|Home | About | Journals | Submit | Contact Us | Français|
Placental trophoblasts are key determinants of in utero development. Mouse trophoblast stem cells (mTSCs), which were first derived over a decade ago, are a powerful cell culture model for studying their self-renewal or differentiation. Our attempts to isolate an equivalent population from the trophectoderm of human blastocysts generated colonies that quickly differentiated in vitro. This finding suggested that the human placenta has another progenitor niche. Here we show that the chorion is one such site. Initially, we immunolocalized pluripotency factors and trophoblast fate determinants in the early-gestation placenta, amnion and chorion. Immunoreactive cells were numerous in the chorion. We isolated these cells and plated them in medium containing FGF and an inhibitor of activin/nodal signaling, which is required for human embryonic SC self-renewal. Colonies of polarized cells with a limited lifespan emerged. Trypsin dissociation yielded continuously self-replicating monolayers. Colonies and monolayers formed the two major human trophoblast lineages—multinucleate syncytiotrophoblasts and invasive cytotrophoblasts (CTBs). Transcriptional profiling experiments revealed the factors associated with the self-renewal or differentiation of human chorionic trophoblast progenitor cells (TBPCs). They included imprinted genes, NR2F1/2, HMGA2 and adhesion molecules that were required for TBPC differentiation. Together, the results of these experiments suggested that the chorion is one source of epithelial CTB progenitors. These findings explain why CTBs of fully formed chorionic villi have a modest mitotic index and identify the chorionic mesoderm as a niche for TBPCs that support placental growth.
Much remains to be learned about the morphological and molecular features of trophoblast (TB) allocation during the early stages of human development. In mouse and human embryos, morphological changes appear following compaction at the morula stage. The outer trophectoderm (TE) layer becomes polarized with tight/adherens junctions and apical microvilli, while the inner cell mass (ICM) contains nonpolarized cells [1, 2]. In mice, many factors with roles in ICM and TE segregation are known (reviewed in [3, 4]). ICM specification involves lineage-restricted transcription factors that include Oct4. Evidence from human embryonic (hE) and induced pluripotent stem cells (SCs) confirms the importance of these factors in establishing pluripotency. In mice, acquisition of a TB fate requires the sequential actions of GATA family members , the TE-specific Drosophila caudal-related transcription factor Cdx2 , and Eomes . Conversely, embryos lacking geminin contain only TE, which undergoes premature endoreduplication . Later, other molecules have roles in TB differentiation. For example, the transcription factor Glial Cells Missing 1 (GCM1) governs formation of the chorionic villi . Due to the lack of experimental models, much less is known about regulators of the early stages of human TB differentiation. Some mechanisms may overlap parallel processes in the mouse (reviewed in ). These include a role for syncytin, an endogenous retroviral gene that encodes a fusogenic protein [11, 12]. But there are also notable differences. For example, human TBs express HLA-G, a nonclassical class I human major histocompatibility complex molecule , and specialized hormones (e.g., chorionic gonadotropin).
In contrast, much has been published regarding the later stages of CTB differentiation . Key morphological aspects are diagrammed in Fig. S1. At a molecular level, the process whereby CTBs withdraw from the cell cycle and fuse to form multi-nucleated syncytiotrophoblasts (STBs) involves the coordinated expression of a number of factors (reviewed by [14–16]). These include molecules with relevant functions, e.g., human fusogenic endogenous retroviruses. Other factors with potential roles in this process that are expressed in the human placenta include ADAM12 , a disintegrin and metalloproteinase. Caspase 8, which could contribute to cytoskeletal remodeling required for fusion, may also be involved . There is strong evidence that other classes of molecules such as connexins (Cx) are critical to STB formation . Placental protein 13, a galectin family member, also plays a role . Hormones—including estradiol, glucocorticoids and hCG—promote STB generation (reviewed in ). Interestingly, CD98, a putative amino acid transporter, has been implicated in this process [21–23]. Finally, physiological factors are key; for example, rising oxygen levels promote fusion .
