A Systematic Study of GelMA‐Carbopol Bioinks for High‐Fidelity Extrusion 3D Bioprinting at Physiological Temperatures

ABSTRACT

Extrusion 3D bioprinting allows depositing cells within hydrogels in well‐defined spatial patterns, facilitating the creation of tissue biomimetics. Gelatin methacrylate (GelMA) hydrogels are biocompatible, biodegradable, and promote cell adhesion, making them a common choice to formulate bioinks. However, the low viscosity of GelMA‐based inks makes it challenging to print at physiological temperatures. Typically, this is overcome with high concentrations (≥ 10%) of GelMA and rheological modifiers (≥ 1%), and using low temperatures, which negatively impact the printing process and cell function. This work develops high performance GelMA bioinks using Carbopol (CBP) as a rheology modifier. Inks containing low GelMA and CBP concentrations exhibit excellent printability at physiological temperatures. Complex constructs, including hollow structures with overhangs, were 3D printed with high shape fidelity. The inks show excellent cytocompatibility toward various cells, including primary human lung fibroblasts (HLF). HLF embedded in bioprinted structures exhibited outstanding viability and proliferated over 14 days of continuous culture. As an application, GelMA‐CBP inks were used to fabricate a stretchable lung tissue model incorporating HLFs, which was used to study fibroblast‐to‐myofibroblast transition. This work resolves limitations of traditional GelMA 3D bioprinting inks and lays the foundation for their use in advanced tissue biomimetics and regenerative medicine applications.

Gonzalez Martinez and collaborators develop a strategy to formulate high performance GelMA‐based bioinks with low solids contents. The resulting bioinks enable 3D bioprinting at 37 °C of high‐fidelity structures with tunable mechanical properties that support high cell viability and function. The new bioinks are validated through the fabrication of dynamic advanced lung tissue biomimetics.

1. Introduction

3D bioprinting is a technology with applications in critical biomedical fields like tissue engineering [1] and regenerative medicine [2]. 3D printed tissue biomimetics are attractive options for replacing traditional 2D culture and animal models [3, 4]. 3D printed tissues can reproduce the physical and biochemical environment of human tissue and better capture signaling pathways and drug responses [5]. In addition, these systems lend themselves to dynamic stimuli that mimic those experienced by cells during regular tissue function. For example, mechanical stimulation of lung cells through cyclic stretching of the lung during breathing is a crucial feature to consider when designing systems to mimic this tissue, this would result in model that more accurately mimic the lung than conventional static cultures performed in 2D [4, 6]. Thus, 3D bioprinting opens the door to the creation of dynamic biomimetic models that can better resemble the mechanical and biochemical microenvironment of in situ tissues [3].

Several technologies are used for 3D bioprinting, with inkjet [7, 8, 9], laser [10, 11, 12], and extrusion [13, 14, 15] bioprinting being the most popular [1]. In the latter approach, a bioink is placed in a container (e.g., cartridge, syringe) and extruded through a nozzle, producing filaments and allowing the fabrication of cell‐laden 3D structures in a layer‐by‐layer fashion [13, 14]. The bioinks used in extrusion‐based bioprinting are formulations containing hydrogels and embedded cells, and must meet specific requirements to simultaneously result in good printability and cell viability [16]. To form stable filaments and yield high structural integrity and shape fidelity structures, it is necessary to use bioinks with high viscosity, shear thinning behavior, and fast thixotropic recovery. The elevated viscosity and thixotropic behavior with rapid recovery avoid the formation of droplets after extruding the ink through the nozzle and cell sedimentation in the cartridge during printing, while the shear thinning behavior allows the extrusion to occur at low pressure, protecting cells from high shear stress [13, 17].

Gelatin is the material of choice for many tissue engineering and bioprinting applications due to its natural origin, biocompatibility, low immunogenicity, biodegradability, gelation properties, and low cost, along with the presence of RGD (Arg‐Gly‐Asp) domains that promote cell adhesion and function [18, 19]. Gelatin is fabricated through acidic (Type A) or basic (Type B) hydrolysis of collagen [18], which produce polypeptides with average molecular weights between 15 and 400 kDa that undergo reversible gelation at low temperatures (T_g = 20 °C–30 °C) [20]. Gelatin hydrogels are thermo‐reversible because the noncovalent interactions between the protein chains can be disrupted by increasing the temperature [20, 21, 22, 23]. Gelatin methacryloyl (GelMA) is a modified form of gelatin that can undergo rapid initiator mediated photo‐cross‐linking [24]. This permits its use in light‐assisted 3D bioprinting, resulting in hydrogels that are covalently cross‐linked and retain their shape independent of temperature, which is not the case for unmodified gelatin. GelMA retains the desirable properties of unmodified gelatin, which has made it a preferred material for bioink formulation [24, 25, 26]. Importantly, the hydrogels obtained from cross‐linked GelMA promote cell adhesion, growth, and proliferation [27].

Despite the numerous advantages that GelMA offers, there are challenges in direct 3D bioprinting of this material via extrusion. At low concentrations and/or physiological temperatures suitable for cell culture (i.e., 37 °C), GelMA solutions have low viscosity and cannot be successfully printed by themselves with extrusion printers. Therefore, low temperatures (to promote physical gelation) and high total solids content are needed to print constructs with structural stability and good shape fidelity. Yet, using a high concentration of GelMA during printing with bioinks containing embedded cells leads to lower cell viability, adhesion, and migration [27, 28, 29]. In addition, decreasing the temperature to improve the viscosity of the ink complicates the printing due to clogging of the extrusion nozzle, and tighter control over temperature is needed, as small fluctuations can lead to significant changes in viscosity that impact the pressure needed for optimal extrusion [30]. The need for higher GelMA concentrations and tight temperature control affect the printing reproducibility and complicate the optimization of printing workflows.

Another strategy to improve the viscosity and printability of GelMA‐based bioinks is to introduce materials that enhance their rheological properties. Examples of rheological modifiers commonly used are methylcellulose [26], alginate [31], hyaluronic acid [35], collagen [36], nanosilicate [38], gelatin [29], and gellan gum [41]. Despite the improvement of the printability with the inclusion of these modifiers, in most cases, it is necessary to use a high concentration of at least one of the components of the ink (modifier or GelMA), leading to a high total solids content in the bioink that can affect normal cell growth (Figure 1A). Furthermore, printing of GelMA using rheological modifiers is not usually done at a physiological temperature (37 °C), making the printing process more challenging and potentially impacting the cell function and viability (Figure 1B). Laponite and gellan gum are notable because low concentrations of these modifiers result in suitable viscosity for extrusion – printed constructs were reported from inks containing 0.5% gellan gum and 3% GelMA at 21 °C [42]. Another approach reported to decrease the GelMA content in the bioink is using gelatin as an additive and taking advantage of its thermal‐cross‐linking [29]. Using this strategy, a bioink containing 5% GelMA and 8% gelatin exhibited good printability, and the unmodified gelatin could be removed post‐printing by dissolving through a gradual increase of the temperature [29]. Although these approaches have reported some success, developing bioinks that contain low concentrations of GelMA and rheology modifier, and achieve high printability and cell viability at 37°C is still an outstanding challenge.

FIGURE 1.

Rheological modifiers, total percent solids, and temperatures used in the literature for GelMA 3D bioprinting through direct extrusion. (A) Graph of total solids vs rheological modifier content reported for GelMA 3D bioprinting. The total solids content is defined as the sum of the GelMA and modifier. (B) Temperatures used to 3D bioprint GelMA bioinks with the respective theological modifiers. Data extracted from References [25, 26, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43].

Recently, our lab developed a versatile strategy to formulate inks for high‐fidelity 3D printing of non‐extrudable polymers using low percentages of Carbopol (CBP) as a rheology modifier [44]. Carbopol ETD 2020 NF (ETD stands for easy to dissolve and NF refers to the US Pharmacopaeia grade assigned to this material) is classified as an interpolymer. Carbopol interpolymers are composed of carbomer homopolymers or copolymers that carry a block copolymer of polyethylene glycol and a long‐chain alkyl acid ester [45]. In this case, it is a copolymer of acrylic acid and an alkyl acrylate, partially crosslinked with allyl pentaerythritol ether [45, 46]. After neutralization with a base, such as NaOH, the carboxylic groups become deprotonated, resulting in swollen hydrogels that can increase by ≈1000 times their dry weight [44]. CBP ETD 2020 NF forms agglomerates of multiple primary particles, estimated at approximately 56 µm in diameter at low concentrations [47]. This “jammed microparticle” hydrogel structure can explain the rheological properties observed in bulk CBP hydrogels. Typically, the gels behave as elastic solids at rest and flow after a stress is applied (yield‐stress material) [48, 49]. The microparticles form units connected by weak hydrogen bonds that can be deformed and broken under a critical shear, permitting the material to flow and giving it its shear‐thinning properties [49]. After the shear is removed, the hydrogen bond network is reformed, leading to a fast thixotropic recovery of the original viscosity.

Our previous work showed that the formulation strategy works for biopolymers with different cross‐linking mechanisms and is amenable to the incorporation of cells and various additives, resulting in reproducible 3D printed structures with high cell viability. In this manuscript, we build on our previous work by developing and optimizing GelMA‐CBP bioinks for extrusion 3D bioprinting of tissue biomimetic constructs. These bioinks present many advantages, including excellent rheological properties for extrusion 3D printing (e.g., shear thinning and rapid thixotropic recovery), cytocompatibility, biodegradability, and the ability to 3D print cell‐laden structures (Figure 2). These new bioinks allow GelMA bioprinting at low concentrations, using low amounts of the rheological modifier, and printing at a temperature of 37 °C while maintaining outstanding printability and cell function, thereby filling a gap in the 3D bioprinting workflows for GelMA‐based bioinks. High cell viability, cell adhesion, proliferation and spreading were observed for A549 epithelial cells, HUVECs, and fibroblasts cultured on top and/or inside the newly formulated matrices. The inks were used to develop a stretchable lung model incorporating primary human lung fibroblasts to study fibroblast‐to‐myofibroblast transition. Our work improves three key issues of GelMA‐based bioinks for extrusion 3D bioprinting, by allowing the printing of complex constructs using i) low concentrations of GelMA, ii) low concentrations of additive, and iii) physiological temperatures. Our work lays the foundation for using 3D printable GelMA materials in tissue engineering, regenerative medicine, and implantable medical device applications.

FIGURE 2.

