Placenta as a source of hematopoieticstem cellsElaine Dzierzak and Catherine Robin

Erasmus MC Stem Cell Institute, Dept of Cell Biology, Erasmus University Medical Center Rotterdam, The Netherlands

The placenta is a large, highly vascularised hematopoie-tic tissue that functions during the embryonic and foetaldevelopment of eutherian mammals. Although recog-nised as the interface tissue important in the exchange ofoxygen, nutrients and waste products between the foe-tus and mother, the placenta has increasingly become afocus of research concerning the ontogeny of the bloodsystem. Here, we describe recent data showing theintrinsic hematopoietic potential and appearance ofhematopoietic cells in the mouse and human placentaand probe the biological rationale behind its hemato-poietic function. As a rest tissue that contains potenthematopoietic stem cells (HSCs), the human placentacould represent (in addition to umbilical cord blood cells)an accessible supplemental source of cells for thera-peutic strategies.

Early placental developmentThe placenta is a mysterious tissue that has been knownsince the earliest times as the ‘‘alter ego’’ or ‘‘external soul’’of the foetus. Aristotle recognised it as a tissue from whichthe embryo receives its nourishment, and Leonardo daVinci and Vesalius depicted it as a disc at the maternal–foetal interface in the uterus [1]. It was only in the mid-1500s that it gained its name placenta (from Latin), mean-ing flat cake. Proposed to be the liver and lungs of thefoetus, only in 1734 was it established that the foetal andmaternal vasculature of the placenta were not continuous[2]. Although this aspect of placental anatomy remainedcontroversial for many years (awaiting improved micro-scopic techniques and molecular markers), the function ofthe placenta was accepted to be the facilitation of nutrientand waste exchange between the mother and foetus, pro-vision of immunoprotection for the foetus and production offactors and hormones for foetal growth [3].

The development of the placenta has been extensivelystudied in the mouse conceptus [4]. At embryonic day (E)8.5, the allantois, a mesodermal outgrowth from theposterior end of the embryo, contacts the chorionic epi-thelium in the extraembryonic part of the conceptus(Figure 1a). This event is called chorio-allantoic fusion.Thereafter, a network of vasculature is generated from theallantois and grows into the chorionic plate. These vesselsundergo extensive branching. The endothelial cells liningthese vessels are in direct apposition with syncytiotropho-blasts, and it is in the small spaces surrounding thesestructures that thematernal blood bathes the foetal villi. Adiagram of the anatomical structure of the mouse placenta

is provided in Figure 1b. Instead of the compact, maze-likevilli in the labyrinth layer found in the mouse, the villousstructures in the human placenta are less branched andthe intervillous space is more open. Placental architecturevaries between species, but there are general similaritiesacross evolution that include placental cell types, functionsand gene and protein expression, reflecting developmentaland regulatory conservation [5,6].

The placenta is connected to the embryo through theumbilical cord. The umbilical artery is contiguous with thedorsal aorta, the main artery of the embryo (Figure 2a).Initially, the connection is at the caudal aspect of the dorsalaorta, but following vascular remodelling the connectionbecomes abdominal [7]. Blood circulation throughout theembryo and extraembryonic tissue is established at E8.25in themouse (Figure 2b) and begins between weeks 3 and 4of gestation in the human conceptus. Following thedevelopment and growth of the placenta at these earlystages, maternal circulation flows through the intervillousspaces beginning at approximately E10.5 and about the11th week of gestation in the mouse and human placentas,

Review

Glossary

Placenta: extraembryonic tissue derived from the chorion and allantois of theearly stage mammalian conceptus.Allantois: posterior outgrowth of the early stage embryo with hematopoieticpotential. It gives rise to the vessels in the umbilical cord of mammals. In birdsand reptiles, it is involved in oxygen exchange and is a reservoir of nitrogenouswaste.Chorion: membrane separating the foetus and mother that is formed by theextraembryonic mesoderm and trophoblast.Endosteal and endothelial niches: specific anatomical locations/microenviron-ments within adult bone marrow that produce factors or provide importantcell–cell interactions for the maintenance, self-renewal and/or differentiation ofhematopoietic progenitors and stem cells. The endosteal niche-containingosteoblasts are juxtaposed to the bone. The endothelial niche is thevasculature within the bone marrow.Hematopoietic stem cells (HSCs): the rare cells existing within adult bonemarrow cavities that contribute to the lifelong production of all blood cells ofthe hematopoietic system. These cells are long-lived, self renewing andpossess the potential to produce all hematopoietic cell lineages.Hematopoietic progenitor cells: intermediate cells within the adult hemato-poietic cell differentiation hierarchy. Having restricted differentiation potential,they expand and differentiate to one or a few hematopoietic lineages. They aregenerally short-lived and do not self-renew.Hemangioblastic cords: in the early-stage human placenta, mesodermal/mesenchymal cells that are the precursors of vascular endothelial andhematopoietic cells.Labyrinth: region of the placenta in which the trophoblast and its associatedfoetal blood vasculature undergoes extensive branching of the villi into adensely packed structure.Pericytes: a relatively undifferentiated cell type that surrounds blood vessels. Itpossesses multilineage differentiation potential and is thought to be theprecursor of mesenchymal stem/stromal cells.Syncytiotrophoblasts: the outer most cell type of the foetal placenta. Theseinvade the uterine wall and provide the large surface area for the exchange ofoxygen, nutrients and waste between the foetus and mother.

Corresponding author: Dzierzak, E. (e.dzierzak@erasmusmc.nl).

