Role of the placenta in the ontogeny of the human hematopoietic system. The sequential role of various intra-embryonic and extra-embryonic tissues in hematopoiesis is represented
We have updated this model of hematopoietic development (Fig. 15.1) to now include the placenta as an active site of hematopoiesis during gestation. Whether the human placenta produces HSCs early in development or, as in the mouse, peaks in its activity during mid-gestation has not yet been discerned. We are currently analyzing data to specifically answer this question and will report our findings in the near future (authors, manuscript in preparation). Nonetheless, one of the key factors governing the itinerant nature of prenatal hematopoiesis is the availability of physical space for hematopoietic activity. Early in embryogenesis, the yolk sac offers more space than exists in the embryo. The liver—the largest internal organ—serves as a hematopoietic organ until the bone marrow develops. In small mammals such as mice, the spleen acts as another major site of hematopoiesis during the perinatal period until the bones develop sufficient size to support hematopoiesis in their marrow cavity. In contrast, splenic hematopoiesis is negligible during human fetal development, presumably because of the prenatal development of the bone marrow allows for an earlier shift in hematopoiesis from the liver. The placenta, however, offers a large potential reservoir for hematopoietic activity throughout gestation. Beyond the considerable blood volume found in the placenta, the placenta also contains significant numbers of hematopoietic elements within its tissues, cells that may have a direct placental origin rather than migrating from the blood.
Many questions remain unanswered, particularity in the human system, about the ultimate reason for the placenta to possess hematopoietic potential. For instance, we do not presently know whether HSCs found in the placenta are produced in situ or they are generated at other locations and migrate to this organ at some point in gestation. Reports in the mouse point to an endogenous production of placental HSCs, since these cells were detected in placentas prior to the onset of fetal circulation and the heart beat begin [29, 30]. However, to unequivocally determine their presence, their functional status and their origin in humans is extremely difficult, particularly early in development, as the embryonic circulation begins at the fifth gestational week . The connection between developing placental vessels and the embryonic circulatory system is established by the connective stalk (the predecessor of the umbilical cord) around 6.5 gestational weeks [32, 33]. Equally enigmatic is the function of the placental HSCs, whether they are differentiating in the placenta to efficiently serve its own needs of hematopoietic cells (there are abundant macrophage-like cells, named Hoffbauer cells, present very early in gestation) while its vascular connection to the embryo is being established or if they are stored in this organ to serve as a pool of HSCs that can migrate out to intra-embryonic sites later in gestation.
What we have observed is that cells resembling primitive hematopoietic progenitors, defined by their cell surface phenotype of CD34++CD45lowCD38− are present in the earliest samples of placenta we have been able to obtain, 5 weeks of gestation . These cells are detected at a very low frequency (0.03–1.2 % of light density placental cells, n = 59) and shared many features with HSCs/progenitors isolated from intra-embryonic hematopoietic organs, such as fetal liver and bone marrow. For instance, they express fetal stem cells markers CD133, CD90, HLA-DR, CD117, and CD31 . We also demonstrated that these cells were of fetal origin capable to produce abundant myeloid colonies and some mixed myeloid-erythroid colonies in in vitro clonogenic colony-forming assays . Moreover, placental cells are capable of full long-term and multilineage engraftment in immunodeficient NOD-SCID mice, thus proving their HSC properties in vivo .
The fact that the placenta is an active hematopoietic organ was established by the presence of not only HSCs but also of more differentiated CD34+CD45+ progenitors as well as mature progeny. These lineage-committed progenitors displayed a limited hematopoietic potential in in vitro clonogenic colony-forming  assays, they were not able to reconstitute immunodeficient NOD-SCID mice  and they contained myeloid-committed (CD34+CD13+CD33+) and erythroid-committed (CD34+CD71+EpoR+) precursors . The finding of a spectrum of hematopoietic progenitors in the placenta is indicative of active hematopoiesis and differs from the profile of hematopoietic precursors found in blood, which is enriched in stem cells and contains relatively fewer committed progenitors.
