Adipose-derived mesenchymal stem cells and regenerative medicine


Author to whom all correspondence should be addressed.



Adipose tissue-derived mesenchymal stem cells (ADSCs) are multipotent and can differentiate into various cell types, including osteocytes, adipocytes, neural cells, vascular endothelial cells, cardiomyocytes, pancreatic β-cells, and hepatocytes. Compared with the extraction of other stem cells such as bone marrow-derived mesenchymal stem cells (BMSCs), that of ADSCs requires minimally invasive techniques. In the field of regenerative medicine, the use of autologous cells is preferable to embryonic stem cells or induced pluripotent stem cells. Therefore, ADSCs are a useful resource for drug screening and regenerative medicine. Here we present the methods and mechanisms underlying the induction of multilineage cells from ADSCs.


Recently, adipose tissue has been reported as a source of adult stem cells (Hombach-Klonisch et al. 2008). A high yield of adipose-derived mesenchymal stem cells (ADSCs) can be obtained with minimal discomfort under local anesthesia (Casteilla & Dani 2006). Following reports by Zuk et al. (2001, 2002), many studies have examined the characteristics, plasticity, and inducibility of ADSCs. Embryonic mesoderm-derived adipose tissue comprises a heterogeneous population of smooth muscle cells, fibroblasts, adipocytes, mast cells, and endothelial cells (Hausman 1981; Hausman & Campion 1982; Pettersson et al. 1984). ADSCs are adherent in vitro, maintaining their mesenchymal phenotype and plasticity toward the mesenchymal lineage after many passages in culture. These cells have been molecularly characterized using a panel of multiple mesenchymal differentiation markers (Lee et al. 2004; Dicker et al. 2005; Wagner et al. 2005; Oedayrajsingh-Varma et al. 2006; Table 1). In addition, they have been found to differentiate into multiple cell types in vitro, including adipocytes, chondrocytes, osteoblasts, cardiomyocytes, and vascular endothelial cells (Majumdar et al. 2000; Zuk et al. 2001; Halvorsen et al. 2000, 2001; Rangappa et al. 2003; Planat-Benard et al. 2004b; ; Konno et al. 2010). Moreover, ADSCs reportedly exert positive effects on patients with graft-versus-host disease occurring after bone marrow transplantation, suggesting an immunomodulatory function (Lombardo et al. 2009). In this review, we describe the methods and mechanisms underlying ADSC differentiation into multilineage cells.

Table 1. Molecular phenotype of adipose tissue-derived mesenchymal stem cells (ADSCs). The ADSCs expression of surface markers and genes is summarized
Positive cellular markers and genes
CD9, CD10, CD13, CD29, CD44, CD49, CD54, CD55, CD59, CD73, CD90, CD105, CD106, CD146, CD166, ASMA, Collagen-1, Endomucin, Fibronectin, Vimentin
Negative cellular markers and genes
CD11b, CD14, CD19, CD31, CD34, CD45, CD79a, CD80, CD177, CD133, CD144, HLA-DR, HLA II, c-Kit, Lin, MyoD88, STRO-1

