Enhanced stem cell retention and antioxidative protection with injectable, ROS-degradable PEG hydrogels

Abstract

Poly(ethylene glycol) (PEG) hydrogels crosslinked with enzyme-cleavable peptides are promising biodegradable vehicles for therapeutic cell delivery. However, peptide synthesis at the level required for bulk biomaterial manufacturing is costly, and fabrication of hydrogels from scalable, low-cost synthetic precursors while supporting cell-specific degradation remains a challenge. Reactive oxygen species (ROS) are cell-generated signaling molecules that can also be used as a trigger to mediate specific in vivo degradation of biomaterials. Here, PEG-based hydrogels crosslinked with ROS-degradable poly(thioketal) (PTK) polymers were successfully synthesized via thiol-maleimide click chemistry and employed as a cell-degradable, antioxidative stem cell delivery platform. PTK hydrogels were mechanically robust and underwent ROS-mediated, dose-dependent degradation in vitro, while promoting robust cellular infiltration, tissue regeneration, and bioresorption in vivo. Moreover, these ROS-sensitive materials successfully encapsulated mesenchymal stem cells (MSCs) and maintained over 40% more viable cells than gold-standard hydrogels crosslinked with enzymatically-degradable peptides. The higher cellular survival in PTK-based gels was associated with the antioxidative function of the ROS-sensitive crosslinker, which scavenged free radicals and protected encapsulated MSCs from cytotoxic doses of ROS. Improved MSC viability was also observed in vivo as MSCs delivered within injectable PTK hydrogels maintained significantly more viability over 11 days compared against cells delivered within gels crosslinked with either a PEG-only control polymer or the gold-standard enzymatically-degradable peptide. Together, this study establishes a new paradigm for scalable creation and application of cell-degradable hydrogels, particularly for cell delivery applications.

Introduction

Poly(ethylene glycol) (PEG) hydrogels are one of the most heavily investigated classes of matrix-mimicking, synthetic biomaterials due to their precedent for safe in vivo usage, intrinsically low protein-adsorption¹, ease of functionalization², and simple tunability of mechanical properties and degradation kinetics³. These hydrogels are typically formed by the covalent cross-linking of hydrophilic PEG macromers, yielding mechanically robust polymer networks that swell upon exposure to water. Recent hydrogel work has focused on in situ cross-linking PEG hydrogels that can be delivered in a minimally invasive manner in the clinic while maintaining an active depot of living cells^4–6, proteins⁷, or drugs⁸.

Strategies for inducing in situ hydrogel formation include UV cross-linking of acrylate or –ene (-C=C-)-terminated units in the presence of a photo-initiator^{4, 7, 9}, enzyme-mediated cross-linking with horseradish peroxidase^10–12, click chemistry-based crosslinking reactions^{13, 14}, or Michael-type addition^15–17. In particular, Michael-type addition does not require potentially-cytotoxic exogenous light sources or free radicals to achieve polymerization, but instead uses a nucleophilic reagent to catalyze the addition reaction between a branched, end-functionalized PEG macromer and a multi-functional nucleophilic cross-linker^{18, 19}. PEG macromers end-functionalized with acrylate and vinyl-sulfone groups have been previously investigated^{15, 17, 20}, and recently, García et al. explored the use of maleimide-functionalized PEG (PEG-MAL) macromers as precursors for injectable hydrogels formed by Michael-addition cross-linking^{5, 6, 21}. Maleimide groups have been extensively used in peptide biochemistry as they quickly and efficiently react with thiol groups with very high specificity at physiologic pH¹⁹. When compared to more conventional Michael-addition hydrogel systems featuring acrylate or vinyl-sulfone groups, these PEG-MAL materials possess superior cytocompatibility, increased cross-linking efficiency, and appropriate in situ gelation kinetics that make this system ideal for in vivo delivery applications^{5, 22}. Furthermore, PEG-MAL macromers demonstrate limited cytotoxicity and inflammation in vivo while possessing rapid renal clearance of the hydrogel degradation products⁶. In a therapeutic capacity, PEG-MAL hydrogels have shown promise as delivery vehicles for both bioactive proteins^{6, 21, 23, 24} and viable cells^{6, 22, 25, 26}.

Importantly for their in vivo translatability, PEG-MAL hydrogels are commonly fabricated with protease-cleavable peptide linkers specifically tailored to degrade in response to cell-produced matrix metalloproteinase (MMP) enzymes²⁷. The MMP-cleavable sequence is flanked by cysteine residues that provide free thiol groups that can participate in the Michael-addition crosslinking with PEG-MAL macromers to form 3D constructs. However, establishing peptide-crosslinked hydrogels as a generalizable tissue engineering platform suffers from a number of difficulties. These peptide sequences are cleaved by specific MMPs that are upregulated to different degrees across diverse pathological environments²⁷; MMP levels and activity are also highly variable across cellular subtypes²⁸ and patient populations²⁹. Moreover, manufacturing peptides on the scale necessary to fabricate large tissue scaffolds is both inefficient and expensive with current strategies^{30, 31}.

The current work was motivated by the need develop synthetic, degradable crosslinker polymers that enable an alternative strategy for creating more readily scalable, cell-degradable PEG hydrogels. To this end, we have developed polymers that can integrate into standard PEG-MAL hydrogels and are selectively broken down by cell-generated reactive oxygen species (ROS). Though produced in many normal biological signaling cascades³², elevated ROS, or “oxidative stress”, is a common hallmark of disease states³³ and biomaterial implantation³⁴, making oxidative stress a promising candidate as a precise, cell-generated signal for triggering material degradation. This has motivated the recent emergence of new classes of ROS-responsive polymers^35–38, including poly(propylene sulfide) (PPS)^39–44, selenium-containing polymers^{45, 46}, arylboronic esters (ABEs)^{47, 48}, oligo(proline) peptide cross-linkers⁴⁹, and poly(thioketal) (PTK) polymers^50–54. Of these ROS-degradable polymers, PTK chemistry in particular possesses many ideal attributes for synthesizing large-scale tissue engineering materials^{53, 55–60}, though it was originally utilized in inflammation-targeted nanoparticles for drug delivery applications^{50–52, 61}. PTK polymers are synthesized by simple condensation polymerization of low-cost, low-toxicity precursors to form polymer chains that are inert to hydrolysis but specifically degraded by ROS⁵¹. Due to their ROS-specific degradation mechanism, PTK-based biomaterials more effectively match in vivo material degradation with tissue regeneration and foster more successful healing outcomes when compared to conventional hydrolytically-degradable, polyester-based biodegradable implants^{53, 55–60}. Moreover, as demonstrated in other oxidation-sensitive biomaterials^62–64, thioketal-containing constructs also possess antioxidative properties and scavenge cytotoxic ROS⁶⁰.

Though other hydrogel chemistries featuring sensitivity to oxidation have been recently developed for stem cell or drug delivery applications^{44, 62–66}, thioketal-based chemistries have not been successfully deployed to date in a hydrogel despite their previous successful implementation as easily-scalable bulk biomaterials in both foam and electrospun fiber formats^{53, 55–60}. Thioketal polymers are notionally amenable to development as crosslinkers for PEG-MAL hydrogels because PTK synthesis can be tuned to produce homobifunctional thiol end groups that efficiently react with maleimides. However, previously developed PTK polymers are relatively hydrophobic due to the non-polar groups in the chemical backbone. The inherent PTK polymer hydrophobicity hinders their deployment as in situ-forming hydrogels since the water-insoluble PTK component cannot fully engage with the water-soluble PEG-MAL to form an effective covalent network under physiologic conditions. Herein, we have developed a fully water-soluble, thiolated PTK component that readily reacts with PEG-MAL macromers for the in situ formation of hydrogels. These novel, covalently-linked PEG-MAL-PTK hydrogels are specifically degraded by cell-generated ROS while retaining the benefits of previously reported PEG-MAL gels, creating a material platform with enhanced translational potential compared to conventional peptide-linked hydrogels.

