Alkaline phosphatase is an enzyme commonly expressed in almost all living organisms. In humans and other mammals, determinations of the expression and activity of alkaline phosphatase have frequently been used for cell determination in developmental studies and/or within clinical trials. Alkaline phosphatase also seems to be one of the key markers in the identification of pluripotent embryonic stem as well as related cells. However, alkaline phosphatases exist in some isoenzymes and isoforms, which have tissue specific expressions and functions. Here, the role of alkaline phosphatase as a stem cell marker is discussed in detail. First, we briefly summarize contemporary knowledge of mammalian alkaline phosphatases in general. Second, we focus on the known facts of its role in and potential significance for the identification of stem cells.

1. Alkaline Phosphatase

Alkaline phosphatase (AP, EC orthophosphoric-monoesterase, alkaline optimum) is a membrane bound enzyme which occurs in almost all living organisms. Mammalian APs have low sequence identity with E. coli, but residues involved in the active site of the enzyme and the ligands coordinating the two zinc atoms and the magnesium ion are conserved; thus, the catalytic mechanism is considered to be similar in prokaryotic and eukaryotic APs [1, 2]. Individual mammalian alkaline phosphatases differ in their heat stabilities and uncompetitive inhibition properties. This is the result of their nonhomologous disulphide bonds, their structural significance, and nonconserved residues [3, 4]. In humans, and probably also in other mammals, four isoenzymes have developed during evolution and are coded by up to four genes. These four isoenzymes can be found in various tissues, where they have different physiological functions. We know three isoenzymes of AP which are specific for certain tissues in humans; these are tissue-specific alkaline phosphatases (TSAP). They are intestinal alkaline phosphatase (IAP; Kasahara isoenzyme), placental alkaline phosphatase (PLAP; Regan isoenzyme), and germ cell alkaline phosphatase (GCAP; Nagao isoenzyme), which are expressed by embryonic cells but also carcinoma cells. The fourth isoenzyme is ubiquitous, and we call it tissue nonspecific alkaline phosphatase (TNAP). TNAP occurs in three isoforms which are bone, liver, and kidney TNAP [5].

In mice, there are also four APs; three of them are tissue specific and the fourth is tissue nonspecific. Intestinal, intestinal-like (IAP-like), and embryonic AP (EAP) are TSAP. The IAP-like AP was identified in mice, which have a knockout gene for IAP and still measurable residual activity in the gut. Murine EAP is similar to human GCAP [6]. Many important facts about the structure, physical and chemical properties, gene localisation, regulation, function, and tissue expression of APs can be obtained from Millán (2006) and McComb et al. (2011) [7, 8]. In the field of stem cell biology, TNAP are the main focus of interest, but EAP or GCAP, which are expressed in pluripotent inner call mass and primordial cells, respectively, cannot be ignored (see below). In general, TNAP and EAP/GCAP are frequently associated with germ layers, progenitors, and naive nondifferentiating cells.

In contrast, TSAPs are expressed with the increasing differentiation and maturation of particular cell lineages and tissues, because of their connection to the functions of such cells, for example, enterocytes (IAP) [911].

2. Expression of AP during Development

Similar to other lineages and developmental-specific genes recognized as determinants of stem cells, the expression pattern of AP during ontogenesis in some parts of tissues also corresponds with particular stem cell precursors and their niches. Thus, here, we briefly summarize the expression and presumable function of particular AP associated with cell differentiation potential (such as pluripotency, e.g.,) and stemness during development in general.

The expression of alkaline phosphatase can already be detected in 2-cell stage preimplantation embryos in mice and it is equally expressed in each embryonic cell to the stage of early blastocyst. At first, it is expressed by both the trophectoderm and inner cell mass (ICM). In the stage of late blastocyst, it seems that AP is strictly expressed in ICM. Lepire and Ziomek [12] described the isoenzyme of AP in the preimplantation stage of embryo-embryonic alkaline phosphatase (EAP). Until this discovery only two isoenzymes of AP had been defined in mice: EAP and TNAP. TNAP is also expressed in the preimplantation embryo, but 10 times less than EAP. Approximately 7 days post coitum (dpc) the expression of Akp5 (EAP) decreases rapidly and Akp2 (TNAP) becomes the major gene of AP, which is expressed between 7–14 dpc in primordial germ (PG) cells [6]. The PG cells exhibit high activity of TNAP during migration to the developing gonad. The activity of TNAP decreases 14-15 dpc in these cells [13]. In humans, it is known that the expression of alkaline phosphatases is detectable prior to 4 weeks of gestation [14, 15].

