Neurogenic Cell Behavior in 3D Culture Enhanced Within a Highly Compliant Synthetic Hydrogel Platform Formed via Competitive Crosslinking

Abstract

Purpose

Scaffold materials that better support neurogenesis are still needed to improve cell therapy outcomes for neural tissue damage. We have used a modularly tunable, highly compliant, degradable hydrogel to explore the impacts of hydrogel compliance stiffness on neural differentiation. Here we implemented competitive matrix crosslinking mechanics to finely tune synthetic hydrogel moduli within soft tissue stiffnesses, a range much softer than typically achievable in synthetic crosslinked hydrogels, providing a modularly controlled and ultrasoft 3D culture model which supports and enhances neurogenic cell behavior.

Methods

Soluble competitive allyl monomers were mixed with proteolytically-degradable poly(ethylene glycol) diacrylate derivatives and crosslinked to form a matrix, and resultant hydrogel stiffness and diffusive properties were evaluated. Neural PC12 cells or primary rat fetal neural stem cells (NSCs) were encapsulated within the hydrogels, and cell morphology and phenotype were investigated to understand cell-matrix interactions and the effects of environmental stiffness on neural cell behavior within this model.

Results

Addition of allyl monomers caused a concentration-dependent decrease in hydrogel compressive modulus from 4.40 kPa to 0.26 kPa (natural neural tissue stiffness) without influencing soluble protein diffusion kinetics through the gel matrix. PC12 cells encapsulated in the softest hydrogels showed significantly enhanced neurite extension in comparison to PC12s in all other hydrogel stiffnesses tested. Encapsulated neural stem cells demonstrated significantly greater spreading and elongation in 0.26 kPa alloc hydrogels than in 4.4 kPa hydrogels. When soluble growth factor deprivation (for promotion of neural differentiation) was evaluated within the neural stiffness gels (0.26 kPa), NSCs showed increased neuronal marker expression, indicating early enhancement of neurogenic differentiation.

Conclusions

Implementing allyl-acrylate crosslinking competition reduced synthetic hydrogel stiffness to provide a supportive environment for 3D neural tissue culture, resulting in enhanced neurogenic behavior of encapsulated cells. These results indicate the potential suitability of this ultrasoft hydrogel system as a model platform for further investigating environmental factors on neural cell behavior.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12195-024-00794-2.

Introduction

The current need for advanced engineered soft tissue constructs is great due to the prevalence of pathologies causing permanent soft neural tissue damage (injury, cancer, neurodegenerative diseases, stroke) [1–5]. Damage to brain tissue due to stroke or traumatic brain injury (TBI) often causes long-term patient disability. Only 10% of stroke survivors experience near-full to full recovery, while TBI is a leading cause of death and disability in young adults [3, 6, 7]. Another impactful type of neural tissue damage is spinal cord injury (SCI). Along with stroke, SCI is a leading cause of paralysis [8]. SCI patients commonly experience chronic pain after injury, which has been correlated with decreased quality of life [9–11]. Low recovery rates for these conditions are likely attributable to the limited regenerative capacity of the adult mammalian central nervous system [12, 13]. Thus, cell therapy is a favorable avenue for functional neural tissue regeneration. For cell-based neural tissue therapy, neural stem cells (NSCs) are considered widely to hold potential for regenerating functional tissue. In animal models of brain tissue pathologies, NSC transplantation has led to positive results including graft survival, donor cell differentiation, endogenous neurogenesis, and sensorimotor function recovery [14–23].

Transplantation of cells within a biomaterial delivery vehicle improves cell retention and promotes integration of the transplanted cells within the host tissue [18, 23–25]. Following results from other regenerative scaffolds, neural tissue cell therapy functional outcomes may be further improved by adding pro-neurogenic cues to the protective delivery matrix [23, 26–28]. The process of NSC neurogenesis is not well understood, however, as the native guide map, the NSC niche, is complex and dynamic [29]; here, a multitude of local signals are heterogeneously presented to NSCs to control their differentiation [30–32]. Improved understanding of neurogenesis will guide the design of tissue engineered delivery constructs to direct and promote regeneration; this indicates a need for appropriate tissue models where neurogenic cues may be investigated.

Many current in vitro culture systems probe the effects of environmental signals on neural cell behavior and NSC fate. One critical environmental property that influences stem cell behavior is local matrix stiffness. Soft tissue models which are more mechanically representative of native soft tissue have been shown widely to better support neurogenic behavior than stiff matrices through multiple mechanisms. NSC fate is biased through mechanotransduction signaling pathways, which are mediated by surface contacts, cell contractility, cytoskeletal tension, and stiffness [33–36]. Axon extension is impacted by mechanosensitive ion channels including piezo1, which are controlled by environmental stiffness [37, 38]. Accordingly, neural cells demonstrate increased neurite extension in softer environments [39–41], and neural stem and progenitor cells favor neurogenesis over astrogenesis on most compliant substrates at stiffnesses of 100–1000 Pa, generally optimized around 500 Pa [33, 42–44] . Even more pluripotent mesenchymal stem cells (MSCs), embryonic stem cells, and induced pluripotent stem cells favor neuronal differentiation on substrates ≤ 1 kPa modulus [45–47]. Differences in stiffness within the 0-1 kPa range can also influence cell behavior [36, 42, 48]. Therefore, successful design of platforms to model soft tissue interactions necessitates a mechanism for controlling modulus within the natural stiffness regime of these tissues (0.1–2 kPa) [49–52].