Materials and Methods

Materials

All chemicals and reagents were purchased from MilliporeSigma (Burlington, MA) except the following. Cobalt chloride hexahydrate, TCEP disulfide reducing gel, calcein-AM, ethidium homodimer, and Tissue-Tek O.C.T. compound were purchased from Thermo Fisher Scientific (Waltham, MA). 10kDa four-arm PEG-maleimide (PEG-MAL) was obtained from Laysan Bio, Inc. (Arab, AL). The cysteine-terminated adhesive peptide GRGDSPC (RGD) was obtained from Genscript (Piscataway, NJ), while the MMP-degradable GCRDVPMSMRGGDRCG (VPM) peptide (received as a hydrochloric acid salt) was custom synthesized by Advanced Automated Peptide Protein Technologies (Louisville, KY). The fluorescent dye MTS-CF640R was purchased from Biotium (Fremont, CA). Cell culture reagents, including Dulbecco’s Modified Eagle Medium (DMEM), fetal bovine serum (FBS), and penicillin/streptomycin were supplied by Gibco Cell Culture (Carlsbad, CA). Primary bone marrow-derived mouse mesenchymal stem cells (mMSCs) from C57BL/6 mice were obtained from Cyagen Biosciences, Inc (Santa Clara, CA). CellTiter-Glo Luminescent Cell Viability Assay was purchased from Promega (Madison, WI). Male BALB/c mice were supplied by The Jackson Laboratory (Bar Harbor, ME).

PTK Dithiol Synthesis

The protocol for PTK dithiol polymer synthesis has been previously outlined by our group^{35, 53} and was adapted from Wilson, et al⁵¹. Briefly, 1 g of p-Toluenesulphonic acid monohydrate (PTSA) was re-crystallized in HCl at −20°C, extracted and dried, and then added to a tri-neck boiling flask and put under positive nitrogen pressure. 1x molar equivalent (9.8mL) of 2,2′-(Ethylenedioxy) diethanethiol (EDDT) was added to the flask, while 0.83x molar equivalent (6.2mL) of 2,2-dimethoxy propane (DMP) was added to an additional funnel attached to the boiling flask. Anhydrous acetonitrile was charged into the flask (100mL) and the addition funnel (50mL), and both were purged with flowing nitrogen for 30min at room temperature. Next, the flask was placed in an oil bath at 80°C, allowed to equilibrate for 15min, and the addition funnel was set so that the acetonitrile-DMP solution was added drop-wise into the continuously stirring boiling flask and allowed to react for 18h. Post synthesis, the acetonitrile was removed by rotary evaporation and the EDDT-PTK polymer was isolated by precipitation into cold ethanol (3x) and dried under vacuum. Polymer molecular weight was analyzed by gel permeation chromatography (GPC, Agilent Technologies, Santa Clara, CA) using a mobile phase of N,N-dimethylformamide (DMF) with 100mM LiBr and quantified using a calibration curve generated from poly(ethylene glycol) (PEG) standards (400 – 4000g/mol). The polymer composition was analyzed with ¹H nuclear magnetic resonance spectroscopy (NMR, Bruker 400 MHz Spectrometer). ¹H NMR chemical shifts were reported as δ values in ppm relative to the deuterated CDCl₃ (δ = 7.26): δ = 1.60 (6H), δ = 2.72 (2H), δ = 2.84 (4H), δ = 3.59–3.78 (8H).

PEG-MAL-PTK Macromer Synthesis

To improve the water-solubility of the synthesized EDDT-PTK polymer, a 4-arm PEG-MAL-PTK macromer was synthesized by combining thiolated PTK polymers with thiol-reactive PEG-MAL units. Briefly, 370mg of EDDT-PTK (8x molar equivalent) was added to a flask and put under nitrogen. The PTK was dissolved in 25mL of anhydrous acetonitrile and then supplemented with 350μL of triethylamine (TEA) to deprotonate the polymer’s thiol groups and increase the efficiency of the Michael addition reaction with the maleimide groups on PEG-MAL¹⁸. 500mg of 10kDa four-arm PEG-MAL (1x molar equivalent, or 4x maleimide equivalent) was dissolved in 60mL of anhydrous acetonitrile and then slowly added to the stirring PTK solution at room temperature and allowed to react for 16h. Following synthesis, acetonitrile was removed by rotary evaporation and the crude product was dissolved in a small amount of dichloromethane (DCM) and precipitated into −80°C diethyl ether to remove unreacted EDDT-PTK. The precipitant was spun down at 3000 × g, after which the ether was removed and the precipitant was collected and dried. The final molecular weight of the PEG-MAL-PTK macromer was determined by GPC, and the chemical composition was determined by ¹H NMR in deuterated CDCl₃. The thiol content of the PEG-MAL-PTK was determined by Ellman’s assay to calculate the actual number of thiol groups per mass of polymer against the theoretical thiol content determined from macromer GPC molecular weight analysis.

PEG-MAL Hydrogel Synthesis

Hydrogels formed from 10kDa 4-arm PEG-MAL macromers were fabricated by first separately pre-dissolving the PEG-MAL and respective thiolated crosslinking component (1:1 ratio of free thiols to maleimide groups) in phosphate buffered saline (PBS) supplemented with 4mM triethanolamine (TEOA)⁵. Three different thiolated crosslinkers were used in experiments: the synthesized PEG-MAL-PTK macromer, a commercially-available 1000Da PEG dithiol (PEG-dt), and an enzymatically-degradable VPM peptide (molecular weight 1697 g/mol)^{5, 27}. Before the addition of TEOA, the VPM peptide solution was buffered to pH 6 with sodium hydroxide to neutralize residual hydrochloric acid (HCl) from the peptide salt. Peptide salts created with trifluoroacetic acid displayed diminished activity compared with HCl salt formulations (data not shown). The PEG-MAL and the thiolated component solutions were then combined and allowed to solidify for 15min, though gelation began immediately following component mixing. The polymer weight percent (wt%) of these hydrogels was also varied (7.5 or 5 wt% in solution) to create hydrogels with different mechanical properties and swelling ratios.

Physical Characterization of PEG-MAL Hydrogels

Mechanical properties of PEG-MAL hydrogels fabricated with PTK or PEG-dt crosslinkers at both 7.5 and 5wt% were determined using parallel plate rheometry. 5wt% gels fabricated with or without 2mM RGD peptide conjugation were also explored. Hydrogels were formed at a volume of 200μL per gel using chambers of a 24-well plate as a mold and allowed to solidify for 15min; after curing, the hydrogels were incubated in deionized water for 16h to ensure that the gels were fully swelled, minimizing any material inhomogeneities⁵. Gels were removed from the plate and then placed between stainless steel 25mm plates on an AR-G2 rheometer (TA Instruments, New Castle, DE). The shear storage modulus (Gʹ) was determined over a frequency sweep from 0.1 to 10Hz at a strain of 1%. To determine the swelling ratio of PEG-MAL materials, 50μL hydrogels made at 7.5 and 5wt% from both PTK and PEG-dt crosslinkers were fabricated in the wells of 96-well plates, cured for 15min, and then allowed to swell in deionized water for 16h. The excess water was removed from each hydrogel before weighing each sample to determine the swollen wet weight. Gel samples were then frozen and lyophilized; the dried gel samples were weighed again to determine the dry weight, with the ratio of wet weight to dry weight determining the swelling ratio values. This metric is given in Equation 1.

Equation 1.

Swelling ratio determination for PEG-MAL hydrogels.

$$HydrogelSwellingRatio\left(SR\right)=\frac{Weigh{t}_{WET}}{Weigh{t}_{DRY}}$$

In Vitro ROS Degradation of PEG-MAL Hydrogels

50μL hydrogels made at 7.5 and 5wt% from both PTK and PEG-dt crosslinkers were fabricated and incubated in deionized water or varying concentrations of hydrogen peroxide (H₂O₂) with added cobalt chloride (CoCl₂) to generate hydroxyl radicals⁶⁷. Hydrogel samples were incubated in 1mL of 65mM H₂O₂ (1mM CoCl₂), 25mM H₂O₂ (0.39mM CoCl₂), 5mM H₂O₂ (77μM CoCl₂), or deionized water with daily degradation media changes for up to 21 days. Three independent gel samples per formulation were used for each time point. At each final time point, the degradation media was removed from each hydrogel sample before determining each sample’s wet and dry weights and subsequent swelling ratio value. To corroborate the swelling ratio data for hydrogel degradation, 200μL hydrogel samples were also fabricated and incubated in the same degradation media and characterized by rheometry as described above.

Hydrogel In Vitro ROS Scavenging

The ROS scavenging potential of PEG-dt and PTK crosslinked hydrogels was assessed through a 1,1-diphenyl-2-picrylhydrazyl (DPPH) assay⁶⁸. 5wt% and 7.5 wt% hydrogels of PEG-dt and PTK hydrogels were fabricated at a 50µl volume. Hydrogels (n=4) were treated with 2mL of 100µM DPPH solution in 80:20 EtOH/H₂0 (v/v) and incubated at 37°C on an orbital shaker for 24hr. 100µl of the incubation solution was sampled at pre-determined time points; absorbance was measured at 517nm and compared to control DPPH solution. The scavenging potential is expressed as % inhibition (Control _DPPH – Sample _DPPH) / Control _DPPH. The antioxidant drug TEMPOL at a 1mg/mL concentration was used as a positive control free radical scavenger.

