Cell-Based Double-Screening Method to Identify a Reliable Candidate for Osteogenesis-Targeting Compounds

Small-molecule compounds strongly affecting osteogenesis can form the basis of effective therapeutic strategies in bone regenerative medicine. A cell-based high-throughput screening system might be a powerful tool for identifying osteoblast-targeting candidates; however, this approach is generally limited with using only one molecule as a cell-based sensor that does not always reflect the activation of the osteogenic phenotype. In the present study, we used the MC3T3-E1 cell line stably transfected with the green fluorescent protein (GFP) reporter gene driven by a fragment of type I collagen promoter (Col-1a1GFP-MC3T3-E1) to evaluate a double-screening system to identify osteogenic inducible compounds using a combination of a cell-based reporter assay and detection of alkaline phosphatase (ALP) activity. Col-1a1GFP-MC3T3-E1 cells were cultured in an osteogenic induction medium after library screening of 1280 pharmacologically active compounds (Lopack1280). After 7 days, GFP fluorescence was measured using a microplate reader. After 14 days of osteogenic induction, the cells were stained with ALP. Library screening using the Col-1a1/GFP reporter and ALP staining assay detected three candidates with significant osteogenic induction ability. Furthermore, leflunomide, one of the three detected candidates, significantly promoted new bone formation in vivo. Therefore, this double-screening method could identify candidates for osteogenesis-targeting compounds more reliably than conventional methods.


Introduction
Tissue engineering is a regenerative approach for tissue regeneration or replacement of damaged tissues using cells, scaffolds, and bioactive factors [1]. Several surgical strategies based on the tissue engineering concept have been used in bone tissue engineering, including autogenous bone grafting and stem cell transplantation, scaffolds, and growth factors [2]. However, unavoidable operative stress and a lack of cost-effectiveness are potential issues with using these surgical and stem cell-based therapies [3].
Ideal and standard regenerative and antiresorptive treatments for bone diseases can be achieved by developing effective, safe, and low-cost drugs and biomaterials [4]. Bones are constantly remodeled through the coupling of bone resorption and formation by osteoclasts and osteoblasts, respectively. Several growth factors, such as platelet-derived growth

Ethical Considerations
All animal experiments in the present study strictly followed a protocol approved by the Institutional Animal Care and Use Committee of the Osaka University Graduate School of Dentistry (approval number: 19-054).

Double Detection of GFP and ALP Expressions
Col-1a1GFP-MC3T3-E1 cells were seeded into 96-well culture plates (black wall and clear bottom) for fluorescence-based assays at a density of 16,000 cells per well in MC3T3-E1 growth medium. The next day, the culture medium was exchanged for a fresh osteogenic induction medium containing 1 or 10 µM of osteogenic inducible small molecules (harmine, phenamil, and resveratrol) or 100 ng/mL BMP2. The medium was changed every two days.
GFP fluorescence in each well was measured 7 days after induction using a fluorescence microplate reader (GloMax-Multi Detection System; Promega, Madison, WI, USA). After the measurement, the cells were cultured in an osteogenic induction medium for another 7 days. After 14 days of osteogenic induction, a standard ALP staining method [15] was used to detect ALP activity in each well. Colorimetric analysis of ALP activity was performed as described previously [26], by measuring the optical density at a wavelength of 405 nm.
After establishing the cell-based double-screening system, 1280 pharmacologically active compounds (10 µM) from a small-molecule library (Lopack 1280 ; Sigma) [27] were used in the screening assay. A list of the compounds in the Lopack 1280 library is shown in Supplementary Table S1.

Cytotoxicity and Cell Proliferation Assays
MC3T3-E1 cells were seeded into 96-well tissue culture plates (1000 cells per well) and maintained in a growth medium for 24 h. The medium was replaced with a growth medium containing 0, 1, 10, 25, and 50 µM candidate compounds (Lef, 1-5, and LFM). The cells were then cultured for 5 and 6 days for the CytoTox-Glo luminescent cytotoxicity assay (Promega) [10] and WST-1 cell counting assay (Dojindo Laboratories, Kumamoto, Japan) [28] to evaluate cytotoxicity and cell proliferation, respectively.

