Stem cell markers are genes and their protein products used by scientists to isolate and identify stem cells. Besides, stem cells can also be identified by functional assays which are considered the gold standard for the identification and therapeutic purposes. Although functional assays are the ideal approach to define stem cell, their molecular markers provide a systematic approach to characterize a healthy and robust stem cell population. Our knowledge on the stem cell identification with respect to their therapeutic application is rather limited due to their extraordinary complexities, specificity, validity, and lack of specific molecular markers. In addition, marker profiles of stem cell population often fluctuate depending on their site of origin, species, and so-called stemness (totipotent vs. multipotent) of the desired. Despite the limited knowledge of marker functionalities, their unique expression pattern and timing provide us useful tool to identify and isolate stem cells. This review explores a number of marker systems for the identification/isolation and characterization of adult and embryonic stem cell in respect to their key applications in the cell-based regenerative medicine.

Embryonic stem cells (ES cells) are pluripotent stem cells derived from the inner cell mass of a blastocyst, an early-stage embryo [1] The ES cells have high potential to differentiate into a wide variety of cell types [2]. In other words, pluripotency, the foundation of mammalian development, endowed with the extraordinary potential to initiate and develop almost all lineages of the mature organism (except those that form a placenta or embryo) [3]. Undifferentiated embryonic stem (ES) and induced pluripotent stem (iPS) cells can be functionally defined by their ability to differentiate into cells derived from the three germ layers the ectoderm, mesoderm, and endoderm that eventually make up to all cell types at a later time. ES cells possess two distinct properties that make them an attractive choice for cell therapy. First, as embryonic stem cells originate from early blastocysts, a very early developmental stage, retain the extraordinary plasticity to become any cell type of approximately 200 cell types that constitute the human body [4]. Given the right combination of signals, embryonic stem cells will develop into mature cells that can function as neurons, muscles, bone, blood or other needed cell types. Another important feature of embryonic stem cells is their ability to remain in an undifferentiated state and to divide indefinitely. These so-called self-renewing cells generate unlimited well-defined, identical, genetically and genomically characterized stem cells [4].

Despite our growing knowledge about their enormous potential, we are limited to exploit ES cell-based therapeutic avenues. One of the challenging tasks is that markers that identify ESC also mark tumor stem cells [5]. Therefore, researchers face a formidable challenge in distinguishing ESCs, especially from TSCs. Another potential issue is that despite their embryogenic origin, human and mice differ in pluripotency mechanisms.

Two homeodomain transcription factors, Oct4 (POU5F1) and Nanog, were the first proteins identified as essential for both early embryo development and pluripotency maintenance in ES cells [6]. Emerging studies indicate that in addition to Oct4, Sox2, and Nanog, many other factors required for pluripotency have been identified, including Sall4, Dax1, Essrb, Tbx3, Tcl1, Rif1, Nac1, and Zfp281 [7]. These pluripotency factors regulate concomitantly to form a complicated transcriptional regulatory network in ES cells [8]. For example, Gifford CA et al confirmed the pluripotency of hiPCS with staining for NANO, OCT3/4, SOX2, TRA-1-60, and SSEA4 [9], J Shi et al with OCT4 and SOX2 antibodies [10], and H Haukedal et al with OCT4 and NANO antibodies.

TRA-1-60 and TRA-1-81 antigens on the human embryonal carcinoma (EC) cells and human pluripotent stem cell surfaces are widely used as markers in identifying and isolating ESCs [11, 12]. Besides, they are also routinely used to assess the pluripotency status of induced pluripotent stem (iPS) cells. They are also expressed in teratocarcinoma and EG cells [13, 14]. Both TRA-1-60 and SSEA4 are both expressed on human embryonal carcinomas and on human embryonic stem cells [11]. Upon differentiation, TRA-1-60 and SSEA4 expression levels decrease and SSEA1 expression increases on human embryonic stem cells over time when treated with Retinoic acid [14]. Besides, they also express CD349/frizzled-9, stage-specific embryonic antigen (SSEA)-4, Oct-4, Nanog, and nestin [15]. Oct-4 and Nanog, as well as several cell surface markers (SSEA-1, SSEA-4, TRA-1-60, and TRA1-81) have been used to characterize mouse and human embryonic stem cells (ESC) [15]. H Haukedal et al characterized hiPSC with a TRA1-81 antibody [16].

