Therapeutic potential of human minor salivary gland epithelial progenitor cells in liver regeneration

Liver disease is a serious problem affecting millions of people with continually increasing prevalence. Stem cell therapy has become a promising treatment for liver dysfunction. We previously reported on human minor salivary gland mesenchymal stem cells (hMSGMSCs), which are highly self-renewable with multi-potent differentiation capability. In this study, keratinocyte-like cells with self-regeneration and hepatic differentiation potential were isolated and characterized, and named human minor salivary gland epithelial progenitor cells (hMSG-EpiPCs). hMSG-EpiPCs were easily obtained via minor intraoral incision; they expressed epithelial progenitor/stem cell and other tissue stem cell markers such as CD29, CD49f, cytokeratins, ABCG2, PLET-1, salivary epithelial cell markers CD44 and CD166, and the Wnt target related gene LGR5 and LGR6. The cells were induced into functional hepatocytes in vitro which expressed liver-associated markers ALB, CYP3A4, AAT, and CK18. Upon transplantation in vivo, they ameliorated severe acute liver damage in SCID mice caused by carbon tetrachloride (CCl4) injection. In a two-thirds partial hepatectomy mouse model, the transplanted cells survived at least 4 weeks and exhibited hepatic potential. These findings demonstrate that hMSG-EpiPCs have potential as a cellular therapy basis for hepatic diseases, physiological and toxicology studies and regenerative medicine.

Major salivary gland-derived epithelial cells are well-studied 11,[16][17][18][19][20][21] and have been successfully differentiated into both salivary gland and endodermal cell lineages, such as hepatocytes and pancreatic beta cells 11,22,23 . However, the major salivary glands do not seem to be an ideal and safe option for stem/progenitor cell harvest because of the invasive procedure to the donor site, whereas the minor salivary glands located in oral mucosa appear to be an optimal source with easy access and less invasion 24 . Therefore, we hypothesized that the minor salivary glands might contain epithelial stem/progenitor cells with ability to differentiate into other endodermal lineages. Previously, we reported the existence of both human minor salivary gland mesenchymal stem cells (hMSGMSCs) and epithelial progenitor cells (hMSG-EpiPCs) derived from human labial salivary glands 25,26 . This present study further details the phenotypes and characterization of the epithelial progenitor cells.

Results
Isolation and self-renewal capacity of hMSG-EpiPCs. Tissue explant method was used for primary cell culture after human minor salivary glands (hMSG) were surgically harvested. After 5-10 days of culture, keratinocyte-like cells migrated from the tissue explants (Fig. 1a). The first passage usually occurred at day 7-10, when the cells around the tissue explant reached 80%-90% confluence. Normally, we obtained approximately 5 × 10 6 homogeneously shaped cells at passage 3 (Fig. 1b). During the passaging, the cell growth and morphology did not show obvious changes until 25 passages (Fig. 1c). MTT and colony formation assays were performed for the cells at passage 5 to test their proliferation ability. The cell population doubling time was 70.85 hours on average, whereas the efficiency of colony formation was 12.3% ± 2.1%. Such results suggest the existence of a population of epithelial progenitor cells derived from hMSGs that could maintain high proliferation capacity during culture in vitro; we named the population as human minor salivary gland epithelial progenitor cells (hMSG-EpiPCs).
Identification and characterization of hMSG-EpiPCs. Flow cytometry analysis of these keratinocyte-like cells at passage 5 demonstrated that the expanded cells highly expressed the epithelial lineage-related markers CD29, CD49f, and CK-KIT (CK14, CK15, CK16, and CK19) as well as the salivary epithelial cell markers CD44 and CD166 (Table 1, Supplementary Fig. S1a). In addition, the cells were negative for the mesenchymal stem cell markers CD90, CD105, CD106, and STRO-1, indicating non-mesenchymal origin. Interestingly, in contrast to other epithelial stem/progenitor cells, these cells included a population positive for the hESC marker SOX2, suggesting consistency with other stem cell types.
