Pulp stem cells derived from human permanent and deciduous teeth: Biological characteristics and therapeutic applications

Abstract Human pulp stem cells (PSCs) include dental pulp stem cells (DPSCs) isolated from dental pulp tissues of human extracted permanent teeth and stem cells from human exfoliated deciduous teeth (SHED). Depending on their multipotency and sensitivity to local paracrine activity, DPSCs and SHED exert therapeutic applications at multiple levels beyond the scope of the stomatognathic system. This review is specifically concentrated on PSC‐updated biological characteristics and their promising therapeutic applications in (pre)clinical practice. Biologically, distinguished from conventional mesenchymal stem cell markers in vitro, NG2, Gli1, and Celsr1 have been evidenced as PSC markers in vivo. Both perivascular cells and glial cells account for PSC origin. Therapeutically, endodontic regeneration is where PSCs hold the most promises, attributable of PSCs' robust angiogenic, neurogenic, and odontogenic capabilities. More recently, the interplay between cell homing and liberated growth factors from dentin matrix has endowed a novel approach for pulp‐dentin complex regeneration. In addition, PSC transplantation for extraoral tissue repair and regeneration has achieved immense progress, following their multipotential differentiation and paracrine mechanism. Accordingly, PSC banking is undergoing extensively with the intent of advancing tissue engineering, disease remodeling, and (pre)clinical treatments.

periapical cyst contains cells with MSCs-like properties, have laid the groundwork for identifying a rich source of stem cells without any interference with the surrounding healthy tissue. Among these stem cells, DPSCs and SHED are noteworthy for their easy accessibility from teeth, which are regarded as disposable during occlusion management. DPSCs are derived from impacted third molars and orthodontic teeth, as well as supernumerary teeth. Compared with DPSCs isolated from permanent teeth, SHED obtained from exfoliated primary teeth show advantage on their painless collection procedures with minimal invasion. Pulp stem cells (PSCs) from human permanent and deciduous teeth, DPSCs and SHED, are herein focused on to emphasize their biological characteristics such as cell markers, multipotency and origin, and potential therapeutic applications, including endodontic regeneration, extraoral tissue repair and regeneration, as well as rising cell banking, in an attempt to shed light on potential translation in clinical settings.

| Biological characteristics
Both credited to MSCs, DPSCs and SHED possess similar features to BMSCs. For example, both cell types are plastic-adherent and can form colonies. However, biological variations related to anatomical localizations might exist. More specifically, PSCs from different dental pulp sources (permanent teeth or deciduous teeth) might exhibit different biological characteristics, which might consequently determine preferable clinical applications. Hence, a better understanding of biological characteristics of PSCs can inform enormous achievements for their potentials in regenerative medicine and novel developments for tissue engineering. Herein, cell markers, multipotency, and origin will be illustrated correspondingly.

Significance statement
Pulp stem cells can be readily harvested from dental pulp tissue of extracted permanent teeth and exfoliated deciduous teeth, respectively. However, a systematic and comprehensive review about pulp stem cells in terms of biological attributes and therapeutic applications is lacking. Accordingly, this review is concentrated on pulp stem cells to emphasize their updated biological characteristics such as cell markers, multipotency and origin, and promising therapeutic applications, including endodontic regeneration and extraoral tissue repair and regeneration, as well as rising cell bank with the intent of enhancing the understanding of dental mesenchymal stem cells and advancing associated tissue engineering and disease treatment.
terms of neural crest derivation, neurogenic markers, such as c-fos, γ-enolase, nestin, βIII tubulin, A2B5, musashi-1, neurofilament heavy and neurofilament light, microtubule-associated protein 2, glial fibrillary acidic protein, and oligo dendrocyte-associated CNPase are also identified on PSCs. 8,13 Given that DPSCs and SHED are derived from dental pulp tissues of different age groups, discriminated gene expressions, such as higher embryonic markers in SHED and higher neurogenic markers in DPSCs, have been observed, determining their lineage propensity toward a specific destination. 9 However, the majority of these markers are known from in vitro propagation studies and they do not actually reflect the properties of  15 Especially 4 days following injury, more NG2 + odontoblasts are detected and extend stubby processes into newly synthesized reparative dentin. This indicates that NG2, commonly identified as a pericyte marker, represents an additional marker of incisor pulp MSCs in mouse. Recently, Gli1 has also been recognized orthotopically as a mouse incisor MSC marker, and Gli1 + cells support both the homeostasis and injury repair of mouse incisor. 16 Notably, the majority of Gli1 + cells in mouse incisor do not express classic in vitro markers of MSCs, including CD44, CD73, CD105, CD146, and nestin, indicating that classic MSC markers may not be efficient to confirm MSCs in vivo. Moreover, Celsr1, a marker known from hematopoietic stem cells, is identified to label a rare and quiescent subpopulation of MSCs in mouse incisor pulp. 14

