Microporous Annealed Particle (MAP) Scaffold Pore Size Influences Mesenchymal Stem Cell Metabolism and Proliferation Without Changing CD73, CD90, and CD105 Expression Over Two Weeks

Abstract

Scaffold pore architecture has been shown to influence stem cell fate through various avenues. We demonstrate that MAP microgel diameter can be tuned to control scaffold pore size and, in turn, modulate MSC survivability, proliferation, metabolism, and migration, thereby enhancing bioactivity and guiding future applications of MAP for regenerative medicine.

Graphical Abstract

Scaffold pore size has been shown to be a bioactive feature and can be tuned in microporous annealed particle (MAP) scaffolds through microgel diameter. In this manuscript, we demonstrate the impact of porosity on mesenchymal stem cell (MSC) survivability, proliferation, metabolism, and migration. We believe that the ability to expand MSCs independent of phenotype shift in a 3D format will provide future routes to manufacturing and study of MSCs.

1. Introduction

Stem cells are a powerful tool in the regenerative medicine toolbox due to their unlimited self-renewal and their potential to differentiate into multiple other cell types.¹ Moreover, stem cells naturally secrete paracrine signals that promote cell growth, prevent apoptosis, and suppress inflammation.^2,3 The regenerative potential of these endogenous super cells can be harnessed for various tissue applications by carefully designing instructional biomaterial constructs.⁴ In particular, scaffold pore architecture (pore size, interconnectivity, and total porosity) has been shown to influence stem cell fate.

Pore architecture plays a role in scaffold permeability, thus modulating the diffusion of oxygen, nutrients, and cell-secreted signals within the scaffold.⁵ Pore architecture also affects vascularization and ECM deposition within a construct, which are critical for neotissue formation.^5,6 Furthermore, individual pore size can impact stem cell adhesion and subsequent mechanotransduction mechanisms that help determine cell proliferation, migration, apoptosis, and differentiation.^5,7,8

Due to this importance, discrepancies in the reported pore size effects on stem cell functions present a challenge for researchers developing biomaterials for tissue regeneration. For example, there are inconsistent results pertaining to the pore-related impacts on MSC proliferation, adhesion, and migration. One study indicates scaffolds with 200 μm pores are best for MSC proliferation when compared to 50-, 100-, 300-, and 400-μm-pore scaffolds.⁹ Contrarily, other reports show MSC proliferation improves with increasing pore size when 94, 130, and 300 μm¹⁰ or 200 and 500 μm¹¹ scaffolds are compared. Regarding cell adhesion, scaffolds with 200 μm pores have been shown to outperform ones with 50, 100, 300, and 400 μm pores⁹, but a different report has shown 325 μm pores are more conducive to cell adhesion than 85-, 120-, 164-, and 190-μm-pore scaffolds.¹² While some researchers demonstrated more effective MSC migration in scaffolds with smaller pores (12 vs 17 μm)¹³, others observed no difference between pore sizes (125–250 vs 425–600 μm)⁶ or that larger pores were more favorable (94 or 130 vs 300 μm)¹². While these discrepancies may be related, at least in part, to the choice of biomaterial¹⁴ and source of MSCs¹⁵, they demonstrate the importance of investigating pore size effects in each unique porous system.

Herein, we elucidate the role of pore size in the microporous annealed particle (MAP) hydrogel biomaterial system. MAP scaffolds are a promising platform for regenerative medicine due to their injectability and unique building-block approach. They are comprised of a flowable slurry of individual hydrogel microspheres that undergo a secondary crosslinking step in situ to form a structurally stable construct with interconnected cell-scale pores.¹⁶ The micro-scale porosity provided by MAP has previously been demonstrated to enhance the proliferation of MSCs compared to nanoporous hydrogels.¹⁷ However, to our knowledge, the functional consequences of varying MAP scaffold pore size have not been fully explored for MSC adhesion, proliferation, migration, and spontaneous differentiation. We investigate these effects by fabricating three distinct populations of microparticles. Due to MAP’s building-block approach, varying the microsphere diameters results in scaffolds with discrete pore sizes.

