Authors: Kwanghyun Baek, Junggeon Park, Eunmi Kim, Ryan Miller, William Ballance, Yongbeom Seo, Yu-Tong Hong, Jaehyun Jeong, Hyunjoon Kong
Categories: Article, hydrogel, reconfigurable hydrogel, minimally invasive, biopatch, drug delivery
Source: ACS applied materials & interfaces
Authors: Kwanghyun Baek, Junggeon Park, Eunmi Kim, Ryan Miller, William Ballance, Yongbeom Seo, Yu-Tong Hong, Jaehyun Jeong, Hyunjoon Kong
Hydrogel is widely used for the sustained delivery of bioactive molecules that can treat various injuries, diseases, and tissue defects. However, inserting hydrogel implants without disrupting their functionality and microstructure often requires a large incision, leading to potential complications, such as infection, scarring, and pain. The gel implant is often manually rolled and inserted through a catheter for a minimally invasive delivery. However, success heavily depends on the user’s skills, which can inadvertently damage the implant. To address this issue, we developed a reconfigurable hydrogel patch that can self-fold into a small tube and unfold spontaneously after implantation through a catheter. The hydrogel path was assembled by layering a drug-releasing poly(ethylene glycol) diacrylate (PEGDA) hydrogel sheet onto a PEGDA and polyethylenimine (PEI) hydrogel sheet, which rapidly swells and degrades homogeneously at controlled rates. The dynamics of the self-folding and unfolding process could be controlled by differences in the expansion ratio and elastic modulus between the two gel layers according to a mathematical model that closely matched experimental results. The unfolding process triggered a sustained release of the protein cargo. Specifically, the reconfigurable gel loaded with angiopoietin 1 significantly enhanced neovascularization, nearly doubling the vascular density compared to the control group following implantation through a tube with 15% smaller diameter than the original shape of the gel patch. This gel biopatch will be broadly useful for the minimally invasive delivery of a wide array of therapeutic molecules, potentially enhancing therapeutic outcomes.
A series of soft material-based bioimplants enabling controlled molecular release, optical/electrical/magnetic sensing and stimulation, and cell deployments have been developed to treat tissue defects and various acute and chronic diseases.^1-3^ These implants are assembled to present sophisticated structures and functional compartments using various microfabrication techniques. Implantation of these implants without disrupting predefined microstructure and functionality inevitably necessitates open surgery and, in turn, causes discomfort, scar formation, and even tissue morbidity. Therefore, extensive efforts have been made to develop methods for delivering implants in a minimally invasive manner.
In the past, minimally invasive delivery systems often utilized injectable hydrogels through sol–gel transition.^4^ This sol–gel transition, induced by various factors such as temperature, ions, enzymes, light, self-healing, and pH, allows for a relatively easy injection process.^5^ However, the resulting gel formed after injection may not permit the presentation of sophisticated microstructures or shapes. To address this issue, the injection of shape-memory hydrogel or the use of rolling up postgel delivered through a catheter has recently been implemented.^6-8^
However, these platforms still have several limitations. First, an injectable preformed hydrogel with shape-memory properties is limited to specific size or structure, which restricts its scope of application. Second, the rolling process sometimes results in material fracture depending on the user’s skill level and is only suitable for relatively small implants. To overcome this problem, certain efforts have been made to prepare a bilayer gel that can self-fold due to the differing swelling ratios of its two layers by using a thermo-responsive polymers or tuning the cross-linking density.^9^ However, few studies demonstrated the gel’s ability to unfold and recover its original shape after delivery.