Reverse Transcription Polymerase Chain Reaction (RT-PCR) Analyses
Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Hilden, Germany). After DNase treatment (Thermo Fisher Scientific), cDNA was synthesized from 1 µg of total RNA using SuperScript III reverse transcription (Thermo Fisher Scientific). Real-time quantitative RT-PCR analysis was performed using Thunderbird SYBR qPCR Mix (Toyobo, Osaka, Japan) on an ABI PRISM 7900 Sequence Detection System (Thermo Fisher Scientific). The mRNA expression of osteogenic marker genes (Osterix, Collagen 1a1, Runx2, and Osteocalcin) was determined using glyceraldehyde-3-phosphate dehydrogenase (Gapdh) as internal control. The primer pairs and sequences are shown in Supplementary Table S2.

Evaluation of the Candidate Molecules Using MSCs
Candidates of osteogenic inducible compounds (10 µM) by the library screening were added to mMSCs and rMSCs in the osteogenic induction medium, and cells were cultured for 21 days. To evaluate the effects of the candidates on the osteogenic differentiation of these cells, ALP/von Kossa staining was performed as described previously [15].

Lef Injection into Rat Calvarial Bone Defects
Eight-week-old male Sprague-Dawley rats were anesthetized, and a circular defect of 5 mm in diameter was formed at the calvaria [28]. Lef injection was performed as previously described [29,30] with minor modifications. Briefly, after the defect was formed, collagen graft material (Terudermis, Olympus Terumo Biomaterials, Tokyo, Japan) containing 0.5 or 5 µg of Lef was transplanted into the bone defects. Total 3 or 30 µg dosage of Lef was applied to each defect site dividing into 5 injections (0.5 or 5 µg for each injection) every three or four days. The same volume of saline was applied to the defect in the control group. Three weeks after the operation, calvariae were extracted for histological (hematoxylin and eosin (H&E) and TRAP staining) [28] and three-dimensional micro-computed tomography (CT) (R_mCT2; RIGAKU, Tokyo, Japan) analyses [31] of new bone formation. Bone density inside the defect was measured using the bone tissue analysis software program (TRI/3D-BON; RATOC System Engineering, Tokyo, Japan). For histological and micro-CT analyses, 10 and 7 samples, respectively, from different mice were used.

Statistical Analysis
One-way analysis of variance (ANOVA) with Dunnett's post hoc test was used to evaluate the statistical significance of the results. Statistical significance was defined as p < 0.05.

Verification of ALP and GFP Expression in Col-1a1GFP-MC3T3-E1 Cells
To verify the enhanced ALP activity and GFP expression in Col-1a1GFP-MC3T3-E1 cells in response to osteogenic induction factors, the cells were cultured in an osteogenic induction medium in the presence of phenamil and BMP2 for 14 days. Col-1a1GFP-MC3T3-E1 cells in the osteogenic induction medium showed slight ALP activity on day 14 ( Figure 1A). In contrast, robust ALP activity in Col-1a1GFP-MC3T3-E1 cells was confirmed in osteogenic induction medium containing phenamil (1 µM) and BMP2 (100 ng/mL). ALP activity was not detected in uninduced cells cultured in the growth medium.

Effects of Identified Candidate Compounds on Cytotoxicity and Cell Proliferation
A proliferation assay showed that 1 and 10 µM Lef, LFM, and 1-5 did not significantly affect the proliferation of MC3T3-E1 cells cultured for 6 days in the growth medium. In contrast, 25 and 50 µM of Lef and 1-5 significantly suppressed the proliferation of MC3T3-E1 cells at 6 days after treatment (Supplementary Figure S1A). Cytotoxicity assay showed that 1, 10, 25, and 50 µM of these candidate compounds did not significantly affect the survival of MC3T3-E1 cells after 5 days of culture under compound stimulation (Supplementary Figure S1B).
ALP/von Kossa staining confirmed a marked osteogenic induction in mMSCs, showing enhanced ALP activity and distinct extracellular matrix calcium deposition by 10 µM of Lef, LFM, and 1-5 on day 14 ( Figure 4A). In addition, 10 µM Lef enhanced ALP activity and nodule mineralization in rMSCs ( Figure 4B,C). LFM and 1-5 did not significantly enhance ALP activity and nodule mineralization of rMSCs; therefore, we selected Lef as a candidate for subsequent investigation in animal experiments.