Advantages of GelMA‐CBP bioinks, the printing process and the cross‐linked materials developed in this work.

2. Materials and Method

2.1. Materials

Gelatin from porcine skin (Type A, ≈300 g Bloom), methacrylic anhydride, lithium phenyl‐2,4,6‐trimethylbenzoylphosphinate (LAP), and ninhydrin were purchased from Sigma‐Aldrich (Oakville, ON, Canada). Sodium hydroxide was purchased from EMD (Darmstadt, Germany). Carbopol EDT 2020 NF (Lubrizol) was kindly provided as a gift from IMCD Canada Limited (Brampton, ON, Canada). Hoechst 33342, anti‐alpha‐smooth muscle actin (anti‐αSMA) primary monoclonal antibody (1A4) conjugated to Alexa Fluor 488 (AB_2574461) or eFluor 660 (AB_2574362), phalloidin conjugated to Alexa Fluor 647 (A22287) or Alexa Fluor 488 (R37110) and MitoTracker Dyes for Mitochondria Labeling (M7512) were purchased from ThermoFisher (Toronto, ON, Canada). Alexa Fluor 555 Anti‐VE‐Cadherin antibody (EPR27436‐55, ab317067, Abcam, Waltham, MA, USA) was ordered from Abcam, and Alexa Fluor 647 anti‐mouse/human CD324 (E‐Cadherin) antibody (DECMA‐1, 147308, BioLegend, San Diego, CA, USA) was acquired from BioLegend.

2.2. Synthesis of GelMA

The introduction of methacrylamide groups in the gelatin was adapted from the method reported by Li [50]. First, 5 g of gelatin were dissolved in 45 mL of PBS at 50 °C under constant stirring. Then, 167 µL of methacrylic anhydride (94%) were added dropwise every 30 min for 2.5 h (6 additions, 1 mL total). After the last addition, the mixture was left to react for another 30 min. Next, the reaction was stopped by adding 100 mL of PBS, and the product was dialyzed against ultrapure water (18.2 mΩ·cm from a Reference A10+ Millipore System, Z00QSVC01 Millipore S.AS., Molsheim, Bas‐Rhib, France, subsequently referred to simply as water) for 7 days at 37 °C, with two daily water changes. Finally, the solution was freeze‐dried for 7 days. Five different GelMA batches were prepared and characterized to determine the degree of functionalization. The scheme describing the synthetic route and the characterization of the final product was shown in Figure S1.

2.3. Degree of Substitution via NMR Spectroscopy and Ninhydrin Test

¹H‐NMR spectra of unmodified gelatin and GelMA were obtained using a Bruker AV 500 spectrometer. The samples were dissolved in deuterium oxide and the experiment was performed at 35 °C. To obtain the degree of substitution (DoS), the spectra of gelatin and GelMA were normalized against the signal of phenylalanine (6.9 – 7.5 ppm). Then, the signal of lysine methylene groups in both spectra was integrated [51]. The calculation was done according to Equation (1).

$$DoS=\left(1-\frac{\int lysinemethyleneinGelMA}{\int lysinemethyleneingelatin}\right)\times 100$$ (1)

The ninhydrin assay was also performed in order to obtain the DoS of GelMA as the percent of modified amines, using the method reported by Zatorski et al. [52]. For this, a calibration curve of unmodified gelatin was obtained using concentrations between 1 and 10 mg/mL. The gelatin was dissolved in PBS and 87.5 µL of each solution were placed in a 96‐well plate. Then, 12.5 µL of a 20 mg/mL ninhydrin solution in ethanol was added to each sample. Next, the well plate was sealed with an optical sealing tape and was kept at 70 °C for 30 min. Finally, the absorbance was determined at 560 nm with a plate reader (Tecan Infinite M200, Männedorf, Switzerland). A solution of 10% GelMA was prepared in triplicate and read under the same conditions as the calibration curve. The DoS was determined as:

$$DoS=\left(1-\frac{C}{10}\right)\times 100$$ (2)

where 10 was the known concentration of the GelMA solution (10 mg/mL) and C was the concentration calculated from the standard curve.

2.4. Ink Formulation

Stock solutions of 1.5 wt% CBP, 15 wt% GelMA and 4 wt% LAP were used for the formulation of the inks (Table 1). The CBP gel was weighed in a 20 mL conical tube and centrifuged at 1050 g for 3 min. Next, the GelMA, water, and LAP were added, and the inks were heated for approximately 5 min at 60 °C and vortexed. Finally, the inks were centrifuged at 2900 g for 5 min to remove bubbles introduced during the mixing process.

TABLE 1.

Final composition of the GelMA‐CBP inks. All values are reported as wt%.

GelMA	CBP	LAP
1	0.5	0.5
3	0.25	0.5
3	0.5	0.5
5	0.1	0.5
5	0.1	0.05
5	0.25	0.5
5	0.5	0.5
7	0.25	0.5
7	0.5	0.5

2.5. Rheological Measurements

Rheological measurements on the inks and the cross‐linked materials were conducted in a TA rheometer (HR20, TA instruments, New Castle, DE, USA) using an 8 mm parallel plate geometry. Plots of viscosity of the ink as a function of the shear rate (0.1 to 100 Hz) were obtained to assess the shear thinning behavior. To determine the thixotropic behavior and fast recovery of the inks, they were exposed to a constant shear of 1 Hz for 15 s, then the shear was increased to 500 Hz for 10 s and finally decreased to 1 Hz for another 15 s. The storage (G') and loss (G″) modulus were measured through oscillation frequency sweeps (angular frequencies between 0.1 – 100 rad/s and shear strain of 0.6%). The values of the plateau region (between 1 and 10 rad/s) were averaged to obtain single G' and G″ values for each sample. The inks were then cross‐linked in a UV chamber (Anycubic Wash and Cure, Shenzhen, China, λ = 365 nm & 405 nm) for 4 min to form puck‐like constructs (10 mm diameter, 1 mm thickness) followed by measurement of the rheological properties of the cross‐linked materials. Compressive elastic moduli were obtained by compressing the cross‐linked samples 100 µm (10% strain) at a constant linear rate of 5 µm/s. The values reported were obtained from the slope of the linear region of the strain vs. stress curve. A minimum of three independently prepared replicate samples were measured per condition.

2.6. Attenuated Total Reflectance (ATR) Characterization

ATR spectra were collected in a Thermo Scientific Nicolet 6700 spectrometer in the 500–4000 cm⁻¹ range. Sterile 5%GelMA‐0.5%LAP and 5%GelMA‐0.5%CBP‐0.5%LAP biomaterial inks were exposed for 30 s to 405 nm light (15% intensity, ≈35 mW/cm²) to achieve cross‐linking. After this, the cross‐linked puck‐like constructs were incubated simulating cell culture conditions (cell culture media, 37°C, and 5% CO₂). Samples were collected and rinsed with water on days 0, 7, and 14 after cross‐linking and freeze‐dried before analysis. Controls of pure CBP hydrogel, LAP and uncross‐linked GelMA were also included.

2.7. Tissue Processing and Characterization

Storage and compressive modulus of samples of porcine brain, lung, liver, and heart were obtained as described above. Fresh tissues kindly donated from a local butcher (J. Waldron Butchers, Hamilton, ON, Canada) were sliced, cut into 8 mm puck‐like samples, and soaked in cell culture media for 1 h before the characterization. Similar data was acquired without submerging the tissue in cell culture media. At least ten replicate samples were prepared and measured per tissue type.

2.8. 3D printing and Printability

The inks described in Table 1 were printed using an Allevi 3 3D bioprinter (Allevi, Philadelphia, PA, USA). The inks were pre‐heated to 37°C, placed in a syringe and centrifuged at 2900 g for 5 min to remove bubbles before printing. The printing parameters used were shown in Table 2.

TABLE 2.

Parameters used to 3D print with the formulated inks.

Parameter	Value
Pressure (psi)	5–25
Printing speed (mm/s)	6–12
Wavelength (nm) a	405
Light intensity (%) b	15
Temperature (°C)	37
Needle	25G
Layer height (mm)	0.2

^a An additional experiment was conducted at a wavelength of 365 nm with 50% intensity.

^b 15% intensity Corresponds to a Power Density of ≈35 mW/cm².

Mesh‐like and cylinder constructs were printed to assess the printability and shape fidelity. The printability was evaluated from the mesh constructs using three printability metrics represented in Equations (3) (combined), (4) (circularity), and (5) (ratio of areas)

$$P=\frac{L^2}{16A_0}$$ (3)

$$P_C=\frac{L^2}{16A}$$ (4)

$$P_A=\frac{A}{A_0}$$ (5)

where L was the experimentally measured perimeter of one square, A was the measured square area,A₀ was the designed square area (taking into consideration the decrease in the area due to the experimentally measured printed filament width) [44]. The mesh constructs were designed with dimensions of 1.2 cm × 1.2 cm width, 2 mm height, and 3 mm × 3 mm internal squares. At least three 3D printed replicates were measured per condition. Next, the width of at least three single fibers extruded at printing speeds of 6, 10, and 15 mm/s was obtained to determine the influence of the printing velocity on the fiber width. Complex constructs were printed using the parameters described in Table 2. The images of the meshes and the fibers were obtained using a Nikon Eclipse Lv100 microscope (Nikon, Melville, NY, USA) with camera model pco.Panda sCMOS (pco., Kelheim, Germany), and were processed with ImageJ software [53] (National Institute of Health, Bethesda, MD, USA).

2.9. Cell Culture

3T3 fibroblasts (NIH/3T3 CRL‐1658, ATCC, USA), A549 epithelial lung cancer cells (A549 CCL‐185, ATCC, USA) and primary human lung fibroblasts (HLF) (PCS‐201‐013, ATCC, USA) were cultured in Dulbecco's Modified Eagle Medium (DMEM/F‐12, HEPES, no phenol red, 11039047, Thermo Fisher Scientific, Toronto, ON, Canada) with 10% FBS (A3840302, Thermo Fisher Scientific, Toronto, ON, Canada) and antibiotic‐antimycotic (A/A, 100 U/mL penicillin/100 µg/mL streptomycin, VWR International Co, Mississauga, ON, Canada). Human umbilical vein endothelial cells (HUVECs) (C0035C, Thermo Fisher Scientific, USA) were cultured in EGM‐2 Endothelial Cell Growth Medium‐2 BulletKit (CC‐3162, Lonza, Walkersville, MD, USA). All the cells were incubated at 37°C and 5% CO₂.

