Reprogrammable 4D tissue engineering hydrogel scaffold via reversible ion printing
Aixiang Ding, Fang Tang, Sriramya Ayyagari, Eben Alsberg
TL;DR
A new hydrogel scaffold can change shape in a controlled way and be reprogrammed, which could help in tissue engineering.
Contribution
A reprogrammable hydrogel system using ion-transfer printing for 4D tissue engineering is introduced.
Findings
Hydrogels with tunable crosslinking gradients enable programmable shape deformation.
Shape morphing can be reprogrammed using a secondary ion-transfer printing process.
The system supports long-term cell differentiation in tissue-like constructs.
Abstract
Shape-morphable hydrogel scaffolds recapitulating morphological dynamism of native tissues represent an elegant tool for tissue engineering (TE) applications. Current morphable hydrogels are predominantly based on multimaterial structures, which involve complicated and time-consuming fabrication protocols, and are often limited to unidirectional deformation. This work reports on the development of a transformable hydrogel system using a fast, simple, and robust fabrication approach for manipulating the shapes of soft tissues at defined maturation states. Simply by using an ion-transfer printing (ITP) technology, a tunable ion crosslinking density gradient across the hydrogel thickness has been incorporated, which enables preprogrammable deformations upon further swelling in cell culture media. Combining with a surface patterning technology, cell-laden constructs (bioconstructs) capable…
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
Topics3D Printing in Biomedical Research · Advanced Materials and Mechanics · Nanofabrication and Lithography Techniques
Introduction
1
Living tissues during development can change their shape to form the final functional tissue architecture, better interact with neighboring tissues and/or achieve better nutrient and waste exchange with the surrounding environment. A representative example of morphologically dynamic tissue development is branching morphogenesis whereby the tissues grow into tree-like architectures through evagination of epithelial tubes to maximize the surface area [1]. Therefore, to biomimic aspects of tissue maturation with tissue engineering (TE), which sometimes utilizes isolated cells and biodegradable scaffolds to produce artificial equivalents for patients with diseased or damaged tissues [2], it may be valuable to not only regulate cell proliferation, cell differentiation, and secretion of extracellular matrix (ECM), but also potentially closely imitate the dynamic architectural form changes that occur during development and the responses of tissues to stimuli [3]. However, traditional TE strategies mostly focus on tuning scaffold macro- and microstructures that resemble mature tissue architectures and often ignore the developmental morphodynamics of native tissues [4]. Since continuous shape evolution occurs during in vivo maturation, new approaches have focused on synthesizing shape transformable (termed 4D) scaffolds to enable 4D biomimetic TE.
Shape memory polymers (SMPs) have been widely utilized as transformable tissue scaffolds due to their robust mechanical properties and structural stability, as well as their ability to undergo reversible shape changes in response to external stimuli, enabling self-deploying and self-fitting implants and minimally invasive delivery [[5], [6], [7]]. The most common approach involves seeding live cells on SMPs in a 2D configuration; however, this fails to provide a physiologically relevant 3D microenvironment for cell growth and differentiation. In contrast, hydrogels synthesized from the crosslinking of hydrophilic polymers and macromers as scaffolding materials have garnered prominence in TE since they are readily available and can be engineered to resemble natural tissue ECM both physically and biochemically [8,9]. Moreover, their ability to encapsulate live cells allows for the creation of 3D microenvironments that closely resemble those found in vivo [10,11]. Geometrically programmed hydrogels with the potential to achieve shape transformations are highly desirable for recapitulating tissue morphodynamics and, as such, work toward this goal has facilitated the development of shape-morphing TE hydrogel scaffolds [12,13]. The engagement of shape transformability breaches the boundary of conventional geometrically inert 3D TE scaffolds and is considered an important next-generation TE technology [14,15]. In addition to the dynamically adjustable architectures, the employment of geometrically transforming hydrogels also presents the potential to fabricate TE scaffolds with more sophisticated structures compared to traditional inanimate hydrogels [16,17].
Strain mismatch acting as the driving force to trigger the shape changes is a primary strategy for designing transformable hydrogel scaffolds [18]. In this regard, transformable hydrogel scaffolds are often devised as multilayers comprised of materials with different swelling properties in different layers [[19], [20], [21]]. Nevertheless, multilayered hydrogels are structurally complex, necessitate multistep preparation, and may encounter delamination issues due to insufficient interfacial adhesion [22]. In contrast, single-layer hydrogel scaffolds are simple in structure and fabrication process but need careful anisotropy incorporation into the structures by controlling the spatial arrangement of the materials used. Therefore, the selection of smart materials with defined responsiveness and the structural design exerts a crucial impact on the reconfigurability capacity of final scaffold products. In this respect, natural polymers such as alginate, hyaluronic acid, silk, gelatin, and chitosan and synthetic polymers such as polyethylene glycol (PEG) represent potential base materials for 4D systems [23], with alginate being of particular interest as the shape of cell-laden hydrogels composed of this material can respond to divalent ions as a physiological stimulus [24]. However, to induce shape morphing, alginate hydrogels usually need to be formed with other materials to generate structural heterogeneity and/or be cast carefully into a single layer with a compositional or microstructural gradient [25,26]. For example, phase separation, polymer diffusion, and spatially defined photocuring are commonly employed for gradient engineering [[27], [28], [29], [30]]. Precise control over the gradient range, location, and orientation is crucial for programming shape transformation, yet it remains challenging to achieve. Moreover, reported single-layered alginate hydrogels typically can only undergo a single unidirectional shape morphing [31], which limits their application in practical scenarios wherein tissues should exhibit complicated developmental dynamism involving multiple different shape changes over time.
