having MW < 3 kDa can suppress the release of cytokines such as inducible nitric oxide synthase (iNOS), tumor necrosis factor (TNF)-α, interleukin (IL)-1β, IL-4 and IL-6 genes that have pro-inflammatory characteristics (Vo et al. 2011; Wu et al. 2015). Also, COS acts as an alleviator to H2O2 -induced oxidative stress injury in ECV304 cells (Liu et al. 2009). COS exhibits antimicrobial properties although the extent of this activity varies depending upon the type of chitosan or pH value. It contains primary amino groups which are thought to be the defining factor in its enhancement of antimicrobial properties compared to chitosan (Chen et al. 2002). COS of MW < 2 kDa demonstrated powerful anti-tumor activity against leukemia cells (Kim et al. 2013). COS also reduced the growth of human gastric and breast cancer cells in vitro, by promoting apoptosis and downregulating the expression of growth factors (Jiang et al. 2021; Mallakuntla et al. 2021).

Chitosan and COS are often used in the development of biomimetic materials, such as hydrogels, which are three-dimensional, highly porous networks applicable as artificial tissue scaffolds, wound dressings, drug delivery depots or cell encapsulating systems (Slaughter et al. 2009). Generally, the more hydrophilic the polymers that make up the hydrogel, the greater the swelling under aqueous conditions. This results in better solute transport (Peppas and Reinhart 1983). In fact, a polymer that is water soluble before being cross-linked can be more suitable as a hydrogel base (Ahmed 2015), which makes COS a better alternative to chitosan. For a high percentage of swelling, preparation in an aqueous solution is useful, which also ensures no adverse reactions occur in animal cells (Ahmed 2015; Shorrock and Rees 1988). Additionally, hydrogel formulations can be developed under neutral pH if the precursor material dissolves under such conditions. This can be an important consideration when certain therapeutic proteins such as fibroblast growth factor 20 (FGF20) are to be incorporated into the hydrogel carriers because highly acidic or alkaline pH can denature the protein (Fan et al. 2007).

COS is certainly more useful in biomedical applications, especially in hydrogels. Unfortunately, chitosan depolymerization techniques to produce COS, although many in number, are not convenient. Physical methods involving gamma, UV or microwave radiation, on their own have not been able to sufficiently break down chitosan to soluble COS (Sato et al. 2003). In most cases, drastic conditions such as high temperatures (up to 90oC (Wu et al. 2015) and pressures (Gao et al. 2004; Savitri et al. 2015), or high-intensity doses (Choi et al. 2002; Jin et al. 2017; Muley et al. 2019b), had to be employed. In addition, most studies discovered a limit to the levels up to which degradation could be achieved. For example, both Muley et al. and Hien et al. reported that gamma radiation at 100 kGy could only achieve a MW reduction of 73–75%, insufficient to achieve the MW of COS. The difficulty in degradation beyond 100 kGy was attributed to the crystalline portions of chitosan (Hai et al. 2003; Muley et al. 2019b). Chemical methods involve either acidic or oxidative degradation. Acidic hydrolysis produces by-products that need to be separated by either fractional precipitation, ultrafiltration or ion exchange (Lin and Lin 2003; Lopatin et al. 2009). Most studies utilized active reactants’ concentrations at values typically above 6 M (Cabrera and Van Cutsem 2005; Kabal’Nova et al. 2001; Kazami et al. 2015; Trombotto et al. 2008), and such concentrations can drastically alter the structure of the units and produce toxic by-products (Uchida et al. 1989). Hydrogen peroxide has emerged as an effective depolymerization agent, used mostly in conjunction with radiation. Using H2O2, one of the first studies found that chitosan MW decreased by over 95% in the presence of 1.5% H2O2 when the reaction temperature was 50oC and duration was 3 h. When the concentration was raised to 3.5%, the same reaction took only 150 min and DP 6–7 COS had the highest yield (Chang et al. 2001). However, peroxide on its own is not as effective, and achieved only limited MW reduction. Najafabadi et al. determined that a reduction of only 22% was possible, even when the peroxide concentration was only 2% (Najafabadi et al. 2018). Hai et al. determined that when 1% H2O2 was used to depolymerize chitosan at ambient temperature, COS was produced after 360 h of reaction. Thereafter the reaction stabilized, and the MW did not decrease appreciably as the reaction time continued until 720 h. The MW decreased from 91  to 4.5 kDa (Hai et al. 2019). The major limitation of the enzymatic approach is the cost, as the enzymes are expensive. Also, large batch and column reactors containing immobilized enzymes did not perform satisfactorily in producing enough COS. Downstream purification and separation of the enzymes and products is also expensive. The reaction rate for immobilized enzymes is also lower than that of free enzymes (Kim and Rajapakse 2005).

Most studies have instead focused on simultaneous radiation and oxidative degradation techniques (known as in situ techniques), which have yielded better results as the degradation was faster (Abd El-Rehim et al. 2015; Duy et al. 2011; Hien et al. 2012; Kang et al. 2007; Nguyen et al. 2017; Xuan Du and Xuan Vuong 2019). With such techniques, however it is cumbersome to irradiate large vessels of aqueous samples simultaneously. It is also beneficial to irradiate chitosan in solid form as the number of radicals observed was larger than when the irradiation was carried out in solution (Yoksan et al. 2004). There is a need to develop suitable depolymerization techniques that produce COS conveniently under mild conditions and have fewer processing steps. In this work, we employed a convenient two-stage physicochemical depolymerization process, where gamma irradiation was followed by oxidative degradation with hydrogen peroxide, to produce COS that can dissolve in physiological pH levels. This study therefore focused on finding a strategy to depolymerize chitosan without resorting to simultaneous radiation and oxidative degradation with hydrogen peroxide, which is a cumbersome strategy. This two-stage depolymerization process had not been analyzed in prior studies and we sought to determine a set of optimum conditions for this process. Furthermore, our goal was to employ processing conditions that did not involve high radiation doses, hydrogen peroxide concentrations, and temperatures. As is common with more recent studies, after producing COS, it is necessary to evaluate its effectiveness. Therefore, in this study, a biocompatible hydrogel was successfully prepared under neutral pH to test the feasibility of the COS in biomedical applications. The main outcome of this investigation is thus to establish a COS production technique that can be carried out under mild operating conditions, has scalability potential, and requires minimal processing steps to reduce the bottlenecks in obtaining biomedically relevant COS. Chitosan can thus be depolymerized by avoiding cumbersome in-situ physicochemical processes. The COS produced can then be used to develop biocompatible hydrogels.