Conventionally, markers used for mESCs, mouse embryonic carcinomas (ECs), or human EC cells, were exploited to identify undifferentiated human embryonic stem cells (hESCs) [11, 12]. At the pre-implantation stage, murine embryos, human germ cells, and teratocarcinoma stem cells express certain molecular receptors known as Stage-Specific Embryonic Antigens (SSEA) on their membrane surface [17]. It is now widely recognized that SSEAs, sphingolipids, identified as a key player in identifying cells endowed with pluripotent and stem cell characteristics. During oogenesis and embryogenesis, SSEA-3 and SSEA-4 are expressed in undifferentiated primate ESC, human embryonic germ (EG) cells, human teratocarcinoma stem cells, and ESC [18]. Currently, Stage-specific embryonic antigen-3 (SSEA-3) and SSEA-4 have been recognized to characterize undifferentiated hESCs but not on undifferentiated mESCs [19, 20]. H Haukedal et al characterized hiPSC with an SSEA3 antibody [16].

A family of Frizzled proteins is expressed in mouse and human ESC [21, 22]. Wnt signals execute their functions via the Fzd family receptors by bind to Fzd and the co-receptors LRP5 or LPR6, and eventually activate the Wnt/β-catenin pathway. Cripto-1 plays a crucial role for early embryonic development and has been associated with the undifferentiated status of mouse ES and human ES cells. During development, Cripto acts as a receptor for TGF-β ligands, including GDF1 and GDF3 [23, 24]. In addition to having essential functions during embryogenesis, as an oncogene, Cripto is upregulated in tumors and promotes tumorigenesis [23, 24]. Frizzled (FZD) proteins comprise a family of transmembrane-spanning receptors (FZD1-10) activated by Wnt ligands. FZD9 can be activated by Wnt 2 and Wnt 8 via the canonical pathway, and by Wnt 7 via non-canonical signaling. Mouse FZD9 is present in the developing brain, in neural precursor cells in the developing neural tube, and in myotomes.

The fundamental implication of transcription factors in the pluripotency maintenance was clearly exemplified by the fact that pluripotent stem cells can be derived from mouse embryonic fibroblasts by inducing transcription factors expression. In addition to extrinsic factors, the pluripotency of ES cells also depends on intrinsic determinants, such as the expression of the POU transcription factor. As such, these induced pluripotent stem (iPS) cells were developed via the overexpression of a set of specific genes Oct4, c-Myc, Sox2, and Klf4. Abundant levels of Oct4, Sox2, and Klf4 reprogram fibroblasts into the iPS cells with a pluripotent state [25, 26]. OCT4, SOX2, NANOG, KLF4, and c-MYC are present in both pluripotent stem cells and cancer stem cells [27], although their subcellular localization profiles might be different [28].

A combination of either Klf4 or c-Myc with Oct4 is sufficient to generate iPS cells from NSC. Oct3/4 is specifically expressed in pluripotent stem cells. The Sox family of genes is associated with multipotent and unipotent stem cells. Sox1 induces iPS cells with similar efficiency as Sox2, and genes Sox3, Sox 15, and Sox 18 generate iPS cells, although with decreased efficiency. Klf5 has been implicated in the transcription of Oct3/4 and Nanog, and ESCs renewal and pluripotency maintenance [29]. Klf4 and Klf2 can regulate the expression of certain transcription factors: Nanog, Tcl1, Esrrb, Sall4, Tcf3, Mycn and Fbxo15 [30]. Nanog is a transcription factor and plays a crucial role in the maintenance of pluripotency and self- renewal in mouse and human ESCs and is downregulated upon ESC differentiation, which is consistent with an intimate association with pluripotent stem cell identity [30]. Nanog has been implicated in pluripotent ES and EG cells, as well as in both mouse and human EC cells.

