Interferon-gamma improves impaired dentinogenic and immunosuppressive functions of irreversible pulpitis-derived human dental pulp stem cells

Clinically, irreversible pulpitis is treated by the complete removal of pulp tissue followed by replacement with artificial materials. There is considered to be a high potential for autologous transplantation of human dental pulp stem cells (DPSCs) in endodontic treatment. The usefulness of DPSCs isolated from healthy teeth is limited. However, DPSCs isolated from diseased teeth with irreversible pulpitis (IP-DPSCs) are considered to be suitable for dentin/pulp regeneration. In this study, we examined the stem cell potency of IP-DPSCs. In comparison with healthy DPSCs, IP-DPSCs expressed lower colony-forming capacity, population-doubling rate, cell proliferation, multipotency, in vivo dentin regeneration, and immunosuppressive activity, suggesting that intact IP-DPSCs may be inadequate for dentin/pulp regeneration. Therefore, we attempted to improve the impaired in vivo dentin regeneration and in vitro immunosuppressive functions of IP-DPSCs to enable dentin/pulp regeneration. Interferon gamma (IFN-γ) treatment enhanced in vivo dentin regeneration and in vitro T cell suppression of IP-DPSCs, whereas treatment with tumor necrosis factor alpha did not. Therefore, these findings suggest that IFN-γ may be a feasible modulator to improve the functions of impaired IP-DPSCs, suggesting that autologous transplantation of IFN-γ-accelerated IP-DPSCs might be a promising new therapeutic strategy for dentin/pulp tissue engineering in future endodontic treatment.

regeneration, and immunosuppression 6,7,9 . However, the properties of DPSCs isolated from irreversible pulpitis tissue, IP-DPSCs, have not yet been fully revealed 11,12,14 . To closely examine the properties of IP-DPSCs, at the beginning of this study, cells were isolated from fresh irreversible pulpitis tissue with a standard CFU-F method 15 . Single cells were independently attached to the plastic culture dishes and then formed cell clusters and CFU-Fs ( Fig. 1c) with different sizes and varied density (Fig. 1d), but the IP-DPSCs showed significantly lower colony-forming efficacy when compared with healthy DPSCs (Fig. 1e). Flow cytometric analysis demonstrated that IP-DPSCs were positive to STRO-1, CD146, CD105, CD73, and CD90, but negative to hematopoietic cell markers CD34, CD45, and CD14, as seen in healthy DPSCs (Fig. 1f). Reverse transcription polymerase chain reaction (RT-PCR) demonstrated that IP-DPSCs expressed genes for embryonic stem cells, NANOG and octamer 4, and for neural crest cells, NESTIN, NOTCH1, and CD271 (Fig. 1g). IP-DPSCs expressed a markedly reduced CD271 when compared with healthy DPSCs. By population-doubling and bromodeoxyuridine (BrdU) assays, IP-DPSCs showed a significantly suppressed proliferation capacity when compared with healthy DPSCs (Fig. 1h-j). IP-DPSCs expressed lower telomerase activity than healthy DPSCs (Fig. 1k). Therefore, these data suggested that IP-DPSCs retained stemness as MSCs, but expressed different essential characteristics when compared with healthy DPSCs.

Multipotency of IP-DPSCs.
When IP-DPSCs were cultured under a dentinogenic condition for 4 weeks, they were capable of forming calcium-deposited nodules positive to Alizarin Red staining (Fig. 2a,b). RT-PCR and quantitative RT-PCR (RT-qPCR) demonstrated that IP-DPSCs at 1 week after the induction expressed odontoblast-specific genes including runt-related gene 2 (Runx2), alkaline phosphatase (ALP), osteocalcin (OCN), and dentin sialophosphoprotein (DSPP) (Fig. 2c, Supplementary Figure 1a) and showed ALP activity (data not shown). IP-DPSCs were then cultured under an adipogenic condition for 6 weeks. Oil Red O staining assay demonstrated lipid accumulation in IP-DPSCs (Fig. 2d,e). RT-PCR and RT-qPCR confirmed the expression of adipocyte-specific genes, including lipoprotein lipase (LPL) and peroxisome proliferator-activated receptor γ 2 (PPARγ 2) (Fig. 2f, Supplementary Figure 1b). Immunofluorescence showed the expression of CD31 on IP-DPSCs that were cultured under an endothelial cell differentiation condition for seven days (Fig. 2g,h). We also treated IP-DPSCs under a neural cell induction condition for 7 days. Immunofluorescent analysis detected the expression of glial fibrillary acidic protein (GFAP), neurofilament M (NF-M), and tubulin β III (β III) on IP-DPSCs (Fig. 2i,j). IP-DPSCs also showed a reduced capability to differentiate into odontoblasts, adipocytes, endothelial cells, and neural cells in comparison with healthy DPSCs (Fig. 2).  Figure 2). Histological analysis showed that dentin/pulp complex-like structures were found in the implant tissues (Fig. 3a,b). This was similar to the findings of previous studies 6,7 . Immunofluorescence showed that human CD146-or DSPP-positive cells were arranged along the surface of the de novo mineralized matrix (Fig. 3c,d). In comparison, control transplants that received implantation of HA/TCP alone (without IP-DPSCs) did not show any mineralized tissue or human CD146 antibody-positive cells (data not shown). However, IP-DPSCs exhibited a reduced capacity for in vivo dentin/ pulp complex regeneration in comparison with healthy DPSCs (Fig. 3e).
