3.1. Rheological investigation at various pH
A hydrogel nanocomposite was designed, comprising a thermoresponsive Alginate-based gelator, encapsulating polymeric nanoparticles (NPs) as nano-carriers. Particularly, the gelator used is the heterograft copolymer constituted of an alginate backbone, grafted by two different types of thermoresponsive polymer chains (ALG-g-HG): (a) pure poly(N–isopropylacrylamide) (PNIPAM) homopolymer and (b) poly[(N–isopropylacrylamide)–
co–(N–tert–butyl acrylamide)] (P(NIPAM–
co–NtBAM)) random copolymer, displaying different lower critical solution temperatures (LCST) [
42].The aqueous dispersion of 4 wt% ALG-g-HG, at pH 7.4, exhibits a sol-gel transition upon heating at T
gel=34.5 °C as determined by oscillatory shear experiments at the G’/G” crossover point (
Figure S1a). The NPs formed by spontaneous self-assembly of 1 wt% P2VP-
b-PEO block copolymers in aqueous media of pH 7.4, forming star-like micelles of core-shell structure with P2VP hydrophobic cores and PEO water-soluble shells. Note that P2VP is a pH-dependent cationic polyelectrolyte, exhibiting pK
a at pH 5.0, above which becomes hydrophobic, driving hydrophobic association [
49,
50,
51]. The hydrodynamic diameter of NPs at pH 7.4 was determined by dynamic light scattering at 23.56 nm (volume average, PDI 0.267) (
Figure 1). Some bigger aggregates appearing in the volume-weighted size distribution are negligible in number as they disappear in the number-weighted distribution.
The final mixed aqueous solution under investigation contained 4 wt% ALG-g-HG and 1 wt% P2VP-
b-PEO-based NPs. Nanocomposite hydrogels regulated at different pH were explored by oscillatory shear measurements to evaluate potential influence of pH to their thermal and mechanical response. The pH window was chosen in the narrow range of 3.5-5.5 which lies in the vicinity of pK
a of P2VP in which its degree of ionization increases by lowering pH (see z-potential versus pH in SI,
Figure S2).
Figure 2 demonstrates the storage (G’) and loss (G”) modulus as a function of temperature at different pH values. The experiments were conducted in the linear viscoelastic regime by applying a successive heating/cooling ramp with a rate of 1 °C/min. In all cases the moduli exhibit strong temperature dependence revealing a thermothickening behavior which depends on the medium pH. At pH 5.4, a thermo-induced sol-gel transition can be observed, providing that the storage modulus increases sharply upon heating, surpassing the loss one at the critical T
gel determined at 34.1 °C. The phenomenon is reversible as revealed by the cooling procedure, with slight hysteresis. The gel-sol transition was shifted at 30.9 °C, a few degrees lower than that observed by heating, due to kinetic effects [
29]. This thermothickening behavior is quite similar with the formulation of 4 wt% ALG-g-HG without NPs (
Figure S1b). This implies that the presence of P2VP-
b-PEO-based micelles at pH 5.4 does not affect the network structure formed by the physical crosslinking of the thermoresponsive side chains of the gelator. However, by lowering pH, remarkable changes on the characteristics of the hydrogel thermoresponsiveness can be observed. The main variation observed in the general thermal behavior of the formulations is that the sol-gel transition observed at pH 5.4 was transformed to a soft gel-strong gel transition at pH 3.5, since the G’/G” crosslinking disappeared and G’ predominates G’’ in the entire temperature range explored. At the intermediate pH 4.5 and 4.0, thermoresponsive behavior exhibits additional significant hysteresis effects. Particularly, at pH 4.5 the transition point T
gel in the heating procedure was shifted to lower values about 8 °C, notably more pronounced than the cooling procedure. While at pH 4.0, T
gel (heating) dropped to 20.2 °C that is about 14 °C lower from that at pH 5.4. Moreover, in the cooling procedure, the gel-sol transition did not occur, provided that the moduli although decreased upon lowering temperature, they never crossed each other, hence preserving the gel-like behavior at low temperatures.
We now focus on the data dealing with the heating process, which is related to potential biomedical applications.
Figure 3 presents the temperature dependence of the elastic part of the complex modulus, G’ (
Figure 3a), together with the loss tangent (G”/G’) at various pHs. By lowering pH two main effects can be observed: a) a decrease of the sol-gel transition (defined at tanδ=1,
Figure 3b) and b) the elasticity of the formulation is expanded at lower temperatures and simultaneously increases. Provided that G’ reflects the degree of crosslinking, this behavior should be ascribed to additional crosslinking leading to the mechanical reinforcement of the hydrogel. This is also reflected in tanδ which decreases with pH alongside the temperature variation.
