The effect of fibroblast growth factor 7 on human dental pulp stem cells for differentiation to AQP5‐positive and αSMA‐positive cells in vitro and in vivo

Abstract Objectives Transplantation of stem cells into wounds has become popular in regeneration therapies. As stem cells for transplantation, human dental pulp stem cells (hDPSCs) are known to be pluripotent cells that are relatively easy to collect from the pulp of deciduous or wisdom teeth. The purpose of this study was to investigate whether hDPSCs treated with fibroblast growth factor 7 (FGF7) would contribute to the regeneration of wounded rat submandibular glands (SMGs). Materials and methods In in vitro studies, hDPSCs were treated with or without FGF7 and mRNA expression levels were examined at days 3, 7 and 14 using qRT‐PCR. The target genes analyzed were BMI1 as an undifferentiated marker, AQP5 as an acinar cell marker, CK19 as a ductal epithelial cell marker, αSMA as a myoepithelial cell marker and VIMENTIN as a fibroblast marker. In in vivo studies, hDPSCs treated with or without FGF7 for 14 days were mixed with type I collagen gels and were transplanted into wounded rat SMGs. Hematoxylin–Eosin and immunohistochemical staining were performed at days 3 and 7, and the numbers of positive cells were counted. The primary antibodies used were against BMI1, AQP5, αSMA, PanCK and VIMENTIN. Results In the in vitro studies, mRNA levels of BMI1 were decreased and αSMA were increased at days 3, 7 and 14, while AQP5 was increased at day 14 in the FGF7 group. In the in vivo studies, the proliferation of hDPSCs and cell islands was observed at day 7 in the FGF7 group. Few BMI1‐positive cells were observed, while numbers of AQP5‐positive and αSMA‐positive cells were increased at days 3 and 7 in the FGF7 group. Moreover, cell islands were AQP5‐positive. Conclusion These results suggest that FGF7‐treated hDPSCs differentiate into AQP5‐positive and αSMA‐positive cells. Moreover, AQP5‐positive cell aggregations were formed.


| INTRODUCTION
The secretion of saliva is related to swallowing, pronunciation, lubrication, antibacterial action, buffering action, taste, and so on (Carlson & Ord, 2008;Vissink, Burlage, Spipijkervet, Jansma, & Coppes, 2003;Vissink, Jansma, Spijkervet, Burlage, & Coppes, 2003). Salivary glands that secrete saliva are tissues that are relatively susceptible to damage from trauma, injury, radiation, drug injury, Sjögren syndrome and so on, and such damage often causes a decrease in salivary secretion, which can significantly reduce the patient's quality of life (Atkinson, Grisius, & Massey, 2005;Fox, 1998;Ship, Pillemer, & Baum, 2002). Generally, salivary glands that are already functionally differentiated are classified as tissues with low regenerative capacity. However, currently, it has been reported that acinar cells regenerate salivary glands by self-renewal (Aure, Konieczny, & Ovitt, 2015) and it has also been reported that stem/progenitor cells present in ductal cells contribute only to ductal cells, not to acinar cells (Kwak, Alston, & Ghazizadeh, 2016). Moreover, there is also a report of salivary gland regeneration using stem cells (Tanaka et al., 2018) and research in this area is being actively conducted. Therefore, research studies on the regeneration of damaged salivary glands are very important and informative.
In the case of salivary gland wound healing, non-specific granulation tissue, which consists of capillaries and fibroblasts, can appear and eventually forms scar tissue that can disrupt salivary gland function (Man, Ball, Marchetti, & Hand, 2001;Takahashi, Schoch, & Walker, 1998). To avoid scar tissue formation in wounded salivary glands, granulation tissue from surrounding areas must be excluded using barrier membrane techniques and stem cells need to be induced in the wound areas for regeneration. However, only a few stem cells are present in salivary glands as noted above. Therefore, tissue engineering techniques in which three elements, cells, scaffolds and growth factors, are required for the regeneration of salivary glands.
It has been reported that regenerative treatments include the transplantation of stem cells into wounds, and mesenchymal stem cells represented by adipose stem cells and bone marrow stem cells are used for salivary gland regeneration (Elsaadany, Zakaria, & Mousa, 2019;Kim et al., 2019). However, human dental pulp stem cells (hDPSCs) are known to be pluripotent cells that can be collected relatively easily from deciduous teeth and from wisdom teeth, but there are no reports of using them for transplantation. We hypothesized that the use of hDPSCs as stem cells would be an effective approach for salivary gland regeneration following salivary gland injury.
There have been many reports that type I collagen gels play important roles in cellular proliferation and differentiation in vitro (Auger et al., 1998;Gospodarowicz, Greenburg, & Birdwell, 1978;Lombaert et al., 2008;Murray, Stingl, Kleinman, Martin, & Katz, 1979;Parry & Creamer, 1980;Sattler, Michalopoulos, Sattler, & Pitot, 1978;Wicha, Liotta, Garbisa, & Kidwell, 1979) and in vivo (Nanduri et al., 2014;Yang et al., 2014). It has also been reported that various types of cells, including stem cells, can easily migrate into type I collagen gels (Kobayashi, Matsuzaka, & Inoue, 2016;. Kobayashi et al. (2016) reported that stem cells migrated into type I collagen gels containing bFGF in wounded salivary glands. Further, they reported that specific proteins that are related to salivary gland components were expressed but the glands did not regenerate structurally, and they concluded that other growth factor(s) might be necessary to stimulate structural regeneration.
In this study, the effects and behavior of transplanting type I collagen gels containing hDPSCs treated with FGF7 into wounded rat SMGs was investigated in vitro and in vivo. Those cells are primary cultures that express CD105, CD166, CD29, CD90 and CD73, but are negative for CD133, CD34 and CD45.
Those hDPSCs have the phenotypic and functional characteristics of bone marrow-derived mesenchymal stem cells.

