Viscoelasticity and injectability

Viscoelastic properties of micro- and nanosphere-based colloidal gels were investigated by rheology. As shown in Fig. 2, the viscoelasticity of micro- or nanospheres dispersions in 1 mM NaCl was evaluated by measuring the elastic modulus (G’) and tan(δ) value as a function of solid content using oscillatory time sweeps measurement. In general, elastic moduli G’ strongly increased with increasing solid content for both gelatin micro- and nanosphere dispersions (Fig. 2A). At the same solid content, nanosphere-based gels were more elastic than microsphere-based gels by a factor of about 100.

Figure 3. (A) Representative rheological profile of gel recovery measurement after network destruction of colloidal gels composed of oppositely charged micro- (MS) and nanospheres (NS). G’ and G” were monitored as a function of time: regions I, II and III show initial gel strength, gel destruction and gel recovery, respectively. 25 w/v% GelA+B microsphere-based gels with the highest initial gel strength for micro-structured gels were compared to 15 w/v% GelA+B nanosphere-based colloidal gels, since these gels exhibited similar initial storage moduli (G’). (B) Percentage of gel recovery within 5min after destruction for micro- and nano- structured colloidal gels with various solid content (5-25 w/v%).

For nanospheres, solid-like colloidal gels were formed in all tested groups as evidenced by tan(δ) values < 1, except for 5 w/v% nanosphere dispersions containing similarly charged GelB nanospheres. Especially for dispersions with solid content > 5 w/v%, highly elastic gels were formed, as evidenced by tan(δ) < 0.1 and G’ > 0.4 kPa. Moreover, the binary mixture of oppositely charged gelatin nanospheres (GelA+B) showed significantly higher G’ values than colloidal gels of similarly charged nanospheres (GelA or GelB), indicating the formation of electrostatic attractions between cationic and anionic nanospheres[10]_{. In general, gel}

water contents than microspheres. This enhanced water uptake by nanospheres can be explained by their exceptionally small size and corresponding large surface area that allows for more ab- and adsorption of water inside the nanospheres as well as water entrapment in the pores formed between nanospheres.

Figure 1. SEM images of lyophilized GelA (A, C) and GelB (B, D) micro- (A, B) and nanospheres (C, D).

Figure 2. Elastic modulus G’ (A) and tan(δ) (B) of colloidal dispersions in 1 mM NaCl solution (pH 7.0) comprising similarly (GelA or GelB) or oppositely (GelA+B) charged micro- (MS) or nanospheres (NS) as a function of solid content. Microsphere dispersions of 5 w/v% solid content were excluded since these dispersions behaved as highly flowable liquid-like

materials, whereas nanosphere dispersions of 25 w/v% solid content were excluded due to the difficulty to form homogeneous gels at such high solid content.

3.2. Viscoelasticity and injectability

Viscoelastic properties of micro- and nanosphere-based colloidal gels were investigated by rheology. As shown in Fig. 2, the viscoelasticity of micro- or nanospheres dispersions in 1 mM NaCl was evaluated by measuring the elastic modulus (G’) and tan(δ) value as a function of solid content using oscillatory time sweeps measurement. In general, elastic moduli G’ strongly increased with increasing solid content for both gelatin micro- and nanosphere dispersions (Fig. 2A). At the same solid content, nanosphere-based gels were more elastic than microsphere-based gels by a factor of about 100.

Figure 3. (A) Representative rheological profile of gel recovery measurement after network destruction of colloidal gels composed of oppositely charged micro- (MS) and nanospheres (NS). G’ and G” were monitored as a function of time: regions I, II and III show initial gel strength, gel destruction and gel recovery, respectively. 25 w/v% GelA+B microsphere-based gels with the highest initial gel strength for micro-structured gels were compared to 15 w/v% GelA+B nanosphere-based colloidal gels, since these gels exhibited similar initial storage moduli (G’). (B) Percentage of gel recovery within 5min after destruction for micro- and nano- structured colloidal gels with various solid content (5-25 w/v%).

