Hypoxic Preconditioning Enhances Bone Marrow–Derived Mesenchymal Stem Cell Survival in a Low Oxygen and Nutrient-Limited 3D Microenvironment

Abstract

Objective

Skeletal tissues such as intervertebral disc and articular cartilage possess limited innate potential to regenerate, in part due to their avascularity and low cell density. Despite recent advances in mesenchymal stem cell (MSC)–based disc and cartilage regeneration, key challenges remain, including the sensitivity of these cells to in vivo microenvironmental stress such as low oxygen and limited nutrition. The objective of this study was to investigate whether preconditioning with hypoxia and/or transforming growth factor–β 3 (TGF-β3) can enhance MSC survival and extracellular matrix production in a low oxygen and nutrient-limited microenvironment.

Design

MSCs from multiple bovine donors were preconditioned in monolayer in normoxia or hypoxia, with or without TGF-β3, and the global effects on gene expression were examined using microarrays. Subsequently, the effects of preconditioning on MSC survival and extracellular matrix production were examined using low oxygen and nutrient-limited pellet culture experiments.

Results

Hypoxic preconditioning resulted in upregulation of genes associated with growth, cell-cell signaling, metabolism, and cell stress response pathways, and significantly enhanced MSC survival for all donors in low oxygen and nutrient-limited pellet culture. In contrast, TGF-β3 preconditioning diminished survival. The nature and magnitude of the effects of preconditioning with either hypoxia or TGF-β3 on glycosaminoglycan production were donor dependent.

Conclusions

These results strongly support the use of hypoxic preconditioning to improve postimplantation MSC survival in avascular tissues such as disc and cartilage.

Keywords

intervertebral disc articular cartilage therapy transforming growth factor–beta glycosaminoglycans

Introduction

Skeletal tissues such as intervertebral disc and articular cartilage possess limited innate potential to regenerate, in part due to their avascularity and low cell density.^1,2 Avascularity results in a local tissue microenvironment characterized by low oxygen tension and limited supply of nutrients. This presents challenges for cell survival and function, even in the absence of pathology.^3,4 Clinically, back and joint pain from disc and cartilage degeneration, respectively, with age or following trauma, are leading causes of disability in the adult population with associated socioeconomic costs estimated to be in the order of many billions of dollars annually.⁵ Current treatment approaches are mostly conservative and involve managing symptoms through physical therapy and administration of pain killers.^6,7 In severe cases, patients may undergo surgical procedures such as spinal fusion or joint replacement, which do not maintain or restore native tissue structure or mechanical function.⁸ In the case of cartilage damage, surgical repair using techniques such as microfracture or autologous chondrocyte transplantation are potential therapeutic options; however, these approaches often have limited efficacy.⁹

To address the limitations of current approaches for treating disc and cartilage damage and degeneration, there has been significant interest in developing biological therapies such as tissue engineered biomimetic scaffolds, targeted drug delivery for inducing endogenous cell activation, and exogenous cell implantation. With regard to the last of these approaches, application of mesenchymal stem cells (MSCs) for tissue repair and regeneration has attracted significant interest. MSCs, which can be derived from many sources including bone marrow, adipose tissue, or synovial fluid, are attractive options for cell-based therapies due to their potential for autologous harvest, expansion, and reimplantation.^10-12 Furthermore, MSCs may exhibit immunomodulatory properties and suppress inflammatory damage, making them highly translational with an excellent safety profile.^13,14 Importantly, MSCs have the potential to differentiate down multiple lineages,¹⁵ and there have been encouraging studies demonstrating the capacity of these cells to adopt phenotypes similar to those of nucleus pulposus cells^16-18 and articular cartilage chondrocytes^19,20 under certain in vitro culture conditions. Furthermore, there have been human and animal studies reporting that MSCs are able to regenerate and repair disc and cartilage tissue in vivo with varying degrees of efficacy.^11,21

Despite these recent advances in MSC-based disc and cartilage regeneration, key challenges remain, including phenotype instability (dedifferentiation or terminal differentiation and mineralization) in long-term culture and inferior functional extracellular matrix (ECM) production compared to native cell types.^22-24 Another key challenge is the sensitivity of MSCs to microenvironmental stress, such as nutrient and oxygen deprivation, to which cells are exposed upon implantation. While native cell types such as chondrocytes and nucleus pulposus cells are adapted to functioning under these conditions,^24-27 low oxygen and nutrient availability both negatively impact MSC survival and ECM production capacity.^28-30 It has also been shown that MSCs isolated from bone marrow are highly heterogeneous, both within and between donors, exhibiting distinct phenotypic characteristics, including sensitivity to microenvironmental stress in culture.^29,31,32

To overcome this challenge, one strategy being explored is the preconditioning of MSCs during monolayer expansion (by exposure to hypoxia or growth factors), in order to enhance their in vivo performance in applications for cartilage and disc regeneration.^33-44 The objectives of this study were to build on the results of these previous studies and to (1) investigate the effects of preconditioning through exposure to hypoxia or transforming growth factor–beta (TGF-β), alone or in combination on the global gene expression profile of MSCs and (2) investigate whether these molecular level changes in response to preconditioning can enhance MSC survival and ECM production capacity when subsequently placed under microenvironmental stress, specifically low oxygen and limited nutrition, in 3-dimensional (3D) culture. We hypothesized that preconditioning using hypoxia and TGF-β would enhance subsequent MSC survival and ECM production under microenvironmental stress by inducing requisite metabolic changes.

Methods

Cell Isolation, Expansion, and Preconditioning

The overall study design is presented in Figure 1 . Bone marrow–derived MSCs were isolated from juvenile (age <6 months) bovine femurs and tibiae (Research 87; Boylston, MA) as previously described.²⁹ MSCs from 3 individual donor animals were maintained as distinct populations (i.e., not pooled) during preconditioning and subsequent microarray analysis or pellet culture, which enabled the analysis of donor-dependent effects on MSC behavior and their response to preconditioning. The 3 MSC donor animals used for the microarray study were not the same as the 3 donor animals used for the pellet culture study (i.e., a total of 6 donor animals across both studies). Following the initial isolation, MSCs were expanded to confluence through a single initial passage in standard tissue culture conditions with normal oxygen tension (atmospheric, 21% O₂), 5% CO₂, in basal medium composed of high glucose (4.5 g/L) Dulbecco’s modified Eagle medium (DMEM; ThermoFisher Scientific; Waltham, MA), 10% fetal bovine serum (FBS; ThermoFisher Scientific; Waltham, MA), and 1% penicillin, streptomycin, and fungizone (PSF; ThermoFisher Scientific). Cells were then passaged and expanded (preconditioned) in basal medium in 1 of 4 different conditions for 1 week: (1) N−: normoxia (21% O₂); (2) N+: normoxia + TGF-β3 (10 ng/mL, R&D Systems; Minneapolis, MN); (3) H−: hypoxia (2% O₂); (4) H+: hypoxia + TGF-β3. Hypoxic culture was carried out continuously in an environmental workstation (Hypoxystation 35; Hypoxygen, MD, USA) that enabled media changes without exposure to atmospheric conditions.