Materials and methods

Materials

Chitosan from crab shell chitin was generously supplied by the Biomaterials and Bioengineering Research team from Center of Bioengineering and NanoMedicine (University of Otago, Dunedin, New Zealand). H2O2 (reagent grade, 30%, Merck), lactic acid (reagent grade, 99%, Sigma-Aldrich) and distilled water was used for the chemical depolymerization reaction. NaCl (reagent grade, Merck) was used for the MW tests. Polyethylene glycol diacrylate (PEGDA 700, Merck), Irgacure-2959 and sodium carboxymethyl cellulose (Na-CMC) were used for the hydrogel preparation. A UV radiation chamber (UVB and UVC, average intensity 10 W/m2) was used for photopolymerization reactions. Glacial acetic acid (99.99%,), Na2HPO4, KCl and KH2PO4 and dimethylsulfoxide (DMSO) were of reagent grade.

Preparation of irradiated chitosan

Irradiation of chitosan (B0) was carried out using Co-60 γ-irradiator (present activity 38.6 kCi) at 5, 15, 25, 50, 75 and 100 kGy dose levels. The average measured dose rate was 13.29 kGy/h. Approximately 8 g of chitosan was packaged in a low-density transparent polyethylene bag (3.8 cm by 6.4 cm) and three such packets were then further encased inside a high-density transparent PVC bag for proper exposure to the radiation at room temperature. The samples were placed roughly equidistant from each other, approximately halfway across the PVC bag (Fig. S1) to ensure radiation exposure was the same for all the samples. During irradiation, the bags were placed directly in line with the source and irradiated at room temperature. The irradiated samples were coded B5, B15, B25, B50, B75 and B100, where the subscript denotes the dose applied. The dose levels were proportionate with the time of exposure, with 5 kGy taking 42 min, 15 kGy taking 68 min, 25 kGy taking 113 min, 50 kGy taking 226 min, 75 kGy taking 339 min and 100 kGy taking 451 min.

Depolymerization with hydrogen peroxide

5% (w/v) solutions of the irradiated chitosan were prepared in a solvent containing 1.5% (w/v) of lactic acid and 0–2% (w/v) of hydrogen peroxide (H2O2). The reaction vessel was wrapped with aluminum foil to prevent exposure to light. Three different temperatures were tested – 10 ± 2oC, 25 ± 2oC and 35 ± 2oC. 4 h and overnight reaction time periods were chosen. Table 1 shows the experimental parameters that were applied (not in the order in which experiments were performed).

Table 1 Conditions tested in the depolymerization reactions
Full size table

After the desired time had elapsed, the reaction was quenched with 0.5 M NaOH solution, and the pH was adjusted to approximately 7. Ethanol (96% C2H5OH) was used to precipitate the COS produced in the reaction. Centrifugation at 4000 rpm for 15–30 min completely separated the precipitate from the solution, and the precipitate was first heated in 90oC for excess ethanol to boil off and then subsequently dried overnight in an oven at 35oC. The dried COS samples were powdered in the presence of ethanol and then after air-drying for a few minutes, were stored in dry containers for the characterization studies.

COS characterization

At different pH levels, the solubility of the COS samples was analyzed by measuring the transmittance at 600 nm using a UV-Visible spectrophotometer (Hach Spectrophotometer DR 3900). The spectrophotometer was set to transmittance mode at a single wavelength. An optical glass cuvette having a path length of 1 inch was used. A 0.1% (w/v) solution of the COS product obtained from each of the depolymerization reaction mixture was prepared in 1.5% (w/v) lactic acid. The pH was adjusted using 0.5 M NaOH and 1 M lactic acid from 2 to 11. At first, the transmittance of the solvent was measured and set as the baseline. The pH was adjusted at increments of 1 unit and the corresponding transmittance of the COS solution was then measured.

The viscosity average molecular weight (Mv) was determined using the Mark-Houwink-Sakurada relationship \([\eta]=K{M}_{v}^{\alpha}\) (1). For chitosan, Wang et al. reported that the values of K and α varied with DD. In the same study, the authors developed a correlation between K and DD (K = 1.64 × 10−30DD14.0 cm3.g−1), as well as for α (α= −1.02 × 10−2DD + 1.82), for a temperature of 30oC (Wang et al. 1991). Using a capillary tube viscometer with a constant temperature water bath (30.0 ± 0.5oC), the viscosity of chitosan, irradiated chitosan and COS samples were determined according to the standard for testing chitosan (ASTM F2103-01) (Zainol, Akil, and Mastor 2009). The capillary had a diameter of 1.2 mm (except in two cases, where 0.8 and 1.5 mm were used). Solutions were prepared using a 0.1 M lactic acid and 0.2 M sodium chloride solvent system (García et al. 2015). Starting with a concentration of 9.6 × 10−3 g/mL, the average fall times for the solution and four dilutions (1.25x, 1.5x, 5x and 10x) were measured. Three triplicate measurements were conducted for each polymer-solvent system. Mv was associated with the number average molecular weight (Mn) using Eq. 1 (Rapado Paneque et al. 2004).