An interesting spectrum of factors regulate CTB differentiation along the invasive pathway. CTB transformation of the uterine vasculature depends on their ability to execute a unique epithelial-to-endothelial transition. In this regard, analysis of CTB invasion in situ and in vitro has established that this switch is a vital component of placental development [25–29]. Initially, CTB progenitors in chorionic villi express E-cadherin (epithelial cadherin) and α6β4 integrin. Upon differentiation, the cells downregulate these molecules and upregulate VE-cadherin, α5β1, αV family members, PECAM-1, and VCAM-1  as well as matrix metalloproteinase (MMP)-9 [30, 31]. Recent data suggest that CTB differentiation/invasion also entails a switch from a venous to an arterial phenotype in terms of the cells’ expression of Eph and ephrin bi-directional signaling molecules . This change is accompanied by the modulation of several growth factors and receptors (e.g., VEGF and angiopoietin [Ang] family members) that function during conventional vasculogenesis and angiogenesis . Oxygen tension also plays an important role favoring CTB proliferation rather than differentiation/invasion [33–35]. In this context, it is not surprising that HIFs are important regulators of CTB differentiation .
Stem and progenitor cell (PC) culture models are powerful tools for analyzing developmental processes in vitro, with added value for human studies because in vivo experimentation is not possible. Pluripotent embryonic stem cells (ESCs) have been isolated in many species including mice and humans . Likewise, lineage-restricted TB stem cells have been established from nonhuman primates and other animals (reviewed in [21, 38, 39]). Although multipotent stem cells have been isolated from human placenta and extraembryonic membranes [40, 41], their ability to give rise to TB stem/progenitor cells has not been explored. The lack of TB stem cell lines from our own species is a critical gap. In their absence, investigators employ BMP-4-treated hESCs with TB-like properties , TBs derived from human embryoid bodies , primary cytotrophoblasts (CTBs; [44, 45]), chorionic villous explants , virally transformed TB lines [47, 48], telomerase-immortalized TBs , or choriocarcinoma cell lines . Here we describe the results of experiments in which we identified the human chorion as a source of continuously self-replicating TBPCs, which we subsequently isolated and characterized. These experiments revealed novel aspects of the core circuitry that enables the self-renewal or differentiation of these cells with important implications regarding mechanisms of placental growth.
The UCSF Committee on Human Research approved this study. Informed consent was obtained from donors who were undergoing elective pregnancy terminations for collection of placentas and associated membranes. See Supplemental Materials and Methods for procedures used to isolate TBPCs.
Colonies from 2 chorionic membranes (73 and 86 weeks of gestation; passages 3 and 4) were incubated (1–2 min) in 1 ml of 0.05% trypsin (Gibco) at 37°C. The reaction was stopped by adding 10 ml of DMEM F12/10% FBS. The cells were isolated by centrifugation for 10 min at 1000 xg and resuspended in DMEM/F12 supplemented with 10 ng/ml basic FGF (R&D Systems), 10% FBS, 10 µM SB431542 (Tocris Bioscience), 1% penicillin/streptomycin and 0.1% gentamycin. The cells were cultured (250,000/6-well plate) on gelatin (Sigma) –coated wells in 20% oxygen. Each line was banked at various intervals (passages 5–17).
Colonies were manually dissected and small clumps were cultured on the surfaces of 50 µl drops of polymerized undiluted Matrigel in differentiation medium: Knockout Serum Replacement Medium (Gibco) supplemented with 10 ng/ml FGF4 (Sigma) and 1% of an antibiotic-antimycotic mixture to which 40 ng/ml EGF (Invitrogen) was added as previously described . The same method was used to differentiate TBPC monolayers except that the cells were cultured on Matrigel (BD Biosciences) -coated coverslips rather than drops.
