Abstract
Skin, the largest organ of the body, is a promising reservoir for adult stem cells. The epidermal stem cells and hair follicle stem cells have been well studied for their important roles in homeostasis, regeneration, and repair of the epidermis and appendages for decades. However, stem cells residing in dermis were not identified until the year 2001, when a variety of stem cell subpopulations have been isolated and identified from the dermis of mammalian skin such as neural crest stem cells, mesenchymal stem cell-like dermal stem cells, and dermal hematopoietic cells. These stem cell subpopulations exhibited capabilities of self-renewing, multipotent differentiating, and immunosuppressive properties. Hence, the dermis-derived stem cells showed extensive potential applications in regenerative medicine, especially for wound healing/tissue repair, neural repair, and hematopoietic recovery. Here we summarized current research on the stem cell subpopulations derived from the dermis and aimed to provide a comprehensive review on their isolation, specific markers, differentiation capacity, and the functional activities in homeostasis, regeneration, and tissue repair.
Introduction
Stroma has attracted much attention recently, since tissue-specific stem cells were identified in adult stromal tissues. The bone marrow stromal tissue-derived stem cells termed as bone marrow mesenchymal stem cells (BMMSCs) were first identified in 1976 (33). Further, MSCs have been found to indwell in many other stromal tissues, such as adipose tissue, periosteum, fetal tissues, and amniotic fluid (15,52,88,157). The discovery of stromal stem cells implied their significant potential roles in tissue repair and regeneration.
Skin, consisting of epidermis, dermis, and appendages, such as hairs and glands, is considered as a natural resource of adult stem cells. Epidermis, derived from the surface ectoderm, attracts the primary focus of stem cell research. Epidermal stem cells, residing in the lower layer of epidermis, are well studied in the past few decades and play important roles in skin regeneration and wound healing (37,48,131,142). Hair follicle stem cells identified in the bulge area are responsible for the hair cycle and hair follicle reconstruction after wounding (86). Melanocyte stem cells reside close to the hair follicle stem cell area in the bulge, and defects cause hair graying (38). Dermis has diverse origins during mammalian embryo development: craniofacial dermis from the neural crest (89), dorsal trunk dermis from the dorsal portion somite formed dermomy-otome (13,79), and ventral trunk dermis from the lateral plate mesoderm (91). However, it is surprising that the dermis, which represents a larger adult stem cell reservoir than the hair follicle and epidermis together, has largely escaped the focus of the majority of the stem cell community. Until 2001, a couple of stem cell subpopulations with multilineage differentiation capacity have been isolated and identified from the mammalian dermis by several research groups independently (116,128,151). Later, other types of stem cell subpopulations were also identified from the dermis, which were regarded as a promising alternative to bone marrow-derived MSCs for stem cell-based therapy. Here we summarized the biological properties and the potential applications of three representative stem cell subpopulations from the dermis, including neural crest stem cells, MSC-like dermal stem cells, and dermal hematopoietic cells.
Neural Crest Stem Cells Derived from the Dermis
The neural crest forms transiently in the dorsal neural primordium during the embryonic development of vertebrate animals and is comprised of abundant stem cells to form neurons, glia cells, pigment cells in the skin, and most of the facial skeleton (80). In adults, many tissues, including the carotid body, dorsal root ganglia, bone marrow, heart, gut, cornea, and skin, still contain neural crest stem cells (87,96). It is confirmed that skin contains diverse subpopulations of neural crest-derived stem cells. In the epidermis, the melanocyte stem cells have been found to migrate from the trunk neural crest during embryo development (71,130). Studies also suggested a developmental plasticity that allows melanocytes to emerge from the cranial neural crest (3). In adult epidermis, the bulge region of the hair follicle has been confirmed to be the anatomic location of the melanocyte stem cell niche (90). Dermis, the subepidermal stromal part of the skin, has been found to be one of the most abundant sources of neural crest stem cells in the past decade. Since the successful isolation of skin-derived precursors (SKPs) (128,129), more subpopulations of neural crest stem cells (Table 1) have been isolated from the dermis. The isolation protocols were similar with SKPs, which in brief were initially discarding the epidermis by tissue decomposition enzymes (59,128,129) and digesting the dermal tissue by related enzymes, such as collagenase, to isolate the dermal cells from the dermal fracture, and obtaining the floating clusters of cells when culturing the isolated dermal cells in the medium with growth factors, such as B-27, epidermal growth factors, and fibroblast growth factors (6,49,59,128,129
The Biological Characteristics of Neural Crest Stem Cells Derived From the Dermis
Abbreviations: SKPs, skin-derived precursors; ND, not determined.