To this end, this study presents a reconfigurable hydrogel patch that can self-fold before implantation, unfold on the implanted tissue after implantation, and finally release bioactive molecular cargo at the implanted site. We hypothesized that a bilayered gel consisting of a drug-carrying gel layer and a sacrificial gel programmed to degrade at a desired rate would self-fold and unfold in a controlled manner. Using the reconfigurable gel system, we delivered the drug-releasing gel patch to the desired site through a small catheter. We examined the hypothesis by using a hydrogel formed by a Michael-type cross-linking reaction between poly(ethylene glycol) diacrylate (PEGDA) and polyethylenimine (PEI) as a model sacrificial gel (Scheme 1). The protonated amine groups of PEGDA–PEI hydrogel expedite osmotic water influx into the gel, leading to rapid swelling and dissolution.^10^ The expansion and subsequent degradation time of the PEGDA–PEI gel could be tuned from 10 min to 1 h by controlling the mass ratio between PEGDA and PEI. This sacrificial gel was attached to a drug-encapsulating PEGDA gel used as a model gel implant.
The transformation rate was characterized by real-time imaging. The underlying mechanism was addressed by calculating the residual stress between the two gel layers. Finally, the PEGDA gel layer loaded with angiopoietin 1 (Ang1), known to stimulate neovascularization, was integrated into the PEGDA–PEI sacrificial gel layer. The resulting construct was implanted on a chicken chorioallantoic membrane (CAM) by using a catheter via a small hole to simulate the minimally invasive implantation procedure of the patch for neovascularization.
A 20% (w/w) PEGDA (M.W. 400 g/mol, Polysciences, Inc.) solution in DI water, including 0.1% Igacure 2959 (Ciba specialty chemicals), was first polymerized by UV irradiation (Jelight Co. Model 20) at a wavelength of 253 nm for 5 min between glass plates with varying gap sizes. Then, a PEGDA–PEI hydrogel layer was prepared on the first layer by coating it with a mixture of varied concentrations of PEGDA and branched PEI (average M.W. 800 g/mol, Sigma-Aldrich). Finally, the bilayered gel was immersed in DI water or PBS to trigger self-folding into a hydrogel tube.
Capillary forces between glasses were utilized to create a thin gel layer less than 200 μm thick. Briefly, the solution was dropped on a 2 cm × 2 cm cover glass, and then, a slide glass with 7.5 cm × 2.5 cm was placed on top, allowing the solution to fill the area between the cover glass and the slide glass. The combined glass setup, held together by capillary force, was suspended on supporters. The volume of the solution was calculated by multiplying the desired thickness by 400 mm^2^. The hydrogels immersed in water were recorded by using a video camera (Leica D-Lux), and the radius of the hydrogel was analyzed from still frames of the videos. The measured values were compared with mathematically calculated radii.
Measurement of expansion The expansion ratio of each hydrogel was calculated by measuring the degree of swelling with eq 1. The degree of swelling was characterized by measuring the mass of water taken up by the gel. The mass of the fully hydrated gel in deionized water for 24 h and the mass of the gel dried through lyophilization were measured. Finally, the degree of swelling was calculated from the mass ratio of the hydrated gel to the dried gel.
Measurement of an elastic To measure an elastic modulus, each hydrogel was prepared in the form of a disc with a 1 cm diameter and 1 mm thickness. Compressive stress and strain of each disc-shaped gel were measured by a uniaxial compression test at a rate of 1 mm/min using a mechanical testing system (Insight, MTS Systems). The elastic modulus was calculated from the initial slope of the stress–strain curve at a 10% strain.
Analysis of protein To characterize the protein release profile of the hydrogel, 2 μg/mL BSA (bovine serum albumin obtained from Sigma-Aldrich) was loaded into the PEGDA gel layer of the bilayered gel. A single-layered PEGDA hydrogel loaded with BSA was also prepared as a control group. Both the bilayered and single-layerd gels were prepared via the fabrication protocol described above. Each hydrogel was incubated in PBS (pH 7.4, Thermo Scientific) at 37 °C. Supernatants were collected at various time points, and the protein concentration was measured using a micro-BCA assay (Thermo Scientific) following the protocol provided by the company. To quantify the cumulative percentage of BSA released, the amount of BSA back-calculated from the measured absorbance was normalized to the total amount of BSA initially loaded into the gel.