In Vivo Degradation and Tissue Response of PEG-MAL Hydrogels Implanted Subcutaneously in Mice

50μL hydrogels made at 7.5wt% using PEG-MAL and the thiolated crosslinkers (PEG-dt and PEG-MAL-PTK) were fabricated with a covalently conjugated fluorescent dye to allow for in vivo quantification of hydrogel persistence. The fluorescent probe MTS-CF640R, which features a methanethiosulfonate (MTS) group connected to the fluorescent molecule CF640R by a disulfide bond, was dissolved in sterile water (2.5mg/mL) and incubated with TCEP disulfide reducing gel for 1h to release the disulfide-conjugated MTS group and generate a thiolated, MAL-reactive probe molecule. Following incubation, the dye/TCEP slurry was transferred to a centrifugal spin cup and centrifuged at 50 × g for 1min to remove the dye solution from the TCEP gel. The slurry was re-suspended in sterile water and the centrifugation dye extraction was repeated 5x to remove the majority of the fluorescent probe from the TCEP gel. The reduced CF640R dye was then frozen, lyophilized, and reconstituted in sterile PBS with 4mM TEOA. An Ellman’s assay was performed on the CF640R dye solution to confirm that free thiol group had been generated from the MTS disulfide reduction.

To prepare dye-conjugated hydrogels, the PEG-MAL component was dissolved at 4x the normal concentration in sterile PBS with 4mM TEOA for gel fabrication. Next, CF640R dye (0.75mg/mL to create a final dye concentration of 85μM dye per 50μL hydrogel sample) was mixed with the PEG-MAL solutions and allowed to react for 15min. To facilitate cellular interactions with the implanted gels, a cysteine-terminated RGD peptide dissolved in sterile water (7mg/mL, 69μg of peptide added per gel) was then added to the CF640R/PEG-MAL solution and reacted for an additional 15min to give a final RGD concentration of 2mM per sample. Finally, each thiolated crosslinker was dissolved in sterile PBS with 4mM TEOA and added to its respective RGD/CF640R/PEG-MAL solution in the chambers of a 96-well plate and allowed to react before implanting in 50µL-sized pockets on the ventral region of male BALB/c mice (four gels per mouse, each gel separated by at least 1cm distance under the skin to prevent signal crossover). Following implantation, the mice were imaged immediately and over time with an IVIS Lumina III (Xenogen, Alameda, CA) using 620/670 nm emission/excitation to quantify fluorescent signal decay as a marker for hydrogel degradation. IVIS images were analyzed using LivingImage software, with a region of interest (ROI) being placed around each hydrogel as guided by anatomical markers (the protrusion of the gel from under the skin). ROIs were kept constant for each gel sample over time and were used to quantify the total radiant efficiency signal of the fluorescent hydrogels over time. Due to a large decrease in fluorescent signal from day 0 to day 1 primarily from loss of unconjugated dye (statistically equivalent signal loss between PEG-dt and PTK implants), quantifications were normalized to day 2 signal values.

Subsets of mice were euthanized at days 14, 21, 28, and 35 for histological analysis. The day 35 explanted gel/tissue samples were cryo-fixed in Tissue-Tek O.C.T. Compound (Fisher Scientific) to cut two succeeding sections per sample. One section was stained with H&E and one left unstained for viewing with fluorescence microscopy to visualize the fluorescently-labeled hydrogel in the tissue. Frozen unstained sections were imaged using a Nikon C1si+ confocal microscopy system on a Nikon Eclipse Ti-0E microscope base (excitation/emission 620/670 nm) and compared to corresponding H&E stained tissue sections imaged with brightfield. Mice with terminal timepoints at days 14, 21, and 28 were euthanized by cervical dislocation before flushing their vasculature with heparinized saline and then 4% formalin. Perfusion fixation prior to hydrogel extraction was done to better maintain gel morphology through histological preparation. Post-fixation, samples were processed, embedded in paraffin, cut into 5µm thick sections, and stained with H&E or immunohistochemical (IHC) staining using rat anti-mouse CD31 (Dianova, Hamburg, Germany) and rabbit anti-mouse Ki67 (Cell Signaling Technology, Danvers, MA) antibodies. IHC sections were counterstained with hematoxylin to enable visualization of cell nuclei. Some sample cohorts did not yield sufficiently intact histological sections to allow for robust analysis; groups without sufficient biological replicates (n=3 or greater) were still included in the data set but were excluded from downstream statistical analysis.

Cell Encapsulation in PEG-MAL Hydrogels

Hydrogel precursor components (PEG-MAL, PEG-dt, PEG-MAL-PTK, and VPM peptide) for 50μL gels at both 7.5 and 5wt% were all dissolved in sterile PBS supplemented with 4mM TEOA. All components were dissolved at 2x their normal concentration for hydrogel formation to allow for the addition of RGD peptide and cell suspension solutions to give a final 50μL hydrogel volume. As described above, cysteine-terminated RGD solution (2mM RGD per 50μL gel) then was added to the PEG-MAL suspension and reacted for an additional 15min. This RGD concentration theoretically consumes 9% and 14% of MAL groups in 7.5 and 5wt% gels, respectively, and the complete reaction of these two groups had been previously validated using an Ellman’s assay to check for free thiols after component mixing. Next, minimally-passaged mouse MSCs (mMSCs) seeded in cell culture flasks (less than 80% confluence) were trypsinized and concentrated to 7 × 10⁶ cells/mL in order to minimize the liquid volume that must be added to the hydrogel precursor solutions. An aliquot of this cell suspension (enough to give 7.5 × 10⁴ cells per 50μL gel) was then mixed with each thiolated component solution and added to its respective RGD-PEG-MAL solution in a 96-well plate and allowed to cure in a cell culture incubator for 20min. After complete gelation was achieved, 200μL of cell culture medium comprised of DMEM with 1g/L glucose, 20% FBS, and 1% penicillin/streptomycin was added on top of each gel. The hydrogel-encapsulated cells were cultured for 72h with daily media changes. To evaluate the morphology and relative viability of encapsulated cells, the culture media was removed from the gels and they were washed three times with PBS before being stained with 2μM calcein-AM and 4μM ethidium homodimer for 40min in PBS. Hydrogels with live/dead stained cells were sandwiched between glass coverslips and then imaged with a Zeiss LSM 710 confocal microscope (Oberkochen, Germany). Images were taken across multiple z-slices and represented with a maximum intensity projection across the z-planes.

To quantify the overall number of viable cells encapsulated in the respective 7.5wt% and 5wt% hydrogel formulations after 72h of culture, 50μL gels with incorporated cells were incubated in 100μL PBS and treated with 100μL of CellTiter-Glo reagent for 10min while being vigorously disrupted with a pipette tip to ensure complete diffusion of the CellTiter-Glo throughout each gel sample. The luminescent signal per hydrogel sample, corresponding with the number of encapsulated viable cells, was then quantified by IVIS imaging.

Cytotoxicity assays were also set up to test for the ROS-protective effects of the PTK hydrogels which are inherently antioxidant due to irreversible reaction with and consumption of ROS, especially hydroxyl radicals⁵⁰. To determine an LC₅₀ of the model oxidative media, mMSCs (2.0 x 10⁴ cells/well) were seeded in 2D in 96 well plate format and allowed to adhere overnight. Complete media was replaced with media containing H₂0₂ supplemented with CoCl₂ to facilitate hydroxyl radical generation. mMSCs were exposed to serially diluted concentrations of H₂0₂/CoCl₂ containing media (20mM/0.25mM to 0.16mM/0.002mM) for 24h. Following incubation, treatments were replaced with 100µl of CellTiter-Glo and quantified for luminescent signal using IVIS imaging and normalized to signal from cells treated with non-oxidative media. Three independent gel samples per formulation were analyzed. After establishing this oxidative LC₅₀, 50μL gels made at 5wt% with the three crosslinker precursors were fabricated containing 7.5 x 10⁴ mMSC cells in a 96-well plate. After 20min of gelation, cells encapsulated in the hydrogels or seeded in parallel into 2D on tissue culture (TC) plastic were treated with 100μL complete culture media or 100μL media containing the LC₅₀ dose of ROS (5mM H₂0₂ supplemented with 0.0063mM CoCl₂). Following a 24h incubation, 100μL of CellTiter-Glo was added directly on top of the hydrogels, incubated for 15min on a shaker plate, then quantified for luminescent signal using IVIS imaging. The respective viable cell signal values from ROS-treated cell samples were normalized to signal from the analogue mMSC 2D or 3D culture samples treated with non-oxidative culture media.