One interesting observation derived from our studies is that the density of CD34++CD45lowCD38− cells changes during gestation, possibly reflecting functional changes of the cells and the hematopoietic output at different stages. The number of HSCs was determined in 59 placental cell suspensions from 5 to 40 weeks of gestation by flow cytometry . Although the total number of CD34++CD45lowCD38− cells increased over gestation as the organ increased in size, the largest numbers of cells per gram of tissue were observed during the embryonic period (up to 8 weeks). Their density in the placental tissue dramatically decreased (sevenfold) from the ninth week onward, remaining quite constant for the rest of the pregnancy (Fig. 15.2).
Changes in the hematopoietic compartment of the placenta during development. (a) Bar chart of the density of CD34++CD45low cells in the placenta expressed as cells per gram of tissue grouped by gestational age. (b) Bar chart of the total number of CD34++CD45low cells contained in the placenta by gestational age. Bar charts represent the median value of measurements made for each gestational age (n = 1–5 observations for each week, n = 59 in total). These data were previously published in a different format 
We presently do not know the physiological reasons behind the substantial change in the density of CD34++CD45lowCD38− cells during the transition from the embryonic to the fetal period. We can speculate that the enhanced hematopoietic function early in gestation may be due to the critical requirement of mature hematopoietic cells in the placenta. Placental hematopoiesis could provide, in situ, critical numbers of erythrocytes to provide O2 to the embryo and Hofbauer cells that could aid in the immunological protection of the embryo/fetus from pathogens while the fetal-placental circulation and the immune system of the embryo are being established. During the period of 5–8 weeks the embryonic liver gradually becomes hematopoietic and thus, more HSC and mature blood cells will be exported into the blood stream. These fetal liver derived hematopoietic cells will presumably be able to reach the placenta, in which case the endogenous production of HSCs might not be as crucial as earlier in development. Another possible explanation is that the placenta contributes to the early embryonic pool of HSCs and once that the intra-embryonic hematopoietic sites, such as the fetal liver, are fully functional, declines its potential accordingly. Finally we should consider the possibility that the embryonic CD34++CD45lowCD38− population contains not only hematopoietic progenitors but also could possess endothelial progenitor potential. The placenta is a highly vascularized organ with a large need of endothelial cells early in gestation. In addition, there are strong evidences suggesting the existence of a common hematopoietic and endothelial precursor, or hemangioblast, in the hemogenic endothelium during embryogenesis in mice and humans [35, 36]. Although direct evidence demonstrating the existence of hemogenic endothelium in placenta is lacking, there are indirect suggestions that it may exist. In Ncx1−/− circulation deficient mice, the emergence of HSC has been shown in the placenta in association with placenta vessels . Our own observations of placental CD34+CD45+ cells in direct contact with endothelial cells also suggest the possibility of hemogenic endothelium in the early gestation placenta . It would be very interesting to assess the endothelial potential of early CD34++CD45lowCD38− placental cells, as it is the aim of the current research in our laboratory.
Another relevant aspect of our studies was to address the question whether placental HSCs/progenitors are functionally similar to intra-embryonic stem cell populations. Although placental HSCs/progenitors are capable to produce multilineage progeny in cytokine-supplemented liquid cultures in vitro and in vivo after being transplanted into mice [18, 34], our in vitro studies point to unique features of placental hematopoietic precursors. Sorted placental CD34++CD45low cells and the more mature population constituted by CD34+CD45low cells were cultured in in vitro clonogenic assays and hematopoietic colonies of different types were counted as well as individually analyzed by flow cytometry . Most of the colonies, at any gestational age, produced only myeloid cells (i.e., colony-forming unit-granulocyte monocyte or CFU-GM) and a small number of colonies showed a mixture of lineages containing both myeloid and erythroid cells (CFU-Mix). The finding of myeloid precursors resonates with the high number of macrophage-like cells, Hofbauer cells, present in the cores of placental chorionic villi; therefore, one could interpret that the hematopoietic potential of placental HSC might be skewed to keep up the number of Hofbauer cells during this organ’s expansion and growth along gestation. One surprising observation is that neither CD34++CD45low cells nor CD34+CD45low cells produce pure burst-forming unit-erythroid (BFU-E), while their intra-embryonic counterparts, in particular the fetal liver, generate abundant BFU-E . This observation is in contrast with a recent report showing that early first trimester placenta is a site for terminal maturation of primitive erythroid cells . The lack of BFU-E activity by placental precursors might be due to the absence of critical signals from the environment that are missing in vitro. Nonetheless, the in vitro behavior of placental erythroid precursors appears to differ from those of the fetal liver or fetal BM.