Polymer hydrogels are widely implemented as cell culture systems because they mimic the hydrated matrix of macromolecules comprising the fibrous extracellular matrix (ECM) which surrounds cells in vivo [53–55]. A multitude of hydrogel materials exist for investigative in vitro culture of neural cells, including naturally-derived and synthetic, and ranging from complex to reductionist [56]. Hydrogel systems currently implemented as neural cell culture platforms are largely composed of soft, naturally-derived polymers which comprise components of the NSC niche ECM: laminin, hyaluronic acid, collagen, or combinations thereof [48, 57–66]. Though they can recapitulate soft tissue stiffness, natural polymers present innate bioactivity with little tunability or control, and mechanical and biochemical properties are inherently linked and cannot be controlled and investigated independently [54, 55, 67].

Synthetic polymer-based hydrogels instead provide greater control over biochemical properties and batch-to-batch consistency, presenting a well-characterized platform for investigating influences of environmental cues on cell and tissue behavior [55, 68, 69]. Poly(ethylene glycol) diacrylate (PEGDA) is a particularly advantageous polymer material for hydrogel cell culture systems because it is bioinert but can be custom-modified with peptides for specific bioactivity, so the matrix may be engineered as desired to have specific controlled interactions with cells. Modular control over biochemical and mechanical hydrogel properties allows for independent investigation of the influence of each parameter on cell behavior [67]. Due to the customizability of this system, PEGDA-based hydrogels have been designed to model many different tissue environments [70–72], including neural culture applications [73–80]. Although standard methods for manipulating PEGDA mechanical properties, altering polymer density or molecular weight between crosslinks [75, 81–83], have proven effective for many tissue applications, they are limited in applicability for soft tissue modeling.

Methods of PEGDA hydrogel stiffness modulation independent of polymer density are scarce. We have recently demonstrated a novel method for decreasing PEGDA hydrogel compressive stiffness to within the neural regime by altering hydrogel crosslinking mechanics. By first incorporating a vinyl group with a higher propensity to terminate addition reactions as a competitive sidechain on the PEGDA polymer backbone that interfered with PEGDA acrylate crosslinking, Schweller and West significantly reduced synthetic hydrogel compressive modulus to within the neural tissue stiffness range without preventing gelation or affecting hydrogel diffusive properties or biochemical signal density [67]. In a newer development, we exerted greater control over soft PEGDA-based hydrogel stiffness by soluble implementation of competitive crosslinking monomers, including Lysine-Alloc (referred to as alloc) to effectively decrease PEGDA-based hydrogel stiffness in a dose-dependent manner to ~ 300 Pa compressive modulus without influencing gelation or biochemical density within the hydrogel. We then demonstrated that this method improved neurogenic behavior of PC12 cells seeded on the surface of these gels beyond the capabilities of the softest PEGDA hydrogels without crosslinking competition [84].

Cells have been shown to respond differently to mechanical forces in 2D and 3D, however, suggesting dimensional effects on mechanotransduction which is a key controller of neural cell behavior [85–87]. Thus, we here expand the allyl-acrylate competition technology to develop a degradable reductionist PEG-based hydrogel for investigative 3D neural cell culture which is more dimensionally representative of cells’ native environment and can provide results that are more clinically translatable [54, 88]. To achieve matrix degradation necessary for cell spreading, migration, and interaction within a nanoporous hydrogel network [89], we conjugated matrix metalloproteinase (MMP)-cleavable peptide GGGPQGIWGQGK (abbreviated PQ), to the PEG polymer backbone to form PEG-PQ-PEG. This collagen-derived sequence, commonly used in 3D tissue engineered constructs, is degraded by MMP-2 and MMP-9, both of which are secreted by many cell types including neural cells [90].

We evaluated this hydrogel system by characterizing the relationship between the mechanical properties and diffusive properties within the gels made with alloc and determining the cytocompatibility of this system for cell encapsulation. We then analyzed the effects of matrix stiffness on PC12 neural cell behavior in 3D culture within our hydrogels by modulating gel stiffness via allyl-acrylate competition and quantifying resultant neurite outgrowth, demonstrating enhanced neurite outgrowth in the softest hydrogels made with alloc. Finally, we cultured primary NSCs in this PEG-PQ-PEG-alloc hydrogel construct, testing the effects of soluble signaling on NSC behavior within the platform and thus demonstrating its functionality as a foundation for hydrogel design to support neurogenesis.

Materials and Methods

Polymer Functionalization

The cell-adhesive, integrin-binding peptide Arg-Gly-Asp-Ser (RGDS; GenScript) was conjugated to 3.2 kDa acrylate-PEG-succinimidyl valerate (acryl-PEG-SVA; Laysan Bio) by dissolving 1 mol acryl-PEG-SVA with 1.2 mol RGDS in anhydrous dimethyl sulfoxide (DMSO, Sigma), then adding 2 moles N,N-diisopropylethylamine (DIPEA, Sigma) per 1 mol acryl-PEG-SVA, and allowing to react overnight at room temperature with constant agitation. The acryl-PEG-RGDS product was dialyzed against ultrapure water using a 3.5 kDa MWCO regenerated cellulose membrane (Repligen) and then the product was lyophilized.