Effects of Lef on Calvarial Bone Defect Regeneration
Three weeks after the operation, Micro-CT images showed superior new bone formation in the bone defects injected with Lef at 3 and 30 µM/site compared to the control condition ( Figure 5A,B). Micro-CT analysis demonstrated that the bone mineral content, volume, and density of bone regenerated by 30 µM/site were significantly higher than those in the control group (p < 0.05) ( Figure 5C-E).
H&E staining showed that the areas of the regenerated bone at the sites injected with 3 and 30 µM/site Lef were significantly larger than those in the control condition (p < 0.01) ( Figure 5F). Magnified images of 3 µM/site and 30 µM/site Lef-injected areas showed superior repair of calvarial defects with new bone formation compared to the control condition ( Figure 5G). The newly formed bone in the Lef-injected areas showed clear cement lines, which is the histological profile of the remodeled compact bone [28]. The number of TRAP-positive multinucleated cells, which represent osteoclasts (Supplementary Figure S2A), in the bone defect regions of the Lef-injected groups was not significantly different from that of the control group (Supplementary Figure S2B).

Effects of Identified Candidate Compounds on Cytotoxicity and Cell Proliferation
A proliferation assay showed that 1 and 10 µM Lef, LFM, and 1-5 did not significantly affect the proliferation of MC3T3-E1 cells cultured for 6 days in the growth medium. In contrast, 25 and 50 µM of Lef and 1-5 significantly suppressed the proliferation of MC3T3-E1 cells at 6 days after treatment (Supplementary Figure S1A). Cytotoxicity assay showed that 1, 10, 25, and 50 µM of these candidate compounds did not significantly affect the survival of MC3T3-E1 cells after 5 days of culture under compound stimulation (Supplementary Figure S1B).

Effects of Identified Candidate Compounds on Osteogenesis In Vitro
At this stage, three candidate compounds (Lef, LFM, and 1-5) were selected for further studies to evaluate their osteogenic induction activity. Quantitative RT-PCR showed an increasing trend in the expression of osteogenic marker genes (Osterix, Collagen 1a1, Runx2, and Osteocalcin) in MC3T3-E1 cells 10 days after osteogenic induction in the presence of Lef, LFM, and 1-5 ( Figure 3A). In particular, 25 µM Lef, 10 µM LFM, and 10-25 µM 1-5 significantly promoted the expression of Collagen 1a1 (p < 0.05), Collagen 1a1 (p < 0.05) and Osteocalcin (p < 0.01), and Osterix and Osteocalcin (p < 0.01), respectively. In addition, these candidate compounds significantly promoted ALP activity in MC3T3-E1 cells on day 14 after osteogenic induction in a concentration-dependent manner (p < 0.05) (Figure 3B).   ALP/von Kossa staining confirmed a marked osteogenic induction in mMSCs, showing enhanced ALP activity and distinct extracellular matrix calcium deposition by 10 µM of Lef, LFM, and 1-5 on day 14 ( Figure 4A). In addition, 10 µM Lef enhanced ALP activity and nodule mineralization in rMSCs ( Figure 4B,C). LFM and 1-5 did not significantly enhance ALP activity and nodule mineralization of rMSCs; therefore, we selected Lef as a candidate for subsequent investigation in animal experiments.

Effects of Lef on Calvarial Bone Defect Regeneration
Three weeks after the operation, Micro-CT images showed superior new bone formation in the bone defects injected with Lef at 3 and 30 µM/site compared to the control condition ( Figure 5A,B). Micro-CT analysis demonstrated that the bone mineral content, ( Figure 5F). Magnified images of 3 µM/site and 30 µM/site Lef-injected areas showed superior repair of calvarial defects with new bone formation compared to the control condition ( Figure 5G). The newly formed bone in the Lef-injected areas showed clear cement lines, which is the histological profile of the remodeled compact bone [28]. The number of TRAP-positive multinucleated cells, which represent osteoclasts (Supplementary Figure  S2A), in the bone defect regions of the Lef-injected groups was not significantly different from that of the control group (Supplementary Figure S2B).

