Chondrogenesis of Adipose-Derived Stem Cells Using an Arrayed Spheroid Format

Abstract

Objective

The chondrogenic response of adipose-derived stem cells (ASCs) is often assessed using 3D micromass protocols that use upwards of hundreds of thousands of cells. Scaling these systems up for high-throughput testing is technically challenging and wasteful given the necessary cell numbers and reagent volumes. However, adopting microscale spheroid cultures for this purpose shows promise. Spheroid systems work with only thousands of cells and microliters of medium.

Methods

Molded agarose microwells were fabricated using 2% w/v molten agarose and then equilibrated in medium prior to introducing cells. ASCs were seeded at 50, 500, 5k cells/microwell; 5k, 50k, cells/well plate; and 50k and 250k cells/15 mL centrifuge tube to compare chondrogenic responses across spheroid and micromass sizes. Cells were cultured in control or chondrogenic induction media. ASCs coalesced into spheroids/pellets and were cultured at 37 °C and 5% CO₂ for 21 days with media changes every other day.

Results

All culture conditions supported growth of ASCs and formation of viable cell spheroids/micromasses. More robust growth was observed in chondrogenic conditions. Sulfated glycosaminoglycans and collagen II, molecules characteristics of chondrogenesis, were prevalent in both 5000-cell spheroids and 250,000-cell micromasses. Deposition of collagen I, characteristic of fibrocartilage, was more prevalent in the large micromasses than small spheroids.

Conclusions

Chondrogenic differentiation was consistently induced using high-throughput spheroid formats, particularly when seeding at cell densities of 5000 cells/spheroid. This opens possibilities for highly arrayed experiments investigating tissue repair and remodeling during or after exposure to drugs, toxins, or other chemicals.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12195-022-00746-8.

Introduction

Three-dimensional (3D) environments that induce a rounded cell morphology, such as hydrogels and micromasses (also known as pellet or aggregate cultures), are critical for directing the chondrogenic differentiation of mesenchymal stem cells (MSCs).^28,40,45 These 3D structures are more physiologically relevant than monolayer cultures as they offer spatial and physical characteristics advantageous for a number of cellular processes, such as proliferation, differentiation, protein and gene expression, and morphology.^24,44 Micromass and spheroid cultures are rapidly gaining interest as a tool for developmental biology, tissue engineering, and regenerative medicine. Cell micromasses have often been used to study chondrogenesis, initially with larger centrifuge tubes (traditional),^6,23,29,42 then moving to 96-well microplates,^38,50,57 and more recently to non-adhesive microwells.^1,5,18,32,56 Cell spheroids/micromasses increase cell-to-cell adhesions,^8,12 and cell–matrix connections^35,39,49,54 making it a supportive environment for cartilage regeneration.^7,19

Recent studies have used spheroid/micromass systems in a variety of ways to investigate cartilage disease and regeneration. These include examinations of growth factor effects,^{13,25,26,34,47} changes in microenvironmental forces during mesenchymal condensation,²² and the influence of compressive forces on chondrogenic differentiation.^16,41 Other studies looked at how cell-to-cell contacts via tight adhesions within these systems served as a precursor to hypoxic conditions for cartilage development.^31,43 Beyond direct study of spheroid systems, many groups have looked towards using these nascent tissues as building blocks to form larger, engineered constructs for implantation.^11,27,53 For the majority of these studies, spheroids/micromasses were created using large cell numbers, ranging in the tens of thousands to hundreds of thousands of cells.^17,42 Past studies have shown that because of their large size, these micromasses can develop necrotic cores due to oxygen diffusion limits, complicating their use as a model system of healthy tissues.³⁶ Furthermore, micromasses that require hundreds of thousands of cells for each sample can be methodologically problematic for highly arrayed experimental designs (e.g., drug/toxin screening) or cases where the cells themselves are a limited resource (e.g., freshly isolated, primary cells). Only a few studies have looked at the lower end of cell numbers and how this relates to chondrogenesis,^30,46 which is the focus of the current work.