In contrast to mice PG cells (which express TNAP), GCAP activity is observed in human migrating PG cells [1517]. In the adult, it is primarily synthesized in the testes, cervix, and thymus. Trace amounts are synthesized in placenta and lung tissues [18]. It is not known whether AP is expressed in the preimplantation stage of human embryos. Also, little is known about the expression of other isoenzymes in human embryonic development because of ethical limitations. On the 8th day post coitum (in mice), TNAP is also expressed in the neuroectoderm [19]. Later (9.5 dpc), TNAP activity is observed in the area of the brain and spinal cord. Between 10.5 and 14.5 dpc, TNAP activity is observed in the mesencephalon and the rhombencephalon, along the entire spinal cord and cranial nerves, and at the end of this stage TNAP positive fibres are in the pons. Fourteen-and-a-half days post coitum, TNAP expression is decreased in the neuroepithelium. In the adult, TNAP positive clusters are observed in the subependymal layer, where the neural progenitors are localized. It is suggested that TNAP is involved in the development of the nervous system partly as an ectonucleotidase in neurogenic zones [20, 21] or specifically by interacting with collagen during neuronal migration [22, 23]. However, TNAP activity is detectable in the choroid plexus up to adult age [23].

However, amongst all vertebrates (including humans), TNAP activity is dominant in the developing skeleton, because TNAP is involved in the mineralization of tissues. The activity within the developing skeleton is associated with the expression of TNAP in chondrocytes and osteoblasts. In mice this arises between 13 and 14 dpc [19].

The expression of another AP, TSAP, is also mostly associated with cell differentiation and such AP activity is generally recognized as a marker of cell differentiation [9, 2426].

3. Role and Function of AP

If we ask why some stem cells express AP, we must know how APs are utilized by particular cells. In general, AP catalyzes the hydrolysis of phosphate esters. AP exhibits three types of activity [2731]:(1)hydrolytic activity R–P + H–OHR–OH + H–P,(2)phosphotransferase activity R–P + R′–OHR–OH + R′P,(3)pyrophosphatase activity R–P–P–R′+ H–OHR–P + R′–P.Hydrolytic activity is considered a general reaction which is catalysed by AP.

The role of individual APs is also apparent from the phenotype of organisms with nonfunctional AP. EAP depletion is not associated with any detectable abnormalities. Also, the depletion of TNAP has no effect on the differentiation or migration of PGC, so we may suggest a similar state also for GCAP in human PGC; however, experimental data are unobtainable due to ethical limitations. PLAP, which is only known in primates, enables the transport of maternal IgG through the placenta and, by an unknown mechanism, improves the growth and development of embryos and cells in general. High levels of GCAP and PLAP are also markers of tumor diseases, typically, such neoplasia as germ cell tumours [32, 33], squamous cell carcinoma of the lung [34], and carcinoma of the gastrointestinal tract and uterus [35, 36].

The key roles of TNAP are in the mineralization of hard tissue (it provides free phosphate for the creation of hydroxyapatite crystals and hydrolyses pyrophosphate, an inhibitor of bone matrix formation) and in the metabolism of vitamin B6, and thus in the metabolism of neurotransmitter -aminobutyric acid (GABA). Therefore, the depletion or recessive mutation of TNAP leads to defects in both the mineralization of hard tissue and the development of the nervous system.

IAP plays a role in the transport of fatty acids and triglycerides from the intestinal tract to the circulation [3740]. IAP also regulates duodenal surface pH [41] and detoxifies bacterial endotoxins in the colon by means of dephosphorylation [42, 43].

4. Regulation of the Expression and Activity of AP

AP activity clearly correlates with its expression, and its fine regulation is mediated by the actual microenvironment rather than by some individual signaling pathways. Thus, AP expression and activity are regulated mainly through the developmental status of cells or tissues. Therefore, as mentioned above, AP expression is a generally suitable marker of differentiation processes both in vivo and in vitro for particular cell types. Although some data on the role of p38 kinase (mitogen-activate protein kinase (MAPK) p38) in the regulation of TNAP expression exist, the precise mechanism remains unknown [4447]. Similarly, we observed both a decreased level of AP and decreased AP activity in p38 −/− ES cells [48, 49] in comparison with their wt counterparts, while the expressions of pluripotent markers such as Oct-4, Nanog, and Zfp42 remained unchanged (our unpublished data).