For nanospheres, solid-like colloidal gels were formed in all tested groups as evidenced by tan(δ) values < 1, except for 5 w/v% nanosphere dispersions containing similarly charged GelB nanospheres. Especially for dispersions with solid content > 5 w/v%, highly elastic gels were formed, as evidenced by tan(δ) < 0.1 and G’ > 0.4 kPa. Moreover, the binary mixture of oppositely charged gelatin nanospheres (GelA+B) showed significantly higher G’ values than colloidal gels of similarly charged nanospheres (GelA or GelB), indicating the formation of electrostatic attractions between cationic and anionic nanospheres[10]_{. In general, gel}

elasticity of nanosphere-based gels was as follows: GelA+B > GelA > GelB. These differences, however, became less pronounced with increasing the solid content, which was related to interparticle cohesions such as hydrophobic interactions starting to dominate the reinforcing contribution of electrostatic interactions to the gel strength.

Figure 4. (A) Injectability of GelA+B micro- (MS, 25w/v% solid content) and nanosphere- based (NS, 15 w/v% solid content) gels. (B) Photographs of MS (left) and NS-based (right) gels after the injection test. (C, D) Photographs of MS (C) and NS-based (D) gels after being injected into deionized water. Nanosphere-based gels with 15 w/v% solid content were compared with microsphere-based gels with 25 w/v% solid content due to their similar initial gel strength.

In contrast, with solid content > 10 w/v%, microsphere dispersions also started to display gel-like behavior characterized by tan(δ) values < 1. It was observed that no significant differences in G’ were found between colloidal gels made of similarly (GelA or GelB) or oppositely (GelA+B) charged microspheres, suggesting that electrostatic interactions between microspheres were negligible.

To compare the injectability of micro- and nano-structured colloidal gelatin gels, a rheological method was adopted to mimic the clinic injection procedure. As shown in

Fig. 3A, both gels displayed much higher G’ than G” in region I, indicating the formation of elastic, solid-like materials. Subsequently, the gel network was destructed by applying a destructive shear strain (1000% for 1 min), which resulted into the transformation of elastic to liquid-like behavior (G’ < G” in region II), thus allowing for extrusion through syringes. Upon removal of the external shear stress, G’ of nanosphere gels restored instantaneously up to ~60% of initial gel strength within 5 seconds, and more than 80% within 5 min. This self-healing behavior after shearing, however, was not observed for microsphere gels, which showed less than 5 % of gel strength recovery within 5 min after network destruction. The percentage of gel recovery for both micro- and nano-structured colloidal gels with solid contents ranging between 5-25 w/v% was depicted in Fig. 3B, which clearly shows that all nanosphere-based gels showed self-healing behavior for all solid contents (>50% recovery of gel elasticity within 5min) whereas none of the microsphere-based colloidal gels was self-healing (<10% recovery of gel elasticity).

A representative measurement of the compressive force as a function of time during injection is given in Fig. 4, which showed that nanosphere-based gels can be easily extruded using syringes with < 10 N compression force until the plunger reached the nozzle as evidenced by a sharp increase of the injection force. Additionally, homogeneous cohesive gel-threads were formed upon injection of nanosphere- based gels from the syringe (Fig. 4B right), which still retained their noodle-like shape after injection into water (Fig. 4D). In contrast, upon injection of microsphere-based gels the compressive force gradually increased during 80 seconds, followed by a rapid force increase form <10 N to >80 N within ~50 s, at which point the plunger of syringe started to deform due to the high compression force. This so-called filter- pressing phenomenon[40] _{involves expulsion of the water phase from the tightly}

packed microspheres and stresses that microsphere-based gels cannot be completely extruded from syringes as reflected by the remaining presence of microspheres in the syringe after completion of the compression test (Fig. 4B left). Furthermore, the microsphere-based gels disintegrated immediately after being injected into deionized water.