Fertilized chicken eggs (Hy-Line W-36) were sourced from the University of Illinois Poultry Farm (Urbana, IL). The eggs were initially incubated horizontally for 7 days at 37 °C and 65% humidity. A 2.3 mm diameter hole was then created at the top of the eggshells, and self-folded gel tubes were placed on the CAMs through a PTFE tube inserted into the hole. The eggs were further incubated for 7 days at 37 °C. Vascular networks in the CAMs were imaged using a Leica S6E stereomicroscope connected to a Leica D-Lux E camera. On the seventh day postimplantation, the embryos were fixed with 10% neutral buffered formalin (3.7% formaldehyde in PBS). The centers of the hydrogels and CAMs of each sample were excised, embedded in paraffin, sectioned, and stained with an α-smooth muscle actin antibody (Abcam). The density and the number of positively stained blood vessels within 2 mm from the center of the gel implants were quantified using NIH ImageJ software. Blood vessel densities were calculated by measuring the total blood vessel area and normalizing it to the total cross-sectional area of the CAM.
All experiments were performed in triplicate unless stated otherwise. Statistical significance was calculated using one-way analysis of variance (ANOVA) with Tukey’s posthoc comparison of the means using Origin Pro 9.1 software at a significance level of 0.05.
Sacrificial PEGDA–PEI hydrogels were formed by mixing aqueous PEGDA with branched PEI at controlled mass ratios. The elastic modulus of the resulting PEGDA–PEI hydrogel was increased from 1 to 3 MPa by altering PEGDA concentrations from 10 to 20 (w/w) % at a constant PEI concentration of 10 (w/w) % (Figure 1a). The elastic modulus of the gel also varied between 0.5 and 1.0 MPa by altering the PEI concentration at a given PEGDA concentration of 10 (w/w) % (Figure 1b).
The resulting gels expanded over the course of several hours and eventually dissolved into the media.^10^ This gel degradation was characterized by quantifying the expansion ratio (S) from the measured swelling ratio of the (1)S=(QfQi)1∕3−1 where Qi and Qf are the swelling ratios of a hydrogel before and after incubation in aqueous media, respectively.^9^ The S values of the gels were linearly increased over time. Increasing PEGDA concentration from 10 to 20 (w/w) % at 10 (w/w) % PEI concentration reduced the increasing rate of S over time (Figure 1c). In contrast, increasing PEI concentration from 5 to 15 (w/w) % at 10 (w/w) % PEGDA concentration resulted in a steeper increase of S over time (Figure 1d).
As the S values of the gel became higher than 0.5, the solid mass of the hydrogel started to decrease significantly due to dissolution. Increasing PEGDA concentrations at a 10 (w/w) % PEI concentration increased the gel dissolution time, marked with a final data point of the S versus time curve in Figures 1c, from 30 to 190 min (Figure 1c). Conversely, the gel’s complete dissolution time decreased from 80 to 8 min with increasing PEI concentration at 10 (w/w) % PEGDA concentration (Figure 1d). Overall, the lifetime of the sacrificial PEGDA–PEI gel could be controlled from 8 to 190 min by varying the mass ratio between PEGDA and PEI.
In parallel with an increase in S, the elastic modulus of the gels also decreased over time. Following the same trend as S, increasing PEGDA concentration at 10 (w/w) % PEI concentration reduced the rate of reduction in the elastic modulus (Figure 1e). Conversely, the elastic modulus of the gel was reduced more quickly by increasing the concentration of PEI at 10 (w/w) % PEGDA concentration (Figure 1f). The increased concentration of protonated amine groups of PEI rapidly absorbed much larger amounts of water, leading to rapid dissolution of the gel. These results suggest that the gel expansion over time is caused by the loss of cross-linked junctions responsible for the gel’s elastic response.