1471-4914/$ – see front matter ! 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.molmed.2010.05.005 Trends in Molecular Medicine 16 (2010) 361–367 361

respectively [8]. It is at this time that the placenta becomesfunctional in its role as the exchange chamber betweenmother and foetus.

The placenta and hematopoiesisCompared with its generally recognised functions, theappreciation of the placenta as a potent hematopoietic site

is relatively recent. Hematopoietic activity was initiallyobserved in themouse placenta in the 1960s and 1970s, butthese findings were not immediately pursued (reviewed in[9]). Studies in human placental villi have suggested thatalready at day 21 postconception, macrophage-like cellsand hemangioblastic cords arise from mesenchymal cells[10]. Tissue-grafting studies in the avian embryo model byDieterlen-Lievre revealed that cells derived from the avianallantois contribute to adult haematopoiesis [11]. Sub-sequent studies by this group established that the mouseplacenta harbours a wide range of clonogenic hematopoie-tic progenitors beginning around E9 [12]. Although this isslightly later than the time such cells appear in the embryoproper or the yolk sac (Figure 2b), the placenta contains themost progenitors of any site up until E12 when the foetalliver surpasses it. The continued presence of hematopoieticprogenitors in the mouse placenta throughout gestationdemonstrates that the placenta it is a highly potent hema-topoietic site.

HSCs are found in highly vascularised tissues includingthe mouse placenta (reviewed in [13–15]). HSCs are thebasis of the adult hematopoietic hierarchy that producesall the blood lineages throughout adult life. Putative HSCsare tested by stringent transplantation assay in which thedonor cells are challenged to provide the complete, long-term hematopoietic repopulation of adult irradiated (HSC-depleted) normal recipients. Using allelic or transgenemarkers to distinguish foetal-derived cells, HSCs aredetectable in the mouse placenta at early E11 and HSC

[(Figure_1)TD$FIG]

Figure 1. Placenta structure at different times in development (a) Early stages ofchorio-allantoic placental formation. The placenta is formed from the fusion of theallantois with the chorion. The allantois is a mesodermal outgrowth emanating fromthe posterior primitive streak. It elongates and upon contact with the chorionicmesoderm gives rise to the labyrinth of the placenta. (b) Mouse placenta structure.The placenta consists of several cell types and layers. The side of the placenta facingthe foetus is the labyrinth. This consists of a vascular network contiguous with theumbilical cord. The network of vessels in the labyrinth form branched villousstructures that are surrounded by mesenchymal stromal and syncytiotrophoblastcells. The maternal vessels run through the spongiotrophoblast region (facing themother) to open into the intervillous space (white area) where the physiologicexchange between mother and foetus occurs.

[(Figure_2)TD$FIG]

Figure 2. Hematopoiesis in the mouse embryo and placenta. (a) Sites of hematopoiesis in the midgestation (E10.5) mouse conceptus. Extraembryonic hematopoieticterritories include the chorio-allantoic placenta and yolk sac. Hematopoietic territories within the embryo body are the aorta (hematopoietic part of the AGM region) andliver. The yolk sac, placenta and AGM are sites of de novo hematopoietic cell generation, whereas the liver is colonised by exogenously generated hematopoietic cells. (b)The temporal appearance of hematopoietic cells in different tissues (from E7.5 until birth).

Review Trends in Molecular Medicine Vol.16 No.8

362

numbers increase dramatically up to E12.5 [16,17]. There-after, HSC numbers in the placenta are superseded by thefoetal liver, and after E15.5 very few or no HSCs are foundin the placenta [16]. Placental HSCs express many of thesame surface marker proteins as adult bone marrow andfoetal liver HSCs, including CD34 and c-kit [16]. Further-more, all placental HSCs express Ly6A (Sca-1) GFP (stemcell antigen-1, green fluorescent protein) [17]. Interest-ingly, Ly6A GFP-expressing cells localise within the vas-culature of the placental labyrinth and the umbilicalvessel, and most of these cells express CD34. Histologicanalyses have shown that the midgestation mouse pla-centa expresses important hematopoietic transcriptionfactors such as Gata-2, Gata-3 and Runx1 [17]. Gata-2 isexpressed in some endothelial cells and cells surroundingthe vessels within the labyrinth, whereas Gata-3 isrestricted to a few cells at the maternal–foetal interface.Runx1 is expressed in cells within the vascular lumen andthe endothelium as well as cells surrounding the vascula-ture of the labyrinth [17,18]. The patterns of Gata-2 andRunx1 expression strongly suggest HSCs and progenitorsare localised within the labyrinth and near the chorionicplate.

Human placental hematopoiesisThroughout development, the human placenta similarlycontains a wide variety of hematopoietic cells, as well asmature and immature hematopoietic progenitors andHSCs (Figure 3). Primitive erythroblasts that morphologi-cally resemble those in the yolk sac fill the placental vesselsbeginning around day 24 [19]. These cells express glyco-phorin-A, GATA-2 and c-KIT, but are not positive for CD34or CD45. As measured by in vitro clonogenic activity,mature and immature hematopoietic progenitors are foundas early as week 6 in gestation through week 17, and atterm [20–23]. These progenitors are multipotent and pro-duce erythroid and myeloid lineage cells, including gra-nulocytes and macrophages. The progenitors are initiallyin both CD34– and CD34+ fractions, but by week 15 allprogenitors are CD34+ [20–22]. Leukocytes begin toexpress CD45 at 12–14 weeks in gestation and at term.