3.3. In vitro degradation

Degradation profiles of micro- versus nano-structured colloidal gelatin gels are given in Fig. 5 which reveals linear degradation profiles for both micro- and nanospheres.

elasticity of nanosphere-based gels was as follows: GelA+B > GelA > GelB. These differences, however, became less pronounced with increasing the solid content, which was related to interparticle cohesions such as hydrophobic interactions starting to dominate the reinforcing contribution of electrostatic interactions to the gel strength.

Figure 4. (A) Injectability of GelA+B micro- (MS, 25w/v% solid content) and nanosphere- based (NS, 15 w/v% solid content) gels. (B) Photographs of MS (left) and NS-based (right) gels after the injection test. (C, D) Photographs of MS (C) and NS-based (D) gels after being injected into deionized water. Nanosphere-based gels with 15 w/v% solid content were compared with microsphere-based gels with 25 w/v% solid content due to their similar initial gel strength.

In contrast, with solid content > 10 w/v%, microsphere dispersions also started to display gel-like behavior characterized by tan(δ) values < 1. It was observed that no significant differences in G’ were found between colloidal gels made of similarly (GelA or GelB) or oppositely (GelA+B) charged microspheres, suggesting that electrostatic interactions between microspheres were negligible.

To compare the injectability of micro- and nano-structured colloidal gelatin gels, a rheological method was adopted to mimic the clinic injection procedure. As shown in

Fig. 3A, both gels displayed much higher G’ than G” in region I, indicating the formation of elastic, solid-like materials. Subsequently, the gel network was destructed by applying a destructive shear strain (1000% for 1 min), which resulted into the transformation of elastic to liquid-like behavior (G’ < G” in region II), thus allowing for extrusion through syringes. Upon removal of the external shear stress, G’ of nanosphere gels restored instantaneously up to ~60% of initial gel strength within 5 seconds, and more than 80% within 5 min. This self-healing behavior after shearing, however, was not observed for microsphere gels, which showed less than 5 % of gel strength recovery within 5 min after network destruction. The percentage of gel recovery for both micro- and nano-structured colloidal gels with solid contents ranging between 5-25 w/v% was depicted in Fig. 3B, which clearly shows that all nanosphere-based gels showed self-healing behavior for all solid contents (>50% recovery of gel elasticity within 5min) whereas none of the microsphere-based colloidal gels was self-healing (<10% recovery of gel elasticity).

A representative measurement of the compressive force as a function of time during injection is given in Fig. 4, which showed that nanosphere-based gels can be easily extruded using syringes with < 10 N compression force until the plunger reached the nozzle as evidenced by a sharp increase of the injection force. Additionally, homogeneous cohesive gel-threads were formed upon injection of nanosphere- based gels from the syringe (Fig. 4B right), which still retained their noodle-like shape after injection into water (Fig. 4D). In contrast, upon injection of microsphere-based gels the compressive force gradually increased during 80 seconds, followed by a rapid force increase form <10 N to >80 N within ~50 s, at which point the plunger of syringe started to deform due to the high compression force. This so-called filter- pressing phenomenon[40] _{involves expulsion of the water phase from the tightly}

packed microspheres and stresses that microsphere-based gels cannot be completely extruded from syringes as reflected by the remaining presence of microspheres in the syringe after completion of the compression test (Fig. 4B left). Furthermore, the microsphere-based gels disintegrated immediately after being injected into deionized water.

3.3. In vitro degradation

Degradation profiles of micro- versus nano-structured colloidal gelatin gels are given in Fig. 5 which reveals linear degradation profiles for both micro- and nanospheres.

Gelatin nanospheres showed much faster degradation, with more than 70% of gelatin degradation after 4 weeks, as opposed to <12% of gelatin degradation for microspheres. Moreover, GelB degraded faster than GelA with ~12% of GelB microspheres degraded after 4 weeks as opposed to 3% degradation for GelA microspheres. These differences between the two types of gelatin were not observed for nanospheres made of GelA and GelB. Furthermore, no significant difference in cumulative degradation of similarly (GelA or GelB) or oppositely charged (GelA+B) micro- and nanosphere-based gels was detected during the enzymatic degradation of gelatin.

