High-throughput assay for the identification of compounds regulating osteogenic differentiation of human mesenchymal stromal cells

Abstract

Human mesenchymal stromal cells are regarded as the golden standard for cell-based therapies. They present multilineage differentiation potential and trophic and immunosuppressive abilities, making them the best candidate for clinical applications. Several molecules have been described to increase bone formation and were mainly discovered by candidate approaches towards known signaling pathways controlling osteogenesis. However, their bone forming potential is still limited, making the search for novel molecules a necessity. High-throughput screening (HTS) not only allows the screening of a large number of diverse chemical compounds, but also allows the discovery of unexpected signaling pathways and molecular mechanisms for a certain application, even without the prior knowledge of the full molecular pathway. Typically HTS is performed in cell lines, however, in this manuscript we have performed a phenotypical screen on more clinically relevant human mesenchymal stromal cells, as a proof of principle that HTS can be performed in those cells and can be used to find small molecules that impact stem cell fate. From a library of pharmacologically active small molecules, we were able to identify novel compounds with increased osteogenic activity. These compounds allowed achieving levels of bone-specific alkaline phosphatase higher than any other combination previously known. By combining biochemical techniques, we were able to demonstrate that a medium to high-throughput phenotypic assay can be performed in academic research laboratories allowing the discovery of novel molecules able to enhance stem cell differentiation.

Figure 1. High-throughput Assay.

A) Schematic representation of HTA: hMSCs were grown on proliferation medium, dissociated using trypsin and plated into 96 well plates at a density of 2000 cells per well in osteogenic medium. Cells were allowed to attach for 24 h, after which medium was refreshed and compounds and controls were added to each plate. After 4 days, both ALP activity and proliferation was measured in a fluorescence plate reader. The compounds activity was evaluated in relation to the controls used. Positive compounds were then further confirmed by FACS and taken for further analysis. B) Representative diagram of the classes of action of the compounds used for the screen (Lopac library, Sigma-Aldrich).

Figure 2. Hit validation.

After being selected from the primary screen, the identified compounds (hits) were subjected to half-logarithmic dilutions in order to assess the optimal concentration of each compound that would induce the highest alkaline phosphatase (ALP) activity. The graphs represent: A1) ALP activity; A2) ACP activity A3) ALP activity/cell, in both basic and osteogenic medium when supplemented with different concentrations of the compound H-8 (as example, since the same was performed for all the hits obtained); B) The table represents the ALP-induction ratio when compared to the control situation (basic medium).

Figure 3. Confirmation of the Hits by flow cytometry on ALP.

A) Evaluation of the osteogenic potential of several treatment conditions by flow cytometry on hMSCS after 4 days treatment. Data is expressed as percentage of ALP positive cells. At least 10,000 cells were measured for each condition and experiments were performed in triplicate. Error bars represent standard deviation. Statistical analysis was performed using (one-way ANOVA and Tukey post-test with a significance level of 0.05. Asterisks represent * P<0.05, ** P<0.01 and *** P<0.001. BM (basic medium); OM (Osteogenic Medium); DEX (Dexamethasone). B) Number of donors in which the respective compound had a positive effect on ALP induction both in basic or osteogenic medium.

Figure 4. Confirmation of the hits by flow cytometry in 3 extra donors.

A,B) Evaluation of the osteogenic potential of several treatment conditions on hMSCS after 4 days. Data is expressed as ALP induction ratio. Data reflects the average of induction of the percentage of ALP-positive cells of each condition divided by the average of induction of the percentage of ALP-positive cells of the respective control (A – basic medium and B – osteogenic medium). Each experimental condition was performed in triplicate and data reflects the results of 3 independent donors). C,D) Table depicting the donor variation on the induction of the percentage of ALP positive cells after each experimental condition when compared to the respective control.

Figure 5. Hit compound activity in the context of reference osteogenic compounds.

The current graphs represents only one of the compounds validated (H-8) and all the possible combinations with the reference osteogenic molecules. Graphs represent the % ALP-positive cells of: the controls (A), the different combinations possible in BM (B) and the different combinations in OM (C). The osteogenic potential of each treatment was assessed by FACS after 4 days, in at least 10000 cells, in triplicate. Error bars represent standard deviation. Statistical analysis was performed using one-way ANOVA and Tukey post-test with a significance level of 0.05. Asterisks represent * P<0.05, ** P<0.01 and *** P<0.001. OM – basic medium supplemented with dexamethasone; cAMP – di-buteryl cyclic AMP, DEX – dexamethasone, NS – non significant.

Figure 6. Results from the Lopac library screen.

Diagram showing the results obtained from the 1280 compounds library screen. The evaluation process was composed of 4 main steps. 1. Selection of the compounds from the primary screen compared to the controls. 2. Dose-curve evaluation of the optimal concentration to be used (fluorescence-based read out). 3. Confirmation by FACS of the hits in cells from 3 other donors. 4. Confirmation of the osteogenic potential and comparison to the reference osteogenic compounds known so far. Out of 1280 compounds, 14 were identified in the primary screen, which were then subjected to dose-response analysis resulting in 5 hits. From these 5 hits, 4 were confirmed by FACS to have enhanced osteogenic potential in most donors. These 5 compounds (leads) were then further evaluated in 3 other donors and compared to reference osteogenic compounds which have resulted in novel osteogenic combinations with increased osteogenic potential that were never tested before. * Compounds not tested were discontinued from the manufacturer or had problems with solubility and therefore were not in conformation with the manufacturers specifications. ** Candidate molecule for in vivo studies, proven to induce osteogenesis in all donors tested and had a synergistic effect with reference osteogenic molecules.