Needle-injectable microcomposite cryogel scaffolds with antimicrobial properties

Porous three-dimensional hydrogel scaffolds have an exquisite ability to promote tissue repair. However, because of their high water content and invasive nature during surgical implantation, hydrogels are at an increased risk of bacterial infection. Recently, we have developed elastic biomimetic cryogels, an advanced type of polymeric hydrogel, that are syringe-deliverable through hypodermic needles. These needle-injectable cryogels have unique properties, including large and interconnected pores, mechanical robustness, and shape-memory. Like hydrogels, cryogels are also susceptible to colonization by microbial pathogens. To that end, our minimally invasive cryogels have been engineered to address this challenge. Specifically, we hybridized the cryogels with calcium peroxide microparticles to controllably produce bactericidal hydrogen peroxide. Our novel microcomposite cryogels exhibit antimicrobial properties and inhibit antibiotic-resistant bacteria (MRSA and Pseudomonas aeruginosa), the most common cause of biomaterial implant failure in modern medicine. Moreover, the cryogels showed negligible cytotoxicity toward murine fibroblasts and prevented activation of primary bone marrow-derived dendritic cells ex vivo. Finally, in vivo data suggested tissue integration, biodegradation, and minimal host inflammatory responses when the antimicrobial cryogels, even when purposely contaminated with bacteria, were subcutaneously injected in mice. Collectively, these needle-injectable microcomposite cryogels show great promise for biomedical applications, especially in tissue engineering and regenerative medicine.

In this study, we fabricated antimicrobial cryogels using methacrylated hyaluronic acid (HAGM) hybridized with CP microparticles at different concentrations (Fig. 1). These microcomposite cryogels exhibited a highly interconnected and dense polymer network that can collapse and be injected through a 16G needle without any mechanical damage (Supplementary Movies 1 and 2). We investigated changes in the physical properties of these cryogels as a result of CP incorporation. Specifically, we determined the effects of CP inclusion on cryogel pore size distribution, interconnectivity, swelling, and mechanical properties. We monitored the release of hydrogen peroxide from CP present within the cryogel scaffolds and how the concentration of CP within the cryogels affected their antibacterial activity. We assessed the antibacterial activity using some of the most virulent bacterial pathogens (i.e., superbugs), including methicillin-resistant Staphylococcus aureus (MRSA) and P. aeruginosa. Next, we assessed the cytocompatibility of cryogels using NIH/3T3 mouse fibroblasts (i.e., cell viability) and the in vitro immunogenic response (i.e., dendritic cell activation) using both CP-free and CPcontaining cryogels. Finally, the in vivo biodegradation and immunogenic response across the cryogels were evaluated using C57BL/6 mice.

Results
Characterization of cryogels. We first characterized the mechanical and structural properties of the fabricated cryogels. Specifically, we examined three cryogel formulations, the first without any CP (i.e., 0% CaO 2 ); the second with 0.1% (w/v) CP (i.e., 0.1% CaO 2 ); and the third prepared with a 0.2% (w/v) CP (i.e., 0.2% CaO 2 ). We first assessed how the cryogel swelling ratio varied as a function of CP concentration. We observed that all cryogel swelling ratios ranged from Q M = 41 to 45, remaining relatively unchanged regardless of the amount of CP incorporated (Fig. 2a). Next, we measured the compressive moduli for each cryogel formulation. The comparison of the compressive moduli revealed that all cryogels exhibited low compressive moduli of approximately 2.2 kPa (Fig. 2b). Similarly, the pore interconnectivity of the cryogels was high (i.e., ~ 80%) and remained similar across the investigated CP concentrations (Fig. 2c). We next characterized the morphological features of the cryogels. Scanning electron microscopy (SEM) imaging show that after cryogelation, a network of large interconnected pores remains (Fig. 2d-f). Overall, the cryogels displayed a continuous, uniform network of interconnected macropores, with the presence of solid CP particles (0.1-0.2% CaO 2 ) embedded within the polymer network ( Fig. 2 h,i). Finally, the average pore sizes ranged from 25 to 30 µm (Fig. 2 g-i) with a similar pore size distribution patterns (Fig. 2j-l) among all three cryogel formulations. Data are presented as the mean ± standard error of the mean (n = 4).