Figure 5. Cumulative in vitro degradation profiles of colloidal gels made of similarly (GelA or GelB) or oppositely (GelA+B) charged micro- (MS) or nanospheres (NS), which were incubated in collagenase-containing PBS to induce enzymatic degradation of gelatin.

SEM images provided additional visual information on the process of enzymatic degradation of gelatin micro- and nanospheres (Fig. 6). Degradation of GelA and GelB microspheres after 2 weeks was hardly visible, which confirmed the slow degradation rate of gelatin microspheres as measured by means of the BCA assay. Microsphere degradation became noticeable after 4 weeks only, with obvious surface erosion for both GelA and GelB microspheres (Fig. 6 I and J). In contrast, obvious morphological changes were observed for nanospheres after 1 week, since both GelA and GelB nanospheres started to fuse together with the neighboring particles and form an interconnected structure (Fig. 6C and D). These integral constructs also experienced a surface erosion process, as evidenced by the loss of sphericity of

GelB nanospheres after 2 week incubation (Fig. 6H). Especially, continuous degradation of GelB nanosphere-based gels resulted into loss of the nano-structure and the formation of irregularly-shaped integral constructs (Fig. 6L).

Figure 6. Representative SEM images of GelA or GelB micro- (MS) and nanospheres (NS) after 1 (A-D), 2 (E-H) and 4 (I-L) weeks of in vitro degradation. Scale bar for images of microspheres is 10µm, for nanospheres 1µm.

Figure 7. Cumulative BMP-2 release from colloidal gels made of similarly (GelA or GelB) or oppositely (GelA+B) charged micro- (MS) or nanosphere-based (NS).

Gelatin nanospheres showed much faster degradation, with more than 70% of gelatin degradation after 4 weeks, as opposed to <12% of gelatin degradation for microspheres. Moreover, GelB degraded faster than GelA with ~12% of GelB microspheres degraded after 4 weeks as opposed to 3% degradation for GelA microspheres. These differences between the two types of gelatin were not observed for nanospheres made of GelA and GelB. Furthermore, no significant difference in cumulative degradation of similarly (GelA or GelB) or oppositely charged (GelA+B) micro- and nanosphere-based gels was detected during the enzymatic degradation of gelatin.

Figure 5. Cumulative in vitro degradation profiles of colloidal gels made of similarly (GelA or GelB) or oppositely (GelA+B) charged micro- (MS) or nanospheres (NS), which were incubated in collagenase-containing PBS to induce enzymatic degradation of gelatin.

SEM images provided additional visual information on the process of enzymatic degradation of gelatin micro- and nanospheres (Fig. 6). Degradation of GelA and GelB microspheres after 2 weeks was hardly visible, which confirmed the slow degradation rate of gelatin microspheres as measured by means of the BCA assay. Microsphere degradation became noticeable after 4 weeks only, with obvious surface erosion for both GelA and GelB microspheres (Fig. 6 I and J). In contrast, obvious morphological changes were observed for nanospheres after 1 week, since both GelA and GelB nanospheres started to fuse together with the neighboring particles and form an interconnected structure (Fig. 6C and D). These integral constructs also experienced a surface erosion process, as evidenced by the loss of sphericity of

GelB nanospheres after 2 week incubation (Fig. 6H). Especially, continuous degradation of GelB nanosphere-based gels resulted into loss of the nano-structure and the formation of irregularly-shaped integral constructs (Fig. 6L).

Figure 6. Representative SEM images of GelA or GelB micro- (MS) and nanospheres (NS) after 1 (A-D), 2 (E-H) and 4 (I-L) weeks of in vitro degradation. Scale bar for images of microspheres is 10µm, for nanospheres 1µm.

Figure 7. Cumulative BMP-2 release from colloidal gels made of similarly (GelA or GelB) or oppositely (GelA+B) charged micro- (MS) or nanosphere-based (NS).