In Vivo Cell Delivery with PEG-MAL Hydrogels

PEG-MAL hydrogel precursor component was dissolved in PBS supplemented with 8mM TEOA at a 2x concentration and incubated for 15 mins with RGD peptides as described above. Crosslinking precursors (PEG-dt, VPM peptide, and PEG-MAL-PTK) were then dissolved in PBS. Mouse MSCs stably transfected with a firefly luciferase reporter gene were trypsinized and concentrated in a cell suspension. Cells (7.5 x 10⁴ cells per gel) were mixed into the crosslinking precursor solution and maintained at 4ºC on ice. For transplantation, the ventral regions of male BALB/c mice were shaved and sterilized with betadine and alcohol before placing four small incisions to allow the creation of subcutaneous pockets. Using two pipettes, a total of 50μL solution (two-parts, one containing PEG-MAL/RGD and another containing crosslinker/cells) was simultaneously injected into the subcutaneous pocket, with four gels being placed per animal. The skin around the pockets was gently agitated for 1 minute to allow for complete mixing and crosslinking of the hydrogel precursors before closing each incision with sutures. 24h post-implantation, 150µL of D-luciferin salt solution in sterile PBS (15mg/ml) was injected subcutaneously into the loose skin over the neck to detect the luminescent emission generated from viable, luciferase-expressing MSCs. Luminescent signal was measured every 5min after injection up to 40min, and peak signal intensity was used for all quantifications. Cellular luminescence was measured at days 1, 3, 5, 7, 9, 11, 13, 15, 18, and 21 post gel implantations, and subsets of the mouse cohort were euthanized through cervical dislocation at days 7, 14 and 21 post implantations (n=3/4 gels, n=2/3 animals per timepoint). Following euthanasia, animals were perfused with a heparinized saline solution followed by 4% formalin before excising the hydrogel/tissue samples. Post fixation, hydrogels were processed, paraffin embedded, and sectioned at 5μm thickness for H&E and Ki67 IHC. Some sample cohorts did not yield sufficiently intact histological sections to allow for robust analysis; groups without sufficient biological replicates are included in the manuscript figures but were excluded from downstream statistical analysis.

To correlate luminescent signal from gel-implanted MSCs to viable cell number, an in vitro calibration curve was generated and used to analyze the in vivo signal generation from hydrogel-delivered cells. First, aliquots of transduced luciferase-expressing mMSC cells at different concentrations (8.0e⁴, 4.0e⁴, 2.0e⁴, 1.0e⁴ and 0.5e⁴ cells) were mixed with dissolved PTK crosslinker and formed into 5wt% hydrogels by mixing with RGD-conjugated 4-arm PEG-MAL. Samples were allowed to cure completely for 20 mins. 200µL of cell culture media was added on top of the cells and allowed to incubate at 37°C for 4 hours. To evaluate cell viability, 200µl of D-luciferin salt solution in cell culture media (150µg/ml) was added on top of the hydrogels and incubated on an orbital shaker for 15 mins before quantifying luminescence by IVIS imaging. A standard curve was generated by plotting total flux versus cell number and used to estimate viable cell number from the in vivo cell delivery experiments.

Histology Quantification

Histological measurements were performed on tissue sections obtained from hydrogel inserts extracted at days 7, 14, and 21-days post-implantation. 10X magnification images were used for quantification of % tissue infiltration, % blood vessel area, and % Ki67+ cells. Three fields of view per section were analyzed and averaged per sample. Image analyses were performed using Metamorph Imaging Software (Molecular Devices Inc., Sunnyvale, CA). For tissue infiltration, area occupied by pink/purple tissue within the hydrogel (in H&E sections) was quantified as a percentage of total area of the field of view. IHC visualization was accomplished using a poly-horseradish peroxidase IgG secondary antibody with 3, 3′-diaminobenzidine (DAB) substrate (BOND Polymer Refine Detection, Leica Biosystems, Buffalo Grove, IL), that appears brown under light microscopy. Blood vessel area was quantified as percent area occupied by CD31 positive brown pixels compared to total area of the tissue/hydrogel in the field of view. Proliferating cells with the hydrogels was quantified as percent of Ki67 positive brown pixels compared to total number of nuclei (quantification of blue pixels).

Statistical Analysis

All data are reported as the mean and standard error of the mean unless otherwise indicated. In particular, some histological cohorts did not yield sufficient biological replicates for statistically powered comparisons due to difficulties in intact tissue section preparation and staining. Any sample set without sufficient replicate numbers are noted in the respective figure captions and were excluded from statistical analysis. Statistical analysis was performed using Student’s t-test for single comparisons or single factor analysis of variance (ANOVA) and Tukey post-hoc tests for multiple comparisons. P-values less than 0.05 were considered statistically significant.

Results

EDDT-PTK and PEG-MAL-PTK Synthesis and Characterization

An EDDT-PTK dithiol polymer was successfully synthesized from the condensation polymerization of ethylene glycol-based EDDT and DMP using a PTSA catalyst (65% total yield), and characterized by GPC and ¹H NMR as shown in Figure 1 and Table 1. Following purification, the EDDT-PTK dithiol polymer was combined with a thiol-reactive 10kDa 4-arm PEG-MAL to conjugate a PTK polymer onto each arm of the 4-arm PEG as demonstrated in Figure 1A. The final product was purified and then characterized by GPC and NMR. GPC chromatograms indicate an increase in molecular weight for the PEG-MAL-PTK compared to the parent PEG-MAL as shown in Figure 1B. NMR spectra of both the precursor components and the final product further confirm successful conjugation, as both EDDT-PTK and PEG-MAL characteristic peaks are present in the PEG-MAL-PTK spectra (Figure 1C). GPC quantification of these polymers’ number average molecular weights (M_n, g/mol) strongly agree with theoretical values (Table 1), and Ellman’s assay quantification of the thiol content of the final PEG-MAL-PTK product further indicates that a 4-arm tetrathiol macromer was produced at a greater than 80% yield.

Figure 1. PEG-MAL-PTK tetrathiol macromer synthesis and characterization.

(A) Synthesis schemes for the EDDT-PTK dithiol polymer and the PEG-MAL-PTK macromer. (B) GPC chromatograms indicating a shift towards higher molecular weight with PTK conjugation to PEG-MAL. (C) ¹H NMR spectra of PEG-MAL-PTK shows the appearance of characteristic peaks from both PTK and PEG-MAL polymers, indicating successful conjugation.

Table 1.

Polymer characterization values.

Polymer	Theoretical Mn	GPC Mn	Thiol Content Mn
EDDT-PTK	850	875	-
10kDa PEG-MAL	11000	10595	-
PEG-MAL-PTK	14400	15890	14000

Physical Characterization and In Vitro Degradation of PEG-MAL Hydrogels

Hydrogels made from PEG-MAL and either a 1000Da PEG dithiol polymer (PEG-dt) or the PEG-MAL-PTK macromer were fabricated at both 7.5 and 5wt%. Hydrogel formation was nearly instantaneous after component mixing and the gelation time could not be calculated by conventional rheometric measurements, though was visually estimated to be approximately 30 seconds. The swelling ratio of hydrogel samples was evaluated from the ratio of each gel’s water-swollen weight divided by the weight of only the dried solid components (Equation 1). The 5wt% hydrogel samples were significantly more swollen than 7.5wt% gels for both formulations (Figure S1), though PTK-based gels were less swollen than hydrogels made with PEG-dt. To evaluate in vitro degradation of these materials, 7.5wt% PEG-MAL hydrogels made with PEG-dt or PEG-MAL-PTK crosslinkers were incubated in media containing different concentrations of ROS. Gel samples were treated with either deionized water or escalating doses of H₂O₂ with CoCl₂, a medium that simulates accelerated in vivo oxidation⁶⁷, for up to 21 days and then weighed (wet and dry) to assess their swelling ratios as an indicator of degradation³. As shown in Figure 2A-B, 7.5wt% PTK-based hydrogels have significantly greater swelling after treatment with varying doses of the H₂O₂ compared to H₂O₂-treated samples made with the PEG-dt crosslinker. PTK gels in 65mM H₂O₂ were only evaluated to day 3 while gels in 25mM H₂O₂ were only evaluated to day 7, as samples incubated for longer times at these ROS concentrations were completely degraded. Hydrogel samples made with PEG-dt were also minimally susceptible to ROS-mediated degradation as samples incubated in the 25 and 5mM H₂O₂ did have significantly increased swelling ratio values at day 7 onward. These same patterns were also seen in 5wt% hydrogels (Figure S2A-B).