Figure 1.

Overall study design. P = passage number.

Microarray Study

MSCs from 3 separate donors were isolated and subjected to preconditioning as described above. MSCs were harvested and RNA was isolated from each sample (n = 3 donors × 4 expansion conditions for a total of 12 distinct samples) using serial TRIzol (Ambion; Austin, TX)-chloroform extractions, in-column treated with RNase-free DNase (Qiagen; Valencia, CA) on miRNeasy columns (Qiagen; Valencia, CA), and eluted following the manufacturer’s protocols. The quality of each RNA sample (RIN > 9) was verified using an Agilent BioAnalyzer and RNA 6000 Pico Kit (Agilent; Santa Clara, CA). Global gene expression was measured using the WTPlus Bovine Gene Chip (Affymetrix; Santa Clara, CA). Gene expression data were normalized using Robust Multi-array Average.⁴⁵ Pathway analysis was performed using Ingenuity Pathway Analysis (IPA; Qiagen; Valencia, CA, USA).

Pellet Culture Study

MSCs from 3 separate donors were isolated and subjected to preconditioning as described above. MSCs were then harvested and cultured as pellets (250,000 cells/pellet) in a low oxygen and limited nutrient environment consisting of 2% O₂ and serum-free, chemically defined media²⁶ comprised of low glucose (1g/L) DMEM (ThermoFisher Scientific) supplemented with 1% PSF, 0.1 μM dexamethasone (Sigma-Aldrich; St. Louis, MO, USA), 50 mg/mL ascorbate 2-phosphate (Sigma-Aldrich), 40 mg/ml l-proline (Sigma-Aldrich), 100 mg/mL sodium pyruvate (Corning Life Sciences, Corning, NY, USA), 6.25 μg/mL insulin, 6.25 μg/mL transferrin and 6.25 ng/mL selenious acid (ITS Premix; Corning), 1.25 mg/mL bovine serum albumin (BSA; Sigma-Aldrich), and 5.35 μg/mL linoleic acid, with no growth factors. Following 2 weeks of culture, pellets were then harvested and either analyzed for DNA and sulfated glycosaminoglycan (GAG) contents (n = 5 per donor for each expansion condition) or fixed in formalin for histology and immunohistochemistry for qualitative assessment of ECM production and distribution (n = 2 per donor for each expansion condition).

Pellet DNA content was quantified using the PicoGreen assay (ThermoFisher Scientific; Waltham, MA) following the manufacturer’s protocol and is presented both as absolute quantity per pellet, and as a percentage of pellet DNA on day 1 of pellet culture (i.e., as an indicator of cell survival after 2 weeks of culture under low oxygen and limited nutrient conditions). GAG content was quantified using the dimethyl methylene blue assay as previously described⁴⁶ and is presented both normalized to DNA content (representative of GAG production per cell), and as the absolute quantity per pellet.

Formalin-fixed pellets were processed into paraffin, and 7-μm thick sections were stained with Alcian blue (GAGs) or picrosirius red (collagen) and imaged using bright field light microscopy (Eclipse 90i; Nikon; Tokyo, Japan). For immunohistochemistry, antigen retrieval was carried out on rehydrated sections using a heat-mediated technique in a 95 °C buffered bath of 10 mM sodium citrate with 0.05% Tween 20, pH 6.0 for 20 minutes. Sections were permeabilized in TBS + 0.025% Triton X-100 (TBS-T) and blocked using Background Buster (Accurate Chemical & Scientific Corporation; Westbury, NY) for 30 minutes at room temperature. Sections were then incubated with primary antibodies against aggrecan (12/21/1-C-6, mouse monoclonal, Developmental Studies Hybridoma Bank; University of Iowa, Iowa City, IA), type I collagen (AB749P, rabbit polyclonal, EMD Millipore; Billerica, MA), or type II collagen (II-II6B3, mouse monoclonal, Developmental Studies Hybridoma Bank; University of Iowa, Iowa City, IA), diluted in Background Buster overnight at 4 °C. Slides were rinsed twice, 5 minutes each in TBS-T, then incubated in 0.3% H₂O₂ in TBS for 15 minutes to suppress endogenous peroxidase activity. Slides were then rinsed 3 times, 2 minutes each with TBS. Antibody staining was visualized using the SuperPicture Polymer Detection kit, diaminobenzidine (DAB), broad spectrum or rabbit (ThermoFisher Scientific; Waltham, MA), according to the manufacturer’s protocols. Slides were rinsed in diH₂O and counterstained with hematoxylin QS (Vector Laboratories; Burlingame, CA) for 30 seconds. Slides were then rinsed with tap H₂O, dehydrated and cleared, and cover-slipped with Permount Mounting Medium (ThermoFisher Scientific; Waltham, MA). Slides were imaged under bright field light microscopy. As negative controls, sections were treated in parallel with the same protocol, except with only Background Buster (no primary antibodies) during the initial overnight incubation.

Statistical Analysis

For the microarray study, statistical analysis was carried out using the Partek Genomics Suite (v6.6, Partek, Inc., St. Louis, MO). Significant differences between groups were established using a 3-way mixed model analysis of variance (ANOVA), with oxygen tension and TGF-β3 treatment as fixed variables and donor as a random variable, in order to account for intrinsic global gene expression differences present between individual animals. P values were adjusted for false discovery rate.

For the pellet culture study, statistical analyses were performed using Systat (Systat Software Inc.; San Jose, CA). Significant differences between groups were established using three-way ANOVA, with oxygen tension, TGF-β3 treatment, and donor as independent variables, and with Bonferroni post hoc tests used to make pairwise comparisons.