$$M_{n} = M_{v} \left[ {\left( {\alpha + 1} \right)\sqrt {\left( {\alpha + 1} \right)} } \right]^{{ - \frac{1}{\alpha }}}$$
(1)

For chitosan and COS, a small amount of the powdered samples were pressed against the sample holder and the FTIR (Fourier Transform IR) spectra were recorded using a diamond-tip ATR tool over wavenumbers ranging from 4000 to 400 cm− 1 at a resolution of 4 cm− 1 and for a total of 40 scans. The spectra were analyzed to identify key chemical groups. In addition, the DD was determined using the Eq. 3.6 (Brugnerotto et al. 2001). IR spectroscopy is a widely used technique for DD determination (Dong et al. 2001), and is usually determined using the absorption ratios of two bands (AR/AM), where AM is the intensity of a band characteristic to N-acetylation and AR refers to the intensity of a band that is independent of DD (e.g., –OH, C–, –CH2 or C–O–C bands) using Eq. 2; A1320 and A1420 represent the absorbances at 1320 (for –CH2) and 1420 cm− 1 (for amide group), respectively. These two peaks were chosen as they do not vary with changes in humidity.

$$DD\left(\%\right)=100-\frac{\frac{{A}_{1320}}{{A}_{1420}}-0.03822}{0.03133}$$
(2)

XRD analysis was conducted on the Rigaku SmartLab SE, which consisted of a Cu-Kα source (with a nickel filter) and was operated at 40 kV and 30 mA. The wavelength was 1.5 Å, the sweep rate was 1.2 o/min and the scattering angle (2θ) range was 5-60o. From the XRD spectrum, using Bragg’s law, the d-spacings for the crystallites in the samples was evaluated, where n is the order of reflection (which is 1), \(\lambda\) is the X-ray wavelength used and θ is the angle of reflection (d = nλ/(2 sinθ). The crystallinity index was also determined using CI (%)=(I20-I16)/I20, where I16 represents the intensity at 2θ = 16o and I20 is the intensity at 2θ = 20o (Chesnutt et al. 2009).

DSC and TGA analyses of native chitosan (B0), sample irradiated at 15 kGy (B15) and the optimum COS product (C) were carried out using a Netzsch STA 449F3 instrument under a nitrogen atmosphere (purge flow rate of 20 mL/min). The heating rate was 30oC/min over the range 30-600oC. The DSC-TGA experiments were carried out using 29.9 mg of B0, 26.3 mg of B15 and 15.8 mg of C and the samples were placed in Al2O3 pans.

B0, B15 and C were mounted on separate SEM stubs and coated with a 10 nm layer of gold using a mini sputter coater by Quorum Testing. The images were taken at a voltage of 10.0 kV in the EVO 18 SEM machine at two different magnifications. Using SEM images, the surface morphology and texture were visualized.

Hydrogel preparation

To prepare the hydrogel, PEGDA was chosen to form cross-links with COS. It is commonly used as an effective synthetic polymer in biomedical hydrogels, as it is hydrophilic, biocompatible, non-immunogenic, and does not adsorb proteins that could cause fouling or pore blockages in the hydrogel (Slaughter et al. 2009). In fact, PEGDA hydrogels have been widely used in tissue engineering applications as it does not induce any unwanted reactions in cells (Lin et al. 2020; Morris et al. 2017). The small chains of COS however cannot be efficiently cross-linked with PEGDA, and the COS-PEGDA hydrogel solution was found to have a watery consistency. To increase the viscosity, CMC was used as a thickener, which forms a poly-ionic complex with COS, holding the chains in place for effective binding with PEGDA. CMC itself has high binding and superabsorbent characteristics (Capanema et al. 2018; Fekete et al. 2017).

0.5 g of the COS product was dissolved in 100 mL of 1% (v/v) acetic acid solution and the pH was adjusted to ~ 7. After overnight stirring, this neutral stock solution was used for the purposes of preparing the COS hydrogel. To 1 mL of this solution, 5 mL of a copolymer solution containing 7.5% (w/v) PEGDA and varying CMC concentration (2 and 4% (w/v) was added. Irgacure 2959, a photoinitiator, was dissolved to make a 1% (w/v) solution in DMSO; from this stock solution, 137 µL was used in the hydrogel formulation such that photocuring required a concentration of 0.02% (w/v) of Irgacure 2959. These low concentrations ensure that the photo-initiator does not induce any toxic effects (Morris et al. 2017). The hydrogel blends were coded COSPCMC-2 and COSPCMC-4 to denote the CMC concentration. Curing was performed under a UVB-UVC combination light source (100–480 nm, 10 W). The curing time (~ 2 h) was chosen so that both types of hydrogels would be cured simultaneously but the exposure would not be too high. Cylindrical hydrogel specimens of 1 cm by 1 cm were prepared for the characterization studies.

Hydrogel characterization studies

Swelling characteristics

Cylindrically shaped gels were freeze dried and weighed (mo). The hydrogel samples were then incubated in a PBS solution (pH ~ 7) for 24 h at 37oC. After incubation, the samples were reweighed (m1) after blotting off the excess surface water using a filter paper. The equilibrium swelling ratio is given by SR = (m1-mo)/m1.

Morphological analysis

Hydrogel samples were cut into thin slices (5 mm thickness) before freeze drying. The gel cross-sections were mounted on SEM stubs and sputter coated with a 10 nm layer of gold for 90 s. Images were obtained at a voltage of 5 kV. Pore structure and surface properties were analyzed using an EVO18 SEM (Carl Zeiss).