The goal of this study was to determine whether a small form factor, limited cell number experimental approach was compatible with MSC chondrogenic induction. We hypothesized that small numbers of human adipose-derived stem/stromal cells (ASCs) grown in non-adherent, agarose microwells to form spheroids would proliferate and differentiate at similar proportional levels to more conventional, large cell number micromasses. To test this hypothesis, we differentiated ASCs in spheroid culture environments compatible with high-throughput testing alongside conventional ASC micromass cultures. Over a 3-week induction period, we examined morphology, proliferation, viability, and the production of chondrogenic molecules [sulfated glycosaminoglycans (sGAGs) and collagen II] by spheroids/micromasses seeded initially with 50 to 250,000 cells per sample.

Materials and Methods

Human Adipose-Derive Stem Cells

All procedures were approved by the internal review board (IRB) at Rhode Island Hospital. Human lipoaspirate was obtained from a healthy female donor (D96), and stromal vascular fraction (SVF) cells were isolated using established protocols.¹⁴ Samples were maintained in humidified incubators at 37 °C, 5% CO₂ and passaged at 80% confluence with 0.25% trypsin–EDTA (HyClone, GE Healthcare Life Sciences, Logan, UT). Following isolation and expansion to passage 3, the enriched cells (now termed ASCs) were stored in liquid nitrogen using freezing medium containing 80% fetal bovine serum (FBS) (ZenBio, Durham, NC), 10% dimethyl sulfoxide (Thermo Fisher Scientific, Waltham, MA), and 10% expansion medium consisting of DMEM/F-12 (HyClone, GE Healthcare Life Sciences, Logan, UT), 10% FBS, 1% antibiotic/antimycotic (A/A) (HyClone, GE Healthcare Life Sciences, Logan, UT), 0.25 ng/mL transforming growth factor-β1 (TGF-β1), 5 ng/mL epidermal growth factor, and 1 ng/mL fibroblast growth factor (R&D Systems, Minneapolis, MN).¹⁵ Prior to experimentation, ASCs were quickly thawed and cultured in expansion medium for one passage (to P4) to reacclimate the cells to a growth environment.

Agarose Gel Microwell Fabrication

Non-adherent microwells were fabricated from 2% molten agarose (Thermo Fisher Scientific, Waltham, MA) using 3D Petri Dish® molds (35 Large, Microtissues Inc., Providence, RI).¹ Microwells were cured at 4 °C for 15 min, transferred to 24-well plates, and equilibrated in control medium for 4 days prior to introducing ASCs.

Cell Culture

Five different cell seeding densities, across three different growth environments, were used to examine spheroid/micromass chondrogenesis. For spheroid conditions, ASCs were seeded at a density of 50, 500, and 5000 cells/microwell in agarose micromolds (35 spheroids/mold). For well plate micromass conditions, ASCs were seeded at a density of 5000 and 50,000 cells/well in non-culture-treated, V-bottomed, 96-well plates. For tube-based micromass conditions, ASCs were seeded at a density of 50,000 and 250,000 cells/tube in 15 mL centrifuge tubes. Spheroids were allowed to self-assemble, while well plates and tubes were centrifuged at 400 g for five minutes to accelerate the formation of initial cell micromasses following previously established protocols.^20,21 Expansion medium was replaced with chondrogenic induction or control media. Chondrogenic medium contained DMEM-HG (HyClone, GE Healthcare Life Sciences, Logan, UT), 10% FBS, 10 ng/mL TGF-β1 (R&D Systems, Minneapolis, MN), 50 µg/mL ascorbate-2-phosphate (Sigma-Aldrich, St. Louis, MO), 39 ng/mL dexamethasone (Sigma-Aldrich, St. Louis, MO), 1% ITS + Premix (BD Biosciences, San Jose, CA), and 1% A/A. Control medium contained DMEM/F-12, 10% FBS, and 1% A/A. Agarose spheroid molds and 15 mL tubes received 1 mL media each, while V-bottomed, 96-well plates received 200 μL. Cell pellets were maintained in culture for up to three weeks with three medium changes per week. Samples were imaged at × 4 and × 20 magnification using phase contrast microscopy at Day 1, 7, 14, and 21 after seeding. ImageJ (NIH) was used to measure the size and morphology of spheroids/micromasses.