The phenotypes of both hMSG-EpiPCs and hMSG tissues were determined via immunofluorescence staining ( Fig. 1d and e). Positive expression of the epithelial lineage-related markers CK15, CK19, ABCG2, and PLET-1 was observed. We also detected human telomerase reverse transcriptase (hTERT), the thymic epithelial progenitor cell markers K5 and K8, the endothelial cell marker CD31, and the basal lamina protein laminin in hMSG-EpiPCs and minor salivary gland tissues. Furthermore, the Wnt pathway target genes LGR5 and LGR6, markers of several lineages of stem/progenitor cells, including liver-derived stem cells, were also expressed. Tissue immunofluorescent staining showed expression of the epithelial lineage-related markers CK18, CK19, K5, and K8 and the Wnt target gene related LGR5 in the excretory duct and acinus; CK15-positive and LGR6-positive cells were located only in the excretory duct, indicating that these cells originated from the excretory duct of the hMSG.
These results demonstrate that hMSG-EpiPCs expressed phenotypes of epithelial stem/progenitor cells and other adult stem cell types, and suggest that these cells originated from the excretory duct part of the minor salivary glands.

Hepatic induction of hMSG-EpiPCs in vitro.
To induce hepatic differentiation of hMSG-EpiPCs in vitro, cells were cultured in 24-well plates containing thick Matrigel with conditioned medium for 16 days. hMSG-EpiPCs formed spheres in Matrigel during the induction procedure (Fig. 2a). After induction, the cells expressed a diverse panel of markers for hepatocytes including ALB, CYP3A4, and CK18 while the uninduced cells were negative (Fig. 2d). Gene expression analysis showed that expression of the hepatocellular markers ALB, CYP3A4, CK18, and AAT increased during differentiation (Fig. 2f). CYP450 activity, indocyanine green (ICG) uptake, and periodic acid-Schiff staining tests were performed to further validate function of the induced hepatocytes (Fig. 2b,c and e). The results suggested that successful hepatic differentiation was achieved in vitro, indicating the hepatic differentiation potential of hMSG-EpiPCs.
Therapeutic potential of hMSG-EpiPCs in CCl 4 -induced acute fulminant liver failure. To test whether hMSG-EpiPCs could promote liver regeneration in vivo, we performed cell transplantation in the animal model of fulminant hepatic injury. By injection of CCl 4 acute liver injury was triggered in severe combined immunodeficient (SCID) mice. The cell-treated group was transplanted with hMSG-EpiPCs, whereas the sham group was treated with PBS, and no disease model or treatment was applied in the control group. Starting at day 1 after CCl 4 injection, the body weight (BW) of the mice in the sham group kept decreasing until day 5 as a result of inflammatory hepatomegaly. Three mice (30% mortality) died 2 days after CCl 4 injection. While all mice treated with hMSG-EpiPCs survived and recovered from liver damage in a shorter time period, along with a lower ratio of liver weight to BW ( Fig. 3a and b).
Human serum albumin (HSA) was detected in the serum of treated mice 2 weeks after injection of the cells (Fig. 3c), whereas it was scarcely detected in the sham group. In addition, in both disease model groups, a dramatic increase in serum alanine transaminase (ALT), aspartate transaminase (AST), and total bilirubin (TBIL) was detected 2 days after CCl 4 injection (1day after hMSG-EpiPC/PBS treatment), indicating severe acute liver damage. ALT, AST, and TBIL levels of animals in the experimental-group decreased to baseline significantly faster than the levels in the sham group (Fig. 3d).