2. Materials and Methods

2.1. Sources of Materials

Four-arm poly(ethylene glycol)-maleimide (PEG-mal) (10kDa) was purchased from NOF America Corporation. RGD cell-adhesive peptide (Ac-RGDSPGGC-NH₂) and the MMP-2-degradable crosslinker (Ac-GCGPQGIAGQDGCG-NH₂) were purchased from WatsonBio Sciences. MethMal annealing macromer was synthesized as previously described.¹⁸ Biotin-labeled maleimide was purchased from TCI America.

2.2. MAP Hydrogel Production and Preparation

Pre-Gel Solution / Chemical Formulation

A 5.5 wt% (w/v) hydrogel was used for all particle sizes and experiments. PEG-mal was dissolved as the backbone in 10X PBS pH=1.20 at a final concentration of 60.84 mg/mL. That solution was used to dissolve the RGD peptide (0.75 mg/mL) and MethMal (7.90 mg/mL) prior to mixing with the crosslinker solution. The MMP-2-degradable crosslinker was dissolved in MilliQ at a final concentration of 12.19 mg/mL, and biotin-labeled maleimide was added at a final concentration of 5 μM. Backbone and crosslinker working solutions (2X above concentrations) were mixed 1:1 to create a pre-gel solution with the above final concentrations.

Macrogel Production

Macro-scale hydrogels (macrogels) were used to determine hydrogel stiffness. 100 μL pre-gel solution was used to form pucks 2 mm in thickness between SigmaCote-coated slides. Pucks were gelled in a humidified environment overnight. Pucks were then weighed and swollen overnight in PBS pH=7.4 overnight at 37 °C. After swelling, the macrogels were wicked of excess moisture and weighed again.

Microgel Production

Microgels of small, medium, and large diameter were produced using a published microfluidic water-in-oil emulsion technique¹⁹. PDMS molds of 5, 11, and 20 μm channel heights were used for small, medium, and large particles, respectively. Surfactant (RAN Biotechnologies) was diluted in NOVEC 7500 (3M) to 1% for medium and large microgels and 2% for small particles. Using syringe pumps, the surfactant and pre-gel solutions were flowed through a 0.22 μm syringe filter (pre-gel only), through the microfluidic device, and then through a cell strainer (Greiner Bio-One) before collection in a 50 mL conical tube. For the medium and large microgels, the surfactant was flowed at a ratio of 2:1 (5 mL/hr to 2.5 mL/hr) to the aqueous pre-gel solution and 100 μm cell strainers were used. For the small particles, the flow rate ratio was 5:1 (5 mL/hr to 1 mL/hr) and a 40 μm cell strainer was used. Triethylamine (20 μL/mL of gel) was diluted in NOVEC 7500 and added to the microgels to ensure complete gelation prior to purification.

Microgel Purification

Microgels were purified as previously described by our lab.²⁰ In summary, particles were washed three times with NOVEC 7500 (1 x gel volume), then PBS pH=7.4 was added (5 x gel volume) for the particles to swell. Leaving the PBS in the conical, three more NOVEC 7500 washes were performed, with the gels being pulsed on a centrifuge at 4696 g for 10 seconds between washes to separate the phases. Next, the oil was removed and microgels were washed simultaneously with PBS (4–5 x gel volume) and hexanes (4–5 x gel volume). For these washes, as well as all future microgel wash steps, particles were centrifuged at 4696 g for 5 min. Microgels were then washed once entirely with PBS to remove all hexanes. Particles were incubated overnight at 37 °C in 100 mM N-acetyl-L-cysteine dissolved in PBS to quench any unreacted maleimides. After quenching, particles were washed once with PBS and then filtered again with 70 μm, 100 μm, and 150 μm cell strainers for small, medium, and large particles, respectively. All non-sterile gels were stored in PBS at 4 °C. Batches of each size were pooled before subsequent testing and use.

