Application of instant assembly of collagen to bioprint cardiac tissues

Abstract

Advancing cardiac tissue engineering requires innovative fabrication techniques, including 3D bioprinting and tissue maturation, to enable the generation of new muscle for repairing or replacing damaged heart tissue. Recent advances in tissue engineering have highlighted the need for rapid, high-resolution bioprinting methods that preserve cell viability and maintain structural fidelity. Traditional collagen-based bioinks gel slowly, limiting their use in bioprinting. Here, we implement TRACE (tunable rapid assembly of collagenous elements), a macromolecular crowding-driven bioprinting technique that enables the immediate gelation of collagen bioinks infused with cells. This overcomes the need for extended incubation, allowing for direct bioprinting of engineered cardiac tissues with high fidelity. Unlike methods that rely on high-concentration acidic collagen or fibrin for gelation, TRACE achieves rapid bioink stabilization without altering the biochemical composition. This ensures greater versatility in bioink selection while maintaining functional tissue outcomes. Additionally, agarose slurry provides stable structural support, preventing tissue collapse while allowing nutrient diffusion. This approach better preserves complex tissue geometries during culture than gelatin-based support baths or polydimethylsiloxane (PDMS) molds. Our results demonstrate that TRACE enables the bioprinting of structurally stable cardiac tissues with high resolution. By supporting the fabrication of biomimetic tissues, TRACE represents a promising advancement in bioprinting cardiac models and other engineered tissues.

INTRODUCTION

Advancing cardiac tissue engineering requires innovative fabrication techniques, including 3D bioprinting and tissue maturation, to enable the generation of new muscle for repairing or replacing damaged heart tissue.¹ Recent advances in tissue engineering have highlighted the need for rapid, high-resolution bioprinting methods that preserve cell viability and maintain complex structural fidelity.^2–4 Traditional collagen-based bioinks are limited by slow gelation speeds, which can result in shape diffusion during bioprinting. We have recently developed TRACE (tunable rapid assembly of collagenous elements), which is a novel technique that uses macromolecular crowding (MMC) to enable the rapid assembly of collagen and other extracellular matrix (ECM) proteins without restrictive conditions, offering greater versatility and preserving native biochemistry.⁴ TRACE enables the fabrication of collagenous tissues, including the use of neutralized collagen solutions of wide-ranging concentrations as bioinks, without the typical requirement for an extended incubation period for collagen gelation.⁴ In this study, we more extensively study how TRACE can be used to study cardiac tissue development. Compared to the gelation methods used in the recent and seminal bioprinting technique FRESH (freeform reversible embedding of suspended hydrogels),^3,5 which relies on either highly concentrated acidic collagen solutions or a cell–fibrin mixture to enhance gelation, our technique leverages macromolecular crowding to rapidly solidify collagen-based bioinks. This approach not only ensures precise printing fidelity but also supports the development of functional tissues. Moreover, we demonstrate that TRACE-bioprinted cardiac tissues maintain greater structural stability and complexity immediately after printing and throughout extended culture periods. This stability is achieved through the use of an agarose slurry, which offers critical mechanical support while enabling nutrient flow through the culture media.⁶ Such structural reinforcement is essential to prevent deformation or collapse of cell-laden bioprinted constructs. By overcoming these challenges, our method enables the bioprinting of more complex geometries that would be otherwise difficult to achieve with gelatin-based support baths or polydimethylsiloxane (PDMS) molds.

Collagen and basement membrane proteins are essential components of cardiac tissues,⁷ providing structural integrity and physicochemical signals. 3D bioprinting-based methods offer better structural control than traditional engineered heart tissue (EHT) fabrication methods that rely on molds.^1,8,9 Achieving effective polymerization is critical for maintaining structural fidelity in freeform 3D bioprinting applications.¹ A recent prominent bioprinting technique, FRESH, minimizes bioink gelation time by either using a high-concentration collagen bioink with a pH-neutralizing support bath⁵ or employing fibrin, a fast gelling hydrogel, as the bioink.³ However, the first approach, which relies on the fast gelation property of a high-concentration (24 mg/ml) collagen bioink, poses challenges for direct cell inclusion due to both the high concentration and acidity of the collagen.⁵ In the second approach, fibrin—while it facilitates gelation and is biodegradable, enabling its gradual degradation in extended culture—primarily appears in native cardiac tissues during injury or wound healing,¹⁰ making it less suitable for replicating the native cardiac environment.³ In another recent study, hyaluronic acid (HA) was added to collagen to thicken the bioink and reduce shape diffusion during the 30-min incubation required for collagen to polymerize. After testing various HA and collagen concentrations, a combination of 3 mg/ml HA and 3 mg/ml collagen was selected to minimize shape diffusion while also considering other necessary parameters for printability.² These studies highlight the need for an instant gelation method in bioprinting that supports versatile bioink compositions and concentrations while allowing direct cell inclusion. TRACE addresses these limitations by leveraging molecular crowding to instantly assemble solubilized ECM bioinks. Collagen and basement membrane proteins, which typically require extended incubation, gel instantly at various concentrations in the TRACE support bath.⁴ TRACE allows us to directly print cell-laden bioinks that are composed of collagen and basement membrane proteins and mimic the native extracellular matrix (ECM) composition of cardiac tissues. We show that stem cell-derived cardiomyocytes survive the rapid gelation process and that the printed tissues remain functional in long-term culture. As an additional benefit, TRACE uses agarose slurry to provide long-term structural support, preventing collapse (e.g., in tubular structures) while allowing media to diffuse through, ensuring sustained nutrient exchange during culture. In contrast, gelatin-based support baths melt at 37 °C, making them unsuitable for extended applications where stability is required.^2,5