Herein, we present a facile, fast fabrication strategy to produce single-layered hydrogel scaffolds with reversible, reprogrammable shape-morphing performance. Oxidized and methacrylated alginate (OMA) hydrogel constructs encapsulating living cells were programmed to perform prescribed shape transformations. Photocured cell-laden OMA constructs (bioconstructs) were physically crosslinked via an ion-transfer printing (ITP) approach to form an ionic crosslinking gradient throughout the hydrogel thickness (Scheme 1), due to the decreasing ion concentration along the diffusion direction [32], with the top surface directly exposed to the Ca^2+^-soaked filter paper forming high-crosslinking density and tight polymer networks and the bottom surface crosslinked by the diffused Ca^2+^ forming low-crosslinking density and loose polymer networks (Scheme S1). The as-prepared large gradient bioconstructs were subsequently tailored into specific geometries to initiate deformation upon culture in cell growth medium (GM) to generate predefined shapes. Furthermore, removing the crosslinker Ca^2+^ ion by a stronger chelator such as ethylenediaminetetraacetic acid (EDTA) at a cytocompatible concentration resulted in internal stress release while maintaining high cell survival. As a result, the bent bioconstructs were straightened and could be reprogrammed by ITP to undergo a different deformation process (e.g., opposite-direction bending). With the proposed approach, we demonstrated “3D-to-3D” morphological conversions and, for the first time, validated its effectiveness to reshape differentiated tissue-like constructs, revealing its capability to program a ready-to-use 3D bioconstruct of interest. Our results suggest the huge potential of this reversible ion printing technology in reprogrammable 4D TE.Scheme 1. The fabrication of a shape-transformable hydrogel TE scaffold and its shape morphing and reprogramming. (a) Soaking filter paper in Ca^2+^ solution for ion transfer. (b) ITP-resultant crosslinking density gradient in OMA hydrogel scaffold and its subsequent shape transformation upon culture in solution and the reprogramming process.Scheme 1
Materials and methods
2
Chemicals, instruments, and general methods
2.1
Unless specified, all solvents and reagents were used without further purification. Sodium alginate (Protanal LF120M, 251 Pa s) was a generous gift from FMC Biopolymer. Photoinitiator (2-hydroxy-4’-(2-hydroxyethoxy)-2-methylpropiophenone), fluorescein diacetate (FDA), ethidium bromide (EB), Dulbecco's Modified Eagle Medium-Low Glucose (DMEM-LG), and fetal bovine serum (FBS) were purchased from Sigma. Insulin transferrin selenium^+^ (ITS^+^) Premix and penicillin/streptomycin (P/S) were purchased from Corning Inc. (Corning, NY). Sodium pyruvate was purchased from HyClone Laboratories (Logan, UT). Non-essential amino acid solution was purchased from Lonza Group (Basel, Switzerland). Ascorbic acid-2-phosphate was purchased from Wako Chemicals USA Inc. (Richmond, VA). Fibroblast growth factor-2 (FGF-2) was purchased from R&D Systems (Minneapolis, MN). Transforming growth factor β1 (TGF-β1) was purchased from PeproTech (Rocky Hill, NJ). N-(2-aminoethyl) methacrylate hydrochloride (AEMA) and methacryloxyethyl thiocarbamoyl rhodamine B (RhB) were purchased from Polysciences Inc. (Warrington, PA), and anti-CD44 antibody (catlog no. CBL154MI) and other common chemicals, such as sodium peroxide, methacrylic anhydride, etc., were purchased from Fisher Scientific (Hampton, NH). ^1^H NMR spectra were obtained on a 400 MHz Bruker AVIII HD NMR spectrometer (Billerica, MA) equipped with a 5 mm SmartProbe™ at 25 °C using deuterium oxide (D_2_O) as a solvent and calibrated using (trimethylsilyl)propionic acid-d4 sodium salt (0.05 w/v %) as an internal reference. Low glucose Dulbecco's Modified Eagle Medium (DMEM-LG) containing photoinitiator (0.05 w/v %) was used to dissolve the OMA and methacrylate gelatin (GelMA). GM consisted of DMEM-LG with 10% FBS and 1% P/S, and chondrogenic medium consisted of DMEM-LG with 1% ITS^+^ Premix, 100 nM dexamethasone, 1 mM sodium pyruvate, 100 μM non-essential amino acids, 37.5 μg/mL ascorbic acid-2-phosphate and 1% P/S supplemented with 10 ng/mL TGF-β1. Hydrogel images were visualized using a Nikon SMZ-10 Trinocular Stereomicroscope equipped with a digital camera. A microplate reader (Molecular Devices iD5, San Jose, CA) was used to read data from the microplates. A UV device (EXFO OmnicureR S1000-1B, Lumen Dynamics Group, Mississauga, Canada) was used for photocrosslinking. All quantitative data was expressed as mean ± SD. Statistical analysis was performed with one-way analysis of variance (ANOVA) with Tukey honestly significant difference post hoc tests using Origin software (OriginLab Corporation, Northampton, MA). A value of p < 0.05 was considered statistically significant.
Synthesis of OMAs and GelMA
2.2
OMA macromers with a theoretical 1% oxidation degree and a theoretical 20% methacrylation degree were synthesized according to a similar method as described in the literature [33,34]. Briefly, 10 g of sodium alginate was dissolved in 900 mL of deionized water (diH_2_O) overnight, and 108 mg of sodium periodate (NaIO_4_) in 100 mL of diH_2_O was rapidly added to the alginate solution under stirring in the dark at room temperature (RT). After reaction for 24 h, 19.52 g of 2-ethanesulfonic acid (MES) and 17.53 g of sodium chloride (NaCl) were added, and the pH was adjusted to 6.5 with 5 N sodium hydroxide (NaOH). Then 1.18 g of N-hydroxysuccinimide (NHS) and 3.89 g of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC·HCl) were sequentially added to the mixture. After 10 min, 1.69 g of AEMA was added slowly. The solution was wrapped with aluminum foil to protect it from light and left to react for 24 h at RT. The mixture was then poured into 2 L of chilled acetone to precipitate out the crude OMA solid, which was further purified by dialysis against diH_2_O over 3 days (MWCO 3.5 kDa, Spectrum Laboratories Inc., Rancho Dominguez, CA). The dialyzed alginate solution was collected, treated with activated charcoal (0.5 mg/100 mL, 50-200 mesh, Fisher Scientific) for 30 min, filtered through a 0.22 μm filter, and frozen at −80 °C overnight. The final OMA polymer was obtained as a white cotton-like solid through lyophilization for at least 10 days. The actual methacrylation of O1M20A was determined to be 6.1% from ^1^H NMR data according to the method described in the literature [35]. Note that the actual oxidation was not provided due to the overlap of the proton peak assigned to the CHO group (∼5.4 ppm) with the polymer proton peak (broad peak located at ∼5.1 ppm).