Microgel Sterilization

Particles for cell-related studies were washed three times with 70% isopropanol (at least 5 x gel volume) and stored in this solution at 4 °C until use. In a biosafety cabinet, microgels were washed three times with sterile 1X PBS pH=7.4 prior to incubation in a photoinitiator solution.

2.3. Mechanical Testing

Compressive stiffness was tested using a Bluehill Universal Instron at a rate of 0.5 mm/min for 2 min (ε=−36% since pucks swelled to 2.8 mm) using a 5.85 mm probe. To account for variations in sample height after swelling, datasets were truncated to 36% engineering strain. All samples were exposed to annealing conditions (0.2 mM LAP in growth medium with 365 nm at 33.4 mW/cm² for 29 seconds) prior to testing. Bluehill Universal software recorded the load (N) and extension (mm) and produced a stress-strain curve. Young’s moduli of n=6 macrogels were calculated using a custom MATLAB code.

2.4. Particle Size Characterization

Oil Phase Particle Sizing

Syringe pumps were paused briefly during production to capture images of MAP particles within each microfluidic device. ImageJ was used to manually determine the diameter of at least 35 microgels per group.

Aqueous Phase Particle Sizing

Hydrogel particles were diluted 1:200 in a PBS solution of fluorophore-conjugated streptavidin. Microgel solutions were sonicated for 5 minutes then imaged on Molecular Devices ImageXpress Micro Confocal. ImageXpress software was used to analyze a minimum of 500 microgels per batch (n=8) for the average particle diameter and polydispersity index (PDI). PDI was calculated by ${\text{D}}_{\text{w}}/{\text{D}}_{\text{n}}$ with weighted average ${\text{D}}_{\text{w}}=\sum {\text{N}}_{\text{i}}{\text{D}}_{\text{i}}^2/\sum {\text{N}}_{\text{i}}{\text{D}}_{\text{i}}$ and number average ${\text{D}}_{\text{n}}=\sum {\text{N}}_{\text{i}}{\text{D}}_{\text{i}}/\sum {\text{N}}_{\text{i}}$.

2.5. MAP Scaffold Pore Size Characterization

Pore Size Imaging

Microgels were incubated 1:1 for at least 30 minutes in a PBS-based photoinitiator solution of 0.2 mM lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP). Particles were dried via centrifugation (18,000 g x 5 min) and aspiration of the supernatant. Custom molds were created by punching a 3 mm hole into DuoDERM and sticking the material to a glass slide. Microgels filled the molds and were annealed together with UV light (365 nm) for 29 seconds at 33.4 mW/cm². A fluorophore-conjugated dextran solution (100 μg/mL, 2,000 kDa) was added to the scaffold, and the gel was covered with a cover slip. Scaffolds were incubated in a dark, humidified environment for at least 2 hours, then z-stack images were captured using multiphoton confocal microscopy (Zeiss LSM 710, Advanced Microscopy Facility at the University of Virginia, RRID: SCR_018736).

Pore Size Analysis

Microscopy image analysis software (Bitplane IMARIS) was used to create a 3D rendering of each z-stack. To do this, a new “Surface” was created for each sample by thresholding to include the maximum number of pores on the top (brightest) and bottom (dimmest) slices. Automated analysis of the Surface revealed the number of voxels making up each pore, which was converted into a volume based on the z-stack dimensions. A small number of excessively large pores skewed individual samples, regardless of particle size, so the geometric mean pore volume was calculated for each z-stack. These values were used to report the arithmetic mean and standard deviation for each group. Total porosity describes the percentage of the entire z-stack image volume that was occupied by rendered pore volumes.

2.6. Seeding of Mesenchymal Stem Cells

Source of MSCs

Human Bone Marrow-Derived Mesenchymal Stem/Stromal Cells isolated from adult human iliac crest from a variety of normal, healthy donors (Rooster Bio Cat# MSC-001). MSCs were expanded according to the manufacturer’s instructions. Growth medium consisted of low-glucose DMEM without phenol red (Gibco) plus 10% fetal bovine serum (Gibco), 2% L-glutamine (Gibco), and 1% antibiotic antimycotic solution (Millipore Sigma). Studies were started using cells at passage 5, meaning they were cultured on MAP scaffolds during passage 6.

