4D Printing of Biocompatible, Hierarchically Porous Shape Memory Polymeric Structures

Abstract

Conventional implants tend to have significant limitations, as they are one-size-fits-all, require monitoring, and have the potential for immune rejection. However, 4D Printing presents a method to manufacture highly personalized, shape-changing, minimally invasive biomedical implants. Shape memory polymers (SMPs) with a glass transition temperature (T_g) between room and body temperature (20–38 °C) are particularly desirable for this purpose, as they can be deformed to a temporary shape before implantation, then undergo a shape change within the body. Commonly used SMPs possess either an undesirable T_g or lack the biocompatibility or mechanical properties necessary to match soft biological tissues. In this work, Poly(glycerol dodecanoate) acrylate (PGDA) with engineered pores is introduced to solve these issues. Pores are induced by porogen leaching, where microparticles are mixed with the printing ink and then are dissolved in water after 3D printing, creating a hierarchically porous texture to improve biological activity. With this method, highly complex shapes were printed, including overhanging structures, tilted structures, and a “3DBenchy”. The porous SMP has a T_g of 35.6°C and a Young’s Modulus between 0.31 and 1.22 MPa, comparable to soft tissues. A one-way shape memory effect (SME) with shape fixity and recovery ratios exceeding 98% was also demonstrated. Cultured cells had a survival rate exceeding 90%, demonstrating cytocompatibility. This novel method creates hierarchically porous shape memory scaffolds with an optimal T_g for reducing the invasiveness of implantation and allows for precise control over elastic modulus, porosity, structure, and transition temperature.

1. Introduction

As commercial 3D printers become more affordable, additive manufacturing, mainly 3D printing, has received a spike in interest [1]. 4D printing, first introduced by Skylar Tibbits [2], describes a subset of additive manufacturing where 3D printed structures are no longer static, and instead can be programmed to change across time, the “4^th dimension”. 4D printing allows for precisely manufactured shape-changing structures, which have a wide range of applications including aerospace [3], manufacturing [4], biomedicine [5], and food [6]. Within the biomedical field, 4D printed shape memory scaffolds (SMS) have desirable properties for the development of drug delivery systems [7, 8], vascular repair [9, 10], tissue engineering [11, 12], and bone reconstruction [13]. Of particular interest is how implantable 4D-printed devices can change the adversity of surgery from being invasive to being minimally invasive [14]. Before 4D printing, SMS fabrication methods had significant limitations in the complexity, customizability, and manufacturing speed of designed structures [15, 16].

Among various materials used for 4D printing, SMPs are a class of smart materials that provide a mechanical action in response to external stimuli, usually temperature. Previous work from our lab explored a photo- and thermal-curable SMP of poly(glycerol dodecanoate) acrylate (PGDA) [14], which is modified from poly(glycerol dodecanoate) (PGD).[17] PGD has been used extensively for tissue engineering [18, 19]. PGDA is a photocurable SMP that improves the rheological properties of PGD, allowing for more effective 3D printing. PGDA is a semi-crystalline polymer that can be crosslinked to establish a permanent shape. Then, if the crosslinked shape is brought above T_g, it becomes rubbery and can be deformed to a temporary shape. Upon cooling, the temporary shape is fixed through crystallization. If the polymer is heated above T_g again, the crystals melt to a highly mobile, amorphous phase, where the chemical crosslinks deform the structure to recover the permanent shape. PGDA solves a primary limitation of the most used SMPs, like polylactic acid (PLA), polycaprolactone (PCL), and polyurethane (PU). They have intrinsic, undesirable transition temperatures, either too much higher than body temperature (38 °C) or far below room temperature (20 °C). Having a transition temperature in this range (20°C-38 °C) is preferred, as programmed structures will undergo temporary shape programming and permanent shape recovery simply by being placed in a physiological environment. A biomedical device with this property could be deformed to a much smaller shape before being implanted, reducing the invasiveness of the procedure, before returning to its original programmed structure.