2.10. Cell Seeding on Top of Crosslinked Constructs

7%GelMA‐0.5%CBP‐0.5%LAP and 7%GelMA‐0.5%LAP biomaterial inks were cast into PDMS molds (10 mm width, 1 mm thick), fully crosslinked, soaked in PBS, and sterilized using UV radiation. Then, the samples were pre‐conditioned in the corresponding cell culture media for at least two hours before seeding A549 epithelial cancer cells, HUVECs, HLF, or 3T3 fibroblasts, at a density of 4.7 × 10⁴ cells/cm². The cells were incubated at 37°C, 5% CO₂, and the culture media was changed every 3 days.

2.11. Cell‐Laden Cast Constructs

The protocol used to obtain cell‐laden cast constructs was like that described for the formulation of the biomaterial inks with minor modifications. GelMA (sterilized by filtration using 0.45 µm filters at 60°C), CBP (sterilized using UV radiation) and LAP (sterilized by filtration using 0.2 µm filters) were mixed to form a pre‐ink. The pre‐ink was heated to 37 °C and weighed in a biosafety cabinet to maintain sterile conditions. Next, cells suspended in culture media were added (pre‐ink to cell culture media at 4:1 mass to mass ratio). Following manual mixing of the cells, the bioink was cast into sterile PDMS molds (10 mm width, 1 mm thick). After 30 s of exposure to 405 nm light (15% intensity, ≈35 mW/cm², using Allevi 3 bioprinter), the constructs were removed from the molds and submerged in cell culture media. The constructs were incubated at 37°C, 5% CO₂, and the culture media was changed every 3 days. Cell‐laden constructs were created with the following final bioink compositions: 5%GelMA‐0.5%CBP‐0.5%LAP‐5×10⁶ A549 cells/mL, 5%GelMA‐0.5%CBP‐0.5%LAP‐5×10⁶ 3T3 fibroblasts/mL, and 3%GelMA‐0.5%CBP‐0.5%LAP‐5×10⁶ HLF/mL. Control bioinks made from pure GelMA (5%GelMA‐0.5%LAP‐5×10⁶cells/mL) were included for all the tested cells.

2.12. 3D Bioprinting of Cell‐Laden Bioinks

The protocol for formulating the bioinks for 3D bioprinting was like that used for preparing the bioinks for cell‐laden cast constructs. After adding the cells, the bioink was heated to 37 °C, placed in a syringe and centrifuged at 500 g for 30 s. HLF (final concentration of 5×10⁶ cells/mL) were 3D bioprinted with an Allevi 3 bioprinter using a G25 needle or nozzle, into mesh constructs with dimensions of 8.8 mm length, 8.8 mm width, 1 mm height, with 2.2 mm × 2.2 mm internal square openings. The hydrogels were crosslinked using 15% 405 nm light with an exposure time of 5 s every two layers. Mesh constructs (9 mm × 9 mm width, 1 mm height, and 3 mm × 3 mm internal squares) containing 3T3 fibroblasts (final concentration of 5×10⁶ cells/mL) were 3D bioprinted in a Cellink BioX printer using a G25 nozzle. The samples were crosslinked using 100% 405 nm light with an exposure time of 5 s every two layers. The 3D printed constructs were incubated at 37°C and 5% CO₂, with culture media changes every 3 days.

2.13. Cell Viability and Imaging

The cell viability on top of cast GelMA‐CBP and pure GelMA puck‐like constructs was obtained on days 1, 3, and 5, while the cells embedded within cast constructs and 3D bioprinted constructs were assessed on days 1, 3, 7, and 14 using a live/dead assay (LIVE/DEAD Cell Imaging Kit, 488/570, Life Technologies Corp., CA, USA) following the manufacturer's instructions. Fluorescence images were taken with an inverted confocal microscope (Nikon A1R HD25, Nikon Canada, Mississauga, ON, Canada) using a 10x/0.45NA objective. Image J (National Institute of Health, Bethesda, MD, USA) software was used to process the images and calculate the cell viability.

2.14. Metabolic Activity Quantification

The cell metabolic activity on top of cast GelMA‐CBP and pure GelMA puck‐like constructs was obtained on days 1, 3, and 5, while the cells embedded within cast constructs and 3D bioprinted constructs were assessed on days 1, 3, 7, and 14 using a WST‐1 commercial kit (Cell Proliferation Reagent WST‐1, 5015944001, MilliporeSigma, Burlington, MA, USA) following the manufacturer's instructions.

2.15. Cell Fixation and Immunofluorescence Labeling

Constructs with cells on top or within were fixed and fluorescently labeled on day 5 or 14. The samples were washed with PBS three times for 5 min each. Then, 0.1% saponin in PBS was added for flash permeabilization, and the samples were incubated for 1 min at room temperature. The constructs were next kept in 4% paraformaldehyde (PFA) in PBS at 4°C for 15 min. The PFA was removed and three rinses with PBS were done for 5 min each. The samples were then incubated in a 3% BSA/0.1% saponin (7/3 v/v) solution at room temperature for 2 h. The solution was removed, and the respective conjugated primary antibody was added to each cell type. Anti‐αSMA primary antibody (1/100, v/v, 5 µg/mL) diluted in 1% BSA was used for HLF, while anti‐E‐cadherin (1/25, v/v, 20 µg/mL) and anti‐VE‐cadherin (1/25, v/v, 20 µg/mL) were used for A549 cells and HUVECs, respectively. The samples were incubated at 4°C for 24 h. Then, the solution was removed and washed three times with PBS for 5 min each. A solution of phalloidin‐Alexa Fluor 488 (two drops per mL) and Hoechst 33342 (two drops per mL) was added, and the samples were incubated for 2 h. Finally, the solution was removed, and the samples were washed with PBS three times for 5 min each.

Fluorescence images were acquired on an inverted confocal microscope (Nikon A1R HD25, Nikon Canada, Mississauga, ON, Canada) using a 40x/1.25NA water immersion objective (high resolution images), 20x/0.75NA objective (Z‐stack) or a 10x/0.45NA objective (Z‐stack).

2.16. Mitochondria Labeling

The mitochondria of 3T3 fibroblasts on top of 7%GelMA‐0.5%CBP‐0.5%LAP and 7%GelMA‐0.5%LAP crosslinked materials were labeled on day 5 post‐seeding. The MitoTracker kit was used following the manufacturer's instructions. Briefly, 50 µg of reagent (1 vial) was dissolved in 94 µL of DMSO to a final concentration of 1 mM. Then the appropriate amount of stock solution was employed to prepare a working solution of 200 nM in cell culture media. The samples were exposed to the MitoTracker‐containing cell culture media for 15 min and then washed three times with fresh media. After that, the samples were fixed and stained as previously described.

2.17. 3D Printed Dynamic Lung Tissue Model

Dog‐bone shape constructs with internal square openings of 1 mm side were printed using 3%GelMA‐0.5%CBP‐0.5%LAP and the conditions shown in Table 2. The samples were sterilized in a UV chamber and soaked in cell culture media for at least 4 h prior to cell seeding. Then, 2.5×10⁵ HLF were seeded on each construct and incubated (DMEM‐10% FBS‐1% A/A) at 37°C and 5% CO₂ for 4 days. The cells were then starved overnight with FBS‐free media (DMEM‐1% A/A) to reset their mechanical memory. Next, the samples were clamped in the CaT stretcher, and subjected to cyclic uniaxial stretching for 24 h at 20% strain and 0.2 Hz while submerged in DMEM‐2% FBS‐1% A/A. To model breathing without pauses, a single stretch cycle included a ramp up to the desired level of strain over a duration of 2.5 s that was followed by a ramp down to baseline level of strain over a duration of 2.5 s. Thereafter, the samples were fixed and labeled for αSMA and F‐actin quantification. Control samples incubated under static conditions were also analyzed.

2.18. αSMA and F‐actin Quantification

The center area of the dog‐bone shape constructs was cut out, fixed, and labeled as previously described. Fluorescence images were acquired on an upright confocal microscope (Nikon A1R HD25, Nikon Canada, Mississauga, ON, Canada) using a 10x/0.45NA objective for quantification images and a 25x/1.1NA objective (water immersion) for high‐resolution images. The channels were optimized independently based on the most intense samples to avoid pixel saturation. All the images from experimental and control samples were acquired using the same conditions on the microscope. Z‐stack images of the center area of the dog‐bone shape constructs were taken and the mean intensity per pixel of the cells was obtained using Image J software [53] (National Institute of Health, Bethesda, MD, USA).

2.19. Statistical Analysis and Graphics

Comparison between multiple groups was conducted using one‐way ANOVA with Tukey post‐hoc test and Bonferroni correction. A two‐tailed t‐test was employed for comparisons between two groups. A value of p < 0.05 was considered a statistically significant difference. Real Statistics Resource Pack for Excel was used to perform the statistical analyses. The software packages OriginPro (2018, OriginLab Corporation, Northampton, MA, USA) and GraphPad Prism 8.4.2 (GraphPad Software, Boston, MA, USA) were used to obtain the graphics.

3. Results and Discussion

3.1. Rheological Properties of GelMA‐CBP Biomaterial Inks

GelMA is a popular biomaterial in the tissue engineering and bioprinting fields because it is derived from a natural biopolymer, contains cell adhesion (RGD) motifs, is biodegradable, biocompatible, affordable, and the presence of methacrylate groups allows fast and controllable photo‐cross‐linking [24, 25, 26]. However, the direct extrusion of inks containing low concentrations of GelMA at physiological temperature is challenging without introducing high concentrations of rheological modifiers (Figure 1). In this work, we focused on developing a simple strategy to formulate GelMA‐based bioinks that overcomes these challenges while achieving high printability and cell viability. The first step was the formulation of GelMA‐CBP inks with low total solids concentrations that exhibited adequate rheological properties for extrusion 3D bioprinting at 37 °C. To do so, the concentration of the photoinitiator (LAP) was fixed at 0.5% and the contents of GelMA and CBP were varied (Figure 3A). As expected, the GelMA inks without CBP were liquid at 37 °C for all the tested concentrations. Although the viscosity significantly increased with the addition of 0.1% CBP, such inks were flowing liquids, inadequate for 3D extrusion bioprinting. However, when CBP percentages were increased to 0.25% and 0.5%, most of the formulations behaved as gels that did not flow after inverting the vials, becoming promising candidates for extrusion bioprinting. The only exception was the ink containing 3% GelMA and 0.25% CBP, where the viscosity was not high enough to prevent the gel from slowly flowing after inverting the vial. It is worth noting that 0.25% CPB dissolved in water behaves as a gel, however, when 0.5% LAP is included, the increase in the ionic strength partially disrupts the gel structure and causes it to flow (Figure 3A, panel 3, vial 1). Nevertheless, we observed that at 0.25% CBP, the inks containing GelMA showed a higher viscosity than CBP‐LAP alone and passed the inversion test. The fact that both GelMA‐0.5%LAP and 0.25%CBP‐0.5%LAP solutions are liquids that flow at 37 °C, but the GelMA‐0.25%CBP‐0.5%LAP inks were gels, suggests that there are GelMA‐CBP interactions that favor gelation. We hypothesize that the free amino groups of GelMA could electrostatically interact with the carboxylic acid groups present in the CBP microparticles, leading to an increase in the viscosity of the formulation prior to any photo‐initiated cross‐linking.