To confer cell-degradability to the polymer backbone, MMP-degradable peptide PQ (GGGPQGIWGQGK, GenScript) was conjugated to 3.2 kDa acryl-PEG-SVA by dissolving 1 mol PQ with 2 mol acryl-PEG-SVA (for PQ to be linked on both ends to acryl-PEG) in anhydrous DMSO, then adding 2 mol DIPEA per 1 mol acryl-PEG-SVA and allowing to react overnight at room temperature with constant agitation. The product was dialyzed against ultrapure water using a 6-8 kDa MWCO regenerated cellulose membrane (SpectrumLabs) and then the product was lyophilized. Conjugation for PEGylated peptides was validated via gel permeation chromatography (GPC) with an evaporative light scattering detector (Polymer Laboratories, Amherst, MA, USA).

Hydrogel Fabrication

Hydrogel precursor solution was made: PEG-PQ-PEG was dissolved at the desired concentration in 20 mM HEPES-buffered saline (pH 8.3) with 10 μM Eosin Y (Sigma), 1.5% v/v triethanolamine (TEOA, Sigma) (pH 8.3) and co-monomer 0.35% v/v n-vinyl-2-pyrrolidone (NVP, Sigma). For hydrogels used to encapsulate cells, PEG-RGDS was also dissolved in the precursor solution and combined with the PEG-PQ-PEG solution. For hydrogels containing alloc, the alloc monomer (H-Lys(alloc)-OH, Sigma) was also dissolved in the precursor solution before crosslinking. This water-soluble monomer consists of a lysine amino acid with an allyloxycarbonyl (alloc) allyl group conjugated to its side group.

To form a hydrogel, this gel precursor solution was pipetted onto a Sigmacote-treated glass slide, between polydimethylsiloxane (PDMS, Electron Microscopy Sciences) spacers. Then a methacrylate-modified coverslip was placed on the spacers, sandwiching the gel precursor in a cylindrical shape between glass surfaces. Glass modification protocols are included in the Supplementary Methods. The gel precursor solution was then exposed to white light from a Fiber-Lite High Intensity Illuminator Series 180 lamp (Dolan-Jenner Industries; 275-280 mW/cm²) for 50 s to form a crosslinked gel. The crosslinking mechanism employed here is free-radical-mediated chain growth polymerization of acrylate chain ends. Both acrylates and allyls are vinyl groups that can undergo chain growth polymerization, but allyls have a greater propensity to self-stabilize and terminate a chain reaction [67, 84, 91–93]. Combining alloc monomers with acrylate-terminated polymers in a gel precursor solution is hypothesized to create competition between groups to incorporate into the vinyl crosslinking centers. Upon incorporation, acrylate efficiently propagates the crosslinking reaction while alloc is more likely to cause termination, resulting in different network crosslinking mechanics. Diagrams of this process have been published previously [84]. The competition between alloc and acrylate is influenced by the molar ratio of the two groups [84].

Mechanical Testing

To determine the compressive moduli of the hydrogels, 4.5% weight/volume PEG-PQ-PEG hydrogels were fabricated with alloc at 0:1, 1:1, 2:1, 3:1, and 4:1 alloc:acrylate molar ratios (n = 4-5), and PEG-PQ-PEG hydrogels were fabricated without alloc at 4.5%, 4%, and 3% weight/volume (n = 3). Gels were made 1 mm in height and were swollen overnight in phosphate buffered saline (1 × PBS, pH 7) at 37 °C, then rinsed with PBS at room temperature. They were placed on the parallel plates of an RSA III microstrain analyzer (TA Instruments) and compressed at room temperature at a strain rate of 0.003 mm/s. From the resulting stress–strain curve, the slope of the linear region immediately following the toe region (65–70% strain) was taken as the compressive modulus.

Protein Diffusion Studies

To determine the effects of competition-mediated changes in bulk mechanical properties on hydrogel permeability, 10 µL 4.5% PEG-PQ-PEG hydrogels were fabricated at 0:1 (4.4 kPa), 3:1 (1.2 kPa), and 4:1 (0.26 kPa) alloc:acrylate molar ratios. All hydrogels were swollen overnight at 4 °C in 1 mg/mL soluble protein solutions (Trypsin Inhibitor and Carbonic Anhydrase, both from Sigma). Then each gel was transferred into a new well with 500 µL PBS and incubated at 37 °C. At each timepoint of 6, 12, 30, 60, and 120 minutes and 24 h, the gels were transferred again to new wells with 500 µL PBS. The amount of protein released into the PBS sink at each timepoint was evaluated via MicroBCA assay (ThermoFisher) (with a standard control curve made with the 0-200 µg/mL protein being tested) and normalized to the total protein diffused over 24 h to determine the rates of release from each gel. 6 gels were tested for each condition.

Cell Culture

NIH 3T3 fibroblasts were cultured in Dulbecco’s Modified Eagle’s Medium, High Glucose (Sigma) with 10% v/v bovine calf serum (BCS, Sigma) and 100 U/mL penicillin, 0.1 mg/mL streptomycin, 0.92 mg/mL L-glutamine (GPS, Corning). PC-12 cells (ATCC CRL-1721, ATCC) were cultured and expanded in suspension culture in complete growth media: RPMI 1640 ATCC modification (Thermofisher,) supplemented with 5% v/v fetal bovine serum (FBS, R&D Systems), 10% v/v heat inactivated horse serum (HIHS, Sigma), 100 U/mL penicillin, 0.1 mg/mL streptomycin (P/S, VWR Lifescience), and 0.1% v/v bovine serum albumin (BSA, Sigma). Before encapsulation, cells were stimulated for 48 h in neural differentiation media: RPMI 1640, ATCC modification + 0.5% FBS + 1% HIHS + 100 U/mL penicillin, 0.1 mg/mL streptomycin (P/S) + 0.1% BSA + 100 ng/mL nerve growth factor 2.5S (NGF, Promega). This priming step, implemented in earlier 2D PC12 culture studies, served to establish a neuronal cell phenotype prior to beginning the study [84]. After encapsulation, cells were maintained in neural differentiation media.

