CD117⁺ amniotic fluid stem cells

Abstract

Broadly multipotent stem cells can be isolated from amniotic fluid by selection for the expression of the membrane stem cell factor receptor c-Kit, a common marker for multipotential stem cells. They have clonogenic capability and can be directed into a wide range of cell types representing the three primary embryonic lineages. Amniotic fluid stem cells maintained for over 250 population doublings retained long telomeres and a normal karyotype. Clonal human lines verified by retroviral marking were induced to differentiate into cell types representing each embryonic germ layer, including cells of adipogenic, osteogenic, myogenic, endothelial, neuronal and hepatic lineages. AFS cells could be differentiate toward cardiomyogenic lineages, when co-cultured with neonatal cardiomyocytes, and have the potential to generate myogenic and hematopoietic lineages both in vitro and in vivo. Very recently first trimester AFS cells could be reprogrammed without any genetic manipulation opening new possibilities in the field of fetal/neonatal therapy and disease modeling. In this review we are aiming to summarize the knowledge on amniotic fluid stem cells and highlight the most promising results.

Keywords: :

• amniotic fluid stem cells
• regenerative medicine
• amniotic fluid
• fetal cells
• tissue engineering
• in utero cell therapy

Amniotic Fluid: Function, Origin and Composition

Amniotic fluid allows the fetus to freely grow and move inside the uterus, and acts as a vehicle for the exchange of body chemicals with the mother.^1,2 In humans, the amniotic fluid starts to appear at the beginning of week 2 of gestation as a small film of liquid between the cells of the epiblast. Between days 8 and 10 after fertilization, this fluid gradually expands and separates the epiblast (i.e., the future embryo) from the amnioblasts (i.e., the future amnion), thus forming the amniotic cavity.³ Thereafter, it progressively increases in volume, completely surrounding the embryo after week 4 of pregnancy. Over the course of gestation, amniotic fluid volume changes markedly from 20 ml in week 7 to 600 ml in week 25, 1,000 ml in week 34 and 800 ml at birth. During the first half of gestation, the amniotic fluid results from active sodium and chloride transport across the amniotic membrane and the non-keratinized fetal skin, with concomitant passive movement of water. In the second half of gestation, the amniotic fluid is constituted by fetal urine, gastrointestinal excretions, respiratory secretions and substances exchanged through the sac membranes.^4-8 The amniotic fluid is primarily composed of water and electrolytes (98–99%) but also contains chemical substances (e.g., glucose, lipids, proteins, hormones and enzymes), suspended materials (e.g., vernix caseosa, lanugo hair and meconium) and cells. Amniotic fluid cells derive both from extra-embryonic structures (i.e., placenta and fetal membranes) and from embryonic and fetal tissues.⁹ Although amniotic fluid cells are known to express markers of all three germ layers,^10,11 their exact origin still represents a matter of discussion; the consensus is that they mainly consist of cells shed in the amniotic cavity from the developing skin, respiratory apparatus, urinary and gastrointestinal tracts.^6,12-14 Amniotic fluid cells display a broad range of morphologies and behaviors, varying with gestational age and fetal development. In normal conditions the number of amniotic fluid cells increases with advancing gestation; if a fetal disease is present, amniotic fluid cell counts can be either dramatically reduced (e.g., intrauterine death and urogenital atresia) or abnormally elevated (e.g., anencephaly, spina bifida and exomphalos).¹⁴ Based on their morphological and growth characteristics, viable adherent cells from the amniotic fluid are classified into three main groups: epithelioid (33.7%), amniotic fluid (60.8%) and fibroblastic type (5.5%).¹⁵ In the event of fetal abnormalities other types of cells can be found in the amniotic fluid, e.g., neural cells in presence of neural tube defects and peritoneal cells in case of abdominal wall malformations. The majority of cells present in the amniotic fluid are terminally differentiated and have limited proliferative capabilities.^16,17 In the 1990s, however, two groups demonstrated the presence of small subsets of cells in the amniotic fluid harboring a proliferation and differentiation potential. First, Torricelli et al.¹⁸ reported the presence of hematopoietic progenitors in the amniotic fluid collected before week 12 of gestation. Then Streubel et al.¹⁹ were able to differentiate amniotic fluid cells into myocytes, thus suggesting the presence in the amniotic fluid of non-hematopoietic precursors. These results initiated a new interest in the amniotic fluid as an alternative source of cells for therapeutic applications.

CD117⁺ Amniotic Fluid Stem Cells

History

The first evidence that the amniotic fluid could contain pluripotent stem cells was provided when Prusa et al.²⁰ in 2003 described the presence of a distinct sub-population of proliferating amniotic fluid cells (0.1–0.5%) expressing the pluripotency marker Oct4 at both transcriptional and protein levels. Oct4 (i.e., octamer binding transcription factor 4) is a nuclear transcription factor that plays a critical role in maintaining ESC differentiation potential and capacity for self-renewal.^21,22 Other than its expression by ESC, Oct4 is specifically expressed by germ cells, where its inactivation results in apoptosis, and by embryonal carcinoma cells and tumors of germ cell origin, where it acts as an oncogenic fate determinant.^23-26 While its role in stem cells of fetal origin has not been completely addressed, it has been recently demonstrated that Oct4 is neither expressed nor required by somatic stem cells or progenitors.^27-29 Despite this, different groups confirmed the expression of Oct4 and of its transcriptional targets (e.g., Rex-1) in the amniotic fluid.^30,31 Remarkably, Karlmark et al.³² transfected human amniotic fluid cells with the green fluorescent protein gene under either the Oct4 or the Rex-1 promoter and established that some amniotic fluid cells were able to activate these promoters. Several authors subsequently reported the possibility of harvesting amniotic fluid cells displaying features of pluripotent stem cells.^33,34 Thereafter, the presence of a cell population able to generate clonal cell lines capable of differentiating into lineages representative of all three embryonic germ layers was definitively demonstrated.³⁵ These cells, named amniotic fluid stem cells (AFS cells), are characterized by the expression of the surface antigen c-kit (CD117), the type III tyrosine kinase receptor of the stem cell factor.³⁶