Seeding MSCs on Top of MAP Scaffolds

Sterile microparticles were incubated 1:1 in a sterile-filtered solution of 0.2 mM LAP in growth medium for at least 30 min at 37 °C. Particles were dried via centrifugation (18,000 g x 5 min) and supernatant aspiration. 20 μL samples of each gel type were pipetted into 96-well plates, and then plates were centrifuged (2,000 g x 10 sec) in both directions to flatten the gel and maintain consistent diffusion of oxygen and nutrients across samples. Scaffolds were then annealed using UV light (365 nm at 33.4 mW/cm² for 29 seconds). Note: The centrifugation step was found necessary to prevent cells from clumping in areas of low height. Cells were seeded on top of scaffolds at a density of 200 cells/μL gel, which resulted in 4000 cells per 96-well sample, directly prior to adding growth media. Notably, media was very carefully added by pipetting slowly down the side of each well to minimize disturbing the cells and the gel.

Seeding MSCs Mixed within MAP Scaffolds

Sterile microgels were incubated 1:1 in a sterile-filtered solution of 0.2 mM LAP in growth medium for at least 30 min at 37 °C. Particles were dried via centrifugation (18,000 g x 5 min) and supernatant aspiration. MSCs were lifted, counted, and aliquoted into separate microcentrifuge tubes for each gel type. The seeding density of the cells was 400 cells/ul gel, which resulted in 8000 cells per 96-well sample. Cells were pelleted (300 g x 5 min) and the supernatant was aspirated. Each pellet was resuspended in MAP particles at a density of 400 cells/μL gel. 20 μL samples of gel and cells were then pipetted into 96-well plates and centrifuged (2,000 g x 10 sec) in both directions. Note: The centrifugation step was found necessary to match the scaffold geometry used in the top seeding method described above. Scaffolds were then annealed using UV light (365 nm at 33.4 mW/cm² for 29 seconds) prior to adding growth media. Notably, media was added immediately and slowly to avoid disturbing scaffold.

2.7. Survivability Assays

Cells were seeded onto or within scaffolds, covered with growth media, and incubated at 37 °C for 24 hours. Media was then removed, and scaffolds were washed once with sterile PBS. The Live/Dead Cell Imaging Kit (Invitrogen) was prepared according to the manufacturer’s instructions and applied at a 1:1 volume ratio to each scaffold. After incubation for 20 minutes, scaffolds were imaged via the ImageXpress Micro Confocal and the number of live and dead cells were counted. Cell viability was calculated using the equation Viability = Live/(Live+Dead).

2.8. Proliferation and Metabolism Assays

MSCs were seeded onto or within scaffolds, covered with growth media, and incubated at 37 °C for up to 14 days with medium changes every 3 days. A different plate was used for each timepoint. At each timepoint, media was removed and replaced with PrestoBlue Cell Viability Reagent (Invitrogen) diluted 1:10 in fresh growth medium to examine cell metabolism. Scaffolds were incubated for 2 hours at 37 °C before measuring the fluorescence (560/590 ex/em) of the supernatant. PrestoBlue solution was then removed, and scaffolds were washed twice with sterile PBS before being frozen at −80 °C. Upon completion of the study, scaffolds were thawed and degraded at 37 °C for 60 min using 100 μL of 100 μg/mL Liberase TM (Roche) dissolved in DI water. CyQUANT Cell Proliferation Kit (Invitrogen) was used to measure DNA content by adding 100 μL of the CyQUANT dye/lysis buffer (prepared at 2X) to each scaffold using a multichannel pipette. After 3 minutes of incubation, a multichannel pipette was then used to transfer 125 μL from each well to a black-bottom 96-well plate. Fluorescence was measured (485/530 ex/em) using the same gain value for each plate.

