Self-Healing Zwitterionic Microgels as a Versatile Platform for Malleable Cell Constructs and Injectable Therapies

Abstract

Injectable and malleable hydrogels that combine excellent biocompatibility, physiological stability, and ease of use are highly desirable for biomedical applications. Here, a simple and scalable strategy is reported to make injectable and malleable zwitterionic polycarboxybetaine hydrogels, which are superhydrophilic, nonimmunogenic, and completely devoid of nonspecific interactions. When zwitterionic microgels are reconstructed, the combination of covalent crosslinking inside each microgel and supramolecular interactions between them gives the resulting zwitterionic injectable pellet (ZIP) constructs supportive moduli and tunable viscoelasticity. ZIP constructs can be lyophilized to a sterile powder that fully recovers its strength and elasticity upon rehydration, simplifying storage and formulation. The lyophilized powder can be reconstituted with any aqueous suspension of cells or therapeutics, and rapidly and spontaneously self-heals into a homogeneous composite construct. This versatile and highly biocompatible platform material shows great promise for many applications, including as an injectable cell culture scaffold that promotes multipotent stem cell expansion and provides oxidative stress protection.

Hydrogels share many properties with natural tissues due to their high hydration and elastic network structure. Their broadening clinical applications—from cosmetic procedures to localized drug delivery and regenerative cell scaffolds—often demand injectability to avoid invasive surgery or malleability to fill unique 3D spaces.^[1,2] Several crosslinking strategies have thus been developed to make hydrogels injectable or malleable (able to be molded into new shapes without cracking) while maintaining their tissue-like elasticity. One such strategy is in situ formation, which uses bioorthogonal “click” reactions to rapidly form a covalently crosslinked network when two components are mixed or injected together. Thiol-ene coupling^[3,4] and azide-alkyne cycloaddition click chemistries such as SPAAC^[5,6] have both been used to make in situ forming hydrogels. This is useful for cell encapsulation, as these gels form without generating free radicals, which can damage sensitive cells.^[7] However, hydrogels cannot typically be reshaped after in situ formation, and producing custom clinical formulations requires specialized expertise and time-consuming optimization.

Other injectable and malleable hydrogels rely on physical crosslinking interactions to switch between solution and gel states under different conditions.^[8,9] For example, some poly(N-isopropylacryl-amide) (PNIPAm) and poly(ethylene glycol) (PEG)-based copolymers can reversibly form thermally responsive supramolec-ular gel structures.^[10] While these have been used for injectable drug formulations, their lack of covalent crosslinks makes them relatively weak and short-lived in vivo and many variations result in toxicity when they disassemble.^[11,12] Similarly, viscoelastic hydrogels based on naturally occurring polymers such as alginate, dextran, or hyaluronic acid can reversibly crosslink through multivalent ion chelation.^[13] These shear-thinning gels can flow when pushed through a needle and self-heal into new malleable and elastic forms.^[14,15] Hyaluronic acid hydrogels are particularly popular in injectable clinical applications and have been modified by Burdick et al. to incorporate reversible “guest-host” interactions for enhanced strength and self-healing properties.^[15,16] However, polysaccharide- and hyaluronic-acid-based hydrogels still face challenges associated with long-term physiological stability and the production of clinical-grade materials.^[17,18]

Among the numerous natural and synthetic polymers available to make hydrogels, polyzwitterions have gained particular attention because of their uniquely biocompatible attributes.^[19–21] These polymers contain balanced pairs of cationic and anionic groups, and mimic the phospholipids comprising cell membranes^[22] or the mixed-charge surfaces of many proteins.^[23] Polyzwitterionic brush coatings,^[24] hydrogels,^[25] and elastomers^[26] provide ultralow levels of nonspecific protein fouling from complex physiological media, exceeding the performance of popular hydrophilic or amphiphilic polymers like PEG. Hydrogels formed from pure zwitterionic polycarboxybetaine (PCB) have been reported to inhibit the foreign body response and resist collagenous encapsulation when implanted in mice,^[27] as well as shield proteins from immunogenic responses in the bloodstream.^[28,29] Moreover, stem cells encapsulated in PCB hydrogels maintain their therapeutic multipotency and avoid nonspecific differentiation.^[30]