In this work, we 4D print porous PGDA structures by mixing porogen (pore-creating medium) microparticles (MPs) with the PGDA ink. The porogen MPs can be leached after printing, thus creating microsize pores throughout the printed PGDA structures. Since macropores can be created by printing self-suspended complex structures, this creates a texture with tunable porosity at different length scales, making them hierarchically porous. This process has several advantages. First, it can manufacture structures that more closely mimic the extracellular matrix of biological macromolecules such as collagen. Second, the porosity increases the available surface area for cells to adhere, as proved by the presence of carboxyl groups within the PGDA network [20]. Third, porosity can influence the mechanical properties of the SMS, and a closer match to the tissues can improve cell behavior [21]. Methods for creating a polymer matrix with micropores include gas foaming [22], freeze-drying [23], electrospinning [24, 25], and humidity-induced phase change [26], which tend to present obstacles when using a thermoset SMP, like PGDA, as a starting material. While 3D-printed porous polymers have been explored to a limited extent, 4D-printed hierarchically porous polymeric scaffolds are even more limited [13, 27, 28]. Moreover, their T_g values are usually out of the desirable range of 20 °C-38°C or they are not biocompatible.

After testing several potential porogens [27, 28], sugar (sucrose) was chosen due to its rapidity of dissolution in water. Rheological properties were then measured with a micro rheometer for varying PGDA: Sugar weight ratios to balance printability with porosity. 3D structures of varying complexity were printed, which demonstrated a shape memory effect (SME). Mechanical properties were tested for hierarchically porous constructs, showing utility within physiological settings. Biocompatibility was then tested via cell culturing to prove that porous 4D-printed PGDA structure could be potentially used for implantation.

2. Materials and Methods:

2.1. Materials Synthesis

The shape memory thermoset PGDA was synthesized according to our previous work [14]. To synthesize Pre-PGD, Glycerol (99%, synthetic, ACROS Organics) and dodecanedioic acid (DDA, 99%, Alfa Aesar) were mixed at a molar ratio of 1:1, then heated to 120 °C in an oil bath under argon flow and magnetic stirring for 24 hours. Next, 40 g of Pre-PGD was dissolved in a solution made up of 0.2 g of 4-methoxyphenol (99%, Acros Organics), 0.4 g of 4(dimethylamino) pyridine (99%, Alfa Aesar), 7.1 mL of triethylamine (Fisher Chemical), and 400 mL of methylene dichloride (DCM, Fisher Chemical). Then, this solution was cooled in an ice bath under nitrogen flow for 10 min. 6 mL of acryloyl chloride (96%, Alfa Aesar), pre-diluted in 60 mL methylene dichloride, was then added. The reaction vessel was then wrapped with aluminum foil during these steps to prevent premature cross-linking. After 12 hours of stirring at room temperature, an additional 0.2 g of 4-methoxyphenol was added. The DCM was removed in a rotary evaporator and a vacuum chamber. Then, the dried mixture was dissolved in 20 mL of ethyl acetate (99%, Acros Organics). The solution was centrifuged at 10,000 rpm for 15 min. Using a rotary evaporator and vacuum chamber, the supernate was dried. It was then melted and magnetically stirred while 0.1 g of the photoinitiator 2,2-dimethoxy-2-phenylacetophenone was added. For the biocompatibility test, 0.4 g of the photoinitiator 2,4,6-trimethylbenzoyldiphenyl phosphine oxide (TPO) was used instead. Finally, the solution was dried to obtain a printable PGDA powder.

The sucrose (>99.5%, at a molecular biology grade, Sigma-Aldrich) was ground by hand using a mortar and pestle and then passed through different sizes of sieves to limit the maximum diameter of micropores. These sieves were 180 μm “coarse”, 125 μm “fine”, and 35 μm “extra-fine”. The PGDA powder was melted and mixed at varying weight ratios with the sugar powder using a conditioning mixer (THINKY AR-100) to prepare a printable mixture.