| Cell culture medium
The dental pulp stem cell growth medium (DPSC-GM) was prepared
This FGF7 stock solution was diluted 100-fold in the DPSC-GM described above to a final concentration of 100 ng/mL.

| Cell culture
hDPSCs were seeded in 100-mm dishes at a cell density of 5000 cells/cm 2 (275,000 cells/dish) and were cultured in DPSC-GM that was changed every 3 days at a temperature of 37 C with 5% carbon dioxide. hDPSCs were detached using Trypsin/EDTA and were passaged.
When hDPSCs reached confluence after 2 passages, the medium was changed to DPSC-GM containing FGF7, and this group was designated as the FGF7 group ( Figure 1). The control group of DPSCs continued to use DPSC-GM but without FGF7. The day when the culture medium was changed is defined as day 0.

| Proliferation assay of hDPSCs
The hDPSCs were seeded in 24 well dishes at a cell density of 5000 cells/cm 2 (10,000 cells/well) and were collected at days 1, 3, 5, 7, 9, 11 and 13 in each group using a cell scraper. The hDPSCs were counted using a Coulter Counter (Beckman Coulter, California) and the number of cells was plotted.
2.6 | Quantitative reverse transcription polymerase chain reaction (qRT-PCR) hDPSCs at days 3, 7 and 14 in each group were collected using a cell scraper. Total RNAs were extracted using a RNeasy Mini Kit (QIAGEN, Hilden, Germany) and were reverse transcribed into complementary DNAs (cDNAs) using ReverTra Ace qPCR RT Master Mix with gDNA Remover (TOYOBO, Osaka, Japan). qRT-PCR was performed using TaqMan Gene Expression Assays (Applied Biosystems, Massachusetts).

| Animals
This study was conducted in compliance with the Guidelines for the

| Preparation of collagen gels
An acid-soluble type I collagen solution derived from porcine tendons was prepared to a concentration of 3.0 mg/mL at a pH of 3.0 (Cellmatrix Type I-A; Nitta Gelatin, Osaka, Japan) for use in this study. This solution was quickly reconstituted into collagen gels with consistently high strength and transparency required for use. 2.9 | Preparation of hDPSCs in collagen gels hDPSCs cultured with or without FGF7 for 14 days were used in the in vivo study. Sheets of hDPSCs were collected using a cell scraper and were separated into single cells by pipetting. The supernatant was removed after centrifugation and a collagen gel solution at a concentration of 100,000 cells/50 μg was added to the remaining hDPSCs.
The collagen gel solution containing hDPSCs, which was in a liquid condition, was reconstituted by incubation for 30 min in an incubator at 37 C to produce a gel and was used as a graft material.