Despite the remarkable biocompatibility, physiological stability, and nonimmunogenicity of PCB hydrogels,^[27,31] no straightforward routes to injectable or malleable formulations have been reported. In addition, all previously reported chemistries required to encapsulate cells in zwitterionic hydrogels generate free radicals.^[30,32] A spontaneously forming PCB hydrogel that combines biomimetic attributes, simplified storage and practical use would be particularly desirable for clinical translation. Here, we demonstrate an injectable and malleable PCB hydrogel platform referred to as zwitterionic injectable pellets (ZIPs), achieved by using zwitterionic microgels as self-healing “building blocks.” This versatile strategy was designed to enable a wide range of injectable formulations and cell-protective constructs without requiring specialized knowledge or laboratory conditions.

To create malleable PCB microgel constructs (Figure 1a), we first produced macroscopic PCB hydrogels using a photopolymerization method previously reported.^[33] All hydrogels in this work were constructed from pure carboxy-betaine acrylamide monomers with either one-carbon (PCB-1) or two-carbon (PCB-2) spacing between the charged groups, and crosslinked with carboxybetaine diacrylamide (CB-X, 0.01–1 mol%) (Figure 1b,c). We designed this microgel platform using exclusively carboxybetaine components to make them analogous to the simple and highly biocompatible PCB hydrogels reported to evade the foreign body reaction^[27] and preserve stem cell multipotency.^[30] The equilibrium water content (EWC) of all hydrogels was >95 wt%. After complete equilibration, we processed the bulk hydrogels into microgels by repeatedly extruding them through micronic steel mesh (Figure S1a,b, Supporting Information); changing the mesh pore size enabled us to easily tune the size of equilibrium-swollen microgels. Similar hydrogel sizing methods have been reported in the manufacture of soft tissue fillers.^[34] In this work, microgels were processed to have a mean diameter slightly greater than most human cells (15–30 μm) (Figure 1d), though the same strategy can be used to target any microscale size. The overall production process is straightforward and amenable to small or large scales. We observed assemblies of these microgels to form malleable but self-supporting constructs, referred to as ZIP. Their spontaneous self-healing behavior is due to the zwitterionic fusion mechanism previously reported (Figure 1e), which combines strong hydration, zwitterion pair attraction, and H-bonding to facilitate self-healing in certain zwitterionic materials.^[35,36] Zwitterionic fusion is unique because it is time- and pH-independent; in single-charged or nonionic self-healing materials, hydrophobic surface reconstruction or a pH-dependent charge barrier limits healing. Here, we have expanded zwitterionic fusion to three dimensions—in ZIP constructs, each microgel participates in dynamic healing interactions with many neighboring microgels in 3D (Figure S1c, Video S1, Supporting Information). Importantly, the individual gels remain large enough to retain strength and elasticity from their internal covalent crosslinking, while their micro-scale results in a high surface area available for dynamic interactions. The chemical makeup of each microgel and the aggregate construct are equivalent to a bulk hydrogel; inter-microgel interactions are key to the tunable behavior of the ZIP material. All the ZIP hydrogels in this work are injectable through a 28-gauge needle and rapidly revert to a self-supporting gel state in an inverted vial or on a flat surface. The photographs in Figure 1f show examples of these features, and the schematic in Figure 1g highlights some of the promising clinical applications of ZIP-based formulations.

Figure 1.

Production, properties, and applications of “ZIP” hydrogels. a) Overview schematic showing the production of viscoelastic zwitterionic injectable pellet (ZIP) gels, which can be lyophilized for simple formulation and mixed with cells or therapeutics. b) Hydrogel components are purely zwitterionic, consisting of: c) carboxybetaine acrylamide polymers (PCB-1 or PCB-2) with carboxybetaine diacrylamide crosslinker (CB-X). d) Covalent crosslinks inside each microgel enable bulk support and elasticity. e) Dynamic zwitterionic fusion interactions enable reconstruction of microgels into new viscoelastic ZIP material. f) ZIP gels can be injected through needles, self-heal, and retain their shape. g) Applications of ZIP gels include injectable soft tissue fillers, therapeutic carriers and cell scaffolds for growth and protected injection.