Functional group analysis

To assess the cross-linking performance effectively, it is necessary to determine the type of functional groups present in the hydrogels. As such, the IR spectra of the hydrogels were recorded. Freeze-dried samples of the gel were cut into small pieces and then analyzed using a diamond-tip ATR tool over wavenumbers ranging from 4000 to 400 cm− 1 at a resolution of 4 cm− 1 and for a total of 64 scans.

Cytotoxicity tests

Although washing and dialysis remove unreacted chemicals from the COS hydrogel, if it is designed to degrade inside the body (such as drug delivery depots or artificial scaffolds that should eventually be replaced by new tissues generated in the body), the chemicals would enter the system. As a result, a cell assay study principally measures the cytotoxicity of any leach-out products in addition to the gel itself is necessary. At first, samples of hydrogel samples of various masses (25 mg, 50 mg and 100 mg) were incubated at 37oC in 10 mLs of PBS solution (pH 7) for the compounds to leach out. The solutions were then filtered through a 0.22-micron syringe filter to sterilize the solutions. To conduct the cytotoxicity analysis, a non-cancer mammalian Vero cell line (kidney epithelial cell lines) extracted from the African green monkey, was incubated in DMEM (Dulbecco’s Modified Eagles’ medium) containing 1% penicillin-streptomycin (1:1), 0.2% gentamycin and 10% fetal bovine serum (FBS). Cells (3.0 × 104/200 µL) were seeded into a 48-well plate and incubated at 37oC and 5% CO2. After 24 h, 50 µL sample (sterilized using a 0.22-micron syringe filter) was added to each well. Cytotoxicity was examined under an inverted light microscope after 48 h of incubation. The results were based on n = 2 duplicate samples.

Mechanical tests

A universal testing machine (UTM) by TestResources (100 series) was used to conduct uniaxial compression tests at a displacement rate of 10.00 mm/min on swollen samples measuring an average of Ø 9 mm by 5 mm (n = 3 samples). Force applied and position were recorded, which were used to evaluate the compressive stress (\(\sigma =F/A)\) applied and the elastic modulus determined using \(\sigma =E(\lambda -{\lambda }^{-2})\), where σ is the applied stress in Pa/m2, F is the measured force, A is the cross-section area and λ is the relative deformation of the sample (\(\lambda =1+/L\)), and ε is the deformation and L is the thickness of the sample.

Results and discussion

Optimization of COS production technique

By measuring the UV transmittance of the solutions of chitosan and COS obtained after radiative and oxidative degradation, the extent of solubility at various pH conditions was determined. Chitosan, and by extension the derivatives, exhibit solubility at pH < 5. This is expected, since the primary -NH2 groups present in the molecule are protonated at such conditions. The defining criteria of solubility at any pH was a transmittance of > 90%, which signaled that no precipitation had occurred, and a clear solution persisted (Zainol et al. 2009). Figure 1 summarizes the optimization studies that were conducted to determine the most optimum temperature, reaction time, hydrogen peroxide concentration and radiation dose that could yield COS of the highest solubility at pH 7–8, which is the physiological pH range. All peroxide reaction tests were conducted on chitosan samples that had been first irradiated at 100 kGy. It is to be noted that solubility tests are conducted in controlled pH conditions instead of pure water as ionic strength of the solution appeared to be a factor governing solubility (Nguyen et al. 2017).

The most optimum combination of peroxide reaction conditions were 25oC, overnight reaction with 2% H2O2. The minimum radiation dose that would be required was then determined by performing oxidative degradation at these optimized conditions for chitosan that had been irradiated at various gamma radiation doses. Again, the transmittance was measured at a range of pH values (Fig. 1d). Co (chitosan without irradiation but degraded with H2O2) precipitated out at pH > 6 as demonstrated by the drop in solution transmittance to 84%, while C1 (5 kGy irradiation) remained soluble up to pH 7.6. C2 to C4 (15–100 kGy dose) exhibited transmittance of > 90% beyond pH 7 and up to pH 10, as no precipitation was observed. This is likely due to the fact that as the radiation dose was increased, the chitosan depolymerization progressed further (Hai et al. 2003). Thus, for C2–C4, the irradiated chitosan was sufficiently depolymerized, degrading to soluble COS after oxidative treatment with hydrogen peroxide. The radiation plays a significant role in aiding the oxidative degradation; studies by Qin et al. and Kabal’nova et al. showed that such dramatic depolymerization required either temperatures above 50oC or hydrogen peroxide concentrations above 0.6 M.

Fig. 1
figure 1

Transmittance of COS solutions at various pH values showing extent of solubility. Optimization of reaction conditions; a impact of hydrogen peroxide concentration (other conditions: 100 kGy irradiation, H2O2 reaction conditions (25oC, overnight); b the impact of temperature (other conditions: 100 kGy irradiation, H2O2 reaction conditions (2% H2O2, overnight), c impact of reaction time (other conditions: 100 kGy irradiation, H2O2 reaction conditions (2% H2O2, 25oC), and d impact of radiation dose (H2O2 reaction conditions (2% H2O2, 25oC, overnight reaction))

Full size image

The mechanism of both radiation breakdown and oxidative degradation have been summarized by other authors. The succeeding mechanism for the gamma radiation breakdown has been proposed by Hai and coauthors (Hai et al. 2003). Reaction (i) occurs as the energy of the gamma rays is sufficient to break down a C-H bond in chitosan. The H free radical attacks chitosan chain in the chain propagation reaction (ii), and a free radical is produced on any of the C atoms of the glucopyranose ring. If the free radical center is on C1 or C4, fragmentation of the chains takes place. The glucosamine groups can also participate and break down into ammonia. In the mechanisms below, R–H and R–NH2 are chitosan chains, R•(Cn) is a chitosan radical, where the unpaired electron is on the Cn atom and F1 and F2 are fragments formed after chain scission.