Till now, neural crest stem cells derived from dermis have been found to exist in various species, including mouse (128), rat (49), pig (25), and human (6,59,129, 143). Fernandes et al. (28) found that SKPs were much more abundant in later stages of fetal development than in the adult dermis, where they were rather scarce (34). Furthermore, SKPs exhibit a clear loss of stem cell competence associated with aging and most likely being explained by cellular senescence (34,77). Neural crest stem cells derived from the dermis distributed ranging from the skin with hair, such as scalp, to the ones without hair, such as foreskin, which promised the potential applications of the neural crest stem cells derived from dermis to the basic and clinical medicine due to their easy acquirement. To achieve the neural crest stem cells as pure as possible, the exact location of the neural crest stem cells in the skin should be fully understood. According to previous studies, the hair and whisker follicle dermal papillae were an endogenous niche for SKPs (9,28). Recently, another research found that about 1,000-fold of the sphere-forming neural crest stem cells were enriched in rodent vibrissal follicle dermal papilla compared with the whole facial skin (49). However, human foreskin with no hair follicle was still an abundant source for SKPs (129). Thus, there should be other niches for neural crest stem cells in the dermis besides the hair follicle papillae. Recently, the dermal capillaries have been proven to be the anatomical niche of the neural crest stem cells in the skin without hair follicles, such as foreskin (106).
Neural crest stem cells derived from the dermis exhibit high proliferation and colony formation potential. SKPs, for instance, possessed similar proliferation ability after 50 passages as the primary ones. When being dissociated into a single cell, they could form floating spheres in the same culture condition (128). Investigation showed that approximately 40% of the isolated single SKP cells were able to self-renew and form clones (129). The SKPs' abilities of self-renewing and proliferation decreased with age and passages in vitro (34,77). The SKPs' abilities of self-renewing and proliferation might also vary from different expansion media, donor source areas, age, and species. Recently, it was reported that SKPs could be derived from primary and passaged adult dermal fibroblasts precultured in conventional monolayers, which might be a better method to enrich SKPs (45). For stem cell-based therapy or tissue engineering, much more efficient methods to gain as many SKPs are also imminent.
Phenotypically, neural crest stem cells derived from the dermis expressed cell markers similar to neural crest stem cells. Nestin, one of the neural crest stem cell markers (1), was expressed in most subpopulations of neural crest stem cells derived from the dermis (Table 1), which promised that nestin might be a potential marker to isolate neural crest stem cells from the dermis. However, nestin was not specific to neural crest stem cells, which was also expressed in some subpopulations of dermal MSCs (112). In addition, neural crest stem cells derived from the dermis showed differences with embryonic neural crest stem cells. Both SKPs and embryonic neural crest stem cells shared the cell markers nestin and vimentin and gene expression of transcription factor Sox9 (28). However, slug, twist, snail, and pax3 were high in SKPs and low or undetectable in embryonic neural crest stem cells. In addition, the mRNA level for p75 neurotrophin receptor (p75NTR, a marker for neural crest stem cells) was low in SKPs (129). In a gap analysis of SKP gene expression, there were other two genes, GAP43 and MAP2, in high expression levels besides the genes nestin, fibronectin, and vimentin (66). Furthermore, SKPs express sex-determining region Y box 2 (Sox2), an embryonic stem cell marker (9). Neural crest stem cells derived from the dermis also expressed mesenchymal cell markers such as vimentin and fibronectin (25,128,129), indicating that neural crest stem cells exhibited stromal cell properties when migrating from neural crest to stroma.
In addition to the markers above, there was another animal model to identify the neural crest origin of the neural crest stem cells derived from the dermis. In the compound transgenic Wnt1-Cre/R26R neural crest reporter mice, neural crest origin cells transiently expressing protein Wnt1 in early development and β-galactosidase reporter could be permanently detected and shared with their progeny. Studies showed that all the SKPs (49) and the p75/ Sox10-positive neural crest-derived cells (143) isolated from the Wnt1-Cre/R26R transgenic mice expressed β-galactosidase, which further confirmed their neural crest origin. This transgenic model together with neural crest stem cell markers would be proper approaches to identify neural crest stem cells derived from the dermis. Another study found that SKPs in the facial and dorsal hair follicles were from distinct developmental origins (58) neural crest and somite using the Cre lineage tracing approach. However, both of the facial and dorsal SKPs expressed the markers fibronectin, vimentin, nestin, and versican. The proliferation, self-renewal, and differentiation of the two populations were highly similar. Further, all the developmentally distinct SKPs from facial, dorsal, and ventral dermis showed a very high degree of similarity in gene expression detected by microarrays, although they retained differential expression of a small number of genes that reflect their developmental origins.