Significance was defined as P < 0.05 and trends defined as P < 0.1. Exact P values were calculated and are presented to 3 decimal places, except in cases where P < 0.001, which are reported as such. All results are expressed as mean ± standard deviation. Significant differences between treatment groups within each donor are presented as solid black lines and trends are presented as dotted black lines.

Results

Microarray Analysis: Effects of Preconditioning on MSC Global Gene Expression

DNA microarrays were used to examine the effects of each monolayer expansion condition and donor animal on the global gene expression profiles of MSCs. Analyses of variance established that TGF-β3 treatment, oxygen tension, and donor animal each exhibited significant effects on gene expression patterns after 1 week of monolayer preconditioning (principal component analysis [PCA] plot, Fig. 2A ). While not affecting the overall statistical findings, retrospective analysis of microarray data revealed that 1 donor lacked genes on the Y chromosome (i.e., it was female). As a result, samples from this donor clustered further from the other 2 donors ( Fig. 2A ). In terms of the total number of genes significantly affected by each variable, TGF-β3 treatment exhibited the greatest effect (3366 genes), followed by donor animal (476 genes) ( Fig. 2B and C ). Oxygen tension had the weakest effect on global gene expression (6 genes). Genes that showed a significant or trending dependence on oxygen tension in the absence of TGF-β3 (H− vs N−) were primarily those involved in growth, cell signaling, or metabolism including β-1,4-galactosyltransferase (B4GALT6), carbonic anhydrase 5B (CA5B), distal-less protein 4 (DLX4), contactin-associated protein-like 4 (CNTNAP4), pyruvate dehydrogenase lipoamide kinase isozyme 1 (PDK1), 6-phosphofructokinase, liver type (PFKL) ( Table 1A ), and cell-stress response pathway genes, including lipoprotein lipase (LPL), nerve growth factor (NGF), pyruvate kinase M (PKM), mitogen-activated protein kinase 13 (MAP3K13), and α-enolase (ENO1) ( Table 1B ). Most of these genes were upregulated in hypoxia, suggesting activation of related pathways. Likely due to the overwhelming effects the addition of TGF-β3 had on gene expression, no genes were significantly differentially expressed as a function of oxygen tension in the presence of TGF-β3 (H+ vs. N+). While only a small number of genes had statistically significant dependence on oxygen tension during monolayer expansion, hypoxic expansion (H−) did result in expected higher (nonsignificant) expression of genes regulated downstream of hypoxia inducible factor–α (HIF-1α)⁴⁷ compared with normoxia expansion (N−), including glucose transporters 1 and 3 (SLC2A1 and 3, 1.7, and 1.8-fold higher, respectively), lactate dehydrogenase A (LDHA, 1.3-fold higher) and glycogen synthase 1 (GYS1, 1.3-fold higher). Furthermore, pathway analysis using IPA confirmed HIF-1α as the top predicted upstream regulator of gene expression changes between H− and N− MSCs (activation score of 4.5, P < 0.001).

Figure 2.

Microarray analysis. (A) Principal component analysis (PCA) plots of microarray data. Lines connect groups from a single donor animal. PCA panel on the right is rotated to show donor clustering. Blue, normoxia; red, hypoxia; −, without transforming growth factor–β3 (TGF-β3); +, with TGF-β3. (B) Representation of total number of genes that exhibited significantly differential expression as a function of oxygen tension, TGF-β treatment, or donor animal. (C) Heat map and hierarchical clustering illustrating the effects of preconditioning using TGF-β under normoxia on global gene expression by mesenchymal stem cells (MSCs) (significantly differentially expressed genes only).

Table 1.

Microarray Analysis: Effects of Hypoxic Preconditioning on Mesenchymal Stem Cell (MSC) Gene Expression.

(A) Growth, Metabolism, and Cell Signaling Pathway Genes.
Gene	Description	Fold Change in Expression H− vs N−	P
B4GALT6	β-1,4-galactosyltransferase	−2.21	0.037
CA5B	carbonic anhydrase 5B	1.63	0.079
DLX4	distal-less protein 4	1.33	0.079
CNTNAP4	contactin-associated protein-like 4	2.68	0.079
PDK1	pyruvate dehydrogenase lipoamide kinase isozyme 1	1.57	0.086
PFKL	6-phosphofructokinase, liver type	1.81	0.086
(B) Cell Stress Response Pathway Genes.
Gene	Description	Fold Change in Expression H− vs. N−	P
LPL	lipoprotein lipase	2.84	0.038
NGF	nerve growth factor	−1.43	0.038
PKM	pyruvate kinase M	1.13	0.038
MAP3K13	mitogen-activated protein kinase 13	1.28	0.079
ENO1	α-enolase	1.20	0.079

The effects of TGF-β3 on MSC gene expression were independent of oxygen tension, and ANOVA confirmed that there was no significant interaction effect between oxygen tension and TGF-β3 treatment for any genes. TGF-β3 treatment resulted in significant differential expression of genes involved in activating growth pathways both in normoxia and hypoxia (i.e., N+ vs. N− and H+ vs. H−) including many genes in the TGF-β pathway, such as bone morphogenetic proteins 2 and 4 (BMP2, 4), SMAD7, TGFB2, and TGF-β receptor 3 (TGFBR3), along with other growth factor pathway genes such as platelet-derived growth factor receptor–β (PDGFRB) and SOX6 ( Table 2A and B ). TGF-β3 treatment, in both normoxia and hypoxia, also resulted in significant differential expression of genes in inflammatory signaling pathways such as interleukin signaling, tumor necrosis factor (TNF), and caspase pathways ( Table 2C and D ), including heat shock protein family B member 7 (HSPB7), interleukin 1 receptor type 1 (IL1R1), interleukins 18 and 36B (IL18, IL36B), TNF-α inducible protein 6 (TNFAIP6), and TNF receptor super family member 1B (TNFRSF1B). Interestingly, many growth and inflammation-related genes were significantly downregulated in the presence of TGF-β3, suggesting a decrease in cell-signaling within these related pathways in the presence of this growth factor. Global pathway analysis confirmed TGF-β as the top upstream regulator of gene expression changes between N− and N+ groups (activation score of 4.1, P < 0.001).

Table 2.

Microarray Analysis: Effects of TGF-β3 Preconditioning on Mesenchymal Stem Cell (MSC) Gene Expression.

