Lithium Chloride Exerts Differential Effects on Dentinogenesis and Osteogenesis in Primary Pulp Cultures

Wnt/β-catenin signaling is known to play essential roles in odontoblast differentiation and reparative dentin formation. Various Wnt activators including LiCl have been increasingly studied for their effectiveness to induce repair of the dentin-pulp complex. LiCl is a simple salt thought to activate Wnt/β-catenin signaling by inhibiting GSK3β. Previous in vitro and in vivo studies showed that LiCl increased odontoblast differentiation and enhanced reparative dentin formation. However, the underlying molecular and cellular mechanisms by which LiCl regulates odontoblast and osteoblast differentiation during reparative dentinogenesis are not well-understood. Our in vitro studies show that exposure of early dental pulp progenitors to LiCl increased the survival and the pool of αSMA+ progenitors, leading to enhanced odontoblast and osteoblast differentiation. The positive effects of LiCl in the differentiation of osteoblasts and odontoblasts from αSMA+ progenitors are mediated by Wnt/β-catenin signaling. Our results also showed that continuous and late exposure of dental pulp cells to LiCl increased the expression of odontoblast markers through Wnt/β-catenin signaling, and the number of odontoblasts expressing DMP1-Cherry and DSPP-Cerulean transgenes. However, unlike the early treatment, both continuous and late treatments decreased the expression of Bsp and the expression of BSP-GFPtpz transgene. These observations suggest that prolonged treatment with LiCl in more mature cells of the dental pulp has an inhibitory effect on osteoblast differentiation. The inhibitory effects of LiCl on osteogenesis and Bsp were not mediated through Wnt/β-catenin signaling. These observations suggest that the effects of LiCl, and GSK3β antagonists on reparative dentinogenesis involve multiple pathways and are not specific to Wnt/β-catenin signaling.

Studies in the pulp showed that the application of LiCl to injured and exposed pulp tissue induced reparative dentinogenesis (7,(19)(20)(21)(22). In vitro studies showed that LiCl increased cell migration and mineralization of human dental pulp stem cells, and increased Dspp expression in human dental pulp cells (DPCs) (hDPSCs) (19).
Although the effects of LiCl on reparative dentin and odontoblast differentiation has been established, much remains to be understood about the underlying cellular events leading to increased reparative dentinogenesis. Therefore, to gain a better . Continuous and late exposure to LiCl had no significant effect on the intensity of XO. Continuous and late exposure to DKK1 decreased the intensity of XO at days 10, 14, and 21 as compared to VH-treated and LiCl-treated cultures. Results represent mean ± SEM of at least 3 independent experiments. Analysis was performed using 1-way analysis of variance. *P ≤ 0.05 relative to VH at each time point; # P ≤ 0.05 relative to LiCl at each time point. XO, Xylenol orange; N.D, not detected.
understanding of the roles of LiCl on odontoblast and osteoblast differentiation, we used well-characterized dental pulp cultures from a series of previously characterized reporter mice (23)(24)(25) and examined the response of cells at various stages of differentiation to LiCl.

Animal Models
All animal experimental protocols were approved by the Institutional Animal Care and use Committee of UConn Health Center. DSPP-Cerulean/DMP1-Cherry, BSP-GFPtpz, pOBCol2.3FP (referred to as 2.3-GFP), αSMA-GFP and nontransgenic mice have been previously described (23)(24)(25). All mice were maintained in the CD1 or C57BL/6 backgrounds. In all experiments, pulp cells from non-transgenic littermates served as negative control.