We studied the rheological behavior of ZIP hydrogels based on PCB-1 and PCB-2 to quantify their viscoelastic properties. First, we conducted oscillatory frequency sweeps on ZIP gels incorporating 0.05 mol% CB-X crosslinker. These data (Figure S2, Supporting Information) reveal the storage modulus (G’, used as a measure of strength or support) to be dominant over the loss modulus (G”) over the full angular frequency range (0.1–100 rad s⁻¹), showing that ZIP materials based on both carboxybetaine monomer variants behave like elastic hydrogels. Summaries of G’ and tan δ (G”/G’, used as a measure of elasticity) values recorded at 1% strain and 10 rad s⁻¹ are shown in Figure 2a,d for PCB-1 and PCB-2-based ZIP gels, respectively. Notably, ZIP gels based on PCB-2 display higher moduli at each crosslinking level compared with PCB-1, especially at very low crosslinker content (0.025 mol% CB-X). In addition, PCB-2 constructs all exhibit lower tan δ values than PCB-1 formulations—around 0.25 and 0.6, respectively— regardless of crosslinking and moduli. These direct comparisons show that PCB-2 gels are generally more supportive and elastic in nature, while PCB-1 gels are more malleable. The differences in tan δ suggest stronger or more numerous dynamic interactions in PCB-2 ZIP gels compared to PCB-1 constructs, which we attribute to the contribution of hydrogen bonding.^[35] Previously, we studied the differences between PCB-1 and PCB-2 at the molecular level, and found carboxylates in PCB-2 to form stronger hydrogen bonds than those in PCB-1.^[37] This increased H-bonding between carboxylate moieties and polymer backbone amides in PCB-2 microgel constructs enhances the overall self-healing associations (zwitterionic fusion) and increases construct elasticity (Figure 1e).

Figure 2.

Rheological properties of ZIP hydrogels. a) Storage moduli (G’) and loss tan δ (G”/G’) of PCB-1-based ZIP hydrogels with different CB-X concentrations (mean ± s.d.). b) Oscillatory strain sweep and c) step-strain test of the lowest-crosslinked PCB-1 sample (0.025% CB-X). Storage (G’, solid markers) and loss (G”, open markers) moduli plotted between 0.1% and 100% strain, or between 1% and 300%. d) Storage moduli (G’) and loss tan δ (G”/G’) of PCB-2-based ZIP hydrogels with different CB-X concentrations (mean ± s.d.). e) Oscillatory strain sweep and f) step-strain test of the highest crosslinked PCB-2 sample (0.5% CB-X).

We then conducted oscillatory strain sweep and step-strain experiments to test the shear-thinning and self-healing behavior of ZIP constructs. Figure 2b,c shows these relationships for a very soft PCB-1 construct, which has a maximum G’ around 20 Pa. Similarly, Figure 2e,f displays comparable plots for a strong PCB-2 sample (G’ = 5 kPa). We found G’ to dominate in these and all intermediate formulations at low strains (0.1–1%). The complex viscosity and G’ began to decrease as we increased the strain towards 100%, with most samples exhibiting a crossover point (tan δ = 1) between 10% and 30% strain. Above this strain, G” becomes dominant and the gels begin to adopt more viscous behavior as intermicrogel associations break and reform. When we toggled strain between 1% and 300%, both PCB-1 and PCB-2 formulations showed inversion of G and G” at high strain and rapidly recovered their elastic properties when strain was reduced. This is indicative of efficient self-healing across all samples. While the viscoelasticity data for a soft PCB-1 construct (Figure 2b,c) differ by several orders of magnitude from a strong PCB-2 construct (Figure 2e,f), the axes are scaled for direct visual comparison of their relative behavior. Overall, PCB-1-based ZIP gels more easily achieve low elastic moduli between 10 and 100 Pa; this softer characteristic is ideal for many cell culture and injection applications.^[38] In addition, the rheological properties of PCB- 2-based ZIP gels are similar to reported measurements for hyaluronic acid dermal fillers,^[39,40] suggesting these hydrogels may be particularly suitable as long-lasting injectable materials for cosmetic and therapeutic applications.

