SEARCH

SEARCH BY CITATION

Keywords:

  • Chondrogenesis;
  • Signal transduction;
  • Mechanical;
  • Transforming growth factor β;
  • Stromal cells

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

The overall objective of the present study was to investigate the mechanotransduction of bovine bone marrow stromal cells (BMSCs) through the interactions between transforming growth factor β1 (TGF-β1), dexamethasone, and dynamic compressive loading. Overall, the addition of TGF-β1 increased cell viability, extracellular matrix (ECM) gene expression, matrix synthesis, and sulfated glycosaminoglycan content over basal construct medium. The addition of dexamethasone further enhanced extracellular matrix gene expression and protein synthesis. There was little stimulation of ECM gene expression or matrix synthesis in any medium group by mechanical loading introduced on day 8. In contrast, there was significant stimulation of ECM gene expression and matrix synthesis in chondrogenic media by dynamic loading introduced on day 16. The level of stimulation was also dependent on the medium supplements, with the samples treated with basal medium being the least responsive and the samples treated with TGF-β1 and dexamethasone being the most responsive at day 16. Both collagen I and collagen II gene expressions were more responsive to dynamic loading than aggrecan gene expression. Dynamic compression upregulated Smad2/3 phosphorylation in samples treated with basal and TGF-β1 media. These findings suggest that interactions between mechanical stimuli and TGF-β signaling may be an important mechanotransduction pathway for BMSCs, and they indicate that mechanosensitivity may vary during the process of chondrogenesis.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

During normal development and skeletogenesis, mesenchymal precursors differentiate into a range of tissues in response to complex and varied environments [1, [2]3]. Various authors have hypothesized that mechanical loading following the onset of muscular contractions provides cues that modulate cell differentiation and functional matrix assembly [4, 5]. Both as models of differentiation and for the development of tissue engineering solutions, much recent effort has focused on studying the in vitro differentiation of mesenchymal precursors, which can be found in almost all mature skeletal tissues [6, 7]. Both biochemical stimulation and biomechanical stimulation have been shown to influence the growth and differentiation of bone marrow stromal cells (BMSCs) [8, [9]10]. Much effort investigating the effect of growth factor and cytokine supplementation on the chondrogenesis of BMSCs has focused around members of the transforming growth factor β (TGF-β) superfamily (TGF-βs, bone morphogenetic proteins, and activins), which have been found to play roles in chondrocyte growth, differentiation, and commitment [11, [12]13]. TGF-β family members have also been shown to participate in the control of proliferation, extracellular matrix synthesis, migration, and apoptosis in many different cell types, including chondrocytes and chondrogenic progenitor cells [11, 14, [15], [16]17].

TGF-β signals from the cell surface via a transmembrane serine/threonine kinase receptor complex [18, 19]. Upon ligand binding, the type II receptor subunit engages and transphosphorylates a type I receptor subunit (TβRI), which in turn phosphorylates the receptor-activated Smad proteins (R-Smads) Smad2 and Smad3. A protein complex with Smad4 forms with the activated R-Smads and translocates into the nucleus, where the complex interacts with additional transcription factors, binding to the promoters of responsive genes and regulating their expression by cooperating with other activators or repressors [20]. In addition to the Smad pathway, TGF-β has been shown to activate other signaling pathways, including p38 mitogen-activated protein kinase (MAPK) [21] and protein kinase C (PKC) [22]. TGF-β signaling has been demonstrated through the p38 MAPK pathway through activation of mitogen-activated protein kinase kinase 1 and subsequent ERK/ELK signaling [23, 24]. It has also been shown that G-protein-dependent activation of PKC results from TGF-β stimulation of growth plate chondrocytes [22]. TGF-β responsiveness may require the activation of the R-Smad2/4 complexes, as well as other signaling pathways [18, 19].

Mechanical stimulation has been shown to be important in the development of many tissues and may influence differentiation of BMSCs [25, 26]. The combination of chondrogenic medium and dynamic compressive loading may enhance chondrogenesis of BMSCs over the addition of exogenous biochemical factors alone [27]. Dynamic compressive loading has been shown to increase the expression of the chondrogenic markers aggrecan and collagen II in rabbit BMSCs in agarose culture [27]. Dynamic compressive loading has also been shown to induce chondrogenesis in chick-bud mesenchymal cells [28, 29]. Studies on human BMSCs cultured under conditions promoting chondrogenesis found that the application of cyclic hydrostatic pressure for multiple days increased proteoglycan and collagen contents after 14 days in culture [30]. Combined with the substantial knowledge base indicating that mechanical stimulation is required for the maintenance of healthy articular cartilage, these studies suggest that controlled mechanical stimulation may direct differentiation and subsequent matrix assembly in engineered cartilage derived from mesenchymal progenitor cells.