Additional training of the cardiac tissue is necessary to promote further maturation.^11,12 Previous studies have demonstrated that isometric constraints can improve contractility.¹³ Our study hypothesizes that fixing both ends of a cardiac strip or inserting a PDMS post in the center of a cardiac ring could promote tissue maturation. A deeper understanding of the cardiac maturation of a ring or tube structure could lead to future applications such as self-pumping vasculature.^14,15

In this study, we implement TRACE, our novel macromolecular crowding technique, to bioprint engineered cardiac tissues. Our study starts by verifying that fibroblasts and stem cell-derived cardiomyocytes can survive through the macromolecular crowding support bath and remain functional in a long-term culture. Our study investigates printing both tissue strips and ring formats for scalable applications. To investigate maturation stimulation, we apply internal PDMS posts as rigid boundaries for tissue rings and characterize maturation signals, including tissue contractility and alignment. Finally, we print cardiac tubes and demonstrate that by leaving agarose slurry inside, we could fabricate fluid-pumping cardiac tubes, a high aspect ratio structure that could be prone to collapsing. We anticipate that TRACE, with its capacity for instant collagen gelation, will serve as a competitive bioprinting method capable of fabricating advanced tissues.

RESULTS

Fabrication of macroscopic tissue strips using instant collagen gelation

Our instant gelation method is based on the effect of macromolecular crowding (MMC), where polyethylene glycol 8000 (PEG 8000) enables instant gelation of collagen [Fig. 1(a)]. For ease of fabrication, we set off this study using molds to make tissue strips [Fig. 1(b)]. 200 mg/ml PEG 8000 solution was first pipetted to fill a PDMS multi-channel device [Fig. 1(b)]. Then, collagen solutions were pipetted to the channels. After a 2-min wait, collagen gel would polymerize. Once the gel polymerized, we used a P200 pipette to flush the channels with 1× PBS, transferring the strips into a Petri dish [Fig. 1(b)]. Following this procedure, we fabricated acellular tissue strips [Fig. 1(c)]. We prefilled the channels with 200 mg/ml PEG 8000 and then 3 mg/ml collagen solution and then waited for polymerization to take place [Fig. 1(d)]. After a 2-min wait, we used a P200 pipette to flush the channels with 1× PBS, transferring the strips into a Petri dish containing 1× PBS [Fig. 1(e)]. We tested a range of channel widths; we fabricated tissue strips in channel widths of 1, 1.5, and 2 mm [Figs. 1(d) and 1(e)]. In subsequent experiments that involved cell-laden tissue strips, 1 mm channels were selected for use.

FIG. 1.

Fabrication of macroscopic tissue strips using instant gelation. (a) Schematic depicting the rapid assembly of collagen by the macromolecular crowding (MMC) agent PEG 8000. We chose PEG 8000 due to its high molecular weight (MW). The schematic is adapted from a previous publication.⁴ (b) Procedure to fabricate tissue strips. 200 mg/ml PEG 8000 solution was pipetted to fill the channels of the PDMS multi-channel device. We then pipetted collagen solutions into the PDMS device channels and waited 2 min for the gel to polymerize. Once the gel polymerized, we used a P200 pipette to flush the channels with 1× PBS, transferring the strips into a Petri dish. (c) Timeline to fabricate and observe acellular strips. (d) When we pipetted 3 mg/ml collagen, which contained red beads for visibility, to the channels prefilled with PEG 8000, the collagen solution polymerized rapidly. We fabricated tissue strips in channel widths of 1 mm (left two channels), 1.5 mm (middle two channels), and 2 mm (right two channels). Subsequent cell-laden tissue strips were fabricated in the 1 mm-wide channels. (e) After 2 min, the polymerized strips were flushed from each channel with 1× PBS using a P200 pipette and transferred into a Petri dish containing 1× PBS. The acellular tissue strips contained microscopic red beads for improved visibility; however, no red beads were used in cell-laden bioinks. (f) Timeline to fabricate the fibroblast strips. On day 0, fibroblast tissue strips were fabricated using a process similar to that of acellular tissue strips. A 3 mg/ml collagen solution containing 300 000 fibroblasts/ml was pipetted into channels prefilled with 200 mg/ml PEG 8000. After 2 min, the polymerized strips were flushed out by injecting a stream of PBS into each channel using a pipette and transferring them into a Petri dish containing 1× PBS. New steps that are worth noting are denoted in blue. (g) and (h) Fibroblast strips were then mounted on a PDMS platform and secured at both ends with fibrin gels. The hollowed-out circular well in the center of the PDMS cutout allows easy observation of the tissue strip without light obstruction. The image in (g) was taken 1 day after the fibroblast strips were fabricated and installed. (i) The progression of a fibroblast strip over 8 days showed that the fibroblasts were healthy after having gone through the instant gelation. Figures (a) and (g) were made using Adobe Illustrator, and (b) used Adobe Illustrator and Biorender.