Discussion
Chemical biology elucidates biological phenomena at the molecular level and is expected to be an essential step for drug discovery. A screening strategy based on a largescale compound library requires more efficient and accurate high-throughput screening : Remaining collagen graft material. Bars: 200 µM.

Discussion
Chemical biology elucidates biological phenomena at the molecular level and is expected to be an essential step for drug discovery. A screening strategy based on a largescale compound library requires more efficient and accurate high-throughput screening systems. To date, several cell-based screening systems have been reported [11], which detect cell proliferation [32], cytotoxicity [33], and differentiation [34]. Most library screening systems for osteoblasts have used the ALP activity index of MSCs and MC3T3 cells [35][36][37][38]. ALP activity is a marker of early osteogenic differentiation [39,40] and can be easily detected using staining. However, ALP staining is necessary to fix cells, making it difficult to evaluate other osteogenic differentiation markers using the same cell culture. Therefore, many studies have described using cell proliferation assays to double-screen indices not related to differentiation [37,38].
A screening system with an index set for several differentiation markers, rather than just one, is ideal for detecting compounds affecting osteogenic differentiation [41]. In addition, the detection methods for cell response must be simple and the detected data are quantitative. In the present study, we applied our new double-screening system using ALP staining and pre-osteoblastic Col-1a1GFP-MC3T3-E1 cell lines stably transfected with the GFP reporter gene driven using a fragment of the type I collagen promoter [12]. Type I collagen is an osteogenic marker; therefore, the cells in the present study robustly emitted GFP fluorescence in response to osteogenic activators such as BMP2 [23][24][25], harmine [16], phenamil [17,18], and resveratrol [19][20][21][22]. However, one limitation of this GFP fluorescencebased screening assay is that detecting only type I collagen expression does not always reflect the activation of the osteogenic phenotype. In addition, proteasome inhibitors, such as MG-132, inhibit the degradation of GFP, resulting in the detection of pseudo-GFP expression [42]. The GFP fluorescence detected in this system was not very high and that of BMP2 was 1.46-fold higher than that in osteogenic-induced cells. Additionally, compound AC-93253 iodide showed the greatest increase in GFP fluorescence (18.8-fold); however, it did not activate ALP in Col-1a1GFP-MC3T3-E1 cells. Therefore, we applied the detection of ALP, another early osteogenic marker, in combination with a GFP fluorescence-based screening assay.
By targeting two early osteogenic markers in the same pre-osteoblast culture, we evaluated the Lopack 1280 library of pharmacologically active small molecules. Such smallmolecule screening assays have attracted substantial attention in recent years as drug discovery tools and for evaluating molecular mechanisms. Alves et al. identified five novel compounds (H-8, GW 5074, propentofylline, pinacidil, and SQ 22,536) with increased osteogenic activity of human MSCs using the Lopack 1280 library based on their ALP activity and a cell proliferation assay [36]. The compounds detected in our study differed from those detected by Alves et al., except for H-8, even though the same library was used. Therefore, cell sources, detection index, time point, or compound concentration might influence the screening results. Therefore, to discover novel compounds using such screening systems, it is important to select suitable cell sources considering running costs, while conducting examinations under optimized conditions to detect the screening indices using easier methods.
The library screening using the present assay identified particularly interesting osteogenic inducible compounds, including Lef, LFM, and 1-5. Lef is a malononitrile derivative that inhibits dihydroorotate dehydrogenase and several protein tyrosine kinases [43,44]. Although Lef is an immunomodulatory agent used to treat rheumatoid arthritis [45], there have been no studies on the effects of Lef on osteoblastic differentiation. Malviya et al. reported that 15 µM Lef inhibited the proliferation of primary human osteoblasts [46]. Similarly, in the present study, 25-50 µM Lef significantly decreased the proliferation of MC3T3-E1 cells. The slight discrepancy in the inhibitory concentration of Lef on the cell proliferation likely resulted from differences in cell