The effects of iron oxide incorporation on the chondrogenic potential of three human cell types

Abstract

Non-invasive monitoring of living cells in vivo provides an important tool in the development of cell-based therapies in cartilage tissue engineering. High-resolution magnetic resonance imaging (MRI) has been used to monitor target cell populations in vivo. However, the side-effects on cell function of the labelling reagents, such as superparamagnetic iron oxide (SPIO), are still unclear. This study investigated the effect of SPIO particles on the chondrogenic differentiation of human bone marrow stromal cells (HBMSCs), neonatal and adult chondrocytes in vitro. Cells were labelled with SPIO for 24 h and chondrogenesis induced in serum-free medium including TGFβ3. For labelled/unlabelled cells, viability, morphology and proliferation were determined using CellTracker™ Green and PicoGreen dsDNA assays. The expression of SOX9, COL2A1 and ACAN was investigated using qRT-PCR after 2, 7 and 14 days. The results showed that viability was unaffected in all of the cells but cell morphology changed towards a 'stretched' phenotype following SPIO uptake. Cell proliferation was reduced only for labelled neonatal chondrocytes. SOX9 and COL2A1 expression decreased at day 2 but not at days 7 and 14 for labelled HBMSCs and adult chondrocytes; ACAN expression was unaffected. In contrast, SOX9 and COL2A1 expression were unaffected in labelled neonatal chondrocytes but a decrease in ACAN expression was seen at day 14. The results suggest that downregulation of chondrogenic genes associated with SPIO labelling is temporary and target cell-dependent. Resovist® can be used to label HBMSCs or mature chondrocytes for MR imaging of cells for cartilage tissue engineering.

Figure 1

Perl's Prussian blue (PPB) staining of HBMSCs, neonatal and adult chondrocytes. Light micrographs showing: (A–C) unlabelled cells; (D–F) Resovist-labelled cells. Blue stain indicates successful cell labelling with Resovist after 24 h. The unlabelled cells were unstained by PPB and only stained for Congo red. Magnification, ×10

Figure 2

Effect of SPIO labelling on cell morphology. Both labelled and unlabelled cells for each cell type were observed to be viable (stained fluorescent green with CFMDA). However, the morphology of labelled neonatal and adult chondrocytes appeared to be more stretched out and flattened in comparison to the unlabelled samples for each cell type. HBMSCs, human bone marrow stromal cells; NC, neonatal chondrocytes; AC, adult chondrocytes

Figure 3

Effect of Resovist labelling on cell proliferation. A significant difference in cell proliferation between labelled and unlabelled cells was observed only in the neonatal chondrocyte group. Results are expressed as mean ± SD (n = 3): *p < 0.05; ***p < 0.001

Figure 4

Effect of Resovist labelling on chondrogenic gene expression by HBMSCs. Resovist labelling affected expression levels of SOX9 and COL2A1 during the initial stages (day 2) in chondrogenic culture. Gene expression levels were then similar for labelled and unlabelled cells at days 7 and 14. Results are expressed as mean ± SD (n = 3): *p < 0.05; **p < 0.01; ***p < 0.001

Figure 5

Effect of Resovist labelling on chondrogenic gene expression by neonatal chondrocytes. Resovist labelling did not affect the expression of SOX9 and COL2A at days 2 and 14; a downregulation of the expression levels of these genes was seen at day 7. ACAN expression levels were similar between labelled and unlabelled cells at day 7, but a significant decrease in expression was seen at days 2 and 14. Results are expressed as mean ± SD (n = 3): *p < 0.05; **p < 0.01; ***p < 0.001

Figure 6

Effect of Resovist labelling on chondrogenic gene expression by adult chondrocytes. Resovist labelling affected expression levels of SOX9 and COL2A1 during the initial stages (day 2) of chondrogenic culture only. Gene expression levels were similar for labelled and unlabelled cells at days 7 and 14. No significant differences were seen in ACAN expression at any of the time intervals used. Results are expressed as mean ± SD (n = 3): *p < 0.05; **p < 0.01; ***p < 0.001