(A) Growth-Related Genes (Normoxia)
Gene	Description	Fold Change in Expression N+ vs. N−	P
BMP4	bone morphogenetic protein 4	−1.99	0.004
SMAD7	SMAD family member 7	1.59	0.020
TGFBR3	transforming growth factor β receptor 3	−3.36	0.003
PDGFRB	platelet-derived growth factor receptor-β	−1.70	0.029
SOX6	SRY-box 6	−1.70	0.049
(B) Growth-Related Genes (Hypoxia)
Gene	Description	Fold Change in Expression H+ vs. H−	P
BMP2	bone morphogenetic protein 2	−2.65	0.046
BMP4	bone morphogenetic protein 4	−1.79	0.005
SMAD7	SMAD family member 7	1.38	0.051
TGFB2	transforming growth factor β 2	2.35	0.013
TGFBR3	transforming growth factor β receptor 3	−2.85	0.004
PDGFRB	platelet-derived growth factor receptor-β	−1.63	0.034
SOX6	SRY-box 6	−1.76	0.042
(C) Inflammation Pathway Genes (Normoxia)
Gene	Description	Fold Change in Expression N+ vs. N−	P
FOS	FBJ murine osteosarcoma viral oncogene homolog	2.02	0.072
CASP8	caspase 8, apoptosis-related cysteine peptidase	−1.67	0.012
HSPB7	heat shock 27-kDa protein family, member 7	2.53	0.013
IL1R1	interleukin 1 receptor, type I	−2.28	0.002
IL6ST	interleukin 6 signal transducer	−1.57	0.006
IL18	interleukin 18	−2.89	0.007
IL36B	interleukin 36, β	−2.83	0.056
TNFAIP6	tumor necrosis factor, α-induced protein 6	3.16	0.041
TNFRSF1B	tumor necrosis factor receptor superfamily, member 1B	2.74	0.021
(D) Inflammation Pathway Genes (Hypoxia)
Gene	Description	Fold Change in Expression H+ vs. H−	P
FOS	FBJ murine osteosarcoma viral oncogene homolog	2.30	0.046
CASP8	caspase 8, apoptosis-related cysteine peptidase	−1.47	0.022
HSPB7	heat shock 27-kDa protein family, member 7	2.40	0.014
IL1R1	interleukin 1 receptor, type I	−1.94	0.003
IL6ST	interleukin 6 signal transducer	−1.41	0.011
IL18	interleukin 18	−2.30	0.012
IL36B	interleukin 36, β	−3.16	0.043
TNFAIP6	tumor necrosis factor, α-induced protein 6	2.88	0.049
TNFRSF1B	tumor necrosis factor receptor superfamily, member 1B	3.04	0.016
TLR6	toll-like receptor 6	−1.98	0.047

Pellet Study: Effects of Preconditioning on MSC Survival and ECM Production

After establishing the nature of the gene expression changes induced in MSCs by preconditioning with hypoxia and/or TGF-β3 through microarray analyses, we carried out pellet culture studies to assess corresponding effects of preconditioning on cell survival and ECM production in low oxygen and limited nutrient conditions. These experiments were repeated with MSCs from 3 separate donors, and results from all 3 are presented. The results of the ANOVA established that oxygen tension, TGF-β treatment, and donor animal during preconditioning each significantly affected pellet DNA and GAG contents.

DNA content ( Fig. 3 ) is presented both as absolute quantity per pellet and as a percentage of pellet DNA on day 1 of pellet culture (i.e., as an indicator of cell survival after 2 weeks of culture under low oxygen and limited nutrient conditions). Effects of preconditioning on pellet DNA content (cell survival) were broadly consistent across donors, although the magnitude of these effects varied between donors. For all 3 donors, after 2 weeks of culture in low oxygen and nutrient conditions, DNA content for pellets formed from MSCs expanded under normoxia without TGF-β3 (N−, i.e., standard expansion conditions) was approximately 20% of pellet DNA content at day 1 ( Fig. 3 ). Similarly, for all 3 donors, for pellets formed from MSCs expanded in hypoxia without TGF-β3, DNA content was significantly higher than for pellets formed from MSCs expanded in normoxia either with or without TGF-β3 (i.e., H− vs. both N− and N+, all P < 0.001) indicating improved survival. Pellets formed from MSCs expanded in normoxia with TGF-β3 (N+) exhibited the lowest cell survival, although the magnitude was donor dependent and only significant for donor 3 compared with pellets formed from N- expanded MSCs (less than 10% cell survival). For donors 2 and 3, for pellets formed from MSCs expanded in hypoxia, there were significant effects of TGF-β3 on survival (i.e., H− vs. H+). Specifically, H+ resulted in significantly lower pellet DNA content compared with H− (P = 0.027 and 0.049 for donors 2 and 3, respectively). Overall, pellets with MSCs from donor 1 were the least sensitive to different expansion conditions with respect to cell survival and only exhibited significant differences in DNA content for N+ versus H−, and N− versus H−. Pellets formed from MSCs from donor 3 were the most sensitive to expansion condition with respect to cell survival, showing significant differences in DNA content for all possible comparisons between monolayer expansion conditions. DNA content was significantly different between donors 1 and 3 for pellets with N+ MSCs.

Figure 3.

Effects of preconditioning on mesenchymal stem cell (MSC) pellet DNA content after 2 weeks of low oxygen and limited nutrient culture. DNA content is presented both as absolute quantity per pellet and as percentage of day 1 DNA as an indicator of cell survival. N = 5 pellets per donor per condition; solid significance bars represent P < 0.05; dotted significance bars represent trends of P < 0.1.

Pellet GAG content is presented both normalized to DNA content ( Fig. 4A , representative of GAG production per cell) and as the absolute quantity per pellet ( Fig. 4B ). In general, the effects of preconditioning on GAG production exhibited considerable variability between donors. When normalized to DNA, GAG production for pellets formed from MSCs from donor 1 exhibited the greatest sensitivity to expansion condition ( Fig. 4A ). For this donor, GAG per DNA was highest for pellets formed from N+ and H+ MSCs (not significantly different) and was significantly higher for these pellets than those formed from either N− or H− MSCs. Pellets formed from MSCs from donors 2 and 3 showed less sensitivity to expansion condition ( Fig. 4A ). For donor 2, the only significant differences were N+ versus N− and N+ versus H−. Similar to donor 1, donor 2 pellets formed from N+ and H+ MSCs exhibited the highest GAG per DNA overall. For donor 3, GAG per DNA was highest for pellets formed from H− and H+ MSCs, and the only significant difference was between N− and H+. There were also significant donor-dependent effects on pellet GAG per DNA, independent of MSC expansion condition. Specifically, for pellets formed from H− MSCs, GAG per DNA was significantly higher for donor 3 compared with both donors 1 and 2 (P = 0.008 and 0.033, respectively).

