Biologically enhanced starch bio-ink for promoting 3D cell growth

Abstract

The excellent rheological property has legitimated the suitability of starch hydrogel for extrusion-based 3D printing. However, the inability to promote cell attachment and migration has precluded the non-modified starch hydrogel from direct applications in the biomedical field. Herein, we develop a novel 3D printable nanocomposite starch hydrogel with highly enhanced biocompatibility for promoting 3D cell growth, by formulating with gelatin nanoparticles and collagen. The rheological evaluation reveals the shear-thinning and thixotropic properties of the starch-based hydrogel, as well as the combinatorial effect of collagen and gelatin nanoparticles on maintaining the printability and 3D shape fidelity. The homogeneous microporous structure with abundant collagen fibers and gelatin nanoparticles interlaced and supplies rich attachment sites for cell growth. Corroborated by the cell metabolic activity study, the multiplied proliferation rate of cells on the 3D printed nanocomposite starch hydrogel scaffold confirms the remarkable enhancement of biological function of developed starch hydrogel. Hence, the developed nanocomposite starch hydrogel serves as a highly desirable bio-ink for advancing 3D tissue engineering.

Graphical Abstract

To fufill the physiochemical requirement of bioprinting and support the cell growth, a novel, printable and shear-thinning nanocomposite starch-based hydrogel is formulated. The developed starch-based nanocomposite hydrogel is mechanically competent to maintain superior printability and shape fidelity, while sustaining long-term cell growth, facilitate cell attachment, migration, and proliferation.

1. Introduction

3D bioprinting offers an unmatched potentiality for creating stand-alone and well-defined tissue architectures with either intricate patterns or customized zones.^[1-3] 3D printed scaffold can greatly improve the mass transport throughout the cultured tissue system, which is highly attractive for various tissue engineering applications.^{[4, 5]} One of the determining factors for successful printing is from bio-ink ^{[6, 7]} which ideally fullfills the requirements of adopted printing technique, for instance, viscosity, shear-thinning, low yield stress and thixotropic property, but, of the utmost importance, supporting the attachment and 3D growth of seeded cells.^{[8, 9]}

Natural extracellular matrix (ECM) biomaterials, such as collagen, undergo irreversible deformation when extruded after gelation without the ability to retain the material shear thinning. Most importantly, the natural ECM materials such as collagen have low stiffness, which significantly limits the ability to form free-standing structures for maintaining good 3D shape fidelity.^[10] Modifications such as the addition of various rheological modifiers and post-printing or on-site chemical/photo-crosslinking have been employed, which resulted in an increasing diversity of mechanically competent materials for 3D bioprinting.^{[11, 12]} However, such crosslinking steps could impair the biocompatibility of developed 3D bio-ink by introducing exogenous chemical groups subjecting to potential toxicity and immunogenicity in both in vitro and in vivo applications, or result a mechanically incompliant network that impedes the desired cellular responses. Given the requirement of both desired mechanical property and biological function, using naturally derived biomaterials without any chemical modification for 3D printing is critically challenging.^[13]

Owing to its cost-effectiveness, availability, biocompatibility and excellent biodegradability under physiological conditions, starch has been identified to be an appealing natural biomaterial for tissue regeneration and drug delivery.^{[14, 15]} In order to improve cell adhesion ability and thermal stability from starch, other polymers such as PVA, chitosan and collagen have been introduced to form starch-based blends.^[16-18] Through physically associated networks, starch-based blends with improved mechanical strength, biocompatibility, and processability have been achieved. Early studies have demonstrated the utility of starch-based scaffolds for bone tissue engineering.^{[19, 20]} Recent years have witnessed an increasing exploration on the therapeutic potential in wound healing of starch-based scaffolds.^{[21, 22]} Various types of starch have been tailored into scaffolding preparation methods such as electrospinning, salt leaching, inject molding.^{[18, 23, 24]} To expand its versatility, recent studies have also demonstrated the suitability of starch hydrogel for 3D bioprinting.^[14] However, the reported studies either utilize starch with a higher concentration over 25% to secure the mechanical stiffness, or sacrifice 3D printability of starch in the low concentration range.^[25]