Rat fetal NSCs (isolated from cortexes of embryonic day 14 Sprague Dawley rats) were purchased from Thermofisher and were cultured as per the manufacturer’s instructions. Briefly, cells were first cultured on tissue culture polystyrene plates coated with poly-l-ornithine (PLO, Sigma) in “growth media” (KnockOut D-MEM/F-12 Basal Medium, 2% StemPro NSC Serum Free Supplement, 2 mM GlutaMAX-I Supplement, 20 ng/mL basic fibroblast growth factor (bFGF), 20 ng/mL human epidermal growth factor (EGF), and 5 µg/mL gentamicin) and passaged once before experimentation. The full components of the manufacturer’s media are listed in Table SI.

All cells were maintained at 37 °C with 5% CO₂.

Encapsulation of Cells Within Hydrogels

Cells were passaged and counted using a hemocytometer. After determining cell density, cells were centrifuged again at 0.5 x G and the cell pellets were resuspended at the desired density in hydrogel precursor solution. Then hydrogels were fabricated as described above, encapsulating the cells homogeneously in 3D within the polymer network.

System Cytocompatibility

3T3 fibroblasts were encapsulated at 2000 cells/µL in 4.5% w/v PEG-PQ-PEG + 3.5 mM PEG-RGDS hydrogels with no alloc (control condition at 4.4 kPa) and with alloc:acrylate ratios that corresponded to soft tissue stiffnesses (3:1 at 1.2 kPa and 4:1 at 0.26 kPa), and a Live/Dead assay (ThermoFisher, Waltham, MA, USA) was performed 24 h after encapsulation: cell media was replaced with warmed 1x PBS containing containing 4 µM ethidium homodimer (EthD-1) and 24 µM calcein AM. After 15 min of incubation in this solution at 37 °C, gels were imaged using the Zeiss 510 inverted confocal microscope using a 10 × EC Plan-Neofluar objective (NA = 0.30) at 517 nm excitation/617 nm emission wavelengths to visualize EthD-1 and 494 nm excitation/517 nm emission wavelengths to visualize calcein AM. The percentage of live cells was quantified, with n = 5-6 gels per condition.

PC12 Encapsulation and Evaluation

For evaluation of gel modulus on neural cell behavior in 3D culture, PC12s were encapsulated at low density (1,000,000 cells/mL) within 4.5% weight/volume PEG-PQ-PEG and 3.5 mM PEG-RGDS gels with different concentrations of alloc corresponding to compressive moduli of 4.4, 1.2, and 0.26 kPa (as well as a 3% PEG-PQ-PEG gel at 1.4 kPa for comparison). Neurite outgrowth was evaluated at different timepoints (4 days, 7 days, 14 days) by quantifying the percentage of neurite (+) cells. Neurite (+) cell = extends at least one neurite ≥ 2 × the cell body diameter. At each of these endpoints, the cells were fixed, permeabilized, and stained with DAPI ((4′,6-diamidino-2-phenylindole (DAPI nuclear stain; Sigma) and Alexa Fluor 488 phalloidin (F-actin probe, Thermo)) for visualization of neurites, and imaged using a Zeiss 510 inverted confocal microscope with a 40X EC Plan-Neofluarobjective (NA = 1.3) with oil immersion. A DAPI filter set (365 nm excitation/420 nm emission) was used to locate individual cell nuclei without visualizing cell morphology, and then the Argon Laser with FITC filter (494 nm excitation/517 nm emission) was engaged and used to image the entirety of each cell stained with 488-phalliodin. Z-stacks were set at 2 µm optical slice thickness, and the number of slices adjusted to capture all neurite extensions for each cell. At least 20 cells were quantified per gel across n = 4-6 gels for greater than 100 cells per condition. All cells quantified were single cells (cell body not in contact with other cell bodies). Neurites were traced using ImageJ software with the semi-automatic tracing segmentation plugin Simple Neurite Tracer (National Institutes of Health, Bethesda, MD, USA) utilized with default settings and Hessian-based analysis [94]. Neurite outgrowth for day 14 samples was also evaluated by tracing and summing the lengths of all neurites for each cell and plotting the distribution of total neurite length per cell for all conditions.

NSC Encapsulation and Evaluation

NSCs were first verified to express neural stem markers and to undergo differentiation in 2D culture in response to soluble signaling. Detailed culture, staining, and imaging methods can be found in the Supplementary Methods, and immunofluorescence images of 2D NSC culture are shown in Figs. S1 and S2. As shown in Fig S1, the NSCs expressed multipotent markers Nestin and SOX2, as well as glial marker GFAP, when cultured in growth media. Upon removal of growth factors, NSCs showed decreased Nestin and GFAP expression and the presence of beta III tubulin (Tuj1) (Fig S2), indicating neuronal differentiation. This validated that the cells were a neural stem/progenitor population.