The second functional difference we have detected is the strict dependence of placental HSCs/progenitors on high fetal bovine serum content in the methylcellulose semi-solid medium used in colony-forming assays (unpublished observations). While fetal liver and fetal BM progenitors grow abundantly in the absence of serum (in fact, the presence of fetal bovine serum—FBS—inhibits the number of CFUs that can be obtained ), placental progenitors generate 60 % less CFU-GM and CFU-Mix in serum free media. The ability of producing CFUs of any type is restored by adding back increasing amounts of FBS, with plateau activity observed around 30 % of FBS.
All together these data suggest different growth requirements and/or different hematopoietic potential of placental HSCs in comparison with intra-embryonic HSC and require further investigation. This aspect is particularly pressing when considering the framework for the potential clinical application of term placental HSCs to expand the number of UCB-HSC for banking and transplantation. Additionally we might take into consideration the influence of the microenvironment where these cells arise. One possible cause of the functional differences between extra-embryonic and intra-embryonic HSC might be the molecular and cellular signals that the hematopoietic niche delivers to the stem cells. Our experiments aimed to characterize the placental hematopoietic niche have led to the conclusion that there are two separate environments where placental HSCs reside: a vascular niche, in which CD34+CD45+ hematopoietic progenitors/HSCs are in direct association with CD34+CD45− endothelial cells and a mesenchymal cell niche, composed by cells expressing stromal markers on the cell surface and vimentin intracellularly (A. Bárcena, M.O. Muench, M. Kapidzic, M. Gormley, S.J. Fisher, manuscript in preparation). Therefore, the placental hematopoietic niches seem, in appearance, similar to the bone marrow niches, although in-depth studies aimed to identify the individual cellular and molecular elements, in particular regarding the mesenchymal cell niche, have not been performed. We think that the information obtained from these studies could help explaining the functional in vitro divergence of placental HSC/progenitors in comparison with other fetal and perinatal stem cell populations and we are currently working on a manuscript that reports some of these data.
3 Term Placenta Contains Hematopoietic Stem Cells/Progenitors
Our research on embryonic/fetal placenta was not only designed to address developmental questions about the emergence of the hematopoietic potential of this transient organ, but also to determine whether HSCs are contained in the term placenta. The prospective of harvesting HSCs from this new hematopoietic source is very exciting, as it could lead to important clinical applications in the field of UCB banking and transplantation. In Table 15.1 we summarize our findings from six placental cell suspensions obtained at birth of full-term neonates. The samples were stained with specific monoclonal antibodies against CD34 and CD45 antigens and subjected to flow cytometric studies. The density of CD34++CD45low cells was fairly similar between samples, ranging from 1 to 11 × 103 cells/g, while the total number of cells recovered ranged from 1 to 12.2 × 106 from the light density (mononuclear cells) fraction. In one of the specimens (sample D) we obtained the corresponding UCB and the light density fraction was also analyzed by flow cytometry. We observed that 75 mL of UCB contained a total of 1.5 × 106 CD34++CD45low cells, while the placenta, weighting 590 g contained 12.2 × 106 CD34++CD45low cells. Although it can be argued that this placenta sample was singularly enriched in CD34++CD45low cells, the yields of CD34++CD45low cells that we regularly observe in UCB and in placenta are very much comparable to those reported in the literature. In a study of 300 UCB samples that were similarly processed and analyzed by FACS, the mean number of total CD34+CD45+ cells was 3.1 × 106 cells per UCB unit  and different groups have published similar results, 0.5–2.6 × 106 CD34+CD45+ cells/UCB unit [41, 42].