In order to overcome above-mentioned limitations, we introduce a novel 3D printable nanocomposite starch hydrogel with highly enhanced biocompatibility for promoting 3D cell growth, by formulating with natural collagen and gelatin nanoparticles. Although a wide series of nanoparticles, including nanosilicates,^{[26, 27]} starch or cellulose nanocrystals,^{[12, 28]} gold nanoparticles,^[29] carbon nanotube and graphene oxide have been investigated and incorporated into hydrogel networks to achieve desired material properties (e.g., electrical conductivity, printability, mechanical stiffness, and stimuli response),^{[30, 31]} using gelatin nanoparticles in starch hydrogel for enhancing 3D cell culture and tissue scaffolding ability has not been explored elsewhere. With the incorporation of gelatin nanoparticles and collagen in starch-based hydrogel to bring nanoscale topological features, cells could reside in a highly biocompatible microenvironment for interactive binding and attachment, which is critically needed in regulating 3D cellular response and tissue growth. We optimized the composition of developed starch hydrogel for achieving the high 3D printing fidelity through dynamic rheological characterization and printability evaluation. The results demonstrated that the combination of gelatin nanoparticles and collagen could tune the mechanical stiffness as desired for extrusion-based 3D bioprinting. Further microstructure analysis revealed that the nanocomposite hydrogel encompassed hierarchical porous structures with distinct fibrous collagen distributing throughout the pore surface, which is highly desirable for cell attachment and migration. In vitro cellular response investigation using mammalian fibroblasts showed excellent cellular viability and interaction with 3D printed nanocomposite starch hydrogel. Therefore, this work introduces a non-modified, low-cost, and handy 3D printable nanocomposite starch hydrogel, with a supplementary network formed by gelatin nanoparticles and collagen for enhancing the biological property. We expect this highly biocompatible; 3D printable nanocomposite starch bio-ink could be broadly employed in the field of 3D tissue engineering and therapeutic development.

2. Results and Discussion

2.1. Evaluations of GNPs

Gelatin nanoparticles were synthesized successfully based on a two-step coagulation process.^[32] NTA and DLS analysis were both employed to determine the size of GNPs, as revealed in Figure 1a and Figure S1. We observed the peak size of GNPs at ~40 nm (diameter), from NTA analysis, which is consistent with the observation from SEM imaging (Figure 1b). Although the results from DLS showed a broader distribution and a peak particle size at ~106 nm (diameter), due to the recognition of the scattering intensity fluctuation from the DLS, the contribution from the smaller particles could be overshadowed by the higher scattering signals from the larger particle aggregates, leading to the overshooting on the nanoparticle size. The cytotoxicity of the GNPs was evaluated using PrestoBlue assay that relies on metabolic activity of the active NIH 3T3 cells. Gelatin nanoparticles with varied concentrations spanning from 0 to 10¹⁰ particles/mL were incubated with NIH 3T3 cells for 5 days. As evidenced in Figure 1c, the cell proliferation rates in the groups with GNPs were comparable to the non-GNPs control group, highlighting the negligible toxicity of GNPs on cultured cells.

Figure 1.

Assessment of gelatin nanoparticles (GNPs) in terms of size distribution, morphology, and toxicity on cell proliferation. a) Size and concentration measurement using Nanoparticle tracking analysis (NTA). Red color indicates the standard deviation. b) FESEM images of gelatin nanoparticles in DI water with lower and higher magnifications (insert). c) Effects of GNP concentration on the cell proliferation over a 5-day culture (n=9 technical replicates), determined using a Prestoblue assay.

2.2. Preparation of GNP Nanocomposite Starch Bio-ink

Starch, as the plant-derived polysaccharide, has been recognized as a natural material in the field of tissue engineering due to its biocompatibility, biodegradability, and ultra-low-cost. The property of starch is generally governed by the ratio of two main structural components: amylose and amylopectin.^[33] The amylose impairs the gel strength, while the amylopectin determines the gel viscosity. The formation of starch hydrogel is a three-step thermal treatment that includes swelling, gelatinization, and retrogradation to form a 3D hydrogel network (Figure S2 and 2b).^[34] To introduce the GNPs into the starch hydrogel, we first suspended the GNPs with the desired concentration into 50 mmol HEPES buffer, then vigorously mixed with starch granules for 20 min to form the GNP-supplemented starch hydrogel (Figure 2a). Additionally, we added collagen as a cell-instructive component to mimic the native tissue due to its ubiquitous presence in most tissues and organs. Collagen forms fibrous networks in the body, which can enhance the tissue structure and function of ECM while promoting cell adhesion, growth, tissue morphogenesis, and biological signaling. We have investigated 6 combinations of starch, collagen and GNPs for developing the best 3D bio-ink recipe in terms of printing fidelity, structural integrity and cell growth. The 3D printed anatomical ear structure was shown as the example in Figure 2c demonstrating the excellent printing fidelity. The printed scaffold also exhibits improved biological property to promote cell attachment and proliferation, as shown in Figure 2d by culturing NIH 3T3 cells for 7 days.

Figure 2.

a) Schematic illustration of the development of 3D printable nanocomposite starch-based hydrogel b) The insert image showed the grid pattern in good shape fidelity with 15 mm × 15 mm × 2 mm cubic structure that can be 3D printed with pure starch hydrogel. c) 3D Printed anatomical ear using nanocomposite starch-based hydrogel with good shape fidelity. Printing temperature was maintained at 6°C. Blue color is food dye. d) NIH 3T3 cell spreading on 3D printed nanocomposite starch hydrogel scaffold over a 7-day culture, indicating the excellent ability for promoting cell growth and attachment.