| Wounded rat SMGs and transplantation of hDPSCs
The in vivo experiment was performed according to the method of Kobayashi et al. (Kobayashi et al., 2016). Each rat was sedated with isoflurane and was then anesthetized by an intraperitoneal injection of three mixed anesthetics (0.5 mL/100 g). The front of the neck of each rat was shaved and an approximately 2 cm incision was made in the skin at the midline of the neck with a surgical knife. The salivary glands were exposed, and defects were created using a 3 mm biopsy punch (Kai Industries, Gifu, Japan) without damaging the main artery or the main salivary duct (Figure 2a,b). The wounded area was covered by a Millipore membrane filter (MF-Millipore, 0.45 μm pore size, HAWP01300; Merck Millipore, Massachusetts) to avoid the invasion of cells from the remaining salivary gland tissue. Each Millipore membrane was cut in advance to an appropriate size, made into a cylindrical shape and inserted into the wounded area. After insertion, the membrane was carefully spread to the wound margin ( Figure 2c).
The wound area lined with a Millipore membrane filter was then filled with collagen gel mixed with hDPSCs treated with FGF7 (FGF7 group) or without FGF7 (Control group) using an injection syringe ( Figure 2d). In addition, Millipore membrane filters were placed on the top and bottom of each wound to avoid the overflow of gels ( Figure 2e). The skin was then sutured using 4-0 nylon thread.

| Histological observations
Three rats in each group were euthanized using an overdose of somnopentyl at days 3 and 7 after the surgery (n = 12

| Statistical analysis
Quantitative data are presented as means ± standard deviation (SD) and were analyzed using two-way ANOVA in GraphPad Prism7 (MDF Corporation, Tokyo, Japan). Differences with a p value <.05 are considered statistically significant.

| Proliferation of hDPSCs
The proliferation of hDPSCs treated with or without FGF7 increased day by day until day 7, after which the cell numbers plateaued in both groups. There was no significant difference in proliferation between the FGF7 group and the control group (Figure 3).

| qRT-PCR analysis of hDPSCs
The expression level of BMI1 mRNA was significantly lower in the FGF7 group compared to the control group at days 3, 7 and 14 ( Figure 4a). The expression level of AQP5 mRNA was not significantly different at days 3 and 7, however it was significantly higher in the FGF7 group than in the control group at day 14 ( Figure 4b). The expression level of αSMA mRNA was significantly higher in the FGF7 group compared to the control group at days 3, 7 and 14 ( Figure 4c).
The expression level of CK19 mRNA was not significantly different in those groups at days 3, 7 and 14 ( Figure 4d). The expression level of VIMENTIN mRNA also did not change significantly at days 3 and 7 but was significantly lower in the FGF7 group compared to the control group at day 14 ( Figure 4e).

| Histological observations
The hDPSCs proliferated in collagen gels in both the FGF7 and the control groups over time ( Figure 5). Aggregations of cells and small island structures were sometimes observed at day 7 in the FGF7 group but were never seen in the control group.