Given the sterility, formulation, and storage benefits of lyophilized formulations, freeze-dried ZIP powder would be a desirable starting point for clinical handling. Therefore, we characterized ZIP formulations through terminal sterilization and lyophilization procedures to study the impact of these processing steps on their material properties. Immediately after processing bulk hydrogels to microgels, all ZIP formulations in this work were sterilized by immersion in 70 wt% ethanol.^[41] Autoclaving, ethylene oxide treatment, and gamma irradiation have also been reported as suitable methods to sterilize zwitte-rionic hydrogels.^[42] Promisingly, when we re-equilibrated gels in sterile water after the ethanol treatment, we observed no visual differences between pre- and poststerilized ZIP gels. The effects of lyophilization on several types of hydrogels have been reported, with the freeze-drying process typically shown to irreversibly change their structure and behavior.^[43] Many lyophilized gels require immersion in water for hours to days to rehydrate or fail to ever reach their original water content, displaying uneven shapes, surface roughness, and modified material properties.^[44] However, due to the particularly strong hydration exemplified by PCB, we hypothesized lyophilization would not have a detrimental impact on ZIP hydrogels. As predicted, we found freeze dried ZIP powders to completely rehydrate to their original EWC within seconds, and retain the transparency, homogeneity, and material attributes (e.g., injectability and self-healing) of fresh samples (Figure 3a). In rheological testing, we found both PCB-1 and PCB-2-based ZIP constructs to retain equivalent moduli (G’) and elasticity (tan δ) before and after the sterilization and lyophilization process (Figure 3b). Their high self-healing efficiency also remains intact, with no difference in step-strain recovery time after multiple strain cycles (Figure 3c). Because these lyophilized formulations enable dramatically simplified formulation of many drug- or cell-encapsulating composite constructs, we used ZIP powders as a foundation to explore some practical applications of this platform.

Figure 3.

Reconstitution of lyophilized ZIP constructs and protected cell injection. a) Lyophilization reduces storage demands of sterile microgel material and simplifies formulation. b) Strength (G’, left axis) and elasticity (tan δ, right axis) of PCB-1 (blue/gray) and PCB-2 (red/gray) ZIP gels before and after the sterilization and lyophilization process (mean ± s.d.); both gels contain 0.05% CB-X. Post-lyophilization gels were rehydrated to their EWC. c) Recovery of G’ and G” by PCB-2 (CB-X = 0.05%) ZIP gels upon reverting from high (300%) to low (1%) strain after step-strain cycles 1–3. Left: fresh ZIP gels prior to sterilization; right: ZIP gels poststerilization, lyophilization, and rehydration to their EWC. d) Injectable protective cell construct made from ZIP gel. e) LIVE/DEAD stained HEK-293T cells before and after injection through a 28-G needle in ZIP gel and PBS. f) Quantified viability before and after injection in ZIP gel and PBS. *P < 0.05, ns = no statistical difference.

Progress in cell-based therapies has ignited demand for more effective methods of storing, culturing, preserving, and administering therapeutic cells in a well-controlled, sterile manner. Based on the demonstrated biocompatibility of zwit-terionic materials and their lack of nonspecific biological interactions, we expected ZIP constructs to be well suited for therapeutic cell culture and scaffold design. Notably, the simple rehydration from powder requires no specialized equipment or expertise, in marked contrast to chemistries common in tissue engineering research.^[45] Reports by Heilshorn et al. have highlighted the protection some shear-thinning hydrogels can provide cells as they pass through a needle^[46,47] (Figure 3d). We speculated ZIP formulations may offer similar protection while being much simpler to prepare. To test this, we suspended healthy HEK-293T cells in phosphate buffered saline (PBS) (10⁶ cells per mL) and gently added this suspension to an appropriate amount of ZIP powder (PCB-1 based, 0.05% CB-X) to form a reconstituted gel construct. This particular ZIP formulation was selected to match the rheological attributes (G’ ≈ 100 Pa) of previously reported cell-protective hydrogels.^[47] Cell-loaded ZIP formulations and control suspensions in PBS were transferred to 1 mL syringes and injected (100 μL s⁻¹) through a 28-gauge needle. While a 25–30% decrease in cell viability was observed in control (PBS-only) samples, no significant change in viability was seen in ZIP-protected formulations. Fluorescent micrographs of stained cells and quantified viability are shown in Figure 3e,f.