2.3. Rheological characterization of nanocomposite starch-based hydrogels

We investigated the rheological profiles of variable combinations of GNPs, starch hydrogel, and collagen, which could inform the development of high-quality bio-ink with the best printability and the 3D shape fidelity. As demonstrated in Figure S6 by printing a 10-layer scaffold in a grid pattern, we observed that starch is the dominant ^factor to maintain the 3D printability, as long as the concentration is larger than 10%. The addition of collagen had adversely affected the printability. This observation was further quantitatively validated by shear rate sweep analysis in Figure 3a, along with the pure starch hydrogels and starch-collagen blend as the control group. As confirmed, significant increase of viscosity was observed along with increasing starch concentration across all the groups. Interestingly, the addition of collagen resulted in the 90.52%, 61.51%, 38.91% reduction of zero-shear viscosity, compared to the corresponding groups with 7.5%, 10%, and 12.5% pure starch hydrogel only. The final concentration of collagen at 1.33 mg/mL in the hydrogel system exhibits extremely low viscosity (~5.3 mPa·s) which is close to water viscosity at low temperature, thus, leading to the reduced viscosity due to the dilution of entire hydrogel system.^[35] However, along with the increased starch concentration overpowering the influence on the overall material, the collagen influence on viscosity reduction of overall material is minimized. The incorporation of GNPs has enhanced the viscosity of starch-collagen blend in a concentration-dependent manner, as shown in Figure S3. With a GNP concentration of 1×10¹⁰ particles/mL, the formed nanocomposite starch-collagen hydrogel exhibits reinforced viscosity, which is comparable to the pure starch but with significantly enhanced cellular interaction sites introduced by collagen material.

Figure 3.

Rheological characterization of starch, the starch-collagen blend, and nanocomposite starch-based hydrogel during the extrusion condition at 6°C a) Shear rate sweep analysis of starch, the starch-collagen blend, and nanocomposite starch-based hydrogel. b) Frequency sweep indicates G' (closed symbol) and G" (open symbol) profiles of starch, the starch-collagen blend, and nanocomposite starch-based hydrogel at varied concentrations. c) Representative recovery performance and thixotropy profiles from tested material groups. (S: starch, C: collagen, G: gelatin nanoparticles. Number in front of each letter indicates the final concentration of each component. Starch concentration: 7.5%, 10%, 12.5% w/v; Collagen: 1.33mg/mL; Gelatin nanoparticle: 5G: 3.33×10⁹ and 10G: 6.67×10⁹ particles/mL)

Frequency sweep offers a well-defined comparison of viscoelastic properties under constant strain. As illustrated in Figure 3b, the storage modulus G' is larger than the loss modulus G" under the applied angular frequency in the entire frequency region for all the groups. Additionally, both G' and G" increased with increasing angular frequency. Although no cross-over point appeared in any group, G" and G' of samples with a starch concentration at 7.5% and 10% are converging at higher values of ω. This result implies the gradual loss in the elastic property in the tested gel-like materials. Overall, frequency sweep analysis cooperatively confirmed the significance of starch in maintaining the mechanical property of all the tested groups, as well as the beneficial effect from the collaboration of collagen and gelatin nanoparticles on manipulating the viscosity and mechanical strength of nanocomposite hydrogel, making it tunable to specific applications.

Thixotropy is a time-dependent shear-thinning behavior that is considered to be associated with the rheological instability of a material. By applying a certain shear rate, the weak physical bonds are mechanically disrupted, and the inner structures disintegrate to separate from aggregates. Such process is shear rate- and time-dependent. Therefore, thixotropy plays a pivotal role in determining the shape fidelity of the 3D-printed constructs. In this study, the thixotropic properties of the nanocomposite hydrogel samples were investigated through a three-phase measurement as shown in Figure 3c. At a given concentration of starch, starch-collagen blend, and the nanocomposite hydrogels all showed the similar recovery trend, in terms of time to reach equilibrium. An increased starch concentration could yield a longer recovery time. With a starch concentration at 10% and 12.5%, introducing GNPs significantly improved the recovery performance in comparison to the starch-collagen blend only, which led to a comparable recovery rate as the pure starch hydrogel. It is worth noting that the dissolving of starch granules could be incomplete at room temperature. The presence of the swollen but non-melted starch granules could cause a retarded and incomplete recovery. In this study we formed starch hydrogel in 45⁰C. Despite the relatively slow recovery, the nanocomposite starch hydrogel showed the shear-thinning and thixotropic property, which is highly desired for extrusion-based 3D bioprinting.