Neural tissue differentiation

Many pathologies once believed to be incurable can now be definitively cured. However, ischemic and degenerative diseases of the central nervous system and traumatic injuries of the spine still lack a therapeutic approach capable of restoring lost function. Therefore, identification of cells capable of neuronal differentiation is of marked interest (Reynolds & Weiss 1992; Richards et al. 1992). Safford et al. (2002) reported neuronal differentiation by ADSCs, showing that murine and human ADSCs differentiate into neuronal tissue when cultured with valproic acid, forskolin, hydrocortisone, and insulin. Under these conditions, murine ADSCs expressed the neural tissue markers nestin, neuronal nuclei (NeuN), and glial fibrillary acidic protein. In human ADSCs, pretreatment with basic fibroblast growth factor (bFGF) and epidermal growth factor (EGF) for 7 days enhanced immunohistochemical changes observed during neuronal induction, causing adoption of a bipolar neuronal phenotype with strong IF-M, NeuN, and nestin expression (Safford et al. 2002). ADSCs from adult donors were treated with EGF and bFGF as neurospheres. The spheres were able to proliferate and induce Schwann and glial-like cells (Zavan et al. 2010). Ashjian et al (2003) reported that human ADSCs differentiate into cells resembling early neurons and glia in neural induction medium containing fetal bovine serum (FBS), insulin, indomethacin (INDO), and isobutylmethylxanthine (IBMX). After 14 days of differentiation, human ADSCs transformed into tropomyosin receptor kinase-A-, vimentin-, NeuN-, and neuron-specific enolase (NSE)-positive cells, adopted a neuron-like morphology, displayed voltage-dependent outward currents active after a brief delay, and did not become inactive during depolarization. Moreover, their steady-state current–voltage relationship exhibited outward reflection, similar to classic, delayed rectifier K+ channels in the mammalian node of Ranvier (Ashjian et al. 2003). The neural induction medium contained three active ingredients: insulin, INDO, and IBMX. Insulin has been shown to promote the maturation of differentiating neocortical cells in rat brains. INDO, a cyclo-oxygenase inhibitor, promotes neural cell survival after ischemic central nervous system injury, and IBMX, a phosphodiesterase inhibitor, increases intracellular cyclic adenosine monophosphate (cAMP), a neural stimulant. However, Ashjian et al. (2003) did not explain the necessity of all three agents. Subsequently, Ning et al. (2006) demonstrated the effect of each of these agents and their respective roles. IBMX induced morphological changes in ADSCs, similar to those induced by all three agents, whereas insulin and IBMX combined had little or no effect. The combination of IBMX and INDO or insulin exerted an effect similar to that of IBMX alone, while the combination of INDO and insulin had little or no effect. ADSCs treated with IBMX alone expressed the neuronal marker NF70 (Ning et al. 2006). Furthermore, 35% of ADSCs treated with IBMX and platelet-poor plasma, a specific inhibitor of insulin-like growth factor 1 (IGF-I) signaling, assumed a neuron-like morphology compared with approximately 95% of cells treated with IBMX. IBMX causes phosphorylation of the IGF-I receptor (IGF-IR) at tyrosine 1136 (Y1136). These results suggest that IBMX-induced neuron-like ADSC differentiation is mediated by IGF signaling through phosphorylation of the IGF-IR receptor at Y1136 (Ning et al. 2008). There are several types of stem cells, including bone marrow derived mesenchymal stem cells (BMSCs), umbilical cord blood cells, and ADSCs. These are attractive stem cell sources for clinical therapies (Kern et al. 2006). Zhang et al. (2012) undertook a comparison of rat ADSCs and BMSCs isolated from the same donor in terms of their proliferation capacity, potential towards neural differentiation, and ability to secret neurotrophin. ADSCs and BMSCs were cultured in neurobasal medium supplemented with EGF, bFGF, and B27 to form neurospheres. The neurospheres were then cultured in neurobasal medium supplemented with all-trans retinoic acid, FBS, horse serum, and N2, on poly-L-lysin and lamina double-coated dishes. Under these conditions, the population of ADSCs expressing nestin protein was significantly greater than that of BMSCs under undifferentiated conditions. Moreover, the expression of neural and glial markers in ADSCs was significantly higher than that of corresponding markers in BMSCs following neural differentiation.