Figure 4.

Effects of preconditioning on mesenchymal stem cell (MSC) pellet glycosaminoglycan (GAG) content after 2 weeks of low oxygen and limited nutrient culture. (A) GAG content normalized to DNA content. (B) GAG content per pellet. N = 5 pellets per donor per condition; solid significance bars represent P < 0.05; dotted significance bars represent trends of P < 0.1.

For GAG per pellet, donor 3 pellets showed the greatest sensitivity to MSC expansion condition, with GAG per pellet highest for pellets formed from H- and H+ MSCs and significant differences between these 2 conditions and all others ( Fig. 4B ). In contrast, MSCs from donors 1 and 2 were relatively insensitive to expansion condition with the only significant differences in GAG per pellet between N− and H+ MSCs for both. There were also donor-dependent differences in GAG per pellet independent of expansion condition. Specifically, for donor 3, GAG per pellet was significantly greater for pellets formed from H- MSCs than for donors 1 and 2 and significantly lower for N+ MSCs than for Donor 1.

Histological results largely supported findings from the above quantitative assays. Pellets formed from MSCs conditioned in hypoxia (both H− and H+) were larger than those conditioned in normoxia (both N− and N+), again reflecting enhanced cell survival ( Figs. 5 and 6 ). All pellets stained positively for GAG (Alcian blue) and aggrecan with staining intensity largely paralleling quantitative results ( Fig. 5A and B ). All pellets also stained positively for collagen; however, there were no clear observable differences as a function of expansion condition or between different MSC donors ( Fig. 6A ). With regard to immunostaining for collagen subtypes, positive staining was evident for both types I and II collagen, with type II collagen staining, in general, being more intense than type I collagen across all expansion conditions and donors ( Fig. 6B and C ).

Figure 5.

Histological assessment of glycosaminoglycan (GAG) and aggrecan composition for mesenchymal stem cell (MSC) pellets. (A) Representative pellet sections stained with Alcian blue for GAG. (B) Representative pellet sections immunostained for aggrecan protein. (C) Representative images of negative controls (no primary antibody) for both sets of secondary antibodies used for immunohistochemistry. D, donor; N−, monolayer expansion of MSCs in normoxia; N+, monolayer expansion of MSCs in normoxia with TGF-β3; H−, monolayer expansion of MSCs in hypoxia; H+, monolayer expansion of MSCs in hypoxia with TGF-β3; scale bar = 200 µm.

Figure 6.

Histological assessment of collagen composition for mesenchymal stem cell (MSC) pellets. (A) Representative pellet sections stained with picrosirius red for total collagen. (B) Representative pellet sections immunostained for type I collagen protein. (C) Representative pellet sections immunostained for type II collagen protein. D, donor; N−, monolayer expansion of MSCs in normoxia; N+, monolayer expansion of MSCs in normoxia with TGF-β3; H−, monolayer expansion of MSCs in hypoxia; H+, monolayer expansion of MSCs in hypoxia with TGF-β3; scale bar = 200 µm.

Discussion

In recent years, MSC-based therapies as clinically viable treatment options for musculoskeletal tissue regeneration and repair have made great progress.⁴⁸ While MSCs are multipotent rather than pluripotent, they are putative progenitor cells for many musculoskeletal tissues and have advantages over pluripotent stem cells because of their relative abundance, ease of isolation, biosafety, and immunomodulatory behavior.^12,15 Preconditioning during monolayer expansion is a promising technique for improving MSC performance upon in vivo implantation for treatment of tissues such as disc and cartilage, which are characterized by low oxygen and limited nutrient supply. Preconditioning may improve MSC performance by acclimating cells to in vivo microenvironmental conditions by inducing requisite metabolic changes and stimulating lineage-specific differentiation. Previous studies have shown that hypoxic expansion of MSCs can enhance chondrogenesis and suppress osteogenesis^35,49-52 and enhance proliferation.^33,36,53 Of note, these prior studies reported somewhat disparate findings with respect to the nature of the effects of hypoxic preconditioning, particularly with regard to promoting chondrogenesis and proliferation.^38,54 These disparate findings may have arisen from experimental differences (e.g., culture conditions), as well as different MSC donor sources, including age and species,³⁸ which ultimately may be of significant importance to clinical outcomes. Monolayer preconditioning has the advantage of adding little or no cell preparation time in the clinical setting and therefore, apart from the cost of growth factors and specialized equipment for hypoxic culture, does not detract from the translatability of MSCs. Here, we build on these prior studies and investigate the potential for preconditioning to improve MSC performance specifically for disc and cartilage regeneration applications. We examine whether preconditioning using hypoxia and/or TGF-β3 improves the survival and ECM production capacity of MSCs in a low oxygen and nutrient-limited pellet culture setting designed to partially simulate the microenvironmental stress cells might be exposed to in vivo in these avascular tissues. Furthermore, unlike most prior studies we examine the donor-dependent effects of preconditioning and use transcriptomic analysis to infer a mechanistic understanding of those effects.

As expected, without preconditioning (i.e., under standard expansion conditions), MSCs exhibited poor survival (measured as DNA content) in low oxygen and nutrient-limited pellet culture. Hypoxic preconditioning consistently improved MSC survival for each of the 3 different donors. These positive effects on survival support prior work using a variety of different in vitro culture conditions.^41,42 In light of the robust positive effects of hypoxic preconditioning on MSC survival, it was somewhat surprising that the corresponding effects on global gene expression from the microarray analysis were modest. This may suggest that the effects of hypoxic preconditioning are predominantly regulated at the posttranscriptional level. The genes that did exhibit significant changes in expression in response to hypoxia are components of various growth, cell signaling, metabolism, or cell stress response pathways.^55-57 The majority of these genes were upregulated in response to hypoxia. These gene expression changes are broadly consistent with the results of recent article by Elabd et al.⁵⁸ that examined transcriptome-wide changes for human MSCS cultured in hypoxia. Combined with pathway analysis results that indicate induction of genes downstream of HIF-1α and a switch to anaerobic metabolism, our microarray results suggest that hypoxic preconditioning acclimates MSCs to microenvironmental stress by inducing the metabolic changes necessary to promote survival. For example, it is possible that with hypoxic preconditioning, upregulation of genes such as LPL, PKM, and MAP3K13, which all play important roles in mitigating the cell stress response, prepares the cells for improved survival and function when subjected to microenvironmental stress. Upregulation of these genes during monolayer expansion likely allows cells to dedicate resources and energy to other metabolic processes more quickly once placed into low oxygen and nutrient-limited 3D pellet culture. In future work, we will carry out gene expression analysis of pellets following culture in low oxygen and limited nutrient conditions to assess whether changes to MSC gene expression as identified in the microarray study are maintained, as well as to identify other gene expression changes that may affect MSC survival.