2.9. Proliferation Staining

On Day 14, scaffolds were washed with sterile PBS and fixed with 4% paraformaldehyde for 30 minutes. Scaffolds were then washed three times with PBS for 5 minutes each. Cells were permeabilized with a 0.1% Triton X-100 solution in PBS for 20 minutes, followed by another three 5-minute PBS washes. Alexa Fluor Phalloidin 647 (1:1000, Invitrogen) was diluted in 1% bovine serum albumin in PBS and added to each well for 1 hour. After three more 5-minute PBS washes, a DAPI solution in PBS (1:1000, Thermo Scientific) was applied for 20 minutes. A final 5-minute PBS wash was performed before covering scaffolds in PBS for imaging via the ImageXpress Micro Confocal.

2.10. Migration Assays

MSCs were resuspended at a density of 400k cells/mL in a solution of CellTracker Green CMFDA Dye (1:750 dilution, Invitrogen), 1% methyl cellulose, and growth media. 20 μL droplets were cultured for 48 hours at 37 °C using the hanging droplet technique.²¹ On the day of the experiment, microgels were incubated 1:1 in sterile-filtered 0.2 mM LAP in growth media for at least 30 min at 37 °C. Particles were dried via centrifugation (18,000 g x 5 min) and the supernatant was aspirated. Dried gel was pipetted into a 96-well plate and centrifuged (2,000 g x 10 sec) in both directions to flatten the gel. Scaffolds were then annealed using UV light (365 nm at 33.4 mW/cm² for 29 seconds). Spheroids were then transferred to the top-center of each gel via pipette and then imaged for the Day 0 timepoint. At 24 hours, the spheroids were imaged again via the ImageXpress Micro Confocal. Fold change in the maximum radial distance from was calculated using ImageJ.

2.11. Spontaneous Differentiation Assays

MSCs were seeded within scaffolds at 400 cells/μL, covered with growth media, and incubated at 37 °C for 14 days with medium changes every 3 days. Media was removed, scaffolds were washed once with sterile PBS, and gels were degraded at 37 °C for 60 min with 100 μg/mL Liberase TM (3X gel volume). Samples of the same group were combined, transferred to a microcentrifuge tube, and pelleted (300 g x 5 min) for Fc Receptor blocking (decrease non-specific) and staining according to manufacturer protocol. 2D controls were grown to the same passage, lifted with 0.05% Trypsin-EDTA (Gibco) or 100 μg/mL Liberase TM, then pelleted for staining. Flow cytometry was performed using the Cytek Aurora Full Spectral 3-Laser (University of Virginia Flow Cytometry Core, RRID: SCR_017829). Fc Receptor Blocking Solution was obtained from BioLegend (Cat# 422302). CD105 (Endoglin) Monoclonal Antibody tagged with PE fluorophore was obtained from Invitrogen (Cat# 505–105-51). CD90 (Thy-1) Monoclonal antibody was obtained from BioLegend (Cat# 328113). CD73 (Ecto-5’-nucleotidase) Monoclonal Antibody was obtained from BioLegend (Cat# 344011). CD45 (LCA, T200) Monoclonal Antibody was obtained from BioLegend (Cat# 368508). Zombie NIR Fixable Viability kit was obtained from BioLegend (Cat # 423106).

2.12. Statistical Analysis

All statistical analyses were performed with GraphPad Prism software. One-way ANOVA was performed when comparing three or more groups, with Tukey’s post-hoc multiple comparisons tests following any results with a p-value <0.05. Significance is indicated by **** p-value <0.0001, *** p <0.001, ** p <0.01, and * p <0.05. All graphs show means with standard deviations.

3. Results and Discussion

3.1. Fabrication of Small, Medium, and Large MAP Populations

Distinct MAP populations were produced according to a published water-in-oil step emulsification technique.^19,22 A hydrogel precursor solution, which generated a stiffness of 51 ± 3.8 kPa, was flowed through microfluidic devices possessing three different channel heights, resulting in three distinct populations (Fig. 1A–C). To minimize polydispersity, each population was sieved through a cell strainer before and after purification (and resultant swelling in PBS). Eight batches of each gel were produced, resulting in mean diameters of 44, 77, and 100 μm for Small, Medium, and Large groups, respectively (Fig. 1F). All polydispersity indices were less than 1.05. Batches were pooled for subsequent testing and use.