Skin maintenance

As individuals age, the skin undergoes changes such as irregular pigmentation, thinning, and loss of elasticity. These occur because of both genetic and environmental factors and may worsen, progressing to precancerous and cancerous diseases. Various medical treatments and topical cosmeceuticals have been used to treat some symptoms of photoaging; however, the results have been less than satisfactory. Recently, the production and secretion of growth factors was reported as an essential function of ADSCs, and diverse regenerative effects of ADSCs have been demonstrated in the skin. ADSCs display multilineage developmental plasticity and secrete various growth factors such as vascular endothelial growth factor (VEGF), IGF, hepatocyte growth factor (HGF), and transforming growth factor (TGF)-β1. These proteins control and manage damaged neighboring cells. Recently, the production and secretion of growth factors have been identified as essential functions of ADSCs, and diverse rejuvenation effects of ADSCs on the skin have also been demonstrated (Kim et al. 2007; Kim et al. 2008a, 2008b, 2009; Park et al. 2008). For example, we verified that ADSCs and conditioned media from ADSCs (ADSC-CM) stimulated both collagen synthesis and migration of dermal Wbroblasts during the wound-healing process (Kim et al. 2007). ADSCs and ADSC-derived secretory factors also protected dermal Wbroblasts from oxidative stress induced by chemicals and UVB irradiation (Kim et al. 2008a). In addition, ADSC-CM inhibited melanogenesis by downregulation of tyrosinase and tyrosinase-related protein-1 (TRP-1) expression in B16 melanoma cells (Kim et al. 2008b). ADSC-CM can be applied in the field of biotechnology for use in cosmetic skin care products as well as in the protein drug industries. Advanced ADSC protein extract (AAPE), which is a conditioned medium cultured under hypoxia with ADSCs was evaluated. Human keratinocytes (HK) play an important role in the processes of skin biology, such as wound re-epithelialization and wound healing and restoration of normal skin (Mansbridge 2008; Nolte et al. 2008; Wright et al. 2009). Keratinocytes with normal dermal fibroblasts lead to the upregulation of mRNA for collagen type I and III and increase fibroblast proliferation and extracellular matrix (ECM) accumulation (Xia et al. 2004). Therefore, keratinocyte proliferation and migration is essential for these functions to take place on the skin surface. Moon et al. (2012) reported the biological function of AARP on HK in vitro and the components of AAPE through proteome and antibody array analysis. HK proliferation was significantly higher in their experimental group than in their control group. DNA gene chips demonstrated that AAPE in keratinocytes notably affected the expression of 290 identified transcripts associated with cell proliferation, cycle, and migration. Increased keratinocyte wound healing and migration was also shown in the experimental group. AAPE treatment significantly stimulated stress fiber formation, which was linked to the RhoA-ROCK pathway. They identified 48 protein spots in two-dimensional gel analysis. Selected proteins were divided into 64% collagen components and 30% noncollagen components as shown by matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) analysis. Antibody array results contained growth factor/cytokines such as HGF, FGF-1, granulocyte-colony stimulating factor (G-CSF), colony-stimulating Factor (GM-CSF), interleukin (IL)-6, VEGF, and TGF-β3, thus differing from the results of 2-D analysis. On the basis of these findings, we can conclude that AAPE activates HK proliferation and migration. These results highlight the potential of topical AAPE in the treatment of skin regeneration (Moon et al. 2012).

Vascular endothelial cell differentiation

Endothelial dysfunction is common to conditions such as coronary artery disease (Tousoulis et al. 2006), diabetes mellitus (Hartge et al. 2007), stroke (Slevin et al. 2006), erectile dysfunction (Gholami et al. 2003), and Peyronie's disease (Agrawal et al. 2008). A strategy to restore endothelial function is therapeutic angiogenesis, wherein proangiogenic agents promote the revascularization of ischemic tissue (Vartanian & Sarkar 2007). Adipose tissue contains a CD34-positive/CD31-negative stromal vascular fraction (SVF), thus exhibiting endothelial progenitor cell characteristics (Miranville et al. 2004; D'andrea et al. 2008). SVF cells demonstrate high proliferative capacity in endothelial growth medium containing IGF-I and VEGF. Under these culture conditions, SVF cells adopt a spindle-shaped morphology. Immunocytochemical analysis has revealed that SVF cells express the vascular endothelial markers CD31 and von Willebrand factor. In addition, a quantitative analysis of laser Doppler imaging findings revealed a time-dependent increase in blood flow after the injection of CD34-positive/CD31-negative human ADSCs into ischemic nude mice (Miranville et al. 2004; D'andrea et al. 2008). Cao et al. (2005) reported a similar effect of Flk1-positive/CD34-negative ADSCs on hind limb ischemia. Moon et al. (2006) also demonstrated ADSC differentiation into endothelial cells in vitro and their proangiogenic action in hind limb ischemia in nude mice. Moreover, during ADSC culture expansion, they observed downregulated CD34 expression. The laser Doppler perfusion index was significantly higher in CD34-, Flk-1-, and CD31-ADSCs-transplanted groups than in controls, even in ischemic hind limbs (Moon et al. 2006). Histological examination revealed that ADSC transplantation recovered muscle injury and increased vascular density compared with control treatment. The effects of ADSCs correlated with transplanted cell numbers but not with CD34 expression (Moon et al. 2006). To elucidate the mechanism of endothelial ADSC differentiation, human and mouse ADSCs were cultured in vascular endothelial differentiation medium with or without bFGF, EGF, VEGF, IGF-I, hydrocortisone, heparin, and vitamin C (Ning et al. 2009; Konno et al. 2010). Of these growth factors, VEGF, EGF, and IGF-I omission had no effect, but bFGF omission greatly diminished the ability of ADSCs to upregulate low-density lipoprotein (Ning et al. 2009) and endothelial marker gene expression (Konno et al. 2010). This finding suggests that bFGF signaling is important for differentiation of ADSCs into vascular endothelial cells.