| Therapeutic applications
Depending on their multipotency and sensitivity to local paracrine activity, DPSCs and SHED exert therapeutic applications at multiple levels beyond the scope of the stomatognathic system, including locally intraoral pulp-dentin complex regeneration and systematically extraoral tissue repair and regeneration. However, the majority of aforementioned applications are conducted in animals, extensive (pre) clinical trials from bench to bedside are thus warranted.

| Pulp-dentin complex regeneration
The most apparent and promising application of DPSCs and SHED is pulp-dentin complex regeneration. Vital pulp, serving as the formative and supportive organ for dentin, is critical for tooth longevity. However, root canal treatment (RCT), as nonregenerative treatment, does not salvage dental pulp and tooth vitality but substitutes pulp tissue for inorganic materials, culminating with a devital and weakened tooth, which is susceptible to tooth fracture or even tooth loss. The emergence of regenerative endodontics brings brighter future and is recognized as a prospective approach for tooth preservation.
According to Murray et al, regenerative endodontics is referred to as biologically based procedures which are designed to replace damaged structures, including dentin and root structures, as well as cells of the pulp-dentin complex. 58 The ultimate purpose of regenerative endodontics is thus to regenerate viable pulp-dentin complex which histologically resembles the native tissue with anticipated physiological functions, including pulp innervation, pulp immunity, vascular perfusion, and tubular dentin formation. Revascularization treatment via blood clotting in root canal, which, although generates ectopic bone and cementum as well as fibrous tissue, shows the absence of histologic pulp-dentin structure, has been refuted as genuine tissue regeneration. In recent years, considerable research efforts have been used to advance the process of regenerative endodontics by virtue of cell transplantation and cell homing, extending our understanding of regenerative endodontics and promoting regenerative endodontics to more concise therapeutic approaches.