The mechanisms through which mechanotransduction occurs in chondrocytes and chondroprogenitor cells remain largely elusive. Investigation of the influence of transforming growth factor β1 (TGF-β1) signaling on the responses of chondrocytes and BMSCs to dynamic compressive loading may provide clues to the mechanisms involved in chondrocyte differentiation and define other potential targets to regulate this process. The overall objective of the present study was to investigate the mechanotransduction of BMSCs through the interactions among TGF-β1, dexamethasone, and dynamic compressive loading. Specifically, the effects of dynamic and static compressive loading on BMSC gene expression, matrix synthesis, and TGF-β signaling through Smad effector molecules were examined.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

Materials

Bovine femora and tibiae were from Research 87 (Marlborough, MA). Collagenase type 2 was from Worthington Biochemical (Lakewood, NJ, http://worthington-biochem.com/). ITS+, high-glucose Dulbecco's modified Eagle's medium (DMEM), phosphate-buffered saline (PBS), MEM Non-Essential Amino Acids Solution, trypsin-EDTA solution, antibiotic/antimycotic, Novex 4%–12% polyacrylamide gels, nitrocellulose membranes, and alkaline phosphatase (AP)-conjugated anti-biotin were from Invitrogen (Carlsbad, CA, http://www.invitrogen.com). Fetal bovine serum was from HyClone (Logan, UT, http://www.hyclone.com). Recombinant human TGF-β1 and basic fibroblast growth factor (bFGF) were from Peprotech (Rocky Hill, NJ, http://www.peprotech.com). Protease Inhibitor Cocktail Set 1 was from Calbiochem (La Jolla, CA, http://www.emdbiosciences.com). Smad2 and phospho-Smad2 antibodies were from Cell Signaling Technology (Danvers, MA, http://www.cellsignal.com). The electrochemiluminescence kit was from Amersham Biosciences (Piscataway, NJ, http://www.amersham.com). Biotin SP-conjugated anti-rabbit IgG was from Jackson Immunoresearch Laboratories (West Grove, PA, http://www.jacksonimmuno.com). Rabbit anti-collagen II and mouse anti-collagen I antibodies were from Abcam (Cambridge, U.K., http://www.abcam.com). Low-gelling-temperature agarose (type VII), ascorbate, chondroitin sulfate standard, l-proline, 1,9-dimethyl-methylene blue, calf thymus DNA, Tween 20, sodium orthovanadate, sodium chloride, Triton X-100, deoxycholate, sodium dodecyl sulfate, sodium fluoride, Tris-base, glycine, and Hoechst 33258 dye were from Sigma-Aldrich (St. Louis, http://www.sigmaaldrich.com). l-5-[3H]Proline and [35S]sodium sulfate were from American Radiolabeled Chemicals (St. Louis, http://www.arc-inc.com). Live/Dead Viability/Cytotoxicity Assay Kit and rhodamine-conjugated phalloidin were from Molecular Probes (Eugene, OR, http://probes.invitrogen.com).

Cell Isolation

BMSCs were isolated from both the femoral and tibial diaphyses of an immature calf on the day following slaughter. After removal of all fascia and muscle, the bones were cut at the mid-diaphysis with a sterile bone saw. Marrow was removed from the medullary canal and transferred to a 50-ml conical tube with sterile PBS plus 1% antibiotic/antimycotic (100 U/ml penicillin, 100 μg/ml streptomycin, and 250 ng/ml amphotericin). Marrow was sequentially passed through large-bore (25 ml) and small-bore (5 ml) pipets to disrupt large pieces and then sequentially passed through 16- and 18-gauge needles and centrifuged at 300g for 15 minutes. The separated fatty layer was removed and discarded. The cell pellet was resuspended in PBS, passed through a 20-gauge needle, and filtered through a 100-μm nylon filter. Mononuclear cells were counted with a Vi-Cell XR Cell Viability Analyzer using the trypan blue exclusion method.

Cells were plated in T-flasks at 5 × 103 mononuclear cells per cm2 in medium consisting of low glucose DMEM, 10% fetal bovine serum, 1% antibiotic/antimycotic, and 1 ng/ml bFGF. Nonadherent cells were removed during the first medium change, 3 days later. Cells were cultured until confluence (∼2 weeks, passage 0), detached with 0.05% trypsin/1 mM EDTA, and replated at 5 × 103 cells per cm2. Cells were grown to confluence and detached two more times and then seeded into 3% agarose gels. Agarose gels were assembled by autoclaving 6% LMP agarose in 1× Ca2+-, Mg2+-free PBS and then cooling the solution to 42°C. An equal volume of cells suspended at 40 × 106 cells per milliliter in 1× Ca2+-, Mg2+-free PBS was combined with the agarose for a final concentration of 20 × 106 cells per milliliter in 3% agarose. Initial mechanical characterization studies indicated that platen liftoff could occur with other formulations but did not occur for 3% agarose gels with the chosen mechanical stimulation protocol, avoiding artifactual stimulation due to enhanced nutrient transport. The agarose was cast in 3-mm sheets between two electrophoresis plates and cooled until polymerized (approximately 10 minutes). Biopsy punches were used to extract 4-mm-diameter disks containing approximately 740,000 cells each.