Following similar steps, we fabricated fibroblast tissue strips [Fig. 1(f)]. Our instant gelation method allows collagen solutions that contain cells to be directly polymerized rapidly, enabling cell-laden tissues to be produced. As such, cells were directly incorporated into the collagen solution. When we pipetted 3 mg/ml collagen containing 300,000 cells/ml of fibroblasts into channels prefilled with 200 mg/ml PEG 8000, the bioink polymerized rapidly. After a 2-min wait, we used a P200 pipette to flush the channels with cell culture media, transferring the strips into a Petri dish containing the media. Each strip was then transferred onto a PDMS platform, where both ends were secured with fibrin, leaving the middle free to move [Figs. 1(g) and 1(h)]. The PDMS platform consisted of a center well that was hollowed out, enabling observation on a microscope without light obstruction [Figs. 1(g) and 1(h)]. Within a week of culturing, the cells aligned, and the tissue became thinner and tauter [Fig. 1(i)].

Cardiac tissue strips are functional but tend to break

We prepared cardiac bioinks and filled them into the PDMS multichannel device prefilled with PEG 8000 solution to start the instant gelation process [Fig. 2(a)]. When human-induced pluripotent stem cell (hiPSC)-derived cardiomyocyte concentration in these bioinks was low, which we tested at 4 × 10⁶ cardiomyocytes/ml, the cardiomyocytes could not connect or lead to synchronous beating [Fig. 2(b) and supplementary Movie 1]. When the hiPSC-derived cardiomyocyte concentration was high, which we tested at 4 × 10⁷ cardiomyocytes/ml, the tissue was dense where cells were connected and were able to beat synchronously [Figs. 2(b) and 2(c) and supplementary Movie 2]. However, cardiac tissue strips tended to break during culturing. Out of ten strips (in two independent repeats) successfully anchored on our PDMS setup, two broke within 1 day of culture, and the rest eight broke within 5 days. The strips broke (ripped apart) likely due to high contractile force and repetitive chafing against the PDMS platform, leading to low output numbers [Fig. 2(d)].

FIG. 2.

Fabrication of cardiac strips using instant gelation. (a) Timeline to fabricate the cardiac strips. Cardiac strips were molded using the multichannel PDMS device as in Fig. 1(b). (b) After 3 days of culturing, the strips that had a low concentration of cardiomyocytes showed no signs of cardiomyocytes being in contact with one another or synchronous beating. The inset shows that cells were not connected and were in isolation. In the inset, the round cells were likely cardiomyocytes, and the spindle-shaped cells were likely HCFs. (c) The cardiac strips printed from bioink composed of 4 × 10⁷ cardiomyocytes/ml, however, showed tissue-level synchronous beating on day 3. The kymograph indicates spontaneous contractions (red arrowheads). (d) The cardiac strips were prone to breaking. The images were taken on day 8.

TRACE printing of functional cardiac tissue rings

Next, we move on to bioprinting to create tissues. The rings were fabricated using the TRACE-enabled printing method rather than molding [Figs. 3(a)–3(c)]. We printed cardiac rings in a support bath that consisted of an optimized agarose slurry and PEG solution [Figs. 3(a), 3(d), and 3(e)]. On day 14, the cardiac rings were able to beat spontaneously [Fig. 3(f)]. When they were electrically stimulated at 1 or 2 Hz, they beat at the stimulated frequencies [Fig. 3(f)]. When we observed the outer perimeter of these rings, we could see aligned sarcomeres [Fig. 3(g)].

FIG. 3.

TRACE (tunable rapid assembly of collagenous elements) printing of cardiac rings. (a) Schematic of the TRACE printing bath, consisting of a mixture of PEG 8000 and agarose slurry, with concentrations optimized to improve printing outcomes. The schematic is adapted from a previous publication.⁴ (b) Design of the construct. Notably, the ring width is determined by the printed fiber width, which in turn depends on both the bioink properties and the flow rate. Our parameters generated cardiac strips that were approximately 0.7 mm in width. (c) Timeline to fabricate the cardiac rings. We only removed the support bath that was outside of the ring on day 0. The slurry inside of the ring prevented the ring from collapsing and was later removed on day 2. New steps that are worth noting are denoted in blue. (d) The extrusion-based bioprinting setup: a glass syringe was filled with the bioink, and the syringe holder had a thermoelectric cooler that prevented gelation of the bioink before being extruded. 24-Well plates were filled with our TRACE printing bath to hold the constructs during printing. (e) After printing, we washed away the external support bath and refreshed it with culture media, leaving the ring and slurry inside intact and ready for culturing. (f) We characterized the contraction and calcium signaling of TRACE-printed cardiac tissue during spontaneous beating and under 1 and 2 Hz stimulation. The kymograph analyses illustrated the dynamic changes of the calcium transient over time. The calcium signaling experiment was conducted on cardiac rings that had been cultured for 14 days after printing. (g) The outer perimeter of the cardiac ring showed aligned sarcomeres. The cardiac ring was cultured for 14 days before fixation and staining.

Constraint training enhances the contractility of the cardiac rings

Furthermore, as a biological characterization of the TRACE-printed cardiac tissues, we compared the cardiac rings that were cultured around a post or without the post [Figs. 4(a) and 4(b)]. Cardiac rings that were cultured with a post could be in contact with the post within several days, whereas the rings that were cultured without the PDMS post would keep on shrinking [Figs. 4(c) and 4(d)]. The constraint training with the PDMS post enhanced the contractility of the tissue, as indicated by a higher percentage of area contraction, which was measured 2 days after the rings had been taken off the PDMS post [Figs. 4(e) and 4(f)]. Compared to free-floating cardiac rings, cardiomyocytes were better aligned on the outer perimeter of the rings that underwent constraint training [Fig. 4(g)].