Myoblast differentiation

Effective treatments are still lacking for muscle disorders such as Duchenne muscular dystrophy (DMD). Mesenchymal stem cells (MSCs) may constitute an attractive cell therapy alternative because they are multipotent and accessible in adult tissues. DMD is an X-linked, recessive, lethal, inherited disease caused by mutations of the DMD gene located at Xp21. DMD is associated with the highest morbidity of neuromuscular disorders. The main symptoms are progressive extremity myasthenia, pseudohypertrophy in the gastrocnemius muscle, and extensive muscular atrophy. Most patients die around the age of 20 because of respiratory failure and/or cardiac insufficiency (Wagner et al. 2005). The cause of DMD is mainly the lack of dystrophin on the muscular cell membrane. ADSCs can differentiate into muscle cells in vitro and in vivo. Therefore, ADSCs are also considered to be a candidate cell that can be transplanted into dystrophin-deficient mdx mice or DMD patients (Rodriguez et al. 2005; Goudenege et al. 2009).

Rodriguez et al (2005) demonstrated that human ADSCs (CD44+, CD49b+, CD105+, CD90+, CD13+, Stro-1-, CD34, CD15, CD117, Flk-1, gly-A, CD133, HLA-DR, and HLA-Ilow) transplanted into mdx mice, an animal model of DMD, results in substantial expression of human dystrophin in the injected tibialis anterior and the adjacent gastrocnemius muscle.

Subsequently, rat ADSCs cultured in media containing 5-azacytidine were reported to differentiate into MyoD+ multinucleated myotube cells (Vieira et al. 2008; Fu et al. 2010). Desmin and myosin were detected using immunofluorescence 21 days after induction. The expression of both proteins peaked at day 28 (Fu et al. 2010). However, under myogenic in vitro culture conditions, human ADSCs fail to reproducibly fuse into myotubes, although myogenic markers are expressed to some extent. Therefore, human ADSCs cells cannot fully differentiate autonomously into the myogenic lineage. Goudenege et al. (2009) demonstrated that the expression of MyoD was forced in human ADSCs. The results of his experiment revealed that the myogenic potential of human ADSCs was drastically increased in vitro and in vivo. Recently, to increase the myogenic differentiation rate, decorin, a small leucine-rich proteoglycan component of the ECM of all collagen-containing tissues (Fisher et al. 1989), was overexpressed in rat ADSCs (DCN-ADSCs). Decorin inhibits the activity of the TGF-β superfamily, improving muscle healing. Nishimura et al. (2007) found that decorin binds myostatin and sequesters it in the ECM during rat skeleton development. Decorin has also been found to neutralize the effects of myostatin in myoblasts (Zhu et al. 2007). Miura et al. (2006) found that immobilized decorin sequesters myostatin in the ECM and prevents its inhibitory action on myoblast proliferation in vitro. Kishioka et al. (2008) showed that decorin overexpressing cells (C2C12) exhibited an increased rate of proliferation compared with control cells. A DCN-ADSC group cultured in media containing 5-azacytidine resulted in an evidently larger single-fiber area, with an increase of 11.7%, compared with nontransplanted and GFP-ADSC groups. Hematoxylin-eosin staining demonstrated a slight decrease in the number of centrally nucleated fibers (CNFs). The proportion of CNFs in control mdx mice was approximately 70.9%, decreasing to approximately 61.6% at 20 weeks after transplantation. However, it dropped to 46.3% in the DCN-ADSC group. Electron microscopy of the mice gastrocnemius revealed better histological manifestations in the DCN-ADSC mdx group (Geng et al. 2012).