We applied this PEGDA–PEI hydrogel as a sacrificial layer that drives the self-folding and unfolding of the PEGDA hydrogel, which has often been used as a drug carrier.^11^ The bilayered hydrogel patch was assembled by first preparing a 0.2 mm thick PEGDA hydrogel. The 20% PEGDA [molecular weight of 400 g mol^−1^] solution was exposed to UV light. Subsequently, the 0.2 mm thickness sacrificial hydrogel layer (PEGDA–PEI) was prepared by placing a PEGDA and PEI solution mixture over the pure PEGDA hydrogel. The chemical reaction between PEI and the remaining acryl functional group of the prefabricated PEGDA hydrogel layer led to a stable adhesion between the two gel layers.
Immersion of the bilayered hydrogel patches with 10 mm length × 10 mm width × 0.2 mm thickness into deionized (DI) water or phosphate buffer saline (PBS) triggered self-folding into multiwalled gel tubes within a minute (Figure 2a and Movie S1). The PEGDA concentration in the PEGDA–PEI gel modulated the tube radius. In particular, increasing PEGDA concentration from 10 to 20 (w/w) % at the 10 (w/w) % PEI concentration decreased the tube radius from 1.4 to 0.6 mm (Figure 2b). In contrast, the tube radius was independent of the PEI concentration (Figure 2c). Increasing gel thickness from 80 to 800 μm linearly increased the tube radius from 0.5 to 4.5 mm (Figure S1).
Afterward, the self-folded gel tube unfolded into a gel sheet due to the dissolution of the sacrificial PEGDA–PEI gel layer (Figure 3a and Movie S2). The changes in the gel tube radii over time were fitted to a mathematical model initially developed to estimate the curvature of a bilayer strip in which each layer has the same thickness [eq 2]: (2)r=h12∈(E1E2+14+E2E1) where r is the inner radius of the gel tube, E1 and E2 are the elastic moduli of the bottom PEGDA gel layer and the top sacrificial PEGDA–PEI gel layer, respectively, h is the thickness of each gel layer, and ε is the difference in the expansion ratio between two gel layers (Figure 3b,c and Supporting Note).^12,13^ Note that E1 remains constant over time, while E2 decreases, as shown in Figure 1e,f. Over time, experimentally measured inner radii exhibited changes comparable to those estimated by using eq 2. We suggest that the initial decrease in the inner radius of the gel tube during folding is related to the increase of ε. The subsequent degradation of the PEGDA–PEI gel decreased E2, leading to an increase in the inner radius. Therefore, the shape change of the PEGDA gel is attributed to continued changes in the ε values and E1∕E2. The gel tube radius, which was also estimated to vary with gel layer thickness and incubation time, showed good agreement with the experimentally measured values (Figures S1 and 3b,c).
The time for the bilayered gel tubes to unfold into a sheet, referred to as complete opening time, was varied from 5 to 180 min by increasing the PEGDA concentration at 10 (w/w) % PEI concentration (Figure 3d). In addition, we could decrease the total opening time from 70 to 5 min by increasing the PEI concentration at 10 (w/w) % PEGDA concentration (Figure 3e). The complete opening time was linearly related to the degradation time of the PEGDA–PEI gel, as determined in Figure 1c,d (Figure 3f). This result suggests that the complete gel opening time could be tuned by altering the composition of the PEGDA–PEI gel. Interestingly, the slope between the degradation time and complete opening time was 0.85, not 1.0, indicating that the gel tubes unfold slightly before the PEGDA–PEI gel degrades fully.