Principally, fully reprogrammed human iPS cells, compared to transiting or incompletely reprogrammed cells, endowed with the important features (i) downregulation of CD13, a fibroblast marker, (ii) upregulation of expression of SSEA-4 and TRA-1-60 like pluripotent markers, (iii) silencing of viral transgenes (iv) endogenous expression of Nanog (v) a low Hoechst retaining or high Hoechst pump out potential. The expression pattern of the fully differentiated iPSCs was remarkably similar to normal or natural pluripotent stem cells ESCs. For instance, similar to expression pattern found in hESC, human iPSCs also express the SSEA-3, SSEA-4, TRA-1-60, TRA-1-81, TRA-2-49/6E, Oct4/POU5F1 [31], and Nanog [31]. Gene expression and genome-wide H3K4me3 and H3K27me3 were found to be extremely similar between ES and iPS cells. For example, Li Z et al confirmed iPS cell transformation from fibroblasts with Tra-1-81, Tra-1-60, SSEA-4 and Nanog labeling [32]. Lee J et al immunostained for NANOG, POU5F1, and SOX2 to examine the iPSC lines [33]. In mouse, iPSCs expressed genes expressed in undifferentiated ESCs, including Oct-3/4, Sox2, Nanog, GDF3, REX1, FGF4, ESG1, DPPA2, DPPA4, and hTERT. Similar to mESCs, mouse iPSCs do not express SSEA-3 and SSEA-4 but both cell types commonly express SSEA-1.

The precursors of primordial germ cells developed from 4–8 cells in E6.25 proximal epiblast express the transcriptional repressor Blimp1 [34]. Over time, these Blimp1-positive cells continuously proliferate and initiate the expression of Fragilis and Stella by E7.5 [35].

[enlarge]

Figure 1. Timeline of Germ Cell Marker Expression. At each germ cell developmental stage, a panel of germ cell markers are expressed at definite time points to drive germline differentiation in vivo. During embryonic development, the specification of primordial germ cells (PGCs) is crucial for the development of the germline, which eventually leads to totipotent zygote upon fertilization. Arrows indicate the period of overexpression of markers.

PGCs eventually trigger migration and colonization of the genital ridge with more cells by E12.5 [36, 37] The coordinated genetic and epigenetic events induce the PGC/germ cells to form male or female germ cells [35, 37]. Reports indicate that the expression of Blimp1, Stella, Fragilis, Piwil2, Dazl, and MVH in ES cells [35, 38], indicating the origin of ES cells could be germline. Depending on the stage, certain markers like Oct4, c-kit are expressed early and downregulated prior to mature Germline Specification stages. As indicated in the schematic figure 1, Tekt1 and GDF9 markers are induced only at later stages. Nanos has been implicated in all development stages from blastocyst to mature sperm or oocyte.

Ectoderm is one of the three primary germ cell layers in the very early embryo. The other two layers are the mesoderm (middle layer) and endoderm (most proximal layer), with the ectoderm as the most exterior (or distal) layer. Certain factors mark the ectoderm including Otx2, Chordin, p63/TP73L, FGF-8, Pax2. FoxJ3, Pax6, GBX2, SOX1, nestin, beta-tubulin, and noggin. Endoderm formation depends on two sequential positive feedback loops mediated by Cripto and Bmp4/Wnt3 that are activated by mature or uncleaved Nodal, respectively, to sustain Nodal signaling from implantation throughout gastrulation [39]. ENDM1 and Flk1 have been used as a definitive mouse endodermal cell marker and a mesoderm cell marker, respectively [40].

Besides what we discussed so far, there are several additional marker systems available to approach stem cell identification and/or isolation. These include enzymatic (alkaline phosphatase and telomerase)-based reaction, small molecules (lectins or short peptides), and quantum dots (QD) or fluorescence dyes, etc. The detailed descriptions of such markers systems are beyond the scope of this review.

It is increasingly evident that cell surface marker system coupled with FACS is the most common approach for CSC isolation for their functional exploitation.