Self-renewal capability of IP-DPSCs. Sequential transplantation is the traditional and gold standard method to determine the self-renewal capacity of stem cells, including DPSCs 6-8 (Supplementary Figure 2). IP-DPSCs were first transplanted with HA/TCP placed under the dorsal skin of immunocompromised mice. Cells were isolated from the primary transplants 8 weeks after the primary implantation, and were then transplanted under the dorsal skin of different immunocompromised mice for a further 8 weeks. Histological investigation revealed that de novo structures in the secondary transplants expressed dentin/pulp complex-like structures in the primary transplants (Fig. 3f,g). Human mitochondria-or DSPP-positive cells were arranged on the de novo mineralized matrix in the secondary transplants (Fig. 3h,i). Population-doubling (Fig. 1h) and telomerase activity ( Fig. 1k) in IP-DPSCs were associated with a self-renewal potential of stem cells 17 . Collectively, these results verified that IP-DPSCs have a self-renewal capacity.
Heterogeneity of IP-DPSCs. Heterogeneity in MSCs 18,19 is one of the unique characteristics of healthy DPSCs 7 . To examine heterogeneity in IP-DPSCs, a total of 12 clonogenic single colonies were obtained from Immunosuppressive function of IP-DPSCs. Healthy DPSCs exhibit T cell suppression 9 . Different numbers of γ -irradiated IP-DPSCs were directly co-cultured with human peripheral blood mononuclear cells (PBMNCs) stimulated with or without concanavalin A (ConA) (10 μ g/ml) (Supplementary Figure 3a). Under a ConA-free condition, any number of IP-DPSCs did not affect the cell viability of PBMNCs (Fig. 4a). In comparison, under a ConA-stimulated condition, the PBMNC viability was suppressed depending on the cell number of IP-DPSCs (Fig. 4b). However, IP-DPSCs exhibited a reduced suppressive function in response to ConA-activated PBMNCs when compared with healthy DPSCs (Fig. 4b). Next, PBMSCs were co-cultured indirectly with IP-DPSCs or healthy DPSCs in a transwell culture system (Supplementary Figure 3). Under concanavalin A (ConA)-free condition, PBMNCs exhibited similar cell viability between direct and indirect co-culture systems (Supplementary Figure 4a). However, under a ConA-stimulated condition, the PBMNC viability did not showed any reduction depending on the cell number of IP-DPSCs in the traswell system (Supplementary Figure 4b).

Cell survival of IP-DPSCs. Activated T cells induce apoptosis of MSCs through the Fas/FasL pathway
When IP-DPSCs and healthy DPSCs were co-cultured with anti-CD3 antibody-activated PBMNCs for 3 days (Supplementary Figure 5b), the IP-DPSCs were also induced into cell death (Fig. 4g). However, the treatment with anti-Fas antibody inhibited cell death of the IP-DPSCs (Fig. 4g). The activated PBMNCs induced terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL)-positive cells in IP-DPSCs one day after the co-culture (Fig. 4h). The number of TUNEL-positive cells was greater in IP-DPSCs than in healthy DPSCs (Fig. 4i).
IFN-γ treatment improved dentinogenic dysfunction of IP-DPSCs. TNF-α and IFN-γ have been known to participate in the immunomodulatory and dentinogenic functions of healthy DPSCs 9,13 . Here, we investigated whether TNF-α or IFN-γ improve the in vitro or in vivo impaired dentin formation of IP-DPSCs. IFN-γ -stimulated IP-DPSCs were capable of forming a significant number of calcium-deposited nodules positive to Alizarin Red staining 4 weeks after the induction (Fig. 5a,b). IFN-γ -stimulated IP-DPSCs showed higher ALP activity (Fig. 5c) and markedly upregulated the expression of odontoblast/osteoblast-specific genes, including Runx2, ALP, osteocalcin, and DSPP (Fig. 5d,e) 1 week after the induction. On the other hand, TNF-α -treated IP-DPSCs showed a markedly reduced dentinogenic capacity in vitro when compared with non-stimulated (intact) and IFN-γ -stimulated IP-DPSCs (Fig. 5a-e).
