Temporal and Spatial Changes in Chondrogenic Gene Expression of MSCs in Response to Dynamic Loading 1Haugh, M G; 1Meyer, E G; 1Thorpe S D; 1Vinardell, T; 2Duffy, G; +1Kelly, D J

+1Trinity Centre for Bioengineering, Trinity College Dublin, Ireland, 2Dept of Anatomy, Royal College of Surgeons in Ireland, Dublin, Ireland Senior author: kellyd9@tcd.ie

INTRODUCTION:

Bone marrow derived mesenchymal stem cells (MSCs) are an attractive cell source for cartilage tissue engineering due to both their ease of isolation and expansion, and their ability to undergo chondrogenesis in 3D culture [1]. Previous work has shown that mechanical signals can be used to regulate chondrogenesis of MSCs [1-3]. It has become apparent that the response of MSCs to loading depends on the temporal application of loading [2, 4]. An understanding of the factors that determine the response of MSCs to loading is key to enabling the use of mechanical signalling to enhance MSC based cartilage therapies. However, investigating the response of MSCs to loading is complicated by the fact that mechanical signals and extra cellular matrix (ECM) accumulation vary both temporally and spatially within cell seeded constructs. Therefore, the objectives of this study were to 1) investigate temporal changes in gene expression as MSCs undergo TGF-β3 induced differentiation in agarose hydrogels, 2) quantify changes in gene expression due to the application of dynamic compression at discrete times during this process and 3) to determine the spatial differences in gene expression in both the absence and presence of dynamic compression and its relationship with ECM accumulation. METHODS:

Isolated porcine bone marrow derived MSCs (3×4 month old donors; 3rd passage) were suspended in 2% agarose Type VII at 15×106 cells/ml and cast in a stainless steel mould to produce cylinders (Ø5×3 mm). Samples were maintained in chemically defined basal media supplemented with 100nM dexamethasone and 10 ng/ml of TGF-β3. In order to investigate the effects of loading four groups were subjected to 1 week of dynamic loading (DL) starting at either day 0, 7, 14 or 21. Loaded samples were subjected to 10% strain at a frequency of 1 Hz for one hour/day, five days/week in a dynamic compression bioreactor [2]. The control group was incubated in a free selling (FS) environment. Samples were assessed at days 0, 7, 14 and 21. Sample cores (Ø3 mm) and annuli were separated and assayed independently for DNA using a Hoechst Bisbenzimide 33258 DNA Quantitation Kit, sulphated glycosaminoglycans (sGAG) using the Blyscan DMMB assay kit, and collagen through measurement of hydroxyproline content (n=3). RT-PCR was used to determine the relative expression of Aggrecan, Collagen type I and Collagen type II, with GAPDH as the house-keeping reference gene (n=3). Samples were stained histologically in alcian blue solution for sulphated mucins, and picro-sirius red for collagen. Collagen type II deposition was identified by immunohistochemistry. Results are presented as mean ± standard deviation. RESULTS: In FS constructs aggrecan and collagen II expression peaks at day 14 and is significantly higher in the core region of the constructs (Fig.1 A-C). At early time points loading did not have a significant effect on mRNA levels. However, at day 14 loading was found to significantly increase aggrecan and collagen II expression in the annulus region of the constructs, while at day 21 loading significantly increased aggrecan expression in the core of the constructs. No significant changes in collagen I expression were observed. Biochemical analysis shows an increase in sGAG and collagen content over the culture period (Fig.1 D&E). sGAG content was significantly higher in the core compared to annulus region at day 21. Loading for 1 week was found to have no significant effect on sGAG accumulation. Collagen content was found to be higher in the annulus region at early time points. However by day 21 there was no spatial difference in collagen content. Analysis of the DNA content shows a significant reduction in cell number between day 14 and 21 (Fig.1 E). The development of a clearly defined pericellular matrix can be seen in the histological images (Fig.2). By day 14 an accumulation of sGAGs can be seen surrounding the cells in addition to a strong diffusion of sGAGs into the surrounding construct. Collagen II staining does not show as defined a pericellular matrix compared to sGAG staining, suggesting that this region is predominantly composed of sGAGs. An accumulation of collagen II can also be seen in the constructs with increasing time.

Fig. 1: Gene expression and biochemical composition within both core and annular regions. A) Aggrecan, B) Collagen type I and C) Collagen type II expression. D) sGAG, E) Collagen and F) DNA content. In the case of gene expression the DL data is gathered directly after loading using samples that have not been previously loaded. The biochemical DL data is gathered from samples that have been exposed to loading throughout the previous week. Significance (p<0.05) compared to: (a) FS core, (b) FS annulus and (c) DL core.

Fig. 2: Histological evaluation of free swelling constructs showing sGAG and collagen type II accumulation. DISCUSSION: Chondrogenic gene expression and matrix accumulation were found to be higher in the core of the constructs in comparison to the annular region. This may be due to the hypoxic environment at the centre of the constructs [5]. In addition, it was observed that loading can significantly enhance chondrogenic gene expression when it is applied two weeks after the application of cytokine induced differentiation. This positive response to loading followed the development of a pericellular matrix, but did not correlate to the spatial differences in the accumulation of ECM, as evidenced by earlier increases in gene expression in the annular region where ECM accumulation was lower. These spatial differences in the response to dynamic compression, which appear unrelated to the accumulation of ECM, suggests that: 1) spatial variations in biophysical stimuli generated in the constructs also regulate gene expression or 2) a complex interplay exists between other regulatory factors (e.g. oxygen tension) and mechanical signals. Understanding how such biochemical and biophysical signals interact to control stem cell differentiation is a central challenge in the field of regenerative medicine. REFERENCES: [1] Huang et al., Stem Cells 2004; 22: 313-23. [2] Thorpe et al., Ann Biomed Eng 2010; In Press. [3] Kelly and Jacobs, Birth Defects Res C Embryo Today 2010; 90: 75-85. [4] Mouw et al., Stem Cells 2007; 25: 655-63. [5] Meyer et al, J Biomech 2010; In Press. ACKNOWLEDGEMENTS: Supported by the President of Ireland Young Researcher Award (08/YI5/B1336).

Paper No. 112 • ORS 2011 Annual Meeting