As a hematopoietic territory, the appearance of hema-topoietic cells in the early gestational stage human pla-centa is slightly delayed compared with the other

hematopoietic sites (Figure 3). The human yolk sac beginsgenerating blood at day 16 with the production of primitiveerythroid cells, and at day 19 the intraembryonic splanch-nopleura (aorta region) becomes hematopoietic. The emer-gence of multipotent progenitors, HSCs and clusters ofcells closely adherent to the ventral wall of the dorsalaorta starts at around day 27 in the developing splanch-nopleura/aorta-gonad-mesonephros (AGM) region [24].Beginning at day 30 and week 13, respectively, the firsterythroid progenitors and multilineage progenitors colo-nise the liver. Thereafter, the bonemarrow becomes hema-topoietic [25]. Thus, the sequential waves of hematopoieticactivity in the human conceptus and placenta are similarto those in the mouse conceptus.

Using the NOD-SCID (non-obese diabetic-severe com-bined immunodeficient) mouse transplantation assay (thesurrogate transplantation assay for human HSC identifi-cation), potent multilineage, high-level repopulating HSCshave been found in the human placenta [22]. Donor humanplacental cells contribute to high percentages of B lympho-cytes and myeloid cells in the bone marrow and spleen ofNOD-SCID recipients 10 weeks postinjection. HSCs can befound in the human placenta beginning at week 6 ofgestation, throughout trimesters 1 and 2 and at the timeof delivery (term). Considering the almost total absence ofHSCs in the mouse term placenta [16], the discovery ofHSCs in the human term placenta was unexpected.Human placental HSCs were verified as being foetal-derived by human forensic identity PCR analysis, indicat-ing that the human placenta still functions to maintain,expand and/or produce the most immature of foetus-derived hematopoietic cells [22].

Interestingly, cell extraction methods allowing theisolation of cells from the placenta vasculature revealthat many potent HSCs are either closely adherent to theendothelium or are within the niches of this compart-ment [22]. At week 6 in gestation, HSCs are found both inthe CD34+ and CD34! fractions. It is as yet uncertainwhether at later stages of development placental HSCsare retained in both these fractions, as they are inumbilical cord blood [26,27]. If so, the vascular endo-thelium might be either considered the cellular source ofHSCs or provide an important HSC growth-supportivemicroenvironment.

Microenvironment of the placentaHematopoietic tissues harbour hematopoietic progenitorsand stem cells in specificmicroenvironments or niches [28].Studies of adult bone marrow have shown that the micro-environment is a complex structure composed of mesench-ymal stromal cells (MSCs), endothelial cells, osteoblasts,adipocytes, the extracellular matrix, growth factors, cyto-kines and adhesive molecules that provide support andregulatory functions for hematopoietic progenitor or stemcell homing, self-renewal, maintenance and differen-tiation. A close relationship exists between HSCs andthe endosteal and endothelial niches in the bone marrow[29,30]. The hematopoietic niche of the placenta within thelabyrinth is likely to be similar, consisting of endothelial,perivascular and mesenchymal cells but also placenta-specific syncytiotrophoblast cells.

[(Figure_3)TD$FIG]

Figure 3. Hematopoiesis in the human conceptus and placenta. The temporalappearance of hematopoietic cells in the yolk sac, AGM region, liver and placentaof the human conceptus from gestational weeks in the first and second trimestersthrough to term. BFU-E (burst forming unit-erythroid) represents the earliesterythroid progenitors. Multipotent progenitors can produce erythroid and myeloidlineage cells.

Review Trends in Molecular Medicine Vol.16 No.8

363

Interestingly, throughout mouse development, thetemporal and spatial distribution of MSCs correlates withhematopoietic territories such as the AGM, foetal liver andneonatal bone marrow [31]. MSC lines have been derivedfrom these tissues, andmany of these cell lines can providesupport for hematopoietic cells in vitro [32,33], underliningthe role that stromal cells play in the proliferation and/orsurvival of HSCs. Such MSCs have osteogenic, adipogenic,chondrogenic and/or myogenic differentiation potential[34,35]. Because the mouse placenta is a potent hemato-poietic territory it is expected that this tissue will also yieldhematopoietic-supportive MSC lines and exhibit similardifferentiation potentials.

MSC lines have already been isolated from humanplacenta and amniotic and chorionic foetal membranes(reviewed in [36]). At the developmental stages tested,ranging from week 3 to term, human MSC lines expressclassical mesenchymal markers as well as markers ofpericytes, CD146 and NG2 [22,37], and after gestationalweek 6, they possess the typical mesenchymal lineagepotentials (osteogenic, adipogenic and/or endothelial)[22,36–46]. Furthermore, some of these placental MSClines also constitute a potent feeder layer for the in vitromaintenance and/or expansion of human umbilical cordblood CD34+ cells and progenitors [22], primate andhuman embryonic stem (ES) cells [23,47,48] and long-term-culture-initiating cells [23]. Immunostained humanplacental sections localise such mesenchymal cell typeswithin this tissue. Moreover, the localised expression ofCD146 and NG2 in the perivasculature of the placentasuggests that placental mesenchymal cells are pericytes[22,37]. Interestingly, a MSC line derived from thematernal part of a gestation week 3 human placenta hasprovided the potent in vitromaintenance and expansion ofhuman umbilical cord blood CD34+ cells and an eightfoldincrease in immature hematopoietic progenitors comparedwith the input number in the sorted CD34+ cord bloodpopulation [22], suggesting that maternal cells contributeto the early hematopoietic supportive microenvironmentby promoting the growth of the placenta as a highlyvascular and hematopoietic territory. Thus, the hemato-poietic inductive/supportive microenvironment of the pla-centa could be unique compared with the other foetal andadult hematopoietic territories, and it most likely consistsof the vascular endothelial, mesenchymal and syncytiotro-phoblast cells that develop in parallel in the villi.