Next, TNF-α -or IFN-γ -pretreated IP-DPSCs were subcutaneously transplanted with HA/TCP into immunocompromised mice. Histological analysis demonstrated that IFN-γ -treatment enhanced the in vivo capability of dentin/pulp complex-like structures in the implant tissues ( Fig. 5f,g). On the other hand, TNF-α -treatment reduced the in vivo capability (Fig. 5f,g). To further examine the effect of TNF-α and IFN-γ on IP-DPSC-mediated dentin/pulp regeneration, TNF-α and IFN-γ -treated IP-DPSCs were transplanted into human tooth root canals (Supplementary Figure 6). The IP-DPSCs formed a dentin-like structure directly on the surface of existing human tooth root dentin (Fig. 5h). IFN-γ -treatment abundantly deposited de novo dentin on a pre-existing human dentin surface, whereas TNF-α treatment did not (Fig. 5h). The newly formed dentin-like structure did not contain dentinal tubules, and the charged cells arranged on its surface and embedded within it. Immunohistochemical assay with anti-human mitochondria antibody revealed that the lining and embedded cells were of human origin (Fig. 5i). These structural findings indicated that IP-DPSC-mediated dentin/pulp complex regeneration on human dentin may be controlled under a similar mechanism as reparative dentin formation occurring in the physiological human dentin/pulp system, suggesting that IFN-γ treatment improved the impaired dentinogenic function of IP-DPSCs.
Next, we examined the effects of TNF-α and IFN-γ on the immunosuppressive functions of IP-DPSCs under co-culture with ConA-stimulated PBMNCs. RT-qPCR analysis and L-kynurenine production assay demonstrated that IFN-γ enhanced IDO mRNA expression and L-kynurenine production in IP-DPSCs co-cultured with ConA-stimulated PBMNCs, whereas TNF-α did not (Fig. 6d,e). Similarly, ELISA showed that IFN-γ induced IL-10 production, whereas TNF-α did not (Fig. 6f). We then co-cultured a greater number of IP-DPSCs with ConA-stimulated PBMNCs under treatment with TNF-α and IFN-γ . The cell viability of the PBMNCs was markedly suppressed when the IP-DPSCs were under treatment with TNF-α or IFN-γ (Fig. 6g). IFN-γ -mediated suppression was more effective than TNF-α -mediated inhibition (Fig. 6g). PBMNC suppression through TNF-α -and IFN-γ -stimulated IP-DPSCs was significantly inhibited by 1-MT and anti-IL-10 antibody treatments (Fig. 6g). These findings suggested that IFN-γ treatment may restore the impaired immunoregulatory function of IP-DPSCs.

Discussion
The present isolation approach reveals the existence of CFU-F clonal populations 15 in human dental pulp tissues diagnosed with irreversible pulpitis. The isolated CFU-Fs also express multipotency and immunological features as MSCs. Moreover, the present sequential transplantation and single colony assays demonstrated that the isolated CFU-Fs express an in vivo dentin/pulp complex regeneration ability, self-renewal capacity, and heterogeneity. Overall, these findings indicate that the present CFU-Fs isolated from irreversible pulpitis tissue display stem Scientific RepoRts | 6:19286 | DOI: 10.1038/srep19286 cell characteristics similar to MSCs and healthy DPSCs 6-8 . Furthermore, IP-DPSCs also demonstrated an immunosuppressive function similar to healthy DPSCs 9 . However, when compared with healthy DPSCs, the IP-DPSCs showed impaired colony-forming capacity, cell proliferation rate, multipotency, in vivo dentin regeneration and immunosuppressive functions.
Recent tissue engineering technology has shown that healthy DPSCs reconstruct a dentin/pulp-like structure on human dentin in a xenograft system using immunocompromised mice 8,27 . In dogs, autologous transplantation of DPSCs from freshly extracted teeth successively regenerates a vascularized dental pulp tissue in pulpectomized teeth 28 . The currently innovated preclinical grade of healthy human DPSCs combined with granulocyte colony-stimulating factor suggests that autologous healthy DPSC-based therapy may almost be ready for clinical application 29,30 . However, the opportunity to obtain healthy donor teeth is quietly limited at a general clinical situation. Therefore, IP-DPSCs isolated from diseased teeth might be a feasible source for dentin/pulp complex regeneration 11,12 . The present IFN-γ -mediated IP-DPSCs facilitated the regeneration of dysfunctional dentin/ pulp and the immunosuppressive function of IP-DPSCs will aid in the development of autologous DPSC-based endodontic therapy.