Cell Cultures, Digital Imaging, and Epifluorescence Analysis
Primary pulp cultures were prepared from the coronal pulp of first and second molars from 5-to 7-d-old mice and grown as described in previous publications (26,27). The cultures were grown in the presence of vehicle (VH; 0.1% bovine serum albumin), various concentrations of LiCl (Sigma), and Dickkopf1 (DKK1) (R&D Systems), with or without simultaneous LiCl between days 3 and 21 (referred to as continuous exposure), days 7 and 21 (referred to as late exposure), and days 3 and 7 (referred to as early exposure).
Mineralization in live primary cultures was examined by Xylenol orange (XO) staining as described before (24). The FIGURE 3 | Effects of continuous exposure to LiCl on the expression of differentiation markers and tissue-and stage-specific reporters in primary dental pulp cultures. Primary pulp cultures were treated continuously from days 3 to 21 with VH (0.1% BSA), DKK1 (50 ng/ml), LiCl (10 mM) or simultaneous LiCl (10 mM) and DKK1 (50 ng/ml). (A-C) Graphs showing changes in the expression of Dmp1 (A), Dspp (B) and Bsp (C) in cultures treated continuously with VH, LiCl, and DKK1 alone or with simultaneous LiCl from days 3 to 17. Continuous exposure to LiCl showed significant increases in Dmp1 (A) and Dspp (B) at days 10-17, which were significantly decreased by DKK1 treatment alone. DKK1 treatment decreased the LiCl-induced increases in Dmp1 and Dspp. However, continuous LiCl treatment, as well as DKK1 treatment decreased Bsp expression. The LiCl-induced decreases were not reversed by DKK1 (C). Expression of Dmp1 and Bsp was normalized to day 7 and Dspp was normalized to day 10 of the VH-treated cultures, which is arbitrarily set to 1 and is indicated by the dashed line.  mean fluorescence intensity of XO staining, DMP1-Cherry, and BSP-GFPtpz was measured as described before (24). Background fluorescence for XO, DMP1-Cherry and BSP-GFPtpz was measured using dental pulp cultures from nontransgenic littermates, and these values were subtracted from the respective XO and transgene measurements. Live or fixed cultures from reporter mice were imaged as described before (24) using Zeiss Axio Observer Z1 inverted microscope and appropriate filter cubes optimized for the detection of various fluorescent protein variants. The exposure times remained the same throughout the each set of experiments.

Immunocytochemistry
Pulp cultures were processed for immunocytochemistry with a 1:1,000 dilution of anti-GFP Alexa Fluor 488-conjugated antibody (Molecular Probes, Invitrogen) (25). DSPP-Cerulean + cells were calculated as the ratio of cells stained with anti-GFP antibody (DSPP-Cerulean + cells) to the total number of Hoechst + cells (26,28). Primary dental pulp cultures derived from the transgenic littermates without the addition of anti-GFP antibody served as controls.

RNA Extraction and Analysis
Total RNA was prepared with TRIzol Reagent (Invitrogen) according to the manufacturer's instructions. After DNase treatment, RNA samples were processed for quantitative polymerase chain reaction (qPCR) analysis with specific primers (Supplementary Figure 1) purchased from Applied Biosystems (28). Gene expression was examined by qPCR analysis using 2 −DDcT method. For TaqMan qPCR reactions, 9 ng of cDNA was combined with 5 µl of TaqMan universal master Mix (Applied Biosystems, Nranchburg, NJ), 2.5 µl H 2 O and 0.5 µl TaqMan primers (total 10 µl). All qPCR reactions were run using Biorad PCR System under the following conditions: 50 • C for 2 min, 95 • C for 10 min, and 40 cycles with denaturation at 95 • C for 15 s and extension at 60 • C for 1 min. We defined the acceptable range of CT values representing gene expression to be between 10 and 35 cycles, according to manufacturer's recommendations.

Flow Cytometric Analysis
Single

Proliferation and Apoptosis Assays
The proliferative cells were detected by incubation of cultures with 10 µM EdU

Statistical Analysis
Analysis was performed by GraphPad Prism 8 software (GraphPad Software) using 1-way analysis of variance (ANOVA). Values in all experiments represented mean ± SEM of at least 3 independent experiments, and P ≤ 0.05 was considered statistically significant.