CTBs were isolated from pooled placental samples and cultured on Matrigel-coated coverslips or 6 well-plates as described . Human embryonic stem cells (H7 line) were cultured in StemPro medium (Invitrogen) supplemented with 8 ng/ml bFGF on Matrigel (3.3%) -coated 6-well plates. They were passaged by disaggregating with Acutase (Millipore), which produced a single cell suspension. Placental fibroblasts were isolated and cultured as described previously .
See SI Materials and Methods.
The experiments were performed using published methods .
Medium was collected from TBPCs that were grown for 72 h at a density of 7–10 colonies/well of a 24-well plate (n=6). TBPCs (100,000 cells; n=9) grown in monolayer form were transferred in 3 ml of differentiation medium to Matrigel-coated 6-well plates. The cultures were terminated (24, 48 or 72 h), the cells were counted, and the release of intact (total) hCG was quantified by using an ELISA kit (Abazyme LLC).
See SI Materials and Methods.
The hESCs, TBPC colonies, and TBPC monolayers were cultured as described above. Freshly isolated primary (second trimester) CTBs were allowed to attach for 45 min, then washed three times with PBS. Either two (TBPC monolayers) or three (hESCs, TBPC colonies, CTBs) independent samples were analyzed. hESCs and TBPCs were cultured for 24 and 36 h, respectively. Freshly isolated CTBs were cultured overnight. Total RNA was purified using a PicoPure RNA isolation kit (Arcturus). Samples were stored at -80°C prior to the start of the experiment. The microarray, data analysis, and RT-qPCR methods we used have been published [54, 55].
Data were calculated as the mean ± SD. The significance was evaluated using a paired Student’s t test and p values ≤ 0.05 were considered statistically significant.
Initially we immunolocalized markers of pluripotency or TB fate specification in tissue sections of villi or chorion. The results showed that cells with the phenotype we sought were more numerous in the chorion. Accordingly, we focused our attention on this region. At 7 weeks this membrane consists of a surface CTB bilayer and an underlying mesenchymal compartment (Fig. 1A). Using cytokeratin 7 (CK7) expression to identify CTBs, we showed that they expressed Oct4 in a nuclear pattern as did mesenchymal components (Fig. 1B). In both locations, subsets of these cells also reacted with antibodies that recognized GCM1 (Fig. 1C) and Eomes (Fig. 1D). Syncytin immunoreactivity was primarily associated with the CTB layer (Fig. S2A) and GATA4 expression was more prominent in nuclei within the mesenchyme (Fig. S2B). These findings suggested that the first trimester chorion could be a source of TBPCs.
Chorionic membranes (Fig. 1E) were stripped of villi (Fig. 1F), and the component cells were released by a step-wise series of enzymatic digestions. The initial digests, which contained primarily differentiated CTBs, were discarded in favor of the later fractions, which we reasoned included less mature progenitors. The resulting cells were resuspended in DMEM/F12 medium containing 10% FBS, FGF2 and the TGF-β/activin/nodal signaling inhibitor, SB431542. Our experimental strategy exploited the observation that FGF is required for mTSC self-renewal  and SB431542, a chemical inhibitor of TGF-β/activin/nodal signaling, has profound effects on hESCs. Specifically, SB431542 downregulates hESC expression of markers of pluripotency  and triggers TB differentiation . In this context, we reasoned that inhibition of TGF β/activin/nodal signaling would act on the placental stem cell population allowing the emergence of progenitors restricted to the trophoblast lineage. The cell suspensions were plated on gelatin substrates and cultured under standard conditions. In the absence of the chemical inhibitor, they grew as a monolayer of mixed cell types (Fig. 1G) that expressed vimentin (Fig. S2C) together with beta3 tubulin (Fig. S2D) and/or smooth muscle actin (Fig. S2E). In the presence of the inhibitor, multiple colonies with an epithelial morphology emerged within 7–10 days (Fig. 1H and Fig. S3A). To date, we have established colonies from 13 individual placentas from 6–11 weeks of gestation.