Figure 2. In vitro degradation of 7.5wt% PEG-MAL hydrogels formed with PEG-dt or PTK crosslinkers.

As assessed by the gel swelling ratio, (A) PEG-dt hydrogel samples display some sensitivity to oxidation while (B) PTK hydrogels are significantly more sensitive to H₂O₂ concentration-dependent degradation. *65mM H₂O₂, ^#25mM H₂O₂, ^$5mM H₂O₂ denotes p<0.05 significance for PTK vs. PEG-dt comparison, n=3 samples per treatment. PTK hydrogel sensitivity to oxidation is further confirmed by frequency sweep rheometry measurements of hydrogel storage modulus, as (C) PEG-dt samples experience marginal stiffness loss after oxidation while (D) PTK gels are nearly fully degraded by 65 and 25mM H₂O₂ treatment.

To confirm the degradation results from swelling ratio assessment, 7.5 and 5wt% hydrogels made with both crosslinkers and incubated in the same escalating doses of H₂O₂ media were mechanically characterized by rheometry. Taken from a frequency sweep of the storage modulus Gʹ, the stiffness of non-degraded 7.5wt% PEG-dt gels was nearly 1000Pa while 7.5wt% PTK gels were approximately 700Pa (Figure 2C-D). The lower wt% samples had correspondingly lower modulus values with 5wt% PEG-dt gels around 500Pa and 5wt% PTK gels near 300Pa (Figure S2C-D). Both PEG-dt and PTK gel samples had dose-dependent reductions in Gʹ values after incubation with the ROS media (Figure 2C-D), but PTK samples had a more substantial decrease in modulus values and mechanical integrity with ~90% modulus loss at 65 and 25mM H₂O₂ vs. ~50% loss at the same doses for PEG-dt. Similar trends in ROS degradation-mediated mechanical integrity loss were also seen in 5wt% gel samples (Figure S2C-D). PTK hydrogels covalently functionalized with the adhesive peptide RGD were also fabricated and evaluated by rheometry, but did not demonstrate a significant change in modulus values after peptide conjugation (Figure S3).

In Vivo Degradation of PEG-MAL Hydrogels in Mouse Subcutaneous Implants

4-arm PEG-MAL polymers were pre-labeled with a thiol-terminated CF640R fluorescent dye and an RGD adhesive peptide. These precursors were then used to fabricate 7.5wt% hydrogels by crosslinking them with PEG-dt or PEG-MAL-PTK. The fluorescently-labeled hydrogels were then implanted into ventral subcutaneous pockets in male BALB/c mice, and the hydrogel-associated fluorescent signal from each implanted sample was measured over time using an IVIS imaging system. All fluorescent readings were normalized to day 2 values as a baseline to account for some diffusional loss of hydrogel components during the first day after implant that are too rapid to be associated with material degradation. Signal quantification over time was used to determine material retention at the implantation site as a measure of hydrogel degradation and resorption. As determined from fluorescent IVIS imaging (Figure 3A), PTK hydrogels underwent significantly more in vivo degradation than PEG-dt implants over 35 days in vivo (Figure 3B). Histological analysis of gel/tissue sections also indicated that PTK hydrogels hosted substantial tissue infiltration and material remodeling in comparison to the minimally-infiltrated PEG-dt implants as demonstrated in Figure 3C. Moreover, comparing H&E-stained histological sections of day 35 hydrogel implants with fluorescent microscopy of the corresponding frozen, unstained sections (Figure S4) further indicates that PTK hydrogel fragments were more interspersed within infiltrating tissue compared with the PEG-dt formulation. Interestingly, many of the PEG-dt histology sections manifest with voids where the hydrogel was located, most likely due to the poor material vascularization and integration with the surrounding tissue, which limited the ability to effectively collect robust histological sections of the gels in situ.

Figure 3. In vivo degradation of 7.5wt% PEG-MAL hydrogels.

(A) Relative gel fluorescence of non-degradable PEG-dt and ROS-degradable PTK hydrogels tracked over a period of 35 days showing significantly more degradation of PTK hydrogels compared to PEG-dt crosslinked gels (*p<0.05), as quantified from (B) IVIS imaging of subcutaneously-implanted, CF640R dye-conjugated hydrogels. (C) Representative histological images of implanted hydrogels (G=gel, implant boundaries outlined in yellow) retrieved days 14, 21, and 28 days post-implantation demonstrate robust tissue infiltration and remodeling of PTK gels (scale bar 250µm). Voids in PEG-dt histology reflect the loss of gels during sectioning due to the poor tissue integration with these implants.

To further examine the in vivo performance of PTK-based, RGD-functionalized PEG-MAL hydrogels, hydrogel tissue infiltration, blood vessel growth, and proliferating cell counts were quantified from histological sections of gel/tissue samples. As shown in Figure 4A,D, tissue infiltration, or the ratio of tissue to hydrogel at the implant site, significantly increased over the first three weeks. IHC staining for the endothelial cell / blood vessel marker CD31 (Figure 4B,D) also demonstrates the presence of blood vessels in the gel-infiltrated tissue growth. Finally, proliferating cells in the remodeling tissue were identified with Ki67 IHC staining and demonstrated a significant increase over the first two weeks of hydrogel implantation (Figure 4C,D)

Figure 4. Integration of PTK-crosslinked PEG-MAL hydrogels into mouse subcutaneous tissue.

Histological quantification of excised gel/tissue sections demonstrates (A) increased tissue infiltration into implants over three weeks, (B) robust blood vessel formation through the in-grown tissue (CD31 IHC staining), and (C) an increased number of proliferating cells (Ki67 IHC staining) in neo tissue over a period of 21 days (Ki67 day 21 data were excluded from statistical analysis due to insufficient replicates). (D) Representative H&E, CD31 (blood vessels) and Ki67 (proliferating cells) stained hydrogel sections at days 7, 14 and 21 post implantation. G=gel, *p < 0.05, scale bar 250µm.

MSC Encapsulation, Cytotoxicity, and Protection from Oxidation with PEG-MAL Hydrogels In Vitro

Next, PTK hydrogels were benchmarked against standard (protease degradable) and control (minimally degradable) hydrogel chemistries for application as cell delivery vehicles. Mouse mesenchymal stem cells (mMSCs) were encapsulated in 7.5 and 5wt% PEG-MAL hydrogels made with either PEG-dt, PEG-MAL-PTK, or VPM peptide crosslinkers and cultured for three days in vitro. All hydrogels were also covalently tethered with 2mM RGD adhesive peptide to encourage cell attachment to the materials. As shown in Figure 5A, maximum intensity projections of multiple z-slices from fluorescent confocal microscopy demonstrate that mMSCs were successfully encapsulated into the different hydrogel formulations and maintained their viability over three days in culture as indicated by characteristic calcein-AM green staining. Furthermore, the encapsulated mMSCs were evenly distributed in the 3D matrix and demonstrated substantial spreading in all the different RGD-containing hydrogel formulations. In particular, cells in the 5wt% VPM and PTK hydrogels demonstrate high levels of cell spreading compared to the other gel formulations, and this finding is validated in a whole-sample measurement of viable cell number as shown in Figure 5B. The 5wt% VPM and PTK hydrogels contain a significantly higher number of viable cells than 5wt% PEG-dt samples. Furthermore, the 5wt% PTK hydrogels also contain a significantly higher number of cells than the gel samples featuring the well-validated VPM crosslinking peptide (Figure 5B). The 7.5wt% formulations had statistically similar numbers of viable mMSCs encapsulated as shown in Figure 5B.

Figure 5. In vitro hydrogel encapsulation of mouse MSCs and assessment of PTK hydrogel ROS scavenging.