Another interesting finding was that the gel tube underwent self-folding and unfolding without any fracture, except for the gel made with 20 (w/w) % of PEGDA and 10 (w/w) % of PEI (Figure 4a). To determine the source of the fracture, we calculated the residual stress at the interface between the PEGDA gel layer and the PEGDA–PEI gel layer. The residual stress at the interface (σ) was calculated by (3)σ=Δ∈E1E2E1+E2+hE1(E1−E2)2r(E1+E2) where E1 and E2 are the elastic moduli of the PEGDA gel layer and the PEGDA–PEI gel layer, respectively, r is the inner radius of the gel tube, h is the thickness of each gel layer, and Δε is the difference in expansion between the two gel layers (Supporting Note).^12,13^ According to the analysis, the self-folded gel tube experienced a continuous increase in σ during degradation of the PEGDA–PEI gel layer. σ decreased during unfolding. The maximum σ increased with an increase in the PEGDA concentration at a given PEI concentration (Figure 4b) or a decrease in the PEI concentration at a given PEGDA concentration (Figure 4c). Notably, the patch made with 20 (w/w) % of PEGDA and 10 (w/w) % of PEI showed an increase in σ to nearly 100 kPa. This residual stress value is higher than the maximal tensile stress of the pure 20 (w/w) % PEGDA gel layer, causing interfacial failure at a strain of 0.16 mm/mm (Figure S2). Therefore, the bilayered gel tube with this composition could not endure the residual stress.
The biopatch was evaluated for cytocompatibility before application in drug delivery. We incubated NIH3T3 fibroblasts with the biopatch for 24 h (Figure S3). Live/dead cell imaging showed that the fraction of viable cells was 94.1 + 1.9%, which was not statistically different from untreated cells and cells incubated with the PEGDA hydrogel patch. This result aligns with previous studies reporting that low-molecular weight PEI, such as PEI (MW 800 g/mol) used in assembling the PEGDA–PEI sacrificial layer, has low cytotoxicity.^14,15^ This result confirms that the bilayered biopatch has high cytocompatibility and is a suitable material for biological applications such as drug delivery and tissue engineering. To assess the effectiveness of the reconfigurable patch as a drug delivery system, we first examined the release profile of molecular cargo from the PEGDA gel layer in the bilayered gel. In this analysis, we included 2 μg/mL of BSA in the PEGDA gel layer to monitor its release profile using the micro-BCA assay (Figure 5a). Interestingly, the PEGDA–PEI gel layer in the self-folded gel tube prevented the initial burst release of BSA from the PEGDA gel layer. Before the gel tube unfolded into a sheet, the gel tube released BSA more slowly than did the pure PEGDA gel patch. Once the unfolding process began with the dissolution of the PEGDA–PEI gel layer, the PEGDA gel layer released BSA at a rate similar to that of the pure PEGDA gel patch. This result indicates that this reconfigurable bilayered gel is advantageous in preventing the premature release of molecular cargo during delivery into a body.
Next, we assembled a bilayered gel patch in which the PEGDA gel layer was loaded with angiopoietin 1 (Ang1). Ang1 promotes neovascularization by binding to the Tie2 receptor on endothelial cells and, in turn, enhances vascular maturation and stability.^16,17^ The PEGDA gel layer with 40 μm thickness and 10 mm diameter was attached to the PEGDA (10%)–PEI (5%) gel layer (Figure 5b). When immersed in phosphate buffered saline (PBS), the bilayered gel self-folded into a tube with a length of 1 cm and an outer diameter of 1.5 mm.
After 30 min of incubation in PBS, the folded tube was transferred onto the CAM of an 8-day fertilized egg via a polytetrafluoroethylene (PTFE) tube (1.8 mm inner diameter and 2.1 mm outer diameter) inserted through a 2.3 mm hole created in the eggshell (Figure 5c). To monitor the unfolding of the gel tube on the implanted CAM, the eggshell was opened and monitored with a video camera. The folded gel tube was completely unfolded on CAM within 30 min (Figure 5d).