Preliminary studies were performed comparing the chondrogenic response to TGF-β1 and dexamethasone stimulation by BMSCs from five bovine donors. Cells from all donors had similar levels of proliferation and matrix production in chondrogenic medium. One representative donor was chosen for which to perform a more rigorous battery of assays for the time-course and mechanical loading studies. A subset of the mechanical stimulation studies were repeated with an additional donor, and the results were found to be consistent with those reported below.

Quantitative Reverse Transcription-Polymerase Chain Reaction

Immediately following loading, agarose gels were dissociated in Qiagen lysis buffer with 1% β-mercaptoethanol. The RNeasy Total RNA Kit (Qiagen, Chatsworth, CA, http://www1.qiagen.com) was then used according to the manufacturer's protocol to purify RNA from the samples. The yield of the purified isolate was read at 260 and 280 nm on a UV-1601 spectrophotometer (Shimadzu, Columbia, MD, http://www.ssi.shimadzu.com), and 1 μg of mRNA was transcribed to cDNA using the Promega reverse transcription (RT) system (Promega, Madison, WI, http://www.promega.com) following the manufacturer's protocol. The SYBR Green PCR Master Mix (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com) was mixed with primers and cDNA for real-time detection of amplification. Real-time, quantitative RT-polymerase chain reaction (RT-PCR) was performed with an ABI Prism 7700 Sequence Detector System (Applied BioSystems, http://www.appliedbiosystems.com) to assess aggrecan, collagen II, and collagen I mRNA expression levels using serially diluted standards of known amplicon concentrations.

Radiolabel Incorporation

Media were supplemented with 20 μCi/ml l-5-[3H]proline and 10μCi/ml [35S]sodium sulfate for the final 20 hours of each culture period to measure protein and proteoglycan synthesis, respectively. At the end of the specified period, the samples were washed four times for 30 minutes each time in PBS supplemented with 0.8 mM sodium sulfate and 1 mM l-proline at 4°C to allow unincorporated radiolabeled precursors to diffuse out of the samples. Samples were then weighed, lyophilized, reweighed, digested in 1 ml of 100 mM ammonium acetate buffer with 250 μg/ml proteinase K at 60°C for 24 hours, and assayed for radiolabel content with an LS5000TD liquid scintillation counter (Beckman Coulter, Fullerton, CA, http://www.beckmancoulter.com). Portions of each digest were assayed for total DNA using the Hoechst 33258 assay with calf thymus DNA as a standard [31] and for sulfated glycosaminoglycan (sGAG) content using the 1,9-dimethyl-methylene blue assay [32, 33] with shark cartilage chondroitin sulfate as a standard.

Viability Staining

After 8 or 16 days of culture, constructs were sectioned, and transverse slices were imaged for cell viability using the Molecular Probes Live/Dead kit (n = 3 per group per time point). Samples were rinsed in three 10-minute PBS washes with gentle agitation, followed by incubation for 1 hour in 4 μM calcein and 4 μM ethidium in PBS. To remove any unincorporated calcein and ethidium, samples were rinsed in three 10-minute PBS washes. Samples were then imaged with a confocal microscope at the requisite excitation and emission wavelengths.

Immunohistochemistry

Constructs were fixed in 10% formalin for 4 hours at room temperature, transferred to 30% sucrose for 48 hours at 4°C, embedded, and frozen in liquid nitrogen-cooled isopentane. Prior to immunostaining, frozen 7-μm sections were thawed and dried for 20 minutes at room temperature. Following fixation in acetone for 5 minutes, slides were dried for 5 minutes and rehydrated in PBS. Sections underwent enzymatic antigen retrieval with 0.5× trypsin at 37°C for 15 minutes. Slides were blocked with 2% normal goat serum, 0.1% gelatin, 0.5% Tween 20, and 1% bovine serum albumin (BSA) in PBS. Samples were incubated primary antibody solution (rabbit anti-collagen II [1:100] and mouse anti-collagen I [1:100] were prepared with 1% BSA/0.1% gelatin/PBS) for 1 hour at room temperature and then rinsed three times with 300 μl of PBS. Samples incubated with nonimmune rabbit IgG and nonimmune mouse IgG were used as negative controls. Samples were incubated in a secondary antibody solution with goat anti-rabbit Alexa Fluor 488 (1:100) and goat anti-mouse Alexa Fluor 594 (1:100) in PBS, with 1.25 μg/ml 4,6-diamidino-2-phenylindole as a nuclear counterstain, for 1 hour. Slides were then rinsed three times with 300 μl of PBS, mounted in gel mount, and coverslipped.