Quantification of sGAG, Collagen, DNA

After 1, 7, 14 and 21 days, samples were fixed in 3.7% paraformaldehyde (Thermo Fisher Scientific, Waltham, MA). For analysis, half of the samples (chondrogenic (n = 3) and control (n = 3)) were digested with papain (Sigma-Aldrich, St. Louis, MO) at 65 °C for 48 h. For microwell and 15 mL tube conditions, each sample was an individual micromass. For agarose micromold conditions, each sample included ~ 35 spheroids from an array, with data reported as the calculated average value for a single spheroid. A microplate reader (Cytation™ 3 Cell Imaging Multi-Mode Reader, Biotek, Agilent, Santa Clara, CA) was used to quantify sGAG, total collagen, and DNA content from all the papain-digested pellets. sGAG content was quantified via a dimethylmethylene blue (DMMB) assay with absorbance measured at 525 nm using established protocols.^4,10 Total collagen content was quantified by using a hydroxyproline assay with absorbance measured at 550 nm following established protocols.³ DNA content was quantified using the Quant-iT Picogreen dsDNA Assay Kit (Invitrogen, Thermo Fisher Scientific, Waltham, MA) with fluorescence measured at 480ex/520em. A standard curve was used to calculate total sGAG, collagen, and DNA amounts per spheroid/micromass.

Live/Dead Viability Assay

A live/dead viability/cytotoxicity kit (catalogue no. L3224, ThermoFisher Scientific) was used to examine general cell viability in spheroids. ASC spheroids were cultured in control medium at 5000 cells/spheroid for three days prior to assessment. In brief, calcein AM (live cells, 1:2000 dilution) and ethidium homodimer-1 (dead cells, 1:500 dilution) were mixed in culture medium, added to wells containing spheroids, and incubated for 60 min at 37 °C in 5% CO₂. After incubation, wells were gently washed three times and re-filled with fresh culture medium. Spheroids were immediately imaged at 488ex/515em for calcein AM and 561ex/617em for ethidium homodimer-1.

Histology and Immunofluorescent Staining

Representative 21-day samples for all spheroid and 250,000 cell micromass groups (n = 1) were fixed in 10% formalin and stored at 4 °C. For spheroids in micromolds, an agarose cap was added to keep them in place before the entire assembly was embedded in paraffin. Micromass samples were directly encapsulated in warm agarose and then embedded in paraffin. For histochemistry, all samples were sectioned (7-μm thick), deparaffinized with xylene, and rehydrated in gradations of deionized water for 5 min each. After staining, sections were dehydrated in gradated ethanol, cleared with xylene, and dried before being mounted with Eukitt mounting medium (Sigma-Aldrich, St. Louis, MO, USA) with coverslips. sGAG deposition was visualized with Safranin O (Sigma-Aldrich, St. Louis, MO, USA). Immunofluorescence staining was done using rabbit, anti-human collagen II (Col2A1, primary, 1:100, MyBioSource, #MBS2111929), anti-human collagen I (Col1A1, primary, 1:100, MyBioSource, #MBS829687) and chicken, anti-rabbit secondary (1:500, Alexa 647 fluorophore, Invitrogen, #903812) antibodies. In general, samples were deparaffinized (xylene and ethanol), rehydrated (deionized water), digested (pepsin), blocked (5% bovine serum albumin), and then stained with primary and secondary antibodies, each with a 1 h incubation period. 4′,6-diamidino-2-phenylindole (DAPI) diluted at 1:1,000 was used for nuclear counterstaining. All histological sections were thoroughly screened, with representative images selected for presentation.

Statistical Analysis

All experimental data were conducted with at least three biological replicates across three iterations. Analysis of variance (ANOVA), Brown-Forsythe, Student’s t-Test, and Shapiro–Wilk were performed using SigmaPlot 12.5 (Systat Software, Inc). Statistical significance among groups was defined as p < 0.05. Results presented as mean ± standard deviation.

Results

Morphology of Spheroids

ASCs were grown in three different culture vessels, with corresponding cell seeding densities, and monitored across four different time points post-plating (day 1, 7, 14, and 21) (Fig. 1). ASCs aggregated into cohesive units within their respective vessels. After a day, spheroids and micromasses coalesced into balls in both chondrogenic and control conditions (Figs. 1a, 1b). ASCs in agarose microwells consistently formed single spheroids regardless of medium condition. Comparatively ASCs in V-bottomed well plates and 15 mL tubes formed a main micromass with smaller, satellite aggregates on adjacent culture surfaces. For all conditions, chondrogenic induction increased spheroid/micromass size over time, reaching a peak at day 7 followed by gradual compaction over the remaining two weeks. Samples in control medium exhibited their largest size on day 1 followed by extensive compaction through the remainder of the study (Fig. 1c).

