In order to learn from each hydrogel development cycle, it is important to properly understand what we have made. The gross characteristics can be understood from visually inspecting the hydrogel and performing mechanical testing, but the most important factors lie in the microscale structuring of the hydrogel. Such factors include pore size of the hydrogel, the distribution of the nHAP particles, and the change in these properties across the gradient gel.
In order to properly understand these microscale factors we chose to pursue scanning electron microscope (SEM) imaging, as standard light microscopy does not have the resolution to investigate the structuring of our protein fibers, or to resolve the nHAP particles. Our sample preparation included freezing the samples at -18°C before lyophilization, which creates a dehydrated version of the gel that retains the microstructural features. After lyophilization the gels were cut to a 75mm square and mounted on SEM stubs, before being coated with ~10nm of AuPd (gold/palladium). Samples were then imaged in the SEM at 3kV, and images were taken from 50 to 80,000x magnification. These images allowed us to learn from each development cycle, highlighting what issues needed to be fixed and what combination of factors produced the intended microstructure. A selection of images are included here, showing the various morphologies we encountered as well as our battle with achieving properly distributed nHAP particles.
The first imaging session looked at a set of gelatin hydrogels frozen at different temperatures, in order to understand the effect various freezing methods had on the final morphology. We tested at -196°C (Fig. 7A) and -18°C (Fig. 7B), the temperature of liquid nitrogen and the temperature of our freezer. From these results, the -18°C hydrogels had pore sizes of 50-200 μm and the -196°C hydrogels had pore sizes of 2-6 μm. From literature reviews we determined the optimal pore size for our bone scaffolds to be 100-500 μm [1], and tendon scaffolds to be 50-700 μm, depending on the source [2]. Due to this result, we chose to move forward with -18 °C freezing as it was closer to the intended pore size.
Early tests with nano hydroxyapatite (nHap) particles in gelatin scaffolds showed issues, as the nHap particles clumped together into large clusters (Fig. 7B’), instead of distributing throughout the gelatin matrix as was intended. Having this information allowed us to better plan how to distribute the nHap, which was achieved through the implementation of an ultrasonication step during hydrogel manufacturing. The SEM images of our final bone-replica collagen hydrogels (Fig. 8A) show the nHap particles distributed and integrated into the collagen matrix as intended (Fig. 8A’), more closely mimicking the natural bone environment. Pore sizes in these gels were on the range of 50-300 μm, close to the ideal range. Our final tendon-replica collagen hydrogels had pores of 50-500 μm (Fig. 8B), also within the acceptable range. They further showed a laminar structure with a preferential axis, which is similar to the form of native tendon tissue. This was unintentional but much appreciated, and was likely due to cohesive forces of the collagen gel attaching to the side of a small mold placing the gel under tension during crosslinking.
Figure 7 SEM images of gelatin hydrogel frozen at -196°C (A), and -18°C (B). B’ and associated inset show clumping of nHap nanoparticles. A & B at 1000x magnification, A’ & B’ at 5000x magnification, B’ inset at 45,000x magnification.
Figure 8 SEM images of collagen hydrogels at 100x magnification (A & B) showing microstructure, and 10,000x magnification (A’ & B’) showing nanostructure.
In order to determine the optimal UV crosslinking times for our collagen and Scl-2 hydrogels, we chose to use a rheology system with an integrated UV curing attachment (TA Instruments DHR3 Rheometer). This allowed us to continuously measure the elastic (G’) moduli of the samples as they were exposed to UV light, in order to stop the procedure when they reached the proper moduli and before they began to degrade from the effects of UV irradiation.
In our first trial with UV modified scaffolds, no increase in elastic modulus was observed during the entire extent of UV irradiation (Fig. 9). While there may have been multiple factors contributing to this, after consultation with Leigh Slyker, a graduate student studying collagen biomimetic cartilage hydrogels in Cornell’s Bonassar research group, we determined the most likely cause was the concentration of the collagen solution we used. His recommended range of concentrations were 10-50x more concentrated than what we had been using, and judging from the rheological data and SEM images he recommended trying an increased concentration of collagen. Our second rheological testing trial used hydrogels prepared in a similar fashion, with a collagen concentration of 10x our previous attempt. During the UV irradiation of these gels they never reached the stiffness required for a proper bone scaffold, but the rheological data showed clear increases in elastic modulus as the test progressed (Fig. 9). This same hydrogel mixture was also crosslinked with glutaraldehyde into a stiff gel, which showed that a secondary issue aside from the collagen concentration was likely within the UV functionalization procedure. As we were working with small volumes during this procedure (<1mL), our current hypothesis for this error is insufficient mixing and incomplete reactions. Although we allowed critical reaction steps to proceed for 18 hours with shaking, the high viscosity of the liquid prevented any significant flow inside the test tubes.
Figure 9 Elastic Moduli of UV Crosslinked Collagen Hydrogels
- 1. Loh, Qiu Li, and Cleo Choong. “Three-dimensional scaffolds for tissue engineering applications: role of porosity and pore size.” Tissue engineering. Part B, Reviews vol. 19,6 (2013): 485-502. doi:10.1089/ten.TEB.2012.0437
- 2. Lutzweiler, Gaëtan et al. “The Overview of Porous, Bioactive Scaffolds as Instructive Biomaterials for Tissue Regeneration and Their Clinical Translation.” Pharmaceutics vol. 12,7 602. 29 Jun. 2020, doi:10.3390/pharmaceutics12070602