Western Blotting

Gels were suspended in 10% wt/vol modified radioimmunoprecipitation assay buffer with protease and phosphatase inhibitors, lysed by freezing with liquid nitrogen, and then homogenized on ice. DNA concentration was determined using the Hoechst 33258 assay. Samples containing 1 μg of DNA were electrophoresed through Novex 4%–12% gels, and then proteins were transferred onto nitrocellulose membranes. After transfer, membranes were blocked and then incubated in primary antibodies (phosphorylated-Smad2/3 [pSmad2/3] or Smad2/3) overnight at 4°C. Membranes were washed and then incubated in SP-conjugated anti-rabbit IgG for 1 hour at room temperature. Membranes were washed and incubated with AP-conjugated anti-biotin for 1 hour at room temperature. Membranes were then washed, developed with electrochemiluminescence solution, dried, and imaged using a Fujifilm FLA 3000 imaging system (Stamford, CT, http://www.fujifilm.com). Images were imported into Photoshop 6.0 (Adobe Systems Inc., San Jose, CA, http://www.adobe.com) and analyzed using Scion Image (Scion Corp., Frederick, MD, http://www.scioncorp.com).

Mechanical Loading System

Dynamic unconfined compression was applied using a custom-designed mechanical loading system. The motion of a 404XR square rail linear table (Parker Automation, Irwin, PA, http://www.parker.com/) was controlled with a BE231 servo motor driven by a VIX500AE servo drive (both from Parker Automation). Using Galil WSDK programming software, the sinusoidal input for table motion was sent to the Galil DMC-2113 servo controller (Galil Motion Control, Rocklin, CA, http://www.galilmc.com/). The position of the bracket was detected using a linear encoder mounted external to the linear table. Samples were placed within individual wells in autoclavable polysulfone chambers. Samples were compressed between the chamber base and stainless steel platens of the chamber lid, which was affixed to the linear table. Four chambers holding up to eight samples each could be simultaneously loaded in the mechanical loading system. To impose static compression, identical chambers were clamped, with compression limited by stainless steel spacers machined to specific heights.

Experimental Design

Unloaded Time Course.

Agarose gels were cultured in basal medium consisting of high-glucose DMEM, antibiotic/antimycotic, nonessential amino acids, 1% ITS+, 50 μg/ml ascorbate, and 0.4 mM proline (BASAL); basal medium plus 10 ng/ml TGF-β1 (TGF-β1); or basal medium plus 10 ng/ml TGF-β1 and 100 nM dexamethasone (TGF-β1 + DEX) for 8 or 16 days. In a preliminary study, dexamethasone alone did not significantly increase DNA content or sGAG accumulation over 8 days relative to BASAL medium, and there were no qualitative differences in viability. As the focus of these studies was on the interactions between TGF-β and compression, the nonchondrogenic dexamethasone-alone condition was excluded from further studies. Media were changed every 2 days (n = 6/group). Gels were analyzed for viability, mRNA expression (aggrecan, collagen I, and collagen II), sGAG and total protein synthesis rates, sGAG accumulation, and DNA content after 8 and 16 days of culture.

Mechanical Stimulation.

Samples were cultured in either BASAL, TGF-β1, or TGF-β1 + DEX medium for either 8 or 16 days before the application of loading. Mechanical stimulation groups included static compression (10%), 1 Hz dynamic compression (10% ± 3%) and free swelling (FS). Gels were analyzed for mRNA expression (aggrecan, collagen I, and collagen II) after 3 hours, for rates of sGAG and total protein synthesis rates after 20 hours, or for Smad2/3 or pSmad2/3 after 1 hour.

Statistical Analysis.

The protein, sGAG, and DNA data were analyzed with a three-factor (day, medium, and loading) General Linear Model, and Western blot data were analyzed with a two-factor (medium and loading) General Linear Model. PCR data for each gene in free swelling gels were first transformed via Box-Cox analysis and then analyzed with a two-factor hierarchical General Linear Model (medium and mechanical loading). Medium was treated as nested within day, with the baseline group at day 0 treated as a separate medium condition, and pairwise comparisons were made on the nested variable. For the loading studies, PCR data were transformed via Box-Cox analysis and analyzed with a three-factor (day, media, and loading) General Linear Model. A value of p < .05 indicated significance, and Tukey's test was used for pairwise comparisons.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

Unloaded Time Course

Viability and DNA Content.

Overall, the samples that were treated with both TGF-β1 and dexamethasone were found to contain a greater amount of DNA than either of the other medium conditions (p ≤ .01). In pairwise comparisons, however, there were no significant differences at day 8 among medium conditions. At day 16, the DNA content for the TGF-β1 + DEX group was 40% greater than in the day 16 TGF-β1 groups (p = .009). There were no overall differences in cellular viability between gels imaged on day 8 and on day 16 for any medium group, so only images from day 8 samples are shown (Fig. 1). Qualitatively, a dramatic difference in viability was seen between the samples treated with BASAL medium and those treated with TGF-β1 or TGF-β1 plus dexamethasone (TGF-β1 + DEX), with a higher proportion of dead cells found in the samples treated with BASAL medium. There were no obvious differences between the groups treated with TGF-β1 and TGF-β1 plus dexamethasone. Viability was also lower in the interior of the constructs compared with the edges for all groups, likely due to diffusional restrictions leading to decreased proliferation or increased cell death in the interior.

thumbnail image

Figure Figure 1.. Representative viability staining for bone marrow stromal cells in agarose cultured for 8 days. Green indicates live cells; red indicates dead cells. Abbreviation: TGF-β1, transforming growth factor β1.