2.4. Printability and the structural integrity

When creating a 3D bioprinted construct, one of the most important considerations is the ability to maintain the desired shape during and after printing. The desired outcome from extruded material is to retain constant width and smooth edges in the shape of the extrusion path without bulging, thinning or breaking. As reported in the previous study, printability of the materials could be quantitatively investigated using equation (1).^[36]

$$Pr=\frac{\pi }{4}\cdot \frac{1}{C},\text{where}C=\frac{4\pi A}{L^2},$$ (1)

L is the perimeter of the enclosed area (labeled with red dash line in Figure s7), and A is area. Therefore, for a square, $C=\frac{\pi }{4}$; while for a circle, C=1. Correspondingly, the Pr for a square is 1, and for a circle is $\frac{\pi }{4}$. The larger the Pr value is, the better the printability is, which may indicate the less impairment on printing resolution and 3D stacking from reduced material viscosity and gelation degree. For an ideal gelation condition or perfect printability status, the interconnected channels of the constructs would demonstrate square shape, and the Pr value is 1. From our study in Figure s6, Pr value larger than 0.8 could show acceptable printability. From published literatures, Pr value within the range of 0.9–1.1 demonstrated satisfied filament morphology and mechanical stability of printed construct. ^{[37, 38]} The printability of the formulated hydrogels with the starch concentration at 12.5% was calculated, as shown in Figure 4a and Figure S7. Starch-collagen blends showed the circular enclosed area between the printed filaments, and thus a relatively lower Pr value (0.857 ± 0.008). Pure starch has the highest Pr value at 0.929 ± 0.032 and the thinnest filaments (as displayed in Table S1), demonstrating the superior 3D printability. Enhanced shape fidelity was observed with the supplement of GNPs, which is in good alignment with rheological investigation in Figure 3.

Figure 4.

a) Characterization of the printability of starch, the starch-collagen blend, and the nanocomposite starch hydrogel (n=6). b) The structural integrity of 3D printed hydrogels in PBS over a 14-day incubation (n=6). (S: starch, C: collagen, G: gelatin nanoparticles. Number in front of each letter indicates the final concentration of each component. Starch concentration: 12.5% w/v; Collagen: 1.33mg/mL; Gelatin nanoparticle: 5G: 3.33×10⁹ and 10G: 6.67×10⁹ particles/mL)

In addition to the 3D printability, the structural integrity and stability during culture process is also crucial. Figure 4b showed the representative images of all the tested material samples over a 14-day immersion in PBS. The 3D constructs from the pure 12.5% starch hydrogel were corrupted and dispersed within 24h of culture, which could be due to the incomplete gelatinization of starch granules. In contrast, other biomaterial samples blended with collagen and GNPs maintained the stable 3D structure. With the thermo-gelation of collagen taking place during the culture, the 3D constructs could further be stabilized for supporting long-term 3D cell culture and tissue growth.

2.5. Morphological microstructure characterization

The porosity plays a pivotal role in scaffolding biomaterials, which can facilitate mass transport associated nutrient and biological exchange for promoting 3D cell growth. We examined the microstructure of the starch, starch-collagen blend, and nanocomposite starch hydrogel using a cryo-scanning electron microscope (Cryo-SEM). ImageJ was used to analyze the Cryo-SEM images to determine the pore size distribution, porosity, and wall thickness within the hydrogel matrix. With a fixed starch concentration at 12.5%, we observed interconnected hierarchical porous structures in all the tested material samples (Figure 5 and Figure S4). This honeycomb-like porous structure implied the capacity to facilitate cell migration, proliferation, oxygen, and nutrient transport, as well as the potential applications for drug loading and release in void space.

Figure 5.

Cryo-SEM images present the hierarchical porous microstructures from a) pure starch, and hybridized hierarchical pores and fibrous sprouts from starch-collagen blend in b) and c). d) The SEM image showing the gelatin nanoparticles in HEPES buffer around 40 nm in size. The unique hydrogel network reinforced by GNPs with interlaced web-like structures was shown by Cryo-SEM imaging in different scales at 5 μm e), 2 μm f), 1 μm f), and 4 μm h). The pore size distribution and porosity from starch i), starch-collagen blend j), nanocomposite starch hydrogel k) were characterized using ImageJ analysis of Cryo-SEM images. m) represents the scaffold wall thickness from all three samples (n=100). (S: starch, C: collagen, G: gelatin nanoparticles. Number in front of each letter indicates the final concentration of each component. Starch concentration: 12.5% w/v; Collagen: 1.33mg/mL; Gelatin nanoparticle: 10G: 6.67×10⁹ particles/mL)

The pure starch hydrogel possesses a smooth surface and relatively thin walls throughout the entire matrix, along with a few starch fibers stretching out from the surface (Figure 5a). However, by adding collagen as the blend (Figure 5b and c), we observed significantly heterogeneous pore surface, with much larger hierarchical pores. The collagen fibers are readily identifiable and randomly stretching out of starch for supplying rich binding sites to promote the cellular attachment. Notably, the addition of collagen can enlarge the matrix pore size, as well as thicken the wall of hydrogel. As evidenced in Figure 5 j and m, both the pore size and wall thickness from the starch-collagen blend are nearly 2-fold higher than that from the pure starch hydrogel. Notably, by further incorporating GNPs into the starch-collagen blend, we observed a higher order of homogeneous microporous structure and abundant collagen fibers, although with the entire hydrogel network compacted in a smaller pore size (Figure 5e and k, and Figure S4). We suspect that GNPs can induce denser hydrogel network to reinforce the scaffolding mechanical strength, which is also supported by our rheological study in Figure 3. The loaded GNPs were found to be merged with the skeleton of the matrix (Figure 5h), as well as attaching to collagen fibers for forming an interlaced web-like structure (Figure 5f and g red arrows). The histogram of pore size distribution and porosity of all three samples were displayed in Figure 5 i, j, and k. Compared with the starch or starch-collagen hydrogels, by introducing the GNPs, the material pore structures are denser to reinforce the overall mechanical stiffness of nanocomposite hydrogel.