Effects of LiCl on Wnt Responsive Cells
We first examined the effects of LiCl on the Wnt-responsive dental pulp cells by examining the levels of Axin2 expression (29) by quantitative polymerase chain reaction (qPCR) analysis. qPCR analysis showed marked concentration-dependent increases in Axin2 expression between 4 and 24 h in the LiCl-treated cultures as compared to the VH-treated and NaCl-treated controls ( Figure 1A). Conversely, DKK1treated cultures showed concentration-dependent decreases in Axin2 expression as compared to controls ( Figure 1B). The changes with the various concentrations of LiCl and DKK1 were detectable at 4 h. Based on these observations, we selected 10 mM LiCl and 50 ng/ml DKK1 for all the following experiments.
The addition of 10 mM of LiCl to pulp cultures at days 3 and 7, induced an ∼2 to 4.5-fold increase in Axin2 levels (Figures 1C,D). DKK1 decreased these increases ( Figures 1C,D). Thus, primary pulp cultures contain Wntresponsive cells and increased expression of Axin2 by LiCl is mediated by Wnt/β-catenin signaling.

Effects of Continuous Exposure to LiCl on Primary Dental Pulp Cultures
The effects of continuous exposure to LiCl on the mineralization and differentiation of pulp cells in vitro were examined by addition of LiCl and DKK1 with or without LiCl between days 3 and 21 (Figures 2A,B).
The extent of mineralization of LiCl-treated cultures was not different from VH-treated cultures. However, continuous exposure to DKK1 markedly decreased the extent of mineralization compared to VH-treated and LiCl-treated cultures (Figures 2A,B).
Continuous LiCl treatment induced marked increases in the expression of Dmp1 and Dspp between days 10 and 17 that were decreased by DKK1 (Figures 3A,B). However, continuous treatment of the cultures with both LiCl and DKK1 resulted in similarly decreased levels of Bsp ( Figure 3C). These findings showed that LiCl increased the expression of Dmp1 and Dspp (markers of odontoblast differentiation) but decreased the expression of Bsp (a marker of osteoblasts) during the mineralization phase of growth in primary pulp cultures. The decreases in the LiCl-induced changes in Dmp1 and Dspp by DKK1 indicated the roles of Wnt/β-catenin signaling in LiClinduced changes. On the other hand, the LiCl-induced decreases in Bsp were not regulated by Wnt/β-catenin signaling.
Primary pulp cultures from 2.3-GFP, DSPP-Cerulean/DMP1-Cherry and BSP-GFPtpz transgenic reporter mice were used to examine if the effects of LiCl on gene expression were related to changes in their levels of transcription and/or changes in the number of cells expressing these markers (Figures 3D-K).
In these experiments, FACS analysis was used to quantify the percentages of cells expressing these transgenes 1-7 days after treatment (days 4-10 of culture) (Figures 3D-G) using gating strategies shown in Supplementary Figure 2.
No significant changes were observed in the percentage of cells expressing the various transgenes between days 4 and 7 in the LiCl-and DKK1-treated cultures compared to VHtreated cultures (Figures 3D-G). On the other hand, following the induction of mineralization, there were marked increases in the percentages of DMP1-Cherry + and DSPP-Cerulean + cells at days 9 and 10 in the LiCl-treated cultures. DKK1-treated cultures demonstrated significant decreases (Figures 3D,E). LiCl-and DKK1-treated cultures showed no significant changes in the percentages of BSP-GFPtpz + cells ( Figure 3F) and 2.3-GFP + cells ( Figure 3G).
The changes in the transgene expression at later time points (days 10-21) were analyzed by fluorometric analysis and immunocytochemistry (Figures 3H-K, 4A) because of the difficulties in obtaining single cells in mineralized cultures. Between days 10 and 21, LiCl-treated cultures showed continuous and marked increases in the intensity of DMP1-Cherry (Figure 3H), and in the percentage of DSPP-GFP + cells ( Figure 3I) compared to VH-treated cultures. DKK1-treated cultures showed continuous and marked decreases in the intensity of DMP1-Cherry ( Figure 3H) and in the percentage of DSPP-GFP + cells (Figure 3I). LiCl-and DKK1-treated cultures showed continuous decreases in the intensity of BSP-GFPtpz ( Figure 3J) and 2.3-GFP ( Figure 3K) compared to VH-treated cultures.