Download figure to PowerPoint

Gene Expression.

Overall, aggrecan, collagen II and collagen I gene expressions were highest in the samples treated with TGF-β1 plus dexamethasone and lowest in BASAL samples (p < .0001; Fig. 2A–2C). Aggrecan, collagen II, and collagen I expressions were upregulated over day 0 levels at both time points for TGF-β1 and TGF-β1 plus dexamethasone samples (p < .001). Aggrecan and collagen II expressions were upregulated over day 0 samples for BASAL samples (p < .0008). For BASAL samples, collagen II expression was upregulated (p < .0008), and aggrecan expression was downregulated from day 8 to day 16 (p = .0004). For TGF-β1 samples, collagen II expression was upregulated (p < .0001). For TGF-β1 plus dexamethasone samples, aggrecan expression was upregulated (p < .0001).

thumbnail image

Figure Figure 2.. Behavior of bone marrow stromal cells in agarose cultured unloaded for 8 and 16 days. (A–C): Gene expression. (D, E): Matrix synthesis rates. (F): Sulfated glycosaminoglycan accumulation. ∗, p < .05 versus day 0; and ★, p < .05 versus day 8 for a given medium condition (mean ± SEM; n = 6). Abbreviations: hr, hour; TGF-β1, transforming growth factor β1.

Download figure to PowerPoint

Matrix Synthesis.

Consistent with the gene expression results, both protein and sGAG synthesis rates increased with the addition of TGF-β and further increased with the addition of dexamethasone (p < .001) (Fig. 2D–2E). Both protein and sGAG synthesis rates increased significantly with time for samples treated with TGF-β1 plus dexamethasone (p ≤ .002). At both days 8 and 16, sGAG accumulation was highest in the gels treated with TGF-β1 plus dexamethasone and lowest in the gels treated with BASAL medium (p < .001 for all) (Fig. 2F). There were no changes in sGAG content for the gels treated with BASAL medium. sGAG content increased over time for the samples treated with TGF-β1 alone (p ≤ .0085) and with TGF-β1 plus dexamethasone (p < .001).

There were visible differences in sGAG deposition patterns among the three medium groups and two time points. Samples treated with BASAL medium showed no observable sGAG staining after 16 days of culture (Fig. 3). TGF-β1 samples had limited sGAG staining in the interiors of the gels even after 16 days, with a marked increase in sGAG deposition from day 8 to day 16 near the edges of the constructs (not shown). sGAG deposition was limited to the pericellular area of individual cells. Samples treated with TGF-β1 and dexamethasone showed no significant sGAG deposition in the interior of the construct after 8 days of culture, with some deposition seen near the edges. After 16 days of culture, samples treated with TGF-β1 and dexamethasone showed significantly more sGAG deposition both internally and near the edge compared with other time points and medium conditions, with diffuse staining and an increased intensity pericellularly.

thumbnail image

Figure Figure 3.. Representative safranin O staining for sulfated glycosaminoglycans after 8 and 16 days of culture. Abbreviations: DEX, dexamethasone; sGAG, sulfated glycosaminoglycan; TGF-β1, transforming growth factor β1.

Download figure to PowerPoint

Localization of both collagen I and collagen II was pericellular, with little interterritorial staining and considerably less staining for collagen I and collagen II in the samples treated with BASAL medium compared with the other medium conditions. There were no observable differences between the TGF-β1 and TGF-β1 + DEX samples in the amount of collagen accumulated. However, whereas samples treated with TGF-β1 had comparable signals for collagen I and collagen II, the samples treated with TGF-β1 + DEX had predominantly collagen II staining.

Overall, the addition of TGF-β1 increased both chondrogenic gene expression and sGAG accumulation compared with BASAL medium, with further increases with both TGF-β1 and dexamethasone. There was a marked increased in sGAG accumulation from day 8 to day 16 for samples treated with TGF-β1 alone and with TGF-β1 plus dexamethasone. Samples treated with TGF-β1 plus dexamethasone accumulated a substantially greater amount of sGAG after 16 days of culture.

Day 8 Short-Term Loading

Gene Expression.