(A) All gels tested maintain a high level of viability and cell spreading as assessed by live/dead staining, though spreading is particularly apparent in 5wt% samples made with degradable VPM and PTK crosslinkers (scale bar 300µm). (B) Quantification of total viable cell number per gel sample, as assessed through a CellTiter-Glo luminescent readout, indicates that cell viability is not significantly different between formulations at 7.5wt%. But, 5wt% VPM gels have significantly more viable cells than PEG-dt gels while 5wt% PTK gels have significantly more viable cells than both 5wt% VPM and PEG-dt samples. (C) Quantification of total viable cell number 24h post exposure to 5mM H₂0₂ / 0.063mM CoCl₂ (signal normalized to non-ROS treated analogue samples), indicating that cell viability is significantly protected for all hydrogels compared to 2D tissue culture (TC) and that PTK formulations provide the highest ROS cytotoxicity protection. (D) To directly quantify ROS scavenging by these materials, PTK gels incubated with the free radical DPPH caused significant ROS inhibition after 24hrs. The PTK gels’ ROS inhibition was equivalent with the antioxidant drug TEMPOL (1mg/mL) under similar conditions, indicating the potent antioxidant effect of these materials. N=3/4 samples per formulation, *p<0.05.

To evaluate the protective capacity of the hydrogel formulations against exogenous oxidative stress, an LC₅₀ dose of cytotoxic ROS was determined for the mMSCs using escalating doses of H₂O₂ / CoCl₂, establishing 5mM H₂O₂ / 0.06mM CoCl₂ as the nominal LC₅₀ at 24h (Figure S5). Cells were encapsulated in 5wt% hydrogels featuring the three crosslinker chemistries, treated for 24h with cell media containing the LC₅₀ dosage of H₂O₂ / CoCl₂, and quantified for viable cell number against 2D tissue culture (TC) plated mMSCs with and without ROS treatment. As demonstrated in Figure 5C, all three hydrogels offered significant protection from ROS-mediated toxicity compared with mMSCs in 2D culture. However, viable cell signal was significantly enhanced in PTK hydrogels measured against both the PEG-dt and VPM formulations, indicating heightened protection from oxidative stress with ROS-reactive PTK samples (Figure 5C).

To complement the cell protection data, the ROS scavenging was also measured directly to confirm the radical scavenging activity of PTK hydrogels in comparison to PEG-dt-crosslinked hydrogels and a positive control antioxidant drug, TEMPOL. Samples were incubated with the colorimetric free radical DPPH for 24hrs to assess the probe’s activity inhibition over time. As demonstrated in Figure S6, all hydrogel samples inhibited ROS accumulation over time though the PTK samples were notably more potent than the analogous PEG-dt gels. At the 24hr time point (Figure 5D), the 7.5wt% PTK hydrogel samples were significantly more inhibitory than the PEG-dt analogues and statistically equivalent to a 1mg/mL dose of the positive control antioxidant drug compound.

In Vivo MSC Delivery and Engraftment with PEG-MAL Hydrogels

Employing minimally-degradable PEG-dt, enzymatically-degradable VPM, and ROS-degradable PTK crosslinker chemistries, luciferase-expressing mMSCs were encapsulated in 5wt%, RGD-functionalized PEG-MAL hydrogels and subcutaneously implanted in BALB/c mice for 3 weeks. Relative viable cell number was determined from IVIS imaging of cellular luminescence (Figure 6A) and quantified relative to day 1 signal values as demonstrated in Figure 6B. Cells encapsulated in the ROS-degradable PTK formulations experienced a significant increase in signal over the first 5 days following implantation before gradually losing viable cells over the next two weeks (Figure 6B). However, mMSCs seeded in PEG-dt gels experienced no early signal increase in cell signal before gradually losing viable cells (Figure 6B). VPM gel-seeded mMSCs experienced significant decreases in viable cell signal after day 1 (Figure 6B), indicating this formulation provided minimal maintenance of encapsulated cells. When calibrated for viable cell number in relation to luminescent signal (Figure S7A), it is estimated that roughly 11% of the initially encapsulated 7.5e⁵ MSCs retain viability in the PTK hydrogels at the end of the 21-day period as shown in Figure S7B.

Figure 6. In vivo delivery of mMSC cells encapsulated in RGD-functionalized PEG-MAL hydrogels with varying crosslinker chemistries.

(A) IVIS imaging of bioluminescent signal from luciferase-expressing mMSCs encapsulated in PEG, PTK, and VPM hydrogels tracked over a period of 3 weeks (B) demonstrates significantly greater cell retention and survival in the PTK formulations. (C) Quantified tissue infiltration from (D) H&E histology shows limited tissue growth into the minimally-degradable PEG-dt samples, nearly complete resorption of VPM-crosslinked gels by day 14, but a more controlled, steady temporal increase in tissue infiltration in PTK crosslinked hydrogels over 3 weeks. (E) Quantification of proliferating Ki67+ cells in infiltrating tissue from (F) IHC-stained sections further demonstrates the enhanced presence of proliferating cells in PTK samples at day 14. Data from tissue infiltration and Ki67 day 14 and 21 time points for VPM samples were excluded from statistical analysis due to insufficient replicates. G=gel, *p < 0.05, **p < 0.005, scale bar 250µm.

The local tissue response to the cell-encapsulated hydrogel implants was also characterized through histological assessment. Quantification of tissue infiltration and hydrogel persistence demonstrated that the minimally-degradable PEG-dt hydrogels were relatively unchanged over 3 weeks, while the PTK samples hosted significant cellular infiltration and underwent considerable material degradation over the same period as shown in Figure 6C-D. Conversely, cell-laden VPM hydrogels were resorbed so completely by day 14 that recovering hydrogel / tissue explants for proper quantification was extremely difficult (Figure 6C-D). Finally, quantification of proliferating cells within the hydrogel margins via Ki67 IHC staining indicated that the PTK hydrogels hosted more Ki67+ cells at day 7 when compared against PTK hydrogel implants without MSCs (Figure 6E-F vs. Figure 4C-D, p = 0.041), suggesting a paracrine, tissue-regenerative benefit of the delivered MSCs. The MSC-containing PTK hydrogels also had more proliferating cells than analogous PEG-dt implants at day 14, although not at statistically significant levels. (Figure 6E-F).

Discussion

Here, we sought to provide an alternative to enzymatically-degradable peptide crosslinkers for formation of cell-degradable PEG hydrogels. To do so, an easily-synthesized, cost-effective, and cellular ROS-degradable PTK macromer was generated featuring reactive thiol end-groups used to crosslink PEG-MAL hydrogels. Initial efforts to synthesize a water-soluble, PTK-based dithiol crosslinker were pursued using the same previously utilized condensation polymerization schemes for PTK synthesis⁵¹ (depicted in Figure 1A, top-left) but utilizing more hydrophilic ethylene glycol (EG)-based dithiol monomers to increase the water-solubility of the final polymer. Dithiol monomers featuring 1, 2, 5, and 12 EG units were first used in attempts to synthesize water-soluble PTK polymers, with only the 12-EG dithiol monomer (600 g/mol) yielding a polymer that readily dissolved in aqueous solutions at concentrations relevant for hydrogel formation. However, it was a challenge to produce the 12-EG-PTK polymer with the appropriate thiol end groups that are necessary for PEG-MAL reaction. Also because of the desired crosslinker molecular weight (~1000–1500 g/mol to compare against the similarly-sized VPM peptide) combined with the inefficiency of the condensation reaction with such a large (12-EG) monomer, the resultant PTK only had a degree of polymerization degree of two. This yielded a crosslinker with only a single degradable TK group on its backbone and was also difficult to purify due to the relatively small difference in monomer and polymer molecular weight. Interestingly, the 1 and 2-EG dithiol monomers both polymerized effectively (polymerization degrees > 3), potentially signifying that the PTK condensation polymerization under these reaction conditions is most effective using relatively small (<200g/mol) dithiol monomers. Unfortunately, PTK polymers made from 1 and 2-EG monomers were both relatively insoluble in aqueous buffers; since the 2-EG dithiol monomer (EDDT) is more hydrophilic than the 1-EG monomer, EDDT-PTK polymers were pursued in the further fabrication of water-soluble PEG-MAL-PTK macromers as outlined in Figure 1.