FIG. 4.

Constraint training enhances the contractility of the cardiac rings. (a) Design of the construct. As the ring contracted and shrank during culture, the cardiac tissue conformed to the PDMS post. (b) Timeline for culturing cardiac rings with PDMS post as a constraint. New steps are denoted in blue. (c) In the absence of PDMS posts, the central void of cardiac rings progressively shrank during culture. The inset schematic shows the respective culturing condition. (d) Quantification of cardiac rings cultured with and without the presence of a PDMS post. The dotted line shows the size of the PDMS post. For control, sample size n = 18; for constrained, n = 22; experimental repeats N = 3. (e) The relative positions of the inner ring during contraction are labeled for both diastole and systole. (f) Quantification of the area change in the inner ring during contraction cycles was measured as the percentage change in the area between systole and diastole. For control, n = 9; for constrained, n = 10; N = 2. (g) Phalloidin staining of the outer perimeter of the rings shows the orientation of the cells without or with PDMS post-constraint.

Agarose slurry in the TRACE printing bath supports the cardiac tube structure

We further TRACE-printed cardiac tubes [Fig. 5(a)]. Of note, these cardiac tubes were not cultured on PDMS posts, and instead, the agarose slurry was kept inside of the tubes on the first 2 days as structural support [Fig. 5(b)]. Both the 5 and 10 mm tubes were able to retain their geometry over a week [Figs. 5(c) and 5(d)]. The cardiac tubes were able to pump fluids during contraction on day 15 [Figs. 5(e) and 5(f) and supplementary Movie 3].

FIG. 5.

TRACE (tunable rapid assembly of collagenous elements)-printed cardiac tubes can pump fluids. Of note, these tubes were not cultured on PDMS posts and were instead cultured in the first 2 days with the agarose slurry kept inside. (a) Design of the 5 or 10 mm cardiac tubes. (b) Timeline of cardiac tube culture. Although we washed away the support bath outside of the tube immediately after printing, the slurry inside of the tubes remained for 2 days and prevented the tubes from collapsing. (c) and (d) 5 and 10 mm cardiac tubes became compact after culture, and their structures were intact. (e) The cardiac cycle (systole and diastole) of a representative 5 mm cardiac tube was indicated by particle image velocity (PIV) on day 15. (f) Mean fluid speed in a region of interest, marked by red rectangles in (e), during a complete cycle of contraction and relaxation. The sign of the speed was assigned based on the direction of the fluid flow, where a positive value means that the average fluid velocity was pointing away from the tissue, and a negative value means an average velocity pointing toward the tissue. The red arrows point to peak speeds during systole or diastole.

DISCUSSION

In this study, we demonstrate the use of TRACE, our innovative macromolecular crowding (MMC) technique, to achieve instant gelation of cell-laden collagen bioinks for bioprinting engineered cardiac tissues. These cardiac tissues could achieve enhanced cellular alignment and improved functionality through subsequent isometric training. We verified the cell viability of the printed tissues after being subject to the macromolecular crowding support bath. We initially bioprinted tissue strips using a mold but moved to tissue rings to improve throughput and also stability from collapse. Furthermore, we explored the fluid-pumping capabilities of cardiac tubes. We successfully printed cardiac tubes and assessed their ability to pump fluid, demonstrating the potential of our approach for creating high aspect ratio (as in long and tall) tissues. Our results demonstrated that TRACE bioprinting can successfully fabricate cardiac tissues without restricting the bioink component selection, allowing for greater versatility. Additionally, agarose slurry in the TRACE printing bath enabled the fabrication of more complex tissue structures while preserving their geometry throughout the culture process. We anticipate that TRACE could be a versatile tool to be used in new cardiac research that involves engineered cardiac tissues.^7,8,16

The tissue strip is a basic geometry that could be a part of many complex geometries. As such, we initially printed cardiac strips. However, we soon realized that the strips had low throughput in our setup for several reasons. First, they were easy to break due to potentially chafing on the PDMS platform during repetitive contractions. Second, installation requires manual manipulation using tweezers to properly position the strip, and this step increases the likelihood of breaking the strip. However, the strip is still a feature that could be vital to future investigations. While using posts to elongate a ring into an extended O-shape is an alternative approach that can simulate strip-like contraction,³ directly printing cardiac strips might still be advantageous when specific applications require longer constructs with uniform uniaxial tension, particularly in the context of creating tissue patches or modeling linear anatomical features. As such, we have some suggestions based on our experimental tests. Rounding the corners of the 3D-modeled designs of the PDMS platform might reduce chafing. Avoiding initial tautness in the strip could enable the ring to stabilize into a more robust structure before being subjected to high tension, potentially preventing the strip from breaking prematurely.

Due to the fast gelation time, the TRACE printing procedure allows minimal incorporation of surrounding supporting materials that may impede the conduction system that transmits electrical signals during muscle contraction. TRACE also minimizes bioink diffusion and loss of shape fidelity.⁴ The agarose slurry support bath also provides a distinct advantage over gelatin-based baths, which are difficult to selectively melt at specific locations during temperature-controlled removal.³ In contrast, our agarose slurry can be easily washed away or, as demonstrated in this study, retained inside the printed rings to support the structure during maturation. This prevents early collapse and helps maintain overall geometry until cell–cell junctions are formed and the overall structure is stabilized.