Effects of Late Exposure to LiCl on Primary Dental Pulp Cultures
Since continuous LiCl treatment demonstrated a lack of significant changes in the various differentiation markers during the first 7 days (proliferation phase), it may suggest that LiCl affects the differentiation of more mature cells. To test this possibility, we examined the effects of LiCl treatment during the mineralization of primary pulp cultures referred to as late treatment (Figures 2, 4, 5).
Late exposure of pulp cultures to LiCl (days 7-21) resulted in changes very similar to continuous treatment, albeit at a lower magnitude. LiCl did not have significant effects on the extent of mineralization (Figures 2C,D), increased the expression levels of Dmp1 and Dspp, but decreased the levels of Bsp (Figures 5A-C). Similar to continuous treatment, late treatment increased in the percentages of DMP1-Cherry + cells ( Figure 5D) and the intensity of DMP1-Cherry expression ( Figure 5H) and increased the number of DSPP-Cerulean + cells (Figures 5E,I).
Late treatment did not have significant effects in the percentage of BSP-GFPtpz + cells and 2.3-GFP + cells (Figures 5F,G) between days 7 and 10 and decreased the intensity of BSP-GFPtpz and 2.3-GFP expression (Figures 5J,K).

Effects of LiCl on Early Progenitors' Primary Dental Pulp Cultures
Our previous observations showed that αSMA + perivascular cells in dental pulp are progenitors capable of giving rise to odontoblasts and osteoblasts during reparative dentinogenesis in vivo (30). Therefore, we examined the response αSMA + perivascular cells to LiCl and DKK1 treatment.
In these experiments, the cultures from αSMA-GFP + transgenic mice were treated with LiCl and DKK1 between days 3 and 7. The effects on the percentage of αSMA-GFP + cells, their proliferation and apoptosis were examined by FACS analysis using gating strategies shown in Supplementary Figure 2  (Tables 1A-E).
LiCl-treated cultures showed markedly increased percentage of αSMA-GFP + population as compared to VH-and DKK1treated cultures (Table 1A). Furthermore, LiCl increased proliferation (Table 1B) and decreased apoptosis (Table 1C) in the αSMA-GFP + population. These changes were not detected in the αSMA-GFP − population (Tables 1D,E) Early and limited treatment with LiCl induced changes during differentiation similar to continuous and late treatment with a few important differences (Figures 6, 7). Unlike the continuous and late treatment, early treatment increased the extent of mineralization (Figures 6A,B). Early treatment also transiently increased the percentage of 2.3-GFP + cells at day 8 ( Figure 7G), unlike continuous and late treatments, which had no significant effects. The effects of early treatment with LiCl on the expression of Dmp1 (Figure 7A), Dspp (Figure 7B), percentages and intensities of DMP1-Cherry + and DSPP-GFP + cells (Figures 7D,E,H,I) were similar to the other treatments, but at lower levels. However, early treatment induced increases in the expression of Bsp and BSP-GFPtpz (Figures 7C,F,J).