NSCs from validated lots were encapsulated in 3.5 mM PEG-RGDS-functionalized PEG-PQ-PEG hydrogels with and without competitive alloc: after 3-4 days of 2D culture in growth conditions (reaching 70-95% confluence), NSCs were passaged using StemPro Accutase (Thermo) and encapsulated at 10,0000 cells /µL density at 4.4 kPa (0:1 alloc:acrylate) and 0.26 kPa (4:1 alloc:acrylate) conditions in “growth media” and “differentiation media,” which consisted of growth media without bFGF and EGF. Note that, as per manufacturer’s instructions, encapsulations in “differentiation media” were first cultured in “growth media” for 2 days before removing growth factors for 7 days of culture in media without growth factor stimulation.

Encapsulations were fixed at day 7, then immunostained for NSC marker Nestin and neuronal differentiation marker beta III tubulin (Tuj1), as well as DAPI (nuclear marker). Detailed protocols for the 3D culture immunostaining process are found in the Supplementary Methods. Hydrogels were imaged with a Zeiss 880 inverted confocal microscope with Airyscan with a 405 nm diode laser to visualize DAPI, Argon/2 laser to visualize Alexa Fluor-488 (filtered to 498-553 nm), and 561 diode laser (filtered to 561-624 nm) to visualize Alexa Fluor 555. 3D cell cultures in hydrogels were imaged with a 20x Plan-Apochromat objective (NA: 0.80). 4 randomly located 30 µm stacks were imaged per gel, with 2.5 µm optical slice thickness. All stacks were summed, and images were processed and analyzed in ImageJ with the same brightness and contrast values across all images for all conditions for each channel. Image processing parameters are provided in the Supplementary Methods. Images were evaluated for cell morphology via quantification of cell circularity and for neurogenic marker expression via Tuj1-stained area normalized to DAPI stained area.

Statistical Analysis

All statistical analyses were performed with JMP Pro software (SAS Institute, Cary, NC, USA). Data sets were evaluated using a student’s t-test for comparisons between two groups or using analysis of variance (ANOVA) followed by a Tukey’s Honest Significant Difference (HSD) post-hoc test to compare across multiple groups. For all tests, unless otherwise specified, p-values of less than 0.05 are considered significant. All data values are reported as mean ± standard deviation. PC12 neurite length distributions were analyzed with pairwise nonparametric Kolmogorov–Smirnov asymptotic tests with a significance threshold of p < 0.0167 (Bonferroni-corrected for the three pairwise tests). NSC circularity value distributions were analyzed with a two-sample Kolmogorov–Smirnov asymptotic test.

Results and Discussion

Degradable Hydrogel Stiffness Reduced by Alloc Competition in Dose-Dependent Fashion

PEG-PQ-PEG hydrogel mechanical properties were analyzed via static compression testing to evaluate the effects of including soluble H-Lys(alloc)-OH on the material modulus. It was found that the addition of H-Lys(alloc)-OH reduced degradable hydrogel stiffness in a similar manner to non-degradable PEGDA gels tested previously [84].

The 4.5% w/v PEG-PQ-PEG (control) condition gels had an average modulus of 4.40 ± 0.89 kPa, which was decreased by adding soluble H-Lys(alloc)-OH to the gel precursor mixture (Fig. 1a). With increasing alloc:acrylate molar ratio, the gel compressive modulus was decreased in a concentration-dependent manner. A compressive modulus of 0.26 ± 0.06 kPa was achieved at 4:1 alloc:acrylate (10.99 mg/mL H-Lys(alloc)-OH), falling within the range of neural tissue stiffness. These hydrogels made with 4:1 alloc to acrylate are 94.1% softer than the control hydrogels made without alloc. This decrease in compressive modulus achieved via soluble competitive monomer was compared to the decrease achieved by reducing the polymer density. (Fig. 1b). As has been shown widely [75, 81], the compressive modulus decreased in direct correlation with polymer density, from 4.40 ± 0.89 kPa at 4.5% w/v to 4.12 ± 0.24 kPa at 4% and 1.39 ± 0.19 kPa at 3% w/v. Unlike the ultra-low stiffnesses achieved with H-Lys(alloc)-OH, 1.39 ± 0.19 kPa (exhibited by 3% gels) was the lowest compressive modulus achieved in PEG-PQ-PEG gels manipulated by decreasing polymer density, falling just outside the range of neural tissue. PEG-PQ-PEG hydrogels did not reliably form at polymer densities lower than 3% w/v. Thus, implementing soluble alloc for crosslinking competition was determined to be effective in reducing hydrogel compressive modulus of this PEG-PQ-PEG degradable matrix system to within the neural tissue compliance range as was previously observed in PEGDA gels [84].

Fig. 1.