The markers expressed in the stem cells might be related to their differentiation potential. SKPs, for example, expressed both neural crest stem cell marker (nestin) and mesenchymal cell marker (vimentin) and could generate both neuroectodermal and mesodermal lineages (128). Previous studies showed that SKPs could give rise to neurons, glia, Schwann cells (129), smooth muscle cells (25,128), adipogenic, osteogenic, and chondrogenic lineages (72). Further, SKPs were able to generate insulin-producing cells in proper culture condition (40). Neural crest stem cells derived from the dermis exhibited high plasticity in neuronal differentiation. In vitro, SKPs could be induced to neuron-like cells with pan-neuronal and peripheral autonomic lineage markers (29); however, no electronic signals could be detected in the induced neuron-like cells from SKPs because these SKPs generated neuron-like cells that were immature neurons (20). Recently, a study improved the culture condition and successfully induced the SKPs into functional neurons with the ability to generate action potentials or synaptic activity (76). In addition, neural crest stem cells derived from the dermis showed neural differentiation superiority. The dermis-derived p75/Sox10-positive cells could generate high neuronal cells, but low in comparison to that of smooth muscle cells, osteogenic, chondrogenic cells, and melanocytes (143). They also exhibited selectivity in neural cell types during differentiating. The AC133+ cells could be induced to differentiate into neurons, astrocytes, and rarely into oligodendrocytes in vitro. In vivo, they gave rise to different neural phenotypes: a most abundant population of well-differentiated astrocytes and immature neurons (6). To some extent, the multipotential, especially neuroectodermal, lineage differentiation ability would be another criterion for identification of the neural crest stem cells derived from the dermis.
The repair of the neurons has been puzzling us for decades. The discovery of the stem cells, especially the neural stem cells, brought new hope. However, the scarce sources and complicated obtainment of the neural stem cells limited their further applications. Based on the processes of high self-renewing and proliferation abilities and multidifferentiation potential, neural crest stem cells derived from the dermis have been applied to tissue repair, especially neural repair. Recent experimental studies have shown the successful application of neural crest stem cells derived from the dermis in the treatment of animal models of neural diseases, including Parkinson's disease (PD), spinal cord injury, and peripheral nervous system diseases. These studies have shown that neural crest stem cells derived from the dermis can ameliorate the symptoms and partly restore neural function. Recently, dopaminergic neuronal cells were generated from SKPs cultured with amino acids 157-171 of von Hippel-Lindau protein (43,67). Transplantation of these dopaminergic neuronal cells derived from SKPs into a PD model of rats resulted in less PD-like symptoms by enabling production of dopamine (43).
Spinal cord injury is a devastating, traumatic event for patients, which often results in neuron loss and the demyelination of intact axons. All of the current treatments of spinal cord injury are not very effective. Stem cell transplantation is one of the promising treatment alternatives. Several types of stem cells, such as neural stem cells, MSCs, and even embryonic stem cells have been transplanted into the injured spinal cord. Many studies have revealed multiple beneficial roles of transplanted stem cells in spinal cord injury. Neural crest stem cells, derived from the dermis, which were easily accessible, capable of rapid expansion in culture, and most importantly neural crest original, were another promising candidate for stem cell- based therapy for spinal cord injury. It was demonstrated that both SKPs and SKP-derived myelinating Schwann cells survived well within the injured spinal cord, reduced the contusion cavity size, myelinated endogenous host axons, and recruited endogenous Schwann cells into the injured sites. The SKP-derived myelinating Schwann cells also provided an environment that was highly conducive to axonal growth, which enhanced locomotor recovery. Results suggested that the SKPs and their derivatives were an autologous source of cells for treatment of the injured spinal cord (10). In addition, neural crest stem cells derived from the dermis had great potential in the peripheral nervous system diseases. Not long before, the protocol for methods of generating and enriching Schwann cells from SKPs was established. The SKP-derived Schwann cells were more than 95% pure and can keep the Schwann cell characteristics for a long time (11). Transplanting the SKP-derived myelinating cells into diseased mice effectively increased the treatment of nervous system injury, such as chronically injured nerves (139), congenital leukodystrophies, and dysmyelinating disorders (81). Whereas the lineage-specific differentiation into cells of neural origin has been well processed in many labs, the transcription factors and molecular key events that initially allocated the neural crest stem cells derived from the dermis to a specific lineage are almost completely unknown. It is of great interest to decode these molecular mechanisms for a more effective development of novel stem cell-based therapies in neural repair.