Cardiomyocyte differentiation

Cardiomyocyte differentiation mainly occurs neonatally and perinatally. In adults, the regenerative potential of cardiac tissue is limited and insufficient to repair damage caused by pathological conditions such as myocardial infarction (Pfeffer & Braunwald 1990); however, cell transplantation can overcome this issue (Orlic et al. 2002). Planat-Benard et al. (2004a) showed that the SVF of primary human ADSC cultures spontaneously differentiates into cells with morphologic, molecular, and functional properties of cardiomyocytes. These cells express cardiomyocyte marker genes, including Gata4, Nkx2.5, Mlc-2v, and Mlc-2a and display spontaneous and triggered action potentials under current clamps.

Chondrogenic differentiation

Osteoarthritis as a result of lesions in the articular cartilage is the most common cause of disability, after trauma or degenerative joint diseases. Cartilage tissue is an avascular aneural tissue with limited capacity for self-repair (Wang et al. 2005). Current treatment methods for cartilage tissue injuries include chondral shaving, subchondral drilling, tissue debridement, microfracture of the subchondral bone, and transplantation of autologous and allogeneic osteochondral grafts. The most common clinical treatment for an articular cartilage lesion is the creation of a bleeding microfracture within the subchondral bone in an attempt to stimulate cartilage regeneration (Gillogly et al. 1998). While some of these approaches show promise, many lead to the formation of fibrous tissue, apoptosis, and further cartilage degeneration (Furukawa et al. 1980; Tew et al. 2000; Mitchell 2004). In addition, the repaired tissue in these circumstances lacks the biomechanical properties of normal articular cartilage, leading to inconsistent long-term outcomes (Minas & Nehrer 1997).

Recently, many efforts have been undertaken to repair articular cartilage lesions by transplantation of human chondrocytes. A major problem associated with human chondrocytes in culture is the phenomenon of dedifferentiation. In addition, autologous chondrocyte transplantation has been associated with significant donor side morbidity and the initiation of osteoarthritic changes in the joint (Lee et al. 1997). Therefore, an alternative cell source is necessary for cartilage repair.

The induction of chondrocytes using ADSCs appears to require a rounded cell shape through pellet culture or a three-dimensional hydrogel (Awad et al. 2004). A pellet culture system originally described as a method for preventing the phenotypic modulation of chondrocytes in vitro (Holtzer et al. 1960) focuses on cell–cell interactions, analogous to what occurs in precartilage condensation during embryonic development (Fell 1925). However, this cell configuration is not sufficient for the induction of chondrogenesis. The chondrogenic differentiation of ADSCs requires certain growth factors. Bone morphogenetic protein 2 and 3 (Johnston et al. 1998; Majumdar et al. 2001) as well as TGF-β 1 and 2 (Mackay et al. 1998; Minguell et al. 2000) can induce the differentiation of ADSCs into chondrocytes under certain culture conditions.