The coordinated development of the microenvironmentwithin the labyrinth requires vessel generation, invasion,branching morphogenesis and syncytiotrophoblast differ-entiation [4]. Genetic studies in mutant mice have ident-ified a panel of genes important for some of these processes[5]. For example, the transcription factor encoded byGcm1(glial cells missing 1) is a pivotal molecule in the initiationof morphogenesis and syncytiotrophoblast differentiation.The germline deletion of Gcm1 in mice is lethal at mid-gestation owing to a failure to develop the placental labyr-inth layer [49]. Originally identified in Drosophila, Gcm isinvolved in macrophage-like cell development [50]. In themouse, it is almost exclusively expressed in the placenta[51]. Similarly, the deletion of Esx1, a homeobox gene,results in a failure to develop the labyrinth layer [52].

PPARg is required for the invasion of foetal vessels intothe presumptive labyrinth at E9.5, as are developmentalfactors such Wnt2 and EphB4/ephrin B2 [4]. It will beinteresting to determine how the placental hematopoieticmicroenvironment and hematopoiesis are affected in thesemouse models, and whether hematopoiesis occurs nor-mally in other hematopoietic sites.

Generator or storage tissue for hematopoietic cells?The hematopoietic system originates in the mesodermalgerm layer of the conceptus. The histologic observation ofyolk sac blood island anatomy has linked the developmentof vascular endothelial and hematopoietic cells [53,54].These studies have suggested a common mesodermal pre-cursor for these two lineages of cells, and the precursor cellhas been named hemangioblast. Other histologic studieshave proposed that the precursors of hematopoietic cells inthe embryonic dorsal aorta are hemogenic endothelialcells, because clusters of hematopoietic cells are closelyassociated with the ventral endothelium of the dorsal aortaat early stages of development [13]. Cell tracing [55,56] inmouse models in conjunction with vital imaging of mouseES cell hematopoietic differentiation cultures and in vitrocultured early posterior primitive streak cells [57,58] haveshown that hematopoietic cells transit through an endo-thelial cell stage before taking on hematopoietic fate. Invivo real-time imaging of the intact midgestational dorsalaorta has shown the emergence of hematopoietic progeni-tor/stem cells from endothelial cells lining this vessel [59].Because the vasculature and other cells of the placentaoriginate from the allantoic mesoderm, which arises fromthe primitive streak at early stages of development, theallantoic mesoderm is postulated to be hemogenic.

Indeed, the allantois of the developing chick can producehematopoietic cells [11,60]. Blood island-like clusters ofhematopoietic cells are found in the prevascularised allan-tois, and upon engraftment into the coelom of host embryosthe cells arising from the donor allantois contribute toadult blood. Results in mouse embryo-grafting exper-iments examining this question are less clear. Afterengraftment and a short culture period, donor allantoiccells contribute to the endothelial lineage but only veryrarely to a small number of erythroid cells [61]. Morerecently, others have found that mouse prefusion allantoisand chorion tissues both possess intrinsic hematopoieticpotential [62,63]; they give rise to clonogenic haematopoie-tic progenitors following a 48-hour organ explant cultureperiod [63]. Moreover, definitive hematopoietic markerssuch asRunx1 (a pivotal hematopoietic transcription factor[64]) and Ly6A GFP (a HSC marker [65]) are expressed inthe early allantois and chorion, indicating their hemato-poietic potential [63]. Thus, before it becomes vascularisedthe early stage placenta is intrinsically hemogenic.

At a slightly later developmental stage, the mouseplacenta generates hematopoietic progenitors. Embryosdeficient for the Ncx1 gene [66] lack a heartbeat and bloodcirculation, which is normally established between theembryo body and the extraembryonic tissues at E8.25. Ifin the absence of circulation a tissue such as the yolk sac,placenta or the body of the embryo contains hematopoieticcells, then the hematopoietic cells must be generated

Review Trends in Molecular Medicine Vol.16 No.8

364

intrinsically [67]. Indeed, erythro-myeloid and lymphoidprogenitors are detected in the Ncx1!/! embryo body,yolk sac and placenta [18,67], indicating their intrinsichemogenic/hematopoietic capacity.

Because Ncx1!/! embryos die before E11, the time atwhich HSC activity is first detected in the placenta, it is asyet uncertain whether the placenta can generate HSCs.Quantitative studies, in which HSC numbers in each of thetissues of the conceptus have been determined [68],suggest that the aorta (the only tissue shown to autonom-ously generate HSCs) cannot produce all of the HSCs thateventually are found in the foetal liver and the adult bonemarrow (tissues that harbour but do not generate HSCs)(Figure 2). In this regard, themidgestationmouse placentacontains an abundance of HSCs [16,17], supporting thenotion that this highly vascularised tissue generates HSCsfrom hemogenic endothelium and/or that it provides aunique supportive growth niche for the expansion ofaorta-derived HSCs. Similarly, the human placenta mightalso autonomously generate HSCs and/or promote theirexpansion to large numbers before theymigrate to the bonemarrow. Recent studies showing the importance of mech-anical stimuli provided by circulation in AGM HSC de-velopment [69] have suggested that mechanical stimuli aswell as hypoxia [70] could be important factors in placentaland HSC development.