Abundant CD34++CD45low cells are present in placenta at birtha
Gestational age (wks)
C-section (C) labor (L)
# CD34++ CD45low (cells/g)
Total # of CD34++ CD45low cells
1.1 × 104
7.2 × 106
1.4 × 103
9.8 × 105
3.0 × 103
1.6 × 106
2.1 × 104
12.2 × 106
1.5 × 103
1.0 × 106
3.2 × 104
8.5 × 106
An example of flow cytometric analysis of light density cells freshly isolated from 50 g of term placental tissue is shown in Fig. 15.3. Although the frequency of CD34++CD45low cells is very low (less than 1 % of the light density fraction), the placenta is quite a large organ and typically weighs 500 g on average. Therefore it is, in principle, possible to purify several millions of CD34++CD45low cells from a single placenta. We investigated their functional status in vivo, by transplanting them intravenously into sub-lethally irradiated immunodeficient NSG mice. Figure 15.4 shows the analysis of a representative mouse 4 months after transplantation of term placental cell suspension (38 weeks of gestational age) enriched on HSCs/progenitors by lineage magnetic depletion. The detection of human mature hematopoietic cells in the mouse bone marrow belonging to multiple lineages (erythroid, myeloid, and lymphoid) proves the long-term engraftment (after 10 weeks post-transplantation) capability of the transplanted cells, which is currently considered the golden standard to demonstrate functional human HSCs. Other researchers in the field have reported similar results, which is highly relevant since they used different methods to dissociate the placental tissue than the one we employ  or, by perfusion of the placenta and in the absence of tissue dissociation . We believe that this is an important consideration: besides the method by which the cells are harvested from term placenta, different groups of investigators have been able to observe their hematopoietic potential. This finding suggests that a side-by-side comparison of different methods for harvesting placental HSCs, with an emphasis on those techniques that could be applicable in the clinical setting, and that allow for an improved yield of cells, is an important step towards the utilization of this organ as a source of transplantable HSCs.
Term placenta contains CD34 ++ CD45 low cells. Flow cytometry analyses of freshly isolated light density fractions from a placenta (39 weeks and 2 days of gestational age), the blood that drips from the placenta without perfusion (placenta wash) and the corresponding UCB. In the upper panels we show our gating strategy, where a gate was set to exclude the abundant Hofbauer cells (CD14+) and stromal cells (CD10+) present in the placenta. The box indicates the presence of CD34++CD45low cells in the different tissues
Hematopoietic reconstitution by placental chorionic villi cells. Chorionic villi cells, harvested at term—38 weeks’ gestation, were transplanted into an immunodeficient NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJ (NSG) mouse and analyzed 4 months later. Bone marrow engraftment by human cells is shown in the top row and staining of the bone marrow of an untransplanted mouse is shown in the bottom row. Total bone marrow cells were analyzed using gates that define live, single cells based on lack of propidium iodide staining and the height and width properties of the forward light-scatter signal, respectively (not shown). The expression of human myeloid, erythroid, and B-lymphoid markers is shown together with the human common leukocyte marker, CD45. Numbers associated with each quadrant represent the percentages of live single cells found in the respective quadrant
We also investigated the anatomical localization or CD34+CD45+ cells in term placenta by immunolocalization studies. The results of these experiments could aid in designing future methods to improve their isolation. In Fig. 15.5 we show a representative experiment of immunostaining and fluorescent microscopy, where several CD34+CD45+ cells can be readily identified in fixed tissue sections of term placenta. The sensitivity of this technique does not allow us to distinguish between cells expressing variable levels of antigens, such as CD34++ and CD34+ cells, as can be done by flow cytometry. Hence we cannot make a strict correlation between the data obtained by FACS versus the immunolocalization results. Nevertheless, the paucity of CD34+CD45+ cells detected by immunostaining is very much in line with the low frequency data obtained by FACS. Hematopoietic progenitors and HSCs are a sparse populations that in term placenta are mostly associated with endothelial cells (CD34+CD45− cells)  lining the abundant vessels that vascularize the chorionic villi. Their contact with endothelium suggests that purification methods based on placental perfusion might be ideal to mobilize the majority of these cells, as discussed in the next section.