Figure 1.

Morphological analysis of adipose-derived stem/stromal cell spheroids/micromasses with different cell seeding numbers on days 1, 7, 14, and 21. Representative brightfield images of samples in (a) chondrogenic and (b) control media. Larger micromass conditions (well plate and 15 mL tube) promoted formation of satellite aggregates. It should be noted that the 15 mL tube samples were removed from their original culture vessels for imaging purposes, so most satellite aggregates are absent. (c) Normalized size changes in spheroids/micromasses over time. Data are presented as mean ± SD. *p < 0.05. All images taken at × 4; Scale bar: 500 µm.

Cell Proliferation and Extracellular Matrix Production

DNA abundance for all samples was tracked over time as an indicator of cell proliferation and used to normalize matrix production by cell number across samples (Fig. 2a). Non-normalized data are also provided since normalization can obscure some trends due to the multiple orders of magnitude difference among sample conditions. In general, DNA abundance increased over three weeks of culture for samples grown in agarose microwells or well plates when exposed to chondrogenic medium but decreased for samples cultured in 15 mL tubes (p < 0.001). When exposed to control medium, all conditions showed either no change or a decrease in DNA (p < 0.001). Since cell proliferation can sometimes mask the development of a necrotic core in micromass cultures,⁹ we examined cellular viability within the largest spheroid condition (5000 cells/spheroid). Results showed that viability was high throughout the spheroid with no increase in cell death moving from the outside edge to the center (Fig. S-1).

Figure 2.

ASC spheroid/micromass content for (a) DNA, (b) sGAG, and (c) collagen for samples cultured in agarose microwells, well plates, and 15 mL tubes for 21 days. Initial cell seeding numbers are designated for each group. Chondrogenic medium samples are shown on the left side, and control medium samples on the right side. sGAG and total collagen content normalized either by spheroid/micromass (top) or DNA (bottom) within each panel. Data are presented as mean ± SD (n = 3). Samples were compared within their group respectively; bars with different letters are significantly different (p < 0.05). If no letters are shown over a group, no differences were detected.

Secretion of chondrogenic molecules by ASC spheroids/micromasses was measured weekly over a 21-day period (Figs. 2b and 2c). Substantial variability was present across the time points, but some consistent patterns were apparent. sGAG and total collagen amounts were generally undetectable at the per sample level for all spheroid conditions. When normalized by DNA amount, spheroid and micromass values were largely similar, with no statistical significance being observed regardless of culture format (p > 0.05). The only exception to this was the 250,000-cell micromass group, which showed higher per-cell production of collagen than other groups. Overall, spheroids/micromasses with greater cell numbers produced more total extracellular matrix. For sGAG production in chondrogenic medium, micromass samples showed a substantial increase by Day 7 followed by no change, or a tapering off, through the remainder of the experiment (Fig. 2b, p < 0.001). This pattern was the same, but delayed slightly, in the 5000-cell spheroid sample. Control medium samples showed little change in sGAG over time. For collagen production in chondrogenic medium, all samples with detectable levels showed an increase over time (Fig. 2c, p < 0.001). This was especially dramatic for the 250,000-cell micromass sample, even when normalized by DNA abundance. As with sGAG, the 5000-cell spheroid samples showed a similar trajectory in collagen levels as the larger micromass samples, increasing steadily over time commensurate with the number of cells present. Control medium samples showed virtually no collagen production.