Kinetics of hydrogen peroxide release from cryogels. Hydrolysis of the CP particles entrapped within the cryogels leads to the formation of hydrogen peroxide. The release profile of hydrogen peroxide within the first 5 h was compared between the two (i.e., 0.1% vs 0.2% CaO 2 ) square-shaped cryogels (dimensions: 4 mm × 4 mm × 1 mm). Both cryogel variants showed similar release patterns with over 60% release within the first 15 min followed by a progressively decreasing release pattern within the next 3 h. The 0.1% CaO 2 cryogels Figure 1. Engineering antimicrobial and injectable microcomposite cryogels. (a) Overview of the fabrication process of antimicrobial CP-containing injectable cryogels: (1) cryogels were fabricated using 4% HAGM with different amounts of CP (0-0.2% CaO 2 ); an initiator system (APS/TEMED) is added to an aqueous HAGM solution prior to cryopolymerization at − 20 °C. (2) Cryotreatment involves phase separation with ice crystal formation, cross-linking and gelation. Thawing of ice crystals (porogens) results in an interconnected macroporous cryogel network. (b) Cryogel partially dehydrated over Kimwipe regains its original shape and size after hydration. HAGM cryogels were stained with rhodamine for visualization. (c) Following injection through a 16G hypodermic needle, cryogels regain their original shape and dimensions. (d) Cryogels retain their encapsulated CP after needle injection as indicated by the Alizarin Red S staining (n = 5). www.nature.com/scientificreports/ Another byproduct of CP hydrolysis is calcium hydroxide (i.e., Ca(OH) 2 ) precipitate which has been reported to have limited antibacterial properties. We next compared the effect of hydrogen peroxide and calcium hydroxide on MRSA growth. The bacterial growth was monitored for 24 h using culture broth containing CP, CP supplemented with bovine catalase, and calcium hydroxide. The role of catalase was to break down hydrogen peroxide into water and oxygen, and as a result, neutralizing its antibacterial effect 86 . We observed bacterial growth inhibition only in the presence of CP, while MRSA grew unhindered under the other conditions. This result suggests that the antibacterial activity of CP is predominantly based on hydrogen peroxide activity and that calcium hydroxide had minimal or no effect on bacteria (Fig. 3b,c).
Antibacterial activity of microcomposite CP-containing cryogels. Two pathogens (i.e., MRSA and P. aeruginosa) were used to study the antibacterial properties of cryogels hybridized with CP. First, partially dehydrated cryogels were contaminated with a known density of bacteria [i.e., colony forming units (CFUs)]. After 6 h of bacterial incubation within the cryogels, the number of CFUs was quantified. Figure 4a-c shows the trend observed for MRSA incubated with CP-free and CP-containing cryogels. In the absence of CP particles, bacteria within the cryogels not only remained viable but also underwent growth and expansion. On the other hand, cryogels containing 0.1-0.2% CP exhibited total bacterial growth inhibition. Similar findings were observed in the case of P. aeruginosa (Fig. 4d-f). For example, in the case of CP-free (0% CaO 2 ) cryogels, the   Figure 4b shows the presence of MRSA within CaO 2 -free cryogels. Conversely, Fig. 4c depicts the absence of bacteria within cryogels containing 0.1% CP. Similarly, the presence of P. aeruginosa can be seen in CaO 2 -free cryogels ( Fig. 4e) but was absent for cryogels containing 0.2% CP (Fig. 4f), clearly demonstrating the antibacterial potential of microcomposite CP-containing cryogels.