On their surface, CSCs of different tissue origin, express certain receptors viz. CD44, CD133, CD24, CD90, CD271, CD49f and CD13. Among these, CD44, a transmembrane glycoprotein, is one of the prominent markers and have been used to identify CSCs of various solid tumors. For instance, the basal cells of normal bladder urothelium express CD44, which is a marker for bladder CSCs [41]. Similarly, CD44 has been implicated in CSCs of breast [42, 43], bladder [41], pancreas [44], gastric [45], prostate [46], head and neck [47], colon [48], ovarian [49]. In colon cancer, stable knockdown of CD44 diminishes the engraftment ability and in vitro clonogenicity of colon CSCs [50, 51]. CXCR4 is now being used in many CSC models including a pancreatic cancer model. In this, only CD133+CXCR4+ CSCs can form distal metastasis, suggestive of a functional role for CXCR4 in the metastatic process [52].

Cytokeratin, the intermediate filaments, can serve as a marker system for epithelial cells at different stages of cellular differentiation. The expression of these cytokeratins relies on the type of epithelium and their expression also changes during the course of normal epithelial cell differentiation [61]. However, cytokeratins are a type of intracellular proteins, which cannot be detected by FACS for prospective isolation of CSCs. Despite this limitation, cytokeratin markers can provide valuable supporting evidence in combination with a standard marker system to sub-fractionate populations of stem, progenitor and differentiated cells in solid cancers [47]. One of the most well-characterized cytokeratin expression patterns is available in skin, where CK15 expression is restricted to the bulge region multipotent stem cells, but replaced by CK5 [62]. As differentiation proceeds, CK5 is replaced by CK10 and eventually the terminal differentiation marker involucrin [47, 59]. Now it is understood that CK14 and CK5 are expressed in urothelial basal cells, CK8 and CK18 mark intermediate cells and CK20 are expressed only in terminally differentiated umbrella cells [41].

Despite the popularity of cell surface markers in prospective isolation of cancer stem cells (CSCs), other markers are also commonly employed to isolate CSCs. The Side Population (SP) discrimination assay is among them for the identification of certain CSCs. It is known that SP assay relied on the unique properties of cell subpopulation endowed with a high level of the ATP-binding cassette (ABC) family of transporters. This ABC reporter system has been implicated to rapidly efflux the dye Hoechst 33342 and leading to this SP phenotype. SP population which enriched for and largely overlaps with CSCs, has been identified in many cancers, such as prostate [63], bladder [62], breast [64], lung [45]. Aldehyde Dehydrogenase (ALDH) methodology was employed to identify stem cells and CSCs. Given that both normal and cancer stem cells express ALDH1, this expression pattern represents poor clinical outcome for breast cancer patients [65]. ALDH marker system has been exploited to isolate tumor cell with CSC properties in glioblastoma, lung, liver and a number of other cancers [65, 66].

Bone homeostasis is a dynamic process relying on the balance of deposition by osteoblasts and resorption by osteocytes. Osteogenesis is not only responsible for the continuous remodeling of bone tissue but also crucial for the maintenance of bone size, shape, and integrity. Disruptions of bone homeostasis accompany disorders that include osteoporosis, arthritis, and many inheritable skeletal diseases. Cell-type specific markers were developed to identify mesenchymal stem cells that have differentiated to osteoprogenitors, osteoblasts, or osteocytes. Certain osteoprogenitors can be distinguished by the expression pattern of TGF-beta, bFGF, BMP-2, and bFGF [67] or gremlin 1 [68]. Additionally, certain markers like ALPP, MCAM, collagen I, collagen I alpha 1, collagen II, RUNX2, decorin, and TPO are also used. A Marconi et al use Sox5/6/9 as the markers for perichondral progenitor cells [69].