Neural crest stem cells derived from the dermis exhibited both neuroectodermal and mesodermal differentiation; thus, besides the application in neural repair, they were also alternatives for tissue engineering. In vitro, SKPs were proven to differentiate into bladder smooth muscle cells in the presence of FBS and bladder organoids or conditioned medium derived from stretched or relaxed bladders. When transplanted into ex vivo culture of mouse embryonic bladders or adult rat bladders exposed to mechanical strain, SKPs differentiated into bladder smooth muscle cells (127). However, further study showed that organoids or diffusible factors derived from stretched bladders attenuated smooth muscle differentiation of SKPs in vitro, which implied that a pathologic bladder microenvironment contained both positive and negative factors for the differentiation of SKPs into smooth muscle cells. In addition to bladder, SKPs also showed great potential for tissue-engineered bone repair (62). However, the application of neural crest stem cells derived from the dermis in tissue engineering is still in the stage of in vitro models, and much more studies of in vivo animal testing or even clinical trials are needed.
MSC-Like Dermal Stem Cells
The presence of nonhematopoietic stem cells in the body was first observed in bone marrow by the German pathologist Cohnheim about 130 years ago (97). Since 1968, many groups have provided the evidence that bone marrow contained the spindle-shaped and plastic-adherent cells, which could differentiate into at least osteoblasts, chondrocytes, and adipocytes (16). According to the International Society for Cellular Therapy, human MSCs are defined as plastic-adherent cells with fibroblast morphology when culturing in standard condition, expressing cell markers CD105, CD73, and CD90, negative for the hematopoietic markers CD45, CD34, and CD14, and processing multidifferentiation potential (23). Besides, MSCs have also been identified in many other tissues, such as adipose tissue, periosteum, fetal tissues, and amniotic fluid (15,52,88,157). Since 2001, distinctive stem cell subpopulations with MSC properties (here termed MSC-like dermal stem cells) have been identified in the dermis (Table 2). The isolation methods of the MSC-like dermal stem cells are similar to bone marrow MSCs. In brief, epidermis and subcutaneous tissues were removed following the tryspin (113,116) or dispase digestion, the remaining dermal tissue was mechanically dissociated (113,116) or continuously digested by collagenase (17,78,151), and the dermal cells were collected and cultured in the proper medium such as DMEM (17,47,78) or IMDM (116) with FBS. The adherent cells were selected for further culture and identification.
The Biological Characteristics of MSC-Like Dermal Stem Cells
Abbreviations: DMCs, dermis-derived multipotent stem cells; MSCs, mesenchymal stem cells; ABCB5, ATP-binding cassette, subfamily B (MDR/TAP), member 5; ND, not determined.
Previous studies suggested that hair follicles might be the priority distribution of the skin stem cells. The epithelial stem cells, for example, have been found to reside in the hair follicle bulge area (141). However, little attention was paid to the hair follicle dermal compartments consisting of dermal papilla and dermal sheath (55). Several decades ago, researchers found that the hair follicle dermal sheath cells participated in skin wound healing and hair follicle reconstruction (36,149). Their studies partially demonstrated that hair follicle dermal cells possessed stem cell characteristics. Further studies found that the adult hair follicle dermal papilla (DP), dermal sheath (DS) cells, and their clonal lines showed highly similar capabilities of proliferation and adipocytic, osteocytic, and neuronal differentiation (56,102). Another study revealed that the DP and DS cells showed a fibroblastic morphology and expressed cell markers CD44, CD73, CD90, nestin, α-internexin, and NG2 (46). Further, the DP cells have been applied for induced pluripotent stem cells by retroviral transduction of octamer-binding transcription factor 4 (OCT4), cMyc, Kruppel-like factor 4 (Klf4), and Sox2 (44). In summary, the hair follicle dermal cells, together with the epithelial stem cells, melanocyte stem cells, and the neural crest stem cells, have verified the hypothesis that hair follicles are the priority distribution of the skin stem cells.
Phenotypically, MSC-like dermal stem cells expressed the markers CD105, CD73, and CD90. In addition, vimentin, a mesenchymal marker, was also generally found in MSC-like dermal stem cells (17,78,85). CD34, negative in bone marrow MSCs, was found in some subpopulations of MSC-like dermal stem cells (151). Besides, some embryonic markers, such as STRO-1, SSEA-4, Sox2, OCT4, and Nanog, were also detected in the MSC-like dermal stem cells (132). Further, MSC-like dermal stem cells could differentiate into the classic mesoderm lineages, including osteocytes, adipocytes, and chondrocytes. Skeletal, smooth, and cardiac muscle cells could also be generated from MSC-like dermal stem cells (151). According to our previous investigations, dermis-derived multipotent stem cells (DMCs) could differentiate into neurons and insulin epithelioid cells (18,117,118). Multipotent fibroblasts derived from dermis also showed high plasticity to differentiate into hepatocytes (17,47,78), neurons (17,47,78), Schwann cells (47), and insulin-like cells (7,78).