This study demonstrates the potential of TRACE to be combined with existing bioprinting technology to promote the development of regenerative medicine. Although vascularization is beyond the scope of this study, given its important role in preventing necrotic cores and improving tissue functions,¹⁷ future studies could integrate TRACE printing with vasculature engineering to enhance the overall functions of engineered cardiac tissues. For example, we have recently demonstrated that incorporating thick collagen fiber bundles, fabricated using macromolecular crowding, into a fibrin-Matrigel co-gel guides vascular network formation around intestinal organoids.¹⁸

This study is focused on the development process to print cardiac tissues using TRACE. The development of TRACE itself and further biological characterization of TRACE-printed cardiac tissues can be found in the separate study.⁴ For example, we characterized the storage modulus and opacity of the gels during TRACE to ensure that the gelation was accelerated during TRACE.⁴ Because TRACE exposes cells to high concentrations of PEG, we also checked cell states, and we found that the cells display high viability and metabolic activities.⁴ In addition, in the separate study, we compared the TRACE-printed cardiac tissues with cardiac tissues that were printed in a regular bath (with only the agarose slurry but without the PEG 8000 solution). The printed tubes in the regular bath break down over the course of culture and could not self-organize into a compacted tissue due to excessive incorporation of agarose.⁴ Overall, this study focuses on the development of TRACE to print cardiac tissues.

Our study demonstrates that TRACE is a versatile method for cardiac tissue bioprinting and can be integrated with other approaches to fabricate functional, biomimetic tissues. Notably, the self-pumping tubes developed in this study could be further enhanced by incorporating one-way valves, paving the way for future advancements in vascular grafts designed for unidirectional flow.^14,15 Additionally, while this work primarily focuses on training cardiac tissues through isometric contraction, future research could investigate the effects of dynamic contractions—such as passive or active stretching,¹³ where the central post undergoes controlled deformation—on the functionality and maturation of bioprinted cardiac tubes.

METHODS

Fabrication of the PDMS multichannel device for molding macroscopic collagen strips

The multichannel device was fabricated using PDMS through a molding process. PDMS and silicone elastomer curing agent (SYLGARD 184, Dow Corning) were mixed in a 10:1 ratio and poured into a Petri dish. Degassing in a vacuum chamber was performed to remove trapped air bubbles. To create parallel channels, blunt needles were aligned and fully immersed in the PDMS liquid mixture. To vary the channel sizes, needles of different gauges were used: 25 gauge [outer diameter (OD) = 0.515 mm; results not shown as these strips were too fragile and ripped apart easily when being flushed], 19 gauge (OD = 1.067 mm), 17 gauge (OD = 1.473 mm), and 14 gauge (OD = 2.109 mm) (McMaster-Carr); channels created with 19 gauge needles were selected for cell-laden experiments. The mixture was subsequently cured overnight in the 60 °C oven. After curing, the device was peeled off the Petri dish, trimmed to the appropriate size, and the needles were demolded. A 30-s treatment in a plasma cleaner (Glow Research) was conducted on the device, to promote the channels' surface hydrophilicity. For cell-laden strip fabrication, the device and its channels were sterilized with 70% ethanol, rinsed with sterile water, and air-dried in the tissue culture hood before use.

Fabrication of a PDMS platform for holding macroscopic collagen strips

The platform was fabricated using PDMS. PDMS and silicone elastomer curing agent (SYLGARD 184, Dow Corning) were mixed at a 10:1 ratio and poured into a Petri dish to a height of 5 mm. Degassing in a vacuum chamber was performed to remove trapped air bubbles. The mixture was then cured overnight in a 60 °C oven. After curing, the PDMS was peeled off the Petri dish, and a 1 ×1 cm² cutout was made using a blade. Each cutout was then punctured with a 6 mm biopsy punch to form a central well. The PDMS cutout was sterilized with 70% ethanol and air-dried before proceeding to the next steps. Both the PDMS cutout and a 35 mm glass-bottom dish with a 20 mm micro-well (Cellvis, D35-20-1.5-N) were plasma-treated and immediately pressed together to ensure adhesion. The assembled setup was incubated overnight at 60 °C to secure the bond.

To ensure proper adhesion of the fibrin gel during the later steps of installing the strip, we pretreated the surfaces where the fibrin would attach with polydopamine. This treatment was applied to the areas outside the central well by immersing the areas with 1.5 mg/ml polydopamine in 10 mM Tris/HCl buffer (pH 8.5) for 30 min at room temperature (RT), followed by a thorough wash with cell culture-grade water. To prevent unintended adhesion of cells or the collagen strip to the surfaces of the center well, the center well was pretreated with Pluronic-F127 (Sigma-Aldrich, P2443). The center well was filled with 3% Pluronic-F127 in the PBS buffer for 30 min, followed by a thorough wash with cell culture-grade water. The device was air-dried in the tissue culture hood before use.

Fabrication of PDMS posts

The post was fabricated using PDMS. PDMS and silicone elastomer curing agent (SYLGARD 184, Dow Corning) were mixed at a 10:1 ratio and poured into a Petri dish to a height of 5 mm. Degassing in a vacuum chamber was performed to remove trapped air bubbles. The mixture was then cured overnight in a 60 °C oven. After curing, the PDMS was peeled off the Petri dish, and several 1 mm diameter posts were punched out using a blunt gauge-19 syringe needle (McMaster-Carr).