Compressive moduli of PEG-PQ-PEG hydrogels in response to H-Lys(alloc)-OH implementation and decreasing polymer density. Compressive modulus = slope of the stress–strain curve in the linear region following the toe region. Values reported as mean ± S.D. a gel modulus reduced by increasing H-Lys(alloc)-OH concentration, neuronal stiffness achieved at highest monomer concentrations. n = 4-5. *p < 0.05, **p < 0.001, ***p < 0.0001, indicating a statistically significant difference. Statistical analysis evaluated by one-way ANOVA followed by Tukey–Kramer HSD tests. b reducing polymer density decreases compressive modulus but does not reach neural tissue compliance. n = 3. **p < 0.001, indicating a statistically significant difference. Statistical analysis evaluated by student’s t-test

Protein Diffusion not Influenced by Alloc-Mediated Stiffness Modulation

In 3D tissue culture, soluble growth factors and other molecules must diffuse through the hydrogel matrix to be accessed by cells. Thus, methods for altering crosslinking mechanics must be evaluated for their effect on diffusivity of key proteins known to influence neural or other soft tissue cell activity.

Protein diffusion was evaluated in 4.5% w/v PEG-PQ-PEG hydrogels fabricated with no alloc (4.4 kPa), 3:1 alloc:acrylate (1.2 kPa) and 4:1 alloc:acrylate (0.26 kPa) by placing hydrogels hydrated with 1 mg/mL protein solution (Trypsin Inhibitor or Carbonic Anhydrase) in PBS sink conditions and measuring the percentage of protein diffused out at set timepoints. These two proteins were selected because they bracket 26 kDa NGF in size. NGF is a key neurotrophic factor.

The protein diffusion rate profiles for both trypsin inhibitor (Fig. 2a) and carbonic anhydrase (Fig. 2b) were comparable between each hydrogel stiffness condition (4.4 kPa hydrogels made without alloc, 1.2 kPa hydrogels made with alloc, and 0.26 kPa hydrogels made with alloc), with all 3 diffusion curves following the same pattern: hydrogels of all stiffness conditions showed rapid protein diffusion at initial timepoints with approximately 50-60% of protein diffused out at 6 minutes and approximately 90% of protein diffused out within 1 h. Altering the hydrogel crosslinking structure by adding alloc to significantly decrease the hydrogel stiffness did not result in a significantly greater percentage of protein released at any of the timepoints evaluated. These results indicate that the alloc-mediated method of tuning hydrogel modulus is orthogonal to material diffusive properties and that inferences may be made about encapsulated cell behavior based on hydrogel mechanical modulus independent of accessibility of these soluble factors.

Fig. 2.

Decreasing PEG-PQ-PEG hydrogel stiffness by addition of H-Lys(alloc)-OH does not alter protein diffusion profiles. a Percentage of 20 kDa protein Tryspin Inhibitor released at 6, 12, 30, 60, 90, and 120 minutes. b Percentage of 29 kDa protein Carbonic Anhydrase released at 6, 12, 30, 60, 90, and 120 minutes. All values are percentages of total protein released over 24 h. Values reported as mean ± S.D. n = 6 hydrogels. For both (a) and (b), protein shows the same diffusion pattern from 4.4 kPa, 1,2 kPa, and 0.26 kPa hydrogels

Encapsulation with H-Lys(alloc)-OH for 3D Cell Culture is Cytocompatible

This 3D cell culture model requires cells to be mixed into the hydrogel precursor in direct contact with the soluble alloc monomer before crosslinking into the hydrogel network; thus, cell viability within this system was assessed. 3T3 fibroblasts were encapsulated in 4.5% w/v PEG-PQ-PEG + 3.5 mM PEG-RGDS hydrogels with no alloc (control condition at 4.4 kPa), 3:1 alloc:acrylate (1.2 kPa), and 4:1 (0.26 kPa), and LIVE/DEAD staining and quantification was performed 24 h after encapsulation. It was found that none of the alloc conditions significantly reduced fibroblast viability. Fig. S3a shows representative LIVE/DEAD staining images of each condition: cell viability was 94.81 ± 3.03% in 0.26 kPa gels and 95.43% ± 1.95% in 1.2 kPa hydrogels, neither of which varied significantly from 96.64% ± 1.29% in 4.4 kPa control gels (Fig. S3b) (p-value = 0.44). These results indicate that this compliant hydrogel system is suitable for cell culture.

PC12 Neurite Outgrowth is Enhanced in Softest Hydrogels

PC12 neural cells were encapsulated in PEG-PQ-PEG + PEG-RGDS hydrogels with and without soluble competitive alloc to evaluate the influence of environmental stiffness on neural cell behavior in the 3D tissue culture model. PC12s were evaluated by quantifying the percentage of cells extending neurites ≥ 2x the length of the cell body at days 4, 7, and 14 after gelation. At all timepoints evaluated, neurite outgrowth was inversely related to hydrogel stiffness, and hydrogels fabricated with alloc supported greater neurite outgrowth than those without alloc. At 4 days after encapsulation, PC12s in soft hydrogels with alloc showed significantly greater neurite outgrowth than those without alloc: 25.99 ± 6.65% of cells in 0.26 kPa gels and 19.45 ± 10.47% of cells in 1.2 kPa extended neurites at least twice the length of the cell bodies, as compared to 8.52 ± 5.97% neurite (+) cells in 1.4 kPa gels and 1.52 ± 2.62% neurite (+) cells in 4.4 kPa gels, indicating accelerated neurite extension (Fig. 3a). At day 7, while neurite growth increased from day 4 for all stiffness conditions, the softest gels continued to show higher numbers of neurite (+) cells by comparison. Percentages of neurite (+) cells were 44.38 ± 2.56 % in 0.26 kPa hydrogels, 30.85 ± 3.41 % in 1.2 kPa gels, 21.57 ± 3.920.9% in 1.4 kPa gels, and 10.94 ± 3.04% in 4.4 kPa hydrogels (Fig. 3b). At 14 days after encapsulation, PC12s in 0.26 kPa hydrogels showed striking neurite outgrowth of 64.99% ± 5.64% of neurite (+) cells, significantly greater than all other conditions (p < 0.0001). The outgrowth in these gels was 993% greater than that of cells in the 4.4 kPa control hydrogels without alloc, which showed just 5.95 ± 4.47% neurite (+) cells, and 191% higher than 1.4 kPa, the lowest stiffness achievable by decreasing polymer density without competing monomers, where 23.32 ± 6.35% of cells were neurite (+) (Fig. 3c).