5. Alkaline Phosphatase and Stem Cells

5.1. Pluripotent Stem Cells

A high level of AP and high AP activity are traditional markers of pluripotent embryonic stem (ES) cells, both mouse and human. This is based on the fact that ICM is highly positive for AP activity, in contrast to trophoblast cells at the blastocyst stage. As ICM is committed to lineage differentiation, AP expression is downregulated and it appears in discrete specialized cell populations such as PG cells and later also in other tissues, for example, in osteoblasts (see above). High AP activity is associated with the majority of pluripotent stem cells. Embryonal cancer (EC also called teratocarcinoma stem cells), embryonic germ (EG), the already mentioned embryonic stem (ES), and induced pluripotent stem (iPS) cells express high activity of AP. Interestingly, the absence of AP activity has been reported in pluripotent epiblast stem (EpiS) cells, which are derived from epiblasts of later developmental stages of the embryo than those from which ES cells are derived. The pluripotency of EpiS cells is partially restricted in comparison with other pluripotent stem cells, which correspond to their more differentiated phenotype, compared to ES cells [50, 51].

Mouse ES cells are derived from pluripotent cells of ICM of early blastocyst 3.5–4 dpc. At this time, the Akp5 gene (coding EAP) is dominantly expressed in the embryo. However, mouse ES cells express the Akp2 gene (coding TNAP), which is used for determination of their undifferentiated state [52]. Mouse PG, EG, and EC cells derived from teratocarcinoma also show high activity of TNAP [13, 53, 54]. The dominant expression of TNAP but not EAP in mouse ES cells supports the hypothesis concerning a shared predecessor of ES and PG/EG cells [55]. On the other hand, the shift in expression of AP isoenzymes may also correspond to the fact that the expression of AP is important for cells, but that, as cell environments are modified during developmental processes, particular AP may work more effectively in new conditions. This, we may suggest, is based on the sensitivity of particular AP to various amino acids and small peptides, which were determined as inhibitors of AP activity; these are presented in Table 1 [5, 5662]. However, further experiments will be required to verify such a hypothesis.

In human embryos, GCAP activity is found in PG cells, in contrast to human ES cells, where TNAP is detected [63, 64]. However, it is uncertain whether GCAP is also expressed, which might be expected. Human EC cells express both isoenzymes, TNAP and GCAP [65]. In human EG cells, the GCAP isoenzyme is expected, because GCAP is expressed in migrating human primordial germ cells and gonocytes [18, 63]. However, precise studies of this phenomenon in humans are lacking.

In addition, some experimental results point to greater similarity between mouse ES/PG/EG/EC and human PG/EG/EC cells, but not human ES cells, which raises another question about the significance and role of AP in these cells [66].

However, this inconsistency in the phenotype of these cells was partially resolved by Brons et al. [51] and Tesar et al. [50]. Both these groups described so-called epiblast stem (EpiS) cells, which correspond to pluripotent cells of epiblast. Mouse EpiS cells are more differentiated than mouse ES cells. Mouse and human ES cells should be equivalent in their properties. However, human ES cells correspond more closely to mouse EpiS cells. Interestingly, EpiS cells do not exhibit detectable AP activity [50]. Thus, we can suggest that the precise determination of AP isoform expression patterns in human ES in comparison to the abovementioned cell populations may further improve our picture of its identity, regulation of self-renewal, and phenotype.

The specific importance of AP for pluripotent cells as well for pluripotent stem cells remains questionable. Andäng et al. suggest the importance of GABA synthesis in ES cells for the regulation of their proliferation and self-renewal [67] and that AP participates in GABA synthesis (see above). We may also suggest the increased need for substrate dephosphorylation associated with the quickly proliferating metabolism of ES cells. In ICM and other types of pluripotent cells, we can also hypothesise a similar role for AP as in ES cells.