Cell transplantation-based regenerative endodontics
Cell transplantation is considered as transplantation of exogenous  59 It is illustrated to be capable of forming reparative dentin-like tissue on the pre-existing dentin layer. Organized dentinal tubules are absent in newly formed dentin; however, DPSCs first show potential to repair tooth structure and seal root canals. As a precursor, the triad of Notably, the availability of pulp tissue as a major obstacle hinders the feasibility and prevalence of autologous transplantation approach.
Except for wisdom teeth or orthodontic teeth, it is not realistic and clinically ethical to extract respective sound teeth for DPSCs. Specifically, in elderly patients, age-related pulp tissue shrinkage due to physiological secondary dentin generation and pathological tertiary dentin generation, as well as mineralization, including denticle and pulpal stones constrains the acquirement of DPSCs. In addition to reduction in quantity, their mobilization, differentiation, and regeneration capacities are adversely disturbed. 73,74 Furthermore, for patients with systemic diseases such as diabetes, systemic lupus erythematosus, and rheumatoid arthritis, intrinsic properties of MSCs are irreversibly altered. 75 As an alternative, allogeneic transplantation seems to be a desirable approach to circumvent those problems. In particular, SHED possess higher proliferation rate and increased population doublings as compared with DPSCs, 18   ) was closed after SHED implantation. Additionally, the amount of dentin (white arrows in [D]) was increased. In the control group, positive indications were not presented after apexification treatment. E, Quantification disclosed that SHED implantation significantly increased vascular formation, sensation, root length, apical foramen width, and dentin thickness. F, Representative H&E image of a human incisor 12 months after SHED implantation showed regenerated pulp tissue with a similar tissue structure to that of normal human pulp tissue. Odontoblasts (black arrows) localized at the margin of the regenerated pulp tissue were observed. SHED, stem cells from human exfoliated deciduous teeth extirpated completely. It is plausible that locally populated MSCs from periapical region, including PDLSCs, SCAP, and alveolar BMSCs account for recruitment. 79,80 In addition, systematically circulated stem/progenitor cells appear clinically available. 75 However, regarding the regenerated pulp-dentin anatomically mimicking native tissue, periapical stem cells appear less therapeutically applicable and feasible as compared with PSCs. Consensus holds that MSCs are distinctive and conserve their identities from their direct tissue sources and therefore tend to differentiate into original phonotypes. 76 The aforementioned revascularization treatment, which generates ectopic bone and cementum as well as fibrous tissue instead of histologic pulpdentin structure, seems indicative of this, especially considering evoked bleeding delivers periapical stem cells into root canal. Accordingly, rather than reparative tissue, the desired regeneration of pulpdentin complex which resembles the native tissue seems more likely to necessitate the presence of PSCs. The remnant viable pulp tissue after pulpotomy is determinant for cell homing-induced regenerative endodontics. Accordingly, more reliable approaches are warranted to distinguish irreversibly damaged pulp from healthy pulp for clinicians.
Optimized treatment protocol combining pulpotomy with cell homing strategy intimately is a prerequisite to enhance the chance of success.
Cell homing encompasses three distinct cellular processes, including cell recruitment, proliferation, and differentiation, which rely on the bioactivity of various growth factors delivered exogenously or endogenously. The growth factor delivery approaches define distinct procedures for cell homing-induced pulp-dentin regeneration. Accordingly, we will discuss the underlying rational and scientific basis in the following part, respectively.  is the most commonly used chelating agent regarding its ability to dissolve smear layer and soften dentin. In addition, EDTA is capable of liberating growth factors sequestered in dentin matrix. It was depicted that 10% EDTA (pH 7) for 20 minutes gives rise to the highest amount of TGF-β1 (923 pg/mL), as compared with 10% EDTA (pH 6) and 17% EDTA (pH 7) (449 pg/mL and 827 pg/mL, respectively). 87 The release of bFGF and VEGF is also revealed under 10% EDTA (pH 7), however, less effective relative to TGF-β1 (bFGF, 10 pg/mL, VEGF, 32 pg/mL after 20-minute exposure). As summarized in Table 2, an abundance of growth factors has been shown to be present in EDTA-soluble fraction of human demineralized dentin matrix. From the perspective of regenerative endodontics, upon releasing from dentin matrix by EDTA conditioning, growth factors are to be reactivated and initiate regenerative capability. As demonstrated in vitro, 10% EDTA (pH 7) for 10 minutes significantly promotes DPSC migration toward conditioned dentin discs. 100   Note: Adapted from Reference 99 with permission. Abbreviations: BDNF, brain-derived neurotrophic factor; bFGF, basic fibroblast growth factor; BMP-2, bone morphogenetic protein 2; EGF, epidermal growth factor; GDF-15, growth/differentiation factor 15; GDNF, glial cell line-derived neurotrophic factor; HGF, hepatocyte growth factor; IGF-1, insulin growth factor-1; PDGF, platelet-derived growth factor; PIGF, placenta growth factor; TGF-β1, transforming growth factor beta 1; VEGF, vascular endothelial growth factor. compared with cells cultured on untreated dentin or without direct contact with EDTA-treated dentin. 102  proliferation, and differentiation. Unfortunately, 5.25% NaOCl not only inhibits DPSC migration toward dentin disks but also prevents their attachment and differentiation on dentin surface. 100 It is not surprising in terms of deproteinization and cytotoxicity effect of NaOCl.
When SHED encapsulated in tooth slice/PLA are transplanted subcutaneously, 93 NaOCl conditioning gives rise to disorganized tissue in pulp cavity and inhibits odontoblastic differentiation, indicative of the deleterious effect that NaOCl imposes on remaining PSCs. As a matter of fact, previous clinical study has revealed that 5% and 0.5% NaOCl do not result in significantly different antibacterial efficacies. 103 Consistently, a latest clinical investigation reports a nonremarkable difference in periapical healing as well as postoperative pain incidence between 5% NaOCl and 1% NaOCl. 104 Based on these results, it can be deduced that there is no tremendous advantage of into newly regenerated tubular dentin should be considered together.
The third case in point is that regenerating tubular dentin remains fairly challenging. Organized dentinal tubules in newly formed dentin are less predictable. This is presumably associated with insufficient signals for differentiation and maturation of odontoblast-like cells and diminished number of odontoblast-like cells. Accordingly, it is beneficial and significant to provide optimal signals for odontoblastic differentiation of recruited cells. To that end, copine 7, a preameloblast-derived factor which enhances generation of organized dentinal tubules in root segment model, seems to be a promising candidate. 106 Last but not least, from a functional regeneration standpoint, the efficacy of pulp-dentin complex regeneration in animal tooth models is predominantly validated by using histological approaches, while in the absence of functional neural and vascular testing. Combination with functional innervation and vascularization appears more persuasive for functional pulp regeneration. 107