The placenta in hematopoietic developmentHematopoiesis in the mouse and human conceptus pro-gresses in wave-like stages (reviewed in [13,25]). The threehemogenic/hematopoietic tissues – the yolk sac, AGMregion (and its precursor tissue the para-aortic splanchno-pleura) and chorio-allantoic placenta – are active at differ-ent but overlapping stages of development. They producevarying repertoires and quantities of hematopoietic cellsand support them in distinct microenvironments. It iscurious that eutherian mammals use the placenta as ahemogenic and hematopoietic tissue, whereas other mam-mals and nonmammalian vertebrates develop a fully func-tional hematopoietic system in the absence of a placenta.However, in these other vertebrate embryos, it is possiblethat hematopoietic cells emerging in the allantois areamplified elsewhere. In marsupials, the allantois is mainlyavascular and does not fuse with the chorion. Many vari-ations in placental structure are seen in mammals. It isdiscoid in mouse and man, but in species such as the pig,horse and whale it is diffuse and distributed over most ofthe uterus inner surface [4]. Interestingly, the evolution ofthe placenta as a complex organ has occurred multipletimes, as found in the fish genus Poeciliopsis [71]. Hence, isthe more fully developed placenta irrelevant as a hemato-poietic tissue or does it play a special role in blood de-velopment in mammals?

One notion is that the large size of the mammalianfoetus and the extended length of the gestational periodmight require a greater quantity of hematopoieticprogenitors and HSCs for foetal growth, and this canonly be accomplished in an extra hematopoietic tissuesuch as the placenta. Indeed, the placenta is largecompared with the size of the other hematopoietictissues of the embryo, and thereby represents a

significant space for producing, amplifying and/orharbouring hematopoietic cells.

Another idea concerning the functional relevance of theplacenta comes from studies of Runx1 haploinsufficientmouse embryos; in thesemice, which have half a dose of theRunx1 transcription factor protein, the AGM region pro-duces fewer HSCs [43,72]. However, Runx1 haploinsuffi-ciency does not have negative effects on HSCs in theextraembryonic tissues (yolk sac and placenta) or foetalliver, and instead results in a surprising increase in HSCnumbers in the placenta [43]. Hence, the placenta seems tobe more resistant to genetic/physiologic changes, implicat-ing it as a highly robust hematopoietic tissue in nonhomeo-static conditions.

Additionally, the provision of growth factors such asinterleukin 3 (IL-3) by the maternal part of the placentacan stimulate hematopoiesis [73]. IL-3 is a potent HSCsurvival and proliferation factor during embryonic devel-opment [43], and the IL-3 gene is a known direct down-stream transcriptional target of Runx1. Interestingly,human placental MSCs produce SCF (stem cell factor),Flt3 ligand, IL-6 and M-CSF (macrophage colony-stimu-lating factor) among other hematopoietic factors [23].Thus, both maternal- and foetal-derived factors can con-tribute to the growth of the placental hematopoietic nicheand the support of hematopoietic cells.

Overall, the human placenta plays an important andmultifaceted role in the development and growth of thefoetus. Its structural complexity and its immense vascularnetwork containing large quantities of circulating bloodcells and progenitors make it a difficult tissue to under-stand, particularly concerning its precise role in the de-velopment of the mammalian hematopoietic system. Theongoing challenge is to determine whether the placentajust contains reserve cohorts of hematopoietic progenitorand stem cells, or whether these cohorts of cells arerequired for adult hematopoiesis and migrate to the bonemarrow at the end of foetal development before the time ofdelivery.

Concluding remarksThe discovery of HSCs in the human placenta throughoutdevelopment and their localisation to the highly vascularcompartment of this tissue opens new lines of enquiry withpotential medical implications (Box 1). As a new source ofHSCs, the term human placenta as yet yields only rela-tively small numbers. The present cell extraction pro-cedures that include dissection, enzymatic digestion withthree enzymes (collagenase, dispase and pancreatin) andmechanical dispersion yield about 10% of the total numberof HSCs found in a unit of umbilical cord blood [22,74].Although placental HSCs can be stored much like umbi-lical cord bloodHSCs, harvest efficiencies togetherwith thecost- and labour-effectiveness must improve before anypotential clinical applications can be considered. In thisregard, a more easily isolatable source of HSCs with thepotential for regenerative medicine has recently beendiscovered, the amniotic fluid [75].

Importantly, knowledge from developmental studiesdemonstrating the hemogenic nature of the early-stagechorio-allantoic placenta [18,62,63] offers a new and

Review Trends in Molecular Medicine Vol.16 No.8

365

exciting challenge in the field. Insights into the signals thatdrive hematopoietic cell generation from mesodermal pre-cursors, endothelial cells and the development of the pla-centa labyrinth and villi could provide new strategies forthe production of hematopoietic cells from the numerousvascular endothelial cells of the foetal part of this tissue. Arecent study with early-stage human embryos has ident-ified a marker ACE (angiotensin-converting enzyme,CD143) recognised by the BB9 antibody, on a subset ofhuman mesoderm that establishes a population of hemo-genic endothelial cells within the foetus [76]. The ACE-expressing cells possess hematopoietic potential in long-term cultures and in SCID mouse in vivo reconstitutionassays. Moreover, ACE-expressing cells isolated fromhuman ES cells differentiated into embryoid bodies arehemangioblastic-like angiohematopoietic cells. It will beinteresting to examine the human placenta for BB9 expres-sion and, if such cells are found, to examine their hemo-genic potential. The characterisation of the molecularprogram of ACE-expressing cells should yield informationon the genes involved in HSC specification, amplificationand maintenance.

