Chondrogenic differentiation of bone marrow-derived mesenchymal stem cells with TGF-BETA, BMP2, GDF5 or combinations thereof does not result in a nucleus pulposus cell phenotype

Abstract

Purpose: Regenerative approaches for intervertebral disc (IVD) degeneration, using multipotent progenitor cells, are thought to be a promising therapeutic approach for low back pain. Chondrogenic differentiation of mesenchymal stem cells (MSC) with GDF5 was shown to induce expression of nucleus pulposus cell phenotypic marker genes. However, the presence or absence of a fibrotic annulus fibrosus cell phenotype after GDF5 stimulation has not been evaluated. In addition, GDF5 has not been compared directly to the more commonly used BMP2 in MSC chondrogenesis and induction of an IVD-like phenotype. Methods: Bone-marrow derived MSC were obtained from 5 donors. Chondrogenic differentiation was performed in pellet cultures with TGF-beta, BMP2, GDF5 or combinations and evaluated by matrix deposition, histology and gene expression changes. IVD-like differentiation was evaluated by qPCR for validated NP (T, KRT19, CD24, CA12, FOXF1, PAX1) and AF (COL5A1, COL12A1, ADAMTS17, SFRP2) phenotypic marker genes. Down-stream signaling was evaluated by phospho-SMAD immunoblotting, target gene activation and the BRE-luciferase reporter assay. Results: Five MSC donors underwent TGF-beta-induced chondrogenic differentiation. After two weeks, the NP marker gene KRT19 was down regulated and the AF marker genes COL5A1, ADAMTS17 and SFRP2 were up regulated (Fig.1A). A detailed time course experiment revealed a rapid down regulation of KRT19/CA12 and up regulation of SFRP2 within 3 days. Up regulation of PAX1, COL5A1 and COL12A1 was observed from 5–7 days onwards. Subsequently, BMP2 (20 ng/ml) and GDF5 (isomolar concentration) with and without TGF-beta (10 ng/ml) were used to induce chondrogenesis. In combination with TGF-beta, only BMP2 enhanced sGAG deposition, which was reflected by increased ACAN expression (Fig.1B). BMP2 and GDF5 stimulation alone was not sufficient to induce chondrogenesis. The NP marker gene CA12 was induced by co-stimulation with GDF5 in comparison to TGF-beta alone. The AF marker gene SFRP2 was suppressed by co-stimulation with BMP2 when compared to TGF-beta alone. As GDF5 exerted little effects on chondrogenesis and gene expression, we evaluated downstream activation of SMAD1/5/8. pSMAD1/5/8C was highest in MSC stimulated with BMP2 followed by TGF-beta and GDF5. Dose-response studies revealed that GDF5 (up to 2 μg/ml) was unable to induce similar BRE-luciferase reporter activity as 20 ng/ml BMP2 (Fig.1C). Finally, MSC were differentiated with 1 μg/ml GDF5, TGF-beta or the combination. At this concentration GDF5 enhanced TGF-beta-induced chondrogenesis (sGAG and DNA content), but failed to induce differentiation alone (sGAG content, SOX9, COL2A1, ACAN). Neither NP or AF marker genes were differentially expressed between TGF-beta or TGF-beta and GDF5 stimulated MSC. Conclusions: Chondrogenic differentiation of bone-marrow derived MSC with TGF-beta, BMP2 or GDF5 failed to induce a specific NP cell phenotype based on 6 NP marker genes. In particular Brachyury/T and FOXF1 expression were difficult to detect in MSC prior to and after chondrogenic differentiation. The observed effects of GDF5 on MSC chondrogenesis and KRT19 down- and PAX1 up regulation by TGF-beta are in line with literature. BMP2 was more potent in inducing SMAD1/5/8 signaling and enhancing TGF-beta induced chondrogenic differentiation then GDF5 at equimolar concentrations. Nonetheless, BMP2 (20 ng/ml) enhanced sGAG content but not DNA content, while GDF5 (1 μg/ml) enhanced both sGAG and DNA content. Importantly, multiple AF marker genes (COL1A1, COL5A1, COL12A1, SFRP2) were consistently induced by TGF-beta-mediated chondrogenesis and only SFRP2 was suppressed by BMP2. Therefore, we conclude that differentiation of bone-marrow derived MSC with TGF-beta, or in combination with GDF5 or BMP2, induces a mixed chondro-fibroblast cell phenotype that does not resemble an intervertebral disc cell phenotype. Hence, we recommend the use of both positive and negative markers for the evaluation of NP-like differentiation of progenitor cells.