In addition to normal dermal tissues, MSC-like dermal stem cells have also been identified in the pathological dermis, such as keloid (85), granulation tissue (95), and burn eschar (133). They expressed the MSC markers (CD90, CD29, CD44, CD13, CD59, fibronectin, and vimentin) and even embryonic stem cell markers (OCT4, Nanog) (95). All of them could differentiate into adipocytes, osteoblasts, and chondrocytes. Granulation tissue-derived MSCs showed much stronger plasticity to generate smooth muscle cells, angiogenic endothelial cells, neurons, and Schwann cells (85). In pathological conditions, MSC-like dermal stem cells seemed to be activated and exhibited an embryonic phenotype. Granulation tissue derived MSCs, for example, expressed embryonic stem cell markers (OCT4 and Nanog) and produced high levels of vascular endothelial growth factor (VEGF).
Fibroblasts, a morphologically defined dermal cell population, synthesize and remodel the extracellular matrix and compose the main part of stromal cells in dermis. MSC-like dermal stem cells are relatively rare in dermis compared with fibroblasts. It is difficult to distinguish the MSC-like dermal stem cells from the primary adherent fibroblasts because fibroblasts exhibit similar characteristics to MSCs (17,47,78). Studies showed that human dermal fibroblasts shared similar morphological appearance, growth rate, phenotypic profile, typical mesenchymal markers, adhesion molecules, and multidifferentiation with human adipose-derived MSCs (2,12). Another study explored a minor difference that fibroblasts exhibited a time lag in the induction of adipogenesis-related genes (PPARγ, C/EBPα, and FABP4) compared with adipose-derived MSCs during adipogenic inducing. The reason might lie in that the preadipocyte transcription factor ZNF423 in dermal fibroblasts was delayed in induction (54).
It is controversial whether fibroblasts possess anti-inflammatory activity. An investigation found that human dermal fibroblasts lacked the ability of angiogenesis and anti-inflammation compared to human adipose-derived MSCs (12). However, some others revealed that both fibroblasts and MSCs exhibited immunosuppressive potential (42,60). Both of them could regulate early T-lymphocyte activation by inducing cell cycle arrest (60). In addition, MSCs and fibroblasts that mediated immunosuppressive effects were promoted by some soluble factors (2, 3-dioxygenase, prostaglandin-E2, nitric oxide) and some inflammatory cytokines (interferon-γ, tumor necrosis factor-α) (24,64,121). The difficulties in distinguishing the fibroblasts and MSCs may lie in the different species, different tissues, and different isolation methods by different groups. The relationship between the dermal fibroblasts and MSC-like dermal stem cells warrants further exploration. Our hypothesis is that dermal fibroblasts are the functional state of the MSC-like dermal stem cells. During wound healing, for example, the resident MSC-like dermal stem cells might transform into functional fibroblasts by proliferating and differentiating. When the wound is healed, these dermal stromal cells become quiescent stem cells again.
Because of the easy accessibility, multidifferentiation, and immunosuppressive effects, MSC-like dermal stem cells have been widely used in a variety of fields, including wound healing/tissue repair, tissue engineering, organ transplantation, and even cell reprogramming. Evidence in animal models showed that MSC-like dermal stem cells could accelerate the healing in cutaneous wounds and spinal cord injury by transdifferentiation and paracrine effects. We have shown that both topical and systemic application of DMCs accelerated wound healing, and the promoting effect of topical application was earlier than systemic. However, for treatment of skin wounds combined with irradiation, only systemic application of DMCs promoted wound healing (19,113). In a model of a wound combined with infection by P. aeruginosa in total body irradiated rats, human β-defensin-2 genetically modified DMCs displayed both enhanced healing-promotion and anti-infection effects (155). Previous studies have shown that the transplanted DMCs accumulated at the wound site and promoted the healing processes by transdifferentiation and paracrine effects similar to bone marrow MSCs (144). The supernatant of DMCs has been found to promote proliferation of fibroblasts, epidermal cells, and endothelial cells. Further, the transcripts tested by microarray assay showed that DMCs expressed a series of cytokines and extracellular matrix molecules including VEGF, platelet-derived growth factor, transforming growth factor- β (TGF-β), hepatocyte growth factor (HGF), vascular cell adhesion protein 1 (VCAM-1), intercellular adhesion molecule 1 (ICAM-1) (19), which are highly related to wound healing. In vivo, DMC transplantation increased the VEGF and HGF expression and capillary density in the wounded tissues. Further, transplanted DMCs were found to incorporate into capillary structures as confirmed by endothelial cell markers (115). Moreover, we found that SDF-1/CCR4 was an important pathway to improve the recruitment of the DMCs in wounded tissues (154). The chemokine receptors (CCR3, CCR4, CCR6, CCR10, CXCR1, and CXCR2) and their ligand CCL5RANTES were also potent in the migration of dermal MSCs in the wounded tissues (65). In the remodeling phase of wound healing, MSC-like dermal stem cells played an important role in reconstruction of the appendages (36).