In the absence of hypoxia, preconditioning with TGF-β3 impaired cell survival. TGF-β3 treatment exerted broad effects on MSC gene expression, which included downregulation of genes involved in tissue growth and response to inflammation and had minimal or negative effects on chondrogenic gene expression. This suggests that TGF-β3 treatment during monolayer expansion may induce the opposite effects of hypoxia, rendering cells less prepared to withstand microenvironmental stressors such as oxygen and nutrient deprivation. Previous studies have shown that exposure of MSCs to TGF-β followed by withdrawal can negatively affect sustained anabolic potential.⁵⁹ In our study, the overall effects of hypoxia and TGF-β3 on cell survival were inferior to hypoxia alone. Microarray results showed no significant interactive effects of hypoxia and TGF-β3 on global gene expression patterns.

In contrast to the effects on cell survival, the effects of preconditioning (in both nature and magnitude) on ECM (specifically, GAG) production by MSCs were more variable across the 3 donor animals. This variability may be attributed to factors such as genetic diversity, age, and gender, the effects of which should be investigated more comprehensively in the future. In the clinical setting, the potential impact of these factors is likely to be highly significant. Identifying specific genetic factors or molecular signatures that predispose MSC populations from certain donors to enhanced performance, both with and without preconditioning, may lay the foundation for the development of improved, patient-specific MSC treatment strategies. Further elucidating the biological basis of donor-dependent effects of preconditioning on both animal and human MSCs will therefore be a focus of future studies.

Matrix deposition in pellets was fairly homogeneous for all MSC expansion conditions and was composed primarily of aggrecan and type II collagen, with lesser amounts of type I collagen, suggesting chondrogenic rather than pro-fibrotic or osteogenic MSC behavior. Previous studies suggest that long-term 3D culture of MSCs can result in phenotype instability and hypertrophy.²² Future work will therefore seek to establish whether preconditioning enhances or diminishes MSC phenotype stability in long-term low oxygen and low nutrient 3D culture.

A limitation of this study was that only juvenile bovine MSCs were used, which did not permit an investigation of age-related changes to cellular function, nor inter-species variations. All this work was carried out in vitro, and while our pellet culture simulated low oxygen and limited nutrition, many other characteristics of the in vivo cellular microenvironment, including inflammation, pH, and physical cell-matrix interactions are likely to impact MSC performance. Co-delivery of anti-inflammatory drugs and physical preconditioning on soft substrates may further enhance in vivo MSC performance on implantation to degenerate disc and cartilage.^60,61

Conclusions

Through these studies, we provide evidence that preconditioning bone marrow–derived MSCs in hypoxia, but not with TGF-β, significantly enhances their survival upon subsequent culture in a low oxygen and nutrient-limited microenvironment. Preconditioning with hypoxia or TGF-β can enhance ECM production, but the nature and magnitude of these effects are likely donor-dependent. Future studies will seek to further elucidate the factors underlying the varying response of MSCs to preconditioning and progress these findings toward clinical application through evaluation in preclinical in vivo models of disc and cartilage degeneration and regeneration.

Footnotes

Authors’ Note

The microarray datasets generated and analyzed during this study are available in the NCBI Gene Expression Omnibus (GEO) repository, GSE109567.

Acknowledgments and Funding

The authors acknowledge technical support provided by staff members at the Molecular Profiling Facility at the University of Pennsylvania. The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was funded by grants from the Department of Veteran’s Affairs (I01 RX001321 and I01 RX001213), and the National Institutes of Health (R21 AR070959, F32 AR071298). Additional support was received from the Penn Center for Musculoskeletal Disorders (P30 AR069619) and the Catherine D. Sharpe Foundation.

Author Contributions

S.H.P. designed and performed the experiments and drafted the manuscript. J.R.B. performed the experiments. J.W.T. carried out statistical analysis of the microarray data. G.R.D., R.L.M, and N.R.M. provided conceptual input on experimental design. L.J.S. designed the experiments and drafted the manuscript. All authors reviewed and approved the manuscript prior to submission.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Animal Welfare

The present study followed international, national, and/or institutional guidelines for humane animal treatment and complied with relevant legislation.

Ethical Approval

Ethical approval was not sought for the present study because post mortem bovine tissues were purchased from a third party vendor (Research 87).

ORCID iD

Lachlan J. Smith

References

Huey

Athanasiou

KA.

Unlike bone, cartilage regeneration remains elusive. Science. 2012;338(6109):917-21.

Grunhagen

Shirazi-Adl

Fairbank

Urban

JP.

Intervertebral disk nutrition: a review of factors influencing concentrations of nutrients and metabolites. Orthop Clin North Am. 2011;42(4):465-77.

Horner

Urban

JP.

2001 Volvo Award Winner in Basic Science Studies: effect of nutrient supply on the viability of cells from the nucleus pulposus of the intervertebral disc. Spine (Phila Pa 1976). 2001;26(23):2543-9.

Williams

Khan

Richardson

Nelson

McCarthy

Analbelsi

, et al. Identification and clonal characterisation of a progenitor cell sub-population in normal human articular cartilage. PLoS One. 2010;5(10):e13246.

Vos

Flaxman

Naghavi

Lozano

Michaud

Ezzati

, et al. Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2163-96.

Reid

Shengelia

Parker

SJ.

Pharmacologic management of osteoarthritis-related pain in older adults: a review shows that many drug therapies provide small-to-modest pain relief. HSS J. 2012;8(2):159-64.

Raj

PP.

Intervertebral disc: anatomy-physiology-pathophysiology-treatment. Pain Pract. 2008;8(1):18-44.

Richardson

Hoyland

Mobasheri

Csaki

Shakibaei

Mobasheri

Mesenchymal stem cells in regenerative medicine: opportunities and challenges for articular cartilage and intervertebral disc tissue engineering. J Cell Physiol. 2010;222(1):23-32.