$${\text{R}} - {\text{ H }} \to {\text{R}} \bullet \left( {{\text{C}}_{4} - {\text{C}}_{6} } \right) + {\text{H}} \bullet$$
(i)
$${\text{R}} - {\text{ H }} \to {\text{H}} \bullet \to {\text{R}} \bullet \left( {{\text{C}}_{1} - {\text{C}}_{6} } \right) + {\text{H}}_{2}$$
(ii)
$${\text{R}} \bullet \left( {{\text{C}}_{1} ,{\text{C}}_{4} } \right) \to {\text{F}}_{1} \bullet + {\text{F}}_{2} \left( {{\text{fragmentation}}} \right)$$
(iii)
$${\text{R}} \bullet \left( {{\text{C}}_{1} ,{\text{C}}_{4} } \right) \to {\text{F}}_{1} \bullet + {\text{F}}_{2}$$
(iv)
$${\text{R}} - {\text{NH}}_{2} + {\text{H}} \bullet \to {\text{R}} \bullet \left( {{\text{C}}_{2} } \right) + {\text{NH}}_{3}$$
(v)

During oxidative degradation with H2O2, the acidic pH of the solution in which the reaction is carried out, causes the unstable H2O2 to form H + and HOO- ions that eventually decompose to HO•, O2−• and H2O. The HO• attacks the chitosan chain, forming free radicals as shown in the previous mechanism (reaction (ii)), except free radical centers on C1, C4 and C5 can lead to fragmentation and hydrolysis reactions, both of which lead to chain breakage.

$${\text H}{}_{2}{{\text {O}}}_{2}\leftrightarrow {{\text {H}}}^{+}+{\text {HO}}{{\text O}}^{-}$$
(vi)
$${\text {HO}}{{\text {O}}}^{-}\to {\text {O}}{{\text {H}}}^{-}+{\text {O}}\bullet$$
(vii)
$${\text{H}}_{2}{{\text {O}}}_{2}+ P{\text {HO}}{{\text {O}}}^{-}\to {\text {HO}}\bullet +{{{\text {O}}}^{\bullet }}_{2-}+{{\text {H}}}_{2}{\text {O}}$$
(viii)
$${\text {R-H}}+{\text {HO}} \to {\text {R}}\left({\text {C}}{}_{1}-{\text {C}}{}_{6}\right)+{{\text {H}}}_{2}{\text {O}}$$
(ix)
$${\text {R}}\left({{\text {C}}}_{1},{\text {C}}{}_{4},{{\text {C}}}_{5}\right)\to {{\text {F}}}_{1}+{{\text {F}}}_{2} \left({\text {fragmentation}}\right)$$
(x)

The summary of the optimization studies is given in S1, where the average transmittance values at physiological pH levels 7–8 have been highlighted to show extent of COS solubility at those pH values.

Molecular weight

Prior studies have shown that increase in gamma radiation dose reduces molecular weights as the chain breakdown occurs more rapidly and to a greater extent (Hai et al. 2003; Yoksan et al. 2004; Zainol et al. 2009). Figure 2 shows that in this study, a similar trend was observed. Irradiation at 15 kGy reduced the MW only slightly (by only 3%). From 15 to 50 kGy, the decrease was rapid. The average molecular weight after irradiation at 75 kGy was 86.8 kDa, which decreased to 79.6 kDa for 100 kGy, which is only an 8% change. This suggests that increasing the dose rate any further would not cause sufficient breakdown of chitosan. An explanation for this breakdown limit was proposed by Yoksan et al.; irradiation only resulted in breakdown of the amorphous regions, while the crystalline portion remained intact. This implies that higher radiation doses cannot produce chitooligosaccharides (COS) of molecular weights < 30 kDa, since the crystalline structure cannot be degraded by radical chain scission reactions (Yoksan et al. 2004). Zainol et al. also confirmed that the MW of chitosan having an initial MW of 576 kDa decreased to 458 kDa at 10 kGy, 242 kDa at 25 kGy, 159 kDa at 50 kGy and finally to 106 kDa at 100 kGy. The correlation graphs are given in supplementary materials (S2).

Fig. 2
figure 2

Effect of radiation on the molecular weight of chitosan (irradiation occurred at room temperature in the presence of air)

Full size image

Similar viscosimetric measurements were carried out for the COS produced after irradiation at 15 kGy and degraded overnight using 2% hydrogen peroxide at 25oC (product C) and an average molecular weight of 12.8 kDa was obtained, a reduction of 95% from the molecular weight of the irradiated precursor (Table 2). Although the most significant molecular weight reduction occurred during the peroxide treatment, these changes required mild conditions. This would not have been possible without prior radiation treatment; Hai et al. treated chitosan having a molecular wight of 36.5 kDa with 2% hydrogen peroxide and determined that although a 93% reduction in molecular weight was possible, it took 720 h (Hai et al. 2019). In another study, a 90 to 70 kDa reduction in molecular weight required 12 h when only 1% H2O2 was used (Hien et al. 2012). It can be speculated that the initial radiation treatment sufficiently degraded the amorphous regions and afterwards the H2O2 could act more efficiently. Based on the results of solubility and MW, this COS product (C) was chosen to be the most suitable for subsequent characterization and application studies.