Histology and Immunofluorescence Staining

After 21 days in culture, all spheroid samples and the gold standard 250,000-cell micromass sample were analyzed by histochemistry for production of sGAG using Safranin O and immunofluorescence staining for production of collagen II (Fig. 3). As opposed to the biochemical assays, smaller spheroid samples could be reliably analyzed for production of matrix molecules through this approach. Safranin O staining showed the presence of sGAG in both control and chondrogenic conditions for 5000- and 250,000-cell seeding densities. Robust, collagen II staining was observed in chondrogenic conditions for 5000- and 250,000-cell samples but not in their matched, control medium samples. Production of collagen I was minimal in 5000-cell spheroids compared to 250,000-cell micromasses where collagen I was more abundant (Fig. 4). This collagen type prevalence reflects more hyaline-like versus fibrocartilage-like tissues, respectively. For the 50- and 500-cell spheroids, sGAG was only visible in the latter for chondrogenic conditions (Fig. S-2). In general, the extremely small cell numbers and physical sizes of these samples made it difficult to measure or visualize sGAG, collagen I, or collagen II rigorously, so their chondrogenic response in this capacity remains inconclusive. Semi-quantitative analysis of the cellular density within spheroid/micromass samples indicated that 5000-cell spheroids in either medium were more compact (40–50% area occupied by nuclei) than 250,000-cell micromasses (~ 25% nuclear area).

Figure 3.

Histological and immunofluorescence staining of ASCs cultured as 5000-cell spheroids or 250,000-cell micromasses in chondrogenic and control medium conditions. Images show matrix accumulation after 21 days in culture. Non-matched (a) Safranin O-stained sGAG and matched (b) DAPI-stained nuclei, (c) fluorescent antibody detection of collagen II, and (d) merged images. Collagen II was abundant in both chondrogenic conditions regardless of culture format. All images taken at × 20; Scale bar: 100 µm.

Figure 4.

Immunofluorescence staining of 5000-cell spheroids and 250,000-cell micromasses in chondrogenic and control medium conditions. Matched images show collagen I accumulation after 21 days in culture. (a) Brightfield, (b) DAPI-stained nuclei, (c) collagen I, and (d) merged images Note the lack of collagen encapsulation at the surface of the spheroid sample compared to micromass. All images taken at × 10; Scale bar: 200 µm.

Discussion

The chondrogenic response of MSCs is often assessed using 3D micromass protocols that use upwards of hundreds of thousands of cells. Traditionally, 15 mL centrifuge tubes are used for these experiments, with the format being limited to one micromass per tube.³⁷ Scaling these systems up for high-throughput testing using 96-well microplates can be technically challenging and remains both expensive and wasteful given the necessary cell numbers and reagent volumes to conduct highly arrayed experimental designs.^38,50,57 Adopting microscale spheroid systems is promising as it can reduce both the number of cells and total liquid volumes.⁴⁸ The objective of the current study was to investigate whether conventional micromass chondrogenic outcomes can be replicated using orders of magnitude fewer cells in a spheroid format. Low cell numbers is a key feature of spheroid system platforms, potentially allowing for high-throughput testing of chondrogenic stimulation and regeneration.²⁶ Results from the current study showed that a 5000-cell spheroid format exhibited similar chondrogenic characteristics as the gold-standard, 250,000-cell micromass culture upon chemical induction.

We tested the effects of non-adherent, 3D culture on ASCs using an agarose microwell environment that contained 35 individual spheroids formed from an average of 50, 500, or 5000 cells/microwell. We compared these samples to more traditional micromasses containing 5000, 50,000, or 250,000 cells/sample in well plate and 15 mL tube formats. The general morphology of samples was tracked over time via optical microscopy. ASCs formed single spheroids when cultured in microwells, while those in well plates and 15 mL tubes formed central micromasses with many satellite aggregates on their periphery. The differences in cellular organization were not purely due to cell numbers since 5000-cell samples exhibited satellites in well plates but not in microwells. Most likely, the non-adherent characteristics and small volume of the agarose microwells encouraged more cell–cell interaction and adhesion.³³ This results in efficient formation of spheroids with few, outlying, non-incorporated cells. Comparatively, the well plate and 15 mL tube culture vessels, resulted in many satellite aggregates, which have previously been noted to form as cells bud out from the initial, primary micromass.² Not having all ASCs come together to form one unit can adversely affect matrix deposition and skew biochemical results since the cells experience more varied microenvironments within the same vessel.^48,51 Regardless of the presence of satellites, all samples exhibited similar changes in primary spheroid/micromass size through the course of the experiment. Following initial self-assembly, spheroids/micromasses in control medium continually compacted over time whereas those in chondrogenic medium could maintain a larger size through a combination of cell proliferation and matrix production. A notable difference between spheroid samples and larger micromasses was that the former had a higher cellular density at 21 days, potentially indicating that more energy is being directed into proliferation over matrix synthesis for these samples. Morphological results were generally consistent across all spheroid sample groups. This suggests that ASC aggregation and spheroid growth proceeds normally even with initial seeding densities down to 50 cells.