The effects of dynamic compression varied greatly among medium conditions and genes for the samples cultured for 8 days prior to loading (Fig. 4A–4C). Dynamic compression for 3 hours downregulated aggrecan gene expression for the samples treated with TGF-β1 (5-fold; p = .0064), with no significant effect on either the BASAL or the TGF-β1 plus dexamethasone samples (Fig. 4A). Dynamic compression upregulated collagen II expression in samples treated with TGF-β1 and dexamethasone (2-fold; p = .02), with no significant effect on either the BASAL or TGF-β1 samples (Fig. 4B). Dynamic compression did not significantly alter collagen I expression for any group.

thumbnail image

Figure Figure 4.. Response of bone marrow stromal cells seeded in agarose and mechanically stimulated after 8 days of unloaded culture. (A–C): Gene expression after 3 hours of mechanical stimulation. (D, E): Matrix synthesis rates over 20 hours of mechanical stimulation. ★, p < .05 versus static for a given medium condition (mean ± SEM; n = 6). Abbreviations: DEX, dexamethasone; hr, hour; sGAG, sulfated glycosaminoglycan; TGF-β1, transforming growth factor β1.

Download figure to PowerPoint

Matrix Synthesis.

There were no significant differences in the pairwise comparisons for any of the medium conditions for either protein or sGAG synthesis rates for the samples cultured for 8 days prior to loading (Fig. 4D–4E). Overall, BASAL samples had the lowest sGAG synthesis rates (p < .001), and samples treated with TGF-β1 plus dexamethasone had the highest (p < .001). Basal samples had the lowest protein synthesis rates (p < .001).

Day 16 Short-Term Loading

Gene Expression.

The effects of dynamic compression also varied greatly among medium conditions and genes for the samples cultured for 16 days prior to loading but were strikingly different from the effects of loading on day 8 (Fig. 5A–5C). Dynamic compression upregulated aggrecan gene expression for samples treated with TGF-β1 plus dexamethasone (3-fold; p < .0001), with no significant effect on either the BASAL or the TGF-β1 samples (Fig. 5A). Dynamic compression upregulated collagen II expression for samples treated with TGF-β1 (5-fold; p = .001) and TGF-β1 plus dexamethasone (60-fold; p < .0001), with no effect on the BASAL samples (Fig. 5B). Dynamic compression upregulated collagen I expression for samples treated with TGF-β1 (30-fold; p < .0001) and TGF-β1 plus dexamethasone (100-fold; p < .0001), with no effect on the BASAL samples (Fig. 5C).

thumbnail image

Figure Figure 5.. Response of bone marrow stromal cells seeded in agarose and mechanically stimulated after 16 days of unloaded culture. (A–C): Gene expression after 3 hours of mechanical stimulation. (D, E): Matrix synthesis rates over 20 hours of mechanical stimulation. ★, p < .05 versus static for a given medium condition; and ∗, p < .05 versus all other groups (mean ± SEM; n = 6). Abbreviations: DEX, dexamethasone; hr, hour; sGAG, sulfated glycosaminoglycan; TGF-β1, transforming growth factor β1.

Download figure to PowerPoint

Matrix Synthesis.

Matrix synthesis rates were consistent with the gene expression results for samples cultured for 16 days prior to loading. Overall, both protein and sGAG synthesis rates were lowest in the BASAL samples (p < .001) and highest in the samples treated with TGF-β1 plus dexamethasone (p < .001) (Fig. 5D–5E). Dynamic compression had no significant effect on either protein or sGAG synthesis rates for BASAL samples. Dynamic compression stimulated protein (twofold; p < .001) but not sGAG synthesis rates for samples treated with TGF-β1 alone. Dynamic compression stimulated both protein (twofold; p < .001) and sGAG (twofold; p < .001) synthesis rates for samples treated with TGF-β1 plus dexamethasone.

Smad 2/3.

Due to the lack of responsiveness to mechanical loading in all medium groups at day 8, examination of Smad2/3 and pSmad2/3 was performed only on samples loaded on day 16. In all medium and mechanical stimulation conditions, there was diffuse Smad2/3 cytoplasmic staining for total Smad2/3 throughout the cells, with no noticeable differences in the level of staining or the numbers of cells stained (not shown). Among medium conditions, there were no significant differences in the levels of Smad2/3 detected by Western blotting (Fig. 6A, 6B). There were no significant differences among loading conditions for any of the medium conditions.

thumbnail image

Figure Figure 6.. Smad and pSmad levels for bone marrow stromal cells seeded in agarose and subjected to 1 hour of mechanical stimulation after 16 days of unloaded culture. (A): Representative Smad2/3 blot for the TGF-β1 samples, with the normalizing standard dynamic in lane 1, static samples in lanes 2–6, and Dyn samples in lanes 7–11. (B): Smad2/3 levels normalized by the standard dynamic after background subtraction. Dashed line indicates 100% of gel standard dynamic. (C): Representative pSmad2/3 blot for the TGF-β1 samples, with the normalizing standard dynamic in lane 1, static samples in lanes 2–6, and Dyn samples in lanes 7–11. (D): pSmad2/3 levels normalized by the standard dynamic after background subtraction. Dashed line indicates 100% of gel standard dynamic. ★, p < .05 versus static for a given medium condition (mean ± SEM; n = 5). Abbreviations: DEX, dexamethasone; Dyn, dynamic; STD, standard; TGF-β1, transforming growth factor β1.