In vitro cytocompatibility studies of antimicrobial cryogels. We next evaluated the cytotoxicity of the antibacterial cryogels using mouse fibroblasts NIH/3T3 cells. Specifically, after 24 h of incubation, cell attachment to the cryogels was observed by confocal microscopy. All CP-free and CP-containing cryogels (0.1- www.nature.com/scientificreports/ 0.2% CaO 2 ) supported the proliferation and attachment of cells along the polymer walls ( Fig. 5a-c). We observed high cell viability (i.e., ~ 90%) for all cryogels (Fig. 5d). These findings indicate that antibacterial cryogels release enough hydrogen peroxide to kill bacterial pathogens but that the level released is below the toxicity threshold of a commonly used mammalian cell line.
In vitro immunogenic response of antimicrobial cryogels. Dendritic cells (DCs) are antigen-presenting cells that can play a crucial role in mounting an effective immune response. They secrete an array of proinflammatory cytokines, some of which are responsible for T cell differentiation 76,77 . Immunogenicity with bone marrow-derived DCs (BMDCs) was assessed using all three cryogel formulations. We further compared BMDC stimulation with cryogel-free (negative control) and lipopolysaccharide (LPS)-containing (positive control) media. Fractions of activated CD11c + CD86 + and CD11c + MHCII + BMDCs were measured by immunostaining in conjunction with fluorescence-activated cell sorting (FACS) analysis. CP-containing cryogels induced basal expression levels of CD86 and MHCII receptors nearly equivalent to those induced by the negative control ( Fig. 6a-c). Finally, we characterized cell culture supernatants by ELISA for the DC-mediated secretion of several proinflammatory cytokines. In good agreement with the minimally activated DCs, the depicted low In vivo immunogenic response and biodegradation of antimicrobial cryogels. Finally, we examined the immunological response of the antimicrobial cryogels in a mouse model. CP-free cryogels, CP-containing (0.1% CaO 2 ) cryogels, and CP-containing (0.1% CaO 2 ) cryogels contaminated with P. aeruginosa, were www.nature.com/scientificreports/ subcutaneously injected into the backs of C57BL/6 mice. Next, the cryogels were explanted on day 4 and subsequently stained with hematoxylin and eosin (H&E) for histological analysis (Fig. 7a-c). We assessed the cellular infiltration into cryogels as well as their integration within the surrounding tissues. Overall, across the three groups tested, cryogels were surrounded by a thin capsule of fibrin and induced a minimal infiltration of leukocytes (e.g., neutrophils and macrophages). It is worth noting that fibrin production and leukocyte infiltration were slightly more observable for the bacteria-contaminated antimicrobial cryogels. Consistent with their high degree of cytocompatibility and minimal in vitro DC activation, these microcomposite antimicrobial cryogels indicated minimal host inflammatory responses in mice. The biodegradation of biomaterials is an essential criterion when designing implantable medical devices. Therefore, we have assessed the extent of cryogel degradation upon subcutaneous injection in mice. CP-free (0% CaO 2 ) and CP-containing (0.2% CaO 2 ) cryogels were tested and after a 2-month implantation, the explanted cryogels displayed signs of degradation (Fig. 7). Compared to their initial dimensions (4 mm × 4 mm × 1 mm), both types of cryogels were clearly smaller, especially for CP-containing cryogels (Fig. 7d-e). This set of data suggests that the incorporation of CP into cryogels may promote biodegradation.

Discussion
Three-dimensional scaffolds that exhibit a highly porous architecture could be very useful for several biomedical applications especially in tissue engineering 1,75 . Macroporous cryogels have been made from a variety of natural polymers meant to recapitulate the composition and structural properties of the ECM 1,37,44 . For example, by varying the polymerization temperature and cooling rate 87 , the mechanical properties of cryogels could be tuned to match those of the native tissues. Similarly, injectable cryogels with improved mechanical properties compared to their hydrogel counterparts can be fabricated 87,88 . However, microbial infections remain a major challenge associated with scaffolds and biomedical implants. Specifically, in clinical orthopedics, complications related to pathogenic bacterial colonization represent a major barrier to tissue repair and healing 89 . Although various approaches have been explored to confer hydrogels with antimicrobial properties, they have been associated with a number of limitations, including cytotoxicity and poor tissue integration [90][91][92] . Here, we report the fabrication of multifunctional cryogels (i.e., needle-injectable, biodegradable, and with antimicrobial activity) that could improve current strategies for developing scaffolds for tissue engineering.