To evaluate the suitability of ZIP constructs for extended cell culture, we formed malleable, 3D cell constructs from a lyophi-lized ZIP powder formulation (PCB-1 based, with 0.5% cRGD-functionalized CBAA-2 monomer and 0.05% CB-X). Simply mixing human mesenchymal stem cell (hMSC), HEK-293T, or NIH-3T3 cell suspensions with ZIP powder produced soft 3D constructs. We used porous Transwell plate inserts to support the cultures while keeping them equilibrated with the media; this model allows media and additives to be refreshed without disturbing the growing cell population evenly distributed throughout the gel. The final cell populations are recovered by thoroughly flushing the construct with PBS through a cell filter, overhydrating the 3D culture to a slurry and allowing size-based separation of cells from microgels. A schematic of this strategy is shown in Figure 4a. All cells in ZIP culture retained high viability (>90%) after 14 days, while the same cell lines grown in standard tissue-culture polystyrene (TCPS) flasks displayed lower viability (Figure 4b).

Figure 4.

hMSC culture and maintained multipotency. a) Schematic showing in vitro Transwell-based ZIP culture system used for hMSCs and other cell lines, b) Viability of hMSC, HEK-293T, and NIH-3T3 cells via MTT assay after 14 days of culture in cRGD-ZIP constructs and TCPS control flasks, c) Cell population (fold expansion) after 14 days of culture in cRGD-ZIP constructs and control flasks, d) Expression of multipotency biomarkers ALCAM and STRO-1 via immunofluorescent staining after 28 days of culture in control flasks and ZIP gels, compared to a fresh population. e) Relative levels of intracellular reactive oxygen species (ROS, via DCFH2-DA) and mitochondrial superoxide (via MitoSOX Red) in cells grown in control flasks and ZIP gels for 28 days, compared to a fresh population. *P < 0.05, ***P < 0.001, ns = no statistical difference.

We were particularly interested in the capability of 3D ZIP culture to expand hMSC populations while maintaining their multipotency—we have previously reported that zwitterionic hydrogels are able to restrain hMSC differentiation,^[30,32] but the simplicity of the ZIP platform and ease of cell recovery makes it much more viable for clinical translation or biomanufacturing. In addition, the physical zwitterionic fusion interactions between microgels allow scaffold reshaping as the cell population expands, suggesting ZIP may be a promising platform for ex vivo stem cell expansion. As shown in Figure 4c, the 3D malleability of ZIP scaffolds resulted in greater population expansion over 14 days than standard flask culture for all cell lines tested. We then compared the differentiation behavior of hMSCs expanded in flasks with those grown in cRGD-func-tionalized ZIP scaffolds. For these experiments, we cultured the cells in bipotential media supporting both adipogenic and osteogenic differentiation. After maintaining hMSCs for 28 days in each platform, we harvested and analyzed all populations. Using immunofluorescence staining, we visualized the expression of multipotency biomarkers ALCAM and STRO-1. As shown in Figure 4d, hMSCs grown in ZIP scaffolds display high expression of both biomarkers, indistinguishable from the fresh seed population. By contrast, flask-cultured hMSCs exhibit reduced ALCAM and STRO-1 expression, suggesting around half the population loses multipotency in this control system. We also used qRT-PCR to quantify the expression of mRNAs characteristic to adipogenic (ADIPOQ, FABP4, LPL, PPARG) or osteogenic (COL1A1, OCN, OPN, RUNX2) lineages, and found none of these genes to be expressed by ZIP- cultured hMSCs at significantly higher levels after 28 days (Figure S3, Supporting Information).

In nature, their physiological niche protects stem cells from excess reactive oxygen species (ROS) exposure. ROS levels must remain low for stem cells to maintain multipotency and self-renewal pathways; excessive ROS can nonspecifically activate differentiation-promoting pathways and inhibit self-renewal through oxidative stress pathways.^[48–50] Moreover, the oxidative stress caused by very high ROS levels can stop the cell cycle and even cause apoptosis. While many hydrophobic biomaterials are known to increase ROS production, culture in a hydrophilic environment has been reported to mitigate and even scavenge ROS.^[51] We therefore wanted to examine whether a highly hydrophilic zwitterionic environment could protect cells from oxidative stress through its ROS-scavenging capacity, playing a similar role to the in vivo stem cell niche. As presented in Figure 4e, after 28 days of culture in ZIP scaffolds or flasks, we measured both intracellular ROS and mitochondrial superoxide levels in expanded hMSC populations. While both levels were heightened by 5–10-fold in flask cultures, ZIP-cultured cells displayed ROS and O₂⁻ levels similar to fresh seed populations. This phenomenon may be a key mechanism behind stem cells favoring self-renewal and mitigating nonspecific differentiation in a zwitterionic environment.