FIGURE 3.

Influence of GelMA and CBP concentration on the rheological properties of the inks and the mechanical properties of the cross‐linked materials. (A) Images of the inks obtained varying the percentage of GelMA (0% colorless, 3% blue, 5% red, and 7% green) and CBP (0, 0.1, 0.25, and 0.5%) using a constant 0.5% LAP. The 0%GelMA‐0%CBP treatment consisted of an aqueous solution of 0.5% LAP. (B) Storage (G′) and loss (G″) moduli of 3%, 5%, and 7% GelMA inks with 0.25% CBP. (C) G′ and G″ of 5% GelMA inks with 0.1, 0.25, and 0.5% CBP. (D) Viscosity vs. shear rate curves of 3%, 5%, and 7% GelMA inks with 0.25% CBP. (E) Viscosity vs. shear rate curves of 5% GelMA inks with 0.1, 0.25, and 0.5% CBP. (F) Thixotropy measurements for 3%, 5%, and 7% GelMA inks with 0.25% CBP. (G) Thixotropy measurements for 5% GelMA inks with 0.1%, 0.25%, and 0.5% CBP. (H) Image of puck‐like cast constructs made of 3%, 5%, and 7% GelMA without and with 0.25% CBP, and puck constructs made of 5% GelMA and 0, 0.1%, 0.25%, and 0.5% CBP. (I) G′ and G″ of cross‐linked pucks made from 3 (blue), 5 (red), and 7% (green) GelMA inks with 0.25% CBP. (J) G′ and G″ of cross‐linked pucks made from 5% GelMA inks with 0.1, 0.25, and 0.5% of CBP. All the experiments were conducted at 37 °C. Data is presented as the mean and standard deviation for n ≥ 3 independently prepared samples. Statistical significance is indicated by * for p < 0.05, ** for p < 0.01, and *** for p < 0.001. A two‐tailed t‐test was done for comparisons between two groups.

To elucidate the influence of GelMA and CBP concentrations on the rheological properties of the ink formulations, we conducted a set of experiments varying either the GelMA (Figure 3C,E,G) or CBP (Figure 3B,D,F) concentrations while fixing the other component. The storage (G′) and loss (G″) moduli of the inks showed a dependency on the concentration of either component (Figure 3B,C). Increasing the CBP and GelMA concentrations resulted in a proportional increase of the storage and loss moduli. The trend was more pronounced in the case of CBP, where G′ increased ≈60‐fold with a slight variation in the total concentration of solids (0.1%–0.5%), indicating that a minimum CBP concentration (i.e., 0.25%) is needed for the ink to present high viscosity and gel‐like behavior.

Two key requirements to develop a functional bioink for high fidelity extrusion bioprinting are: i) that the formulation presents a shear thinning behavior to significantly reduce the viscosity during the extrusion process, helping preserve the cell viability, and ii) that the formulation presents rapid thixotropic recovery to favor the formation of a stable filament after the ink is extruded through the nozzle [44]. The viscosity vs. shear rate curves for GelMA‐CBP formulations indicated that all the inks possessed shear thinning behavior at 37°C (Figure 3D,E). Increased shear rate (such as that experienced by the ink during extrusion through a nozzle) resulted in a reduction in viscosity, which produces more favorable conditions for cells embedded in a bioink during extrusion [17]. In addition, the viscosity vs. shear curves shifted to higher values as the concentrations of GelMA or CBP were increased, with a more significant influence of the latter (Figure 3D,E), indicating that the CBP concentration largely dictates the resulting viscosity of the inks. To assess the thixotropic behavior of the GelMA‐CBP inks, the process of extrusion through a nozzle was mimicked by alternating shear rates between 1 and 500 Hz (Figure 3F,G). We observed that the formulations exhibited the expected decrease in viscosity as the shear rate was increased and an instantaneous (< 1 s) and complete recovery after the high shear was removed, independent of the percentages of GelMA and CBP used (Figure 3F,G; Figure S2). These results show that although both GelMA and CBP play a role in the rheological properties of the inks, CBP is the dominant component given that minor variations of its concentration resulted in large changes in the rheological properties of the inks.

To evaluate the influence of temperature on the rheological properties of the inks, the shear thinning and thixotropy experiments were repeated at 22 °C (Figure S3). At this temperature, all the samples (even those with 0% CBP) produced well‐defined gels that passed the inversion test (Figure S3A), however, samples without CBP did not form robust filaments when extruded. The trends for G′, G″, viscosity, and thixotropic recovery were similar at both temperatures (Figure S3B–G). Higher concentrations of GelMA or CBP resulted in increased G′, G″ and viscosity while the inks also presented rapid thixotropic recovery. As expected, given the gelation (physical cross‐linking) of GelMA at 22 °C, when evaluated at this temperature the inks showed higher modulus and viscosity. While the GelMA‐CBP formulations could be extruded at 22 °C, printing with inks containing low concentrations of GelMA at 37°C is more relevant for cell‐laden bioinks. Therefore, the printing experiments going forward were all conducted at this temperature, which approximates human physiology. These results show that GelMA‐CBP inks present ideal rheological properties for bioink formulation for extrusion 3D printing at physiological temperatures.

3.2. GelMA and CBP Tune the Mechanical Properties of Cross‐Linked Constructs

Pucks cast from GelMA‐CBP inks were used to determine the influence of both components on the storage and loss moduli of the cross‐linked materials. Our data showed that the values of G′ and G″ directly correlated with the concentration of GelMA and CBP (Figure 3I–J). Changing the GelMA concentration from 3% to 7%, while keeping the CBP concentration at 0.25% allowed tuning the mechanical properties of the cross‐linked hydrogels across an order of magnitude, from G' values of 0.41 ± 0.05 to 4.7 ± 0.6 kPa, respectively (Figure 3I). These G′ values show that it is possible to bioprint constructs from GelMA‐CBP bioinks to mimic the bulk mechanical properties of a range of soft tissues (e.g., brain, or lung) [54]. In addition, we observed that G′ and G″ values plateaued, showing no statistically significant differences, for pucks made from inks containing 5% GelMA and CBP concentrations > 0.1% (Figure 3J). However, when CBP was absent, the G′ value was almost half that of the samples containing CBP, and the differences were statistically significant. We further compared the G' and G″ of cross‐linked materials measured at 22  C and 37°C (Figure S4), and observed no statistically significant differences between values, indicating a full cross‐linking of GelMA under the experimental conditions used. Unlike the inks, the physical gelation of GelMA does not play a significant role in the mechanical properties of the constructs after photo‐cross‐linking. Altogether these results suggest that, even at low concentrations, CBP plays a relevant role in the mechanical properties of the cross‐linked materials, and the mechanical properties do not change within the CBP concentration range tested. This can be explained by CBP serving the role of an agent that fills in the pores of the GelMA gels, leading to increased mechanical properties of the cross‐linked materials. This is supported by images of cross‐linked GelMA with and without CBP (Figure 3H), where it can be observed that the incorporation of CBP into the structure of the pucks drastically improved their transparency. This is relevant if cells incorporated within 3D printed constructs made from GelMA‐CBP bioinks need to be imaged using optical microscopy.

To understand the behavior of CBP within the cross‐linked materials over the time, pure 5%GelMA and 5%GelMA‐0.5%CBP samples were incubated for 14 days in conditions that simulated cell culture (cell culture media, 37 °C, 5% CO₂) and evaluated through FTIR. ATR‐FTIR spectra were collected on days 0 (just after cross‐linking), 7, and 14. We found that the signals of CBP were still present after 14 days of incubation, indicating that it remained entrapped within the GelMA network, which is in agreement with our previous results for PEGDA‐CBP hydrogels [44]. A more detailed analysis of the ATR results can be found in Figure S5. The fact that CBP does not leave the constructs after incubation is a limitation of these materials. However, the low concentrations used and the cytocompatibility [55] of CBP mitigate this issue. It is also necessary to consider that the present study was limited to 14 days, and in implantation scenarios, the time span is more prolonged. Further experiments are needed to fully understand the behavior of this rheology modifier over time and its impact on biocompatibility of the GelMA constructs, including implantation studies in animal models.

3.3. Printability of GelMA‐CBP Inks

The effect of GelMA and CBP concentrations on biomaterial ink printability at 37 °C was assessed by printing cylinder and mesh constructs (Figure 4). The cylinders were used to qualitatively evaluate the ability of the inks to produce structurally robust constructs, while the meshes were used to quantitatively assess the printability figure of merit for each ink (Figure 4A,B). Several approaches have been reported to determine the printability of a biomaterial ink [56, 57]. We used a previously described equation that combines printability based on circularity and printability based on the ratio of the measured area and the theoretical area of the openings in the printed construct [44]. More details and comparisons with other printability assessment methods are included in Figure S6. In general, all biomaterial inks printed with high shape fidelity, structural stability, and printability at physiological temperature. Figure 4C shows that, at a fixed 0.25% CBP, cylinder‐like constructs with 1 cm height were successfully printed at all the concentrations of GelMA tested. The cylinders were able to support their weight and did not collapse, indicating that the mechanical properties were sufficient to maintain structural stability. Furthermore, the 3D printed constructs could stretch and return to their original dimensions (Figure S7). This suggests that constructs made from these materials are suitable for dynamic tissue biomimetic constructs, e.g., for lung and muscle studies, where the stretching of a cell‐laden structure is desired. In addition, the thickness of the cylinder walls decreased with increasing GelMA concentrations, which implies that at lower concentrations of the cross‐linkable polymer the filaments can deform, resulting in broader structures that deviated from the designed shape, thinner, better‐defined, and more accurate constructs were obtained from the biomaterial inks with 7% GelMA. This observation was supported by the numerical values calculated for the printability parameter (Figure 4D). A significant increase in printability from 0.3 ± 0.1 to 0.72 ± 0.03 (with a value of 1 representing a perfect reproduction of the designed shape) was seen when the concentration of GelMA increased from 3% to 7%. This was expected, considering the improved rheological properties of the inks and the formation of a higher density cross‐linked polymer network after photo‐cross‐linking. Qualitatively, the images of the center square openings of the mesh structures show the influence of the GelMA concentration on printability, where better defined openings were obtained at higher GelMA concentrations. The squares obtained with 3% GelMA were poorly defined due to the significant deformation and broadening of the extruded filaments prior to cross‐linking, while higher GelMA contents produced less deformable fibers, resulting in well‐defined squares. The width of the extruded fiber was also correlated with the printing speed, with a tendency to decrease at higher speeds (Figure 4D). This result also highlights the importance of the printing speed on the fiber width and, therefore, on the shape fidelity of the bioprinted construct.