Isolation and culture

The proportion of c-kit positive/hematopoietic lineage negative cells in the amniotic fluid varies over the course of gestation roughly describing a Gaussian curve; they appear at very early time points in gestation (i.e., at 7 weeks of amenorrhea in humans and at E9.5 in mice) and present a peak at mid gestation corresponding to 90 × 10⁴ cells/fetus at 20 weeks of pregnancy in humans and to 10,000 cells/fetus at E12.5 in mice (Fig. 1).³⁷ Human AFS cells have been derived either from small volumes (5 ml) of 1st (10–12 gestational weeks) and 2nd trimester AF (14–22 gestational weeks) or from confluent back-up amniocentesis cultures.^35,38 Murine AFS cells have been obtained from the AF collected during the second week of gestation (E11.5−E15).^35,37,39 Rat AFS cells have been retrieved at a mean gestational age of 16 d.p.c.⁴⁰ AFS cell isolation is based on a two-step protocol consisting in the prior immunological selection (via flow cytometry or magnetic cell sorting) of CD117 positive cells from the AF (approximately 1% of total AF cells) and in the subsequent expansion of these cells in culture.^35,37,41-43 Isolated AFS cells can be expanded in feeder layer-free, serum-rich conditions without evidence of spontaneous differentiation in vitro. Cells are cultured in basic medium containing 15% of fetal bovine serum and Chang supplement.^35,37,43 In a single study, murine AFS cells have been also successfully cultured onto feeder layers (i.e., mitomycin C-treated mouse embryonic fibroblasts) in basic medium supplemented with 15% fetal bovine serum, mercaptoethanol, stem cell factor (SCF), bone morphogenetic protein 4 (BMP4) and leukemia inhibitory factor (LIF).⁴⁴ Whether, these different culture protocols influence AFS cell features and potential remains to be established.

Figure 1. CD117 positive/hematopoietic lineage negative cells in the amniotic fluid of mice and human along the course of gestation. (A) Percentage of mouse AFS cells (mAFSC) as a function of gestational age. (B) Total number of mouse AFS cells (mAFSC) per embryo equivalent (EE) at different gestational ages. (C) Percentage of human AFS cells (hAFSC) as a function of gestational age. Means are represented by bars. Adapted from reference 37.

Characterization

Karyotype analysis of human AFS cells deriving from pregnancies in which the fetus was male revealed the fetal origin of these cells.^35,38 AFS cells proliferate well during ex vivo expansion. When cultivated, they display a spectrum of morphologies ranging from a fibroblast-like to an oval-round shape (Fig. 2A). AFS cells possess a great clonogenic potential.^35,39,42,45 Clonal AFS cells lines expand rapidly in culture (doubling time = 36 h) and, more interestingly, maintain a constant telomere length (20 kbp) between early and late passages (Fig. 2B).^35,43 Despite their high proliferation rate, clonal AFS cells show a homogeneous, diploid DNA content without evidence of chromosomal rearrangement even after expansion to 250 population doublings (Fig. 2C). Clonal AFS cells lines express markers of a pluripotent undifferentiated state: Oct4 and NANOG.^35,39 However, it has been shown that they do not form tumors when injected undifferentiated into severe combined immunodeficient (SCID) mice.^35,38,46-49 The cell-surface antigenic profile of AFS cells has been determined through flow cytometry by different investigators (Table 1). Cultured human AFS cells are positive for ESC (e.g., SSEA-4) and mesenchymal markers (e.g., CD73, CD90 and CD105), for several adhesion molecules (e.g., CD29 and CD44) and for antigens belonging to the major histocompatibility complex I (MHC-I). They are substantially negative for hematopoietic and endothelial markers (e.g., CD14, CD34, CD45, CD133 and CD31), and do not express antigens belonging to the major histocompatibility complex II (MHC-II), thus presenting a low immunogenicity profile.^{35,37,47,50,51} As stability of cell lines is a fundamental prerequisite for basic and translational research, AFS cell capacity of maintaining their baseline characteristics over passages has been evaluated based on multiple parameters. Despite their high proliferation rate, AFS cells and derived clonal lines maintain constant morphology, doubling time, apoptosis rate, cell cycle distribution and marker expression (e.g., Oct4, CD117, CD29 and CD44) up to 25 passages.^43,52 During in vitro expansion, however, cell volume tends to increase and significant fluctuations of proteins involved in different networks (i.e., signaling, antioxidant, proteasomal, cytoskeleton, connective tissue and chaperone proteins) can be observed using a gel-based proteomic approach.⁵² The significance of these modifications warrants further investigations and needs to be taken in consideration.

Figure 2. (A) Human AFS cells mainly display a spindle-shaped morphology during in vitro cultivation under feeder layer-free, serum-rich conditions. (B and C) Clonal human AFS cell lines retain long telomeres and a normal karyotype after more than 250 cell divisions. (B) Conserved telomere length of AFS cells between early passage (20 population doublings, lane 3) and late passage (250 population doublings, lane 4). Short length (lane 1) and high length (lane 2) telomere standards provided in the assay kit. (C) Giemsa band karyogram showing chromosomes of late passage (250 population doublings) cells. Adapted from reference 35.