2.6. In vitro cell culture on 3D printed nanocomposite starch-based hydrogel scaffold

Current starch hydrogel has not been widely used as 3D bio-ink in tissue engineering yet, due to its inability for cell adhesion and migration and thermal instability during culture period. To further evaluate the potential of the developed nanocomposite starch-based hydrogel for serving as ideal bio-ink in tissue engineering, NIH 3T3 fibroblasts were seeded on the top of the 3D printed nanocomposite starch hydrogel in grid pattern. As mentioned obove, pure starch scaffolds were not suited for tissue culture due to its instability and rapid collapse during culture conditions. Here we used 3D-printed starch-collagen blend as a control group for 3D cell culture. Figure S5 showed the cell distribution after seeding for 3h at day 1 with consistent seeding density across all samples (seeding density ~10⁶ cells/mL). The majority of the cells were resting in the cavity of printed grid pattern. During culture, cells grow and start to attach, spread, and migrate on the surface of the scaffolds. As shown in Figure 6, on day 4, cells seeded on the material of starch-collagen-nanoparticle hydrogel exhibited excellent attachment and migration. In contrast, cells on the starch-collagen hydrogel mainly stayed at the interconnected channels, and only a few cells were observed on the edge of the scaffold, indicating limited migration. On day 7, as illustrated in Figure 6 i and j, the attached cells further stretched and elongated on the scaffold of starch-collagen-nanoparticle hydrogel and achieved better uniformly distribution on the nanocomposite scaffold with network formation (Figure 6f and h), while cells on starch-collagen scaffold showed poor migration (Figure 6b and d). Figure 6 i and j gave a better view of comparison to control group, regarding cell distribution, stretch, and elongation within our developed nanocomposite starch-based hydrogel. The cell metabolic activity analysis in Figure 6k quantitatively demonstrated a much faster cell proliferation rate with nearly 2-fold enhancement in our developed nanocomposite starch hydrogel. Collectively, the results highlighted the enhancement on cell attachment, elongation, and migration by using developed material blend, which could be ideal 3D tissue scaffolding for improving 3D cell growth.

Figure 6.

Evaluation of cellular response on 3D printed nanocomposite starch-based hydrogel scaffold Starch-collagen blend hydrogel scaffold serves as the control group. Limited cell growth on starch-collagen blend hydrogel scaffold on Day 4 (a and c) and Day 7 (b, d and i), compared with abundant cell attachment and spreading on 3D printed nanocomposite starch bio-ink on Day 4 (e and g) and Day 7 (f, h and j). k. Evaluation of cell proliferation on both scaffolds over a 7-day culture (n=6). (S: starch, C: collagen, G: gelatin nanoparticles. Number in front of each letter indicates the final concentration of each component. Starch concentration: 12.5% w/v; Collagen: 1.33mg/mL; Gelatin nanoparticle: 10G: 6.67×10⁹ particles/mL)

3. Conclusions

Starch is an appealing natural biopolymer for versatile tissue engineering applications, owing to its cost-effectiveness, scalable production, biocompatibility and biodegradability. However, due to the mechanical instability and the absence of cell binding sites, starch has not been widely used as the 3D bio-ink in the field of tissue scaffolding. In this work, we present a novel formulation of 3D printable starch-based hydrogel with desired biological property and improved stability by blending starch with collagen as well as the gelatin nanoparticles. Through the rheological characterization, this novel nanocomposite starch hydrogel exhibits highly desired shear-thinning and thixotropic property, as well as the mechanical strength suited in high-fidelity 3D printing. Gelatin nanoparticles also functioned as a rheological modifier, which can tune the mechanical strength over a wide range for tailoring to variable tissue engineering applications. By formulating starch with gelatin nanoparticles, the resulted bio-ink showed significantly enhanced mechanical property and enabled the preservation of shape fidelity, which compensated the loss of mechanical strength when blended with collagen. The unique homogeneous microporous structure with abundant collagen fibers and GNP interlaced web-like structure not only support the efficient mass transport, but also supplies rich attachment sites for promoting 3D cell growth, as evidenced by culturing fibroblast cells with fold increased proliferation rate. Due to the enhanced mechanical strength, the developed starch nanocomposite hydrogel scalffods also can maintain good 3D structural integrity with reduced degradation rate for being gradually replaced along with cell growing, and remodeled to match the complexity of real tissue microenvironment, which is critically needed for long-term tissue culture.