Osteoblasts are the skeletal cells which constitute the extracellular matrix of bone, typically arise in the embryo. Runx2, a transcription factor, was identified upregulated in immature osteoblasts but downregulated in mature osteoblasts. It has been reported that osteoblast lineage was determined by Runx2, whereas Sp7 and canonical Wnt-signaling are downstream determinants of the fate of mesenchymal cells to osteoblasts [70]. Runx2 triggers the expression of major bone matrix genes during the early stages of osteoblast differentiation, but Runx2 is not essential for the maintenance of these gene expressions in mature osteoblasts. Some relevant factors also used as osteoblast markers: alkaline phosphatase/ALPP/ALPI, osteocalcin, BAP1, OPN, BAP31, osterix/Sp7, collagen I, SCUBE3, fibronectin, SPARC, and IGFBP-3 [71, 72]. Noda S et al measured osteoblastic differentiation from primary dental pulp stem cells with alkaline phosphatase (ALP) expression [73].

Osteocytes are the most abundant cells in bone, and their death by microdamage has been suggested to be the major event leading in the initiation of osteoclastic bone resorption. As osteocytes secrete several factors: TGF beta, RANKL, and MCSF, which may affect the recruitment and function of osteoclasts after being released in the marrow compartment [74] osteocytes produce sclerostin which negatively interacts with Wnt signaling and thereby inhibits bone formation [75]. Osteocytes also produce DKK which targets the Wnt signaling pathway to control bone formation negatively.

Myogenesis is the generation of muscle tissue when committed muscle stem cells or myoblasts proliferate into multinucleated myotubes in the embryo. In adult tissue, muscle-derived stem cells are a distinct population of cells from muscle satellite cells, which promote muscle regeneration in response to injury and disease. In the presence of FGF and other growth factors, myoblasts proliferate but do not differentiate. Following the depletion of growth factors, myoblasts stop dividing and secrete fibronectin onto their extracellular matrix to promote the subsequent stages of muscle cell fusion. A cocktail of the myogenic cell marker CD56, the endothelial cell marker UEA-1 receptor (UEA-1R), and the perivascular cell marker CD146 can be used to mark the myogenic precursor marker [76]. The isolation and identification of differentiated smooth (VE-cadherin, alpha-smooth muscle actin), skeletal (FABP3, Integrin alpha 7), and cardiac muscle (desmin, myosin heavy chain) require myogenesis markers. Liu L et al published a well-validated protocol for isolating skeletal muscle stem cells using FACS with VCAM+CD31−CD45−Sca1− markers [77, 78]. Those myogenic markers also represent an important tool for the detection of muscular cancers and for degenerative muscular diseases and dystrophies.

Neural stem cells (NSCs) are unique cells endowed with self-renewing, multipotent potential, responsible for the generation the main phenotypes of the nervous system. NSCs have the unique potential to produce clonally related progeny that upon differentiation, constitute the central nervous system developed by neurons, astrocytes, oligodendrocytes, and the ependymal cells. Many neural cell types were identified and isolated by using cell surface marker expression. CD133 is expressed on the surface of NSC and has been widely used to isolate NSC from human brain [79]. CD15, a stage-specific embryonic antigen-1 is now identified as a marker of NSC and radial glia from the subventricular zone (SVZ) in mice [80]. CD24 was used as a marker to enrich NSCs [81, 82]. Mauffrey P et al used PSA-NCAM, CD24 and EGFR to ascertain the neural origin of prostate stromal cells [83]. Peh et al utilized CD133, CD15, and GCTM-2 combination to enrich neurosphere-forming NSC from neural induction cultures [82]. Nestin and Sox-2 serve as useful markers which predominantly expressed in stem cells of the central nervous system (CNS) and also implicated in neural stem cell proliferation and differentiation. In addition to intracellular molecules, products are available to study proteins which are expressed at the cell surface, including ABCG2, FGF R4, and Frizzled-9. M Coolen et al identified neural progenitors in killifish using Musashi-1 (Msi1) and nestin [84].