Table 1. Surface markers expressed by human AFS cells: results by different groups

Markers	Antigen	CD no.	De Coppi, 2007*	Chiavegato, 2007*	Ditadi, 2009*	Guan, 2011*	Rota, 2012*	Moschidou, 2012°
ES cells	SSEA-3	none	-	nt	nt	nt	nt	+
	SSEA-4	none	+	nt	nt	-	+	+
	Tra-1–60	none	-	nt	nt	nt	nt	+
	Tra-1–81	none	-	nt	nt	nt	nt	+
Mesenchymal	SH2, SH3, SH4	CD73	+	+	nt	+	+	+
	Thy1	CD90	+	+	+	+	+	+
	Endoglin	CD105	+	+	nt	+	+	+
Endotelial and hematopoieic	LCA	CD14	nt	nt	nt	nt	nt	-
	gp105–120	CD34	-	-	-	-	nt	-
	LPS-R	CD45	-	-	+	-	nt	-
	Prominin-1	CD133	-	-	-	nt	nt	nt
	MCAM	CD146	nt	+	nt	+	-	nt
Integrins	β1-integrin	CD29	+	+	nt	+	+	+
Ig-superfamily	HCAM-1	CD44	+	+	+	+	+	+
MHC	I (HLA-ABC)	None	+	+	+	+	+	-
	II (HLA-DR,DP,DQ)	None	-	-	-	nt	-	-

Nt, not tested; *AFS cells obtained during the second trimester of gestation; °AFS cells obtained during the first trimester of gestation.

Potency

AFS cell broad multipotency has been established by demonstrating that clonal AFS cell lines: (1) can differentiate into cell lineages originating from all three primary germ layers under specific inducing conditions and (2) are able to form embryoid bodies (EB) when cultured in suspension. Starting from monoclonal lines, AFS cells (human, CD117-selected) could be differentiated in vitro along six distinct lineages representative of the three germ layers (adipogenic, osteogenic, myogenic, endothelial, neurogenic and hepatic) under specific culture conditions (Fig. 3). Cell clones were obtained by limited dilution and the single cell origin of clonal lines was confirmed via retroviral vector integration.³⁵ When cultured under conditions in which they are unable to attach to the surface of culture dishes and without anti-differentiation factors, embryonic stem (ES) cells harbor the potential to form three-dimensional aggregates, named EBs, which recapitulate the first steps of early mammalian embyogenesis.^53,54 In 2010, for the first time, it was reported that starting from clonal human CD117-selected and AFS cell lines, embryoid bodies can be formed. The frequency of EB formation (i.e., % of number of EB recovered from 15 hanging drops) was estimated to be around 28% for AFS cell lines and around 67% for AFS cell clonal lines. Similarly to ES cells, EB generation by AFS cells is regulated by the mTor (i.e., mammalian target of rapamycin) pathway and is accompanied by a decreased expression of stem cell markers (e.g., nodal and Oct4) and by an increased expression of differentiation markers (e.g., nestin-ectodermal, GATA4-endodermal and Brachyury-mesodermal).^43,55,56 In 2012, it has been demonstrated that first trimester human CD117⁺ AFS cells represent a more undifferentiated and potent sub-population of AFS cells as they share 82% transcriptome identity with ES cells, express markers of pluripotency (OCT4, NANOG, SOX2 and c-MYC, KFL4) at a molecular and protein level and are able to form beating EBs with high efficiency. Within these EBs, AFS cells express markers representative of the three embryonic germ layers, with concomitant downregulation of OCT4.³⁸ Despite the aforementioned results, however, both 1st and 2nd trimester AFS cells cannot be rigorously considered pluripotent stem cells, as they do not form tumors when injected into nude mice and, so far, have not been shown to generate germline-competent chimeras.^{35,38,39,46-49,51,57} Very recently, however, Moschidou et al.³⁸ have shown that human 1st trimester AFS cells can be reprogrammed to functional pluripotent stem cells by culture on Matrigel in ES cell medium supplemented with the histone deacetylase inhibitor (HDACi) valproic acid (VPA) but without ectopic reprogramming factors [i.e., amniotic fluid induced pluripotent stem (AFiPS) cells]. AFiPS cells show signs of true pluripotency as not only they differentiate into functional lineages of the three germ layers (bone, fat, cartilage, definitive endoderm, hepatocyte, ectoderm, neurons and oligo-dendrocytes) and readily form EBs in vitro, but also form teratomas after injection into immunocompromised mice. AFiPS cells express more than 200 genes in common with ES cells but uniquely express around 312 genes including those involved in spermatogenesis that are only active in germ cells. Complementary analysis revealed, in fact, that AFS cells share several common characteristics with primordial germ cells (PGC) including expression of CD117, T, FGF8, SOX17, STELLA, DAZL, NANOS, VASA, SSEA1, FRAGILIS and PUM2, thus suggesting that AFS cells might originate from PGC or PGC progenitors, which have been retained in the AF at the start of their migration to the genital ridge.³⁸ In vivo studies are needed to test the functional differentiation and tissue repair capacity of AFiPS cells in animal models of disease. However, these preliminary data suggest that AFiPS cells could provide an alternative source of pluripotent cells for allogeneic clinical use in cell-based therapies as they can be collected early in pregnancy during termination procedures and are not at potential risk of virally induced tumorigenicity as they are reprogrammed with VPA.