types, because the reactivity of primary cells/cell lines to Lef differs in the same species and among the carcinoma cell lines [46][47][48]. In this study, 1-50 µM Lef did not show cytotoxicity on MC3T3-E1 cells. Our results showed, for the first time, the positive effects of Lef on osteoblastic differentiation. Lef is associated with the Janus kinase (JAK)/signal transducer and activator of the transcription (STAT) signaling pathway [49]. The JAK/STAT pathway regulates osteogenic differentiation by activating STAT5b in osteoblasts and bone marrow-derived MSCs [50]. Although speculative, the enhanced osteogenic differentiation by Lef in the present study might be due to the activation of the JAK/STAT pathway. Further studies are required to investigate the mechanisms underlying these effects.
LFM is a potent and selective inhibitor of Bruton's tyrosine kinase (Btk) [51][52][53]. Btk suppresses the osteogenic differentiation of MC3T3-E1 cells, primary calvarial osteoblasts, and bone marrow stromal ST2 cells [54], which supports the present results. Btk regulates osteoblastic differentiation through the MAPK, NF-κB, and protein kinase C (PKC) α signaling pathways [54]. In addition, Btk is a negative regulator of Wnt-β-catenin signaling in B cells [55]. The Wnt signaling pathway plays an important role in promoting osteogenic differentiation [56]. The osteogenic effects of LFM in the present study might involve these signaling pathways because of Btk inhibition.
1-5 is an isomer of H-7 dihydrochloride, which is a potent and selective inhibitor of PKC [57]. This family comprises more than 10 isoforms that regulate apoptosis isoforms specifically [58]. Recent studies have shown that the PKC family inhibits osteogenic differentiation [59], whereas a PKC α inhibitor promotes osteogenic differentiation via the p44/42 MAPK signaling pathway [60], which may partly explain our data.
In the present study, Lef preferentially promoted the in vitro osteogenic differentiation of rMSCs compared to LFM and 1-5; therefore, the effects of Lef on in vivo bone regeneration were investigated using a rat calvarial defect model. H&E staining showed that Lef effectively repaired calvarial bone defects, which was supported by the existence of cement lines in the area of newly formed bone, indicating active bone metabolism [28]. In addition, micro-CT analysis showed that bone mass, mineral density, and volume significantly increased following Lef administration, suggesting that Lef promotes bone regeneration. Immune response and enhanced osteoclast activity caused by inflammation affect bone resorption during bone remodeling [61]. Lef is an effective compound targeting not only osteoblasts, but also osteoclasts, and inflammatory T cells because it inhibits osteoclastogenesis [62] and activates T lymphocytes [63]. In the present study, a histological analysis using TRAP staining showed that Lef did not significantly affect the number of osteoclasts. Although Lef has the potential to regulate osteoblastic and osteoclast differentiation in vivo, the concentration of Lef used in the present study preferentially affected osteogenesis rather than osteoclastogenesis during new bone formation. It is important to analyze the effects of Lef on immune cells, such as T cells, during bone regeneration.

Conclusions
The cell-based double-screening method using the Col-1a1/GFP reporter and ALP staining assays could reliably identify candidates for osteogenesis-targeting compounds compared to conventional methods. The small-molecule compounds Lef, LFM, and 1-5 detected using this screening system promoted osteogenic differentiation, with Lef particularly promoting bone regeneration in a rat calvarial defect model, highlighting this screening method as a promising tool for identifying the novel synthetic regulators of osteogenesis.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biomedicines10020426/s1, Table S1: List of compounds in the Lopack 1280 (Sigma) library, Table S2: Primers used for SYBR-based real-time quantitative RT-PCR analyses, Figure S1: Effects of identified candidate compounds on cytotoxicity and cell proliferation, Figure S2: Histological analysis for the effects of leflunomide (Lef) on osteoclast formation in the rat calvarial bone defect. Institutional Review Board Statement: The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Animal Care and Use Committee of the Osaka University Graduate School of Dentistry (approval number: 19-054).

Data Availability Statement:
The datasets generated and/or analyzed during the current study are available from the corresponding authors on reasonable request.