The cells that comprised the colonies expressed Oct4 as well as ZO-1, a marker of apical-basal polarity (Fig. 1I). They also stained for cytokeratin (Fig. S3B), GATA4 (Fig. S3C), and nestin (Fig. S3D), a marker of hESC neuronal differentiation, but failed to express TB stage-specific antigens (e.g., GCM1, hCG). To assess their potential, we plated them on Matrigel in differentiation medium. After 72 h, the cells continued to express CK7 (Fig. S3E). Fusion was observed, particularly among the cells that migrated furthest from the aggregates. Multi-nucleated cells tended to have the highest reactivity with anti-hCG (Fig. S3F), which was also released into the medium (Fig. S3G). Those that migrated from the original attachment site stained for GCM1 (Fig. S3H) and HLA-G (Fig. S3I). When colony-derived cells were plated on Matrigel-coated transwell filters, CK7-positive cells traversed the pores and reached the undersides (Figs. S3J, K). They also upregulated the expression of integrins α1 (Fig. S3L) and α5 (Fig. S3M), which accompanies CTB invasion in vivo . When the cells were grown under conditions that promote embryoid body formation, they aggregated but the components continued to express TB markers (e.g., CK; Fig. S3N). Together, the results of these experiments were evidence that the colonies contained chorion-derived TBPCs. However, the cells, which had a normal karyotype, gradually lost Oct4 expression and differentiated after 8–10 passages. This finding suggested that the TBPC colonies might be heterogeneous and/or that polarity might be associated with a limited lifespan.
Accordingly, we used trypsin to dissociate the colonies into single cell suspensions, which disrupted polarity and enabled differential adhesion, a method we and others have used to purify primary cells . As shown in Fig. 2A, the resulting cells grew as dispersed monolayers rather than as colonies. Thus far, we have established 2 cell lines from colonies that were derived from chorion samples obtained at 73 and 86 weeks of gestation. They have been cultured continuously without differentiating for up to 25 passages (P) and banked at 5P intervals after which they were recoverable with minimal loss.
Marker analysis of the undifferentiated monolayers failed to detect staining for Oct4, nestin, or ZO-1 a molecule required for apical-basal polarity. The cells retained nuclear localization of GATA4 (Fig. S4A) and continued to express CK7 (Fig. 2B). They also exhibited reactivity with antibodies against Eomes (Fig. 2C), geminin (Fig. 2D), GATA3 (Fig. S4B) and GCM1 (Fig. S4C). They stained for other human TB antigens such as the neonatal FCγ receptor (Fig. S4D) and integrin α6 (Fig. S4E). Together these results suggested that trypsinization enabled isolation of cell lines with characteristics of of trophoblast progenitors (TBPCs) that can be propagated in vitro.
To test whether TBPCs can give rise to the mature human trophoblast subpopulations, we plated the cells on Matrigel in differentiation medium. By 72 h the cells formed aggregates (Fig. 2E) as do primary CTBs when they are cultured on this matrix (Fig. S4F ). The aggregates were comprised of mononuclear and multinuclear cells (Fig. 2F). In additional experiments, we compared the antigen repertoire of the two cell types. The staining patterns of primary CTBs are shown in Figs. S4G-K. The differentiated TBPCs expressed CK7 (Fig. 2G), geminin (Fig. 2H) and GCM1 (Fig. 2I); Eomes immunoreactivity became cytoplasmic (Fig. S4L). They upregulated staining for markers of differentiated CTBs including HLA-G (Fig. 2J) and hormones such as human placental lactogen (hPL; Fig. 2K), and hCG (Fig. 2L), which between 48 and 72 h of culture was secreted into the medium (2424 ± 591 pg/300,000 cells). In addition, the aggregates stained for syncytin (Fig. 2M). Immunoblotting confirmed that expression of the latter antigen was significantly upregulated as compared to undifferentiated TBPCs (Fig. 2N). To explore the cells’ invasive potential, they were plated in the same medium on Matrigel-coated transwell filters. The number of cells that traversed the pores was ~8 fold higher than the number of primary CTBs that served as positive controls with placental fibroblasts as negative controls (Fig. 2O). During invasion, they switched their integrin profile (α6 to α1) as do invasive CTBs (Fig. S4M; ). Together these results suggested that the monolayer TBPCs could differentiate into syncytiotrophoblasts (STBs) and CTBs.