FIGURE 4.

Printability of GelMA‐CBP inks. (A) 3D model of the cylinder and a mesh printed with the GelMA‐CBP inks. Scale bars represent 5 mm. (B) Zoomed in view of a center square of the mesh and a schematic representation of the square under three possible printing conditions: ideal printing, over gelation, and filament deformation. The equation used to calculate the printability is included, where L is the measured perimeter of the printed square and A₀ its designed area, considering the fiber size. (C) 3D printed structures used for qualitative (cylinder) and quantitative (mesh) assessment of the printability of the inks with varying GelMA concentrations (3% – blue, 5% – red, and 7% – green) at a fixed CBP concentration (0.25%). (D) Printability values and fiber size (printing speeds of 6, 10, and 15 mm/s) of the inks with varying GelMA concentrations (3% – blue, 5% – red, and 7% – green) at a fixed CBP concentration (0.25%). Optical images of the center squares of the meshes and fibers obtained at the highest printing speed are shown for each ink. Scale bars represent 1 mm. Panels (E) and (F) are similar to (C) and (D) but for inks with varying CBP concentrations (0.1%, 0.25%, and 0.5%) at a fixed GelMA concentration (5%). All experiments were conducted at 37 °C, using 0.5% LAP, and cross‐linking with 15% 405 nm light. Data is presented as the mean and standard deviation for n ≥ 3 independently prepared samples. Statistically significant differences are indicated by * for p < 0.05, ** for p < 0.01, and *** for p < 0.001. Comparison between multiple groups was conducted using one‐way ANOVA with Tukey post‐hoc test and Bonferroni correction.

The influence of CBP on the printability of GelMA‐CBP inks was assessed at 0.1%, 0.25%, and 0.5% CBP, while keeping the concentration of GelMA at 5% (Figure 4E,F). The cylinders printed with 0.25% and 0.5% CBP showed good shape fidelity and structural integrity, and both inks had good printability values. On the other hand, we found that 0.1% CBP was not enough to print 5% GelMA under the experimental conditions used (15% 405 nm light and 0.5% LAP). The cylinder‐like construct collapsed, and it was not possible to print a mesh without a considerable deformation of the filaments. Therefore, the printability of 5%GelMA–0.1%CBP ink was assigned the value of 0. This was an expected result since this ink formulation flowed (Figure 3A) and formed droplets when extruded. Our data indicates that the influence of CBP concentration on printability was considerably stronger than that of GelMA. Varying the CBP concentration from 0.1 to 0.5% resulted in biomaterial inks that ranged from not being printable (0.1%) to having outstanding printability (0.5%). This trend was also observed in the rheological characterizations of the inks and confirmed that CBP concentration is critical to achieve good printability and shape fidelity through extrusion 3D bioprinting.

We further evaluated the influence of LAP concentration, and the intensity and wavelength of light used for cross‐linking on the ink printability (Figure S8). A 5%GelMA‐0.1%CBP ink with 0.05% LAP, 10‐fold lower concentration than the amount used in the inks described above, presented adequate viscosity for extrusion. This highlights the impact of the ionic strength on the properties of CBP‐based gels, given that the biomaterial ink with the same concentration of GelMA and CBP but 0.5% LAP flowed and was not extrudable (Figure 4E,F). Because LAP is an ionic compound, its concentration directly impacts the hydrogen bonding between CBP microparticles, which is critical for the formation of a gel. Although the formulation with low LAP content presented suitable viscosity for extrusion, the constructs obtained at 15% of 405 nm light (condition used for the printing described above) had very poor shape fidelity (Figure S8). The material did not cross‐link fast enough to retain the shape during printing, indicating that a higher concentration of LAP is needed under these irradiation conditions. However, when 50% intensity for 365 nm (UV) light was used, constructs with good shape fidelity and integrity were obtained (Figure S8). These results show that 0.1% CBP and 0.05% LAP can be used for the extrusion 3D printing of GelMA inks if high intensity UV light is employed. However, prolonged exposure to high intensity UV light would lower the viability of cells in the printed construct. Thus, we elected to keep LAP at 0.5% and used 0.5% CBP as the optimal concentration for the 3D printing of cell‐laden inks (bioinks) using low intensity 405 nm light and printing at 37°C.

3.4. 3D Printing of Complex Constructs

After demonstrating the high printability and the tunability of the mechanical properties of GelMA‐CBP constructs, we explored the ability to print complex structures and tested the limits of each formulation. A largely untapped and exciting application of 3D bioprinting is to use materials with properties resembling those of ECM to build structures, that mimic tissues and organs. To assess the suitability of our bioinks for this application, formulations composed of 0.5% CBP and 1%, 3%, 5%, or 7% GelMA were used to print structures with increasing levels of complexity at 37 °C. While 0.5% CBP concentration is not the minimum that can be used for each formulation, it is a good compromise because it guarantees high printability for this range of GelMA concentrations at a constant CBP content, which simplifies the analysis and allows direct comparisons between inks printed at 37 °C.

The first step was to print our standard cylinder and mesh where we found that all these formulations had acceptable shape fidelity and reproducible printability (Figure 5A; Figure S9). The ink with the lowest GelMA concentration (1%) was used to test the lower limit of total solids that provided adequate printability, since this ink had a water content of 98%. We observed that with this formulation it was possible to print mesh‐like constructs and achieve acceptable shape integrity and printability. In the case of the cylinder, the construct could be printed but deformed under its weight when we tried to handle it, showing that tall structures (1 cm) were not structurally sound and easily deformed. Nevertheless, the construct was stretchable (Figure S10), indicating that there was enough cross‐linking in the GelMA network to provide cohesion to the soft material and to allow printing of low‐profile structures.

FIGURE 5.

Printability of the inks with 1%, 3%, 5%, or 7% GelMA and 0.5% CBP. (A) Cylinder‐like and mesh‐like constructs printed with 1% (colorless), 3% (blue), 5% (red), and 7% (green) GelMA. Model and resulting structure of (B) maple leaf outline 3D printed with 1%GelMA‐0.5%CBP‐0.5%LAP ink, (C) ear 3D printed with 3%GelMA‐0.5%CBP‐0.5%LAP ink, (D) hollow nose 3D printed with 5%GelMA‐0.5%CBP‐0.5%LAP ink, and (E) ghost figuring 3D printed with 7%GelMA‐0.5%CBP‐0.5%LAP ink. Scale bars represent 5 mm. Comparison of the (F) storage, (G) loss, and (H) compressive moduli of the cross‐linked GelMA‐CBP inks with porcine brain, lung, liver, and heart tissues (shaded band represents the range between the minimum and maximum values measured). All printing was performed at 37 °C using 405 nm light. Data is presented as the mean and standard deviation for n ≥ 3 independently prepared samples. Statistically significant differences are indicated by * for p < 0.05, ** for p < 0.01, ** for p < 0.001, and *** for p < 0.0001. Comparison between multiple groups was conducted using one‐way ANOVA with Tukey post‐hoc test and Bonferroni correction.

After confirming the high printability of the four formulations selected, we printed a 5 mm‐tall maple leaf outline, an anatomically sized solid human ear, an anatomically sized hollow nose, and a hollow ghost figurine using 405 nm light and at 37 °C (Figure 5B–E). Increasingly more complex models could be printed as concentration of GelMA was increased, since this improved the robustness of the printed materials. The printed structures reproduced the models with high fidelity for the inks with GelMA contents between 3 and 7%. On the other hand, the maple leaf outline was the only structure accessible to the softest ink (1%GelMA‐0.5%CBP‐0.5%LAP, Figure 5B), where the final construct presented acceptable shape fidelity and stability. All other constructs printed with 3%, 5%, and 7% GelMA were structurally sound and could be handled without damage (Figure S11). The cross‐linked constructs showcase the ability of the biomaterial inks to print complex models that are challenging, or impossible, to achieve with other GelMA formulations by direct extrusion 3D printing. The human ear shape, printed with the 3% GelMA biomaterial ink, presented layers that are not fully supported by underlying layers, producing slight overhangs that lead to hollow spaces (Figure 5C). The nose, printed with the 5% GelMA biomaterial ink, increased the complexity by presenting a completely hollow dome‐like structure with soft angles (Figure 5D), while the hollow ghost figurine, printed with the 7% GelMA biomaterial ink, presented more pronounced angles and overhanging features (Figure 5E). Overall, these results demonstrate that complex structures can be printed with GelMA‐CBP inks and achieve outstanding printability and shape fidelity while maintaining the desired properties for bioprinting: low concentration of GelMA and additive, and printing at physiological temperature of 37 °C. The ability to print complex constructs can be critical when using these materials for tissue engineering applications.

