We developed a bioreactor for automated cell seeding of three-dimensional scaffolds by continuous perfusion of a cell suspension through the scaffold pores in oscillating directions. Using quantitative biochemical and image analysis techniques, we then evaluated the efficiency and uniformity of perfusion seeding of Polyactive foams as compared to conventional static and spinner flask methods. Finally, we assessed the efficacy of the perfusion seeding technique for different scaffolds and cell types. Perfusion seeding of chondrocytes into Polyactive foams resulted in "viable cell seeding efficiencies," defined as the percentages of initially loaded cells that were seeded and remained viable, that were significantly higher (75 +/- 6%) than those by static (57% +/- 5%) and spinner flask seeding (55% +/- 8%). In addition, as compared to static and spinner flask methods, cells seeded by perfusion were respectively 2.6-fold and 3.8-fold more uniformly distributed and formed more homogeneously sized cell clusters. Chondrocytes seeded by perfusion into Hyaff-11 nonwoven meshes were 26% and 63%, respectively, more uniformly distributed than following static and spinner flask seeding. Bone marrow stromal cells seeded by perfusion into ChronOS porous ceramics were homogeneously distributed throughout the scaffold volume, while following the static method, cells were found only near the top surface of the ceramic. In summary, we demonstrated that our cell seeding perfusion bioreactor generated constructs with remarkably uniform cell distributions at high efficiencies, and was effective for a variety of scaffolds and different mesenchymal cell types.

Adult human articular chondrocytes were expanded in a medium with 10% serum (CTR) or further supplemented with different mitogens (i.e., EGF, PDGFbb, FGF-2, TGF beta 1, or FGF-2/TGF beta 1). Cells were then induced to redifferentiate in 3D pellets using serum-supplemented medium (SSM), serum-free medium (SFM), or SFM supplemented with factors inducing differentiation of chondroprogenitor cells (i.e., TGF beta 1 and/or dexamethasone). All factors tested during expansion enhanced chondrocyte proliferation and dedifferentiation, as assessed by the mRNA ratios of collagen type II to type I (CII/CI) and aggrecan to versican (Agg/Ver), using real-time PCR. FGF-2/TGF beta 1-expanded chondrocytes displayed the lowest doubling times, CII/CI and Agg/Ver ratios, averaging, respectively, 50, 0.2 and 15% of CTR-expanded cells. Redifferentiation in pellets was more efficient in SFM than SSM only for EGF-, PDGFbb- or FGF-2-expanded chondrocytes. Upon supplementation of SFM with TGF beta and dexamethasone (SFM TD), CII/CI ratios decreased 4.4-fold for EGF- and PDGFbb-expanded chondrocytes, but increased 96-fold for FGF-2/TGF beta 1-expanded cells. Chondrocytes expanded with FGF-2/TGF beta 1 and redifferentiated in SFM TD expressed the largest mRNA amounts of CII and aggrecan and generated cartilaginous tissues with the highest accumulation of glycosaminoglycans and collagen type II. Our results provide evidence that growth factors during chondrocyte expansion not only influence cell proliferation and differentiation, but also the cell potential to redifferentiate and respond to regulatory molecules upon transfer into a 3D environment.

No abstract

We developed and used real-time RT-PCR assays to investigate how the expression of typical osteoblast-related genes by human bone marrow stromal cells (BMSC) is regulated by (i) the culture time in medium inducing osteogenic differentiation and (ii) the previous expansion in medium enhancing cell osteogenic commitment. BMSC from six healthy donors were expanded in medium without (CTR) or with fibroblast growth factor-2 and dexamethasone (FGF/Dex; these factors are known to increase BMSC osteogenic commitment) and further cultivated for up to 20 days with ascorbic acid, beta-glycerophosphate and dexamethasone (these factors are typically used to induce BMSC osteogenic differentiation). Despite a high variability in the gene expression levels among different individuals, we identified the following statistically significant patterns. The mRNA levels of bone morphogenetic protein-2 (BMP-2), bone sialo protein-II (BSP), osteopontin (OP) and to a lower extent cbfa-1 increased with culture time in osteogenic medium (OM), both in CTR- and FGF/Dex-expanded BMSC, unlike levels of alkaline phosphatase, collagen type I, osteocalcin, and osteonectin. After 20 days culture in OM, BMP-2, BSP, and OP were more expressed in FGF/Dex than in CTR-expanded BMSC (mRNA levels were, respectively, 9.5-, 14.9-, and 5.8-fold higher), unlike all the other investigated genes. Analysis of single-colony-derived strains of BMSC further revealed that after 20 days culture in OM, only a subset of FGF/Dex-expanded clones expressed higher mRNA levels of BMP-2, BSP, and OP than CTR-expanded clones. In conclusion, we provide evidence that mRNA levels of BMP-2, BSP, and OP, quantified using real-time RT-PCR, can be used as markers to monitor the extent of BMSC osteogenic differentiation in vitro; using those markers, we further demonstrated that only a few subpopulations of BMSC display enhanced osteogenic differentiation following FGF/Dex expansion.

Our results provide evidence that gene patterns related to chondrocyte differentiation discriminate between CTR and OA human cartilage with higher sensitivity than single ECM genes. The method described here has the potential to improve understanding of the progression of OA and could become a valuable diagnostic tool.