To overcome the severely limited regenerative capacity of cartilage, bone marrow mesenchymal stromal cells (MSCs) are an attractive cell source that is accessible less invasively and in higher quantity than articular chondrocytes (ACs). However, current in vitro chondrogenic protocols induce MSCs to form transient cartilage reminiscent of growth plate cartilage that becomes hypertrophic and is remodeled into bone. In contrast, under the same conditions, ACs form stable articular-like cartilage. Developmental studies in mice have revealed that TGF-beta/BMP, Wnt, and Hedghog/PTHrP signaling are the major regulators of both, articular cartilage and endochondral bone formation. While the differential regulation of TGF-beta/BMP and Hedgehog/PTHrP in endochondral MSC versus AC chondral differentiation is established knowledge, little is known about Wnt in these cells. Aim of this study was therefore to compare in vitro levels of Wnt network components in MSC-derived endochondral versus AC-derived articular cartilage.

Whole genome expression data comparing human MSCs and ACs at days 0 and 28 of in vitro chondrogenesis were screened for differential expression of Wnt ligands, receptors, co-receptors, activators/inhibitors and signaling molecules. Expression of the most strongly differentially regulated Wnt network genes was studied in detail during in vitro chondrogenesis of MSCs vs ACs via qPCR at days 0, 7, 14, 21, 35, and 42.

During early chondrogenesis, most Wnt components were expressed at low levels in both MSCs and ACs, with two exceptions. MSCs started into chondrogenesis with significantly higher levels of the non-canonical ligand WNT5A. ACs on the other hand expressed significantly higher levels of the canonical antagonist FRZB on day 0. During advancing and late chondrogenesis, MSCs downregulated WNT5A but still expressed it at significantly higher levels at day 42 than ACs. Strong regulation was also evident for WNT11 and the receptor PTK7 which were both strongly upregulated in MSCs. Unlike MSCs, ACs barely regulated these non-canonical Wnt genes. With regard to canonical signaling, only the transcription factor LEF1 showed strong upregulation in MSCs, while FZD9 and FRZB were only slightly upregulated in late MSC chondrogenesis. Again, these genes remained unregulated in ACs.

Our data suggest that a dynamic Wnt network regulation may be a unique characteristic of endochondral MSC differentiation while during AC chondral differentiation Wnt expression remained rather low and stable. Overall, mRNA of the non-canonical Wnt network components were stronger regulated than canonical factors which may indicate that primarily non-canonical signaling is dynamic in endochondral differentiation. Next step is to assess levels of active and total beta-catenin, the canonical Wnt mediator, and to use Wnt antagonists to establish a causal relationship between Wnt signaling and endochondral differentiation.

Dynamic compressive loading of cartilage can support extracellular matrix (ECM) synthesis whereas abnormal loading such as disuse, static loading or altered joint biomechanics can disrupt the ECM, suppress the biosynthetic activity of chondrocytes and lead to osteoarthritis. Interactions with the pericellular matrix are believed to play a critical role in the response of chondrocytes to mechanical signals. Loading of intact cartilage explants can stimulate proteoglycan synthesis immediately while the response of chondrocytes in tissue engineering constructs dependent on the day of culture. In order to effectively utilize mechanical signals in the clinic as a non-drug-based intervention to improve cartilage regeneration after surgical treatment, it is essential to understand how ECM accumulation influences the loading response. This study explored how construct maturity affects regulation of ECM synthesis of chondrocytes exposed to dynamic loading and unraveled the molecular correlates of this response.

Human chondrocytes were expanded to passage 2, seeded into collagen scaffolds and cultured for 3, 21, or 35 days before exposure to a single loading episode. Dynamic compression was applied at 25% strain, 1 Hz, in 9 × 10 minute-intervals over 3h. Gene expression and protein alterations were characterized by qPCR and Western blotting. Proteoglycan and collagen synthesis were determined by radiolabel-incorporation over 24 hours.