To better understand the molecular characteristics of the TBPC colonies and monolayers, we used a microarray approach. The most highly differentially expressed genes are shown in Fig. 3 and the entire data set is included as Fig. S5. As compared to the H7 hESC line, the colonies highly upregulated mesenchymal-type extracellular matrix components, PAPPA, NR2F1 and 2 (Coup TFI and II), cadherin 11, HGF, IL13RA2, imprinted genes (PEG3 and H19), EPAS1 (HIF2α and RARRES1 (Fig. 3A). Downregulated genes included neurotensin and pluripotency factors (HESRG, LEFTY1, NANOG, OCT4 and SOX2). Overall, the TBPC colonies and monolayers had similar expression patterns with a few notable differences (Fig. 3B). Genes that were dramatically downregulated in the monolayers included PEG3, MIB1 (which regulates Notch signaling; ), WASF2 (which plays a role in regulating cytoskeletal rearrangement; ), chromosome maintenance factors (SMC3 and -5, SMARCC1) and JAK1. SSTR1 was the most highly upregulated gene along with a detoxifier, EPHX1. Also upregulated were a LIM domain protein that is involved in cardiac development (PDLIM7; ) and ADAMTSL1. Finally, we compared the gene expression patterns of the monolayers to primary CTBs isolated from second trimester placentas (Fig. 3C). In TBPC monolayers, claudin 11 was the most highly upregulated gene along with HMGA2, an architectural transcription factor , N-cadherin (CDH2), another imprinted gene (IPW), SSTR1 and F2RL2 (PAR3). The downregulated genes included an amelioride binding protein (ABP1), chorionic somatomammotropin (CSH1 and 2) and its ligand (CSHL1), GCM1, IGFBP1, PEG3, PAPPA2, LAIR2 and CCL4. Ingenuity pathway analysis highlighted the importance in TBPC monolayers of genes that are involved in neuronal development and stem cell pluripotency as well as integrin and thrombin signaling (Fig. S6). In comparison, primary CTBs upregulated the PPAR and growth hormone signaling pathways among others. Together these results give us interesting new insights into the factors that could play a role in human TB differentiation, which included molecules involved in mouse TB differentiation as well as novel regulators.
The microarray data were validated by using antibody-based and RT-qPCR approaches. We were particularly interested in verifying the expression pattern of HMGA2. In the chorion, an antibody that specifically recognizes this molecule stained the nuclei of cells in the mesenchymal and CTB layers (Fig. 4A). Immunoblot analysis showed that a band of the expected molecular mass (12 kDa), which was detected in lysates of TBPC monolayers and their differentiated progeny, was absent in primary CTBs (Fig. 4B). Analysis of cultured TBPCs revealed nuclear expression (Fig. 4C), which in some cells became cytoplasmic during the initial stages of differentiation (Fig. 4D). Thus, the expression pattern of HMGA2 was consistent with the microarray data. RT-qPCR also confirmed the expression patterns of CLDN11, CDH11, CSH1, H19, NR2F1, NR2F2, PEG3, and WASF2.
Our previous work showed that adhesion molecules play seminal roles in CTB differentiation . Therefore, we were particularly interested in their expression patterns and possible functions. We focused on N-cadherin and integrin α4, which were upregulated at the RNA level in TBPC monolayers as compared to CTBs (Figs. 3 and S7). In the chorion, anti-N-cadherin stained the mesenchymal layer including a subset of Eomes-positive cells, and CTBs (Fig. 4E). Integrin α4 immunoreactivity was solely associated with cells of the mesenchymal layer that expressed Eomes (Fig. 4F). To explore possible functions, TBPCs were plated on Matrigel-coated transwell filters in differentiation medium ± function blocking anti-N-cadherin or anti-integrin α4. Inclusion of either antibody blocked aggregate formation (by 60 ± 5.1% for N-cadherin and by 77 ± 6.7% for integrin α4; p ≤ 0.001). Furthermore, the cells, which were no longer invasive, failed to express stage-specific antigens that are indicative of TB differentiation (hCG, HLA-G, hPL, and integrin α1). Addition of the isotype control antibody (anti-TRA-1-60) had no effect.