The PDMS posts were sterilized with 70% ethanol and air-dried before the next steps. Both the PDMS posts and a glass-bottom 24-well plate were plasma-treated and then immediately positioned upright using tweezers to facilitate adhesion. The assembled setup was incubated overnight at 60 °C to ensure a secure bond.

Instant gelation of tissue strips via MMC in the PDMS multichannel device

The instant gelation solution was prepared by dissolving high molecular weight polyethylene glycol (PEG 8000, Sigma) in sterile PBS at a concentration of 200 mg/ml. The solution can be stored at room temperature. For tissue culture applications, the PEG bath was sterilized by a 0.45 μm syringe filter. Sterile PEG 8000 solution was pipetted to fill the channels of the PDMS multi-channel device. We then pipetted bioinks into channels of a PDMS mold that was prefilled with 200 mg/ml PEG 8000 and waited for 2 min for the gel to polymerize. Once the gel polymerized, we used a P200 pipette to flush the channels with 1× PBS, transferring the strips into a Petri dish containing 1× PBS or cell culture media.

Bioinks preparation

To demonstrate the feasibility of instant gelation, we prepared the acellular bioink composed of 3 mg/ml collagen mixed with 5% red beads for visibility (v/v; ThermoFisher Scientific, F8801), following previous protocols.¹⁹ Specifically, rat tail collagen type I (stock concentration 8–10 mg/ml; Corning, 354249) was mixed with 10× PBS (making up 10% of the final bioink volume) containing phenol red. The mixture was neutralized with NaOH (pH 7.2–7.4), followed by the addition of the red bead solution (making up 5% of the final bioink volume), and then diluted to the desired concentration using 1× PBS.

The fibroblast bioink was composed of 3 mg/ml collagen and 300,000 cells/ml. The preparation steps were similar to those used for acellular strips. Rat tail collagen type I was mixed with 10× PBS (making up 10% of the final bioink volume) containing phenol red and then neutralized using NaOH. Cell culture media containing the fibroblasts was added to the bioink, making up 20% of the final bioink volume. Finally, 1× PBS was used to adjust the bioink to the desired concentration.

The cardiac bioink was prepared by combining rat tail collagen I with Geltrex (15% v/v), 10× PBS containing phenol red (10% v/v), dissociated cardiomyocytes (30% v/v), and human cardiac fibroblasts (HCFs) (10% v/v). The mixture was then adjusted with 1× PBS to achieve a final collagen concentration of 3.0 mg/ml. The final cell concentrations in the hydrogel were 4 × 10⁷ cells/ml for cardiomyocytes and 4 × 10⁶ cells/ml for HCF unless otherwise noted. The cardiac bioink was adapted from an established protocol.⁷

Tissue strip installation on the PDMS platform

Fibroblast and cardiac strips were mounted on a PDMS platform and secured at both ends with fibrin gels. Before the transfer, the central well of the PDMS platform was filled with the appropriate culture medium so as to keep the newly transferred strip moist later. During the transfer, the strips were carefully pipetted up along with some culture media using a wide-bore pipette tip. The strip was then laid gently on the PDMS platform, ensuring it remained mostly straight. After the transfer was completed, a fibrin gel was prepared by mixing 20 mg/ml fibrinogen with 13.3 units/ml thrombin at a 1:1 ratio. The mixture was applied to the edges of the PDMS platform in less than 30 s, covering the ends of the strip before the fibrin polymerized. After allowing 5 min to ensure complete fibrin polymerization, additional culture media were added to fully immerse both the platform and the strips.

To prepare sterile 20 mg/ml fibrinogen, we dissolved 5 mg of fibrinogen (Sigma-Aldrich, F8630-1G) in 250 μl of DPBS (-/-, without calcium or magnesium; Gibco, 14190-144), kept the mixture at 37 °C water bath for 3 h to ensure complete dissolution, and sterile-filtered it through a low protein-binding filter (Fisher Scientific, SLGVR33RS). Thrombin (Sigma-Aldrich, T9549) was dissolved in DPBS to a concentration of 13.3 units/ml. The preparation of the fibrinogen and thrombin followed a previous publication.²⁰

Cell culture and maintenance

The fibroblasts used in this study were normal human lung fibroblasts (NHLF) from Lonza (CC-2512; tissue acquisition number is 29729, which correlates with a patient). Fibroblasts were cultured in DMEM (ThermoFisher, 11965118) supplemented with 10% fetal bovine serum (FBS; ThermoFisher, 16000044) and 1% pen-strep (Life Technologies, 15140122). The cells were cultured at 37 °C in 5% CO₂ with media changed every other day. The cells were lifted by 0.25% Trypsin-EDTA (Gibco, 25200056) once the confluency reached 70%–80%. Passage numbers lower than 20 were used for experiments.

Human cardiac fibroblasts (HCFs; PromoCell, C-12375) were cultured in human cardiac fibroblast media (Cell Applications, 316-500), and the media was changed every other day. Passage numbers lower than 6 were used for experiments.