Interestingly, as mentioned above and expected in EpiS cells, all known mouse and human pluripotent stem cells express TNAP preferentially. The expression of TNAP is also quickly upregulated during the process of reprogramming somatic cells into iPS cells. In the original protocol, the iPS phenotype is induced through the exogenous expression of four genes: Oct-4 (Pouf5), Sox-2, and Klf4, which are responsible for the maintenance of pluripotency, and c-myc, responsible for the induction and/or increasing rate of proliferation [68]. TNAP expression increases directly after the transfection of somatic cells by these four genes. This corresponds with the conception of the direct regulation of TNAP expression by Oct-4 and Sox-2 [69]; see Table 2. Thus, TNAP expression appears long before real cell reprograming and the reexpression of endogenous genes for Oct-4 or Sox-2 [70]. Further, our in silico analysis identified several binding sites for both Oct-4 and Sox-2 and further factors associated with pluripotency such as Nanog, Tcf3, Sa4b, and FoxD3 in promoters of TNAP (Table 2). For all the above-mentioned facts, AP activity is a widely accepted marker of pluripotent stem cells and it has been shown that the maintenance of AP activity in AP-positive colony formation positively correlates with the clonogenic and self-renewal potential of undifferentiated human ES cells in cultures [64]. In addition, AP activity is downregulated reciprocally with differentiation processes involving pluripotent stem cells [71]. On the other hand, depletion of the TNAP gene Akp2 has no effect on the formation of either ICM or PG cells or their expansions in mouse [52]. Surprisingly, at the same time, both ICM and PG cells may be considered at least precursors of pluripotent stem cells, as are ES or EG cells. Unfortunately, little is known about the inability to isolate ES cells from TNAP−/− embryos. More importantly, a consequence of TNAP depletion is a negative effect on bone development and vitamin B turnover (see above).

Interestingly, although EAP is the dominant form of AP in pluripotent ICM, we recognized only one Nanog binding site within the mouse EAP gene, and no binding sites were observed for the other pluripotent genes mentioned above. In human GCAP, one binding site for the mentioned pluripotent genes, apart from FoxD3, was recognized (Table 2). The mechanism concerning the shift from EAP/GCAP expression in pluripotent ICM to TNAP expression in the other pluripotent cells and PG cells is still not understood.

However, it seems clear that TNAP expression is not closely associated with pluripotent stem cells or with pluripotency alone, because ICM cells express another AP dominantly and they have no pure self-renewal potential (and such ICM cells cannot be recognized as stem cells) [19]. Furthermore, PG cells are other cells which strongly express TNAP and other so-called pluripotent markers (Oct-4, Nanog), but they are not pluripotent.

Paradoxically, the regulation of TNAP expression in pluripotent stem cells is also influenced by the effect of retinoic acid (RA). RA has a prodifferentiation and pleiotropic effect on many cells both in vivo and in vitro [72, 73]. The RA effect is fast and strong. Cells undergo the differentiation process after the addition of RA to a culture independently of other conditions. The final phenotype depends on the RA concentration, type of culture (adherent, suspension), and presence of particular signaling molecules [72, 7477]. It is interesting that TNAP activity increases with this differentiation, although both stemness and pluripotent markers decrease. This has been proven in both EC and ES cells ([78, 79] and Figure 2). The upregulation of TNAP expression is clearly mediated by RA receptors binding motifs in the TNAP gene (Table 2). The role and significance of this RA-induced TNAP expression in EC and ES cell differentiation are unknown. However, this observation goes against the hypothesis concerning the role of TNAP in fast proliferating cells. RA induces a decrease in the proliferation and accumulation of cells in the G1 phase of the cell cycle in both ES and EC cells [8082].

Further, the in vitro study of the differentiation of ES cells to PG cells yields a remarkable observation. RA is required for the induction of PG cell formation during the EBs-mediated differentiation of ES cells [83]. In this system, 5-day-old EBs, in which TNAP, Oct-4, and Nanog expression have been downregulated, are treated by RA. After this treatment, PG cells can be determined within EBs. However, PG cells also express high levels of Oct-4, Nanog, and TNAP [84]. Increased TNAP may simply be explained by the presence of RA receptor (RAR/RXR) binding sites in the TNAP gene, as mentioned above. However, the expression of Oct-4 is directly inhibited by RA [85]. The expression of Nanog, the second most important pluripotent gene, can be positively regulated by RA, as has been shown in some recent studies [77, 86]. In this case, Nanog expression is probably induced through the induction of insulin-like growth factor (IGF) expression by RA or other retinoids [86]. IGF is a strong inductor of the PI3-K pathway, whose activity leads to the increasing expression of Nanog in ES cells [87, 88]. Thus, RA mediates the auto- and paracrine stimulation of PI3-K pathways by IGF. Importantly, Nanog itself regulates the expression of Oct-4 [89]. Thus, we suppose that, in the mentioned model of the RA-induced formation of PG cells, Oct-4 expression is upregulated and maintained by Nanog. The feedback in the regulation network Oct-4/Nanog and the positive role of IGF in the maintenance of ES cells are well known [87, 89, 90]. However, the factors that enable the above supposed RA-induced creation of Oct-4 positive cells (PG cells) in contrast to the fast differentiation of RA-treated pluripotent stem cells are still unknown. We can hypothesize a small degree of balance between the expression of RA-targeted genes and modulation of the Oct-4/Nanog network. The reciprocal ordering of Oct-4 and Nanog expression during the establishment of their network may also play an important role here. Recently, the preferential role of Nanog in PG cell formation rather than in pluripotency itself has also been demonstrated [91].