Download figure to PowerPoint

In all medium and mechanical stimulation conditions, diffuse pSmad2/3 cytoplasmic staining throughout the cells was comparable to background staining using fluorescent microscopy. However, when imaged using confocal microscopy, all pSmad2/3 localization was nuclear, with no diffuse cytoplasmic staining apparent (not shown). Among medium conditions, there were significant differences in the levels of pSmad2/3 detected by Western blotting (p < .001; Fig. 6C, 6D). Samples treated with BASAL medium had significantly less pSmad2/3 protein compared with the other medium groups (p < .001); samples treated with TGF-β1 plus dexamethasone had significantly more pSmad2/3 protein compared with the other medium groups (p = .002). In a pairwise comparison, dynamic compression increased pSmad2/3 protein levels over static compression slightly for the BASAL (1.9-fold; p = .01) and TGF-β1 (1.4-fold; p = .005) groups, but not for the TGF-β1 plus dexamethasone group.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

Overall, TGF-β1 increased cell viability, extracellular matrix gene expression, matrix synthesis, and sGAG construct content over BASAL medium, and the addition of dexamethasone further enhanced extracellular matrix gene expression and protein synthesis. There was little stimulation of gene expression or matrix synthesis at day 8 with mechanical loading in BASAL or chondrogenic media. In contrast, on day 16, there was significant stimulation by dynamic loading for chondrogenic medium groups. The level of stimulation was also dependent on the medium condition, with the samples in BASAL medium being the least responsive and the samples supplemented with both TGF-β1 and dexamethasone being the most responsive at day 16. Overall, both collagen I and II gene expressions were more responsive to dynamic loading than was aggrecan expression.

A number of factors could be responsible for the observed differences in responsiveness to mechanical stimulation of samples cultured 8 and 16 days before loading. One contributing factor could be differences in the differentiation states of the cells in the various groups, as a greater level of chondrogenic differentiation may result in a response more similar to that of a terminally differentiated chondrocyte. Articular chondrocytes respond favorably to dynamic compression, either in their native matrix or in tissue-engineered constructs. The application of dynamic compression maintains articular cartilage integrity and stimulates cartilage-specific gene and protein expression in chondrocytes seeded in three-dimensional constructs [34, 35]. Different levels of matrix accumulation could also have affected both the biochemical signaling from the extracellular matrix involved in dynamic loading and the local mechanical stimuli resulting from the macroscopic loading. Knight et al. have demonstrated that the stiffness of isolated chondrons (chondrocytes with attached pericellular matrix intact) seeded into agarose is higher than that of the surrounding extracellular agarose environment, leading to stress shielding of the chondrocytes during loading [36, 37]. This stress shielding has been proposed as an explanation for the apparent loss of mechanoresponsiveness of chondrocytes in agarose, with mechanosensitivity regained after an interconnected ECM is established. The possibility exists that the level of cellular deformation within the BMSC population cultured in different medium formulations differs depending on the culture time and differentiation state. In addition, histological staining suggests that cells within a given construct had variable levels of pericellular matrix accumulation. Such inhomogeneous matrix accumulation could similarly lead to inhomogeneous responses to loading within a given construct, meaning that the fraction of cells that did respond to mechanical loading may have responded even more robustly than is suggested by the homogenized gene expression and matrix synthesis results. Safranin O staining of BMSCs in agarose showed the beginning of a more interconnected endogenously produced ECM only after 16 days of culture in the TGF-β1 + DEX group (Fig. 3). Irrespective of differentiation state, the degree of mechanostimulation can be affected by these varying amounts of accumulated matrix [36, [37]38].

The addition of dexamethasone to the chondrogenic medium greatly influenced both the gene expression and matrix synthesis of the differentiating BMSCs. Glucocorticoids, including dexamethasone, can affect gene expression through transcriptional and posttranscriptional mechanisms by binding to specific receptors that belong to the superfamily of nuclear receptors (classic mechanism). The glucocorticoid/receptor complex acts as a ligand-dependent transcriptional factor to either activate or repress the transcription of certain genes [39, 40]. Dexamethasone is typically added to both chondrogenic and osteogenic cultures to stimulate the osteochondral phenotype [41, 42] and has been found to promote chondrogenic differentiation of adult human mesenchymal stem cells by enhancing cartilage-specific gene expression [43]. In addition, Locker et al. found that dexamethasone induced Sox9 upregulation in the pluripotent mesoblastic C1 line [44]. Although the mechanisms of dexamethasone in the chondrogenic studies presented here are unknown, positive regulation through Smad2/3 or Sox9 and negative regulation of the inhibitory Smads and Smad ubiquitin regulatory factors are possibilities.