Together, these data give us novel insights into the mechanisms that drive human placental development and a new experimental system for studying the early steps of human TB differentiation. Immunolocalization experiments showed that the CTB layers of the chorion stained for pluripotency factors and TB fate determinants/stage-specific antigens. However, we also found that cells of the chorionic mesenchyme co-stained for the same molecules. To assess their developmental potential, we developed a method for isolating the latter cells and culture conditions that supported their propagation in vitro. Upon differentiation they formed the TB cell types of the mature placenta. Taken together, our results identify the chorion as a TBPC niche, a finding that is consistent with the fact that villi sprout from this membrane. We also found phenotypically similar cells in the mesenchymal cores of chorionic villi suggesting another location where TBPCs reside or through which they transit. This theory is bolstered by our recent success in isolating TBPCs from second trimester villous mesenchyme as well as chorion (manuscript in preparation).
These findings help us to better understand other data from our group and other investigators. For example, we have repeatedly tried to derive TB stem cells from the trophectoderm of blastocyst-stage human embryos. Occasionally we obtained colonies, but they inevitably differentiated without generating lines of self-replicating cells. Possible explanations include our lack of understanding of their growth requirements. Alternatively, this population might be relatively short-lived giving way to the precursors that we isolated from the chorion. Our data also suggest that chorionic TBPCs have the potential to form STBs and invasive CTBs. This might explain why the placenta continues to grow in the absence of CTB proliferation after the first trimester . Finally, the villous mesenchyme, where the organ’s intrinsic vasculature arises, is a niche for hematopoietic SCs that form the major blood cell lineages . We recently discovered that the chorion also contains these cells (manuscript in preparation). How development of one progenitor population might influence formation of the other is an open question that we are actively investigating as these types of mutually beneficial associations have been observed previously, e.g., endothelial and neuronal stem cells . In this regard, it is interesting to speculate that the specialized placental niche could play a role in programming human TBs, which upon differentiation along the invasive pathway acquire many vascular-like characteristics .
At a molecular level, the derivation conditions that we employed suggested that TBPC self-renewal involves FGF pathways, as in the mouse, with the additional requirement of blocking activin/nodal signaling. Transcriptional profiling studies identified numerous molecules that could play roles in specifying TBPC identity and/or enabling their self-renewal, a portion of which have been implicated in placental development and/or function. In this regard, the antigenic phenotype of the colonies, which lacked expression of TB markers, suggested that they were a more primitive population than the monolayers, which expressed several TB stage-specific antigens. In this context, TBPC colonies, as compared to hESCs, expressed numerous transcription factors and other molecules with relevant functions in embryonic or extraembryonic development. For example, the colonies were distinguished by their differential expression of Coup TF transcription factors, which were recently implicated in CTB differentiation , and HIF 2α, which plays oxygen-dependent and independent roles in regulating mouse and human TB differentiation . Their gene expression patterns also highlighted the functions of other molecules, such as hepatocyte growth factor, (which plays important roles in human placentation ), cadherin 11 (a mediator of cell-cell interactions), and FOXF1 (which regulates the human growth hormone variant gene that is exclusively expressed by the placenta; ). Upregulation of the imprinted genes, PEG3 and H19, is additional evidence that this phenomenon plays an important role in human placentation (reviewed in ). Although we detected staining for Oct4, this molecule and other pluripotency factors were expressed at much lower levels than in hESCs. Unexpectedly, neurotensin was the most highly downregulated gene, possible evidence of the neuroepithelial-like properties of these cells. Thus, these data gave us insights into both hESCs and TBPCs. The expression patterns of the TBPC colonies vs. monolayers suggested an important role for PEG3 in colony maintenance and somatostatin in continued self-renewal of the monolayers, which may also entail inhibiting Notch signals as suggested by the upregulation of MIB1, a ubiquitin ligase that targets ligands in this pathway for proteosomal degradation .