FTIR analysis

Table 2 shows the variation in DD after irradiation at different doses and of the optimized product. Because the values of DD are comparable, it can be concluded that neither radiation nor H2O2 increased the number of carbonyl groups (C = O), which would have increased DD. Therefore, the functional groups remain intact in these processes. Although previous studies have also concluded that irradiation has no impact on DD (García et al. 2015; Nguyen et al. 2017; Zainol et al. 2009), H2O2 has been observed to increase the DD to a much higher degree, as it modified the chemical structure (Nguyen et al. 2017; Qin et al. 2002). Our results however demonstrate that H2O2 did not have such a drastic impact (DD of C is 62.3%, whereas for B0, it is 65.5). This indicates that only the polysaccharide backbone of the chitosan chain participates in the depolymerization reactions, while the functional groups remain intact (García et al. 2015; Zainol, Akil, and Mastor 2009). Yoksan et al. proposed that the DD can remain unaltered if •COCH3 radicals are not formed during scission. The results exhibited in our experiments are in agreement with Yoksan et al.’s findings [48].

The FTIR spectra of B0, B15 and C show almost similar characteristic peaks (Fig. 3a). All spectra show the broad peak for −OH at 3320 cm− 1, the narrow band for C=O (amide I) at 1630–1640 cm− 1, the characteristic peak for −NH2 at 1570–1590 cm− 1. The peak at 1029–1069 cm− 1 arise due to the stretching vibration of C−O in the C−O−C bond, while that at 1148 cm− 1 is associated with C–O–H bond stretching (Bernardino-Nicanor et al. 2017; Muley et al. 2019). The peak at 885–890 cm− 1 corresponds to the glucopyranose ring (Yoksan et al. 2004). In other words, the backbone structure did not change significantly. All peaks have been normalized to the same baseline after automatic baseline correction. However, peak analysis revealed slight differences. According to the peak absorption ratio comparison method outlined by Zainol et al. (S3), the ratio between the C=O and –OH absorption peaks differed between the three types of samples. Referring to Fig. S3, the e-f peak height is a measure of the C=O bond and b-c measures the –OH group. The ratio for B0 was determined to be 0.504, whereas for B15 and C, it increased to 0.635 and 0.782 respectively. This is because B0 had more –OH groups than the others (Muley et al. 2009), while more C=O bonds formed after the chain scission reactions and oxidative degradation of chitosan, which resulted in a slight decrease in DD as shown on Table 2 (Zainol et al. 2009).

XRD analysis

Values of the crystallinity index describe whether the irradiation and oxidative treatment had any impact on the crystalline regions of the samples. The CI of chitosan and the irradiated samples are similar (Table 2), which indicates that irradiation could not sufficiently induce radical chain scission in the crystalline regions; the depolymerization only occurred in the amorphous zones. However, treatment with hydrogen peroxide led to a more drastic reduction, as H2O2 greatly accelerates radical formation.

Table 2 Average MW (± standard deviation, n = 3 samples), DD, CI and d-spacings of chitosan, irradiated chitosan and COS
Full size table

Samuels identified two forms of chitosan. Both are extended 2-fold helix structures and form I has a distinctly sharp characteristic peak at 2θ = 11o, while the one for form II, mostly appears at 2θ = 20.5o (Samuels 1981). Both B0 and B15 show the first peak respectively at 11.4o and 11.9o, and the second one at 20.4o and 20.7o (Fig. 3b), with the form II peak being more prominent. This indicates that the radiation scission had minimal impact on the crystalline portion, which also explains why the crystallinity index decreased only slightly (Table 2 ) (Muley et al. 2019). In form II crystals, the main chain is constrained due to the H bonds, which maintains the packing structure of chitosan (Fan et al. 2009; Yoksan et al. 2004). After the oxidative treatment, a dramatic decrease in crystallinity occurred (C). The peak at 11o disappeared, while the form II peak diminished in intensity and shifted to 19.8o. This behavior is consistent with the XRD patterns of tendon hydrated water-soluble COS found in other studies (Fan et al. 2009; Hai et al. 2019). The decrease in the angle reflects the expansion of the crystal lattice due to the formation of a greater number of acetyl groups after depolymerization (characterized by the lower DD), which is also confirmed by the larger d-spacing (Table 2 ). Chitosan depolymerization first proceeds in the amorphous region that covers its surface. Once the surface has been sufficiently degraded, the crystalline regions can then be attacked. It is possible that irradiation first depolymerized only the amorphous regions without much impact on the crystal structure. Afterwards, the H2O2 could affect the crystalline regions (F. Tian et al. 2003). D-spacings of the crystalline forms found in the B0, B15 and C are summarized in Table 2.

Fig. 3
figure 3

a Comparison between the FTIR spectra of B0, B15 and C, highlighting only the major characteristic peaks; b XRD spectra of B0, B15 and C (a.u.: arbitrary units), showing that the peak at 11.4–11.9o has disappeared for C

Full size image

DSC-TGA analyses

Figure 4a shows the TGA behavior of B0, B15 and C, alongside the derivative of the mass loss. All three exhibit two distinct decomposition events. The first event, corresponding to the loss of hydrogen bonded water molecules, occurred at 53.7oC, 49.6oC and 38.0oC respectively. For B0, the mass loss was 8.6%, B15 and C exhibited a loss of 3.1% and 4.3%. The second decomposition was observed at 233.6oC, 234.6oC and 163oC respectively, corresponding to a 31.6%, 43.8% and 45.1% decrease, and occurred when the saccharide rings of the molecules underwent complete degradation (García et al. 2015). The progressive decrease in the decomposition temperatures is due to the decrease in the thermal stability that occurred after the depolymerization reactions (and the decrease in crystallinity) (Timur and Paşa 2018). It is evident from the dramatic differences shown by C that the depolymerization was more severe after irradiation and oxidative degradation by peroxide (Qin et al. 2002). The TGA derivative curve in Fig. 4b correspond to the derivatives of the TGA data in Fig. 4a and confirm the temperatures at which these major mass changes occurred.