Figure 1 –

Synthesis and characterization of Small, Medium, and Large MAP populations and scaffolds. A) Schematic of microfluidic step emulsification process. B) Representative images of microgels fabricated using devices of varying channel heights. Scale bars are 200 μm. C) Representative images of swollen particles with visibly distinct differences in diameter. Scale bars are 100 μm. D) Representative 3D reconstructions of confocal z-stacks created using Imaris software. Scale bars are 100 μm. E) Quantification of particle diameters in the oil phase using ImageJ. F) Swollen particle diameter quantification in PBS using fluorescent microscopy. G) Average pore size and H) total scaffold porosity analyses were performed using Imaris. All graphs show means +/− standard deviations. One-way ANOVAs followed by post-hoc multiple comparison tests (Tukey’s HSD) were used to determine significance. **** p-value <0.0001.

3.2. Characterization of MAP Scaffold Pore Architecture

After synthesizing distinct particle size populations, the corresponding MAP scaffold pore architectures were characterized. Scaffolds formed with each set of microgels were incubated in a PBS-based solution of fluorescently labeled dextran of high molecular weight (2,000 kDa). The vast size of the dextran prevents it from entering the nanoscale pores of the hydrogel particles, but the scaffold’s micron-scale interconnected porosity allows for free diffusion of the solution. Soaked scaffolds were imaged using a Zeiss multiphoton confocal, and z-stack images of each scaffold were reconstructed with IMARIS image visualization and analysis software (Fig. 1D). Small, Medium, and Large particle scaffolds possessed mean pore sizes of 3107, 4637, and 7547 μm³, respectively (Fig. 1G). The total scaffold porosity was calculated by dividing the sum of all pores in a scaffold by the total z-stack image volume (Fig. 1H). The total scaffold porosity (approximately 30%) was consistent across scaffolds. This maintenance of total porosity across a range of pores is important for comparison (i.e., lack of consistency would imply variance in the particle fraction, which has been shown to affect bulk scaffold stiffness, macromolecule diffusion, and cell growth²³). This uniformity is difficult to replicate in non-spherical microgel approaches, as seen in the change in total porosity for techniques that use varying sizes of fractionated microgels²⁴ or dilutions of rod-shaped microgels²⁵.

3.3. MSC Survival

For our studies, cell survivability was focused on the capacity of cells to survive the seeding process. Cells were seeded either on top of or intermixed within scaffolds. MSC survivability was tested by applying a viability stain 24 hours after seeding. When cells were seeded on top of scaffolds, all groups showed at least 96% viability (Fig. 2A). When MSCs were mixed into scaffolds and exposed to the centrifugation and annealing steps, cell survivability dropped to 85%, 90%, and 93% for Small, Medium, and Large scaffolds, respectively (Fig. 2B). As there was no statistically significant drop in viability between the large particle groups regardless of cell seeding technique, we hypothesize that the drop in viability is primarily a result of mechanical damage sustained during cell pellet resuspension. Since MSC pellets were resuspended using the same volume of each gel, the Small group inherently possessed more particles, a greater surface area, and an increased frequency of collisions for the cells during mixing²⁶.

Figure 2 –

Analyses of the impact of MAP scaffold particle size on MSC survival and growth reported as fold changes from Day 0. Survivability of MSCs seeded A) on top of MAP scaffolds or B) within MAP scaffolds was determined by Live/Dead staining. Quantification of DNA content over two weeks for cells seeded C) on top of or D) within scaffolds. E) MSCs showed no difference in proliferation at Day 14 when seeded on top of scaffolds. F) Cells proliferated the most by Day 14 when seeded within Large particle scaffolds. Quantification of metabolic activity over time using a resazurin-based assay for MSCs seeded G) on top of or H) within scaffolds. I) Differences between groups were observed at Day 14 when cells were seeded on top of scaffolds. J) MSCs seeded within Large particle scaffolds displayed the greatest fold change in metabolic activity at Day 14. All studies were performed in duplicate with n≥3. All graphs display means +/− standard deviations. One-way ANOVAs were used to determine significant differences, followed by post-hoc Tukey’s multiple comparisons tests when applicable. **** p-value <0.0001, *** p <0.001, ** p <0.01, * p <0.05.