The GelMA was synthesized following our previously reported protocol [12], with slight modifications from the original method [36]. Briefly, 20 g of gelatin was dissolved in 200 mL of PBS (pH 8.0) at 50 °C under constant stirring until fully dissolved. Methacrylic anhydride (20 mL) was then added dropwise at a rate of 1 mL/min while maintaining vigorous stirring. The reaction mixture was kept for 1 h at 50 °C, after which it was cooled to RT and allowed to react overnight. The crude product was precipitated by adding excess acetone and subsequently purified by dialysis against deionized water using 14 kDa MWCO dialysis tubing for 7 days at 50 °C. The final GelMA product was obtained after lyophilization for 2 weeks.
The ^1^H NMR spectra for OMA and GelMA are shown in Supporting Information.
Cell expansion
2.3
Human mesenchymal stem cells (hMSC) were isolated according to the literature [37]. hMSC cells were expanded in GM supplemented with 10 ng/mL FGF-2 and NIH3T3 cells were expanded in GM. The culture was performed in a humidified incubator at 37 °C and 5% CO_2_ with medium changes every 2 or 3 days. The cells were harvested when they reached ∼80% confluence.
OMA hydrogelation
2.4
Solution of OMA (3 w/v %) with or without GelMA (0.5 w/v %) in DMEM-LG containing photoinitiator (0.05 w/v %) was placed between two quartz plates with 1.0 mm spacers and UV crosslinked at ∼18 mW/cm^2^ for 45 s to form cell-free hydrogels, which were then ion printed using the ITP protocol described in Section 2.8 below and cut into specific geometries to culture in medium. For the mechanical property studies, dual-crosslinked OMA hydrogels were prepared by culturing the photocrosslinked hydrogels in a 0.5 M Ca^2+^ aqueous solution for 10 min. To encapsulate cells inside hydrogels, macromers in DMEM-LG containing cells were subjected to photocrosslinking as described above, generating cell-laden hydrogels.
Degradation and swelling tests
2.5
Photocrosslinked cell-free OMA hydrogels were fabricated as above (Section 2.4) and used for degradation and swelling tests. Circular hydrogel samples with a diameter of 8 mm (d0) were obtained via punching the bulk hydrogels using a biopsy punch. For the degradation test, the samples were frozen for 4 h at −80 °C and lyophilized for 2 days. The masses of the dried gels were measured as initial weights (Wi). The dried hydrogels were rehydrated by culturing in 5 mL of GM under cell culture conditions (37 °C, 5% CO_2_), and the medium was changed every 3 days. At predetermined timepoints, the hydrogels were collected and dried by lyophilization to obtain dried mass (Wd). Degradation expressed as the mass loss was quantified as (Wi-Wd)/Wi × 100% (N = 3). For the swelling test, the samples were cultured for 4 h in 5 mL of H_2_O, PBS (pH 7.4), GM, or 0.5 M Ca^2+^ solution under RT. The diameters of the swollen hydrogels (dS) were measured. The swelling was calculated with the following equation: ds/d0 (N = 3).
Rheology
2.6
Dynamic rheological examination of the as-prepared and swollen/shrunk photocrosslinked cell-free OMA hydrogels was performed to measure the hydrogel storage moduli (G′) with a Kinexus ultra + Highest specification rheometer (Malvern Panalytical, Malvern, United Kingdom). In oscillatory mode, a parallel plate geometry (8 mm diameter) measuring system was employed, and the gap was set to 1 mm. After each hydrogel was placed between the plates, all the tests were carried out at RT (N = 3). Oscillatory frequency sweep (0.1∼100 Hz at 1 % strain) tests were performed to measure G’.
Compressive modulus measurement
2.7
The elastic moduli of the as-prepared and swollen/shrunk photocrosslinked cell-free OMA hydrogels were determined by performing uniaxial, unconfined constant strain rate compression testing at RT using a constant crosshead speed of 0.8%/sec on a mechanical testing machine (225lbs Actuator, TestResources, MN, USA) equipped with a 5 N load cell. The compressive modulus of each sample, under the specified sample dimensions and testing conditions, was determined using the first non-zero slope of the linear region of the stress-strain curve within 0 ∼ 10% strain (N = 3).
ITP
2.8
ITP was used to create a crosslinking density gradient across the thickness of photocrosslinked OMA hydrogels. In a typical protocol, the filter paper (FisherBrand, qualitative P5, medium porosity, Fisher Scientific) was soaked in a Ca^2+^ bath at a concentration of 0.1 − 0.5 M (specified in experiments) for 30 min. The soaked filter paper was taken out of the ion bath and placed onto the surface of the photocrosslinked OMA hydrogel for 30 − 120 s (specified in experiments). The hydrogel with a Ca^2+^-crosslinking gradient was tailored into specific shapes (e.g., square, circle, and bars) and then cultured in medium for 5 min at RT. Hydrogel images were then taken for analyzing the shape changes. The bending angles were quantified according to a method previously described in the literature [20].
Reprogramming and reversibility study
2.9
Reprogramming was realized through EDTA treatment and a subsequent re-ITPing. Briefly, ITPed hydrogels after deformation were transferred from the culture medium to 10 mL of EDTA solution (5 mM) in H_2_O (for cell-free hydrogels) or DMEM-LG without adding NaHCO_3_ (for cell-laden hydrogels), which were further cultured at RT (no shaking) or 37 °C under shaking (2.37 rev/s) (Bellco Glass 7744-01010 orbital shaker, Bellco Biotechnology, NJ, USA) for over 1.0 h to recover the shapes. Subsequently, the EDTA-treated samples were transferred into H_2_O (for cell-free hydrogels) or DMEM-LG (for cell-laden hydrogels) for 5 min to rinse away the residual EDTA within the hydrogels. Then, the recovered hydrogels were reprogrammed via another ITP process on either the same side or the other side as described in Section 2.8, and this was followed by culture in medium to obtain another deformed shape. Reversibility was assessed by repeating the above process for five cycles.