| DISCUSSION
Generally, when salivary gland tissues are damaged and atrophy, degenerate, undergo necrosis and lose their original functions, nonspecific fibroblasts usually migrate into the damaged area and eventually scar tissue is formed. Only in cases when tissue stem cells migrate into the damaged area, incomplete regeneration might occur. Thus, it is important to avoid the migration of non-specific fibroblasts from the surrounding tissue. To avoid this influx of contaminating fibroblasts, barrier membranes, such as Gore-tex membranes, have been used to cover the defect areas. This barrier membrane technique is usually used in periodontal surgery and is called guided tissue regeneration (GTR) (Nyman, Lindhe, Karring, & Rylander, 1982). In this study, Millipore membrane filters that have a proper pore size and readily transport tissue fluids, including growth factors, were used as a barrier membrane instead of Goretex membranes, which are waterproof and do not transport some types of molecules. This prevents cell migration but allows the transport of many kinds of growth factors from the wounded remaining salivary glands of the experimentally created defects. The results of this study showed that no capillaries or elongations of duct structures were observed and that transplanted hDPSCs could proliferate in the defect areas. These results suggest that this F I G U R E 4 mRNA expression levels. qRT-PCR analysis showed that the mRNA expression levels of cultured hDPSCs in vitro were as follows; BMI1 was significantly lower at days 3, 7 and 14 (a), AQP5 was significantly higher at day 14 (b), αSMA was significantly higher at days 3, 7 and 14 (c), CK19 was not significantly different at days 3, 7 and 14 (d) in the FGF7 group compared to the control group. VIMENTIN was significantly lower at day 14 (e) in the FGF7 group compared to the control group. Bars: means ± standard deviation. Significance levels: *p < .05 F I G U R E 5 Hematoxylin-eosin stain. Wounded areas in rat SMGs assessed by hematoxylin-eosin staining in the in vivo study. hDPSCs proliferated in collagen gels in both groups, and the number of cells increased with time. Aggregations of cells were observed at day 7 in the FGF7 group. Original magnification; ×400. Scale bars: 100 μm method is useful to investigate the transplanted hDPSCs treated with growth factors from the remaining salivary gland tissue.
In this study, the effects of hDPSCs treated with FGF7 on wound healing of rat SMGs were evaluated in vitro and in vivo.
FGF7 is a member of the FGF family, which plays important roles in regulating the proliferation and differentiation of a wide range of cells (Morita & Nogawa, 1999). Currently, 22 distinct FGF proteins have been identified in humans, all of which are known as signaling molecules with structural similarities (Blaber, Disalvo, & Thomas, 1996;Ornitz & Itoh, 2001). It is known that FGF1 to The results demonstrated that the expression levels of αSMA mRNA increased significantly in the FGF7 group compared to the control group at all time periods. Furthermore, αSMA-positive cells were significantly increased in the FGF7 group compared with the control group at all time periods. αSMA is a cytoskeletal protein of smooth muscle cells and is a myoepithelial cell marker. This suggests that FGF7 causes hDPSCs to differentiate into αSMA-positive cells, which could be myoepithelial cells.
In addition, AQP5 mRNA expression levels increased more in the FGF7 group at day 14 than in the control group in the in vitro study.
Furthermore, the number of AQP5-positive cells increased in the FGF7 group compared with the control group in the in vivo study (Ma et al., 1999). AQPs are proteins with pores that are present in acinar cell membranes (Agre, 2006). Since AQP5 can selectively pass only water molecules, it is involved in regulating water uptake into cells (Gonen & Walz, 2006). There are currently 13 known types of AQPs, and AQP5 is found in salivary glands (Cotroneo, Proctor, Paterson, & Carpenter, 2008;Nielsen et al., 2002) and is used as a marker of acinar cells. This suggests that FGF7 affects the differentiation of hDPSCs not only to αSMA-positive cells, but also to AQP5-positive cells, which could be acinar cells.
There F I G U R E 7 The number of immunopositive cells. Immunohistochemical staining showed that the numbers of immunohistochemically positive cells of hDPSCs in the in vitro study were as follows; at days 3 and 7, few BMI1-and PanCK-positive cells were observed in both groups (a, d), the numbers of AQP5-and αSMA-positive cells were significantly higher in the FGF7 group compared to the control group (b, c), the number of VIMENTIN-positive cells was significantly lower in the FGF7 group compared to the control group (e). Bars: means ± standard deviation. Significance levels: *p < .05

| CONCLUSION
These results suggest that hDPSCs treated with FGF7 differentiate into AQP5-positive and αSMA-positive cells, and form AQP5-positive cell aggregates when transplanted into experimentally damaged rat SMGs.

ACKNOWLEDGMENTS
We would like to thank members of the Department of Pathology for their technical assistance.