Beyond morphology, we also quantified DNA, sGAG, and total collagen amounts for the 21-day culturing period. DNA abundance served as a proxy for measuring ASC proliferation and also served to normalize matrix production across all samples. Cell numbers for almost all groups in chondrogenic medium stayed relatively constant over the course of the experiment. The largest micromass condition was the only group with a statistically significant decrease in cell numbers. This might be due to fracturing of the primary micromass into smaller satellites, or alternatively, more cell death in response to limited oxygen and nutrients in the culture vessel or micromass itself. Tissues greater than ~ 400 µm in size have been reported to develop necrotic cores due to oxygen and nutrient diffusion limitations.^30,55 Because an equivalent, total number of cells in the microwells are split among 35 spheroids, the surface area-to-volume ratio is much greater than a single, large micromass, and distances from the spheroid core to its surface are much less than oxygen diffusion limits. This is a positive feature of the spheroid system, resulting in healthy spheroids with very few dead/dying cells and no necrotic cores (Fig. S-1).

Comparisons of matrix production, specifically sGAG and collagen, indicated similarities between the 5000-cell spheroid and the 250,000-cell micromass samples. This was most apparent when analyzing histological characteristics, which was a more reliable means of assessment than conventional biochemical assays. ASCs in chondrogenic medium robustly secreted collagen II in both the spheroid and micromass culture formats, confirming that chondrogenesis was successful. Interestingly, in the spheroid samples, collagen II staining is present only in the 3D spheroid and not in the peripheral cells coating the bottom/sides of the microwell. By comparison, collagen I production was sparse in the 3D spheroid but present in the peripheral cells. The large micromass also produced collagen I, although not to the same extent as collagen II, and the protein was more predominant at its surface. This fibrous encapsulation is a common feature in large engineered tissues, and the lack of its presence in small spheroids may help with their integration in tissue engineering efforts using them as building blocks.^27,53 Overall, collagen type comparisons between spheroid and micromass systems showed a more hyaline-like phenotype in the spheroid systems. Compared to collagen, production of sGAGs was much greater in micromass samples than spheroids, although on a per-DNA basis there was no difference across culture format. Histology indicated sGAGs were deposited in both cases, but quantitative comparisons are challenging to make since the conventional DMMB biochemical assay had a relatively high lower limit of detection. As expected, the total amount of sGAGs produced by 250,000-cell micromasses far exceeded that produced by any of the smaller cell seeding densities. Interestingly, we found that combining 35 spheroids into a single sample for the DMMB assay still resulted in very small sGAG levels. This suggests that sGAG accumulation may benefit from having a lower surface-to-volume ratio, perhaps because molecules at the edge of a spheroid/micromass/engineered construct are often lost to the surrounding medium. Future work can investigate the amount of sGAG released to the surrounding medium vs. retained in the spheroids to test this hypothesis. Another related possibility for this discrepancy is a disproportional production of sGAGs by more rounded cells in the core/interior of a sample compared to flattened cells in the shell/exterior. Large micromasses would have a higher proportion of the former over the latter.

These results suggest culturing ASCs within a non-adherent microwell at 5000 cells/spheroid appropriately fosters key, chondrogenic characteristics present in conventional culturing conditions. Furthermore, spheroids have an advantage in promoting more uniform cell viability, maintaining of a single condensate of cells, and successfully inducing chondrogenic differentiation of stem cells. Culturing within microwells makes feasible high-throughput studies aimed at screening greater quantities of treatments and sample replicates. This is relevant for more extensive investigations of how growth factors,²⁶ oxygen levels,³¹ cell interactions,⁵² and compressive forces/pressures^16,41 influence chondrogenic differentiation. Further reduction in cell numbers may even be feasible, particularly if the assay readouts are visual markers that can be detected through optical measurements of individual spheroids.

Supplementary Information

Below is the link to the electronic supplementary material.