Adipose-derived mesenchymal stem cells also have strong immunosuppressive properties (Prunet-Marcassus et al. 2006; Niemeyer et al. 2007), and their therapeutic effects have been proved in an experimental animal model of rheumatoid arthritis (RA; Gonzalez et al. 2009). RA is characterized by an excessive immune response accompanied by progressive joint tissue destruction. In a rheumatic joint, there is an imbalance between cartilage regeneration and degradation. Fibroblast-like synoviocytes (FLS), macrophage-like synoviocytes, and chondrocytes secrete aggrecanases and MMPs, which are enzymes capable of cleaving cartilage constituents. Local synthesis of tissue inhibitors of MMPs is downregulated and the expression of genes supporting chondrogenesis is also decreased (Andreas et al. 2009). Skalska et al.(2012) investigated whether the chondrogenic function of ADSCs from rheumatoid joints is maintained and whether TNF, the key cytokine in RA pathogenesis, may have specific influence on this type of cell. The results of alcian blue staining showed effective chondrogenesis in RA and healthy ADSCs. TNF inhibited GAG deposition in both RA and healthy samples. Sox9, Acan, and Col2a mRNA expression significantly increased in chondrogenic-medium-treated cells and decreased after TNF exposure. No statistically significant differences were observed between RA and healthy ADSCs (Skalska et al. 2012).

Adipogenic differentiation

Human and mouse ADSCs can differentiate into adipocytes. Adipocyte regeneration is useful for the reconstruction of breast tissue after surgery for breast cancer and asymmetry as well as soft tissue and subdermal defects created by trauma, surgery, or burn injury (Zuk et al. 2002; Rodriguez et al. 2005; Choi et al. 2006; Konno et al. 2010). ADSCs cultured in media containing IBMX, insulin, dexamethasone, mesenchymal cell growth supplement (MCGS), and L-glutamine for 2 weeks became lipid-retaining cells stained by oil red O (Konno et al. 2010).

Pancreatic β-cell differentiation

Insulin-dependent diabetes mellitus (IDDM) is characterized by the rapid development of severe metabolic abnormalities caused by insulin deficiency (Gepts 1965; Herold et al. 2005; Keymeulen et al. 2005; Santana et al. 2006; Voltarelli et al. 2007). IDDM results from deficient insulin secretion due to β-cell destruction in the pancreatic islets of Langerhans. Pancreatic β-cells do not regenerate and xenotransplantation is problematic; therefore, IDDM is impossible to cure using standard tissue failure therapies. Transplantation treatment was established by Shapiro et al. (2000), and many patients receiving islet transplantation required no further therapy (Robertson 2004). However, islets need to be obtained from donors, and resources are restricted. Furthermore, not all patients are eligible for transplantation. Regeneration of insulin-producing cells was thus explored to address these problems. Okura et al. demonstrated that human ADSCs differentiate into islet-like cluster cells under the following sequential culture conditions:

  1. Dulbecco's modified eagle medium (DMEM) (60%) and MCDB 201 (40%) medium containing EGF, dexamethasone, ascorbic acid, and FBS
  2. Knockout DMEM, including FBS, glutamine, and nonessential amino acids
  3. ITS- and fibronectin-supplemented DMEM/F-12
  4. DMEM/F-12 with N-2 supplement, B-27 supplement, and bFGF
  5. Glucose-free DMEM/F-12 with N-2 supplement, B-27 supplement, nicotinamide, and exendin-4.

The generated islet-like clusters secrete insulin on glucose and nonglucose secretagogue stimulation and express molecules characteristic of pancreatic β-cells, including Isl-1, Pax4, Pax6, Pdx1, PC1/3, Pc2, Kir6.2, Glut2, glucokinase, and insulin (Okura et al. 2009). Kajiyama et al. (2010) suggested that the generation of mature pancreatic β-cells is unnecessary. Transplantation of Pdx1-overexpressing ADSCs, which do not exhibit pancreatic β-cell-like characteristics in vitro, into mice with streptozotocin-induced IDDM caused engraftment of transplanted Pdx1-overexpressing ADSCs in the pancreas, amelioration of hyperglycemia, and improvement of survival rate (Kajiyama et al. 2010).