2.2. Printing Process

A standard 10 mL syringe was loaded with the liquid PGDA-sugar mixture, then placed inside the Allevi 2 Bioprinter. An M19 gauge nozzle (Internal Diameter = 0.686mm) was attached. The syringe was heated to 48–56 °C using Allevi’s corresponding online software. The precise temperature was varied with each print to optimize the print speed and layer solidification. A supply of air was used pneumatically at pressures between 5 and 20 psi depending on the print conditions. The tip of the nozzle was occasionally heated with a heat gun to prevent clogging before each print. The programmed shape of each deformable 3D shape was designed with Computer Aided Design (CAD) software (SolidWorks), then processed to create a G-code script using a slicing program (Slic3r). G-code files were uploaded to the online Allevi software. The designed structure was then 3D printed.

2.3. Porogen Leaching and Curing

3D printed samples were initially brittle with no SME. After printing, the samples were placed in a deionized water bath (Water, Deionized Distilled, ASTM Type II, Sigma-Aldrich) for one week, with the water changed at least three times, to allow the sugar to fully dissolve into the water, leaving behind a structure of micropores. To measure this dissolving rate, 3D printed samples were weighed before dissolving, then soaked in DI water and weighed each day. This weight change, which was used to determine the soaking time for samples, is tabulated in Figure S1. After leaching, the samples were crosslinked under Ultraviolet (UV) light with an intensity of 10 mW cm^−2 for 10 minutes. Then, they were thermally cured in a vacuum oven at 140 °C for 4 hours to produce the structure with desirable material properties. According to our previous work, this post-curing duration results in a T_g between the room temperature and body temperature [14].

2.4. Rheological and Thermal Characterization

Samples of the PGDA-sugar mixture at different weight ratios were loaded into a micro rheometer (Anton Paar Modular Compact Rheometer MCR 302) with a 10 mm flat plate geometry and a 1 mm gap to characterize the flow properties and how this impacted printability. Rheology was tested at varying temperatures, ramping from 65 °C down to 35 °C at a 5 °C increment. TA Q20 Differential Scanning Calorimetry (DSC) was performed on the PGD-Asugar mixtures with varying ratios at cyclically ramped temperatures from −20 °C to 100 °C at a rate of 20 °C/minute.

2.5. Mechanical Testing

Storage modulus, loss modulus, and tangent delta were measured using Dynamic Mechanical Analysis (DMA) (Hitachi Dynamic Mechanical Analyzer DMA7100) with a 3D printed rectangular 50 mm × 10 mm × 1.5 mm sample. Tensile testing was performed using an ASTM-D638 Type IV dog bone shape specimen. The specimen was printed, crosslinked, and made porous using the above procedure. Pore size was varied based on the size of sugar MPs used, either “extra-fine”, “fine”, or “coarse”. Each porosity used the same 1:1 weight ratio of PGDA to sucrose. Tensile mechanical properties were measured with a digital motorized stand (Mark-10 ESM303 with Series 5 Force Gauge). The tensile testing was performed both at room temperature and after heating the dog bone sample to the rubbery state above T_g with a heat gun.

2.6. Shape Memory Characterization

Shape recovery testing was performed with a 3D-printed rectangular 30 mm × 5 mm × 2.5 mm sample. The sample was heated in a 60 °C water bath before a 90° deformation was created and it was submerged in a room-temperature water bath. After five minutes, the fixity angle was measured. The sample was then returned to the 60 °C water bath and the one-way SME occurred. Finally, the sample was cooled, and the recovery angle was measured. Shape fixity ratio (R_f) and Shape recovery ratio (R_r) were calculated as described previously [29].

2.7. Porosity Testing

Scanning electronic microscopy (SEM, FEI Quanta 600F Environmental SEM) was used to examine the surface and internal porosity of printed scaffolds, both at different sugar particle sizes and different weight ratios. The sugar particle sizes of “extra-fine”, “fine”, and “coarse” delineations were used in the tensile testing, at a 1:1 PGDA-to-sucrose weight ratio. Surface porosity with the “fine” particle size was also compared to a 2:1 PGDA-to-sucrose ratio.