3.4. MSC Growth

MSC proliferation and metabolism were evaluated using the same seeding processes and densities as the survivability studies. The combination of DNA content and metabolic activity should provide a deeper understanding of overall cell growth. For cells seeded on top of scaffolds, there was no significant difference or trend among Small, Medium, and Large groups at Day 7 or 14, with all three groups experiencing a fold change of about 4.3 from Day 0 (Fig. 2E). Regarding metabolism, the Large group (fold change of 3.6) showed a greater increase in metabolic activity than the Small (2.9) and Medium (3.2) groups at Day 14 (Fig. 2F). Although the acellular controls displayed no evidence of pore-related effects on assay results (Fig. S1), the discrepancies between our proliferative and metabolic measurements may be due to cells clogging the smaller pores, thus slowing diffusion and/or metabolic activity without affecting proliferation.

The pore architecture played a more prominent role when cells were seeded within the scaffolds. Small particle scaffolds reported the slowest rate of proliferation with a fold change at Day 14 of 2.7 compared to Medium (3.2) and Large (4.4) scaffolds (Fig. 2I). The analysis of metabolism corroborated this trend, showing significant increases in metabolic activity as particle size increased (Fig. 2J). Although the subpar performance of the Small particles may be due to the impairment in survivability (e.g., induction of a lag phase of growth), the Large group also significantly outperformed the Medium group, demonstrating a greater increase from both Day 0 to Day 7 and Day 7 to Day 14 for both metrics of growth.

3.5. MSC Migration

Cell migration was investigated via a modified published technique²¹. In brief, we first create a dense spheroid of fluorescently labeled MSCs using a hanging drop approach before placing the spheroids on top of annealed scaffolds. A z-stack maximum intensity projection is acquired at 0 and 24 hours (Fig. 3A). Placing the spheroid at the surface is analogous to the interactions when placed in contact with tissue (e.g., in situ implantation). Calculating the average fold change from Day 0 established a trend of migration decreasing with increasing pore size (Fig. 3B). Importantly, MSC migration between the particles (i.e., not as a monolayer) can be observed in high resolution images (Fig. S2). Migration in the x-y plane for MAP scaffolds composed of Small microgels (fold change of 2.4) was significantly farther than migration in the Large group (fold change of 2.0). An explanation of this could be that the pores of the Small scaffolds get crowded faster and therefore drive migration. Migration in the z-plane (see Fig. S3) showed no significant difference between groups. Notably, this migration was much lower than in the x-y plane. For example, for the Small group z migration was ~140μm, while x-y migration was ~340m or ~2.5x the z migration (note: the initial diameter of the spheroids is 339.51 μm +/− 24.3 um, thus a fold change of 2 equates to a 24 hour diameter of 679 um, or a migration of ~340μm). A limitation of our approach is the lack of an exogenous chemoattractant to direct migration into the scaffold and this resulted in preferential migration in the x-y plane. Our lab has previously demonstrated the creation of chemotactic gradients within MAP scaffolds²⁰, which could be utilized to better mimic chemoattractant-driven migration of MSCs through scaffolds of different particle sizes. Alternatively, future studies placing a spheroid within the scaffold (i.e., increasing the influence of pore size) would be an observation of interest. Nevertheless, this study describes the preference of cells to migrate differentially in contact with scaffolds of varying pore size, which is necessary for complete and uniform tissue regeneration.²⁷

Figure 3 –

MSC migration through MAP scaffolds of Small, Medium, and Large particles. A) Representative images of cell migration at 0 and 24 hours. All scale bars represent 164 μm. B) Fold change analysis of radial cell migration distance at 24 hours via fluorescent microscopy. Study was performed in duplicate with n≥3. Graph displays means +/− standard deviations. One-way ANOVA followed by post-hoc Tukey’s multiple comparisons test was used to determine significance. ** p-value <0.01.