Of future interest is the prospect of using this knowledgeof the genetic program to induce the differentiation of earlypopulations of hemogenic cells to HSC fate, particularly iflarge numbers of such cells can be isolated from theplacenta. Alternatively, with the advancement of the tech-nology of reprogramming somatic cells to pluripotent stemcells, it might be possible, with our complete knowledge ofthe pivotal factors involved in HSC generation, to obtainother patient-specific cells that can be reprogrammed intotherapeutically potent HSCs [77]. In this regard, the

human placental allantoic and chorionic foetal membranes[36] provide an abundant source of developmentally youngsomatic cells such as MSCs that can be stored, differen-tiated and/or reprogrammed for regenerative medicine.

AcknowledgementsThe authors thank lab members and particularly K. Bollerot, S. Mendes,E. Haak, M. Crisan, F. Cerisoli, I. Lauw, P. Kaimakis, R. Jorna, P.Imanirad, R. van der Linden, E. Steegers and T. Cupedo who contributedto these studies. We also thank Mihaela Crisan for critical reading of themanuscript. Support was provided by the Landsteiner Society for BloodResearch (0614), NIH R37 (DK51077), Dutch BSIK Stem Cells inDevelopment and Disease (03038) and NWO VIDI (917.76.345).

References1 Longo, L.D. and Reynolds, L.P. (2010) Some historical aspects of

understanding placental development, structure and function. Int. J.Dev. Biol. 54, 237–255

2 Pijnenborg, R. and Vercruysse, L. (2007) Erasmus Darwin’senlightened views on placental function. Placenta 28, 775–778

3 Gude, N.M. et al. (2004) Growth and function of the normal humanplacenta. Thromb. Res. 114, 397–407

4 Rossant, J. and Cross, J.C. (2001) Lineage Specification andDifferentiation: Extraembryonic Lineages. In Mouse development:patterning, morphogenesis and organogenesis (Rossant, J. and Tam,P.L., eds), pp. 155–174, Academic Press

5 Cross, J.C. et al. (2003) Genes, development and evolution of theplacenta. Placenta 24, 123–130

6 Rawn, S.M. and Cross, J.C. (2008) The evolution, regulation, andfunction of placenta-specific genes. Annu. Rev. Cell Dev. Biol. 24,159–181

7 Garcia-Porrero, J.A. et al. (1995) Potential intraembryonic hemogenicsites at pre-liver stages in the mouse. Anat. Embryol. (Berl.) 192, 425–

4358 Georgiades, P. et al. (2002) Comparative developmental anatomy of the

murine and human definitive placentae. Placenta 23, 3–199 Mikkola, H.K. et al. (2005) Placenta as a site for hematopoietic stem cell

development. Exp. Hematol. 33, 1048–105410 Demir, R. et al. (1989) Fetal vasculogenesis and angiogenesis in human

placental villi. Acta Anat. (Basel.) 136, 190–20311 Caprioli, A. et al. (1998) Blood-borne seeding by hematopoietic and

endothelial precursors from the allantois. Proc. Natl. Acad. Sci. U. S. A.95, 1641–1646

12 Alvarez-Silva, M. et al. (2003) Mouse placenta is a major hematopoieticorgan. Development (Cambridge, England) 130, 5437–5444

13 Dzierzak, E. and Speck, N.A. (2008) Of lineage and legacy: thedevelopment of mammalian hematopoietic stem cells. Nat.Immunol. 9, 129–136

14 Ottersbach, K. and Dzierzak, E. The placenta as a haematopoieticorgan. Int. J. Dev. Biol. In press

15 Gekas, C. et al. (2008) Isolation and visualization of mouse placentalhematopoietic stem cells.Current protocols in stem cell biologyChapter2, Unit 2A 8 1–2A 8 14

16 Gekas, C. et al. (2005) The placenta is a niche for hematopoietic stemcells. Dev. Cell 8, 365–375

17 Ottersbach, K. and Dzierzak, E. (2005) The murine placenta containshematopoietic stem cells within the vascular labyrinth region.Dev. Cell8, 377–387

18 Rhodes, K.E. et al. (2008) The emergence of hematopoietic stem cells isinitiated in the placental vasculature in the absence of circulation. CellStem Cell 2, 252–263

19 Challier, J.C. et al. (2005) Immunocytological evidence forhematopoiesis in the early human placenta. Placenta 26, 282–288

20 Barcena, A. et al. (2009) The human placenta is a hematopoietic organduring the embryonic and fetal periods of development. Dev. Biol. 327,24–33

21 Barcena, A. et al. (2009) A new role for the human placentaas a hematopoietic site throughout gestation. Reprod. Sci. 16,178–187

22 Robin, C. et al. (2009) Human placenta is a potent hematopoietic nichecontaining hematopoietic stem and progenitor cells throughoutdevelopment. Cell Stem Cell 5, 385–395

Box 1. Outstanding questions

" Does the placenta have the ability to generate HSCs? The placentacontains HSCs [22], but it is as yet undetermined whether theplacenta can intrinsically produce these potent, therapeuticallyimportant stem cells [18].