Cardiac differentiation from hiPSCs

We maintained human-induced pluripotent stem cells, hiPSCs, as previously described.²¹ In brief, hiPSCs (Coriell Institute for Medical Research, GM23338) were cultured with mTeSR1 (Stemcell Technologies, 85850) on Geltrex (Gibco, A14132-02)-coated plates until just before colonies became connected, which was about 80% confluent. Cells were detached with ReleSR (Stemcell Technologies, 05872) following its commercial protocol and re-plated on Geltrex-coated 12-well plates. Cells were maintained in mTeSR1 until they had above 98% confluence before starting differentiation.

Differentiation to cardiomyocytes was performed as previously described^21,22 and summarized in supplementary Table 1. In brief, on day 0 of differentiation, we prepared 12.5 μM CHIR99021 (Stemcell Technologies, 99021) in a mixed media of mTeSR1 and RPMI/B-27 without insulin (v/v was 1:3); RPMI (Gibco, 11875-093) and B-27 without insulin (Gibco, A18956-01) were mixed beforehand at a ratio of 49:1. Twenty-four hours later, the media was replaced with RPMI/B-27 without insulin. On day 3 at the same hour when CHIR99021 was added, media was replaced with RPMI/B27 without insulin, containing 5 μM IWP4 (Stemcell Technologies, 72552). An additional 48 h later (on day 5), the media was replaced with RPMI/B27 without insulin, and the media was again refreshed on day 7. On day 9, media was replaced with RPMI/B27 (the B27 used in this combined media contained insulin; Gibco, 17504-044), and media was refreshed on day 11. On day 12, we purified cardiomyocytes using RPMI without glucose (Gibco, 11879-020), containing 4 mM lactate (Sigma-Aldrich, L1750-10G),²³ and the media was refreshed on day 14. On day 16, media was replaced with RPMI/B27, which contains insulin. On day 18, cardiomyocytes were dissociated using TrypLE (Gibco, A1217702) following its commercial protocol.

Generation of TRACE printing bath

Agarose, a biocompatible inert hydrogel, was used to create a supportive slurry following our TRACE protocol.⁴ In brief, the agarose slurry was prepared by dissolving 0.7% (w/w) agarose (Sigma-Aldrich) in boiling 50 ml PBS (Gibco) in a 100 ml glass media bottle, followed by gradual cooling from 85 to 20 °C on a hot plate while stirring with a magnetic bar at 700 rpm. TRACE support bath was created by thoroughly mixing 5.8 ml of the resulting agarose slurry with 0.6 ml of 800 mg/ml PEG 8000 solution in PBS. Mixing was achieved by pumping the two materials in two 10 ml syringes (BD Emerald) back and forth through a 90-degree Luer-lock connector (McMaster-Carr) at least ten times or until thorough mixing. For cell-laden collagen bioprinting, the 0.7% agarose solution was autoclaved before cooling, and PEG 8000 was filtered through a 0.45 mm filter (Millipore) for sterilization. The mixing accessories were also autoclaved.

TRACE printing

The TRACE printing process followed a previous protocol.⁴ Extrusion-based bioprinting was performed using a commercialized bioprinter, LulzBot Bio Printer (LulzBot, North Dakota, USA), with a custom temperature-controlled print head. All digital models were designed in SolidWorks and exported as stereolithography (.stl) files. The files were printed using Cura LulzBot Edition (Aleph Objects, Colorado, USA). We set the layer height to 0.2 mm, and the flow rate was set to 180% to encourage the merging of layers and avoid delamination. 25-Gauge nozzles with 1 inch in length (McMaster-Carr) were used to print at a speed of 7 mm/s. Before printing, a container large enough to hold the printed structure was filled with the TRACE support bath and secured to the printing bed. The temperature-controlled print head was set to 4 °C. Bioink was loaded into a pre-chilled printing syringe (Hamilton, 81420) through a straight socket-to-socket connector (McMaster-Carr, 51525K423) and then stored on ice until use. To ensure sterile conditions during cell-laden collagen printing, the printer was wiped using 70% ethanol and transferred into a biosafety cabinet. The nozzles, printing syringes, and needles were sterilized by autoclaving. After the printing was completed, the printed structures were secured in the TRACE bath for at least 3 min and then released carefully by removing and pipetting the TRACE support bath. In this step, we tried to remove as much support bath as possible that was surrounding the printed ring or tube. The support bath inside the ring or tube helped maintain the printed structure and prevented the newly printed structure from collapsing. We kept the support bath within the ring or tube until day 2. At that point, we removed the remaining support bath by repeatedly and vigorously flushing the structure using fluid streams from a pipette. Then, the desired culture media was added, and the printed tissues were cultured in a humid incubator at 37 °C with 5% CO₂ for future applications. On day 2, we removed the remaining support bath and installed the rings on PDMS posts, and we used tweezers to carefully position a ring around a post.

Maintenance of cardiac tissues

Following a previously established protocol, the bioprinted cardiac tissues were cultured in seeding media [10% FBS (Gibco, 10438-026), 1% P/S (Gibco, 15140122), 1% NEAA (Gibco, 17 11140-050), 1% L-glutamine (Gibco, 25030-081), and 1% sodium pyruvate (Gibco, 11360-070), and 10 μM Y27632 (Stemcell Technologies, 129830-38-2)] on day 0. On day 1, media was refreshed with DMEM (Gibco, 11995065)/B27 (Gibco, 17504-044) mixed at a ratio of 49:1 (v/v), and media was refreshed every other day from then on.

To track the macroscopic change of cardiac tissues across days, images were taken with a coin microscope purchased from Amazon.