Harnessing endogenous repair mechanisms to promote tissue regeneration has long been a goal in biomedical science. However the recent focus is mainly upon stem cell-based transplantation approaches to disorders that range from diabetes to spinal cord injury. As the evolving body of work shows that many adult tissues contain resident stem cell populations, another therapeutic approach would be promising. Epidermal stem cells, for example, were activated after wounding and responsible for reepithelialization. In a collagen-promoter green fluorescent protein (GFP)-reporter transgenic bone marrow chimeric mouse model, no GFP signal was found in the skin granulation tissues after wounding, which suggested that the fibroblasts/myofibroblasts involved in wound healing originated from resident dermal stem cells rather than circulating bone marrow stromal progenitor cells (4). Though the underlying mechanisms of the recruitment of resident dermal stem cells into skin wound healing are still unclear, it would be another interesting approach to recruit resident dermal stem cells for further chronic wound treatment and skin regeneration.
In addition to cutaneous wound healing, MSC-like dermal stem cells have been applied to other tissue repair. In vitro, DMCs stably transfected with Tailless-like (TLX)-expressing plasmid showed enhanced proliferation and preferential differentiation into NF200-positive neurons in contrast to GFAP-positive astrocytes. The TLX-expressing DMCs improved locomotor recovery and healing of spinal cord injury significantly in a rat spinal cord injury model (140). In vitro and in vivo, MSC-like dermal stem cells have shown myocardial differentiation (151), which implied the potential application in the cardiac disease treatment. Another study showed that the ABCB5+ MSCs derived from the dermis gave rise to human spectrin- and delta-sarcoglycan-positive skeletal myofibers and accelerated skeletal muscle regeneration in a human-to-mouse skeletal muscle injury xenotransplantation model (63).
Studies both in vitro and in vivo have shown that MSC-like dermal stem cells exhibited immunosuppressive potential. In an autologous melanocyte transplantation clinical trial, the efficiency of the transplantation was closely associated with skin-homing CD8+ T-cell activities. Further study found that dermal MSCs inhibited the isolated skin-homing CD8+ T-cell proliferation and regulated their cytokine/chemokine production in the coculture system, which implied that DMSCs might be used as an auxiliary agent to improve transplantation efficacy (152). In a fully MHC-mismatched cardiac allograft murine model, the transplanted ABCB5+ MSCs were found to localize in recipient immune tissues, including bone marrow, spleen, and blood (110). The survival of the mice that received ABCB5+ MSC transplantation was apparently prolonged. The ABCB5+ MSCs suppressed the rejection by upregulating expression of PD-L2 and PD-1 and increasing levels of CD4+FoxP3+ Treg generation (109,110). In our recent study, transplanted MSC-like dermal stem cells decreased the incidence and severity of acute graft-versus-host disease after MHC-mismatched hematopoietic stem cell transplantation in mice by the inhibiting of splenic cell proliferation and enhancing Treg cell activities (35). We also found that transplantation of DMCs diminished intestinal damage by modulating the inflammatory cytokine secretion, increasing the migration activity, and inhibiting the apoptosis of LPS-stimulated macrophages in a mouse sepsis model (unpublished data).
With a high degree of plasticity, MSC-like dermal stem cells would be another interest for tissue engineering (8). Recently, MSC-like dermal stem cells were shown to form self-assembled tissue-engineered constructs and cause upregulation of collagen type II and COMP gene expression, when being cultured in chondrogenic medium. This process could be enhanced by application of TGF-β1 or bone morphogenetic protein 2 (BMP-2) (107). Although sparse data was provided, the success of bone marrow MSCs would arouse our great excitement in the MSC-like dermal stem cells for tissue engineering.