To better analyze qualitatively the data, we can distinguish three temperature zones as indicated in
Figure 4: the low temperature zone, the high temperature zone and between them the transition zone. The latter zone is defined between the temperature at the onset of the abrupt increase of the elastic modulus G’ namely T
c,thermothickening and T
f, at the temperature where G’ becomes independent of T or increases slightly and smoothly. A number of factors characterizing the thermothickening phenomenon based on the elastic modulus G’, i.e., T
c,thermothickening, T
f, T
gel, ΔT(transition zone) and extend of the thermothickening effect exemplified by the ratio of G’ augmentation [G’(T=50)/G’(T
c,thermothickening)] are gathered in
Table 1 at various pHs.
One of the important effects of pH concerns the transition zone. As indicated in
Figure 5a, the onset of thermothickening is shifted to lower temperature more than 13 °C from pH 5.4 to 3.5. This causes the broadening of the transition zone ΔΤ from 9.4 to 19.5 degrees and the shifting of the sol-gel transition to lower temperatures. As it is known, the thermothickening effect is due to the thermal-induced hydrophobic interactions of the PNIPAM-based sticky side chains of the gelator, which are activated above their LCST, leading to intermolecular association. The fact that this association starts occurring at lower temperatures implies that the solubility of the side chains was obstructed by other interactions. Indeed, focusing in the low temperature zone we observe a remarkable increase of G’ more than two order of magnitude (
Figure 3a and
Figure 5b) which leads to gel-like behavior even at low temperatures below T
gel at pH 3.5. This suggests the formation of a network different than that induced by heating, which is formed from the interactions, responsible for the variations (T
c,thermothickening, T
gel, ΔT) observed in the transition zone. Moreover, these new interactions seem to contribute as well to the G’ augmentation in the high temperature zone (
Figure 5b). Provided that the G’ augmentation is more pronounced in low temperatures it eventually weakens significantly the extend of the thermothickening effect, as reflected by the G’(T=50)/G’(T
c,thermothickening) ratio, from 110.5 to 3.5 (
Table 1). This likely suggests that a smaller number of side chains are now available for hydrophobic association.
In order to elucidate the nature of the interactions, other than the hydrophobic one induced by heating, that contribute to the reinforcement of the network, we also explored the behavior of the ALG-g-HG without the presence of NPs.
Figure S3 demonstrates the data from the temperature ramp experiments performed in the linear viscoelastic regime (strain amplitude 0.1%) at 1 Hz of 4 wt% ALG-g-HG at pH 4.5 and 3.5, keeping the same conditions with the composite formulations. For the sake of comparison these data have been plotted together with the data of
Figure 3a and
Figure S1b in
Figure 6. Surprisingly, the behavior of the bare ALG-g-HG gelator is quite like those of the formulations incorporating the P2VP-
b-PEO micellar NPs, since the data for the corresponding formulations at the same pH almost coincide. This suggests that the additional crosslinking of the ALG-g-HG, observed at low temperatures in the absence of PNIPAM hydrophobic association, must be attributed to the intermolecular interaction between the alginate backbone and the PNIPAM-based side chains. This is corroborated from the shift of the characteristics of the transition zone (T
c,thermothickening, T
gel) to lower temperatures discussed above. Note that gel-like behavior was not observed in pure alginate formulations of 4 wt % at the same pH range. At low pH, although the degree of ionization of the alginate backbone decreases, a number of negatively charged units persists. These units might interact with the positively charged units of P2VP lying within the micellar cores (
Figure S2). It seems that this is not the case, which should be attributed to the thick PEO shells which protect the P2VP cores from contacts with the grafted alginate. However, some differentiation in T
c,thermothickening between the composite and the pure ALG-g-HG, i.e. shift at lower temperatures for the composite at pH<5, below the pK
a of P2VP (
Table S3), shows that ionic interactions between the charged units of P2VP and alginate should not be excluded.
In a previous paper [
18] dealing with the thermothickening of ALG-g-P(NIPAM–co–NtBAM) hydrogels, it was shown that the ionic interactions between the negatively charged alginate units with the divalent cations of Ca
2+ resulted to strong reinforcement of the network, as revealed by a significant increase of the elasticity below and above the transition temperature of thermothickening. Interestingly, the presence of the additional ionic interactions slightly affected the transition temperature, i.e. a few degrees lowering for the highest Ca
2+ concentration. This means that the ionic interactions act independently and do not affect the hydrophobic ones exerted by the PNIPAM-based side chains, inducing thermothickening. The above corroborate our interpretation that the significant shift of the characteristics of the transition zone (
Table 1 and
Table S3) is due to the interactions between the alginate backbone and the side chains, the nature of which is likely of H-bonding, providing that at low pH most of the carboxylic units are in their protonated form.