Hepatocyte differentiation

Most liver diseases cause hepatocyte dysfunction and potential organ failure. Liver transplantation may be the only treatment for severe liver damage; however, few livers are available and rejection remains an unresolved problem. Replacement of diseased hepatocytes with stem cells is the main aim of liver-directed cell therapy. Growing evidence suggests that reservoirs of stem cells capable of transdifferentiation from one phenotype to another may reside in several types of adult tissue (Kashofer & Bonnet 2005). Reportedly, CD13-, CD34-, CD45-, CD90-, and CD105-positive cells, evident among ADSCs, differentiate into hepatocyte-like cells when cultured with HGF, bFGF, and nicotinamide (Talens-Visconti et al. 2007). Differentiated human ADSCs adopt hepatocyte morphology and express the hepatocyte cell markers albumin, transthyretin, cytochrome 2E1, and enhancer-binding protein β; however, they do not express cytokeratin 18 or 19 (Talens-Visconti et al. 2007). Banas et al. (2007) selected ENG-positive ADSCs to achieve a multipotent and homogenous cell subpopulation. CD105-positive ADSCs were cultured with FGF1, FGF4, and HGF-containing medium for 3 weeks, followed by culture with oncostatin M and dexamethasone-containing medium for 2 weeks. ADSC-derived hepatocyte-like cells expressed ALB, TTR, TDO2, CYP7A1, and HNF4-α according to reverse transcription–polymerase chain reaction (RT–PCR) and ALB, CYP3A4, TTR, CYP3A4 according to Western blot analysis. To verify the function of these cells, the hepatocytes were transplanted into CCl4-injured mice. When transplanted cells were integrated into host livers, some liver functions improved, including a decrease in ammonia level and damaged liver marker glutamic-pyruvic aminotransferase (GPT) level in peripheral blood (Banas et al. 2007). Sgodda et al. (2007) demonstrated hepatocyte differentiation from ADSCs cultured in medium containing HGF, EGF, and TGF-α. Hepatocyte-specific transcripts of CD26, CYP1A1, and CX32 were detected. Transient transgenic luciferase expression was stimulated by cAMP when driven by the hepatocyte-specific promoter of the cytosolic PCK1 gene. ADSC-derived hepatocytes were transplanted into the liver and donor cells engrafted into the host liver parenchyma, predominantly in the periportal region (Sgodda et al. 2007). These findings indicate that ADSCs exhibit the potential to differentiate into hepatocyte-like cells in vitro and in vivo.

Conclusions and perspective

In recent years, induced pluripotent stem cells (iPSCs) have constituted an important breakthrough in the field of regenerative medicine. Tissue stem cells, ADSCs in particular, are equally useful in the field of regenerative medicine for the following reasons:

  1. ADSCs are multipotent, with the potential to differentiate into neurons, adipocytes, osteoblasts, cardiomyocytes, vascular endothelial cells, pancreatic β-cells, and hepatocytes in vitro and in vivo (Fig. 1). Methods of inducing each cell lineage are being established and mechanisms of differentiation elucidated.
  2. ADSC extraction is easy compared with that of other stem cells such as BMSCs. The procedure is minimally invasive and safe.
  3. There is no risk of immune rejection with autologous cells, unlike that with embryonic stem cells or iPSCs.
Figure 1.

Directly induced differentiation of adipose tissue-derived mesenchymal stem cells (ADSCs) the diverse differentiation ability of ADSCs is highlighted. This characteristic is based on the technology of defined factor-mediated direct differentiation of multipotent ADSCs into many cell types within the body. Several strategies for direct induction into three germ layers are depicted. AAPE, adipose-derived stem cell protein extract; BMP, bone morphogenetic protein; EGF, epidermal growth factor; FBS, fetal bovine serum; FGF, fibroblast growth factor; IBMX, isobutylmethylxanthine; IGF, insulin-like growth factor; INDO, indomethacin; MCGS, mesenchymal cell growth supplement; SVF, stromal vascular fraction; TGF, transforming growth factor; VEGF, vascular endothelial growth factor.

Therefore, ADSCs are a useful resource for drug screening and research in the field of regenerative medicine.


This work was supported by a Grant-in-Aid for Young Scientists (B) of the Japan Society for the Promotion of Science (M.K.), Akaeda Medical Research Foundation (M. K.) and SENSHIN Medical Research Foundation (H.I).