Figure 3. AFS cells differentiation. (A) Hepatogenic: urea secretion by human AFS cells before (rectangles) and after (diamonds) hepatogenic differentiation in vitro. (B) Neurogenic: secretion of neurotransmitter glutamic acid in response to potassium. (C) Osteogenic: mouse micro-CT scan 18 weeks after implantation of printed constructs of engineered bone from human AFS cells; arrow head: region of implantation of control scaffold without AFS cells; rhombus: scaffolds seeded with AFS cells. Adapted from reference 35.

Preclinical studies

In the last ten years AFS cell research has become a fast-growing and promising area of interest. Many studies contributed to the today’s picture of the AF containing multipotent stem cells for different therapeutic applications. Hereafter, the most relevant papers investigating AFS cell therapeutic potential in the translational research setting are summarized; only studies explicitly employing CD117⁺ AFS cells have been considered.

Heart: in vitro differentiation toward the cardiac lineage and therapeutic potential in animal models of myocardial infarction

Cardiovascular diseases are the first cause of mortality in developed countries despite advances in pharmacological, interventional and surgical therapies. Cell transplantation is an attractive strategy to replace endogenous cardiomyocytes lost by myocardial infarction. Fetal and neonatal cardiomyocytes are the ideal cells for cardiac regeneration as they have been shown to integrate structurally and functionally into the myocardium after transplantation.⁵⁸ However, their application is limited by the ethical restrictions involved in the use of fetal and neonatal cardiac tissues.⁵⁹ Chiavegato et al.⁴⁷ were the first to investigate human AFS cell (human, 2nd trimester and CD117-selected) plasticity toward the cardiac lineage. Undifferentiated AFS cells express cardiac transcription factors at a molecular level (i.e., Nkx2.5 and GATA-4 mRNA) but do not produce any myocardial differentiation marker. The authors proved that, under in vitro cardiovascular inducing conditions, human AFS cells express cardiomyocyte (i.e., Nkx2.5, MLC-2v, GATA-4 and β-MyHC), endothelial (i.e., angiopoietin and CD146) and smooth muscle (i.e., smoothelin) markers. However, when xenotransplantated in a rat model of myocardial infarction, with or without cyclosporine treatment, or in intact heart from immuno-competent or immuno-deficient animals, AFS cells were acutely rejected and, in some animals with myocardial infarction, formed chondro-osteogenic masses.⁵⁹ Recently, Delo et al.⁶⁰ demonstrated that the chondro-osteogenic lesions reported by the latter study are independent from cell injection and related to infarction severity/size in the ischemic rat heart. AFS cell capacity of differentiating toward the cardiac lineage has been subsequently confirmed by several studies. Guan et al.⁵⁰ showed that undifferentiated AFS cells (human, 2nd trimester and CD117-selected): (1) express several cardiac genes including the transcription factor mef2, the gap junction connexin43 and H- and N-cadherin, (2) acquire a cardiomyocyte-like phenotype (i.e., morphological changes, upregulation of cardiac-specific genes troponin I and troponin T, redistribution of connexin43) after a 24-h incubation with 5-aza-2′-deoxycytidine (5-AZA-dC) and (3) form mechanical and electrical connections with neonatal rat cardiomyocytes when cultured together. Bollini et al.⁴⁰ proved that the co-colture with neonatal rat cardiomyocytes activates in AFS cells (rat, CD117-selected) a “myocardial gene program:” expression of myocardial antigenic markers (cTnI, MyHC and α-actinin) and development of electrical excitability and activity. In vivo, AFS cells (rat, CD117-selected) directly transplanted in an undifferentiated state in the heart of rats with ischemia/reperfusion injury (i.e., ligation of the left anterior descending coronary artery) engraft in the myocardium, differentiate into endothelial and smooth muscle cells and improve heart ejection fraction as measured 3 weeks after injury by MRI.⁴⁰ As proven in a similar model of acute myocardial infarction, AFS cell (human, 2nd trimester and CD117-selected) beneficial effects on infarct size and myocardial cell survival are exerted rapidly (few hours) by the secretion of paracrine effectors such as the cardioprotective and proangiogenic factor thymosin β4 (Tβ4).⁶¹ AFS cells display a cardiogenic potential in vitro and in vivo; their potential use in the autologous setting (e.g., tissue engineering approaches to treat congenital cardiac malformations) as in the allogeneic context (e.g., ischemic injury) needs further evaluation.

Skeletal muscle: myogenic differentiation in vitro and muscle regeneration in animal models of disease