Relative to CTBs, TBPCs expressed much higher levels of HMGA2, which promotes mouse SC self-renewal and is downregulated during differentiation. The mechanisms include let7-mediated repression in concert with the enhanced expression of p16 (reviewed in ). Our data suggest that this molecule plays a similar role in the human TB lineage. Other upregulated molecules included F2RL2, IL13RA2, N-cadherin, and another imprinted gene, IPW. We also detected relatively lower expression in TBPCs of molecules that play important roles in TBs or are expressed upon their differentiation, e.g., CCL4 and CSH (placental lactogen), respectively [73, 74]. For example, CTB differentiation was associated with upregulated GCM1 expression. In mice, deletion of this transcription factor revealed its role in formation of chorionic villi . It is interesting to note that integrin α4, which is expressed in cells of the chorion, is a transcriptional target of this factor . The importance of this relationship is highlighted by the fact that integrin α4 function is required for differentiation of TBPCs. However, we failed to detect expression of Cdx2 and Hand1, which regulate earlier steps in TB differentiation in this species (reviewed in ). The downregulated expression of CSH suggested a role for this molecule in CTB formation. Finally, PEG3 was again upregulated in CTBs, one of only a handful of genes with a biphasic expression pattern.
Beyond molecular insights into the mechanisms that govern human TB specification and self-renewal, this new culture system will have other interesting applications. For example, the TBPC model enables molecular and functional analyses of the early stages of human TB differentiation–the period between allocation of the trophectoderm and formation of the CTB progenitors that populate the chorionic villi. Very little is known about this facet of human placental development, which is key to the growth of this organ. Conversely, we want to use this approach to gain a better understanding of the origins of pregnancy complications that entail abnormal TB differentiation. For example, preeclampsia (e.g., maternal hypertension and proteinuria with vascular dysfunction as the common denominator) is associated with shallow CTB invasion and consequently, poor placental perfusion. Some investigators attribute the anatomical findings to fundamental defects in CTB differentiation . Isolation of TBPCs from affected pregnancies will enable tests of this hypothesis. Additionally, it is also possible that TBPCs could be closer to an ES-like state, particularly at an epigenetic level, than adult cells. Accordingly, this population may be much easier to reprogram for use as research tools and for therapeutic applications. Overall, the derivation of TBPCs that can differentiate into STBs and invasive CTBs is a significant addition to the cell-based tools that can be used to study many aspects of human embryonic and placental development.
Diagram of the histological organization of the human maternal-fetal interface at midgestation. In this location, cytotrophoblasts (CTBs), which are specialized (fetal) epithelial cells of the placenta, differentiate and invade the uterine wall, where they also breach maternal blood vessels. The basic structural unit of the placenta is the chorionic villus, composed of a stromal core with blood vessels, surrounded by a basement membrane and overlain by cytotrophoblast progenitor cells. As part of their differentiation program, these cells detach from the basement membrane and adopt one of two lineage fates. They either fuse to form the syncytiotrophoblasts that cover floating villi or join a column of extravillous cytotrophoblasts at the tips of anchoring villi. The syncytial covering of floating villi mediates the nutrient, gas, and waste exchange between fetal and maternal blood. The anchoring villi, through the attachment of cytotrophoblast cell columns, establish physical connections between the fetus and the mother. Invasive cytotrophoblasts penetrate the uterine wall up to the first third of the myometrium. A portion of the extravillous cytotrophoblasts home to uterine spiral arterioles and remodel these vessels by destroying the muscular wall and replacing the endothelial lining. To a lesser extent, they also remodel uterine veins. At term, few villous cytotrophoblast progenitor cells remain, the syncytiotrophoblast layer thins, and the stromal cores expand. VC, villus core. (Diagram modified from ).