Fig. 3.

Neurite outgrowth is increased within softest hydrogels at all timepoints evaluated and continues over the longest time period. PC-12 neurite outgrowth is shown as a function of gel modulus. a Percentage of neurite (+) cells at 4 days after encapsulation. b Percentage of neurite (+) cells at 7 days after encapsulation. c Percentage of neurite (+) cells at 14 days after encapsulation. For all graphs, data is shown as mean ± S.D. *p < 0.05, **p < 0.001, ***p < 0.0001, indicating a statistically significant difference. Statistical analysis evaluated by one-way ANOVA followed by Tukey–Kramer HSD tests, n = 4-6 gels, ≥ 20 cells counted per gel. d PC-12 neurite outgrowth compared across day 0, 4, 7, and 14 post-encapsulation. Data is shown as mean ± S.D. and statistical comparison was made by pairwise student’s t-tests between time points within each stiffness condition. **p < 0.001 indicates a statistically significant difference between values at day 7 and day 14, while n.s. indicates no significant difference in values between day 7 and day 14. n = 4-6 hydrogels, ≥ 20 cells counted per gel

This consistent concentration-dependent direct relationship between hydrogel compliance and neurite outgrowth is indicative of the ability to tune neurite outgrowth by adjusting the hydrogel stiffness. Alloc-mediated control over cell behavior holds promise for influencing cell activity in other soft tissue constructs. Comparison of percentage neurite (+) cells across timepoints also demonstrates an influence of stiffness on the timescale over which neurites extend. From 4 days to 7 days, neurite outgrowth for every stiffness condition increased significantly; however, at all conditions stiffer than 0.26 kPa, neurite outgrowth at day 14 was not significantly different than that at day 7 as evaluated by pairwise student’s t-tests for each time point (Fig. 3d; p = 0.07 for 4.4 kPa gels, p = 0.84 for 1.4 kPa gels, p = 0.42 for 1.2 kPa). In contrast, PC12s encapsulated in 0.26 kPa hydrogels showed a statistically significant (p < 0.001) 46.43% increase in neurite (+) cells from day 7 to 14. This suggests that PC12s reach their maximum percentage of neurite (+) cells at earlier timepoints in a stiffer environment, while the softest hydrogel is supportive of new neurite outgrowth over a longer period.

Additionally, the length of neurite extensions was characterized at day 14. For each individual cell evaluated, the lengths of all neurite projections from that single cell were measured and summed, and the distribution of these length sums for all cells across all gels was graphed for each stiffness condition. The results followed the same trends as neurite (+) cell quantification: the sum of neurite lengths for each cell was greater in the softer gels, and the distribution of neurites in the 0.26 kPa gels was significantly different than those in all stiffer conditions by pairwise Kolmogorov–Smirnov tests with Bonferroni correction (of *p < 0.0083), showing a distribution trending towards longer neurites in comparison to those within stiffer gels (Fig. 4a). These results demonstrate that neurites are not only more numerous, but also longer in the softest gels, which serve as a highly supportive environment for neurite extension from neural cells. Representative images of day 14 PC12s at each hydrogel stiffness stained with DAPI and Phalloidin are shown in Fig. 4b-4e.

Fig. 4.

Neurite length distribution is influenced by hydrogel stiffness, with an inverse relationship between gel stiffness and neurite length. a Distributions of total neurite length per cell for cells in 4.4 kPa, 1.4 kPa, 1.2 kPa, and 0.26 kPa hydrogels. Cells binned by neurite length in bins of 25 µm and graphed as a percentage of total cells for each condition. > 120 cells total counted per condition. **p < 0.001, ***p < 0.0001 indicating a statistically significant difference between distributions by pairwise Kolmogorov–Smirnov asymptotic tests with Bonferroni correction. b Representative confocal images of PC12s in 4.4 kPa hydrogels c 1.4 kPa hdyrogels d 1.2 kPa hydrogels e 0.26 kPa hydrogels. All images = 18 µm stacks (z-projections). Scale bars = 50 µm

NSCs Exhibit Neural-Type Behavior in Softest Gels

To evaluate the translation potential of this system, the more therapeutically relevant primary cell type rat fetal NSCs were next implemented in culture within the alloc hydrogels. NSCs were first cultured in growth media in 4.4 kPa hydrogels without alloc and 0.26 kPa with alloc to determine the effect of alloc-controlled matrix compliance on NSC behavior and morphology in growth media conditions. 30 µm stacks of images were summed and then merged across the DAPI, Tuj1, and Nestin channels, then the composite images were made binary, and a particle analysis was performed to evaluate roundness of cells and contiguous cell clumps, in order to compare cell morphology.