2.8. Biocompatibility Testing

To test the biocompatibility of the printed materials, co-culturing was performed with C3H10T1/2 cells in DMEM supplemented with 10% FBS, 2mM L-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C in a humidified atmosphere with 5% CO₂. The materials were first sterilized in ethanol and then incubated in DMEM for 48 hours before being put in the co-culture assay. Cell viability was assessed after 1, 2, 3, and 4 days of co-culturing with the materials, and images were captured using a Nikon microscope. Fluorescent images were taken on the indicated day, with cells on coverslips incubated in propidium iodide (PI) staining solution at 4 °C for 15 minutes. The samples were then fixed in 10% formalin at RT for 10 minutes, and fluorescent images were captured using a Keyence microscope and processed using ImageJ. Percent survival was calculated by dividing the number of cells positive for PI by the total number of cells.

2.9. Statistical Analysis

All experiments were performed a minimum of three times. Experimental data were then analyzed using single-factor analysis of variance (ANOVA). Results were listed as mean ± standard deviation. A p-value < 0.05 was considered statistically significant.

3. Results and Discussion

3.1. Porous PGDA Synthesis

Printing hierarchically porous PGDA structures while maintaining a shape memory effect can be difficult. Extrusion-based 3D printing, while it is highly accessible, requires the printing ink to reach elevated temperatures, which has the potential to prematurely cross-link the thermoset. This is the primary reason why PGD was modified to the photocurable PGDA. By exposing 3D printed samples to UV light, photo-crosslinking occurs, preventing the PGDA from melting at the elevated temperatures necessary for thermal crosslinking and setting the T_g. To generate a porous matrix, a method must be selected to function even with the restrictions of both extrusion and thermal crosslinking. In this context, porogen leaching was selected as a candidate. A sacrificial porogen (sucrose) was added to the printing ink and then could be removed by dissolution before thermal crosslinking to leave behind a porous texture, which is desirable for tissue engineering applications [20, 21]. By tuning the size and concentration of the porogen, different porosities could be generated for different applications.

The process outlined in Figure 1a was the method used for printing hierarchically porous structures. First, the PGDA was melted and mixed with sugar MPs, typically at a 1:1 weight ratio. Second, the mixture was 3D printed using a pneumatic extrusion printer. Third, the printed samples were left in a DI water bath to allow for the sugar particles to dissolve. Finally, the samples were crosslinked under UV light and then in a vacuum oven to set the T_g and generate 4D-printed structures with SME. Both a 2D “square clover” (Figure 1b) shape and a 3D “pentagonal flower” structure (Figure 1c) demonstrated one-way shape memory response where a shape could be recovered via heating above T_g after being deformed to a different temporary shape.

Figure 1.

(a) Scheme showing the printing process of porous PGDA structures. PGDA is synthesized and then mixed with sugar as a 3D printable ink. After printing, the structures are soaked in water to allow the sugar to dissolve. Porous constructs are then crosslinked with UV light and then in a vacuum oven. (b) Photographs showing a shape memory cycle of a “square clover”. (c) Photographs showing a shape memory cycle of a “pentagonal flower”. Scale bar: 1 cm.

3.2. Properties of Pre-Crosslinked PGDA-Sucrose

Different PGDA-to-sugar weight ratios were explored to optimize printability and porosity. Herein, the smallest “extra-fine” sugar particle size (< 35 μm) was used. First, the rheological properties of the melted mixture were characterized using a micro rheometer. Pure PGDA was compared to PGDA-sugar mixtures at a 2:1 ratio and a 1:1 ratio at various temperatures (Figure 2a–c). Higher sugar ratios were also explored but could neither be consistently extruded from the 3D printer nozzle nor effectively measured with the rheometer setup. As demonstrated, a lower temperature results in a higher viscosity, and a higher weight ratio of sugar leads to a higher viscosity. Viscosities as a function of the temperature for inks with different sugar weight ratios at a constant shear rate of 100 s^‒1 are summarized in Figure 2d. A Shear thinning effect (Figure 2e) was also prevalent in each tested sample at 35 °C, but considerably more prevalent at the higher sugar weight ratios, with the 1:1 ratio achieving a maximum viscosity of 19.4 kPa s. This shear thinning effect at a temperature lower than the printing temperature is useful as the ink rapidly cools while exiting the nozzle. Recent work has shown the shear thinning effect to be a desirable property in extrusion-based printing [30, 31], improving shape fidelity and printability. The 1:1 ratio was then used for all further printing, as it significantly increased the shear thinning effect, even at higher temperatures. Characteristic DSC curves of the sample printed from inks with different sugar weight ratios are shown in Figure 2f. Please note that because these inks are not yet UV or thermally crosslinked, the peaks in heat flow represent the melting temperature rather than T_g. Samples must be printed prior to setting T_g, both to preserve extrusion and to allow for a SME.