Stem cell therapy is an attractive method to treat muscular degenerative diseases because only a small number of cells, together with a stimulatory signal for expansion, are required to obtain a therapeutic effect.⁶² The identification of a stem cell population providing efficient muscle regeneration is critical for the progression of cell therapy for muscle diseases.⁶³ Despite early data reporting AFS cell capacity of expressing molecular markers of muscle differentiation under specific culture conditions,³⁵ high impact papers have just been very recently published. First, Gekas et al.⁴⁸ demonstrated that human, 2nd trimester, CD117⁺-selected AFS cells can acquire a myogenic-like phenotype (i.e., expression of desmin and MyoD) after culture in myogenic-specific induction conditions (i.e., culture on plastic plates coated with Matrigel and 3 μM 5-aza-2′-deoxycytidine). However, when these cells were transplanted undifferentiated into uninjured muscles of SCID mice, they failed to differentiate toward the myogenic lineage in vivo. In 2012, Ma et al.⁴⁹ demonstrated for the first time that human, 2nd trimester, clonal AFS cell lines expressing (not selected for) CD117⁺ could differentiate into skeletal myogenic cells in vitro and incorporate into regenerating skeletal muscle in vivo. In vitro, after temporary exposure to the DNA demethylation agent 5-aza-2′-deoxycytidine or by co-culture with C2C12 myoblasts, AFS cells differentiated into skeletal myogenic cells expressing specific markers such as Desmin, Troponin I and α-Actinin. In vivo, four weeks after transplantation into cardiotoxin-injured and X-ray-irradiated muscles of NOD/SCID mice, AFS cells survived, differentiated into myogenic precursor cells (i.e., Desmin, Laminin and Myf5 expression) and fused with host myofibers, thus contributing to skeletal muscle regeneration. The authors hypothesized that undifferentiated AFS cells first migrated into the regenerating myofibers driven by cytokines or chemokines, then converted into myogenic precursors residing in the satellite cell niche as demonstrated by Myf-5 expression, and finally were activated to fuse with the myofibers. Very recently, Piccoli et al.⁴⁴ demonstrated for the first time the functional and stable long-term integration of AFS cells into the skeletal muscle of HSA-Cre Smn_F7/F7 mutant mice, which closely replicate the clinical features of human muscular dystrophies. AFS cells were obtained from E11.5–13.5 GFP⁺ mice through immediate CD117 selection after amniotic fluid collection. Approximately 25,000 freshly isolated AFS cells were directly injected into the tail vein of each animal without previous expansion in culture. Transplanted mice displayed enhanced muscle strength, improved survival rate by 75% and restored muscle phenotype in comparison to untreated animals. Not only, dystrophin distribution in GFP⁺ myofibers was similar to that of wild type animals, but also GFP⁺ cells were found engrafted into the muscle stem cell niche as demonstrated by their sub-laminal position and by Pax7 and α7integrin expression. Functional integration of AFS cells in the stem cell niche was confirmed by successful secondary transplants of GFP⁺ satellite cells derived from AFS cell-treated mice into untreated Smn_F7/F7 mutant mice. In order to progress toward their application for therapy, the therapeutic potential of cultivated AFS cells was also investigated and 25,000 AFS cells previously expanded in culture were intravenously injected into Smn_F7/F7 mice. Despite maintaining a therapeutic potential, cultivated AFS cells regenerated approximately 20% of the recipient muscle fibers vs. 50% when employing freshly isolated AFS cells, thus, highlighting the importance of optimizing AFS cell expansion conditions. In summary, available data support that AFS cells have the potential: (1) to differentiate toward the myogenic lineage, (2) to participate into muscle regeneration in animal models of muscle injury and (3) to engraft and to participate into the muscle stem cell niche. Hence, AFS cells do constitute a promising therapeutic option for muscle skeletal muscle degenerative diseases. Understanding the process of AFS cell myogenesis in vivo and improving cell expansion in vitro will be crucial requisites before this could happen.

Bone: differentiation in vitro and bone repair in vivo

Critical-sized segmental bone defects are one of the most challenging problems faced by orthopedic surgeons. Autologous and heterologous bone grafting are limited respectively by the small amount of tissue available for transplantation and by high re-fracture rates.^64,65 Tissue engineering strategies that combine biodegradable scaffolds with stem cells capable of osteogenesis have been indicated as promising alternatives to bone grafting;⁶⁶ however, bone regeneration through cell-based therapies has been limited so far by the insufficient availability of osteogenic cells.⁶⁷ The potential of AFS cells to synthesize in vitro mineralized extracellular matrix within porous scaffolds has been investigated by different groups. After exposure to osteogenic conditions in static 2D cultures, AFS cells differentiate into functional osteoblasts (i.e., activate the expression of osteogenic genes such as Runx2, Osx, Bsp, Opn, Ocn and produce alkaline phosphatase) and form dense layers of mineralized matrix.^35,67-70 When seeded into 3D biodegradable scaffolds and stimulated by osteogenic supplements (i.e., rhBMP-7 or dexamethasone), AFS cells show an even more significant osteogenic differentiation and production of mineralized matrix in comparison to 2D cultures.^35,67-70 In vivo, AFS cells are not able to mineralize at ectopic sites unless previously pre-differentiated in vitro.⁶⁷ However, when seeded into 3D scaffolds and pre-commited in vitro toward the osteogenic lineage, AFS cells are able to repair critical sized bone defects. Rodrigues et al.⁷¹ initially deposited AFS cells (human, 2nd trimester and CD117-selected) into starch-poly(ε-caprolactone) (SPCL) scaffolds and cultured them for three weeks in osteogenic conditions; then implanted the constructs composed by AFS cells plus SPCL scaffolds into critical sized femoral defects of rats. In vivo formation of new bone was observed by micro-CT imaging with complete repair of bone defect after 16 weeks after implantation. Similarly, pre-differentiated AFS cells (human, 2nd trimester and CD117-selected) seeded into silk fibroin scaffolds were able to repair critical-size cranial bone defects in immunocompromised rats via formation mature bone in a four week time.⁷² These results demonstrate the potential of AFS cells to produce 3D mineralized bioengineered constructs and suggest that AFS cells may be an effective cell source for the autologous and heterologous repair of large bone defects. Further studies are needed to confirm the long-term functional performance of these grafts into bone injury sites.

Cartilage: in vitro differentiation into functional chondrocytes

Enhancing the regeneration potential of hyaline cartilage is one of the most significant challenges for treating damaged cartilage.⁷³ The capacity of AFS cells to differentiate into chondrocytes has been tested in vitro. Human CD117-selected AFS cells treated with TGF-beta1 have been proven to produce significant amounts of cartilaginous matrix (i.e., sulfated glycosaminoglycans and type II collagen) both in pellet and alginate hydrogel cultures.⁷⁴ Via cultivation of AFS cells (human and CD117-selected) into a defined osteochondral medium, Rodrigues et al.⁷¹ generated bilayered osteochondral constructs consisting, on one side, of a SPCL scaffold containing AFS cells differentiated toward the osteogenic lineage and, on the other side, of a agarose hydrogel seeded with AFS cells committed toward the chondrogenic lineage. The regenerative potential of pre-differentiated cells and engineered constructs should be tested in vivo.

