Engineering a Pro-Osteogenic Secretome through the Transient Silencing of the Gene Encoding Secreted Frizzled Related Protein 1

Abstract

The evidence sustaining the regenerative properties of mesenchymal stem cells' (MSCs) secretome has prompted a paradigm change, where MSCs have shifted from being considered direct contributors to tissue regeneration toward being seen as cell factories for producing biotech medicines. We have previously designed a method to prime MSCs towards osteogenic differentiation by silencing the Wnt/β-Catenin inhibitor Sfpr1. This approach produces a significant increase in bone formation in osteoporotic mice. In this current work, we set to investigate the contribution of the secretome from the MSCs where Sfrp1 has been silenced, to the positive effect seen on bone regeneration in vivo. The conditioned media (CM) of the murine MSCs line C3H10T1/2, where Sfrp1 has been transiently silenced (CM-Sfrp1), was found to induce, in vitro, an increase in the osteogenic differentiation of this same cell line, as well as a decrease of the expression of the Wnt inhibitor Dkk1 in murine osteocytes ex vivo. A reduction in the RANKL/OPG ratio was also detected ex vivo, suggesting a negative effect of CM-Sfrp1 on osteoclastogenesis. Moreover, this CM significantly increases the mineralization of human primary MSCs isolated from osteoportotic patients in vitro. Proteomic analysis identified enrichment of proteins involved in osteogenesis within the soluble and vesicular fractions of this secretome. Altogether, we demonstrate the pro-osteogenic potential of the secretome of MSCs primmed in this fashion, suggesting that this is a valid approach to enhance the osteo-regenerative properties of MSCs' secretome.

Keywords: Sfrp1; bone regeneration; mesenchymal stem cells; osteogenesis; secretome.

Figure 1

Effect of the CM from C3H10T1/2 cells where Sfrp1 has been transiently silenced (mCM-Sfrp1) and control CM (mCM-Ctrl) on the osteogenic potential of the murine cell line C3H10T1/2. All analyses were performed at day 21 of osteogenic differentiation. (a) Semi-quantitative PCRs showing the relative expression levels of different osteogenic markers (Runx2, Alpl, Bglap and Isbp) pre-treated during 48 h with either the mCM-Sfrp1 or mCM-Ctrl prior to the induction of osteogenic differentiation. Graphs represent average values of three independent experiments (n = 3). mCM-Sfrp1 values are normalized to mCM-Ctrl values. (b) Graph showing the alkaline phosphatase activity in C3H10T1/2 under the different experimental conditions analyzed. Graph shows average values of three independent experiments (n = 3). (c) In vitro mineralization of C3H10T1/2 cells 21 days after being treated with the two CM revealed by Alizarin Red staining. mCM-Sfrp1 values are normalized to mCM-Ctrl values. Results from one representative experiment out of three different experiments performed (n = 3). For all experiments, technical triplicates were also performed with each individual sample. In all graphs, bars represent standard error of the mean values (SEM). *** p-value < 0.0005.

Figure 2

Effect of CM from the C3H10T1/2 cell line on the murine osteocytic cell line MLO-A5. Bars show gene expression in the osteocyte cell line MLO-A5 after being exposed for 48 h to different amounts (25% and 75%) of CM produced in the murine cell line C3H10T1/2 (mCM-Sfrp1 or mCM-Ctrl). All values were normalized against the values obtained in cells growing in fresh media (value of 1). Values shown correspond to average values of three biological replicates (n = 3). In all graphs, bars represent standard error of the mean values (SEM). t-test vs. Fresh M p-value is represented with *. t-test mCM-Sfrp1 vs. mCM-Ctrl p-value is represented with #. * and # p-value < 0.05; ** p-value < 0.005; ** and ### p-value < 0.005; *** and ### p-value < 0.0005.

Figure 3

Gene expression of bone turnover markers in ex vivo cultures of murine calvaria bone treated with mCM-Ctrl y mCM-Sfrp1. (a) Experimental design. Gene expression analysis was performed from mRNA extracted from all calvarial disks after 11 days exposed to a 50% CM. mRNA was extracted and converted to cDNA prior to gene expression analysis by real time-PCR. The entire experiment was repeated three times. Each time, six calvarial bone discs were analyzed for each experimental condition (Fresh M, mCM-Ctrl and mCM-Sfrp1) (b) Expression of osteocytes related genes normalized to Fresh M. (c) Expression of osteoclast-related genes normalized to Fresh M. (d) Expression of osteoblast-related genes normalized to Fres M. In all graphs, error bars represent standard error of the mean values. t-test mCM-Sfrp1 vs. Fresh M. p-value is represented with *. t-test mCM-Sfrp1 vs. CM-Ctrl p-value is represented with #. * and # p-value < 0.05; ** p-value < 0.005; *** and ### p-value < 0.0005.

Figure 4

Effect of the CM from the human MSC cell line ASC57Telo where SFRP1 has been transiently silenced (hCM-SFRP1) on the osteogenic potential of primary human osteoporotic MSCs (hMSCs-OP). All analyses were performed at 20 days of osteogenic differentiation. All graphs reflect average values from six different female patients with ages between 77 and 89 years old (n = 6). (a) Semi-quantitative PCRs showing the relative expression levels of different osteogenic markers: RUNX2 (Runt Related Factor 2), Alkaline phosphatase (ALPL) and Osteocalcin, (BGLAP) pre-treated during 48 h with either the hCM-SFRP1 or a CM from cells transfected with a control GapmeR (hCM-Ctrl). Technical triplicates were also performed with each individual sample. (b) Graph showing the average alkaline phosphatase activity in hMSCs-OP under the different experimental conditions analyzed. (c) In vitro mineralization of hMSCs-OP cells 20 days after being treated with the two conditioned media revealed by Alizarin Red staining. Results from one representative sample are shown in the pictures. Bar graph represents quantification of the staining. In all graphs, bars represent standard error of the mean values. * p-value < 0.05.

Figure 5

(a) Volcano plot showing changes in protein abundance in the different secretome fractions in mCM-Sfrp1 compared with those of mCM-Ctrl. Both CM were obtained after transfecting the murine MSC line C3H10T1/2 with either a GapmeR for Sfrp1 silencing or a Control GapmeR, respectively. Left: soluble fraction; right: exosomal fraction. (b) Image shows the hierarchical clustering of the significant hits found in the exosomal and soluble fraction. Left: soluble fraction; right: exosomal fraction. Hierarchical clustering was performed using Euclidean distance, and color shows the relative z-scored abundance of each hit among samples. In blue, statistically significant hits with two-fold enrichment in CM-Sfrp1 compared to CM-Ctrl; in red, statistically significant hits with two-fold enrichment in wild-type cells. Samples corresponding to soluble fractions are named with the suffix SF-CM. Samples corresponding to exosomal fractions are named with the suffix EF-CM. (c) Validation of three of the hits by qPCR. Dvl1 (Segment Polarity Protein Dishevelled Homolog DVL-1) is over-represented in the exosomal fraction of CM-Sfrp1, Mgp (Matrix Gla Protein) is under-represented in the soluble fraction of CM-Sfrp1, and Ccn4 (Cellular Communication Network Factor 4) is over-represented also in the soluble fraction of CM-Sfrp1. Graphs represent average values of three independent experiments (n = 3). mCM-Sfrp1 values are normalized to mCM-Ctrl values. For all three experiments, technical triplicates were also performed with each individual sample. In all graphs, bars represent standard error of the mean values (SEM). *** p-value < 0.0005.