Gradient structure visualization via Scanning Electron Microscopy (SEM) imaging and dye incorporation
2.10
The ITPed and EDTA-treated hydrogels were sectioned with a blade to expose their cross-sectional surfaces. The samples were then frozen in liquid nitrogen for 5 min, followed by lyophilization for 24 h. Prior to imaging, the specimens were sputter-coated with a 10 nm layer of gold. SEM was conducted using a JSM-IT500HR system (JEOL Ltd., Tokyo, Japan) operated at an accelerating voltage of 5 kV with a secondary electron detector (SED).
To directly visualize the presence of gradients within the ITPed hydrogels, fluorescent dyes were incorporated into the polymer precursor solutions. Methacryloxyethyl thiocarbamoyl rhodamine B (Polysciences Inc., 0.002% w/v) or fluorescein isothiocyanate (FITC)-dextran (Mw 2,000,000 Da, Sigma, 0.5 mg/mL) were added to the precursors prior to crosslinking and subsequent ITP processing (0.5 M Ca^2+^, 1 min). The resulting hydrogels were then transferred to deionized water and incubated for 5 min at RT before imaging under either an optical microscope (Nikon SMZ-10 Trinocular Stereomicroscope, Tokyo, Japan) or a fluorescence microscope (Zeiss LSM 710, Oberkochen, Germany).
Live/dead staining
2.11
The viability of cells was assessed using live/dead staining comprised of FDA and EB. The staining solution was freshly prepared by mixing 1 mL of FDA solution (1.5 mg/mL in DMSO) and 0.5 mL of EB solution (1 mg/mL in PBS) with 0.3 mL PBS (pH 8). 20 μL of staining solution per 1 mL of culture medium was added to the sample medium and incubated for 5 min at RT. Fluorescence images of the samples were taken using a Nikon Eclipse TE300 fluorescence microscope (Nikon, Tokyo, Japan) equipped with a 14 MP Aptina Color CMOS digital camera (AmScope, Irvine, CA).
In vitro cartilage-like tissue engineering, biochemical quantification, and histological staining
2.12
The hMSCs from donor 1 at passage 4 (P4) were used to evaluate the bio-orthogonality of Ca^2+^ and EDTA treatments on stem cell phenotype and stemness, whereas the hMSCs from donor 2 at passage 5 (P5) were used for the chondrogenic cartilage-like TE and programming studies. The cells were cultured in T175 flasks and harvested for encapsulation when they reached ∼80% confluence. Square cell-laden hydrogels with a length of 18 mm were fabricated as above (Section 2.4). For bio-orthogonality evaluation, the constructs were divided into three groups: (i) positive control (PC), receiving no treatment; (ii) ITPed, treated with Ca^2+^ transfer printing (0.5 M, 5 min, RT); and (iii) Recovered, treated with EDTA following ITP (5 mM, 10 min, 37 °C). The bioconstructs were cultured in 8 mL of GM for 24 h, after which a portion (N = 2) was collected for immunofluorescence staining, and the remaining samples (N = 3) were transferred to 8 mL of chondrogenic medium for differentiation studies over 21 days. For cartilage-like tissue formation and programming, the hMSC-laden hydrogels were cultured in 8 mL of chondrogenic medium in a humidified incubator at 37 °C with 5 % CO_2_ over a course of 21 days, and 4 mL of medium was changed every other day. The cartilage-like tissues were obtained at day 21 and used for biochemical analysis, histological staining, and ITP-induced shape programming and reprogramming. For controls, cell-laden hydrogels of identical dimensions, prepared under the same conditions, were cultured in GM and collected on day 21 for further investigation.
For biochemical analysis, the engineered cartilage tissues were first homogenized in 0.5 mL of papain buffer (Sigma) at 0 °C for 1 min and then digested in a total of 1.0 mL of papain solution (Sigma) at 65 °C for 24 h and centrifuged for 10 min at 15,000 rpm, and then the supernatants were collected for DNA and glycosaminoglycan (GAG) quantifications (N = 3).
Per the manufacturer's instructions, a Picogreen assay kit (Invitrogen) was used to quantify the DNA content in the supernatant. Fluorescence intensity of the dye-conjugated DNA solution was measured using a microplate reader with an excitation of 480 nm and emission of 520 nm.
The GAG content was quantified using a 1,9-dimethylmethylene blue (DMMB) assay [38]. 40 μL of supernatant from the digested samples was transferred into a 96-well plate, to which 125 μL of DMMB solution was then added. Absorbance at 595 nm was recorded on a microplate reader. GAG content was normalized to DNA content.
The cartilage-like tissues were fixed in 10% neutral buffered formalin (NBF) overnight at 4 °C, dehydrated, and embedded in paraffin. Briefly, tissue samples were cut into 5 μm thick sections using a Leica RM2255 rotary microtome (Leica Microsystem Ltd., Milton Keynes, UK). Slide sections were then deparaffinized and stained with Hematoxylin and Eosin (H&E) to observe gross cell and tissue morphology [39], Safranin O (SafO) with a Fast Green counterstain [40] and Alcian Blue (pH 1.0) for GAG indication [41], anti-CD44 antibody for confirming surface CD44 expression [42]. Stained samples were imaged using a fluorescence microscope under bright field.
Results and discussion
3
Dual-crosslinkable OMA macromers [34,43] were synthesized as the main hydrogel material (Scheme S2). The OMA macromers in the presence of a photoinitiator were covalently crosslinked into a stable hydrogel network via photocuring. This was evidenced by the largely weakened methacrylate proton signals (H_a_ and H_b_ in dotted rectangles) in Fig. S1 after photoirradiation. The OMA hydrogels exhibited a dramatic mass loss at Day 1, likely due to the removal of uncrosslinked macromers during incubation. Thereafter, the degradation proceeded gradually under culture in GM in an incubator at 37 °C with 5% CO_2_ (Fig. S2). This photocured OMA hydrogel can be further ionically crosslinked by Ca^2+^ ions to form dual-crosslinked networks. Culturing only photocured OMA hydrogels in aqueous solutions such as water (H_2_O), phosphate-buffered saline (PBS) (pH 7.4), and GM induced volumetric expansion (swelling), accompanied by a small range of variation in storage modulus (G′) and a large decrease in compressive modulus (Figs. S3 and S4). In contrast, culturing the hydrogels in 0.5 M Ca^2+^ solution brought about obvious volumetric shrinkage (deswelling), which was accompanied by a large increase in G’ and compressive modulus. These results confirmed the effective binding of OMAs with Ca^2+^ ions, leading to tightening and toughening of the hydrogel networks.