Orth

Rey-Rico

Venkatesan

Madry

Cucchiarini

Current perspectives in stem cell research for knee cartilage repair. Stem Cells Cloning. 2014;7:1-17.

10.

Wei

Shen

Williams

Diwan

Mesenchymal stem cells: potential application in intervertebral disc regeneration. Transl Pediatr. 2014;3(2):71-90.

11.

Vadala

Russo

Ambrosio

Loppini

Denaro

Stem cells sources for intervertebral disc regeneration. World J Stem Cells. 2016;8(5):185-201.

12.

Noel

Djouad

Jorgense

Regenerative medicine through mesenchymal stem cells for bone and cartilage repair. Curr Opin Investig Drugs. 2002;3(7):1000-4.

13.

Teixeira

Pereira

Ferreira

Maia

Gomez-Lazaro

Barbosa

, et al. Immunomodulation of human mesenchymal stem/stromal cells in intervertebral disc degeneration: insights from a proinflammatory/degenerative ex vivo model. Spine (Phila Pa 1976). 2018;43(12):E673-82.

14.

Lo Sicco

Reverberi

Balbi

Ulivi

Principi

Pascucci

, et al. Mesenchymal stem cell-derived extracellular vesicles as mediators of anti-inflammatory effects: endorsement of macrophage polarization. Stem Cells Transl Med. 2017;6(3):1018-28.

15.

Pittenger

Mackay

Beck

Jaiswal

Douglas

Mosca

, et al. Multilineage potential of adult human mesenchymal stem cells. Science. 1999;284(5411):143-7.

16.

Wei

Chung

Tao

Brisby

Lin

Shen

, et al. Differentiation of rodent bone marrow mesenchymal stem cells into intervertebral disc-like cells following coculture with rat disc tissue. Tissue Eng Part A. 2009;15(9):2581-95.

17.

Stoyanov

Gantenbein-Ritter

Bertolo

Aebli

Baur

Alini

, et al. Role of hypoxia and growth and differentiation factor-5 on differentiation of human mesenchymal stem cells towards intervertebral nucleus pulposus-like cells. Eur Cell Mater. 2011;21:533-47.

18.

Risbud

Albert

Guttapalli

Vresilovic

Hillibrand

Vaccaro

, et al. Differentiation of mesenchymal stem cells towards a nucleus pulposus-like phenotype in vitro: implications for cell-based transplantation therapy. Spine (Phila Pa 1976). 2004;29(23):2627-32.

19.

Caplan

AI.

Adult mesenchymal stem cells for tissue engineering versus regenerative medicine. J Cell Physiol. 2007;213(2):341-7.

20.

Barry

Murphy

JM.

Mesenchymal stem cells: clinical applications and biological characterization. Int J Biochem Cell Biol. 2004;36(4):568-84.

21.

Goldberg

Mitchell

Soans

Kim

Zaidi

The use of mesenchymal stem cells for cartilage repair and regeneration: a systematic review. J Orthop Surg Res. 2017;12(1):39.

22.

Farrell

Fisher

Huang

Shin

Farrell

Mauck

RL.

Functional properties of bone marrow-derived MSC-based engineered cartilage are unstable with very long-term in vitro culture. J Biomech. 2014;47(9):2173-82.

23.

Vinardell

Sheehy

Buckley

Kelly

DJ.

A comparison of the functionality and in vivo phenotypic stability of cartilaginous tissues engineered from different stem cell sources. Tissue Eng Part A. 2012;18(11-12):1161-70.

24.

Acosta

Jr Metz

Adkisson

Liu

Carruthers-Liebenberg

Milliman

, et al. Porcine intervertebral disc repair using allogeneic juvenile articular chondrocytes or mesenchymal stem cells. Tissue Eng Part A. 2011;17(23-24):3045-55.

25.

Risbud

Guttapalli

Stokes

Hawkins

Danielson

Schaer

, et al. Nucleus pulposus cells express HIF-1 alpha under normoxic culture conditions: a metabolic adaptation to the intervertebral disc microenvironment. J Cell Biochem. 2006;98(1):152-9.

26.

Gorth

Lothstein

Chiaro

Farrell

Dodge

Elliott

, et al. Hypoxic regulation of functional extracellular matrix elaboration by nucleus pulposus cells in long-term agarose culture. J Orthop Res. 2015;33(5):747-54.

27.

Schipani

Ryan

Didrickson

Kobayashi

Knight

Johnson

RS.

Hypoxia in cartilage: HIF-1alpha is essential for chondrocyte growth arrest and survival. Genes Dev. 2001;15(21):2865-76.

28.

Wuertz

Godburn

Neidlinger-Wilke

Urban

Iatridis

JC.

Behavior of mesenchymal stem cells in the chemical microenvironment of the intervertebral disc. Spine (Phila Pa 1976). 2008;33(17):1843-9.

29.

Farrell

Shin

Smith

Mauck

RL.

Functional consequences of glucose and oxygen deprivation on engineered mesenchymal stem cell-based cartilage constructs. Osteoarthritis Cartilage. 2015;23(1):134-42.

30.

Naqvi

Buckley

CT.

Extracellular matrix production by nucleus pulposus and bone marrow stem cells in response to altered oxygen and glucose microenvironments. J Anat. 2015;227(6):757-66.

31.

Cote

McLeod

Farrell

McClanahan

Dunagin

Raj

, et al. Single-cell differences in matrix gene expression do not predict matrix deposition. Nat Commun. 2016;7:10865.

32.

Kim

Erickson

Huang

Garrity

Mauck

Steinberg

DR.

Donor variation and optimization of human mesenchymal stem cell chondrogenesis in hyaluronic acid. Tissue Eng Part A. 2018;24(21-22):1693-703.

33.

Krinner

Zscharnack

Bader

Drasdo

Galle

Impact of oxygen environment on mesenchymal stem cell expansion and chondrogenic differentiation. Cell Prolif. 2009;42(4):471-84.

34.

Zscharnack

Poesel

Galle

Bader

Low oxygen expansion improves subsequent chondrogenesis of ovine bone-marrow-derived mesenchymal stem cells in collagen type I hydrogel. Cells Tissues Organs. 2009;190(2):81-93.

35.

Muller

Benz

Ahlers

Gaissmaier

Mollenhauer

Hypoxic conditions during expansion culture prime human mesenchymal stromal precursor cells for chondrogenic differentiation in three-dimensional cultures. Cell Transplant. 2011;20(10):1589-602.