DISCUSSION
Using a well-characterized primary pulp culture from various reporter mice, we investigated the underlying mechanisms regulating the stimulatory effects of LiCl on odontoblast and osteoblast differentiation. Our results showed that exposure of pulp cultures to LiCl during the proliferation phase of in vitro growth and before induction of differentiation increased the pool of αSMA + progenitors by regulating their survival. This limited and early treatment also transiently increased the number of pre-odontoblasts expressing 2.3-GFP that rapidly differentiated into odontoblasts expressing DMP1-Cherry and DSPP-Cerulean transgenes and osteoblasts expressing BSP-GFPtpz. Our results also showed that the LiCl-induced changes in the expression of Dspp, Dmp1, and Bsp in these early progenitors were reversed by DKK1. DKK1 is a secreted protein that, by binding to LRP5/6 and Kremen receptors, functions as a negative regulator of Wnt/βcatenin signaling (31). These observations indicated that the positive effects of LiCl in the differentiation of osteoblasts and odontoblasts from αSMA + progenitors are mediated by Wnt/βcatenin signaling. The effects of LiCl on early progenitors is very similar to the effects of WNT3a (27).
Our results also showed that continuous and late exposure of dental pulp cells to LiCl increased the expression of odontoblast markers through Wnt/β-catenin signaling and the number of odontoblasts expressing DMP1-Cherry and DSPP-Cerulean transgenes. However, unlike the early treatment, both continuous and late treatments decreased the expression of Bsp and the expression of BSP-GFPtpz transgene. These observations suggest that prolonged treatment by LiCl in more mature cells in the dental pulp has an inhibitory effect on osteoblast differentiation. The LiCl-induced decreases in the levels of Bsp during the mineralization phase of in vitro growth were not reversed by DKK1, indicating that the inhibitory effects of LiCl on osteogenesis and Bsp are not mediated through Wnt/β-catenin signaling. The inhibitory effects of prolonged exposure of pulp cells to LiCl are different from the stimulatory effects of WNT3a on pulp cells (27).
Although the stimulatory effects of Wnt/β-catenin signaling on osteoblast differentiation has previously been wellestablished (32), the effects of LiCl on osteogenesis remains controversial. Several studies have shown the positive effects of LiCl on osteogenic differentiation. LiCl enhanced osteogenic differentiation by bone marrow-derived mesenchymal stem (33), human periodontal ligament fibroblasts (34), and stem cells from the apical papilla (17).
In contrast, several studies reported inhibitory effects of lithium on osteogenic differentiation. LiCl inhibited the osteogenic differentiation of MC3T3-E1 pre-osteoblasts and a pluripotent mesenchymal cell line C2C12 (35,36), bone marrowderived mesenchymal stem cells (37), cementoblasts (38), SHED (39) and osteogenesis in the cultured embryonic palatal shelves Results represent mean ± SEM of at least 3 independent experiments; Analysis was performed using 1-way analysis of variance. *P ≤ 0.05 relative to VH at each time point; @ P ≤ 0.05 relative to DKK1 at each time point; # P ≤ 0.05 relative to LiCl at each time point. N.D, not detected. (40). In both MC3T3-E1 and C2C12 cell lines, LiCl inhibition of osteogenesis was mediated through BMP2/SMAD signaling pathway (35). Additionally, in MC3T3-E1 cells, the negative effects of LiCl osteogenesis were mediated through GSK3b, whereas in C2C12 cells, these effects were independent of GSK3mediated signaling (35). Furthermore, LiCl modulates the activity of several other pathways, including inhibiting inositol monophosphates (IMPase) in the phosphatidylinositol (PI) signaling pathways, with essential positive roles in osteoblast differentiation and bone formation (41,42). These observations suggest that the inhibitory effects of LiCl on osteogenesis is complex and may involve more than one pathway.
Considering the effects of LiCl on multiple signaling pathways, the results of the studies in which LiCl has been used to block the GSK3 activity should be interpreted carefully. GSK3β, a serine/threonine kinase, modulates many signaling pathways and transcription factors (43), and the effects of LiCl on reparative dentinogenesis cannot be considered specific to Wnt/β-catenin signaling.
A few studies have shown that reparative dentinogenesis induced by LiCl and small-molecule GSK3 inhibitor drugs following pulp exposure resulted in the formation of reparative dentin that was more similar to physiological dentin in tubularity and chemical composition (7,8). Our observations suggest that the improved reparative dentin structure in these studies as compared to osteodentin is related to LiCl-mediated inhibition of osteogenesis.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by Institutional Animal Care and use Committee of UConn Health Center.