To encourage a fast conjugation of EDDT-PTK polymers to PEG-MAL while limiting uncontrolled crosslinking of multiple PEG-MAL groups, three strategic aspects were incorporated into the synthesis of PEG-MAL-PTK macromers: using an eight-fold molar excess of PTK to PEG-MAL (two-fold molar excess of EDDT-PTK polymers to each maleimide-terminated arm of the PEG macromer), diluting the concentration of PEG-MAL compared to the more highly concentrated PTK, and using a slow, drop-wise addition to apply PEG-MAL to the PTK solution. These synthetic protocol features were all designed to quickly react single thiol groups on different EDDT-PTK polymers with the four arms of more dilute PEG-MAL, thereby rapidly consuming all the available maleimide groups on single PEG-MAL units (Figure 1A). This approach spares the thiol group on the unconjugated end of each attached PTK polymer, generating a four-arm PEG-MAL-PTK macromer with terminal sulfhydryl groups (Table 1) that remain available to participate in subsequent thiol-ene reactions. The large hydrophilic PEG component of the macromer (10,000Da PEG vs. 3400Da PTK) pushes the entire polymer into aqueous solubility despite the hydrophobic PTK content and makes the water-soluble PEG-MAL-PTK tetrathiol macromer amenable to hydrogel formation. This simplistic yet high-yield functionalization reaction, coupled with the efficient EDDT-PTK synthesis from inexpensive monomers, further highlights the ease and scalability of this approach for generating cell-degradable hydrogels.

Gels fabricated using the PEG-MAL-PTK macromers were directly compared against control materials made with the same PEG-MAL but utilizing a commercially-available 1000Da PEG dithiol (PEG-dt) crosslinker to determine these samples’ physical attributes and degradation properties in vitro. Gelation following mixing of PEG-MAL and either crosslinker took place very rapidly, in agreement with studies on a similar PEG-MAL system that showed complete gelation in less than 10 minutes⁵. The PEG-dt features ethylene glycol units, which also comprise most of the EDDT-PTK backbone, but does not contain the ROS-sensitive thioketal moiety. As expected, due to the decreased crosslink density, it was found that 5wt% hydrogels were significantly more swollen and had approximately 2-fold lower storage modulus values than 7.5wt% samples for both formulations (Figure S1, Figure 2C-D, Figure S2C-D). Moreover, assuming the Poisson’s ratio of these hydrogels is around 0.5, the elastic modulus of these hydrogels is approximately three times higher than the Gʹ storage modulus⁶⁹. Given that these hydrogels range between 300 and 1000Pa for Gʹ values, it can be assumed these materials have elastic moduli values between 1–3kPa, well in line with previously cited PEG hydrogel systems used for cell delivery^{5, 70}. Interestingly, the covalent functionalization of the RGD peptide onto the PEG-MAL macromer did not impact hydrogel stiffness values despite the consumption of crosslinkable maleimide units (Figure S3). The lack of effect of pre-conjugation with RGD on mechanical properties is most likely due to the low overall amount of RGD conjugation.

When placed in degradation media containing H₂O₂ and hydroxyl radical-generating CoCl₂, it was found that the swelling ratio of PTK hydrogels significantly increased compared to the PEG-dt samples as displayed in Figure 2A-B. Both PEG-dt and PTK hydrogels incubated in water over the same time frame were also relatively unchanged, signifying both the selective ROS-mediated cleavage of thioketal bonds in the PTK crosslinks and the inherent degradability of these materials as similarly seen in other thioketal-based systems^{50, 55}. These hydrogel degradation data were also confirmed by rheometric mechanical testing of hydrogel samples incubated in escalating doses of H₂O₂ (Figure 2C-D), as PTK hydrogels displayed substantial dose-dependent decreases in Gʹ values compared with PEG-dt samples. Interestingly, the PTK hydrogels exhibit substantially more sensitivity to oxidation than hydrophobic scaffolds fabricated with PTK crosslinkers; after incubation in 65mM H₂O₂, the gels are ~90% degraded (Figure 2D) over three days while PTK-urethane scaffolds lose only ~10% of their mass over 30 days at the same dose⁵³. This enhanced sensitivity is presumably caused by the increased interaction of ROS with a hydrated gel network compared against the limited accessibility of the TK bonds within hydrophobic constructs in aqueous environments.

Though to a significantly lesser degree than seen in the PTK hydrogels, the PEG-dt samples were also sensitive to the ROS media compared to water (Figure 2A, C), potentially indicating ROS-mediated degradation of ether bonds in the PEG-dt crosslinker or the PEG-MAL macromer. Ether bonds have known susceptibility to ROS, particularly hydroxyl radicals⁶⁷, though the ROS-mediated degradation of thioketal groups is significantly faster than covalent ether bond cleavage in biomaterials⁵³. However, these data further indicate that the PEG-dt crosslinked hydrogels, and PEG-based materials in general⁷¹, should be considered “minimally-degradable” as opposed to “non-degradable” due to ether bond sensitivity to ROS over longer time-frames. The thiol-maleimide linkage is also a potential source of hydrogel destabilization due to this bond’s sensitivity to thiol exchange or retro-Michael reactions with free thiol groups, particularly in vivo⁷². Though evidence of thiol-maleimide bond cleavage is not readily apparent in vitro due to the relative stability of the PEG-dt formulations in purely aqueous conditions (Figure 2A), it is probable that longer time scales of material incubation could eventually lead to gel degradation through this mechanism.

Fluorescently-labeled biomaterials have been previously used for the longitudinal evaluation of in vivo hydrogel degradation^73–75, and this strategy was employed here to explore the degradation differences between 7.5wt% PEG-MAL gels crosslinked with either PEG-dt or the PEG-MAL-PTK polymers. Instead of labeling the RGD peptide or thiolated crosslinker, the PEG-MAL was pre-conjugated with a thiol-terminated fluorescent probe to collect the most accurate determination of hydrogel degradation since the PEG-MAL component makes up the majority of the gel and the multi-armed PEG more effectively integrates the probe into the bulk of the material. Both fluorescently-labeled hydrogel formulations implanted subcutaneously in mice had decreased signals over time as visualized in Figure 3A, though the PTK hydrogels had significantly lower signal compared to PEG-dt samples across nearly the entire time course as quantified in Figure 3B. These data portend that, as expected, the minimally-degradable PEG-dt hydrogels are relatively inert from ether oxidation or thiol-maleimide destabilization over this time scale in vivo while the ROS-sensitive PTK implants undergo significant bioresorption.

The enhanced degradability of implanted PTK hydrogels is further confirmed by histological evaluations of explanted samples from various time points after in vivo implantation. As shown in Figure 3C and Figure S4, the PEG-dt implants host minimal cellular infiltration as evidenced by the large gaps in the histology images arising from the difficulty in fixing and sectioning these poorly integrated materials. Conversely, PTK hydrogels appear fragmented and interspersed between substantial tissue growth, indicating that these materials undergo extensive cellular remodeling over time. In further quantifying the tissue infiltration into implanted PTK hydrogels, histological measurements of the tissue area within the hydrogel boundary (visualized in Figure 4D) demonstrates a significant increase in tissue growth and corresponding decrease in hydrogel presence over three weeks (Figure 4A). Moreover, this gel-infiltrating tissue is well vascularized (Figure 4B,D) and features substantial numbers of proliferating cells (Figure 4C,D), further highlighting the active remodeling of the implanted PTK materials by the local tissue environment. Though these PTK samples do not achieve full resorption over this period, it is anticipated that these implants will continue their degradation trajectory and will eventually be fully cleared from the tissue as non-toxic degradation byproducts via renal passage⁶.

Though PEG hydrogels are demonstrably useful in their naïve state, their potential also manifests as in situ gelling cell-delivery vehicles that encapsulate and protect cells after deployment of the hydrogel solution. Furthermore, PEG’s inherent non-fouling properties block protein adsorption¹ and enable these materials to be specifically functionalized with chemical motifs to direct the behavior of encapsulated cells. In particular, RGD-based peptides are particularly effective at promoting cell viability and spreading within PEG hydrogels^{4, 76}. As a proof of concept, mouse MSCs were encapsulated in RGD-containing 7.5 and 5wt% PEG-MAL hydrogels made with PEG-dt, PEG-MAL-PTK, and VPM peptide crosslinkers and cultured for 72h. PEG-MAL hydrogels made with VPM crosslinkers and RGD are non-cytotoxic and promote robust cellular attachment^{5, 6}, making these materials a benchmark for cellular encapsulation applications. MSCs embedded in 7.5wt% gel samples achieved statistically similar viability levels across all three hydrogel formulations (Figure 5B), and roughly similar levels of cell spreading as qualitatively assessed from fluorescent microscopy images in Figure 5A. However, MSCs encapsulated in 5wt% PTK hydrogels had a significantly higher signal from viable cells than both 5wt% VPM and PEG-dt samples, while the VPM cell signal was also significantly higher than PEG-dt (Figure 5B). These results are reflected in the microscopy images in Figure 5A as the encapsulated cells in 5wt% VPM and PTK gels were notably more spread through the hydrogel matrix compared to cells in 5wt% PEG-dt samples. These collective data, in accordance with previous reports⁵, indicate that the higher crosslinking density of 7.5wt% hydrogels limits spreading and expansion of encapsulated MSCs, while lower wt% hydrogels with a degradable crosslinker (VPM or PTK) allow for greater cellular migration through and remodeling of the matrix compared to the minimally degradable PEG-dt samples.