In our study, we incorporated CP that imparts antibacterial activity to the cryogel scaffolds. The hydrolysis of CP leads to the formation of calcium hydroxide and hydrogen peroxide. Although calcium hydroxide is used as an antimicrobial medication in dentistry 83 , it did not appear to drive the antibacterial activity of CP-containing cryogels at the concentration reported in our study (Fig. 3b). The presence of catalase (i.e., an enzyme that degrades hydrogen peroxide) resulted in the total loss of the antibacterial activity of CP, pointing to the production of hydrogen peroxide as the primary driver of the observed antibacterial effects. Conversely, in medical settings, hydrogen peroxide is used as an antiseptic solution for wound disinfection and irrigation at concentrations as high as 30%. Lower concentrations of 3-6% also have bactericidal action 93 . Unlike cryogels that are loaded with antibiotics 94,95 , our cryogels based on hydrogen peroxide-mediated bacterial inhibition have a minimal risk of eliciting resistance 96 . Although bacteria possess multiple antioxidant defenses to fight reactive oxygen species 86 our CP-containing cryogels can release > 450 µmol of hydrogen peroxide producing hydroxyl radicals that cannot be detoxified by bacterial enzymes 86,96,97 .
Previously, scaffolds loaded with CP have been used for oxygen release 98 . Converting hydrogen peroxide from CP to water and oxygen can be useful for tissue engineering applications 99,100 . However, the concentration of CP has to be finely tuned to prevent cytotoxic side effects. In our study, CP-containing cryogels producing hydrogen peroxide were sufficient to inhibit bacterial growth, while there was minimal to no observable toxicity to mammalian cells. We showed that NIH/3T3 cells cultured with CP-containing cryogels exhibited high cell viability (Fig. 5). Similarly, these cryogels seemed unlikely to induce an inflammatory response 101,102 . We observed that these cryogels do not activate or trigger the secretion of proinflammatory cytokines when cultured with primary DCs (i.e., BMDCs). Similarly, in the in vivo setting, the cryogels induced minimal inflammatory reactions even when purposely contaminated with bacteria prior to implantation. This result reaffirms our hypothesis that CPcontaining cryogels can provide strong antimicrobial activity and prevent bacterial colonization. In addition to its antimicrobial activity, hydrogen peroxide has been reported to have antiviral and antifungal properties 81,82 . We hypothesize that CP-containing cryogels could potentially have some biocidal actions against other microorganisms (e.g., virus, yeasts, molds) and further studies are needed. Another advantage of this platform is that the injectable scaffolds are preformed, enabling them to recover their shape at the site of injection and thus eliminating the need for in situ gelation. Furthermore, CP-containing cryogels exhibited signs of biodegradation after 2 months when subcutaneously injected in mice. The degradation may be the result of enzymatic hydrolysis and oxidation 88,89 . These unique characteristics make these microcomposite cryogels very attractive for a wide range of biomedical applications, including designing scaffolds for tissue engineering and regenerative medicine.
Multiple CP-containing cryogels were formulated for this study and the resulting concentration of microparticles was relatively low. While for the purpose of our study (i.e., to prevent implant-associated infections), the transient antimicrobial activity of our cryogels is acceptable, a more prolonged biocidal action may be required for other applications such as wound healing. However, a finely tuned balance needs to be achieved between hydrogen peroxide generation and its possible cytotoxicity on mammalian cells. To that end, adjusting CP hydrolysis, and ultimately hydrogen peroxide production within its biocompatible range, is necessary. Using other metal peroxides such as magnesium peroxide could potentially provide a more sustained yet low level of hydrogen peroxide release. Although it may be not needed due to the potential biocidal action of hydrogen peroxide across several microorganisms, another potential limitation of these CP-containing cryogels is that they might not be able to undergo autoclave sterilization, as previously reported by our group 88  www.nature.com/scientificreports/ studied whether the highly dense and interconnected polymers in cryogels can provide sufficient protection for CP hydrolysis during autoclave treatment or whether alternative sterilization techniques could be used. Similarly, when combining CP-containing cryogels with other drugs or proteins to be released from the cryogels, possible interactions or damage resulting from the presence of hydrogen peroxide might need to be taken into consideration. Finally, the needle-injectable antibacterial cryogels reported here were 16 mm 3 . However, for some tissue engineering applications, it might be necessary to deliver larger cryogel constructs that are above the currently reported volume. In such cases, it might be best to use catheters instead of standard needle-injections.