5.2. Other Stem Cells

The high expression/activity of AP in ES cells leads to the presumption that it is a universal marker of SC. Recent research has not confirmed the accuracy of such an assumption. Except for spermatogonia stem cells, a high level of AP has not been definitively demonstrated as a general SC marker in other recognized tissue-specific stem (embryonal or adult) cells or stem cell rich populations. Nevertheless, Langer et al. [20] determined some TNAP positive cells in embryonal and adult central nervous systems. These neural TNAP positive cells were observed within cell populations of the subventricular zone, which is rich in neural stem cells and various neural progenitors. Some of these progenitors were TNAP positive. Here, TNAP was not associated with particular progenitors, but with subpopulations of several identified progenitors. It seems that TNAP expression in the subventricular zone (SVZ) is not associated with a specific cell phenotype, but with cell behaviour, for example, proliferation or migration [20]. This hypothesis is supported by the mitogenic effect of adenosine generated by TNAP through the hydrolyzation of nucleoside triphosphates and diphosphates [20, 92]. Further examples of AP activity in somatic stem cells can be found in reports of various sets of mesenchymal stem cells (MSC) [9395]. Sobiesiak et al. [96] identified TNAP as a stemness marker of MSC, signed as MACS1. Recently, however, Kim et al. [97] presented proosteogenic properties of MSC with higher TNAP activity. MSC derived from adipose tissue exhibit significantly lower levels of TNAP transcript and activity (Figure 1). This is in contrast to the situation in pluripotent stem cells, in which it may be suggested that improved stemness correlates with high TNAP activity, excluding RA effects (see above). The notion that TNAP in mesenchymal cells is a mark of progenitors rather than MSC themselves is also supported by further studies documenting high TNAP in differentiated osteoblasts and odontoblasts [97, 98] (Figure 1). In addition, side population (SP) cells derived from human dental pulp tissue, which are rich in stem cells, exhibit lower levels of TNAP mRNA than the major population [98]. However, further detailed study in this field and on tissue-derived SP cells would be useful. Data comparing the expressions and activities of TNAP in various stem cell-enriched populations are presented in Figure 1.

6. Conclusion

TNAP expression is a suitable marker of pluripotent stem cells, but with some limitations. A high level of AP correlates very well with pluripotency and undifferentiated pluripotent stem cell phenotypes. A low level of AP activity denotes the restriction of pluripotency, which was also observed in EpiS cells [50, 51] and with differentiation. Some other limitations are mentioned above. In all other cases, that is, in adult SC, and so forth, a high level of AP is associated with the process of differentiation rather than with stemness. Thus, particular cell types are able to regulate the expression of TNAP, and probably also other APs, through various combinations of transcriptional regulatory networks. Accordingly, AP such as TNAP, for example, may be expressed under the control of Oct-4 in pluripotent cells and also in Oct-4 negative cells, such as mesenchymal cells and their progeny. Unfortunately, the importance of AP activity for pluripotent cells and/or stem cells (principally for pluripotent stem cells) is still generally unclear. Similarly, we do not understand the role of the shift in expression of EAP/GCAP to TNAP between ICM and ES cells. Does it play a role in the pluripotency of stem cells or does it only represent a marker of the common ancestor/precursor of ES and PG cells? Curiously enough, no PG or ICM cells are considered genuine pluripotent stem cells. PG cells are not pluripotent and both ICM and PG cells have limited self-renewal potential [55]. To answer the question of whether AP/TNAP level and activity are common markers of pluripotent stem cells, further work is required.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.


This work was supported by grants from the Ministry of Education, Youth and Sport of the Czech Republic (MSM0021622430, HistoPARK CZ.1.07/2.3.00/20.0185), and the European Regional Development Fund (KI-MU; CZ.1.07/2.3.00/20.0180).