By introducing a defined ion diffusion pathway through the thickness of the photocrosslinked OMA hydrogel using the ITP strategy, a crosslinking gradient and consequently a porosity gradient were established, as evidenced by SEM analysis (Fig. S5a), and a substantial increase in storage modulus (G′) with longer ITP durations (Fig. S6a). Removal of Ca^2+^ with EDTA resulted in a uniform distribution of enlarged pores (Fig. S5b) and a substantial reduction in G’ values (Fig. S6b). The formation of the gradient structure was further validated through dye incorporation experiments using rhodamine B and FITC, which enabled visual confirmation of the presence of spatially distinct regions (Fig. S7). Since the swelling of OMA hydrogels decreases with the increasing polymer network crosslinking density [12], it was conjectured that the gradient OMA hydrogels resulting from ITP would exhibit out-of-plane deformation with the higher-crosslinking side on the concave side (Fig. S8).
The shape-morphing behavior of the gradient hydrogels strongly depends on the incubation solution, macromer concentrations, Ca^2+^ concentration of the ion reservoir, and ITP time. To examine the impact of those parameters on the shape-morphing behaviors, hydrogel bars were fabricated as a simplified prototype. As can be seen in Fig. 1, all those parameters exerted a clear impact on the shape morphing. Generally, parameters contributing to larger swelling and/or greater gradient range across the bar thickness give rise to more pronounced hydrogel deformation. For example, hydrogel bars cultured in H_2_O displayed a significantly higher bending due to the significantly higher swelling (Fig. S3b) than those cultured in PBS (pH 7.4) and GM (Fig. 1a–S9), while the diminished hydrogel swelling due to the increase in macromer concentration (Fig. S10) resulted in a rapid decrease in the bending angle (Fig. 1b). Likewise, increasing Ca^2+^ concentration (Fig. 1c) and extension of ITP time (Fig. 1d) that benefits a larger gradient range led to larger bending angles. The results suggest that the deformation output could be programmed by tuning these parameters.Fig. 1. The impact of (a) incubation solution (macromer concentration: 3%; Ca^2+^ concentration: 0.5 M; ITP time: 1 min; culture time: 5 min), (b) macromer concentration (Ca^2+^ concentration: 0.5 M; ITP time: 1 min; culture media: GM; time: 5 min), (c) Ca^2+^ concentration of ion reservoir (macromer concentration: 3%; ITP time: 1 min; culture media: GM; culture time: 5 min), and (d) ITP time (macromer concentration: 3%; Ca^2+^ concentration: 0.5 M; culture media: GM; time: 5 min) on the shape-morphing behaviors of hydrogel bars at RT. Hydrogel dimensions: 15 mm (length) × 2 mm (width) × 1 mm (thickness). N = 3, data are presented as mean ± SD.Fig. 1
Previous studies have demonstrated that hydrogel dimensions, including length, width, and thickness, play a role in shape-morphing behavior, offering a useful strategy for programming hydrogel constructs [44,45]. To evaluate the effects of these parameters in these gradient-crosslinked hydrogels, hydrogel strips with varying lengths, widths, and thicknesses were tested (Fig. S11). The bending angle increased with both length (10–20 mm) and thickness (0.6–1.4 mm), indicating a strong dependence on these dimensions. In contrast, variations in width (1.5–2.5 mm) showed a negligible impact on bending deformation.
Next, gradient OMA hydrogels as cell-laden scaffolds were subjected to incubation in GM to investigate their shape-morphing properties. NIH3T3 cells with a density ranging from 5 to 100 million (M) cells/mL macromer solution were encapsulated and the bending behavior of the resulting hydrogel bars was examined. Similar to the cell-free counterparts, the cell-laden hydrogel bars quickly morphed into a “C” shape in about 3 min in cell culture medium (Fig. 2a). There was a downward trend in the bending angle of the cell-laden hydrogel bars with increasing cell density (Fig. 2b). Extended culture in GM led to gradual strain relaxation, resulting in a progressive reduction in bending angle (Fig. S12a and b) and a concurrent decline in compressive modulus (Fig. S12c). This behavior is likely attributed to limited crosslinking depth and dynamic rearrangement of the ionic crosslinks, resulting from the exchange of Ca^2+^ with monovalent cations in the surrounding media during incubation [46].Fig. 2. Shape-morphing properties of cell-laden hydrogel bars. (a) Representative photographs show the bent shapes. (b) Quantitative bending angles. (c) Photomicrographs of bright field images. (d) Photomicrographs of representative live/dead fluorescence images, captured immediately after shape programming. ∗p < 0.05 compared with other groups, except for the “0” cell density group. Experimental conditions: macromer concentration: 3%; Ca^2+^ concentration: 0.5 M; ITP time: 1 min; culture media: GM; culture time: 5 min. Hydrogel dimensions: 15 mm (length) × 2 mm (width) × 1 mm (thickness). N = 3, data are presented as mean ± SD.Fig. 2
With the same method, deformed large bioconstructs were obtained through programming large bioconstructs, including cell-rich hydrogel discs and square hydrogel slabs (Fig. S13). As expected, the cells within the bent hydrogel bars were presented in round morphology as shown in Fig. 2c and maintained high viability indicated by the predominantly green-colored cells in the live/dead staining results (Fig. 2d and S14), wherein the live cells were visualized fluorescently with green color by FDA, while dead cells were visualized with red color by EB. The cytocompatibility of the shape-morphing system, together with its preprogrammable deformability, suggests its reliability for use in 4D TE.