Conclusion
In this work, we engineered injectable microcomposite cryogels with antimicrobial properties. The cryogels with an interconnected network of large pores (10-100 µm) could be delivered using minimally invasive approaches (i.e., injected through hypodermic 16G needles) without any structural alteration. The incorporation of CP at low concentrations within the cryogels inhibited the growth of multiple pathogenic bacterial strains that are commonly associated with implant failures. The antibacterial activity seemed to be primarily driven by the production of hydrogen peroxide, a byproduct of CP degradation found in the cryogels. While hydrogen peroxide retarded or prevented bacterial growth, it had little to no effect on NIH/3T3 cell viability, and these cryogels did not seem to trigger detectable BMDC activation. Similarly, antimicrobial cryogels did not induce an inflammatory response when tested in mice even when purposely contaminated with pathogenic bacteria prior to subcutaneous injections. Furthermore, cryogels indicated clear signs of degradation in vivo. Collectively, these results indicate that needle-injectable and biodegradable CP-containing cryogels exhibit antimicrobial activity that holds great promise for a wide variety of biomedical applications, particularly when designing tissue engineering scaffolds. Chemical modification and characterization of HA. Methacrylate groups were added to HA to yield the HA-glycidyl methacrylate (HAGM) conjugate as per the procedure mentioned by Rezaeeyazdi et al. 103 , which allows cross-linking. Quantification of the degree of methacrylation was performed by 1 H NMR analysis using a Varian Inova-500 NMR spectrometer. The degree of methacrylation was calculated according to a method previously described 55 . The HAGM macromonomer was found to have a degree of methacrylation of approximatively 20% ( Supplementary Fig. 5).

Fabrication and characterization of antimicrobial microcomposite cryogels. The HAGM 4%
(w/v) macromonomer was dissolved in deionized water, mixed with an appropriate amount of CP, and subjected to free radical polymerization induced by 0.5% (w/v) TEMED and 0.25% (w/v) APS. This prepolymer mix was distributed into precooled Teflon molds. Polymerization was then allowed to complete at a subzero temperature, i.e., − 20 °C, for 17 h 33 . After completion of the polymerization, the gels were brought to room temperature (RT) to remove ice crystals and washed with deionized water. The swelling ratio was determined using a standard gravimetric procedure. To calculate the swelling ratio of each sample, cryogel samples 8 mm in diameter and 6 mm in height (n = 4) were prepared and immersed in PBS at pH 7.4 and 37 °C. The equilibrium mass swelling ratio (Q M ) was calculated by the equation Q M = ms/md, where ms and md were the fully swollen gel and freeze-dried gel weights, respectively. The Young's modulus was determined using an Instron testing system (Instron 5944). Cylindrical cryogels (6 mm in diameter, 8 mm in height) were deformed between two parallel plates with a strain rate of 10% per minute for multiple cycles. To measure the degree of interconnectivity, disc-shaped cryogels (13 mm in diameter, 1 mm in height) were fabricated. The interconnected void volume was calculated as the ratio of weight of water wicked from the gels by a Kimwipe to the wet weight of fully soaked cryogel disks 33 .