Mesenchymal stem cells (MSCs) are multipotent mesoderm-derived progenitor cells with the capacity to differentiate into adipose, bone, cartilage, and muscle tissues providing wide-ranging therapeutic avenues. Adult mesenchymal stem cells can be isolated from the stroma of the bone marrow. MSCs express a panel of key markers including markers CD10, CD13, CD73, CD105, and CD271. Other markers include CD140b), HER-2/erbB2 (CD340), and frizzled-9 (CD349) [15]. Noda S et al characterized primary dental pulp stem cells with CD44, CD73, CD90, and CD105 [73]. MSC derived from bone marrow (BMMSCs) are postnatal stem cells capable of self-renewing and differentiating into osteoblasts, chondrocytes, adipocytes, and neural cells. Study report that BMMSCs express key factors: STRO-1, CD29, CD73, CD90, CD105, CD146, Oct4, and SSEA4 [85, 86]. However, BMMSCs do not express hematopoietic cell markers especially CD14 and CD34 [87].

Multipotent skin stem cells or epidermal stem cells are responsible for tissue homeostasis during normal cell turnover and wound healing. Such skin stem cells were found to reside within a niche within the hair follicle. Although, predominantly under quiescence or a slow cell cycle state, epidermal stem cells can be stimulated to proliferate and differentiate into the specialized cells that compose a hair follicle during wound healing. K15 protein expression not only marks an epidermal stem or progenitor cell subpopulation and also represents basal-like cells during the epidermal differentiation program [88] (please note, the commonly used k15 antibody EPR1614Y is specific for western blot, but not specific for immunocytochemistry and immunohistochemistry; antibody LHK15 should be used as the skin stem cell marker [89] ). Recent studies indicate that several proteins including CD34, nestin, follistatin, p63, integrin alpha 6, tenascin C, and frizzled factors are useful for the identification of epidermal stem cells [90], in addition to EGFR, IGFR, Delta1, and TBRII [91]. Y Shwartz et al, for example, isolated mouse hair follicle stem cells as CD45^-, integrin alpha 6⁺, CD34⁺ and Sca-1^- with a cell sorter and immunestained mouse hair follicle stem cells with P-cadherin [92].

Adipose tissue is composed primarily of adipocytes, cells that function to store energy in the form of lipid droplets. Additionally, pluripotent progenitor cells, called adipose-derived stem cells, can be identified in adipose tissue by the expression of markers such as Integrin family members, CD44, and ICAM-1/CD54. Like mesenchymal stem cells, multipotent adipose-derived stem cells have been shown to differentiate into cells of the mesodermal lineage including adipocytes, chondrocytes, osteoblasts, osteoclasts, and myoblasts in vitro. Other in vitro studies suggest that adipose-derived stem cells have additional plasticity and can transdifferentiate into cells of other germ layers such as hepatocytes and neuronal-like cells. Typically, adipose-derived stem cell markers (ASCs) express characteristic adhesion and receptor molecules, surface enzymes, extracellular matrix and cytoskeletal proteins, and proteins associated with the stromal cell phenotype. Reports indicate that the surface phenotype of ASCs resembles that of bone marrow-derived mesenchymal stem or stromal cells (MSCs) [93] and skeletal muscle-derived cells. However, one exception might be the glycoprotein CD34 which is present on human ASCs but absent on MSCs.

It is now widely recognized that multipotent intestinal stem cells that reside between villi within the crypts drive continuous replenishment of the epithelial cells. Intestinal stem cells or crypt cells have the ability to self-renew and regenerate the intestinal tissue by differentiating into various intestinal cells including endocrine cells, enterocytes, goblet cells, and Paneth cells. Wang et al demonstrate that Lrig1 is implicated in maintaining the intestinal epithelial homeostasis and marks stem cells in the intestine [94]. Lgr5 is used as a stem cell marker of the intestinal epithelium and the hair follicle [95-97], however, intestinal organoids can also develop from Lgr5- cells [96]. Another study demonstrates that a small number of slow-cycling epidermal SCs are localized to the basal layer of the epidermis. Bmi1, Tert, Hopx, and Lrig1 are also shown to be robustly expressed in Lgr5+ intestinal stem cells [98]. Intestinal stem cells also express BMI-1 and TERT [98, 99] and intestinal reticular stem cells which generate periepithelial intestinal mesenchymal sheath express gremlin 1 [68].

Lgr5 is a stem cell marker for multiple epithelia, and has recently been shown to be the marker of stem/progenitor cells in ovary and tubal epithelia [100].