| Extraoral tissue repair and regeneration
PSCs share similar characters with BMSCs. However, compared with BMSCs, harvesting PSCs from disposable teeth is less invasive and more feasible. Moreover, the administration of PSCs involves minimal immune objection and ethical issues. Accordingly, PSCs may represent a prospective therapy and appealing candidate for BMSC transplantation for extraoral tissue engineering and regenerative medicine. As shown in Table 3 Table 3, such as calvarial defect, 110 cerebral ischemia, 116,117 and limbal stem cell deficiency, 38  It was also reported that SHED are capable of differentiating into hepatocyte-like cells directly without fusion in mouse models of carbon tetrachloride-induced liver fibrosis, therefore recovering liver disfunction. 41 In a recent mouse in vivo study, 12

108
Acute renal injury SHED Cryopreserved SHED were transferred intravenously or intraperitoneally into glycerol-induced acute renal failure rats. SHED homed to kidney and accelerated renal tubule epithelial cell regeneration.

109
Nephritis SHED SHED were transferred intravenously into systemic lupus erythematosus MRL/lpr mice.
Hypercellularity, mesangial matrix hyperplasia, and basal membrane disorder were prevented histologically, while serum creatinine, urine protein, and C3 were significantly reduced, serum albumin was elevated. SHED-CM attenuated pro-inflammatory response, and induced anti-inflammatory M2-like microglia, substantially improving cognitive function.

112
Cerebral ischemia SHED SHED-CM was injected intranasally into middle cerebral artery occlusion-induced ischemia rats. SHED-CM promoted neurogenesis and angiogenesis, ameliorating ischemic brain injury.

113
Cerebral ischemia SHED SHDE and SHED-CM were transplanted into hypoxia-ischemia-injured neonatal mice. Both SHED and SHED-CM remarkably suppressed brain loss, while augmented survival rate and neurological function.
114 Traumatic brain injury SHED SHED or SHED-Ex were injected into external mechanical force-injured rat brains.
SHED-Ex significantly ameliorated behavioral score and lesion recovery, while suppressed pro-inflammatory M1microglia.
DPSCs were superior to BMSCs in terms of reducing infarct volume and reactive gliosis as well as promoting angiogenesis.

117
Parkinson's disease SHED dSHED or SHED were transplanted into striatum of 6-hyroxydopamine-induced Parkinsonian rats. dSHED were more efficient to improve dopamine level and promoted neurological recovery in contrast to SHED.
SHED reduced neurotoxicity, while enhanced behavioral performance and olfactory function.

119
Spinal cord Spinal cord injury SHED SHED were transplanted intraspinally into NYU-impactor-induced spinal cord injury rats.
SHED reduced astrocyte hyperplasia, inhibited neuronal apoptosis and T cells entrance into parenchyma as well as TNF-α expression.

120
Spinal cord injury SHED SHED or iSHED were injected into spinal cord-injured rats.
SHED and especially iSHED promoted functional recovery with neuronal and glial differentiation. The transplantation significantly suppressed liver fibrosis and restored alanine transaminase, aspartate transaminase, and ammonia levels.

125
Heart Ischemia-reperfusion injury SHED SHED-CM was intravenously injected in left anterior descending artery ligation-induced ischemia-reperfusion injury mice.
SHED-CM reduced myocardial infarct size as well as decreased apoptosis and inflammatory cytokine levels.

126
Acute myocardial infarction DPSCs DPSCs were injected intramyocardially in coronary artery ligation-induced myocardial infarction rats. Cardiac function was improved with thickened anterior wall of left ventricle, reduced infarct size, and increased angiogenesis.