Scientific RepoRts | 6:19286 | DOI: 10.1038/srep19286 Inflammatory pulp reactions including reversible and irreversible pulpitis are carried out by several etiologies/stimuli 5 . Human dental pulp tissue with irreversible pulpitis expresses high levels of TNF-α and IFN-γ 4 . Major inflammatory helper T cell cytokines, TNF-α and IFN-γ , have been known to affect the immunomodulatory and dentinogenic functions of healthy DPSCs 9,13,31-34 . However, the effects of TNF-α and IFN-γ on the immunomodulatory and dentinogenic capabilities of IP-DSPCs have not yet been clarified. Recent studies of short-term TNF-α exposure-enhanced dentinogenic activity in healthy DPSCs suggest that reparative dentin formation occurs in response to reversible pulpitis 32,35,36 . Meanwhile, prolonged and heavy exposure of TNF-α diminishes the mineralization potential of IP-DPSCs and healthy DPSCs 13,37 . In comparison, IFN-γ accelerates the osteoblastic differentiation of human MSCs in vitro and in vivo 38 . IFN-γ -treated bone marrow-derived mesenchymal stem cells (BMMSCs) and DPSCs also accelerated IDO production to enhance the suppressive effect on T cell proliferation 9,39,40 . The present IFN-γ treatment successfully recovered the dysfunction of IP-DPSC-mediated dentin regeneration in vitro and in vivo and the suppression of T cells via IDO, whereas TNF-α treatment did not. These data suggest that exogenous IFN-γ treatment to IP-DPSCs is a novel approach to improve its ability of dentin/pulp complex regeneration and immunosuppression.
BMMSCs derived from patients with systemic lupus erythematosus, which have impaired bone forming potential, express enhanced phosphorylation of NF-κ B under accelerated TNF-α signaling 41 . The inhibition of NF-κ B activation induces the osteogenic capacity of normal BMMSCs 42 . In healthy DPSCs, suppression of activated NF-κ B also enhances odontogenic ability under TNF-α stimulation 31,43 . In the present study, TNF-α treatment caused NF-κ B activation with improved dentin/pulp complex regeneration and T cell immunosuppression in IP-DPSCs 13,14 . On the other hand, IFN-γ treatment enhanced the ability for IP-DPSCs by activation of NF-κ B independently. These findings suggest that a practical control for NF-κ B activation may enable regulation of the impaired functions of IP-DPSCs, including dentin/pulp regeneration and T cell immunosuppression.
The ectopic human telomerase reverse transcriptase (TERT) gene in human BMMSCs induces bone formation in vitro and in vivo by enhancing a critical master transcription factor for osteoblast differentiation, Runx2 44,45 . TERT modulates FasL expression on BMMSCs to induce T cell apoptosis 46 . Recently, the non-steroidal anti-inflammatory drug aspirin (acetylsalicylic acid) has been shown to enhance telomerase activity and stimulate bone formation in vitro and in vivo as well as induce T cell apoptosis in BMMSCs and stem cells derived from human deciduous teeth 25,47,48 . In this study, IFN-γ -treated telomerase activity in IP-DPSCs facilitated in vitro and in vivo dentin formation and T cell suppression, whereas TNF-α treatment in IP-DPSCs did not. Therefore, regulation of telomerase activity in IP-DPSCs may accelerate their impaired dentin/pulp regeneration and immunosuppressive functions.
In conclusion, the present findings indicated that IFN-γ treatment improved the impaired IP-DPSC functions of dentin/pulp regeneration and immunosuppressive regulation. These effects were modified by telomerase activity, but independent of the NF-κ B pathway. This functional gain study of disease-derived DPSCs may contribute to achieving not only autologous stem cell-approached pulp tissue engineering, but also to develop novel pulp-capping agents targeted to recipient diseased stem cells in future endodontic therapy (Supplementary Figure 7).