Figure 2.

(a) Viscosity vs. shear rate plot for pure PGDA at varying temperatures. (b) Viscosity vs. shear rate plot for an ink with PGDA-to-sugar weight ratio of 2:1 at varying temperatures. (c) Viscosity vs. shear rate plot for an ink with PGDA-to-sugar weight ratio of 2:1 at varying temperatures. (d) Viscosities vs. temperature plot of inks at different sugar weight ratios at a constant shear rate of 100 s^−1. (e) Viscosities vs shear rate plot of inks at different sugar weight ratios at 35 °C, demonstrating a shear thinning behavior. (f) Characteristic heat flow vs. temperature curves of inks at different sugar weight ratios.

3.3. 3D Printed Structures

After achieving sufficient extrusion characteristics with the PGDA-sugar thermoset, more complex 3D structures were printed to demonstrate the possible printing of a wide array of geometries from ink made of 1:1 PGDA-sugar mixture. Figure 3 compares a variety of 3D printed structures to the computer models used for generating G-code. Successful prints included a 45-degree tilted conic structure, which represents a difficult angle for thermosets to maintain the shape during the print process [32](Figure 3a). Both a cylindrical lattice, which was used for later SEM imaging and sugar removal testing (Figure 3b) and a standard lattice (Figure 3c) were also printed. A thin-walled tube (Diameter > 10·thickness) (Figure 3d), an overhanging structure (Figure 3e), and a complicated 3DBenchy model used for benchmarking 3D printing capabilities (Figure 3f) were also printed successfully.

Figure 3.

3D models (left) and corresponding photographs (right) for (a) a 45° hollow truncated cone; (b) a cylindrical lattice; (c) a 3D lattice; (d) a thin-walled tube; (e) an overhanging bridge; (f) a 3D Benchy boat. Scale bar: 1 cm.

3.4. Mechanical Properties

After printing, the samples were further UV and thermally crosslinked. The mechanical properties of these crosslinked porous PGDA were measured via DMA and tensile testing, shown in Figure 4. DMA results (Figure 4a) show a clear drop of the storage modulus when the temperature crosses T_g. The loss modulus follows a similar trajectory, while the tangent delta peaks first at T_g (~35.6 °C) and again at ~ 54 °C. This T_g value confirms its possibility as an implantable biomedical device, as it could undergo an SME simply triggered by body temperature. If necessary, T_g can be tuned lower by increasing the thermal crosslinking time [14].

Figure 4.

Mechanical properties of 4D printed porous PGDA samples. (a) DMA curves showing storage modulus, loss modulus, and tangent delta at varying temperatures. (b) Tensile stress-strain curves for the porous PGDA samples below T_g. (c) Tensile stress-strain curves for the porous PGDA samples above T_g.

Characteristic stress-strain curves, both below T_g (Figure 4b) and above T_g (Figure 4c), were measured for samples with varying pore sizes (“extra-fine”, “fine”, or “coarse”). Key mechanical properties of Young’s Modulus, strain at fracture, and stress at fracture are compared in Table S1 both below T_g and above T_g. Below T_g, there was not a statistically significant difference for the Young’s Moduli, strain at fracture, or stress at fracture between the three pore size groups. Because of existing imperfections caused by the stress concentrations of printed infills, the mechanical properties of the samples were measured to be substantially lower than PGDA itself [14]. Above T_g, there still was not a statistically significant difference for the measured strain at fracture or stress at fracture for the three pore size groups. However, there was a substantial difference in Young’s Modulus, decreasing from 1.22 MPa to 0.31 MPa between the “extra-fine” and “coarse” samples. Though the samples have the same composition, this difference indicates that the orientation and size of internal pores can alter the mechanical property. This is like an occurrence in 3D printed structures, where samples can have the same infill ratio but different Young’s Moduli depending on their infill pattern and orientation [33] or depending on the nozzle diameter and printing speed [34]. This range of Young’s Moduli effectively matches the stiffnesses of human soft tissues. Matching stiffness to the host tissue is important in preventing inflammation caused by mechanical mismatch [35, 36]. These 3D-printed porous constructs have an elastic modulus that aligns well with the soft tissues in the cardiovascular system [37] and the Achilles tendon [38]. Additionally, it was previously demonstrated that these values can be increased substantially by increasing the thermal curing time [14].