The fabrication of cell-laden bioconstructs with complex configurations was explored by harnessing a “patterning ITP” approach. NIH3T3 cells at a density of 20 M/mL hydrogel precursors were used to fabricate bioconstructs unless otherwise stated, representing a relatively high cell density within the hydrogels. In a simple demonstration, we segmentally patterned hydrogel bars to enable asymmetrical, multi-directional bending, yielding “1D-to-2D” shape transformations to afford complex structures with “S” or “S”-like shapes (Fig. 3). We next advanced the pattern design to create more complex 3D architectures. Ca^2+^-soaked filter papers with specific geometries were used to create patterned Ca^2+^ crosslinking within disc-shaped bioconstructs (Fig. 4). Attributed to the locally suppressed swelling within the Ca^2+^ patterns and the significantly larger swelling in the non-patterned regions, various complex 3D bioconstructs with “wave” or “flower” like structures were obtained via “2D-to-3D” shape transformations after being cultured in GM for 5.0 min. Thus, this simple patterning process was demonstrated to be effective in manipulating the bioconstruct shapes via a set of programmed cooperative deformations.Fig. 3. Segmental patterning of cell-laden hydrogel bars and the corresponding resultant multi-directional bending complex 2D bioconstructs. Hydrogel length: 30 mm (upper), 39 mm (middle), and 39 mm (bottom), hydrogel width (2 mm), pattern length: 15 mm. Experimental conditions: macromer concentration: 3%; Ca^2+^ concentration: 0.5 M; ITP time: 1 min; culture media: GM; culture time: 5 min. Scale bars = 5 mm.Fig. 3. Fig. 4Patterning of cell-laden hydrogel discs and the resulting complex 3D bioconstructs after deformation. The gray areas represent the patterned regions. Hydrogel diameter: 20 mm; diameter of disc pattern: 10 mm; width of “Y” and “+” patterns: 0.3 mm. Experimental conditions: macromer concentration: 3%; Ca^2+^ concentration: 0.5 M; ITP time: 5 min; culture media: GM; culture time: 5 min. Scale bars = 5 mm.Fig. 4
The anisotropic internal strain caused by the gradient Ca^2+^ crosslinking can be erased by removing the Ca^2+^ from the hydrogel networks with EDTA, which in turn relaxes the bent hydrogel bars and further leads to hydrogel straightening. A subsequent reprogramming via a second ITP process could induce the hydrogel re-bending in a prescribed manner (e.g., bending in the opposite direction). As such, shape reprogramming could be achieved. Fig. 5a illustrates a typical reprogramming process. Cell-free hydrogel bars were first employed to verify reprogrammability. Our previous study demonstrated that cell-laden constructs maintained high viability following treatment with 5 mM or 10 mM EDTA [20]. Thus, a concentration of 5 mM EDTA was employed in this study. As hypothesized, bent hydrogel bars in H_2_O were able to recover to their relaxed states (near-straight shapes) after soaking in EDTA (5 mM) for about 1 h at RT. The straightened hydrogel bars at RT regained their deformed states (“C” shapes) in about 10 min of culture after a second ion printing (Fig. S15a). The results demonstrated that the shape transformation was reversible. Switching the hydrogel bars back and forth between the two distinct states occurred easily and the response time could be shortened by increasing the incubation temperature to 37 °C (Fig. S16). However, signs of fatigue were observed in the following relaxation–reprogramming cycles (Fig. S15b), which could be ascribed to the combined effects of the rearrangement of local networks by the repetitive breaking and reforming of the physical bonding [47,48] and the already swollen networks reducing the range of the resultant gradient crosslinking density. A similar repetitive switching between the relaxed and deformed states by the same reprogramming approach was also observed for cell-rich hydrogel bars (Fig. 5b and c), in which opposite-bending programming, quantified by the negative bending angles, was also demonstrated. The encapsulated cells remained highly viable after cycling five times (Fig. 5d and S17). Unlike previous work on 4D biofabrication that often demonstrates only “2D-to-3D” shape transformations, unique “3D-to-3D” shape transformations by taking advantage of the reprogrammability were realized in large bioconstructs as shown in Fig. 6. In comparison to “2D-to-3D” transformations, the implementation of “3D-to-3D” transformation is deemed much more challenging [49,50] and may be valuable in 4D TE for its powerful potential to better mimic the deformations and movements occurring in 3D-structured living organisms [51].Fig. 5. Reprogramming of cell-rich hydrogel bars at RT. (a) Schematic illustration of the reprogramming process. (b) Switch of a representative hydrogel bar between stretched and bent states under cell culture conditions. (c) Quantitative bending angles of hydrogel bars at the two states over repetitive cycles. (d) Representative live/dead photomicrographs from different cycles. Experimental conditions: macromer concentration: 3%; Ca^2+^ concentration: 0.5 M; ITP time: 1 min; culture media: GM; culture time: 5 min. Hydrogel dimensions: 15 mm (length) × 2 mm (width) × 1 mm (thickness). N = 3, data are presented as mean ± SD.Fig. 5. Fig. 6Reprogramming of (a) a hydrogel slab and (b) a hydrogel disc. Length/width of hydrogel slab: 15 mm; pattern parameters: 3 mm width for cross pattern and 10 mm sides for square pattern. Diameter of hydrogel disc: 15 mm; pattern parameters: 10 mm diameter for disc pattern and 3 mm width for “Y” patterns. Experimental conditions: macromer concentration: 3%; Ca^2+^ concentration: 0.5 M; ITP time: 5 min; culture media: GM; culture time: 5 min.Fig. 6
The remarkable shape-morphing capability, together with the reprogrammability, inspired us to explore the system's capacity for driving reprogrammable 4D shape transformation TE. Specifically, 4D chondrogenesis of hMSCs at a density of 10 M/mL hydrogel precursors was employed as a proof-of-concept model. To promote better cell adhesion to hydrogel networks and better cell survival during long-term culture, 3% (w/v) OMA mixed with 0.5% (w/v) cell-adhesive GelMA [52,53] was used as scaffolding material. The incorporation of a small amount of GelMA did not influence the shape morphing ability of the OMA hydrogels (Fig. S18). hMSC-laden bioconstructs in the shape of square slabs were subjected to differentiation in chondrogenic medium over a course of three weeks (21 days), during which time the bioconstructs exhibited no morphological conversions but only linear expansions (Fig. S19), and cell viability was demonstrated to be high (Fig. S20). Biochemical quantification and histological staining were performed to assess the differentiation. The results of the biochemical analysis revealed that, in comparison with the control group (Ctrl, bioconstructs cultured in