" Does the placenta contain hemangioblasts and/or hemogenicendothelial cells? Results from studies of the mouse chorion andallantois have shown that these tissues possess intrinsic hema-topoietic potential [62,63]. In the absence of circulation, the mouseplacenta intrinsically generates multipotent progenitors [18];however, it is as yet unknown whether the vasculature of thelabyrinth and/or the mesenchymal cells within the villi arehemogenic.

" How many HSCs does a human placenta contain? If placenta-derived cells are to be contemplated for hematopoietic transplan-tation purposes, they must be at least as abundant as thoseavailable in umbilical cord blood. Improvements in extractionprocedures should optimise both the quantitative yields andviability of these cells.

" How robust are placental HSCs in transplantation scenarios?Xenotransplants of human placental cells have revealed thepresence of multipotent, long-term engrafting HSCs. However, itis uncertain whether placental HSCs are qualitatively as potent (inproliferation, differentiation and self-renewal) as those fromumbilical cord blood and adult bone marrow.

" Does the placenta contain a unique hematopoietic inductive/supportive microenvironment? The placenta is an extraembryonicand transient tissue; it is uncertain whether unique placental celltypes and genetic programs regulate the development ofhematopoietic cells differently than in the other embryonic andadult hematopoietic tissues.

Review Trends in Molecular Medicine Vol.16 No.8

366

23 Zhang, Y. et al. (2004) Human placenta-derived mesenchymalprogenitor cells support culture expansion of long-term culture-initiating cells from cord blood CD34+ cells. Exp. Hematol. 32, 657–664

24 Tavian,M. etal. (2001)Thehumanembryo,butnot itsyolk sac,generateslympho-myeloid stem cells: mappingmultipotent hematopoietic cell fatein intraembryonic mesoderm. Immunity 15, 487–495

25 Tavian,M. and Peault, B. (2005) Embryonic development of the humanhematopoietic system. Int. J. Dev. Biol. 49, 243–250

26 Bhatia,M. et al. (1998) A newly discovered class of humanhematopoieticcells with SCID-repopulating activity. Nat. Med. 4, 1038–1045

27 Wang, J. et al. (2003) SCID-repopulating cell activity of human cordblood-derived CD34- cells assured by intra-bone marrow injection.Blood 101, 2924–2931

28 Yaniv, I. et al. (2006) The tale of early hematopoietic cell seeding in thebone marrow niche. Stem Cells Dev. 15, 4–16

29 Calvi, L.M. et al. (2003) Osteoblastic cells regulate the haematopoieticstem cell niche. Nature 425, 841–846

30 Lo Celso, C. et al. (2009) Live-animal tracking of individualhaematopoietic stem/progenitor cells in their niche. Nature 457, 92–96

31 Mendes, S.C. et al. (2005) Mesenchymal progenitor cells localize withinhematopoietic sites throughout ontogeny. Development (Cambridge,England) 132, 1127–1136

32 Moore, K.A. et al. (1997) In vitro maintenance of highly purified,transplantable hematopoietic stem cells. Blood 89, 4337–4347

33 Oostendorp, R.A. et al. (2002) Stromal cell lines from mouse aorta-gonads-mesonephros subregions are potent supporters ofhematopoietic stem cell activity. Blood 99, 1183–1189

34 Durand, C. et al. (2006) Mesenchymal lineage potentials of aorta-gonad-mesonephros stromal clones. Haematologica 91, 1172–1179

35 Zipori, D. (2004) Mesenchymal stem cells: harnessing cell plasticity totissue and organ repair. Blood Cells Mol. Dis. 33, 211–215

36 Parolini, O. et al. (2008) Concise review: isolation and characterizationof cells from human term placenta: outcome of the first internationalWorkshop on Placenta Derived Stem Cells. Stem Cells 26, 300–311

37 Crisan, M. et al. (2008) A perivascular origin for mesenchymal stemcells in multiple human organs. Cell Stem Cell 3, 301–313

38 Fukuchi, Y. et al. (2004) Human placenta-derived cells havemesenchymal stem/progenitor cell potential. Stem Cells 22, 649–658

39 Igura, K. et al. (2004) Isolation and characterization of mesenchymalprogenitor cells from chorionic villi of human placenta. Cytotherapy 6,543–553

40 Li, C.D. et al. (2005) Mesenchymal stem cells derived from humanplacenta suppress allogeneic umbilical cord blood lymphocyteproliferation. Cell research 15, 539–547

41 Miao, Z. et al. (2006) Isolation of mesenchymal stem cells from humanplacenta: comparison with human bone marrow mesenchymal stemcells. Cell Biol. Int. 30, 681–687

42 Portmann-Lanz, C.B. et al. (2006) Placental mesenchymal stem cells aspotential autologous graft for pre- and perinatal neuroregeneration.Am. J. Obstet. Gynecol. 194, 664–673

43 Robin, C. et al. (2006) An unexpected role for IL-3 in the embryonicdevelopment of hematopoietic stem cells. Dev. Cell 11, 171–180

44 Wulf, G.G. et al. (2004) Mesengenic progenitor cells derived fromhuman placenta. Tissue Eng. 10, 1136–1147

45 Yen, B.L. et al. (2005) Isolation of multipotent cells from human termplacenta. Stem Cells 23, 3–9

46 Zhang, X. et al. (2006) Mesenchymal progenitor cells derived fromchorionic villi of human placenta for cartilage tissue engineering.Biochem. Biophys. Res. Commun. 340, 944–952