127
Muscle Muscular dystrophy SHED SHED were transplanted singly or consecutively into golden retriever muscular dystrophy dogs via intra-arterial or intramuscular injection. SHED were capable of engrafting, differentiating, and persisting in the affected muscle in the absence of immunosuppression. Intra-arterial and consecutive delivery was more effective.  41 and indirect paracrine mechanism-mediated antifibrosis and anti-inflammation potentials. 124 Consequently, in terms of mechanisms underlying PSC transplantation-derived benefits, it is important to note that paracrine mechanism initiated by therapeutic factors secreted from transplanted PSCs, at least partially, participates in treatment and enhances endogenous tissue repair capability.
The attachment and survival of transplanted PSCs at the lesion site is determinant on the following multipotential differentiation.
Once PSCs fail to survive in the injury sites, their direct contribution to tissue recovery will be diminished or even abolished, which, however, is not implicated in their indirect paracrine-dependent efficacy on tissue recovery in some instances. As revealed by SHED transplantation for acute lung injury treatment, 111 a progressive reduction of surviving SHED occurs in lung, from 10% over a 2 to 24 hours time period after transplantation to less than 1% 1 week after transplantation. However, SHED promote such significant recovery as SHED-CM thereafter to the end of the study, ascertaining that powerful paracrine mechanism elicited by SHED mediates durable therapeutic potential. In another instance, irrespective of scarce homing and distribution in bone, SHED intravenous transplantation pronouncedly reduces ovariectomy-induced bone loss in the trabecular bone of the distal femur metaphysis. 132 In contrast, it is extraordinarily noteworthy that survived cells do not imply they are definitely capable of differentiating and compensating for the lost cells. 138 In a study which revealed remarkable therapeutic effects of both SHED and SHED-CM on the recovery of neonatal hypoxia-ischemia brain injury, 114  rheumatoid arthritis, 139 and spinal cord injury. 122 Treatment of autoimmune encephalomyelitis mice with recombinant ED-Siglec-9 recapitulates therapeutic efficacy of SHED-CM treatment, indicating that ED-Siglec-9 is a paramount factor mediating SHED-CM potential for encephalomyelitis treatment. 140 Hepatocyte growth factor, originally known as mitogen for hepatocytes, is more highly expressed in SHED-CM compared with BMSC-CM. 126 It is reported to play a pivotal role in SHED-CM-mediated inhibitory actions on myocardial infarct size and apoptosis following ischemia-reperfusion.
In addition to soluble factors, exosomes derived from SHED-CM (SHED-Ex) are reportedly attributable to the functional recovery of diabetes, 135 traumatic brain injury, 115  propagation on bench, cell survival concern at lesion site, and such potential safety hazards as tumorigenesis compared with cell transplantation.

| Cell banking
Due to easy accessibility and favorable therapeutic applications of PSCs, cell banking has attracted additional attention. Clinically acquired teeth and harvested whole pulp tissues as well as isolated DPSCs and SHED can be cryopreserved to maintain their stemness capacity for many years, providing a banked stem cell source for further use. Furthermore, with the induction of GCSF, mobilized dental pulp stem cells (MDPSCs) are isolated from DPSCs. 145 Extraordinarily, MDPSCs from aged human donors are as potent as those from young donors, revealing no obvious distinction in terms of migration, multipotent differentiation as well as pulp-like tissue regeneration in ectopic tooth. Hence, these cells might be proper candidates for PSC banking, irrespective of the MSC aging-associated decrease in regenerative potential. Recently, cell/tissue banks in the dental field have been taken into practice in several countries, such as BioEDEN (Austin, Texas) and Store-A-Tooth (Lexington, Kentucky).

| CONCLUSION
Following the innovative discovery of DPSCs almost 20 years ago, PSCs have been paving the way for the substantial advancement of tissue engineering and regenerative medicine, ranging from endodontic regeneration to extensive extraoral applications. Tremendous efforts have been made to simplify PSCs-dependent endodontic regeneration procedures, from the absence of cell transplantation to dependence on endogenous growth factors in an attempt to facilitate clinical practice.
Orthotopic pulp-dentin regeneration has been realized through SHED transplantation, however, whether DPSC transplantation or solo growth factor induction possess the similar capability is still elusive.
Alternatively, recruiting periapical MSCs is highly appreciated if epigenetic regulation of odontogenic capability of PDLSCs, SCAP, or BMSCs is achieved to regenerate pulp-dentin complex in a true sense. Identification of epigenetic factors warrants further investigation, such as Wnt3a. 146 Actually, dental MSC inspired-endodontic regeneration approach seems more conservative in comparison with total tooth organ regeneration strategy, which holds the highest promise for dis-

CONFLICT OF INTEREST
The authors declare no potential conflicts of interest.

AUTHOR CONTRIBUTIONS
X.S.: organization and design, manuscript writing, final approval; J.M., Y.L.: organization, manuscript review and editing, financial support, final approval.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author.