MSC-like dermal stem cells also exhibited potential application to cell reprogramming. The elite and stochastic models have been considered as the mechanism of induced pluripotent stem cell (iPSC) generation (146). The stochastic model purports that every cell type has the potential to be reprogrammed to be an iPSC (100). However, the differentiation stage and the starting cell types were also reported to affect the generation efficiency (14,26), which supports the elite model proposing that iPSC generation occurs only from a subset of cells. Dermis is one of the most abundant resources of such cell subsets. A study showed that the human dental dermal tissue isolated mesenchymal-like stem/progenitor cells could be reprogrammed into iPSCs, and the efficiency was much higher than fibroblasts (148). Kuroda et al. isolated a population of SSEA-3+ adult stem cells (termed muse cells) from cultured skin fibroblasts (69). Further studies showed that the muse cells scattered sparsely in the connective tissue in dermis. When the four factors OCT3/4, Sox2, Klf4, and c-Myc were transferred into the muse cells and nonmuse cells, only the muse cells formed iPSCs (137). The muse cells seemed to be the initial cells of the iPSCs (124,135,137).
Dermal Hematopoietic Cells
In mammals, skin and bone marrow share many similarities. Both of them contain a large amount of stem cell types. In bone marrow, MSCs are the keystone of the hematopoietic stem cell niche (32). As mentioned above, a large amount of MSC subpopulations also resided in the skin. Skin might be another source of hematopoietic stem cells. Ectopic extramedullary hematopoiesis (EMH), defined as the formation of blood cells outside the bone marrow, usually occurs in a scenario of chronic anemia when, even after conversion of the bony yellow marrow to red marrow, the body is still unable to meet the demand for blood cells. EMH most commonly occurs in the liver and spleen but, in fact, occurs almost anywhere in the body. Skin is one of the ectopic hematopoiesis tissues in the body, which has escaped the attention of the majority of the hematopoietic community. A few decades ago, it was reported that a patient with myelofibrosis had a complication of cutaneous EMH. A soft bluish nodule like a hemangioma was found in the skin. Microscopically, both myeloid and erythroid cells, but not megakaryocytes, were found (68). Later, more cases were reported to exhibit cutaneous EMH in primary myelofibrosis. In the cutaneous EMH compartments, all three marrow elements (myeloid, erythroid, and megakaryocytic series) have been present (21,22,30,83,84,94,98,101,103-105). Recently, a study analyzed the related genes and found that the Janus kinase 2 gene mutated, which might be responsible for the cutaneous EMH (31). Besides the reports above, some interesting experiments have been carried out to verify the hematopoiesis in the skin. Gurevich et al. implanted the homologous demineralized tooth matrix subcutaneously with human BMP-2, and 1 month later, osteohemopoietic foci similar to the bone marrow of the skeletal bones by the cellular composition and morphological parameters were found in the skin. All three directions of myelopoiesis (myeloid, erythroid, and megakaryocytic series) were detected in the osteohemopoietic foci. The induced hemopoietic foci functioning continued at least for 1 year (41). Recently, another study demonstrated that subcutaneous transplants of bone marrow MSCs were capable of generating ectopic bone and organizing functional hematopoietic marrow elements in immunocompromised mice (147).
All of the above studies implied that hemopoietic stem cells might reside in the skin (Table 3). Previously, we found that DMCs were able to promote the recovery of the hematopoietic system (118,153,154,156). The mechanism might be that DMCs reconstruct the microenvironment of the hematopoietic system and secrete growth factors to help the growth of the residual hematopoietic stem cells. Lako et al. also reported that the hair follicle DP and DS cells could be induced into cells of erythroid and myeloid lineages. Furthermore, transplanted hair follicle cells reconstructed the bone marrow hemopoietic system in lethally irradiated mice. One year after transplantation, colony assays from bone marrow of primary recipients demonstrated that over 70% of clonogenic precursors were derived from donor hair follicle cells. When bone marrow from primary mice was harvested and used to repopulate secondary myeloablated recipients, multilineage hematopoietic engraftment was also observed (70). The CD45+ cells that were identified in DP also showed a high expansion and macroscopic colony-forming abilities in medium with GM-CSF and granulocyte colony-stimulating factor (G-CSF) (114). Meindl et al. isolated a population of cells expressing CD45, Sca-1, CD34, and CD117 from cultured mouse dermal cells. Further study showed that the majority of the CD45+ dermal cells were mast cell precursors. After lethal irradiation, mice injected with the CD45+ dermal cells survived up to 44 weeks. Donor dermal cells were detected in nearly all the hematopoietic tissues (bone marrow, spleen, liver, lymph nodes, and thymus) between 2 and 11 months posttransplantation (82).