Although there are many scaffold biomaterials developed in the field of tissue engineering and regenerative medicine, current clinical translation is still hampered due to the concerns from the toxicity and biocompatibility. The chemical and mechanical properties of the biomaterial scaffold must be optimized to suite the interaction with cells and surrounding tissue microenvironment, including the efficient mass transport and host tissue integration. Our developed nanocomposite starch bio-ink has the ability to tune the mechanical strength by varying the concentration of each component, which offers great opportunity to tailor to the specific tissue environment. Such flexibility also allows the potential for incorporating live cells during the 3D extrusion printing process which is undertaking in our research lab as one of the future direction. Additionally, the developed nanocomposite starch is completely non-toxic and non-chemically modified, therefore, the starch nanocomposite is still highly biodegradable via converting carbohydrates back into forms that are usable for various biosynthetic and metabolic routes in vivo. We foresee this novel material platform holds great potential for clinical use and commercialization in regenerative medicine, as well as promoting biomimetic tissue engineering and research.

Experimental Section

Materials:

Commercial pregelatinized native common corn starch (Cargill Gel-Instant 12030) was gratefully supplied by Cargill, USA. Ham's F-12K medium, Dulbecco’s Phosphate Buffered Saline, heat inactivated fetal bovine serum, penicillin/streptomycin, Cell Tracker Red CMTPX, LIVE/DEAD™ Viability/ Cytotoxicity Kit, PrestoBlue™ Cell Viability Reagent, DAPI and Alexa Fluor™ 488 Phalloidin for cell culture and characterization were purchased from Thermo Fisher Scientific. Collagen Type I from rat tail was purchased from Advanced BioMatrix, Carlsbad, CA (Catalog #5279). Type A gelatin from porcine skin (300g Bloom), glycine, HEPES and 25% glutaraldehyde solution were obtained from Sigma-Aldrich, USA. Other chemicals including acetone, sodium hydroxide and hydrochloric acid (1N HCL solution) were purchased from Fisher Sci, USA.

Preparation of Gelatin Nanoparticles (GNPs):

Gelatin nanoparticles were prepared by a two-step coagulation process based on a previously reported study.^{[32, 39]} Briefly, 0.625g gelatin was dissolved in 12.5 mL DI water at 40°C under constant stirring for 1h. 12.5 mL acetone was introduced to the solution at a syringe pump-driven constant speed 0.1 mL/s under vigorous stirring (300 rpm). Thereafter, the solution was kept sitting for 15 min for the high molecular weight gelatin to precipitate, and the supernatant that containing low molecular weight gelatin was decanted. 7 mL DI water was subsequently added to redissolve the precipitated gelatin at 40°C. Upon complete dissolution, the pH of solution was brought down to a range of 2.7-3.0 using hydrochloric acid. 16.5 mL acetone was further added to the solution at a rate of 1 mL/min under constant stirring (600 rpm). 1mL acetone containing 60 μL 25% glutaraldehyde solution was gradually fed to the solution at a rate of 0.05 mL/min to crosslink the gelatin nanoparticles and the solution was stirred overnight. The reaction was deactivated by adding of 200 μL 1m Glycine solution. Acetone was then removed, and the solution was filtered through a 0.22 μm syringe filter. The achieved solution was stored in 4° C for future use.

Nanoparticle tracking analysis:

Gelatin nanoparticle’s size distribution and concentration were determined by Nanoparticle Tracking Analysis (NTA) using a NanoSight NS300 system (Malvern Technologies, Malvern, UK) configured with a 488 nm laser. Samples were diluted 50X in dPBS to an acceptable concentration. Samples were analyzed under constant flow conditions (flow rate = 50 uL/min) at 25°C with a camera level of 8. Data were analyzed using NTA 3.4 software with a detection threshold of 5.

Dynamic light scattering:

Dynamic light scattering allows for the characterization of particle size distributions with nanoparticles in various solutions through the analysis of the electrophoretic mobility. A Malvern Zetasizer (Malvern Panalytical Inc. 117 Flanders Road Westborough MA 01581-1042 United States) was used in conducting dynamic light scattering analysis of gelatin nanoparticles with concentration of 10⁹ particles/mL suspended in deionized water in 50-fold dilution. In preparing the sample for dynamic light scattering (Malvern Zetasizer), the scattering mode and angle was automatically selected as 90° inside scattering mode, with an equilibration time at 1 minute and temperature of 25°C, before measuring 5 runs per sample. The sample is prepared in the pH 7.2 solution, with the wavelength of incident laser at 633 nm. A cuvette loading cell was used and all results are listed with an average of 5 runs.