Histological analysis revealed that hepatocyte necrosis occurred 2 days after injection of CCl 4 in both hMSG-EpiPC-and PBS-treated groups (24 h after hMSG-EpiPC or PBS injection) (Fig. 4a). Hepatocytes presented as vacuolar, hydropic, karyorrhexic, and karyolytic, and erythrocyte diapedesis was also observed. At 8 days post CCl 4 injection (7 days post hMSG-EpiPC or PBS injection), both groups had recovered from acute liver damage, but diseased hepatocytes still could be observed in the PBS-treated group while no such cells existed in the hMSG-EpiPC-treated group, suggesting enhanced liver regeneration in the hMSG-EpiPC-treated mice.  (Table 2). No evidence of hepatic tumour formation in the hMSG-EpiPC-treated livers was found at any time point.

Discussion
The lack of an easily accessed population of human adult endodermal stem/progenitor cells with multi-differentiation potential is a challenge for cell therapy and regenerative medicine. The hMSGs locating in the oral mucosa may be promising candidates as cell source for their easy accessibility and adequate cell quantity 24 . In a previous study, we reported the existence of hMSGMSCs with phenotypes of embryonic and mesenchymal stem cells 26 . These hMSGMSCs are highly self-renewable and have multi-potent differentiation potential. The hMSG-EpiPCs we reported here were also isolated from primary cultured hMSG. However, these keratinocyte-like hMSG-EpiPCs showed a different shape and phenotype that suggested an epithelial lineage. These hMSG-EpiSCs showed high efficiency in endodermal differentiation in vitro, which inspired us to further research their therapeutic potential in liver regeneration.
In the current study, cells expressing CK18, CK19, K5, K8, and/or LGR5 were found in both the excretory duct and acinus, whereas expression of CK15 and LGR6 was observed only in the excretory duct. Thus, we propose the existence of an epithelial progenitor cell population mainly derived from the hMSG excretory duct; these oval-shaped keratinocyte-like cells were named hMSG-EpiPCs.
In contrast to mesenchymal stem cells, there are no universal standard phenotypic markers for stem/progenitor cells of epithelial-lineage. Various markers are used to mark epithelial stem/progenitor cells with various origins. Flow cytometry analysis detected high expression of the epithelial lineage-related markers CD29, CD49f, and CK-KIT in hMSG-EpiPCs, together with markers previously identified in salivary-related cells such as CD44 and CD166 27 . ABCG2, which has been reported to be positively expressed in progenitor cells isolated from mouse submandibular glands 28 , was also found to be expressed in epithelial progenitor cells isolated from hMSGs. Immunofluorescence staining revealed that hMSG-EpiPCs expressed hTERT, which exerts control over cell proliferation 29 . Additionally, the thymic epithelial progenitor cell markers K5, K8, and PLET-1 were expressed in hMSG-EpiPCs and in minor salivary gland tissue [30][31][32] . Cytokeratins are proteins found in the intracytoplasmic cytoskeleton of epithelial tissue that were previously reported to identify mouse postnatal mammary gland epithelial cells 33 , human hair follicle stem cells 34 , and human liver progenitor cells 35 . In addition, Lombaert IM previously reported that salivary gland cells isolated from murine submandibular glands were positive for CD117, while hMSG-EpiPCs were negtive 36 . These observations help to confirm the existence of hMSG-EpiPCs. In addition to these epithelial-lineage markers, hMSG-EpiPCs were also positive for the endothelial cell marker CD31, which suggest their applicability to vascular tissue engineering. Expression of the basal lamina protein laminin was also observed. In rat salivary glands, ligation of main ducts caused the disappearance of acinar cells and proliferation of ductal cells around laminin + cells, which suggests activation of tissue progenitor cells in a damaged state 24 . Cultured hMSG-EpiPCs included a population of cells positive for the hESC marker Sox2,which maintains pluripotency of hESCs and regulates formation of various epithelial tissues during foetal development 37,38 , which may explain the high proliferation ability of hMSG-EpiPCs.