3.2. Self-healing and Injectability
As it is known, this kind of physically crosslinked hydrogels exhibit finite linear viscoelastic regime, the extent of which depends on the network structure. Above a certain critical strain (γ
c) a gel-sol transition occurs due to the mechanical disruption of the physical crosslinks. This γ
c was determined at 51% conducting a strain sweep test at 37 °C on the formulation of pH 4.0 (
Figure S4), where a reinforced network was observed.
To evaluate the self-healing of this hydrogel, successive oscillatory, time sweep experiments were performed by applying stepwise strains below and above γc. Particularly, the hydrogel was subjected to strain amplitudes at 1% and 300% for 600 s at each step, at 37 °C. As seen in
Figure 7, the formulation exhibits gel-like behavior (G’>G”) at 1% strain which turns instantly to sol (G’<G”) upon applying 300% strain. The hydrogel was recovered instantly in the third step by switching the strain amplitude again to 1%, within the linear regime. The hydrogel recovery was reproducible as shown in the fourth and fifth steps. The above results confirmed the self-healing ability of the hydrogel since it is recovered after sudden network disruption.
Concerning the hydrogel performance (e.g. 3D printing), injectability is one of the critical properties of hydrogels as carriers of payloads. This property can be easily afforded when the thermoresponsive hydrogel exhibits sol-gel transition above room temperature, since it flows below the gelation temperature. This is the case of all the formulations for pH>4.0, while for the system at lower pH (pH 3.5), a gel-like behavior is valid even at room temperature. In the latter case shear thinning behavior is demanded. Hence, shear flow experiments were conducted to evaluate the response of the hydrogels to shear.
Figure 8 displays the apparent shear viscosity (η/Pa.s) as a function of shear rate (increasing/decreasing) obtained at 20 °C for all pHs. Obviously, the viscosity increased by lowering pH in any shear rate examined, exhibiting shear thinning except for that at pH 5.4 where a Newtonian behavior was observed. Notably, the shear thinning is perfectly reversible practically without hysteresis since the viscosity values were recovered by decreasing shear rate. Moreover, the extent of shear thinning augments when the viscosity increases by lowering pH. For instance, the viscosity drops more than 500 folds by progressively increasing the shear rate from 0.01 to 100 s
-1 for the formulation of pH 3.5, exhibiting the highest elasticity, while at pH 4.0, the viscosity decrease is about 200 times lower.
Finally, the response of viscosity to shear was evaluated by sudden changes of the applied shear rate as demonstrated in
Figure 9. Specifically, the formulation of the highest elasticity (pH 3.5) was subjected to successive stepwise changes of shear rates from 0.01 (approaching situation at rest) to 20 s
-1 (situation during injection). As observed, the viscosity drops about two orders of magnitude, and it is recovered promptly and reproducibly. In conclusion, the results of the experiments demonstrated in
Figure 7,
Figure 8 and
Figure 9 showed that this kind of hydrogels exhibit reversible shear thinning, responding perfectly to sudden changes of shear and self-healing properties, meeting the requirements of 3D printable hydrogels.
3.3. Evaluation as drug delivery system
In order to evaluate the effectiveness of the nanocomposite hydrogel as pH-controlled drug delivery system, the micellar P2VP-b-PEO NPs were loaded with the hydrophobic Nile Red which was encapsulated in their hydrophobic cores and its release kinetics was conducted at the physiological temperature at two different pH.
It was found that the P2VP-b-PEO@Nile Red micelles displayed Nile Red loading efficiency LE=1.79% (w/w) for a theoretical (feed) loading Lth= 9.98% (w/w) and encapsulation efficiency EE=17.93% (w/w).
Nile Red-loaded P2VP-
b-PEO micelles (P2VP-
b-PEO@Nile Red) and ALG-g-HG/ /P2VP-
b-PEO@Nile Red nanocomposite hydrogels were studied to evaluate the Nile Red release.