Methods
Ethics statement and human participants. All human samples were obtained as discarded biological/ clinical samples from systemically healthy donors (19-25 years old) in Kyushu University Hospital. Extracted dental pulp tissue (n = 4) from permanent teeth that were diagnosed with irreversible pulpitis was aseptically collected under pulpectomy. Healthy pulp tissue (n = 3) was obtained from participants with healthy impacted third molars (19-23 years old). Human PBMNCs were provided by healthy volunteers (26-38 years old). The procedures for using human samples were conducted in accordance with the Declaration of Helsinki and approved by the Kyushu University Institutional Review Board for Human Genome/Gene Research (Protocol Number: 393-01). Written informed consent was obtained from all subjects. All experimental protocols were approved by the Institutional Animal Care and Use Committee of Kyushu University (Protocol Number: A21-044-1). The methods were carried out in accordance with the approved guidelines.
Histology of dental pulp tissue of human teeth diagnosed with irreversible pulpitis. Histological analysis with hematoxylin and eosin (H&E) staining and immunohistochemistry for human dental pulp tissue was performed as described in Supplementary Methods 49 . Mouse anti-STRO-1 IgM antibody (R&D Systems, Minneapolis, MN, USA) and non-immune mouse IgM (R&D Systems) were used for immunohistochemical analysis.
Isolation and culture of DPSCs. Stem cells from the dental pulp tissue of human permanent teeth were isolated as described in Supplementary Methods 7,50 . Briefly, the dental pulp tissue was digested with 0.3% collagenase type I (Worthington Biochemicals, Lakewood, NJ, USA) and 0.4% dispase II (Sanko Junyaku, Tokyo, Japan) for 60 min at 37 °C. The single cells obtained were seeded. After 3 hours, the cultures were washed and cultured with a regular medium consisting of 15% fetal bovine serum (Equitech-Bio, Kerrville, TX, USA), 100 μ M L-ascorbic acid 2-phosphate (Wako Pure Chemicals, Osaka, Japan), 2 mM L-glutamine (Nacalai Tesque, Kyoto, Japan), and antibiotics containing 100 U/ml penicillin and 100 μ g/ml streptomycin (Nacalai Tesque) in Alpha Modification of Eagle's Medium (Invitrogen, Waltham, MA, USA). After forming well-attached colonies, the cells were passed for expansion.
CFU-F assay. The CFU-F assay was analyzed as described in Supplementary Methods 49-51 . Immunophenotype analysis. Cell surface antigens of IP-DPSCs and healthy DPSCs were assayed by flow cytometric analysis according to Supplementary Methods [49][50][51] . All primary antibodies are summarized in Supplementary Table 1.

Population-doubling and BrdU incorporation assays.
Population-doubling and BrdU incorporation assays were performed as described in Supplementary Methods 49-51 . Telomerase activity assay. Telomerase activity was measured by a telomere repeat amplification protocol (TRAP) assay using a quantitative telomerase detection kit (Allied Biotech, Ijamsville, MD, USA) with RT-qPCR as described in Supplementary Methods 49-51 . In vitro multidifferentiation capacity assay. Assays for multipotency into odontoblasts/osteoblasts, adipocytes, endothelial cells, and neural cells were performed as described in Supplementary Methods 49,50 . Assays for in vivo dentinogenic ability and self-renewal capacity. To analyze in vivo dentinogenic capacity, the cells were implanted subcutaneously with HA/TCP ceramic powders (40 mg, Zimmer Inc., Warsaw, IN, USA) into BALB/cAJcl-nu/nu mice as described in Supplementary Methods [49][50][51][52] (Supplementary Figure 2). Eight weeks after the surgery, the implants were analyzed histologically. For self-renewal assay, the cells were sequentially transplanted as described in Supplementary Methods 49,50 (Supplementary Figure 2).
In vivo dentin regeneration on human dentin. In vivo dentin regeneration on human dentin was performed as described in Supplementary Methods 8 (Supplementary Figure 6).
Histological assay for implant tissue. Harvested implant tissue was treated for histochemical, immunohistochemical, and immunofluorescent assays as described in Supplementary Methods [49][50][51] . The newly formed mineralized tissue area in each field was measured and shown as a percentage of the total tissue area 44,52 . Single colony-derived cell assay. Single colony-derived cells were assayed by population-doubling, BrdU incorporation and in vitro dentinogenesis as described in Supplementary Methods 19,49,50 . Treatment with IFN-γ and TNF-α. IP-DPSCs were stimulated with IFN-γ (100 ng/ml; PeproTech, Rocky Hill, NJ, USA) or TNF-α (100 ng/ml; PeproTech) at 37 °C for 0, 30, 60, and 120 minutes before analyzing an expression of NF-κ B and its phosphorylated NF-κ B under a serum-depleted condition. IP-DPSCs were