3.5. Cross‐Linked GelMA‐CBP Materials Can Mimic the Bulk Mechanical Properties of Tissue

Recapitulating the mechanical properties of human tissue is key when developing biomimetic models for basic research and for drug development applications. To determine the potential of GelMA‐CBP inks to mimic tissue, we measured the bulk mechanical properties of porcine tissues and compared them with those of 1%–7% GelMA formulations. Fresh porcine brain, lung, liver, and heart were sliced and cut to form 8 mm puck‐like pieces that match the size of pucks cast from the GelMA‐CBP inks. The storage (G′), loss (G″) and compressive (E) modulus were obtained from the naïve tissue samples and after conditioning them in cell culture media (Figure S12). We found statistically significant differences between G′ and G′’ of the naïve and the media conditioned samples, with the conditioned tissues showing lower moduli than fresh naïve ones. In the case of the brain and liver, there were no differences in the values of E between naïve and conditioned samples. These differences point to interactions between the tissue of interest and the cell culture media components (e.g., proteins, divalent cations) that lead to apparent softening of soft tissues, where the extracellular matrix network swells or is not strong enough to overcome these treatments. In addition, we observed that the samples measured without a previous conditioning step showed a larger replicate to replicate variability, suggesting that the tissues could be drying during processing and data acquisition, yielding variable and progressively higher moduli. These are key observations that, together with the use of different techniques for the determination of the mechanical properties, could explain the high variability observed in the moduli reported in the literature for fresh tissues (the interested reader is referred to Reference [54], where the mechanical properties of human tissue as reported by different techniques are discussed). To ameliorate these issues and obtain the most reproducible and comparable data, we elected to do a conditioning step in cell culture media for the GelMA‐CBP materials prior to performing the mechanical property measurements.

Puck‐like constructs (8 mm diameter, 1 mm thickness) from 1%, 3%, 5%, and 7% GelMA inks were incubated in cell culture media for 1 h before shear and compression rheology data were acquired. A comparison between the bulk mechanical properties of samples with and without conditioning in cell culture media was also performed (Supplemental Information, Figure S13). The comparisons of the storage, loss, and compressive moduli for porcine tissues and the cross‐linked GelMA‐CBP materials are shown in Figure 5F–H (shaded bands present the range of values for the tissues measured). It can be observed that compressive moduli and storage moduli of the cross‐linked GelMA‐CBP constructs closely matched those of the porcine organs, indicating that they could be used for building biomimetic models. However, when a deeper analysis is conducted, it is evident that the GelMA‐CBP materials do not exactly match the three moduli simultaneously. For example, despite the match of the G′ and E values between the heart tissue and 7% GelMA, the range of values for G″ of the heart were well above that of any of the cross‐linked materials. Furthermore, most of the G″ values of the cross‐linked materials were lower than those of the tissues, indicating a limitation in capturing the viscous properties of porcine tissues. This mismatch of the viscoelastic properties between the GelMA‐CBP materials and the porcine tissues depending on the type of rheological measurement points to the importance of performing a complete analysis of the samples of interest and biomaterial inks when developing biomimetic models. It has been previously reported that tissues, including the brain [58], lung [59], liver [60], and heart [61], present viscoelastic properties that result in a time‐dependent mechanical response and stress‐relaxation in response to a deformation [62]. This means that, to mimic tissues accurately, it is necessary to use hydrogels with the appropriate viscoelastic characteristics. Covalently cross‐linked GelMA hydrogels are predominantly elastic [63, 64], meaning that despite their advantages, these materials require further enhancements to fully capture the mechanical properties of tissue.

3.6. Evaluation of the Cytocompatibility of GelMA‐CBP Materials

The development of bioinks and biomaterial inks with high printability and suitable cell compatibility is essential for advancing 3D cell culture, tissue models, and regenerative medicine. Bioinks are defined as formulations that contain embedded cells that can be directly 3D printed (“cells inside” approach), while biomaterial inks are used to 3D print acellular scaffolds designed to be seeded with cells in a subsequent step (“cells on top” approach) [3, 16]. Both strategies are valuable for specific use cases, for example, the study of epithelial and endothelial barriers formed on different substrates is usually conducted by seeding cells on top of pre‐fabricated constructs [65]. Conversely, cancer models and fibroblast culture are more relevant when cells are encapsulated within the printed material [66, 67]. In this section, we evaluate the advantages and limitations of GelMA‐CBP formulations when used as biomaterial inks, as well as their capacity to support the fabrication of cell‐laden constructs.

The 7%GelMA‐0.5%CBP‐0.5%LAP biomaterial ink was used to assess the behavior of different cell types seeded on top of the fully crosslinked material. Epithelial cells (A549), endothelial cells (HUVECs), and fibroblasts (3T3 murine fibroblasts and primary human lung fibroblasts (HLF)) were seeded at a density of 4.7 × 10⁴ cells/cm² on puck‐like constructs. Cell viability, morphology, and metabolic activity were evaluated on days 1, 3, and 5 post‐seeding (Figure 6). Cell viability, as assessed using a fluorescent Live/Dead assay and confocal microscopy, was > 95% across all time points for all the cells seeded on top of GelMA‐CBP and pure GelMA, with no statistically significant differences between them. In line with this, the observed cell spreading and the increase in metabolic activity confirmed that the addition of CBP did not compromise the cytocompatibility of the crosslinked material. A549 cells, HUVECs, and HLF exhibited no statistical differences in metabolic activity when seeded on GelMA‐CBP or pure GelMA constructs. Interestingly, 3T3 fibroblasts presented a higher proliferation rate on GelMA‐CBP compared to pure GelMA. This could be due to a lower number of cells attaching to the GelMA‐CBP puck during seeding, which would provide more available surface area for subsequent proliferation relative to day 1 (Figure 6D).

FIGURE 6.

Epithelial cells (A549), endothelial cells (HUVECs), primary human lung fibroblasts, and 3T3 fibroblasts cultured on top of puck‐like constructs. (A) Confocal images for live (green)/dead (magenta) assay, cell viability, and metabolic activity of A549 cells cultured on top of 7%GelMA‐0.5%CBP‐0.5%LAP and 7%GelMA‐0.5%LAP puck‐like constructs for 1, 3, and 5 days. Panels B, C, and D are similar to panel A for HUVECs, HLF, and 3T3 fibroblasts, respectively. Statistical differences are indicated by ** for p < 0.01. Comparison between multiple groups was conducted using one‐way ANOVA with Tukey post‐hoc test and Bonferroni correction. Data is presented as the mean and standard deviation for n ≥ 3 independently prepared samples.

To evaluate the morphology and phenotype of the seeded cells, samples were fixed and stained on day 5 and imaged using confocal microscopy (Figure 7). Cells cultured on GelMA‐CBP hydrogels exhibited morphologies similar to those observed on pure GelMA, confirming that the addition of CBP did not significantly modify the cell behavior. A549 epithelial cells showed typical cobblestone‐like morphology and were positive for E‐cadherin at cell–cell borders, indicating the establishment of adherens junctions and epithelial sheet formation (Figure 7A). Similarly, HUVECs expressed VE‐cadherin at intercellular contacts, consistent with endothelial barrier formation (Figure 7B). HLF presented elongated morphology with well‐defined F‐actin stress fibers and positively stained for αSMA, suggesting partial activation toward a myofibroblast phenotype, which can be associated with their interaction with the stiff hydrogels (Figure 7C). Finally, 3T3 fibroblasts presented their typical morphology, and mitochondrial staining highlighted robust mitochondrial networks, suggesting active metabolism (Figure 7D). These results demonstrated that GelMA‐CBP supports epithelial, endothelial, and fibroblast cell attachment and proliferation with typical cell morphologies and marker expression. Furthermore, the capacity to perform high‐resolution immunofluorescence and mitochondrial labeling of different cells seeded on these constructs demonstrated that GelMA‐CBP is compatible with confocal imaging, expanding the utility of the material for applications that require cell visualization.

FIGURE 7.

Fluorescence labeling of cells cultured for 5 days on 7%GelMA‐0.5%CBP‐0.5%LAP and 7%GelMA‐0.5%LAP puck‐like constructs. (A) Fluorescence labeling of the nucleus (DAPI – cyan), E‐cadherin (primary antibody – green), F‐actin (phalloidin – magenta), and merged image of all channels for A549 epithelial cells. (B) Fluorescence labeling of the nucleus (DAPI – cyan), VE‐cadherin (primary antibody – green), F‐actin (phalloidin – magenta), and merged image of all channels for HUVECs. (C) Fluorescence labeling of the nucleus (DAPI – cyan), αSMA (primary antibody – green), F‐actin (phalloidin – magenta), and merged image of all channels for HLF. (D) Fluorescence labeling of the nucleus (DAPI – cyan), mitochondria (MitoTracker – green), F‐actin (phalloidin – magenta), and merged image of all channels for 3T3 fibroblasts.

The ability of GelMA‐CBP hydrogels to support cell encapsulation was assessed using A549‐laden puck‐like constructs (Figure 8). Solutions containing A549 cells were mixed with GelMA‐CBP or pure GelMA pre‐inks (concentrated versions of the final inks) at a 1: 4 w/w cell suspension: ink ratio and a density of 5 × 10⁶ cells/mL, resulting in final bioinks of 5%GelMA‐0.5%CBP‐0.5%LAP or 5%GelMA‐0.5%LAP, respectively. Cell viability within both matrices remained consistently high (>95%) across all time points (Figure 8C). In addition, metabolic activity increased significantly from day 1 to day 3 before reaching a plateau (Figure 8D). The encapsulated cells tended to agglomerate and form spheroids that were positive for E‐cadherin immunostaining (Figure 8B). This behavior is consistent with prior reports of A549 cells cultured in GelMA hydrogels, validating the biological relevance of the system [68]. Interestingly, cells within GelMA‐CBP formed larger spheroids compared to those inside pure GelMA. This could be related to differences in the mechanical properties of the hydrogels or variations in the microenvironment due to the presence of CBP in the matrix. Z‐stack imaging of up to 200 µm depth confirmed a homogeneous distribution of spheroids throughout the constructs, demonstrating the structural integrity and uniformity of the encapsulation process (Figure 8E). These results establish GelMA‐CBP hydrogels as a robust platform for epithelial cancer cell encapsulation, maintaining high viability, good metabolic activity, and biologically relevant spheroid formation. The ability of GelMA‐CBP to promote spheroid growth showcases its potential for recapitulating tumor microenvironments in vitro. In addition, the outstanding printability of these formulations positions them as a promising tool for 3D bioprinting of cancer models, enabling future studies in a physiologically relevant context.

FIGURE 8.