Cryogel injectability. The ability of cryogels to pass through a conventional-gauge needle and then regain their original shapes was checked by injecting the cryogel through a hypodermic needle. First, square-shaped HAGM cryogels (dimensions: 4 mm × 4 mm × 1 mm) were suspended in 0.2 ml of PBS and syringe-injected through a 16G needle. ARS, an anthraquinone dye, has been widely used as a sensitive technique for the semi quantification of calcium deposits 63 . A 1% ARS solution was prepared in deionized water, and the pH was adjusted to 4.2. Fresh solution was used for the assay. Cryogel samples were dipped in the ARS solution for 20 min and then washed with deionized water multiple times until there was no color in the surrounding liquid. To confirm the antibacterial potential of cryogels after lyophilization as well as injection, cryogel samples (dimensions: 4 mm × 4 mm × 1 mm) with 0.2% CP were fabricated and lyophilized immediately. They were briefly sanitized, washed, partially dried over sterile gauze and inoculated with 1000 cells of bacterial pathogen in TSB. Experiments were performed with MRSA and P. aeruginosa. After 4 h of incubation, the cryogels were flushed with sterile PBS, and the PBS used for washing was subjected to CFU determination ( Supplementary  Figs. 6 and 7). Assessment of biocompatibility. Cryogels with 0%, 0.1% and 0.2% CP were fabricated as mentioned in previous sections. They were supplemented with 0.8% w/w ACRL-PEG-G4RGDSP to promote mammalian cell adhesion. The cryogels were briefly sanitized in 70% ethanol, washed, and partially dried over sterile gauze. The collapsed cryogels were placed in a multiwell plate. A total of 50 k 3T3 cells (NIH/3T3, CRL-1658, ATCC) suspended in 10 µl of DMEM supplemented with 10% FBS and 1% penicillin-streptomycin were uniformly distributed onto each cryogel piece. They were allowed to adhere to the cryogel for 2 h. Then, 200 µl of culture medium was added to the wells. After 24 h of incubation, the cryogels were recovered, stained with ViaQuantTM Far-red according to the manufacturer's instructions, and fixed with 4% paraformaldehyde in PBS for 30 min. The cells were permeabilized with 0.1% Triton X-100 for 5 min in PBS and stained with Alexa Fluor 488-phalloidin and DAPI according to the manufacturer's protocol. The cryogels were observed under a Leica TCS SP5 X WLL confocal microscope, and images were recorded. Four representative samples were used to calculate the % viability.

Assessment of in vitro immune response with dendritic cells (DCs).
For this study, BMDCs were obtained as previously described 105 . A 24-well plate was seeded with 200 × 10 3 cells per well in RPMI 1640 medium supplemented with FBS. Cryogels with 0%, 0.1%, and 0.2% CP of size 4 mm × 4 mm × 1 mm were used. They were briefly sanitized in 70% ethanol, washed, partially, and dried over sterile gauze, followed by 24 h of coincubation with DCs.
The supernatant was checked for the presence of secreted cytokines by ELISA. The supernatants were suitably diluted for IL-6, TNF-α, and IL-12, and assays were performed according to the manufacturer's instructions. DCs were recovered from the well plate as well as from within the cryogels with the help of a cell scraper and were fixed with 4% paraformaldehyde treatment. The activation of DCs was assessed with flow cytometry (BD FACS Calibur DxP), and the expression levels of CD11c, CD86, and MHCII were quantified. Expression levels were interpreted based on BMDCs cultured in media alone as a negative control and BMDCs incubated in media with 100 ng/mL LPS as a positive control.
Assessment of in vivo biodegradation. CP-free (0% CaO 2 ) and 0.2% CaO 2 cryogels (dimensions: 4 mm × 4 mm × 1 mm) were assessed for their capacity to degrade in the body. Cryogels suspended in sterile PBS (0.2 mL) were syringe-injected through 16G needles in both dorsal flanks of 12-weeks old female C57BL/6J mice (n = 5/condition, The Jackson Laboratory, Bar Harbor, ME, USA). Each mouse received both cryogel types, one on each flank. After 2 months, mice were euthanized and the cryogels explanted with the surrounding tissues to perform histological analysis. Cryogels were then fixed in 4% PFA, embedded in paraffin, and then sectioned (5 μm thick). The slices were first stained with H&E or MT and subsequently imaged (iHisto, Woburn, MA, USA). Animal work was performed under an approved protocol by the Northeastern University Standing Committee on Animals in compliance with the National Institutes of Health guidelines.