The interactions between TGF-β1 signaling and mechanical stimulation could be due to one or more of a wide range of potential mechanisms. Stimulation of the TGF-β signaling pathway could modulate mechanotransduction either directly or indirectly by increasing the sensitivity of the BMSCs to loading. For example, TGF-β signaling may lead to pSmad activation of mechanosensitive proteins, such as focal adhesion kinase and paxillin, which might therefore directly increase the mechanosensitivity of the cells [45, [46]47]. Alternatively, downstream targets of TGF-β signaling may be necessary components of mechanotransduction in chondrocyte progenitor cells. For example, TGF-β1 and TGF-β3 have been shown to upregulate Sox9 gene expression in differentiating cells [26, 48, 49]. Sox9 binds to the promoter regions of type II collagen and aggrecan, enhancing transcription of collagen II and aggrecan mRNA expressions [50, [51]52]. In addition, Smad3 enhances the transcriptional activity of Sox9 and the expression of the α1 (II) collagen gene by forming a transcriptional complex with Sox9 and binding to the promoter region [48]. By increasing transcription of the downstream targets of mechanical stimulation, TGF-β signaling may indirectly amplify the effects of mechanotransduction.

Conversely, mechanical stimulation may modulate TGF-β signaling. One direct mechanism may involve the production of TGF-β or its receptors through upregulation of mRNA expression, efficiency in translation, or a combination of both by mechanical stimulation. Huang et al. found that cyclic compressive loading promoted gene expressions of Sox9, c-Jun, and both TGF-β receptors and productions of their corresponding proteins in rabbit BMSCs in three-dimensional agarose culture [53]. TGF-β receptor I has been shown to cause receptor-activated Smad2/3 phosphorylation at the C-terminal SSXS motif, causing dissociation from the receptor and association with the common mediator Smad4 [54, [55]56]. Upon heteromeric complex formation, translocation to the nucleus leads to interactions with various DNA-binding cofactors and comodulators to activate transcription. An increase in the available supply of TβR1 could lead to a greater level of Smad2/3 activation, amplifying the signaling occurring as a result of TGF-β1 exogenous stimulation. In addition, mechanical stimulation may upregulate the production of proteolytic moieties such as plasmin and stromelysin-1, leading to the activation of endogenously produced, latent TGF-β1 [57, [58]59]. Mechanical stimulation might also increase the phosphorylation of the Smad2/3 proteins and/or their translocation to the nucleus via cell shape changes due to the physical compression of the cell, either separating or bringing together molecules and organelles within the cell and therefore indirectly enhancing the effects of TGF-β signaling.

The Smad signaling pathway has been shown to interact with various other TGF-β-influenced pathways in fully differentiated chondrocytes. Multiple intracellular signaling cascades, particularly those involving the mitogen-activated protein kinases p38, ERK1, and JNK, have been shown to be activated by TGF-β in promoting cartilage-specific gene expression [60, 61]. TGF-β stimulation induces the rapid transient phosphorylation of Smad 2, ERK1/2, and p38 necessary for upregulation of aggrecan gene expression in chondrogenic ATDC5 cells. Smad2 was found to be upregulated in the initial activation of aggrecan expression but not required for long-term expression, whereas both ERK1/2 and p38 were found to be necessary for sustained aggrecan expression [24]. Investigators have also shown that the TGF-β induction of biglycan expression in pancreatic cells requires activation of MKK6-p38 MAPK signaling downstream of Smad signaling [62]. This evidence suggests a complicated, multifaceted TGF-β signaling network that may participate in the response of BMSCs to dynamic compression. The initiation of chondrocyte-specific gene and protein upregulation by dynamic compression may require Smad signaling, and other signaling mechanisms may participate in the long-term regulatory response, as seen in ATDC5 cells. Further studies would be necessary to elucidate the temporal signaling response of these cells to mechanical stimulation.

This study has illustrated a relationship between dynamic compression and TGF-β signaling in chondroprogenitors, providing potential targets for manipulating cell differentiation and for treating diseased or injured cartilage. Understanding the effects of loading at different stages of progenitor differentiation could be important in understanding the amount of pretreatment these cells require prior to introduction into a cartilage defect. If effects of mechanical compression are neutral or negative during early stages of chondrogenesis, it may be necessary to predifferentiate BMSCs in vitro prior to implantation or to rigorously control postoperative weight bearing. Further studies are necessary to elucidate the specific mechanisms involved in the response to short-term loading and to determine whether the effects continue with sustained loading. Ultimately, these studies will be useful in understanding influences of loading on BMSC differentiation and in enhancing the in vitro development of tissue engineered constructs for cartilage regeneration.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

This work was supported in part by the Georgia Tech/Emory Center for the Engineering of Living Tissues, an Engineering Research Centers Program of the National Science Foundation under Award Number EEC-9731643, National Science Foundation graduate fellowships (J.K.M. and J.T.C.), and a Cellular and Tissue Engineering Training Grant Program under NIH award 5 T32 GM008433-13 (C.G.W.).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References