A549‐laden puck constructs. (A) Confocal images for live (green)/dead (magenta) assay of the cells cultured for 1, 3, 7, and 14 days inside cross‐linked pucks made from 5% GelMA‐0.5%CBP‐0.5% LAP and 5% GelMA‐0.5% LAP bioinks. (B) Fluorescence labeling of the nucleus (DAPI – cyan), E‐cadherin (primary antibody – green), F‐actin (phalloidin – magenta), and merged image of all channels for A549 epithelial cells inside the pucks. (C) Cell viability over time of the A549 cells inside the pucks. (D) Metabolic activity of A549 cells cultured inside the pucks measured using the WST‐1 assay. (E) Z‐stack of fluorescently labeled (nucleus (DAPI – cyan) and F‐actin (phalloidin – magenta)) A549 cells inside the pucks on day 14. Fluorescence images were acquired using a confocal laser scanning microscope. Data is presented as the mean and standard deviation for n ≥ 3 independently prepared samples. Statistically significant differences are indicated by ** for p < 0.01. Comparison between multiple groups was conducted using one‐way ANOVA with Tukey post‐hoc test and Bonferroni correction.

It is broadly acknowledged that for a cell to thrive while encapsulated in a hydrogel, the 3D microenvironment must present specific attributes (e.g., porosity, rigidity, adhesiveness) that meet the specific type of cell's requirements. In this sense, while excellent results were obtained for A549 cell encapsulation using 5%GelMA‐0.5%CBP‐0.5%LAP as a matrix, we observed that this bioink was not entirely suitable for fibroblast encapsulation. Fibroblast viability was ≥ 90% over 14 days for both GelMA‐CBP and pure GelMA formulations (Figure S14B) and the LAP concentration used did not affect the cell viability at day 0 (just after cross‐linking the viability was ≈90%, Figure S15). Yet, cells in pure GelMA elongated and proliferated only in the layers that were close to the surface of the hydrogel and remained round in the bulk (Figure S14A, C), while the fibroblasts encapsulated in GelMA‐CBP, remained round independently of their position within the construct (Figure S14A, D). Cell elongation in the layers close to the surface has been previously reported in pure GelMA hydrogels for encapsulated keratocytes [69, 70] and human periodontal ligament fibroblasts [71]. These works attributed the round cell morphology in the bulk to insufficient access to nutrients and oxygen, which was explained by the limited porosity in the tight crosslinked network of the GelMA hydrogels. This argument has been supported by observations made on human dermal fibroblasts inside different GelMA concentrations [72, 73]. This work showed that fibroblasts encapsulated in 10% and 15% GelMA presented limited spreading and remained rounded within the bulk material, while those within 5% GelMA elongated and proliferated, albeit this was only true close to the surface. In the case of GelMA‐CBP hydrogels, the CBP microparticles act as fillers that reinforce the material's mechanical properties (as seen by the 2‐fold increase in G′ of GelMA‐CBP over GelMA alone) and can limit nutrient and waste diffusion. This significantly impacts the response of encapsulated fibroblasts but not that of other cells, like A549. The observed result could be explained by some cells being more tolerant to lower nutrient availability and an accumulation of waste, as well as having a higher capacity to remodel a stiffer hydrogel microenvironment [73]. Based on these results, it became evident that fibroblasts required a softer and more porous matrix to thrive while encapsulated.

To provide more favorable conditions for encapsulated fibroblasts, primary human lung fibroblasts (HLFs) were embedded within 3% GelMA‐0.5%CBP‐0.5% LAP and were compared with cells embedded in 3% GelMA‐0.5% LAP bioinks (Figure 9). HLFs presented high viability in either formulation, maintaining >97% viability in GelMA‐CBP and >92% viability in pure GelMA throughout the 14‐day experiment (Figure 9A). Similar to the behavior observed for A549 cells inside GelMA‐CBP, the metabolic activity of HLFs increased from day 1 to day 3 and then plateaued (Figure 9C). The cells in pure GelMA hydrogels elongated and rapidly organized into interconnected fibroblast networks spanning most of the hydrogel volume as early as day 1. Consequently, this immediate spreading was accompanied by only negligible changes in normalized metabolic activity. Such behavior can be associated with the very low stiffness, high porosity, and susceptibility to extensive cell remodeling of pure 3% GelMA hydrogels, which were extremely soft and difficult to handle. The soft and porous matrix facilitated nutrient and waste diffusion, allowing HLFs to reach their maximum proliferative potential within the first days of culture when encapsulated at a relatively high cell density (5 × 10⁶ cells/mL). This was reflected in a normalized metabolic activity that barely changed during the 14 days tested. Conversely, the cell encapsulated in 3%GelMA‐0.5%CBP encountered a soft but mechanically stronger matrix than pure 3%GelMA. This condition slowed initial spreading and proliferation, but over time led to a more pronounced increase in metabolic activity, indicating progressive remodelling and adaptation to the microenvironment.

FIGURE 9.

HLF‐laden puck constructs. (A) Confocal images for live (green)/dead (magenta) assay of the cells cultured for 1, 3, 7, and 14 days inside cross‐linked pucks made from 3% GelMA‐0.5%CBP‐0.5% LAP and 3% GelMA‐0.5% LAP bioinks. (B) Fluorescence labeling of the nucleus (DAPI – cyan), αSMA (primary antibody – green), F‐actin (phalloidin – magenta), and merged image of all channels for HLF cells inside the pucks. (C) Cell viability over time of the HLF inside the pucks. (D) Metabolic activity of HLF cultured inside the pucks measured using the WST‐1 assay. (E) Z‐stack of fluorescently labeled (nucleus (DAPI – cyan) and F‐actin (phalloidin – magenta)) HLF inside the pucks on day 14. Fluorescence images were acquired using a confocal laser scanning microscope. Data is presented as the mean and standard deviation for n ≥ 3 independently prepared samples. Statistically significant differences are indicated by **** for p < 0.0001. Comparison between multiple groups was conducted using one‐way ANOVA with Tukey post‐hoc test and Bonferroni correction.

The encapsulated HLFs were labeled and imaged using confocal microscopy (Figure 9B). Their fibroblast‐like morphology and the presence of well‐defined F‐actin stress fibers were confirmed in both hydrogel types. In addition, the cells were positive for αSMA staining, with the protein mostly diffuse within the cytoplasm. This contrasts with the behavior observed when HLF cells were seeded on top of 7%GelMA‐0.5%CBP‐0.5%LAP, where αSMA organized into stress fibers was observed in most of the cells. Diffuse cytoplasmic αSMA is commonly observed, even if the matrix used is soft, when HLF are cultured for extended number of passages on hard tissue culture plastic (TCP) due to their mechanical memory. During the activation of fibroblasts, they transition to a proto‐myofibroblast phenotype characterized by the expression of αSMA in the cytoplasm. During the maturation process, this protein organizes into stress fibers, which are typically observed in diseases like idiopathic pulmonary fibrosis (IPF) [74]. Our results point out the capacity of GelMA‐CBP hydrogels to support HLF encapsulation, survival, and proliferation throughout the bulk hydrogel with homogeneous distribution (Figure 9E), while maintaining αSMA expression at low, diffuse levels. This phenotype suggests that 3%GelMA‐CBP can provide a microenvironment that mimics aspects of a healthy lung tissue, offering a promising platform to study fibroblast activation and to investigate how biochemical cues, such as TGF‐β, drive the transition toward more fibrotic, disease‐relevant states.

3.7. 3D Bioprinting of HLF‐Laden and 3T3‐Laden GelMA‐CBP Bioink

After demonstrating the capacity of GelMA‐CBP materials to support cell encapsulation, we next assessed their performance in a 3D bioprinting workflow using 3T3 fibroblasts and HLFs. We first used murine 3T3 fibroblasts incorporated within 3%GelMA‐0.5%CBP‐0.5%LAP, 5%GelMA‐0.5%CBP‐0.5%LAP, and 7%GelMA‐0.5%CBP‐0.5%LAP inks to test the hypothesis that the stiffness of GelMA‐CBP hydrogels plays a major role in the behavior of these cells after encapsulation via 3D bioprinting. Mesh structures of 9 mm x 9 mm width, 3 mm x 3 mm internal squares, and 1 mm height (five layers, each 0.2 mm thick) were fabricated at physiological temperature (37°C) using a G25 nozzle on a Cellink BioX bioprinter. Crosslinking was achieved by exposing the constructs to 405 nm light for 5 s every two layers (Figure S16). High cell viability was observed across all conditions from day 1, with 3%GelMA‐0.5%CBP‐0.5%LAP and 5%GelMA‐0.5%CBP‐0.5%LAP showing higher viability than 7%GelMA‐0.5%CBP‐0.5%LAP at this time point, consistent with the higher GelMA concentration in the latter formulation. Cell morphology evolved distinctly according to the stiffness of the matrix. In softer matrices (3%GelMA‐0.5%CBP‐0.5%LAP), cells elongated and spread rapidly, reaching confluency by day 7 (Figure S16A). Fibroblasts within the 5%GelMA‐0.5%CBP‐0.5%LAP formulation exhibited a mixed morphology at day 7, with both rounded and elongated cells distributed throughout the construct. In contrast, cells within the stiffest condition (7%GelMA‐0.5%CBP‐0.5%LAP) remained largely rounded by this time point. However, by day 14 all hydrogels displayed the typical elongated fibroblast morphology. The spreading and proliferation were markedly delayed in stiffer hydrogels, which highlights the critical impact of the total solid content of the bioink on cell behavior, likely mediated by differences in nutrient, oxygen, and waste diffusion.

The bioprinted 3T3 fibroblasts were fixed, stained, and imaged within the GelMA‐CBP constructs (Figure S17). The 3T3 fibroblasts displayed a marked preference for spreading, proliferating, and/or migrating near the hydrogel surface (Figures S18–S20). This behavior was similar to that observed in 3T3 fibroblasts encapsulated within pure 5%GelMA pucks and was consistent with improved mass transport at the construct periphery. It is worth noting that 3T3 fibroblasts encapsulated in bulk 5%GelMA‐0.5%CBP hydrogels remained rounded throughout the 14‐day culture, whereas, when they were bioprinted with the same formulation, they exhibited spreading and elongation over time. We attribute the differences in cell morphology outcome to the size of the bioprinted constructs, where the width of the meshes are in the sub‐millimeter range, while the bulk puck hydrogels are millimeters in diameter and thickness. These differences impact the transport of nutrients within the 3D microenvironments as well as change the proximity of the cells to an external boundary “surface” that could promote cell elongation. This is supported by literature reports that show that 3T3 fibroblasts display more elongated morphologies within micropatterned GelMA constructs than within a bulk hydrogel [75]. These results reveal a clear advantage of using thin 3D bioprinted scaffolds with sub‐millimeter features over mold cast materials for 3D cell culture experiments and biomimetic models.