As shown in the graphed distributions of circularity values (calculated as $\frac{4\times \pi \times area}{{\mathit{perimeter}}^2}$, with 1.0 as a perfect circle) for all “particles” and accompanying representative images in Fig. 5a, as well as the average circularity values per gel in Fig. 5b, NSCs largely remained rounded and/or clumped together in 4.4 kPa hydrogels but showed significantly more spreading and early neurite-like extension in multiple dimensions in the 0.26 kPa environment. Additional representative images are shown in Fig S4. These results indicate that the alloc-controlled softest gels present a more favorable environment for NSC culture than gels without alloc, which fall outside the neural stiffness regime. This finding is consistent with the previous literature which has shown that matrices closest in stiffness to the neural range are more supportive of 3D NSC survival, growth, and neuronal differentiation than matrices which are stiffer [56, 62, 95].

Fig. 5.

NSCs respond to stiffness and growth factor signaling within degradable PEG-PQ-PEG hydrogels a NSCs display increased spreading and elongation within softest alloc hydrogels. Distributions of “particle” circularity values; values trend more towards circular for 4.4 kPa, ***p < 0.0001 indicating a statistically significant difference between distributions by two-sample Kolmogorov–Smirnov asymptotic test. Fluorescent images demonstrate more elongated morphology in 0.26 kPa hydrogels. Images = 30 µm stacks (z-projections). Scale bars = 100 µm. b Average circularity values, n = 3-4 hydrogels per condition. ***p < 0.0001 indicates a statistically significant difference between values via student’s t-test. c Within 0.26 kPa hydrogels, NSCs cultured in “differentiation media” showed more Tuj1 expression per DAPI staining. Confocal images of Tuj1 expression, 30 µm stacks (z-projections), scale bars = 100 µm. d average Tuj1:DAPI area ratios, *p < 0.05 indicates a stastically significant difference in Tuj1 (+) area per DAPI area values (p = 0.0494). n = 4 gels per condition

Additionally, this soft alloc hydrogel culture system was tested as a platform for evaluating NSC response to biological signaling by comparing these NSCs in 0.26 kPa hydrogels in growth media with those cultured in growth-factor-starved “differentiation media.” At 7 days of culture in either growth or differentiation media, gels were fixed, stained, and imaged. Within each summed 30 µm stack (representative images shown in Fig. 5c), measurements were performed in ImageJ to determine the Tuj1 (+) area (green channel, area measurement function) normalized to DAPI (+) area (blue channel, particle analysis function) for each gel. The detailed image processing parameters for these measurements are provided in the Supplementary Methods. The average normalized Tuj1 expression was greater in the differentiation media gels than in the growth media gels (p < 0.05, Fig. 5d). These results show that the NSCs are responsive to biological signals within the alloc gels: they express more neural marker Tuj1 when growth factors are removed to create differentiation-favoring conditions.

Although these results do not show optimized neurogenesis, they do show that NSC behavior can be influenced by environmental factors within this hydrogel system. This suggests the applicability of the degradable PEG-alloc hydrogel as a NSC culture platform which can be further modified to enhance neurogenesis. Potential future studies include adjusting immobilized and soluble biochemical signaling within this system and extending the NSC culture period to optimize differentiation into mature neurons, evaluated by expression of mature neuronal markers such as MAP2 and NeuN.

While many degradable hydrogels have previously been designed for 3D NSC culture [62, 96–100], including those containing PEG [74, 80], this system is the first to our knowledge which presents a fully synthetic hydrogel that can be modularly tuned between 4.4 and 0.26 kPa, implemented to support primary NSC culture in 3D.

Conclusion

Towards developing physiologically relevant 3D soft tissue culture models, we extended the application of allyl-acrylate competition in cell-degradable PEG-PQ-PEG hydrogels to confer soft-tissue representative stiffness to reductionist degradable synthetic hydrogels. Here we demonstrated that this method of stiffness modulation does not influence diffusion rates of key proteins through the matrix and is biocompatible. We then implemented this hydrogel system for neural cell culture and showed that stiffness modulation with alloc can be used to control PC12 neurite outgrowth, with the softest hydrogels supporting extensive neurite projection in both number and length. Finally, having developed a reductionist soft synthetic hydrogel matrix for neural cell culture, we next evaluated this system as a platform for culturing more physiologically representative neural cells: primary rat fetal NSCs. The NSCs showed elongated morphology and responded to soluble cues within the soft degradable hydrogels, establishing these hydrogels as an appropriate soft tissue culture system for modeling interactions of the NSC niche. In addition to improving biochemical signaling within the matrix to optimize neurogenesis, a compelling future application is implementing this hydrogel system as a platform for further investigating the role of NSC mechanotransduction in neurogenesis. NSC mechanical signal transduction and mechanosensitive ion channel activity may be evaluated within the soft network, and atomic force microscopy may be implemented to characterize stresses created between cells and polymer matrix. With fine control over network stiffness within the soft tissue regime, this model presents an advantageous platform for evaluating mechanical forces in neural cell culture. Additionally, the modular nature of this system allows independent investigation of single and multiple environmental factors on NSC behavior in a reductionist environment, which may improve understanding of neurogenesis and inform design of future therapeutic NSC delivery matrices.

Supplementary Information

Below is the link to the electronic supplementary material.

Data availability

The data that support this published work are available upon request from the corresponding author, RC.

Declarations

Conflict of interest

The authors state no conflicts of interest or competing interests for this work.