normal GM), the experimental group (Exp, bioconstructs cultured in differentiation medium) had comparable DNA levels (Fig. 7a) but significantly higher GAG production, a primary cartilage ECM component (Fig. 7b and c). Also, Exp exhibited a significantly higher storage modulus than Ctrl (Fig. S21), indicative of the generation of cartilage-like tissue constructs. Histological staining with H&E and SafO showed that the cells on D21 were uniformly distributed throughout the bioconstructs and resided primarily in small pockets surrounded by hydrogel (Fig. 7d and e). Moreover, the intense staining of SafO further confirmed substantial GAG production of the engineered cartilage-like tissues. However, the evident SafO staining observed in the Ctrl group may result from nonspecific dye binding to the alginate hydrogels, although the staining intensity was visibly weaker than that of the Exp constructs.Fig. 7Evaluation of chondrogenesis. Biochemical quantification of (a) DNA levels, (b) GAG production, and (c) GAG/DNA ratios in the control (Ctrl) and experimental (Exp) groups. Photomicrographs of representative H&E and SafO histologically stained samples from (d) the Ctrl group and (e) the Exp group. ∗p < 0.05. N = 3, data are presented as mean ± SD.Fig. 7
Constructs in different shapes, obtained by cutting the tissue-like specimen, were ITPed to evaluate the shape morphability and reprogrammability of differentiated constructs after 3 weeks of culture in chondrogenic pellet medium (CPM). Like the undifferentiated samples, ion-printed tissue discs and square slabs could also deform into curled geometries (Fig. 8a). The results of reprogramming experiments also demonstrated the reprogrammability of these engineered tissues (Fig. 8b and S22). However, the deformation extent of these tissue-like bioconstructs was much less pronounced compared to that of undifferentiated bioconstructs, due to the morphing constraints stemming from the extensive stiff ECM produced. Such manipulation of differentiated bioconstructs allows for the construction of functional living tissues with high structural dynamism and complexity, which could be highly valuable for personalized biomedicine for its ability to facilitate seamless integration of engineered tissues into tissue defects of specific patients [54].Fig. 8. Programming the shape of engineered tissue-like bioconstructs after 21 days of culture in CPM. (a) Deformation of the ion-printed tissue disc and tissue slab. (b) Multiple deformations of a tissue bar via reprogramming. Experimental conditions: Ca^2+^ concentration: 0.5 M; ITP time: 5 min, cultured in GM for 5 min. Construct dimensions: disc, 20 mm in diameter × 1.0 mm in thickness; square sheet, 15 mm in length × 1.0 mm in thickness; bar (prior to culture), 18 mm in length × 2.0 mm in width × 1.0 mm in thickness.Fig. 8
Although Ca^2+^ and EDTA were employed to program the shape of hMSC-laden constructs, these reagents could potentially influence cell phenotype and stemness due to fluctuations in calcium concentration. Specifically, EDTA can chelate intracellular Ca^2+^, which is essential for maintaining cell viability and physiological functions. To evaluate the effects of Ca^2+^ and EDTA treatment on hMSC chondrogenic potential, hMSC-laden bioconstructs were subjected to ITP-based programming followed by recovery with EDTA. The expression of the hMSC surface biomarker CD44 was examined via immunofluorescence staining, and chondrogenic differentiation was assessed, with untreated (PC) constructs serving as controls. Strong CD44 staining was observed across all groups (Fig. S23). Furthermore, all groups exhibited comparable levels of chondrogenic differentiation after 21 days of culture, with no statistically significant differences observed (Fig. S24). These results suggest that Ca^2+^ and EDTA treatments, under the applied conditions, did not exert a discernible impact on hMSC differentiation. Nevertheless, although Ca^2+^/EDTA proved effective for programming cell-laden hydrogels, the resulting morphologies underwent gradual relaxation during culture (Fig. S12). Therefore, the use of ITP programming prior to differentiation is constrained by the material system in this study. Future implementation of this strategy using cytocompatible polymers capable of forming stronger crosslinked networks through ITP may overcome this limitation.
The focus on biomimicry of native tissue development has recently aroused the need for the development of 4D TE technologies that can achieve dynamic morphological changes over time [55,56]. Emerging 4D TE accounts for not only regenerating tissue substitutes with complex configurations but also offers opportunities to imitate the intrinsic dynamism of living tissues [[57], [58], [59]] and thus may be advantageous in some applications over traditional morphologically static TE, which generally lacks the feature of tissue dynamics [60]. For example, through simple self-rolling mechanisms, flat 2D hydrogels can transform into 3D tubular tissues with tunable lumen sizes for potentially partially replicating the geometry of portions of native blood vessels, trachea, esophagus, intestine and urinary tract that may be difficult to fabricate using other technologies [17,61]. Additionally, morphable scaffolds may enhance minimally invasive surgery by enabling in situ deployment and potentially better integration into irregular tissue defects. A prominent example includes self-expanding scaffolds that are delivered in a temporary configuration and recover their functional shape post-deployment [[62], [63], [64]]. In the present system, rapid post-programming deformation enables delivery in a compact state, followed by spontaneous deployment through intrinsic strain relaxation. However, if long-term shape retention is required, additional structural stabilization strategies would be necessary.
Geometrically dynamic hydrogels can also conform to curved tissue topographies through preprogrammed deformation, while reprogrammable scaffolds offer on-demand self-adaptation to varying curvatures, making them particularly useful for complex defect reconstruction [65,66]. Notably, such adaptability has recently enabled sutureless anastomosis for resected tissues, streamlining surgical procedures [67]. Beyond structural advantages, 4D scaffolds could offer additional biofunctional benefits with the potential to generate anisotropic and evolving mechanical cues to guide cell behavior and subsequently promote tissue remodeling and regeneration [21,68,69]. Although still largely in the conceptual stage, such dynamically reprogrammable systems offer distinct advantages over static scaffolds by enabling spatiotemporal modulation of geometry and internal strain, thereby better recapitulating aspects of tissue morphogenesis and providing a controllable platform to study complex tissue regeneration processes in vitro [59].