Kidney: AFS cell differentiation in organotypic cultures and therapeutic potential in animal models of acute and chronic disease

The incidence and prevalence of end stage renal disease (ESRD) continues to increase worldwide. Although renal transplantation represents a good treatment option, the shortage of compatible organs remains a critical issue for patients affected by ESRD. Therefore, the possibility of developing stem cell-based therapies for both glomerular and tubular repair is receiving intensive investigation.⁷⁵ Different stem cell types have shown some potential in the generation of functional nephrons^76-81 but the most appropriate cell type for transplantation is still to be established.⁸² The first evidence that AFS cells could be of relevance for renal stem cell therapy has been published in 2007. Perin et al.⁴² demonstrated that monoclonal CD117⁺ human 2nd trimester AFS cells: (1) do not basally express kidney markers and thus are not specifically pre-committed to kidney progenitor cells and (2) if injected into isolated murine embryonic kidneys can integrate into renal tissue, participate in all steps of nephrogenesis and express molecular markers of kidney differentiation such as ZO-1, GDNF and claudin. In 2010, Siegel et al.⁴⁵ definitely proved that monoclonal human CD117-selected AFS cells could contribute to kidney development. Chimeric organotypic renal structures were generated via mixing single-cell suspensions of murine embryonic kidneys with human AFS cells. Descending from one cell, pluripotent Oct4- and CD117-positive AFS cells were able to contribute to renal tissue formation and to express renal specific markers (e.g., Pax2-ureteric bud and nephron structures marker, E-cadherin-distal tubule marker, calbindin-ureteric bud marker and laminin-marker of epithelial structures in the developing kidney). Moreover, it was proven that AFS cell contribution to nephrogenesis is regulated by components of the mammalian target of rapamycin (mTOR) signaling pathway (i.e., mTORC1 and mTORC2).⁴⁵ The first successful therapeutic application of AFS cells in an animal model of renal disease was published in 2010. Clonal cell lines derived from CD117⁺ human 2nd trimester AFS cells were directly injected into the kidneys of nude mice affected by acute tubular necrosis (ATN) due to glycerol-induced rhabdomyolysis. When administered at the time of damage, AFS cells provided a strong protective effect toward ATN injury, as reflected by the more rapid resolution of tubular structural damage, decreased tubular cell apoptosis, increased tubular cell proliferation and reduced creatinine and BUN levels. Despite these striking results, the efficiency of AFS cell integration into the renal tissue was low and just few integrated cells showed a commitment toward renal differentiation. Multiplex cytokine assays demonstrated a lower expression of pro-inflammatory molecules (e.g., IL16 and IL27) and a higher expression of anti-inflammatory cytokines (e.g., IL10, IL1Ra and IL6) in the kidneys of mice treated with AFS cells vs. placebo, thus suggesting an immune-modulatory effect of AFS cells in the prevention/resolution of renal damage during ATN.⁵⁷ Comparable results have been obtained by Hauser et al.,⁸³ who injected a heterogenous mixture of 2nd trimester human amniotic fluid-derived cells (non CD117-selected and non-clonal) into nude mice with glycerol-induced acute kidney injury (AKI) obtaining normalization of renal function. Similarly to the previous study, the injected cells were found to persistently accumulate within the peritubular capillaries and the interstitium and to induce enhanced tubular cell proliferation and reduced apoptosis in a differentiation-independent mechanism. However, as a heterogenous cell population was employed, the authors could not exclude the employment of renal-progenitor cells (known to be contained in human amniotic fluid) and could not guarantee the utilization of a homogeneous population of stem cells. Another proof of AFS cell beneficial effect in AKI models was recently given by Rota et al.⁵¹ Infusion of human, 2nd trimester, CD117⁺ AFS cells into NOD-SCID mice with cisplatin-induced AKI improved renal function, limited tubular damage and prolonged animal survival. AFS cells predominantly engrafted in the peritubular regions without showing signs of tubular differentiation. They exerted anti-apoptotic effect and stimulated proliferation of tubular cells via local release of proteic factors, including IL6, VEGF and SDF1/CXCL12. The therapeutic potential of AFS cells was enhanced by cell pretreatment with glial cell line-derived neurotrophic factor (GDNF), which markedly ameliorated renal function and tubular injury by increasing AFS cell production of growth factors, motility, and expression of receptors involved in cell homing and survival.⁵¹ After proving the beneficial role of AFS cells in acute kidney damage (i.e., ATN), Perin’s group has also recently demonstrated AFS cell therapeutic potential in a model of chronic kidney disease. CD117-selected murine clonal AFS cell lines were systemically injected into genetically modified mice affected by Alport Syndrome (Col4a5^−/−) during the very early phase of disease (i.e., before onset of proteinuria). A single injection of AFS cells delayed interstitial fibrosis and progression of glomerular sclerosis, prolonged animal survival, and ameliorated the decline of kidney function. AFS cells did not differentiate into podocyte-like cells and did not stimulate production of the collagen IVa5 needed for normal formation and function of the glomerular basement membrane. Instead, the mechanism of renal protection consisted in the paracrine/endocrine modulation of pro-fibrotic cytokine expression and of macrophages recruitment to the interstitial space.³⁹ Taken together, the above reported studies provide evidence that AFS cells harbor the potential to differentiate upon renal lineages and warrant further investigation of their potential use for acute and chronic kidney diseases.⁸⁴ The discrepancy concerning renal differentiation potential of AFS cells in in vitro vs. in vivo experiments as well as AFS cell mechanism of action for kidney repair need further investigation.