Scanning electron microscopy:

Field Emission-Scanning electron microscopy (FE-SEM) was used for visualizing and probing the gelatin nanoparticles’ morphology, topology, and size. Samples were first prepared through a serial dilution to a final concentration of 1×10⁹ particles/mL, followed with homogenization by light vortexing at 3000 rpm for 10 seconds, then directly aliquoted onto a 100% acetone-cleaned, Ted Pella aluminum pin stub mount (4595 Mountain Lakes Blvd, Redding, Ca 96003), and solution was allowed to evaporate completely. Immediately, the samples were sputter coated for 60 seconds with an Au/Pd target using a Denton Desk V Sputter Coater (1259 North Church St. Bldg 3 Mooretown, NJ USA 08057), and loaded into the Hitachi SU5000 Schottky Field-Emission Scanning Electron Microscope (20770 W. Nordhoff Street, Building 4, Chatsworth, CA 91311, United States) at a high negative vacuum pressure of 10⁻⁸ torr. An incident electron beam was applied onto the samples at 7 keV and beam current of 16.7 nA. Aperture and stigmata corrections were done before sample images were obtained.

Cytotoxicity of GNPs:

The cytotoxicity of GNPs was determined using PrestoBlue assay. Nanoparticles were suspended into culture medium to achieve final concentrations at 0, 10², 10⁴, 10⁶, 10⁸, 10¹⁰ particles/mL, respectively. Afterwards, NIH-3T3 cells were plated at a density of 5 ×10⁴ cells/well in 24 well plate for culturing with nanoparticle-loaded medium for 5 days. PrestoBlue reagent was added to the medium for a final concentration of 10% and incubated for 2h at 37 °C. The absorbance was measured using Cytation 5, BioTek plate reader (λ=570 nm). The intensity was then converted to cell numbers according to the plotted standard curve at each time point (1, 3, and 5d).

Preparation of nanocomposite starch-collagen hydrogel:

The starch powder was sterilized by UV radiation thrice before use. The GNPs in DI water were loaded to HEPES buffer to achieve 50 mmol HEPES buffer (1 X PBS and pH=7.4) containing a series of particle concentrations at 5×10⁹ and 10×10⁹ particles/mL. Starch was stirred into GNPs-containing HEPES buffer at 45°C with constant stirring at 400 rpm for 20 min using an overhead mixer. After degassing by centrifugating for 30 min at 3000 xg, the GNPs-containing starch gels were kept in 4°C fridge for 1h. Prior to mixing, collagen was neutralized with HEPES buffer in 2× PBS in a 1: 1 ratio. ^{[40, 41]} The buffer system is prepared by adding HEPES powder to PBS buffer to achieve a 50mM HEPES, pH=7.4. Briefly, the HEPES buffer was prepared by adding HEPES powder into 2× PBS and was subsequently mix with collagen at concentrations of either 8mg/mL or 10.9 mg/mL thoroughly to neutralize the collagen. The homogeneous nanoparticle/hydrogel composites were formed by mixing GNPs-loaded starch hydrogel and collagen through a three-way stop cock in a 2:1 ratio. Thus, nanoparticle/hydrogel composites with 6 different combinations (starch final concentrations at 7.5%, 10%, and 12.5% w/v; GNP final concentrations at 3.33×10⁹ and 6.67×10⁹ particles/mL; collagen at 1.33 mg/mL) were obtained (as shown in Table 1 with the labels of formulations used in the following experiments). The nanoparticle/hydrogel composites were kept in the fridge for 3h before use. Starch gels at 7.5% w/v, 10% w/v, 12.5% w/v in HEPES buffer and the starch-collagen hydrogel blends were prepared as the parallel control.

Table 1.

Composition of the samples and control groups and their corresponding labels.

	Samples	Control
		Pure starch	Starch-collagen blend
Starch (S)	7.5%, 10%, 12.5%	7.5%, 10%, 12.5%	7.5%, 10%, 12.5%
Collagen (C)	1.33mg/mL	/	1.33mg/mL
GNP (G)	5×109, 10×109	/	/
Sample formulations	7.5 S-1.33 C-5 G	7.5 S	7.5 S-1.33 C
	7.5 S-1.33 C-10 G	10 S	10 S-1.33 C
	10 S-1.33 C-5 G	12.5 S	12.5 S-1.33 C
	10 S-1.33 C-10 G
	12.5 S-1.33 C-5 G
	12.5 S-1.33 C-10 G

Rheological evaluation:

Rheological evaluation of the nanocomposite hydrogels was implemented on the MCR 702 MultiDrive rheometer (Anton Paar), using a 25 mm measuring plate. Strain sweeps, frequency sweeps and shear rate sweeps were carried out at 6°C. Strain sweeps in the range of 0.1%–100% at the frequencies of 1 Hz at 6°C were performed to identify the linear viscoelastic region (LVR) of the hydrogel samples, followed by frequency sweeps from 0.1 to 100 rad s⁻¹ at a constant 1% strain. Shear rate sweeps in a range of 0.1–500 s ⁻¹ was tested subsequently. The thixotropic property was investigated by exerting a shear rate at 0.1 s⁻¹ for 60s (before printing) to the material, followed by increasing the shear rate to 100 s⁻¹ and maintained for 10s (bio-inks being extruded through the needle tip causes the high shear rate), and finally, decreasing the shear rate to 0.1 s ⁻¹ for 60s (recovery). The viscosity over time was recorded.