Maturation of constructs during culture significantly elevated ECM deposition according to histology and GAG/DNA content and chondrocytes redifferentiated as evident from raising COL2A1 and ACAN expression. Loading of d3 constructs significantly reduced proteoglycan synthesis and ACAN expression compared to controls while the identical loading episode stimulated GAG production significantly (1.45-fold, p=0.016) in day 35 constructs. Only in mature constructs, pERK1/2 and its immediate response gene FOS were stimulated by loading. Also, SOX9 protein increased after loading only in d21 and d35 but not in d3 constructs. Interestingly, levels of phosphorylated Smad 1/5/9 protein declined during construct maturation, but no evidence was obtained for load-induced changes in pSmad 1/5/9 although BMP2 and BMP6 expression were stimulated by loading. Selected MAPK-, calcium-, Wnt- and Notch-responsive genes raised significantly independent of construct maturity albeit with a generally weaker amplitude in d3 constructs.

In conclusion, construct maturity determined whether cells showed an anabolic or catabolic response to the same loading episode and this was apparently determined by a differential SOX9 and pERK signaling response on a background of high versus low total pSmad1/5/9 protein levels. Next step is to use signaling inhibitors to investigate a causal relationship between Smad levels and a beneficial loading response in order to design cartilage replacement tissue for an optimal mechanical response for in vivo applications.

Objective

In order to effectively utilize mechanical signals in the clinic as a non-drug-based intervention to improve cartilage defect regeneration after surgical treatment, it is essential to identify crucial components of the cellular response that are typical to the anabolic process. The mechanisms behind the effect of mechanical stimulation are, however, not fully understood and the signaling pathways involved in the anabolic response of chondrocytes to mechano-transduction are not well described. Therefore, a genome-wide identification of mechano-regulated genes and candidate pathways in human chondrocytes subjected to a single anabolic loading episode was performed in this study and time evolution and re-inducibility of the response was characterized.

Cell-based tissue engineering is a promising approach for treating cartilage lesions but the optimal cell-scaffold combination for hyaline cartilage regeneration has yet to be identified. Novel hydrogels allow including tailored tissue type specific modifications with physiologically relevant peptides, by this selectively influencing the cell response. Aim of this study was to modify a poly(ethylene glycol) (PEG)/heparin hydrogel by functionalization with cell instructive peptides introducing matrix-metalloprotease (MMP)-degradability, the cell adhesion motif RGD, or collagen binding motifs (CKLER, CWYRGRL) to improve cartilage matrix deposition in tissue engineering constructs.

The hydrogels were formed by mixing thiol-endfunctionalized (MMP-insensitive) starPEG or starPEG-MMP-conjugates carrying MMP-sensitive peptides at every arm and maleimide-functionalized heparin [1] in the presence or absence of cell instructive peptides. Human mesenchymal stromal cells (MSC) or porcine chondrocytes were grown in the hydrogels for up to 4 weeks in vitro under chondrogenic conditions, and in vivo in subcutaneous pockets of immunodeficient mice.

MMP-sensitive and –insensitive starPEG/heparin hydrogels supported chondrogenic differentiation of MSC according to induction of COL2A1, BGN and ACAN mRNA expression. Enhanced MMP-sensitivity and therefore degradability increased cell viability and proliferation. RGD-modification of the hydrogels induced cell-spreading and an intensively interconnected cell network. Other than hypothesized, CKLER and CWYRGRL were unable to raise collagen deposition in constructs in vitro. Matrix deposition in chondrocyte-containing peptide-functionalized hydrogels was high and the instructive effect of the hydrogels on chondrocytes appeared stronger in vivo where the merely pericellular cartilaginous matrix deposition was overcome in RGD-functionalized starPEG/heparin hydrogels.

Peptide-functionalized starPEG/heparin hydrogel altered cell morphology, proliferation and differentiation with MSC being similar sensitive to cell-matrix interaction peptides like articular chondrocytes. We also demonstrated that in vivoperformance of cell instructive hydrogels can exceed results gained by in vitromodels. Altogether, the manipulation of hydrogel constructs with signaling cues is considered promising for functional cartilage tissue engineering.