Figure 10 presents the cumulative release of Nile Red versus time for both nanocarrier systems, at 37 °C and at two different buffer solutions of pH 3.5 (similar to the pH of gastric acid in the human stomach) and pH 7.4 (simulates the pH of blood and pH of healthy tissues). From the release study of the P2VP-
b-PEO@Nile Red micelles, it appears that the release process is highly influenced by the pH of the medium. The micelles in buffer pH 3.5 release a higher amount of Nile Red (16% of Nile Red is released at 120 hours) than the micelles in buffer pH 7.4 (4% of Nile Red is released at 120 hours). This is due to the protonation of the P2VP moieties of the P2VP-
b-PEO polymer that makes them less hydrophobic, leading to micelles with a less dense P2VP core. This way, the Nile Red molecules can escape easier from the hydrophobic P2VP micelle cores.
Table S4 in Supplementary Material summarizes the particle size distributions of the P2VP-
b-PEO NPs at three different pH values of the aqueous medium. As observed when the pH was adjusted to a lower value of pH 3.5, the micellar diameter was substantially decreased to ~3.5 nm. At such acidic environment, the degree of ionization of P2VP blocks increases by protonation, leading to disassembly of the micelles into smaller associates or even unimers. These results are in good agreement with those previously reported for copolymers with similar PEO/P2VP(P4VP) hydrophilic/hydrophobic balance [
52,
53,
54]. Thus, one would expect acceleration of the Nile Red release from the P2VP-
b-PEO micelles at pH 3.5. where the P2VP blocks are partially charged (
Figure S2). However, only 16% of Nile Red was achieved at 120 hours, at pH 3.5. This probably could be attributed by the fact that Nile Red is a molecule of high hydrophobicity which could also form π–π stacking interactions with the PVP blocks, making it firmly retained in the polymer micelles.
At pH 7.4, an initial burst release can be observed, which may imply that some Nile Red molecules are weakly bound within the micelle cores and that a fraction of the encapsulated Nile Red is located at the core/corona interface of the micelles and therefore it is more accessible and prone to release. In this pH, high above the pK
a of P2VP at pH 5.0, the P2VP moieties are hydrophobic. The low release 4% of Nile Red in buffer pH 7.4 could be attributed to the highly hydrophobic nature of Nile Red and possible strong, non-covalent interactions of Nile Red with the P2VP hydrophobic micellar cores causing a more firmly Nile Red entrapment in the P2VP-
b-PEO micelles. These findings are in agreement with the pH-dependent release profiles of various hydrophobic drugs from poly(ethylene glycol)-
b-poly(2-vinylpyridine) micelles. It should be mentioned that the release kinetics of hydrophobic drugs is remarkably influenced by their chemical structure except of the size and the degree of protonation (controlled by pH) of the P2VP blocks forming the hydrophobic cores [
54].
Inspired by the study of Chroni et al. [
55], which proved that ultrasound has significant influence on the release profile of curcumin (CUR) from CUR-loaded nanocarriers, we proceeded to a similar study. So, after the 120 hours of the release experiments, the samples were subjected to ultrasonication. It can be clearly seen that for pH 3.5, after 8 hours under ultrasound stimulation, a high release of Nile Red occurred (release rate increased from 16% to 31% at 128 hours (total experiment hours). After further sonication, 16 more hours (to reach 144 total experiment hours), Nile Red release increased further to 35%. The increase of Nile Red release from the micelles submerged in the PB buffer of pH 7.4, was marginal, showing a rise of only 2% after 24 hours of sonication (from 4% at 120 hours, the release of Nile Red reached a 6% at 144 hours). This release resistance under sonication confirms the strong attractive interactions between Nile Red and the vinyl pyridine moieties of the micellar cores (see also reference [
56]), which justifies the observed release kinetics.
When the P2VP-
b-PEO@Nile Red micelles were loaded in the ALG-g-HG hydrogel, it can be seen, that for both PB buffer solutions at pH 3.5 and pH 7.4, Nile Red release is lower, around 6% and 3.3%, suggesting noticeable slowdown of the Nile Red release rate which is attributed to the high viscosity of the hydrogel medium [
46]. Importantly, this effect was more pronounced at pH 3.5. This behavior is related to the pH-induced alteration of the network structure formed by the ALG-g-HG gelator. As already discussed in
Figure 2 and
Figure 3, the network is reinforced by decreasing pH as reflected to G’ augmentation, due to additional crosslinking induced by H-bonding. Particularly, the complex viscosity η* increased remarkably with the pH reduction at 37 °C, where the release experiments were conducted, as demonstrated in
Figure S5. This effect might justify the deceleration of the Nile Red release rate due to the decrease of drug diffusion within the higher viscosity hydrogel medium.