The fast, facile biofabrication approach described herein to establish a reprogrammable scaffold by using inexpensive techniques and simple materials opens a new paradigm for 4D TE applications. Our model provides a robust platform to manipulate construct architectures in multiple ways on demand for both undifferentiated and differentiated bioconstructs at defined times. The shape transformation is driven by the differential swelling caused by the crosslinking density gradient across the hydrogel thickness. We engineered large bioconstructs with sophisticated topologies, such as wavy and flower-shaped structures. High cell survival was detected before and after shape morphing. Impressively, “3D-to-3D” shape transformations have been achieved simply by releasing and rebuilding the internal anisotropic strain. The potential to reprogram the shape of engineered tissues was also demonstrated. The developed system outperforms conventional multilayering approaches on 4D biofabrication when it comes to simplicity, efficiency, cytocompatibility, and deformability, and provides morphodynamical means to possibly regulate the biomaterial mechanical properties in a spatiotemporal manner, enhancing the capacity to use scaffolding materials to direct the fate of encapsulated cells [70,71]. Furthermore, the ability to program and reprogram the geometry of cell-laden hydrogels, without reconstructing entire scaffolds, introduces a unique strategy for engineering adaptive living materials, in which structural transformation can be coupled with biological function for applications such as targeted drug delivery, biosensing, soft robotics, bioelectronics, and bioactuators [72,73].
While this platform shows great promise for engineering tissues with complex and tunable architectures, it is important to note that shape transformations using this strategy may need to primarily occur prior to in vivo implantation, as the ITP process and shape recovery are not currently readily applicable to constructs that have already been implanted. Moreover, many tissue defects do not provide sufficient space for large out-of-plane deformations. The ability to directly manipulate in vivo implanted constructs could enable adaptive integration with host tissues that dynamically evolve in shape [74], and the current limitation of the ITP strategy in this context warrants further development. Nonetheless, the self-relaxation behavior of ITPed (deformed) structures in cell GM presents unique opportunities for self-deploying implants that can conform to targeted sites upon delivery.
Conclusion
4
On the basis of ITP technology, this work has demonstrated a simple strategy to enable multi-scale structural morphing of soft tissues via physiologically relevant stimuli. Preprogrammable and reprogrammable shape morphing of bulk cell-laden alginate hydrogels occurs as a result of forming reversible gradient crosslinking. Deliberately tuned parameters, including culturing medium, ITP time, macromer concentration, and Ca^2+^ reservoir concentration, in conjugation with specific local crosslinking incorporation by means of surface patterning, allow for controlled, multiple, multi-directional deformations, generating complex tissue constructs at various stages of maturation. Multi-dimensional switching between 1D, 2D, and 3D structures endows the system with excellent biofabrication capabilities, making it adaptable for customized, complex biological and physiological environments. The results collectively presented in our study reveal a reliable platform for 4D TE, which could also be a valuable tool for investigating the role of dynamical architecture in mechanobiology, developmental morphogenesis, bioactuation, and beyond.
CRediT authorship contribution statement
Aixiang Ding: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Conceptualization. Fang Tang: Writing – original draft, Methodology, Formal analysis. Sriramya Ayyagari: Methodology, Formal analysis. Eben Alsberg: Writing – review & editing, Supervision, Resources, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization.
Ethics approval and consent to participate
No direct animal or human experiments were performed in this study. Human mesenchymal stem cells (hMSCs) used in the research were obtained from the Core Facility of Case Western Reserve University under an Institutional Review Board (IRB)-approved protocol, and provided in a de-identified form with no identifiable personal information disclosed. All experimental procedures involving biological materials were carried out in accordance with the relevant ethical guidelines and regulations.
Declaration of competing interest
Eben Alsberg is an associate editor for Bioactive Materials and was not involved in the editorial review or the decision to publish this article. All authors declare that there are no competing interests.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Varner V.D.Nelson C.M.Cellular and physical mechanisms of branching morphogenesis Development 141142014275027592500547010.1242/dev.104794 PMC 4197615 · doi ↗ · pubmed ↗
- 2Ikada Y.Challenges in tissue engineering J. R. Soc. Interface 31020065896011697132810.1098/rsif.2006.0124 PMC 1664655 · doi ↗ · pubmed ↗
- 3Tamay D.G.Dursun Usal T.Alagoz A.S.Yucel D.Hasirci N.Hasirci V.3D and 4D printing of polymers for tissue engineering applications Front. Bioeng. Biotechnol.7201910.3389/fbioe.2019.00164 PMC 662983531338366 · doi ↗ · pubmed ↗
- 4Lee Y.B.Jeon O.Lee S.J.Ding A.Wells D.Alsberg E.Induction of four-dimensional spatiotemporal geometric transformations in high cell density tissues via shape-changing hydrogels Adv. Funct. Mater.31242021201010410.1002/adfm.202010104 PMC 832384534335134 · doi ↗ · pubmed ↗
- 5Peng M.Zhao Q.Wang M.Du X.Reconfigurable scaffolds for adaptive tissue regeneration Nanoscale 15132023610561203691956310.1039/d 3nr 00281 k · doi ↗ · pubmed ↗
- 6Zhao Q.Wang J.Wang Y.Cui H.Du X.A stage-specific cell-manipulation platform for inducing endothelialization on demand Natl. Sci. Rev.7320206296433469208210.1093/nsr/nwz 188PMC 8289041 · doi ↗ · pubmed ↗
- 7Montgomery M.Ahadian S.Davenport Huyer L.Lo Rito M.Civitarese R.A.Vanderlaan R.D.Wu J.Reis L.A.Momen A.Akbari S.Pahnke A.Li R.-K.Caldarone C.A.Radisic M.Flexible shape-memory scaffold for minimally invasive delivery of functional tissues Nat. Mater.16102017103810462880582410.1038/nmat 4956 · doi ↗ · pubmed ↗
- 8Nuttelman C.R.Rice M.A.Rydholm A.E.Salinas C.N.Shah D.N.Anseth K.S.Macromolecular monomers for the synthesis of hydrogel niches and their application in cell encapsulation and tissue engineering Prog. Polym. Sci.33220081671791946194510.1016/j.progpolymsci.2007.09.006PMC 2390836 · doi ↗ · pubmed ↗