In the present study, the hepatic cells differentiated from hMSG-EpiSCs expressed ALB, CK18, and CYP3A4. The expression of ALB, CK18, and CYP3A4 demonstrated fully mature hepatocyte function. In addition, Q-PCR showed that expression of AAT increased during differentiation. AAT is mainly produced in the liver and protects the lungs against proteolytic damage 47 . However, expression of AAT was not observed on immunofluorescent staining, and this result thus needs further study. Cell functionality was validated through analyses of CYP450 activity, ICG uptake, and periodic acid-Schiff staining, indicating that fully functional hepatocytes were acquired from hMSG-EpiPCs.
To further test the clinical potential of hMSG-EpiPCs, we used different animal models to validate the therapy effect. CCl 4 is a model substance for elucidating hepatotoxic mechanisms such as hepatocellular death, fibrosis, fatty degeneration, and carcinogenicity 48 . After injection of 20% CCl 4 , the levels of ALT, AST, and TBIL dramatically increased as a result of enzyme release caused by massive hepatocyte damage, indicating acute and massive death of hepatocytes, which also caused the body weight of the mice to decrease. Hepatomegaly occurred, according to the pathological sections. In the PBS-treated group, 3 mice (30% mortality) died 2 d after CCl 4 injection. The ALT, AST, and TBIL levels decreased faster in the hMSG-EpiPC-treated mice than in their PBS-treated counterparts, along with faster body weight increases, faster hepatocyte regeneration, a higher survival rate, and lower ratio of liver weight to BW. Moreover, the detection of HSA in the serum of hMSG-EpiPC-treated mice and the immunofluorescence assays result two weeks after cell injection implied proliferation and hepatic differentiation of transplanted hMSG-EpiPCs. The mouse model of two-thirds partial hepatectomy is a relatively "clean" model of hepatic damage, with no massive hepatocyte death and mild inflammation; this model does not cause inflammatorycellinfiltration or negatively impact the engraftment of transplanted stem/progenitor cells 49 . Weber et al. reported that after 70% hepatectomy, liver regeneration was activated immediately, and the weight of the remaining liver had increased to a normal level 20 d after the surgery 50 . Our result shows that after hMSG-EpiPC injection, the number of ALB/CM-Dil, CK18/CM-Dil and CYP3A4/CM-Dil double-positive cells increased daily, whereas AFP/CM-Dil double-positive cells occurred 1 week after transplantation, reached their highest level 2 weeks after transplantation, and then gradually decreased in number, which suggests both proliferation and hepatic differentiation of the transplanted hMSG-EpiPCs. Both differentiated and non-differentiated hMSG-EpiPCs expressed CK19 in vitro, which is a marker of both the minor salivary gland excretory duct and the bile duct. After injection of hMSG-EpiPCs, most of the CM-Dil-positive cells were negative for CK19; a few positive cells occurred only in the portal area. Therefore, we speculate that hMSG-EpiPCs promote liver regeneration through two major mechanisms. One is by differentiating into hepatocytes expressing ALB, AFP, CK18, and CYP3A4. The other is by migrating into the portal area and participating in the reconstruction of the intrahepaticbileduct. Therefore, hMSG-EpiPCs showed a positive effect in both disease models, indicating that they are a promising candidate for treatment of liver diseases.

Conclusion
In the current study, we isolated and characterized a population of epithelial progenitor cells from hMSGs. These cells have self-renewal and hepatic differentiation capacity. We also demonstrated the ability of these cells to ameliorate liver damage through transplantation in acute liver injury models. These cells have potential for hepatic differentiation. Our study highlights the importance of hMSG-EpiPCs in hepatic disease modelling, physiological and toxicology studies, regenerative medicine, and clinical therapy.