36.

Adesida

Mulet-Sierra

Jomha

NM.

Hypoxia mediated isolation and expansion enhances the chondrogenic capacity of bone marrow mesenchymal stromal cells. Stem Cell Res Ther. 2012;3(2):9.

37.

Mastrogiacomo

Cancedda

Quarto

Effect of different growth factors on the chondrogenic potential of human bone marrow stromal cells. Osteoarthritis Cartilage. 2001;9(Suppl A):S36-40.

38.

Pei

Environmental preconditioning rejuvenates adult stem cells’ proliferation and chondrogenic potential. Biomaterials. 2017;117:10-23.

39.

Hudson

Bonassar

LJ.

Hypoxic expansion of human mesenchymal stem cells enhances three-dimensional maturation of tissue-engineered intervertebral discs. Tissue Eng Part A. 2017;23(7-8):293-300.

40.

Elabd

Centeno

Schultz

Lutz

Ichim

Silva

FJ.

Intra-discal injection of autologous, hypoxic cultured bone marrow-derived mesenchymal stem cells in five patients with chronic lower back pain: a long-term safety and feasibility study. J Transl Med. 2016;14:253.

41.

Beegle

Lakatos

Kalomoiris

Stewart

Isseroff

Nolta

, et al. Hypoxic preconditioning of mesenchymal stromal cells induces metabolic changes, enhances survival, and promotes cell retention in vivo. Stem Cells. 2015;33(6):1818-28.

42.

Burian

Probst

Palla

Riedel

Saller

Cornelsen

, et al. Effect of hypoxia on the proliferation of porcine bone marrow-derived mesenchymal stem cells and adipose-derived mesenchymal stem cells in 2- and 3-dimensional culture. J Craniomaxillofac Surg. 2017;45(3):414-9.

43.

Chiou

Longaker

MT.

Mitogenic and chondrogenic effects of fibroblast growth factor-2 in adipose-derived mesenchymal cells. Biochem Biophys Res Commun. 2006;343(2):644-52.

44.

Hagmann

Moradi

Frank

Dreher

Kämmerer

Richter

, et al. FGF-2 addition during expansion of human bone marrow-derived stromal cells alters MSC surface marker distribution and chondrogenic differentiation potential. Cell Prolif. 2013;46(4):396-407.

45.

Irizarry

Hobbs

Collin

Beazer-Barclay

Antonellis

Scherf

, et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003;4(2):249-64.

46.

Farndale

Buttle

Barrett

AJ.

Improved quantitation and discrimination of sulphated glycosaminoglycans by use of dimethylmethylene blue. Biochim Biophys Acta. 1986;883(2):173-7.

47.

Liu

Shen

Zhao

Chen

GQ.

Targeted genes and interacting proteins of hypoxia inducible factor-1. Int J Biochem Mol Biol. 2012;3(2):165-78.

48.

Richardson

Kalamegam

Pushparaj

Matta

Memic

Khademhosseini

, et al. Mesenchymal stem cells in regenerative medicine: focus on articular cartilage and intervertebral disc regeneration. Methods. 2016;99:69-80.

49.

Ranera

Remacha

Álvarez-Arguedas

Castiella

Vázquez

Romero

, et al. Expansion under hypoxic conditions enhances the chondrogenic potential of equine bone marrow-derived mesenchymal stem cells. Vet J. 2013;195:248-51.

50.

Henrionnet

Liang

Roeder

Dossot

Wang

Magdalou

, et al. *Hypoxia for mesenchymal stem cell expansion and differentiation: the best way for enhancing TGFβ-induced chondrogenesis and preventing calcifications in alginate beads. Tissue Eng Part A. 2017;23(17-18):913-22.

51.

Choi

Pingguan-Murphy

Abas

WAW

Azmi

MAN

Omar

Chua

, et al. Impact of low oxygen tension on stemness, proliferation and differentiation potential of human adipose-derived stem cells. Biochem Biophys Res Commun. 2014;448(2):218-24.

52.

Sheehy

Buckley

Kelly

DJ.

Oxygen tension regulates the osteogenic, chondrogenic and endochondral phenotype of bone marrow derived mesenchymal stem cells. Biochem Biophys Res Commun. 2012;417(1):305-10.

53.

Pilgaard

Lund

Duroux

Lockstone

Taylor

Emmersen

, et al. Transcriptional signature of human adipose tissue-derived stem cells (hASCs) preconditioned for chondrogenesis in hypoxic conditions. Exp Cell Res. 2009;315(11):1937-52.

54.

Boyette

Creasey

Guzik

Lozito

Tuan

RS.

Human bone marrow-derived mesenchymal stem cells display enhanced clonogenicity but impaired differentiation with hypoxic preconditioning. Stem Cells Transl Med. 2014;3(2):241-54.

55.

Papandreou

Cairns

Fontana

Lim

Denko

NC.

HIF-1 mediates adaptation to hypoxia by actively downregulating mitochondrial oxygen consumption. Cell Metab. 2006;3(3):187-97.

56.

Wang

Kaufman

RJ.

The impact of the unfolded protein response on human disease. J Cell Biol. 2012;197(7):857-67.

57.

Whitmarsh

AJ.

Regulation of gene transcription by mitogen-activated protein kinase signaling pathways. Biochim Biophys Acta. 2007;1773(8):1285-98.

58.

Elabd

Ichim

Miller

Anneling

Grinstein

Vargas

, et al. Comparing atmospheric and hypoxic cultured mesenchymal stem cell transcriptome: implication for stem cell therapies targeting intervertebral discs. J Transl Med. 2018;16(1):222.

59.

Martin

Gullbrand

Mohanraj

Ashinsky

Kim

Ikuta

, et al. *Optimization of preculture conditions to maximize the in vivo performance of cell-seeded engineered intervertebral discs. Tissue Eng Part A. 2017;23(17-18):923-34.

60.

Gorth

Martin

Dodge

Elliott

Malhotra

Mauck

, et al. In vivo retention and bioactivity of IL-1ra microspheres in the rat intervertebral disc: a preliminary investigation. J Exp Orthop. 2014;1(1):15.

61.

Heo

Szczesny

Kim

Saleh

Mauck

RL.

Expansion of mesenchymal stem cells on electrospun scaffolds maintains stemness, mechano-responsivity, and differentiation potential. J Orthop Res. 2018;36(2):808-15.