After 7 days of incubation at 37 °C, the CAM implanted with the Ang1-loaded bilayered gel patch (Condition III in Figure 5e) showed the highest vascular density compared with the untreated CAM (Condition I in Figure 5e) and the CAM implanted with a blank bilayered gel patch (Condition II in Figure 5d). We further evaluated whether Ang1 released from the bilayered gel patch improves endothelial maturity by identifying the alpha-smooth muscle actin layer on the endothelium with immunohistochemical staining (Figure 5f). As expected, the Ang1-loaded bilayered gel patch led to the largest number of mature blood vessels compared to other conditions (Figure 5f,g). The CAM model provided a relatively simple and efficient experimental platform to evaluate the basic functionality and potential of our system. However, to further substantiate our findings, it is essential to conduct systematic studies using more complex mammalian models under relevant pathophysiological conditions. In future research, we plan to pursue this direction, aiming for a more comprehensive validation. Moreover, while this study focused on establishing a design rule for reconfigurable biopatch, exploring the effects of other therapeutic proteins beyond Ang1 is an important further study to enhance the diversity and effectiveness of our system. Our findings are a solid foundation for these subsequent studies and will serve as crucial preliminary data for evaluating efficacy in more advanced in vivo models.
The results of this study demonstrate that the reconfigurable hydrogel system significantly reduces the invasiveness required for the hydrogel patch implantation. The self-folding of the hydrogel patch into a tube, driven by the difference in the swelling ratio between PEGDA and PEGDA–PEI gel layers, reduces the surface area of the original gel patch by up to 70%, thereby reducing the required invasion area to 2.3% of the area required for implantation of the original gel patch. The subsequent degradation of the sacrificial PEGDA–PEI gel layer, tuned by the mass ratio between PEGDA and PEI, drives the unfolding of the gel tube into a sheet, completing the minimally invasive delivery process. The mechanism underlying these self-folding and unfolding processes, which occur without any gel fracture, could be explained well by computational models developed to predict the change in diameter of the gel tube and interfacial stress.
We propose that this reconfigurable gel patch is advantageous in preventing complications and infections that often result from implantation through open surgery.^18,19^ This gel system will also promote faster wound recovery and reduce the discomfort from pain and inflammation. Combined with robotic-assisted surgery, this patch system will significantly enhance the quality of patient-friendly medication.^20,21^
This reconfigurable hydrogel patch also offers significant advantages in controlling the release rate of the protein drug, Ang1. First, the PEGDA–PEI sacrificial layer envelops the PEGDA gel layer in the self-folded gel tube, limiting the initial burst release of the drug during delivery to the target tissue. Second, the PEGDA–PEI gel dissolves at a controlled rate, thus reducing stress on the self-folded PEGDA gel and allowing it to revert to its original shape. Unlike other material systems that degrade slowly and nonhomogeneously, this system ensures controlled unfolding and maintains the drug release profile of the original gel patch, resulting in desired, enhanced neovascularization. Therefore, this biomaterial is broadly useful for delivering a wide array of bioactive molecules of interest after minimally invasive delivery.
In conclusion, a reconfigurable hydrogel patch was developed for the minimally invasive delivery of Ang1, which is capable of enhancing neovascularization. The gel patch was assembled by layering an Ang1-releasing PEGDA hydrogel layer onto a PEGDA–PEI hydrogel that degraded homogeneously at a controlled rate. This bilayered gel self-folded into a multiwalled gel tube in media, enabling delivery through a plastic tube with a diameter only 15% of the original disk-shaped gel patch. Upon implantation, the gel tube unfolded into its original shape due to degradation of the PEGDA–PEI gel layer. This unfolding process also triggered the sustained release of Ang1 from the PEGDA gel, significantly enhancing neovascularization on the implanted CAM compared to that under control conditions. A mathematical model, which aligned well with experimental results, confirmed that the differences in elastic moduli and expansion between the two gel layers control the degree of self-folding and unfolding as well as residual stress at the interface that can lead to structural disintegration during self-folding. This reconfigurable hydrogel system can be broadly useful for the minimally invasive delivery of various protein drugs, offering several benefits, such as enhanced therapeutic outcomes, reduced pain, and fast wound recovery.