3.5. Porosity Characterization

The porosity of the printed hierarchically porous matrix was explored. Figure 5 shows the SEM images, which illustrate how the sugar weight ratios and particle sizes impact the porosity of the printed PGDA. By progressively zooming in upon the surface of the construct (Figure 5a), we can see different scales of pores ranging from the macroscale lattice, the mesoscale leached sugar MPs, and the microscale PGDA’s structure. Their sizes can have a significant impact on cell migration [39] depending on the size of the cell and the host tissue. The pores created by varying the sugar particle sizes were imaged in the cross-section (Figure 5b) and statistically analyzed (Figure 5c). The “extra-fine” MPs led to a pore diameter of 17.3± 5.5 μm, while the “fine” ones produced pores with a diameter of 61.9 ± 18.7 μm, and the “coarse” ones resulted in pores with a diameter of 114.6 ± 28.9 μm. These pores are correlated with many different cells throughout the human body. The “extra-fine” ones match the myocytes and monocytes, while the “fine” ones match endothelial cells [40]. These pore sizes were also used to select the cells used for culturing.

Figure 5.

Porosity characterization of the printed PGDA. (a) SEM images of a 3D lattice at different magnifications: (i) Scale bar: 1 cm; (ii) Scale bar: 100 μm; (iii) Scale bar: 10 μm. They display hierarchical porosity progressively zoomed on the surface. (b) Cross-sectional SEM images showing pore distribution when using (i) extra fine, (ii) fine, and (iii) coarse sugar MPs as the porogen. Scale bar: 100 μm. (c) Distributions of pore diameters recorded via SEM imaging when using (i) extra fine, (ii) fine, and (iii) coarse sugar MPs as the porogen.

The result of altering sugar concentration on the pore distribution was also investigated. Figure S2 shows SEM images of the surface porous PGDA made from inks with PGDA-to-sugar weight ratios of 1:1 and 2:1. Clearly, a higher concentration of sugar resulted in a higher concentration of pores in the final construct. This ratio can be tuned to alter the mechanical properties or total porosity based on the host tissue that is being mimicked. However, the sugar concentration cannot be tuned excessively higher than 1:1, as issues with the extrusion such as clogging were created.

3.6. Shape Memory Effect

3D printed “square clover” and “pentagonal flower” structures were used as demonstrations of the SME (Figure 1b–c). The structures were first raised to an elevated temperature of above T_g, before being deformed to a temporary shape, then allowed to cool below T_g. This temporary shape hardens and can then be held for an extended period. If the stimulating temperature is above T_g, the structure returns to the originally programmed shape within seconds.

To quantify the shape memory characteristics of the printed samples, R_f and R_r were measured according to previous work [29]. Data for R_f and R_r are tabulated in Figure S3. R_f was calculated to be 98.5% ± 1.4% and R_r was calculated to be 98.6% ± 0.8%. This cycle can be reproduced consistently, and previously a very high R_f and R_r were maintained even after 100 cycles of PGDA’s SME [14]. Additionally, the shape memory effect demonstrated by PGDA is rapid, occurring in under two seconds. This was demonstrated previously [12], with recovery time of PGDA structures compared to stimulating temperature. As stimulating temperature increased, the recovery time of PGDA decreased from 1.4 seconds to 0.4 seconds. This shape fixing and recovery is possible because the dodecanoic acid (DDA) becomes crystalized and holds the polymer network for shape fixation when below T_g. Above T_g, it melts to allow the network to freely move, causing shape recovery.