47 Kim, S.J. et al. (2007) Human placenta-derived feeders supportprolonged undifferentiated propagation of a human embryonic stemcell line, SNUhES3: comparison with human bone marrow-derivedfeeders. Stem Cells Dev. 16, 421–428

48 Miyamoto, K. et al. (2004) Human placenta feeder layers supportundifferentiated growth of primate embryonic stem cells. Stem Cells22, 433–440

49 Anson-Cartwright, L. et al. (2000) The glial cells missing-1 protein isessential for branching morphogenesis in the chorioallantoic placenta.Nature genetics 25, 311–314

50 Lebestky, T. et al. (2000) Specification of Drosophila hematopoieticlineage by conserved transcription factors. Science (New York, N.Y.)288, 146–149

51 Basyuk, E. et al. (1999) Murine Gcm1 gene is expressed in a subset ofplacental trophoblast cells. Dev. Dyn. 214, 303–311

52 Li, Y. and Behringer, R.R. (1998) Esx1 is an X-chromosome-imprintedregulator of placental development and fetal growth. Nature genetics20, 309–311

53 Murray, P. (1932) The development in vitro of the blood of the earlychick embryo. Proc. Roy. Soc. London 11, 497–521

54 Sabin, F. (1920) Studies on the origin of blood vessels and of red bloodcorpuscles as seen in the living blastoderm of chicks during the secondday of incubation.Carnegie Inst. Wash. Pub. # 272, Contrib. Embryol. 9,214

55 Chen, M.J. et al. (2009) Runx1 is required for the endothelial tohaematopoietic cell transition but not thereafter. Nature 457, 887–891

56 Zovein, A.C. et al. (2008) Fate tracing reveals the endothelial origin ofhematopoietic stem cells. Cell Stem Cell 3, 625–636

57 Eilken, H.M. et al. (2009) Continuous single-cell imaging of bloodgeneration from haemogenic endothelium. Nature 457, 896–900

58 Lancrin, C. et al. (2009) The haemangioblast generateshaematopoietic cells through a haemogenic endothelium stage.Nature 457, 892–895

59 Boisset, J.C. et al. (2010) In vivo imaging of haematopoietic cellsemerging from the mouse aortic endothelium. Nature 464, 116–120

60 Caprioli, A. et al. (2001) Hemangioblast commitment in the avianallantois: cellular and molecular aspects. Dev. Biol. 238, 64–78

61 Downs, K.M. et al. (1998) Vascularization in the murine allantoisoccurs by vasculogenesis without accompanying erythropoiesis.Development (Cambridge, England) 125, 4507–4520

62 Corbel, C. et al. (2007) Hematopoietic potential of the pre-fusionallantois. Dev. Biol. 301, 478–488

63 Zeigler, B.M. et al. (2006) The allantois and chorion, when isolatedbefore circulation or chorio-allantoic fusion, have hematopoieticpotential. Development (Cambridge, England) 133, 4183–4192

64 North, T.E. et al. (2002) Runx1 expression marks long-termrepopulating hematopoietic stem cells in the midgestation mouseembryo. Immunity 16, 661–672

65 de Bruijn, M.F. et al. (2002) Hematopoietic stem cells localize to theendothelial cell layer in the midgestation mouse aorta. Immunity 16,673–683

66 Koushik, S.V. et al. (2001) Targeted inactivation of the sodium-calciumexchanger (Ncx1) results in the lack of a heartbeat and abnormalmyofibrillar organization. Faseb. J. 15, 1209–1211

67 Lux, C.T. et al. (2008) All primitive and definitive hematopoieticprogenitor cells emerging before E10 in the mouse embryo areproducts of the yolk sac. Blood 111, 3435–3438

68 Kumaravelu, P. et al. (2002) Quantitative developmental anatomy ofdefinitive haematopoietic stem cells/long-term repopulating units(HSC/RUs): role of the aorta-gonad-mesonephros (AGM) region andthe yolk sac in colonisation of the mouse embryonic liver. Development(Cambridge, England) 129, 4891–4899

69 North, T.E. et al. (2009) Hematopoietic stem cell development isdependent on blood flow. Cell 137, 736–748

70 Dunwoodie, S.L. (2009) The role of hypoxia in development of themammalian embryo. Dev. Cell 17, 755–773

71 Reznick, D.N. et al. (2002) Independent origins and rapid evolution ofthe placenta in the fish genus Poeciliopsis. Science (New York, N.Y.)298, 1018–1020

72 Cai, Z. et al. (2000) Haploinsufficiency of AML1 affects the temporaland spatial generation of hematopoietic stem cells in the mouseembryo. Immunity 13, 423–431

73 Chamley, L.W. et al. (2001) Is interleukin-3 important inantiphospholipid antibody-mediated pregnancy failure? Fertil. Steril.76, 700–706

74 Bhatia, M. et al. (1997) Purification of primitive human hematopoieticcells capable of repopulating immune-deficient mice. Proc. Natl. Acad.Sci. U. S. A. 94, 5320–5325

75 Ditadi, A. et al. (2009) Human and murine amniotic fluid c-Kit+Lin-cells display hematopoietic activity. Blood 113, 3953–3960

76 Zambidis, E.T. et al. (2007) Emergence of human angiohematopoieticcells in normal development and from cultured embryonic stem cells.Annals of the New York Academy of Sciences 1106, 223–232

77 Graf, T. and Enver, T. (2009) Forcing cells to change lineages. Nature462, 587–594

Review Trends in Molecular Medicine Vol.16 No.8

367