Method
Ethical Statement. All   Human Salivary Gland Tissue and Cell Culture. Tissues specimens containing salivary glands were transported in 15-mm centrifugetubes filled with regular medium (Dulbecco's modified Eagles medium (DMEM, HyClone), supplemented with 1% penicillin and streptomycin (5,000 UmL −1 ) (Sigma), 1% GlutaMAX supplement (Gibco), and 10% foetal bovine serum (Gibco)). When isolating hMSG-EpiPCs for primary culture, excess connective tissues were removed by scissors. The glands were cut into small pieces no larger than 1 mm 3 and placed at the base of the culture flask and then incubated for 3 to 4 h, before 5 mL of keratinocyte medium (ScienceCell) was added.
The medium was first changed at day 5, and then changed every 3 days. Cells at 80% confluence were harvested with 0.25% trypsin and re-seeded in a cell culture dish. The media were changed every 3 days. Cells were passaged using 0.25% trypsin (HyClone) at 80% confluence in a 1:3 split ratio.
MTT Assay. The MTT assay (Sigma) was performed to determine the proliferation ability of the cells. At 80% confluence, cells at passage 5 were trypsinized and transferred to a 96-well plate at a density of 2 × 10 3 per well. MTT solution was added to each well of the 96-well culture plate, which was then incubated at 37 °C for 4 h. After the MTT solution was removed, samples were dissolved in DMSO (Sigma) and transferred to new 96-well plates. The absorbance of each well was measured at a wavelength of 490 nm for 7 continuous days. The cell population doubling time was calculated as previously described 26 .
Cell Colony Formation Assay. Cells at passage 5 were trypsinized and seeded in a 60-mm dish atdensity of 200 cells and then cultured for 2 weeks. The media were changed every 3 days. The surviving colonies (>50 cells/ colony) were counted using crystal violet staining.
The cells were washed, fixed and incubated with antibodies as previously described 26 . We used a FACSAria II system (BD) for analyses. For each marker, 3 independent biological experiments were performed. FlowJo 7.6.1 software (Tree Star Inc. OR) was used to analyse the acquired data.

Transplantation Assay.
To determine whether transplanted MSG-hEpiPCs could promote liver regeneration in vivo, liver damage was produced in the mice by intraperitoneal injection of 10 mlkg −1 CCl 4 dissolved in olive oil (1:4). At 24 h after injection, 150 μL suspensions of hMSG-EpiPCs at passage 5 labeled with CM-Dil were injected via the caudal vein(approximately including 2 × 10 6 cells). The body weight of each mouse was measured, and liver tissues and serum were collected 1 day, 3 days, 1 week, and 2 weeks after hMSG-EpiPC injection for assays.
To test the proliferation of transplanted hMSG-EpiPCs in vivo, a mouse model of two-thirds partial hepatectomy was established as previously described, after which hMSG-EpiPCs at passage 5 were labelled with CM-Dil and injected immediately, as described above. The liver tissues were collected at 1 week, 2 weeks, 3 weeks, and 1 month after hMSG-EpiPC injection for immunofluorescence staining. The quantity of CM-Dil-labeled double positive cells in the immunofluorescence staining sections was measured under a Nikon TE2000-S microscope. Three researchers blinded to the groups counted the number in 15 randomized filelds on each section in 40 × magnification.
Immunohistochemistry and Immunofluorescence. Slides of hMSGs, cells, and liver tissues were fixed for 15 min in precooled 4% paraformaldehyde. The hMSG sections were stained with hematoxylin and eosin. For immunofluorescence, slides of tissues and cells were stained using standard techniques as described previously 26 . Images were acquired using a Nikon TE2000-S microscope, and an Olympus laser scanning confocal microscope.
Quantitative Reverse Transcriptase-polymerase Chain Reaction. Total RNA was obtained from cultured cells using the RNeasy Mini kit (Qiagen) and reverse-transcribed using M-MLV (Invitrogen) according to the manufacturer's protocol. Real-time PCR was performed using SYBR Green I Master (Roche) on a LightCycler 480 system (Roche Molecular Systems, CA). Expression levels of desired genes were normalized to those of GAPDH as a reference gene. The samples were collected from three independent biological replicates.