Along with applications relevant to cell therapy, we explored other clinical areas in which ZIP formulations may provide unique benefits. Injectable hydrogels are commonly sought in the formulation of poly(lactic-co-glycolic acid) (PLGA) drug depots, as they facilitate accurate volumetric dosing and keep the depot localized at the injection site.^[52] To formulate a model ZIP-based depot, we mixed doxorubicin (DOX)-loaded PLGA microspheres with ZIP powder; this formulation reconstituted to a homogeneous mixture in seconds, and the microspheres were held in place with no obvious settling or separation. Both ZIP-formulated and control samples released ≈90% of their total drug cargo within two weeks in vitro, with the ZIP formulation exhibiting lower “burst” release over the first 24 h and similar release kinetics thereafter (Figure S4, Supporting Information). This indicates the ZIP gel did not significantly inhibit or accelerate PLGA hydrolysis and erosion, and that small molecule drugs can freely diffuse out through the zwitterionic matrix.

As a final example, we used the ZIP platform to create a practical injectable and spreadable protein formulation containing a model enzyme, β-Lactamase. While PEGylation remains the most common strategy to improve the pharmacological properties of biologics, PEG reduces protein activity and can induce antipolymer antibodies.^[31,53] ZIP formulations consist of pure PCB, which is itself based on glycine betaine, a common osmolyte known to stabilize proteins and prevent denaturation and aggregation.^[54–56] PCB-protein conjugates or nanogels show enhanced stability, maintained bioactivity, and mitigated immunogenic responses in vivo,^[28,29] and we were curious if a simple ZIP excipient could provide some of these benefits for topical or localized applications. When we reconstituted ZIP powder with β-Lactamase, we found the enzyme-loaded gel to fully maintain its protein activity (V_max) in buffer (Figure S5, Supporting Information). By contrast, a hydrogel formed from the thermally responsive PEG-based poloxamer P407, which has been widely used for preclinical drug formulations,^[57,58] reduced enzyme activity by over half.

In summary, we have developed a simple and versatile strategy to create shear-thinning and self-healing zwitterionic hydrogel formulations based on reconstructed microgel assemblies and zwitterionic fusion. Importantly, these gels consist purely of carboxybetaine polymers and crosslinker, are straightforward to make at any scale, and can be simply sterilized and lyophilized for long-term storage and facile reconstitution. Injectable and malleable ZIP formulations—containing therapeutic cells, drug-loaded microspheres, or biologics—can easily be created for many clinical applications. Whether as an injectable tissue filler, drug delivery vehicle, or protective stem cell culture scaffold, ZIP hydrogels present a promising new platform for a wide variety of clinical demands requiring biocompatible injectable materials.

Experimental Section

Material Formulation:

Bulk zwitterionic hydrogels were produced using a photopolymerization method previously reported.^[33] Briefly, carboxybetaine acrylamide monomer with either a one-carbon (CBAA-1) or two-carbon (CBAA-2) distance between the charged groups was dissolved in deionized water to 2.5 M. Carboxybetaine diacrylamide (CB-X) crosslinker (0.01–1 mol% relative to monomer) and photoinitiator 2-Hydroxy-4’-(2-hydroxyethoxy)-2-methylpropiophenone (Irgacure 2959; 0.05 wt%/v%) were added to the monomer solution, which was thoroughly mixed, degassed, and cast between two glass slides with a 1 mm polytetrafluoroethylene spacer. Polymerization was initiated by UV exposure (Spectroline XL-1500; 302 nm, 15 min), and the resulting PCB-1 or PCB-2 hydrogels were dialyzed in DI water for at least 5 days. Microgels were produced by extruding bulk gels through progressively finer micronic steel meshes (TWP; 500 μm down to 25 μm pores) using a stainless-steel piston and cylinder apparatus. Microgels were passed through the final mesh size at least three times for size homogeneity of the final ZIP material. All batches were then sterilized by precipitation into ethanol, re-equilibrated in sterile water, and lyophilized (Labconco FreeZone) before further use.