3D printing and structural integrity evaluation:

All the 3D bioprinting were performed with the EnvisionTEC 3D Bioplotter^®. Material was loaded into a sterile 30cc cartridge and extruded through a tapered 25G needle tip (ID = 260 μm). The temperature of printing cartridge was maintained at 6°C during printing. Printing pressure and speed were manually optimized for each combination. Each construct contains 10 layers. 15 mm × 15 mm × 2 mm scaffolds in grid pattern were printed directly into 6-well culture plates. As proposed in the previous study,^[36] printability, which demonstrates the gelation of the bio-ink, could be calculated using the following equation:

$$Pr=\frac{L^2}{16A}$$ (2)

Where L is the perimeter of the pore, A means the area of the pore. To obtain the Pr value of each bio-ink, optical images of printed constructs were captured and analyzed in Image-J software. The perimeter and area of the pores were achieved through threshold adjustment (n = 4). To demonstrate the printability of the selected bio-ink, a human anatomical ear shape was printed. The structural integrity of the nanoparticle/hydrogel composites in DPBS were examined for a consecutive 14 days. 3D printed rectangular samples (15 mm × 15 mm × 2 mm) were prepared as well. Samples were observed at various time intervals. The experiment was conducted (N = 6) under identical conditions.

Morphological microstructure characterization:

Cryo-FE-SEM (CFE-SEM) was also used for visualizing and probing the pore structures, shapes, sizes and content for a series of hydrogel composites: Starch, Starch-Collagen, Starch-Collagen-GNPs and gelatin nanoparticles (GNPs). A 5 mm thick hydrogel (in the z-component) was flash-frozen within a tank of liquid nitrogen and immediately passed into the cryo-box attachment on the Hitachi FE-SEM. Fractionation within the cryo-box, by cleavage on the longitudinal plane, immediately proceeded and the sample was allowed to sublimate off the water under high negative pressure at 10⁻⁸ torr, analogous to lyophilization. Aperture and stigmata corrections were done before sample images were obtained. ImageJ was used to analyze the microstructures of the samples.

Cell culture and characterization:

NIH 3T3 mouse fibroblasts (American Type Culture Collection, Manassas, VA) were cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin in a humidified incubator at 37 °C with 5% CO₂. The cell culture medium was changed every 2-3 days, and the cells were harvested once they reached 80–90% confluence. Cells were labelled with cell tracker Red CMTPX for 30 min prior to seeding for the ease of visualization. For cell seeding, all the printed scaffolds were soaked in complete cell suspension for 3h before seeding. Cells were then prepared with a density at 5 × 10⁶ cells/mL, 40 μL cell suspension droplet was carefully placed onto the center of each scaffold surface, thereafter, the scaffold was incubated for 3-4h in a CO₂ incubator to allow the cell attachment to the scaffolds. 1 mL medium was gently added to the culture well plate after the cells were attached. Medium was changed every day.

Similarly, cell metabolic activity was examined with the PrestoBlue assay. Samples (n=6) were incubated with PrestoBlue at a ratio of 9:1 for 2h at day 1, 4 and 7. To assess cell attachment and morphology, printed constructs were stained for phalloidin (F-actin in green) and 4′,6-diamidino-2-phenylindole (DAPI; nuclei in blue) at day 4 and 7. Samples were fixed with 4% (v/v) paraformaldehyde for 30 min at room temperature, rinsed 3 times with PBS, then permeabilized with 0.2% (v/v) Triton X-100 for 30 min at room temperature. Afterwards, the samples were thoroughly rinsed 3 times. Blocking buffer containing 5% bovine serum albumin (BSA) and 0.1% (v/v) Triton X-100 in PBS was added and incubated for 1h. The samples were washed thrice after the removal of blocking buffer. Alexa Fluor™ 488 Phalloidin (0.8 U/mL) and DAPI in PBS were added to each sample and incubated for 2h and 30 min away from light to stain F-actin and nuclei, respectively.

The Cytation 5 Cell Imaging Multi-Mode Reader equipped with an inverted fluorescent microscope (widefield) was used for fluorescence cell imaging. Objectives with 4× and 20× magnifications were used. The fluorescence imaging channel was in the order of FITC followed with DAPI. Images were processed by Gene 5 software with stitching function to achieve Figure 6a, b, e and f. The scaffold material has very light autofluorescence background which is much lower compared to fluorescence signals from labeled cells as shown in Figure S8. By applying image contrast, cells can be easily differentiated from scaffold materials and their stretching morphology was presented in Figure 6 and Figure S8.

Statistical analysis:

All results are expressed as mean value ± standard deviation (SD) with number of replicates indicated elsewhere in the Material and methods or in the figure captions. The results are evaluated by one-way ANOVA analysis coupled with a post-hoc Tukey test. Differences are considered statistically significant when p ≤ 0.05 and greatly significant when p ≤ 0.001.