3.7. Biocompatibility

The biocompatibility of the porous PGDA prepared with “coarse” and “fine” sugar MPs was evaluated through in vitro culturing with 10T1/2 cells. The microscope images taken after each day of the four-day culturing show that the PGDA prepared with “coarse” sugar allowed significant cell proliferation (Figure 6a). However, the PGDA prepared with fine sugar showed lower cell density on Day 4, leading us to hypothesize that there may be sucrose remaining inside the scaffold that could cause cell death. To test this hypothesis, we used glucose-free DMEM and found that while control cells began starving to death after Day 2, cells co-cultured with PGDA prepared with fine sugar continued to proliferate (Figure S4a), indicating that the remaining sucrose in the scaffold was indeed causing cell death.

Figure 6.

(a) Representative microscope images of 10T1/2 control cells and cells co-cultured with porous PGDA captured on day 1, 2, 3, and 4, respectively. Scale bar: 100 μm. (b) Representative images of 10T1/2 control cells or cells co-cultured with the PGDA captured on day 1, 2, 3, and 4, respectively. Propidium Iodide (PI, red) was used to detect dead cells (white arrows), and nuclei were stained with DAPI (blue). Scale bar: 10 μm. (c) Quantification of survival rates of control cells and cells co-cultured with PGDA. Statistical analysis was conducted for each day (Kruskal-Wallis test followed by Dunn’s multiple comparisons test) (p-value for Day 4 Sugar F = 0.003 v.s. Control). Note: Sugar F and Sugar C indicate the fine and coarse size sugar, respectively.

Fluorescence images of 10T1/2 cultured in dishes with 70% PGDA SMP films after 1, 2, 3, and 4 days are shown in Figure 6b, with living and dead cells stained blue and red, respectively. The number of cells increased, and there was no significant difference in cell survival between the control and co-cultured with PGDA prepared with the coarse sugar, indicating great biocompatibility of the materials. The cell viability (Figure 6c and Figure S4b) showed that the cell survival rates were above 90% (the death rates were lower than 10%) after 4 days, indicating that the printed porous PGDA fabricated has an excellent cytocompatibility.

4. Conclusion

In this work, by using sugar MPs as the sacrificial material (porogen), we report 4D printing of hierarchically porous SMS with a T_g appropriate for application in biomedical implantation. 3D structures like an overhanging bridge and a 3DBenchy were successfully fabricated, showing the wide applicability of this printing methodology. Successful biocompatibility was demonstrated through cell culture, indicating the method’s potential in generating multifunctional, minimally invasive biomedical implants. A repeatable SME was shown, with shape fixity and recovery ratios of 98.5% and 98.6% respectively. Additionally, comparable mechanical properties to soft biological tissues were also demonstrated, including a tunable elastic modulus between 0.31 MPa and 1.22 MPa. By tuning the size and concentration of sugar MPs, different porosities, elastic moduli, and cell responses could be generated. Each printed material can be precisely controlled for glass transition temperature, porosity, elastic modulus, and programmed shape.

Porous 4D-printable PGDA presents a useful addition to the biomedical field, particularly in tissue engineering. Using this method allows manufacturers to precisely control both porosity and elastic modulus by altering the porogen ratio. While current implanted biomedical devices are often one-size-fits-all, require removal or monitoring, have the potential for immune rejection, and use invasive surgical procedures, this method gives users the flexibility to personalize every aspect of an implant. Hierarchically porous shape memory scaffolds can match the extracellular matrix of host tissues more effectively, while 4D printing allows the scaffolds to undergo a programmed shape change to decrease invasiveness. Future studies could be investigated using the hierarchically porous texture for aneurysm occlusion or cardiovascular stents. The tunable, hierarchically porous structure may also allow for potential loading with different substances, including magnetic materials and drugs for application in on-demand drug delivery.

Highlights:

• Porous shape memory scaffolds were prepared using 4D printing and porogen leaching.

• Shape memory scaffolds matched the mechanical properties of soft tissues.

• Biocompatibility was demonstrated via cell culturing.

• Shape memory effect enables minimally invasive implantation.