Rheological Characterization:

Rheological measurements were performed using a Physica MCR 301 rheometer (Anton Paar) with a parallel plate geometry (40 mm plate diameter, 900 μm gap) and Peltier plate for temperature control. Oscillatory frequency sweeps (0.1–100 rad s⁻¹, 1% strain), oscillatory strain sweeps (0.1–100% strain, 10 rad s⁻¹), and step-strain experiments (toggling between 1% and 300% strain for three or more cycles) were all conducted at 25 °C. Each formulation was analyzed before and after the sterilization— lyophilization-rehydration process.

Cell Culture and Injection:

Cell lines (HEK-293T, NIH-3T3, and hMSC; all from Lonza) were cultured in basal medium (low-glucose DMEM, 1% penicillin/streptomycin) under standard tissue culture conditions (37 °C, 5% CO₂) unless otherwise specified. PCB hydrogels were formulated with cRGD as previously described^[59] to promote cell attachment. ZIP gels used for cell constructs were PCB-1 based, with 0.5 mol% cRGD-functionalized CBAA-2 and 0.05 mol% CB-X. To evaluate cell protection during needle flow, HEK-293T cells were suspended in PBS (10⁶ cells mL⁻¹) and constructs were formed by gently mixing the cell suspension (1 mL) with ZIP powder (50 mg). The ZIP-cell constructs and control suspensions were carefully transferred to 1 mL syringes and injected at 100 μL s⁻¹ through 28-gauge needles. Cells were LIVE/DEAD stained with calcein-AM and ethidium bromide homodimer (Thermo Fisher) and imaged with an inverted fluorescence microscope (Nikon T2000U) to assay viability. For longer-term culture, ZIP-cell constructs were formed in Transwell tissue culture inserts (Corning, 8 μm pores) and placed in 12-well plates. Appropriate media and supplements were added to the outer compartment of each well until flush with the top of the constructs. Cell viability and population growth experiments were conducted in basal media for 14 days; media was changed every third day. Final cell populations were recovered by flushing the constructs with PBS through a cell filter (40 μm pore size).

Stem Cell Analysis:

The differentiation behavior of hMSCs was studied during culture in ZIP constructs or standard TCPS flasks over 28 days. As previously described,^[32] bipotential media with adipogenic and osteogenic supplements was used. To analyze phenotype, cells were labeled with Alexa Fluor 647 anti-STRO-1 (Biolegend) and FITC anti-ALCAM (AbD Serotec) and assayed via fluorescence flow cytometry (Becton Dickinson LSR II, 488/530 filter). Quantitative real-time PCR was used to quantify mRNA expression characteristic to adipogenic (ADIPOQ, FABP4, LPL, PPARG) or osteogenic (COL1A1, OCN, OPN, RUNX2) lineages. To test oxidative stress levels, a DCFH2-DA assay (Abcam) was used to detect intracellular H₂O₂ and a MitoSOX Red assay (Life Technologies) was used to quantify mitochondrial superoxides.

Drug formulations:

PLGA (Evonik) microspheres containing chemotherapeutic drug DOX (Sigma) were produced using a double (W/O/W) emulsion method adapted from several similar protocols^[60,61] and characterized via SEM. Drug loading was quantified by dissolving microspheres in DMSO and measuring DOX absorbance (480 nm) with a microplate reader (BioTek Cytation 5). To produce ZIP formulations, DOX-PLGA microspheres (40 mg, containing 2 mg DOX) were mixed with ZIP powder (50 mg) and reconstituted with PBS (1 mL). Drug release rates were evaluated in PBS at 37 °C by dispensing the ZIP-DOX-PLGA depot (or DOX-PLGA microspheres without gel) into Transwell inserts (Corning, 8 μm pore size). Buffer was sampled and replaced at selected time intervals and the cumulative DOX release was assayed spectroscopically. Enzyme formulations were produced by reconstituting ZIP powder with a solution of TEM-1 β-Lactamase; nitrocefin (Life Technologies) was used as a model substrate. Kinetic parameters of ZIP-enzyme formulations were measured under saturating substrate conditions, with enzyme activity (V) defined as the initial rate of change in substrate absorbance at 490 nm.