3.6. MSC Differentiation

Flow cytometry was used to evaluate MSCs for potential spontaneous differentiation following 14 days in culture. A flow cytometry panel was constructed from the manufacturer’s quality control report as well as literature^28,29. MSCs should stain positive for CD73, CD90, and CD105, but negative for CD45. Liberase was significantly more efficient in degrading our MAP scaffolds and was used for their isolation. Due to a reported effect that the enzyme used for degradation of our hydrogels, Liberase TM, can result in loss of CD105²⁹, a Liberase TM-harvested 2D control was included in addition to a standard Trypsin-EDTA-harvested 2D control. However, no difference was observed between the controls (Fig. S4). This may be because the enzyme was deployed at a much lower concentration than that at which deleterious effects were recorded. The gating strategy is provided in greater detail in Fig. S4.

Following cell isolation from MAP scaffolds via Liberase and cytometry analysis (Fig. 4A–C), it was observed that the percentage of cells expressing the phenotypic profile CD73+CD90+CD105+CD45- was not significantly different between scaffolds made from different particle sizes (Fig. 4D). At least 98% of the MSCs retained this profile across groups, indicating that seeding MSCs within MAP scaffolds does not lead to their spontaneous differentiation within 14 days. Future studies confirming tri-lineage differentiation ability will be beneficial to confirm these results²⁸.

Figure 4 –

Maintenance of stem cell surface marker expression via flow cytometry. A) Analysis of Trypsin EDTA-harvested 2D-control MSCs. i) Cells were first gated away from debris, then ii) gated based on viability. iii) Single cells were identified to remove clumps from analysis. Using the singlet population, iv) cells were screened in series to be CD73+, v) CD45-, vi) CD105+, and vii) CD90+. B) Liberase TM-harvested 2D-control MSCs showed no loss of stemness surface markers. C) In each sample, the number of cells that fit the entire expression profile was divided by the number of singlet cells to determine the percentage of the population that avoided spontaneous differentiation. Graph displays means +/− standard deviations of n≥3. Representative gating analyses are shown for scaffolds with D) Small, E) Medium, and F) Large particles.

4. Conclusions

In conclusion, we demonstrated the synthesis and characterization of MAP scaffolds of varying microgel diameters and investigated the impact of microparticle size on MSC function. These functional studies represent the first comparisons of three distinct particle sizes using the MAP platform, which should help guide future applications of MAP. This is also the first demonstration of stem cell surface marker maintenance using MAP scaffolds, which should increase confidence in the platform for stem cell-based treatments.

Recently, MSC proliferation, metabolism, and spreading (i.e., migration) were compared in MAP scaffolds of two different particle sizes.³⁰ While this work will also undoubtedly help guide future MAP applications, cells were seeded as spheroids for all experiments, which we argue is not as relevant for clinical translation compared to the monodispersed cells used herein.

Our results indicate that MAP microgel diameter can be tuned to modulate MSC survivability, proliferation, metabolism, and migration, thereby enhancing bioactivity and increasing the feasibility of these scaffolds for regenerative medicine. By seeding the cells within MAP scaffolds, significant differences were observed for survivability, proliferation, and metabolism assays, with the Large microparticle hydrogels displaying favorable results in all cases. A potential mechanism that these findings suggest is that, when working with low cell densities, the effects of pore size-based diffusion of oxygen and nutrients supersede the raw potential for cell growth provided by increased surface area. Because both MSC paracrine signaling^31,32 and contact inhibition^33,34 impact cell survival, proliferation, and migration, the results obtained at low seeding densities may not be the same at high seeding densities. Future studies could confirm this by exploring higher cell densities that are based on the theoretical surface area of each group rather than gel volume.