The DSC thermograms (Fig. 4c) of the three molecules are similar to the characteristic thermograms of chitosan (García et al. 2015; Kittur et al. 2002; Muley et al. 2019a). The first broad endothermic peaks occurred due to moisture loss, occurring at 109oC for both B0 and B15, but at 113oC for C. Depending on the structure of the polysaccharide, the water retention capacity changes; the increase in the first endothermic peak’s area is associated with the ability to bind more water. Due to depolymerization, the structure became more disordered, which has been correlated with the increase in water retention capacity (Kittur et al. 2002). Decomposition events (breakdown of the amine groups) are depicted by the prominent exothermic peaks, which are observed at around 299oC for B0 and B15 and at 303oC for C. Additionally, the peak height for C is 1.29 mW/mg, an almost 50% reduction than those for B0 (2.53 mW/g) and B15 (2.23 mW/g), indicating that C decomposed more easily.

Fig. 4
figure 4

a Thermal degradation behavior of B0, B15 and C; b Evaluated values of derivative weights from the TGA data showing regions of major weight loss; c DSC thermograms showing phase changes in B0, B15 and C

Full size image

Morphological characteristics

The SEM analyses (Fig. 5) confirm that major structural changes arose due to the depolymerization. B0 and B15 had a more amorphous structure, characterized by a nonporous texture. Noticeable cracks and fractures appeared on C, including micropores. The fact that B0 and B15 were both viewed on a scale of 10 micron, whereas C could only be resolved at a scale of 100 micron, indicated that particles of C were much smaller than those of B0 and B15.

Fig. 5
figure 5

SEM images showing surface morphology of B0, B15 and C (first row of images at 5 kX, 10 kX and 1 kX magnifications respectively; second magnification at 30 kX, 20 kX and 20 kX respectively)

Full size image

Synthesis of COS-PEGDA-CMC hydrogel

In this study, to make COS-based hydrogels, COS was first mixed with only poly-ethylene glycol diacrylate (PEGDA). Incorporation of PEGDA has been shown to improve mechanical strength of chitosan/COS hydrogels (Jung and Yi 2012). However, as COS had shorter chains (which explains the lower molecular weight), the extent of crosslinking was negligible during the time the polymerization reaction was allowed to proceed (~ 2 h). The condition that the reaction should occur within a reasonable duration was imposed to avoid long operation times that might be damaging to cell-laden hydrogels. Because of the watery consistency of the COS-PEGDA blend, it was decided that carboxymethyl cellulose (CMC) would be used as a thickener and an additional complexing agent. Using the two individual reactions between CMC-PEGDA, and PEGDA-COS, a possible reaction mechanism can be postulated (Fig. 6). At first, when COS was added to the CMC-PEGDA solution, the small positively charged chains of COS were electrostatically attracted to the negative domains of CMC (Fig. 6A), forming a self-assembling poly-ionic complex (Zhu et al. 2021). Under aqueous conditions, PEGDA does not crosslink with CMC or COS to any appreciable extent. Subsequent photopolymerization resulted in free radical formation on both COS and PEGDA (Fig. 6B), leading to crosslinking between these chains, in addition to cross-linking between PEGDA chains themselves. Such extensive networks resulted in gelation and formation of a cross-linked polymeric complex (Fig. 6C). As the COS chains are smaller, by being attracted to the long CMC chains, COS fragments became organized, which allowed ease of interaction with PEGDA.

Fig. 6
figure 6

A Poly-ionic complexation between COS and CMC; B Radical formation on COS chains ( • indicate free radical centers due to UV radiation; wavy lines indicate cross-linking ends of PEGDA, formed after UV irradiation); C Poly-ionic COS-CMC complex cross-linked with PEGDA.

Full size image

Physical characterization of COS-PEGDA-CMC hydrogel

Swelling ratio

After freeze drying, the two different COS-PEGDA-CMC (COSPCMC) hydrogels containing varied concentrations of CMC (2% and 4%), were incubated for 24 h at 37oC in PBS and the swelling behavior was analyzed. For COSPCMC-2 and COSPCMC-4, the ratios were 6.44 and 10.24 (average of n = 3 samples) respectively. The increase in swelling ratios are likely due to the presence of a greater number of COO- groups in a more concentrated solution of CMC, which caused the chains to stiffen and allowed better uptake of water (Barbucci et al. 2000; Reeves et al. 2010). High swelling ratios are associated with the water solubility and hydrophilicity of the precursor material (Ahmed 2015; Hafezi 2019; Li et al. 2015; Tian et al. 2015); in this case, both PEGDA and COS were observed to be soluble in pH 7. Morris et al. determined that COS-PEGDA photopolymerized hydrogels had a swelling ratio of 8 when the COS:PEGDA ratio was 1:5. Such hydrogels were capable for use as artificial tissue scaffolds (Morris et al. 2017). Chitosan-based hydrogels were found to have swelling ratios ranging from 4 to 8 in previous studies (Yao et al. 2019; Zhou et al. 2011). High water swelling ratios improve the water retention capability of the hydrogel and are related to larger mesh sizes in the gel network, thus allowing enhanced cell proliferation due to effective nutrient diffusion into the network (Park et al. 2009). Swelling ratio depends on several factors, such as the concentration of the precursor polymers, crosslinker and photoinitiator, as well as the extent of crosslinking (Slaughter et al. 2009).