Next, leveraging the learnings from the experiments with 3T3 fibroblasts, we 3D bioprinted primary HLF embedded in 3%GelMA‐0.5%CBP‐0.5%LAP bioink. Mesh structures similar to those described above were fabricated at physiological temperature (37°C) using a G25 needle on an Allevi 3 bioprinter (highlighting the applicability of the printing methodology to different bioprinters). Crosslinking was achieved by exposing the constructs to 405 nm light for 5 s every two layers (Figure 10). The bioprinted HLF showed excellent post‐printing viability, with values of ≈94% on day 1, increasing to >99% by day 7 (Figure 10A,B). In contrast to the observations of 3T3 fibroblasts, the morphological analysis of HLF showed cells that were initially mostly round, with only a few of them showing elongation at day 1. Over time, the cells progressively transitioned into fully elongated, spread fibroblasts that formed interconnected networks by day 14 (Figure 10A). In addition, metabolic activity was initially low on day 1, likely reflecting the limited number of cells in the small bioprinted constructs and the stress from the printing process (Figure 10C). However, robust proliferation and fibroblast spreading led to a ≈45‐fold increase in metabolic activity by day 14, highlighting the ability of HLF cells to remodel GelMA‐CBP bioprinted matrices and thrive within them. Consistent with encapsulation results, bioprinted fibroblasts developed well‐defined F‐actin stress fibers, while αSMA expression remained low and diffusely distributed in the cytoplasm (Figure 10D). Finally, confocal Z‐stack imaging confirmed that HLF cells were homogeneously distributed throughout the hydrogel volume, forming dense, interconnected fibroblast networks by day 14 (Figure 10E). The excellent viability and spreading of HLF inside 3%GelMA‐0.5%CBP‐0.5%LAP bioink was confirmed in an additional experiment performed employing G25 nozzles instead of needles, demonstrating the versatility of the new formulations to be bioprinted under a variety of conditions (Figure S21). Overall, these results establish GelMA‐CBP as a versatile bioink capable of supporting both high‐resolution scaffold fabrication and robust 3D bioprinting under physiological conditions. The ability to achieve high printing fidelity and adequate cell behavior at 37 °C expands the range of applications beyond conventional GelMA systems. Collectively, these findings position GelMA‐CBP as a next‐generation biofabrication platform that combines outstanding printability, high cytocompatibility, and structural versatility, offering clear advantages over existing GelMA‐based inks for 3D bioprinting for tissue engineering and regenerative medicine.

FIGURE 10.

3D bioprinted HLF. (A) Confocal images for live (green)/dead (magenta) assay of the 3D bioprinted cells cultured for 1, 3, 7, and 14 days inside mesh‐like constructs made from 3% GelMA‐0.5%CBP‐0.5% LAP. (B) Cell viability over 14 days of the 3D bioprinted HLF. (C) Normalized metabolic activity over 14 days of the 3D bioprinted HLF measured using WST‐1 assay. (D) Fluorescence labeling of the nucleus (DAPI – cyan), αSMA (primary antibody – green), F‐actin (phalloidin – magenta), and merged image of all channels. (E) Z‐stack of fluorescently labeled (nucleus (DAPI – cyan) and F‐actin (phalloidin – magenta)) 3D bioprinted HLF on day 14. Fluorescence images were acquired using a confocal laser scanning microscope. Data is presented as the mean and standard deviation for n ≥ 3 independently prepared samples. Comparison between multiple groups was conducted using one‐way ANOVA with Tukey post‐hoc test and Bonferroni correction.

3.8. Development of a 3D Printed Dynamic Lung Tissue Model

Many organs in the human body, like the heart or lungs, are constantly subjected to cyclic mechanical stress. Thus, 3D printing bioinks that result in stretchable hydrogel structures are attractive for dynamic tissue modelling. GelMA‐CBP inks show tunable mechanical properties and excellent stretchability, opening the door to building dynamic tissue biomimetic constructs. To assess GelMA‐CBP constructs as lung tissue model that can undergo cyclic mechanical stress and result in relevant biological cell readouts, we 3D printed dog‐bone meshes using 3%GelMA‐0.5%CBP‐0.5%LAP inks. The cross‐linked structures made from the 3% GelMA‐CBP biomaterial ink had mechanical properties that recapitulated the storage and compression moduli (G' and E) of porcine lung tissue, making it the best candidate to mimic the lung tissue (cf. Figure 5F–H). Primary human lung fibroblasts were seeded on top of the dog‐bone structures and cultured for five days under static conditions before starting cyclic stretching. A cell and tissue (CaT) stretcher [6] was used to subject the cells to uniaxial stretching for 24 h (Figure 11A). The stretching amplitude and frequency were chosen to mimic a strained human breathing process.

FIGURE 11.

Cyclic stretching of HLFs on 3D printed constructs made from 3%GelMA‐0.5%CBP‐0.5%LAP biomaterial inks. The cells were stretched for 24 h using a sawtooth pattern with 0% to 20% strain at 0.2 Hz to mimic the process of breathing. (A) Schematic representation of the seeding, stretching, and imaging process. (B) Fluorescence images of the nucleus (DAPI – cyan), αSMA (primary antibody – green), F‐actin (phalloidin – magenta), and a merged image of all channels for HLFs cultured under static and stretched conditions. (C) Zoomed‐in image of static and stretched cells, the images present the merged fluorescent channels. Quantification of (D) αSMA and (E) F‐actin expression from HLFs cultured under static and cyclic stretch conditions. Statistically significant differences are indicated with * for p < 0.05. A two‐tailed t‐test was employed for comparisons between two groups. Data is presented as the mean and standard deviation for n ≥ 3 independently prepared samples.

The expression of alpha‐smooth muscle actin (αSMA), an indicator of fibroblast‐to‐myofibroblast transition, and filamentous actin (F‐actin) were quantified by fluorescence staining and confocal imaging. Intensity values were extracted from the images through background subtraction and thresholding (details on the image processing steps are presented in Figure S22). HLF grown under static or stretching conditions showed the expected spread out and narrow elongated morphologies typical of lung fibroblasts (Figure 11B,C). Additionally, the cyclic stretching induced a statistically significant increase in αSMA expression, indicating a fibroblast‐to‐myofibroblast transition, which is seen during the genesis of lung fibrosis (Figure 11D). On the other hand, the F‐actin expression was not changed by exposure to cyclic stretching (Figure 11E). Overall, these results show that the 3D printed GelMA‐CBP models could mimic the lung dynamic environment and produce relevant biological readouts. The protocols developed further open the door to research focused on studying the mechanism of action of existing fibrosis treatments and the screening of new drug candidates that can slow down or revert the progression of fibrosis.

The differences observed in the αSMA expression, the high printability, and tunable mechanical properties of the GelMA‐CBP bioinks are promising for the development of dynamic tissue models. However, further research is needed to fabricate complex models that more accurately reflect the microenvironment in the tissue. Follow up research studying the impact of the time, frequency, strain, stretching pattern, and uniaxial vs volumetric stretching is paramount to understand how cells respond to constant mechanical stimulus on top of these 3D printed materials and select the conditions that replicate in vivo biology. Other important factors to consider for future studies are how the dimensionality of the microenvironment (2D – on top vs. 3D – within the material) and the presence of soluble biochemical cues (e.g., cytokines and growth factors such as TGF‐β1) impact cells. Including lung epithelial cells on top of the material and HLF within it and exposing the co‐culture to immune cells and biomolecules of interest could dramatically improve the accuracy of the biomimetic lung model. Refining the mechanical properties of the constructs is also essential, since a limitation of the current model is that the viscous component of tissue is not fully recapitulated by the inks presented. Using more viscoelastic (e.g., double network) instead of mostly elastic hydrogels (i.e., covalently cross‐linked) may be key to mimic the mechanical properties of the extracellular matrix in humans. This additional work is crucial for developing fully functional dynamic tissue models that can be used in lieu of animal models. We anticipate that the GelMA bioinks can be refined to overcome all these limitations and help answer critical questions related to ageing and disease tissue states.

4. Conclusions

GelMA‐CBP biomaterial inks and bioinks were formulated, showing suitable rheological properties for extrusion 3D bioprinting at low total solids concentrations at a physiologically relevant temperature of 37 °C. The biomaterial inks presented both shear thinning and thixotropic behavior with fast and complete recovery, two properties that are key to facilitating extrusion 3D bioprinting and high cell viability. The rheological properties of the biomaterial inks were determined by the concentration of both GelMA and CBP, with the latter having a greater impact. CBP not only acts as a rheological modifier but also improves the mechanical properties of the cross‐linked material, which was reflected in an increase in the G' values compared to GelMA without CBP. The concentration of both GelMA and CBP affected the printability. Small variations in CBP content (from 0.1% to 0.5%) resulted in large improvements in printability, with 0.5% being the optimal concentration for the LAP concentration used. It was possible to print complex constructs with high shape fidelity at low concentrations of GelMA (1%–7%) and CBP (0.5%) at 37 °C. The newly developed GelMA‐CBP inks presented excellent cytocompatibility as evaluated using A549 epithelial cancer cells, HUVECs, HLF, and 3T3fibroblasts. It was possible to culture, fluorescently label, and analyze through confocal microscopy different cell types with adequate morphology, spreading, and metabolic activity. In addition, the incorporation of CBP into the GelMA network did not affect the capacity of the hydrogels to support cell encapsulation, showing that embedded A549 and HLF cells had an excellent response to the materials over 14 days. HLF were successfully 3D bioprinted with outstanding cell proliferation and spreading over the assessed time points. In addition, the bioinks allowed dynamic tissue modelling. Specifically, fibroblast‐to‐myofibroblast transition was demonstrated in a lung tissue model after 24 h of stretching of HLF. The addition of epithelial and immune cells and interrogating the impact of the type of stretching (uniaxial vs volumetric) are necessary next steps to model the lung more accurately. In closing, our work solves the three key issues of GelMA‐based bioinks for extrusion 3D bioprinting, by allowing the printing of complex constructs using: i) low concentrations of GelMA, ii) low concentrations of additive, and iii) physiological temperatures. Our work lays the foundation for using GelMA materials in tissue engineering, regenerative medicine, and implantable medical device applications.

Conflicts of Interest

The authors have nothing to report.

Supporting information

Supporting File: adhm71248‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.