In exploring why cells encapsulated in 5wt% PTK gels outperformed those in analogous VPM formulations, it was hypothesized that the intrinsic oxidation-sensitive components in the PTK samples conferred some oxidative protection of the encapsulated cells. Encapsulation of cells within hydrogels has been previously shown to increase cellular oxidative stress⁷⁷, and culturing cells with hydrogels featuring chemistries that irreversibly react with or scavenge ROS can confer protection from endogenous and exogenous ROS^{44, 60, 62, 63}. By incubating MSC-laden hydrogels in culture media with an LC₅₀-dose of ROS, the PTK hydrogels were confirmed to provide significantly more protection against exogenous oxidative stress than the PEG-dt and VPM formulations (Figure 5C), though importantly all three hydrogels provided some defense against ROS compared with cells in 2D culture. To directly explore this antioxidant mechanism, in vitro measurements of ROS inhibition by PTK hydrogels indicated that free radicals were significantly inhibited after interacting with these materials as demonstrated in Figure 5D and S6. Collectively, the inherently enhanced viability and the active defense against exogenous, cytotoxic ROS further motivates the usage of these in situ-forming PTK hydrogels in cell encapsulation applications.

To conclusively demonstrate the potential of these ROS-sensitive PEG-MAL hydrogels as MSC delivery vehicles, RGD-functionalized 5wt% gels featuring the three crosslinker chemistries (PEG-dt, VPM, and PTK) were fabricated with embedded luciferase-expressing mMSCs and implanted subcutaneously in mice for three weeks. The 5wt% hydrogel formulations were chosen for the in vivo cell-delivery study based on the favorable in vitro encapsulation data for these materials (Figure 5B). IVIS imaging of cell-generated luminescence signal, as visualized in Figure 6A, demonstrated a significant expansion in viable cell signal from PTK implants while the PEG-dt and VPM sample signals either remained flat or immediately decreased (Figure 6B). Though the viable cell number in PTK gels does eventually decrease by day 21, the PTK MSC retention is significantly enhanced over the first 11 days compared against the other delivery vehicles. It should be noted that all three formulations appear to lose a substantial portion of their encapsulated MSC population over the first 24hrs when analyzing the luminescence/viable cell calibration (Figure S7), though the PTK implants retain the highest proportion of their initial cellular payload. This improved PTK hydrogel functionality is plausibly explained by two factors: 1) the inherent antioxidative capacity of PTK-based materials⁶⁰ as demonstrated in Figure 5C-D, and 2) the balanced degradation kinetics of the PTK formulations.

As shown in Figure 5A-B, the minimally-degradable PEG-dt samples are not amenable to cellular motility and expansion in vitro and unsurprisingly perform similarly in vivo. However, as demonstrated in Figure 6C-D, enzymatically-degradable VPM hydrogels are extremely sensitive to the local tissue environment and are nearly completely resorbed by day 14 in vivo. Previous work has established a close correlation between prolonged hydrogel retention with lengthened survival times of encapsulated stem cells^{75, 78}, and the relatively short in vivo persistence of VPM hydrogels reasonably explains the poor retention of the associated mMSCs. However, the performance of a future MMP-degradable formulation could conceivably be enhanced by using a different peptide sequence that features decreased susceptibility to enzymatic degradation given the VPM peptide’s extremely high sensitivity to MMP activity²⁷. For the PTK hydrogels, the relatively controlled in vivo resorption kinetics offer enough degradability for encapsulated MSC expansion while maintaining an inherently protective barrier against the local inflammatory response to allow for significantly longer survival times (Figure 6B). However, the degradable nature of the PTK implants does eventually cause material resorption and native immune cell infiltration (Figure 6D), presumably leading to MSC apoptosis and the subsequent decrease of the implanted population seen in Figure 6A-B. Despite this eventual reduction in therapeutic MSCs, the cell-laden PTK hydrogels host significantly more proliferating Ki67⁺ cells at day 7 than gels without encapsulated MSCs (Figure 6E-F vs. Figure 4C-D), suggesting the pro-regenerative response elicited by the elevated number of viable, gel-encapsulated MSCs.

Conclusion

We have successfully developed a PEG-based hydrogel that is crosslinked with ROS-degradable PTK polymers. These hydrogels feature water-soluble PTK macromers that possess terminal thiol groups, allowing covalent bonding with maleimide-capped, multi-armed PEG molecules for hydrogel fabrication. PTK hydrogels are mechanically robust and undergo ROS-mediated, dose-dependent degradation in vitro, while hydrogels fabricated with a control PEG-dt crosslinker are only minimally affected by ROS. This hydrogel degradability also extends in vivo as PTK-based implants undergo significant more bioresorption than minimally-degradable PEG-dt gels while also promoting robust cellular infiltration and tissue regeneration. Furthermore, these ROS-sensitive hydrogels can successfully encapsulate mesenchymal stem cells and maintain significantly more viable cells than PEG-dt formulations or even gold-standard hydrogels crosslinked with enzymatically-degradable peptides. Critically, this cellular survival enhancement is supported by the inherently antioxidative behavior of PTK polymers which were shown to protect encapsulated MSCs from toxic levels of exogenous ROS. The oxidation-sensitive hydrogel formulations further demonstrated their utility as injectable cell delivery vehicles in vivo, hosting viable encapsulated MSCs for significantly longer periods than the peptide-linked or minimally-degradable control materials. Together, these results represent a paradigm shift in the creation of selectively cell-degradable hydrogels due to the enhanced performance of ROS-sensitive PTK hydrogels in comparison to conventional benchmark materials, and further motivates the continued exploration of these materials for clinically-relevant applications.

Supplementary Material

1

Figure S1. Swelling ratio characterization of freshly fabricated hydrogels. Swelling ratios of 7.5 and 5wt% PEG-dt and PTK hydrogels indicates that PEG-dt gels swell with more water than PTK gels (*p < 0.05). The lower wt% gels also swell more than the 7.5wt% materials.

Figure S2. In vitro degradation of 5wt% PEG-MAL hydrogels formed with PEG-dt or PTK crosslinkers. As assessed by the gel swelling ratio, (A) PEG-dt hydrogel samples display some sensitivity to oxidation while (B) PTK hydrogels are significantly more sensitive to H₂O₂ concentration-dependent degradation. (*65mM H₂O₂, ^#25mM H₂O₂, ^$5mM H₂O₂ denotes p<0.05 significance for PTK vs. PEG-dt comparison, n=3 samples per treatment). PTK hydrogel sensitivity to oxidation is further confirmed by frequency sweep rheometry measurements of hydrogel storage modulus, as (C) PEG-dt samples experience marginal stiffness loss after oxidation while (D) PTK gels are nearly fully degraded by 65 and 25mM H₂O₂ treatment.

Figure S3. Frequency-sweep measurements of 5wt% PTK hydrogel mechanical properties with RGD peptide functionalization. Hydrogels formed with and without 2mM RGD peptide did not experience a significant change in storage modulus compared to unconjugated samples.

Figure S4. Histology of fluorescently-labeled 7.5wt% PEG-MAL hydrogels from day 35 tissue explants.

Figure S5. Viability of mMSC 24 hours after treatment with media containing H₂0₂/CoCl₂ used to determine LC₅₀

Figure S6. ROS scavenging kinetics of PEG-dt and PTK hydrogels. PTK hydrogel samples incubated with the free radical DPPH increasingly inhibited ROS activity over time, leading to significant DPPH inhibition compared to PEG-dt formulations at 24hr. The PTK gels’ ROS inhibition at 24hr was equivalent with the antioxidant drug TEMPOL (1mg/mL) under similar conditions, indicating the potent antioxidant effect of these materials.

Figure S7. Calibration of luciferase signal from transduced MSCs against viable cell number. (A) Calibration curve of increasing number of PTK hydrogel-encapsulated MSCs vs. their luciferase signal output. (B) Plot of predicted viable cell number from the luminescent signal quantitation of PTK hydrogel-encapsulated MSCs from Figure 6A-B.