Blood: hematopoiesis and hematologic diseases

Hematopoietic stem cells (HSC) lie at the top of the hematopoietic ontogeny and, if engrafted in the right niche, can theoretically reconstitute the entire blood supply. Thus, the generation of autologous HSCs from patient-specificic stem cells offers real promise for cell-therapy of genetic and malignant blood disorders. The hematopoietic potential of AFS cells was explored by Ditadi et al. in 2009.³⁷ Murine AFS cells were obtained from E12.5 to E13.5 GFP⁺ mice. Human AFS cells were derived from 2nd trimester amniotic fluid. CD117⁺/hematopoietic lineage negative cells were immediately selected after collection of the amniotic fluid. In vitro, AFS cells exhibited strong multilineage hematopoietic potential; in semisolid media, these cells were able to generate erythroid, myeloid and lymphoid colonies. Moreover, murine cells exhibit the same clonogenic potential (0.03%) as hematopoietic progenitors present in the liver at the same stage of development. In vivo, mouse AFS cells (2 × 10⁴ cells/animal i.v. injected) were able to generate all three hematopoietic lineages after primary and secondary transplantation into immunocompromised hosts (i.e., sublethally irradiated Rag^−/− mice), demonstrating their ability to self-renew. These results clearly show that AFS CD117⁺ cells present in the amniotic fluid have true hematopoietic potential, both in vitro and in vivo and support the idea that they may be a new source of stem cells for therapeutic applications.

Lung: AFS cell differentiation in organotypic cultures and therapeutic potential in animal models of disease

Chronic lung diseases are common and debilitating; medical therapies have restricted efficacy and lung transplantation is often the only effective treatment.⁸⁵ The use of stem cells for lung repair and regeneration after injury holds promise as a potential therapeutic approach for many lung diseases; however, current studies are still in their infancy.⁸⁶ AFS cells (human, 2nd trimester and CD117⁺-selected) ability to integrate into the lung and to differentiate into pulmonary lineages has been investigated in different experimental models of lung damage and development.⁴⁶ In vitro, AFS cells injected into mouse embryonic lung explants engrafted into the epithelium and the mesenchyme and express the early pulmonary differentiation marker TFF1. In vivo, when injected into nude mice in the absence of lung damage, systemically administered AFS cells homed to the lung but did not differentiate into specialized cells. When injected into nude mice with lung damage (i.e., hyperoxia or naphthalene injury), AFS cells not only exhibit a strong tissue engraftment but also express specific alveolar and bronchiolar epithelial markers (e.g., TFF1, SPC and CC10).⁴⁶ A recent work demonstrated that, in case of hyperoxia, alveolar epithelial cell type 2 (AEC2) produced chemotactic substances that increase AFS cell migration to the site of damage and that stimulate AFS cells to express phenotypic markers of distal alveolar epithelial cells (i.e., SPC, TTF-1 and ABCA3), thus leading to tissue repair.⁸⁷ These results illustrate AFS cell capacity of responding to lung damage by expressing specific alveolar or bronchiolar epithelial cell lineage markers, depending on the type of injury.

Nervous system: neurogenic differentiation

Stem cells represent a promising therapeutic strategy for the cure of different neurologic diseases including neurodegenerative disorders and central nervous system acute injuries.^88,89 The demonstration that AFS cells are capable of entering the neuroectodermal lineage in neuronal conditions was given in 2007. Monoclonal CD117⁺ human AFS cells were induced to differentiate upon the neurogenic lineage as proven by specific marker expression (e.g., GIRK potassium channels), morphology changes, exhibition of barium-sensitive potassium current and glutamate release after stimulation.³⁵ However, CD117⁺ AFS cell capacity of producing mature, functional neurons in vitro and in vivo is still under investigation.^41,90,91

Discussion and Conclusions

AFS cell biological properties

The AF contains a broad range of cells, the majority of which are terminally differentiated and have limited proliferative capabilities. Since the first evidence that the AF could contain stem cells in 2003,²⁰ many groups have reported the isolation of multipotent and pluripotent cells.^33,34 In many studies, however, mixtures of unselected, non-clonal cells have been employed,⁹⁰ leading to conflicting results and uncertainty regarding the identity of the cell population employed. Indeed, stem cells within the AF harbor a spectrum of potency ranging from multipotent cells to unipotent committed progenitors (e.g., CD24⁺OB^-cadherin⁺ renal progenitors, SOX2⁺ neural progenitors and CD133⁺ endothelial progenitors).^92-95 AFS cells are a subpopulation (approximately 1%) of cells within the AF expressing the type III tyrosine kinase receptor of the stem cell factor (i.e., c-kit or CD117). Since their first isolation in 2007,³⁵ succeeding studies have provided evidence concerning their broad multipotency and therapeutic potential in animal models of disease. Morphological, immunocytochemical, biochemical and molecular investigations have revealed that AFS cells represent a novel and specific entity, being distinct from other stem cell populations including those deriving from extra-embryonic tissues. AFS cells present intermediate characteristics between ES cells and mesenchymal stem cells (MSCs)^17,96 as they, on the one hand, express embryonic markers, harbor a great proliferative and clonogenic potential, differentiate into multiple lineages and generate EBs and, on the other hand, express mesenchymal markers and do not form teratomas after implantation in vivo. Despite AFS cell capacity of differentiating into lineages representative of all embryonic germ layers and of generating EBs, their lack of tumorigenesis has been frequently advocated against their pluripotency. To this regard, the recent work of Moschidou et al.³⁸ has demonstrated that AFS cells can be reprogrammed to functional pluripotent stem cells (i.e., able to generate teratomas after injection in vivo) with the exclusive employment of a histone deacetylase inhibitor in a transgene-free approach, thus suggesting that only epigenetic modifications are required to obtain a whole pluripotent state in these cells. It is to be emphasized that this result was obtained with the exclusive employment of 1st trimester human AFS cells; further studies are therefore needed to explore eventual differences among AFS cell obtained at different gestational ages and to evaluate the efficacy of valproic acid reprogramming on 2nd trimester AFS cells. Moreover, the differentiation capacity and the tissue repair potential of AFiPS cells need to be investigated in vivo in animal models of disease.