Antigenic profile of the chorion and cells isolated from this membrane. Components of the chorion stained for (A) syncytin and (B) GATA4. When chorion-derived cells were cultured in the presence of FGF2 and in the absence of SB431542, they grew as a monoloyer of mixed cell types that expressed (C) vimentin together with (D) β3 tubulin and/or (E) smooth muscle actin.
(A-D) Phenotypic characterization of TBPC colonies and (E-N) their differentiated progeny. (A) The colonies had an epithelial morphology. (B) They expressed cytokeratin 7 (CK7), (C) GATA4, and (D) nestin. (E) Upon differentiation, they immunostained for CK7 and (F) hCG, (G) which they secreted into the medium. (H) They upregulated expression of GCM1, and (I) HLA-G. (J) When they were plated on Matrigel-coated transwell inserts, CK-positive cells reached the undersides of the filters. (K) Quantitation of the results showed that their invasion levels were approximately equal to primary cytotrophoblasts (CTBs) and much greater than placental fibroblasts (Pl. Fibs.). Upon invasion they also upregulated the expression of (L) integrinα1 and (M) integrin α5. (N) When the cells were cultured under conditions that support formation of embryoid bodies, they continued to express CK7. Panels E and F show the same cells co-stained for CK7 and hCG respectively. Scale bars: L, 2 µm; A-K, M, N, 10 µm.
(A-E) Phenotypic characterization of TBPC monolayers, (F-K) primary cytotrophoblasts (CTBs) and (L, M) the differentiated progeny of TBPC monolayer cultures. TBPC monolayers expressed (A) GATA4, (B) GATA3, (C) GCM1, (D) FCγ receptor and (E) integrin α6. (F) In culture, primary CTBs aggregated and reacted with antibodies that recognized (G) CK7, (H) geminin, (I) GCM1, (J) hPL and (K) hCG. (L) Upon differentiation of TBPC monolayer cultures Eomes immunoreactivity became cytoplasmic. During invasion, they upregulated expression of (M) integrin α1. Panels J and K show the same aggregate co-stained for hPL and hCG respectively. Scale bars, 10 µm.
Global gene expression profiling of TBPC colonies and monolayers. Included are genes that were differentially expressed by ≥ 2-fold. (A) Comparison of human embryonic stem cells (hESCs) vs. TBPC colonies (TBPCc). (B) TBPCm vs. primary cytotrophoblasts (CTB). Blue, decreased gene expression; red, increased gene expression.
Ingenuity pathway analysis highlighted the mechanisms that are involved in TBPC self-renewal and CTB differentiation.
Global gene expression profiling revealed the integrin adhesion receptor repertoire of human embryonic stem cells (hESCs, H7 line), TBPC colonies (TBPCc), TBPC monolayers (TBPCm), and primary cytotrophoblasts (CTB). Blue, decreased gene expression; red, increased gene expression.
This work was supported by CIRM grant RC1-00113 and NIH grant U54HD055764. We thank the members of the Fisher group for stimulating discussions, helpful suggestions and technical support.
Author contributionsOlga Genbacev: Conception and design, Performance of experiments, Collection and/or assembly of data, Data analysis and interpretation
Matthew Donne: Performance of experiments
Mirhan Kapidzic: Provision of study material or patients
Matthew Gormley: Data analysis and interpretation
Julie Lamb: Performance of experiments
Jacqueline Gilmore: Performance of experiments
Nicholas Larocque: Performance of experiments
Gabriel Goldfien: Provision of study material or patients
Tamara Zdravkovic: Performance of experiments
Michael McMaster: Conception and design, Collection and/or assembly of data, Data analysis and interpretation, Manuscript writing, Final approval of manuscript
Susan Fisher: Conception and design, Collection and/or assembly of data, Data analysis and interpretation, Manuscript writing, Final approval of manuscript.
The authors declare no conflicts of interest.