Calcium imaging and electrical stimulation

We followed a previous protocol for calcium imaging³ with some modifications. Tissues were first incubated at 37 °C and 5% CO₂ for 90 min containing 5 μM calcium indicator Cal 520 AM (AAT Bioquest, 21130) and 0.025% Pluronic-F127 (Sigma-Aldrich, P2443). The media was then refreshed with DMEM/B27. We waited for 30 min so that the calcium signal could reach equilibration before we started imaging in a 37 °C, 5% CO₂ chamber.

A custom bioreactor, kindly lent to us by Dr. Stuart Campbell, was used to electrically stimulate the cardiac tissues.²¹ In brief, the bioreactor was assembled by fitting a laser-cut Teflon frame with computer numerical control (CNC)-machined graphite electrodes (Graphitestore) on top of a 12-well plate. This assembly was further connected to an Arduino Uno microcontroller and a printed circuit board to provide bipolar pulses to individual tissues. We applied a 2.5-V/cm across individual tissues. The bioreactor was cleaned with 70% ethanol and de-ionized water and autoclaved before use.

Calcium signal and kymograph analyses were performed using Fiji. The calcium signal was analyzed by calculating the average intensity within a field of view approximately 2.6 × 2.6 mm², with the cardiac ring positioned at the center. The average intensity was then normalized by dividing the average intensity at diastole (F/F0).

Particle imaging velocimetry imaging and measurements

Particle imaging velocimetry (PIV) measurements were performed using a Leica DMi8 system at 10× magnification with a mercury lamp and a high-sensitivity Hamamatsu camera. Cell culture media consisted of 0.5% (v/v) 1 μm-in-diameter fluorescent microspheres (ThermoFisher Scientific, F8821) to visualize the fluid flow. The time interval of imaging was set to 17 ms. Imaging was conducted in DMEM/B27 media and an environmental chamber at 37 °C, 5% CO₂. PIV analysis was performed using PIVlab²⁴ in Matlab (MathWorks).

Immunofluorescent staining

The cardiac rings were fixed on day 14 with 4% paraformaldehyde (PFA, Santa Cruz Biotechnology) for 40 min and then thoroughly washed with PBS. Samples were permeabilized by 0.3% Triton-X100 (Sigma) for 1 h at room temperature and then thoroughly washed with PBS. 1% BSA in PBS was used to block the samples at room temperature for 1 h. The samples were incubated with a primary antibody of α-actinin (Millipore Sigma, A7811) overnight at 4 °C. The samples were then thoroughly washed with PBS and incubated with Hoechst (1:2000), rhodamine phalloidin (1:1000; Abcam), and a secondary antibody for 3 h at room temperature. The samples were thoroughly washed with PBS before imaging.

To visualize the outer perimeter of the rings, we used a PDMS post to go through the center of the ring, allowing it to stand on its side. Images were acquired using a Leica DMI8 confocal microscope.

Inner ring area measurements

We measured the inner areas of the cardiac rings to track tissue progression over time. The measurements were performed manually using Fiji.

We also measured the inner areas at systole and diastole to assess the extent of area change. Measurements were also performed manually using Fiji. Percentage area contraction was calculated as (A_diastole − A_systole)/A_diastole × 100, where A stands for the inner areas of the rings.

Statistical analysis

Statistical significance was determined by t-test, with significance indicated by ^*P ≤ 0.05, ^**P ≤ 0.01, ^***P ≤ 0.001, and ^****P ≤ 0.0001. Error bars represented the standard error of the mean (SEM). Statistical analysis was performed using GraphPad Prism 10.

SUPPLEMENTARY MATERIAL

See the supplementary material for additional information on the comparison between cardiac strips that could not beat synchronously due to low cell density (4 × 10⁶ cardiomyocytes/ml; Movie 1) and strips that could beat synchronously due to high cell density (4 × 10⁷ cardiomyocytes/ml; Movie 2). HCFs were seeded at a 1:10 ratio to cardiomyocytes in both tissues. Movie 3 shows a cardiac tube that pumps fluids. Table 1 summarizes the timeline for differentiating cardiomyocytes from hPSCs.

AUTHOR DECLARATIONS

Conflict of Interest

The authors have no conflicts to disclose.

Ethics Approval

Ethics approval was not required for this study.

Author Contributions

Hugh Xiao: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Project administration (equal); Visualization (equal); Writing – original draft (equal). Zixie Liang: Conceptualization (equal); Investigation (equal); Methodology (equal); Visualization (supporting); Writing – original draft (supporting). Xiangyu Gong: Conceptualization (equal); Investigation (equal); Methodology (equal); Visualization (supporting); Writing – original draft (supporting). Seyma Nayir Jordan: Conceptualization (supporting); Investigation (supporting). Alejandro Rossello-Martinez: Methodology (supporting). Ilhan Gokhan: Methodology (supporting). Xia Li: Methodology (supporting). Zhang Wen: Methodology (supporting). Sein Lee: Methodology (supporting). Stuart Campbell: Investigation (supporting). Yibing Qyang: Investigation (supporting). Michael Mak: Conceptualization (equal); Formal analysis (supporting); Funding acquisition (lead); Investigation (equal); Methodology (supporting); Project administration (lead); Supervision (lead); Writing – original draft (equal); Writing – review & editing (equal).

DATA AVAILABILITY

The data that support the findings of this study are available within the article and its supplementary material.