Based on available data, AFS cells carry specific advantages in comparison to other stem cell types such as adult stem (AS) cells, ES cells, or induced pluripotent stem (iPS) cells.

(1) Obtainment

While AS cells can be difficult to isolate from their niches and ES cells derivation requires the disruption of embryos, AFS cells can be easily obtained without ethical concerns from the AF during termination procedures or diagnostic amniocenteses. Actually, the safety of amniocentesis has been established by several studies documenting an extremely low overall fetal loss rate (0.06–0.83%) related to this procedure^97,98 and AFS cells can be obtained from AF samples without interfering with diagnostic procedures.^99-101

(2) Safety

While ES and iPS cells are tumorigenic, AS cells can harbor genetic alterations occurring with aging and iPS cells accumulate genetic alterations during propagation in culture, AFS cells are genomically stable and do not form tumors when injected in vivo.^99-102

(3) Reprogramming

The reprogramming of adult somatic cells is a complex process that requires the introduction of exogenous factors. First trimester AFS cells, otherwise, can be reprogrammed with high efficiency and with a stable maintenance of the pluripotent phenotype only with the employment of a chemical substance (i.e., valproic acid, an FDA-approved drug for the treatment of epilepsy), thus reducing the risks associated with the random integration of the reprogramming transgenes into the host genome.^103-108 Many questions concerning AFS cell biological properties remain to be solved including origin, biological function in vivo, epigenetic state, and immunological reactivity. Answers to all of these issues are necessary before AFS cell employment in therapy.

AFS cell therapeutic potential

Despite the very recent identification of AFS cells, many studies have investigated their actual differentiation and therapeutic potential in vivo and in experimental models of organ development (i.e., organotypic cultures).^42,46 Second trimester human AFS cells or rodent cells of equivalent gestational age have been employed in all cases; to our knowledge, no data resulting from the administration of 1st trimester AFS cells or reprogrammed AFiPS cells are currently available. In the majority of the existing studies AFS cells have been utilized after previous expansion in culture; only in the works by Ditadi et al.³⁷ and Piccoli et al.⁴⁴ the cells were administered in vivo immediately after isolation from the AF without any passage in vitro.

Pre-expanded in culture AFS cells

In organotypical cultures of embryonic organs (kidney and lung), AFS cells integrate in the developing tissue, participate to organogenesis and express markers of differentiation toward committed cells.^42,46 In the absence of organ damage, AFS cells diffuse systemically and home to different organs but do not show signs of commitment.^46,109 In animal models of disease, systemically administered AFS cells integrate into the damaged tissue and differentiate toward endoderm- and mesoderm-derived specialized cells (data concerning AFS cell capacity of differentiating toward the ectodermal lineage in vivo are still lacking). Despite a low efficiency of tissue engraftment, AFS cells significantly improve animal survival and reduce tissue damage both at functional and morphological levels. The common hypothesis is that AFS cells act by modulating inflammation, apoptosis as well as cell proliferation via the production of paracrine/endocrine effectors.^39,51,57,61

Non-cultured AFS cells

In the work of Ditadi et al.³⁷ and Piccoli et al.⁴⁴ low numbers of murine AFS cells were injected in mice (around 2 × 10⁴ cells/animal) respectively affected by severe immunodeficiency and muscle dystrophy. Freshly isolated AFS cells not only improved animal morbidity and mortality but also showed high rates of tissue engraftment and cell differentiation, thus regenerating all hematopoietic lineages and muscle fibers after primary and secondary transplantation.^37,44 The discrepancy among results obtained with cells pre-expanded in culture vs. freshly isolated AFS cells in terms of regeneration potential may be determined by cell changes induced by culture conditions (e.g., reduced survival of the stem cell fraction, beginning of differentiation processes and modifications of cell epigenetic state). Thus, optimization of AFS cell culture conditions is mandatory in order to progress toward the therapeutic application of these cells. Although further studies are needed to better understand their biological properties and to define their therapeutic potential, AFS cells appear to be promising candidates for cell therapy and tissue engineering. In a future clinical scenario, AFS cells collected during routinely performed amniocenteses and termination procedures could be banked and, in case of need, subsequently expanded in culture or engineered in acellular grafts.^17,110 The autologous employment of AFS cells looks particularly appealing for the treatment of perinatal disorders such as congenital malformations (e.g., congenital diaphragmatic hernia) and acquired neonatal diseases requiring tissue repair/regeneration (e.g., necrotizing enterocolitis). In this way, affected children could benefit from having autologous expanded/engineered cells ready for implantation either before birth or in the neonatal period. Moreover, considering that a bank of 150 donor cell lines would provide a beneficial match for up to 37.9% of the population,¹¹¹ AFS cell banks would also benefit their heterologous employment as multi/pluripotent cells of fetal origin.

Acknowledgments

M.C. is supported by Citta’ della Speranza, Malo Italy. P.D.C. is supported by the Great Ormond Street Hospital Children’s Charity.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.