The Biological Characteristics of Dermal Hematopoietic Cells
Abbreviations: DMCs, dermis-derived multipotent stem cells; CFU-GM, colony-forming units-granulocyte-macrophage; ND, not determined.
Cell reprogramming would be another way to generate hematopoietic cells from the dermis. The dermal cells would be firstly reprogrammed into iPSCs by definite factors, and then hematopoietic stem/progenitor cells would be induced from the iPSCs in proper conditions (39,51,73,111,125. However, this process was very complex to finish by establishing stable pluripotent cellular intermediates. Thus, another relatively simple method was developed by directly converting fibroblasts into hematopoietic cells. PU.1 and C/EBP-α have been used to induce dermal fibroblasts directly into macrophage-like cells. These fibroblast-derived macrophage-like cells displayed macrophage functions, including phagocytosing small particles and bacteria, mounting a partial inflammatory response, and exhibiting strict CSF-1-dependent growth (27). Another study revealed that ectopic expression of OCT4 could activate human skin-derived fibroblasts expressing hematopoietic transcription factors CD45 together with specific cytokines (stem cell factor, G-CSF, fms-related tyrosine kinase 3 ligand, IL-3, IL-6, and BMP-4). These fibroblast-derived CD45+ cells generated granulocytic, monocytic, megakaryocytic, and erythroid lineages in vitro, and demonstrated engraftment capacity in vivo (123). According to Ono's study, functional megakaryocytes/platelets could be directly converted from the skin-derived fibroblasts by transferring a combination of nuclear factor erythroid-derived 2 p45 unit (p45NF-E2), Maf G, and Maf K (92). Sandler et al. fused embryonic human fibroblasts with human fetal liver CD34+ cells, and the hybrid cells showed the potential to differentiate into several hematopoietic lineages (108).
Implications and Prospects
Ideal stem cell sources are thought to be easily accessible, capable of rapid expansion in culture, immunologically compatible, and amenable to stable differentiation or transdifferentiation. Dermis-derived stem cells have evoked great excitement in regenerative medicine. Skin is the largest organ of the body, and the dermis-derived stem cells have a high potential for expansion. In addition, the subpopulations of stem cells within the dermis could be directly applied to a large range of medical fields (Fig. 1). Neural crest stem cells derived from the dermis, for instance, were proven to be a good candidate for neural repair and neurogenesis. Furthermore, dermis-derived stem cells are highly plastic. They could not only differentiate into all ectoderm, mesoderm, and endoderm lineages but also show great potential in dedifferentiation or transdifferentiation. The last but not the least, dermis-derived stem cells also possess immunosuppressive activity, which suggests their potential application to prevent transplant rejection and treat autoimmune diseases. All of the above characteristics promise the therapeutic implications of dermis-derived stem cells in future regenerative medicine. However, the current literature is often difficult to interpret due to uncertainty about the cells being used. The underlying mechanisms of the dermis-derived stem cells in tissue repair, anti-inflammation, tissue engineering, and even hematopoietic recovery remain to be further clarified. Another consideration of the therapeutic implications of dermis-derived stem cells is the risk of tumorigenesis. Although no tumor formation in the animal model and clinical trails of MSCs has been shown, the expansion and enrichment of the stem cells before transplantation might add to the risk of malignant transformation. Genetically unmodified murine MSCs have been shown to undergo abnormal chromosomal changes (fusion, fragmentation, and ring formation) even at early passages and then form malignant tumors in vivo (57). Our study also demonstrated that DMCs could undergo spontaneous transformation after long-term in vitro culture (119). Further, the iPSCs also sustained the risk of tumor formation. Taken together, controlling dermis-derived stem cell-based clinical applications is a challenge that necessitates a thorough understanding of the molecular and cellular events involved in development, regeneration, and tumor initiation. The development of the epigenetics might provide novel insights in the underlying mechanisms. Recently, the technology to generate iPSCs and directly converted cells from dermal cells has made great progress (50,53,93,134), which provides a new platform for future research and applications of dermis-derived stem cells.

The potential therapeutic applications of dermis-derived stem cell subpopulations. Abbreviations: iPS, induced pluripotent stem; MSC, mesenchymal stem cell; MHC, major histocompatibility complex.
Footnotes
Acknowledgment
This work was supported by State Key Basic Research Development Program (2012CB518103), Natural Science Foundation Programs (81372727 and 81000831), and Program of New Century Excellent Talents in University (NCET-11-0869) from Ministry of Education and Innovation team building program of Chongqing University (KJTD201338). The authors declare no conflicts of interest.