Hydrogel morphology

SEM micrographs of the hydrogel cross-sections revealed irregular porous networks in both gels (Fig. 7). However, the pores in COSPCMC-2 were sparse, but of definite shapes. The surfaces also appeared to be smooth. COSPCMC-4 exhibited a highly irregular interconnected macroporous network and a rough texture (Fig. 7B). Such textures have previously been observed in COS-PEGDA hydrogels prepared via photopolymerization (Morris et al. 2017; Muñoz and Zuluaga 2017). It is possible that the increased CMC content in COSPCMC-4 interfered with the crosslinking between COS and PEGDA chains, as previous studies have found that presence of CMC in PEGDA hydrogels resulted in such textures (He et al. 2017; Kono 2014). The loosely packed network also explained why this hydrogel exhibited better swelling. Due to the superior swelling capacity and porous structure of COSPCMC-4, it was chosen for subsequent analyses.

Fig. 7
figure 7

SEM images of the cross-sections of the two hydrogels; a COSPCMC-2 (containing 2% CMC), b COSPCMC-4 (containing 4% CMC). Pores have been marked inside white boxes. Images at a magnification of 5 kX

Full size image

Functional group analysis

FTIR spectra of the two hydrogels (after washing in PBS buffer and freeze-drying) revealed important characteristic peaks (Fig. S3), which are summarized in Table S4. The peaks were identified using the peak analyzer feature of the OMICS Spectral Analysis Software. After interpretation of the peaks, the presence of all three polymers in the hydrogel could be ascertained, which implied that cross-linking was successful.

Cytotoxicity assays

The use of photoinitiators (Irgacure 2959) and synthetic polymers such as PEGDA increase the risk of cytotoxicity of hydrogels. In addition to the toxicity of the hydrogel itself, a concern in biomedical applications is the possibility of chemicals leaching from the hydrogel. Cytotoxicity studies conducted on Vero cells revealed that the leach-out products were not cytotoxic to any extent, as the cell viability was more than 95% (Fig. 8).

Fig. 8
figure 8

Vero cells observed under an optical microscope. a With leachant solution from 2.5 mg/mL hydrogel, b With leachant solution from 5 mg/mL hydrogel, c With leachant solution from 10 mg/mL hydrogel

Full size image

Mechanical tests

Uniaxial compression tests of COSPCMC-4 under swollen conditions and room temperature revealed that the material failed at forces above 0.2–1.8 N. The curves shown are representative of one sample. The stress against λ-λ−2 curve (which is associated with the shear strain) (Fig. 9) is similar to those obtained for cross-linked materials experimentally (Anseth et al. 1996; Muniz and Geuskens 2001). It can be seen that the curve exhibited non-linear characteristics, which is indicative of the viscoelastic behavior of hydrogels, that is a resistance to shear stress while maintaining elastic deformability (Anseth et al. 1996). The viscosity is a result of the free water and amorphous regions (Li et al. 2016). The curve exhibited small slopes at low deformations, as the surface of the sample was not perfectly flat. Curve fitting in the linear portion of the graphs was used to evaluate the compressive modulus, which was 4.5 ± 2.7 kPa (n = 3, ±S.D). Such elasticities are ten-fold lower than soft skeletal muscles (25 kPa) (Mathur et al. 2001). Banerjee et al. developed alginate hydrogels for the encapsulation of neural stem cells that mimicked soft neural tissues. The elastic modulus values were in the range 0.18-20 kPa; however, brain tissues have a modulus range of 100–1000 Pa, and as such, hydrogels having an elastic modulus of 180 Pa effectively promoted neuronal differentiation (Banerjee et al. 2009). Saha et al. determined that glial cell (responsible for brain structure, plasticity and nerve functions) differentiation was higher when hydrogel elastic modulus was of the range 1–10 kPa (Saha et al. 2008). In fibrin hydrogels, when the Young’s modulus was 30 kPa, Chondrocyte differentiation was the highest, in addition to type II collagen levels (Bachmann et al. 2020). It can thus be inferred that the stiffness of the gels must be such that it supports the activity of the cells for which it has been designed. Nevertheless, additional reinforcement using other polymers or improved cross-linking should be considered for COSPCMC-4 to increase the modulus to levels that are strong enough for in vivo use as artificial scaffolds (Morris et al. 2017).

Fig. 9
figure 9

Compressive stress as a function of λ - λ−2. Regression line of best fit in the linear portion of the graph was used to evaluate Young’s modulus

Full size image

Conclusion

Despite the numerous applications and benefits of COS, the main deterrent in its use is the lack of convenient production techniques. This work was intended to fill a critical research gap and develop a facile method for chitosan depolymerization. To test the applicability of the COS, a biocompatible hydrogel was developed. Among several different processing routes, physicochemical methods were found to be the most well-characterized and convenient. In this study, the ease and efficiency of irradiating a solid powder was taken into consideration and a two-stage method was devised, where gamma radiation would precede oxidative degradation with hydrogen peroxide. Several combinations of reaction conditions were tested to optimize the process so that dose level, temperature of H2O2 reaction and the concentration of the peroxide itself would be minimum. We determined that irradiation at 15 kGy, followed by oxidative degradation at 2% H2O2, at 250 C in an overnight reaction yielded COS that was soluble in pH 7–8. MW was reduced by 95%, from 249  to 12.8 ± 1.6 kDa, while DD was found to be 62.3%. Physical and chemical characterization revealed that the precursor chitosan was depolymerized, yet its functional groups remained unchanged in the process. This ascertained that the depolymerization process was suitable for COS that could be used subsequently, and a photopolymerized hydrogel was developed. The hydrogel exhibited high porosity and swelling ratios (6.44–7.17), comparable to those designed for tissue engineering and the compressive modulus was 4.5 ± 2.7 kPa, similar to that of soft tissues. Thus, the viability of COS in hydrogels was proven through the experiment. In addition of COS biomedical applications, we anticipate this newly developed hydrogel could be a potential candidate for the production of 3D printable bioink suitable for soft tissue engineering.