Late stage definitive endodermal differentiation can be defined by Daf1 expression

Background Definitive endoderm (DE) gives rise to the respiratory apparatus and digestive tract. Sox17 and Cxcr4 are useful markers of the DE. Previously, we identified a novel DE marker, Decay accelerating factor 1(Daf1/CD55), by identifying DE specific genes from the expression profile of DE derived from mouse embryonic stem cells (ESCs) by microarray analysis, and in situ hybridization of early embryos. Daf1 is expressed in a subpopulation of E-cadherin + Cxcr4+ DE cells. The characteristics of the Daf1-expressing cells during DE differentiation has not been examined. Results In this report, we utilized the ESC differentiation system to examine the characteristics of Daf1-expressing DE cells. We found that Daf1 expression could discriminate late DE from early DE. Early DE cells are Daf1-negative (DE-) and late DE cells are Daf1-positive (DE+). We also found that Daf1+ late DE cells show low proliferative and low cell matrix adhesive characteristics. Furthermore, the purified SOX17low early DE cells gave rise to Daf1+ Sox17high late DE cells. Conclusion Daf1-expressing late definitive endoderm proliferates slowly and show low adhesive capacity. Electronic supplementary material The online version of this article (doi:10.1186/s12861-016-0120-2) contains supplementary material, which is available to authorized users.


Background
The definitive endoderm (DE) gives rise to the gastrointestinal and digestive system. In the mouse embryo, the DE progenitors reside at the posterior region of the epiblast that derived from the inner cell mass [1]. During gastrulation, as the cells ingress through the primitive streak, the epiblast segregates into the three germ layers that form the somatic cell lineages of the ectoderm, mesoderm, and definitive endoderm (DE). DE arises from the Forkhead box A2 (Foxa2)-expressing anterior primitive streak (APS) [2][3][4][5] and is then regionalized into the fore-, mid-, and hindgut [6].
The DE is identified by the expression of SRY (sex-determining region Y)-box 17 (Sox17) [7][8][9] and chemokine (C-X-C motif ) receptor 4 (Cxcr4) [10][11][12]. Sox17 mutant mouse embryos have a reduced DE, apoptosis of the foregut, and abnormal morphogenesis of the mid-and hindgut [9]. Sox17 is also required for the assembly of the basement membrane, as the Sox17 mutant embryo fails to segregate the DE from the mesoderm [13]. Activin is a frequently used inducer for DE differentiation from pluripotent stem cells, embryonic stem cells (ESCs), and induced pluripotent stem cells (iPSCs) [14][15][16]. When SOX17 is overexpressed, human ESCs spontaneously differentiate into the DE, independent of Activin [17]. In zebrafish embryos, cxcr4a regulates directional migration [11,18] and DE proliferation during gastrulation [19]. In chick embryos, Cxcr4 is expressed in the DE and angioblasts. Cxcr4 and its ligand Cxcl12 form a reciprocal signaling loop that triggers angioblast migration to the pancreatic endoderm and induces pancreatic development. Inhibition of Cxcr4 suppresses angioblast migration into the pancreatic endoderm region. As a result, the size of the pancreas decreases [10]. Although Cxcr4 is also expressed in the mesodermal cells, it is often used in combination with E-cadherin for purifying ESC-derived DE cells using flow cytometry [12].
Daf1 is an inhibitor of complementary activation [20]. Daf1 is expressed by immune cells and DE-derived tissue, such as intestine and airway [21]. Using microarray analysis and in situ hybridization, we previously identified Daf1 as a DE cell surface marker based on its expression in ESC-derived and embryonic DE. Daf1 is also expressed in pancreatic progenitor cells [22,23]. However, the role of Daf1 in the DE is not well understood. In this study, we found that the DE population that expresses Daf1 (Daf1 + DE) has slow cell cycling and low cell-matrix adhesive characteristics. Furthermore, Daf1-negative DE cells (Daf1-DE) turn out to be Foxa2 + Sox17 low cells and Daf1-positive DE (Daf1 + DE) cells to be Foxa2 + SOX17 high cells. Our results therefore suggest that E-cadherin + Cxcr4 + DE is composed of two populations: Sox17 low early DE and Sox17 high late DE. Sox17 high late DE cells were positive for Daf1, and were slow proliferative and low cell-matrix adhesive cells.

Daf1 + DE are slowly proliferating cells
Previously, we reported Daf1 as a surface marker, expressed in a subpopulation of DE [23]. DE are defined as E-cadherin+/Cxcr4+ cells [12]. When cultured in Activincontaining medium [24,25], ESCs sequentially give rise to APS cells on day4 (defined as E-cadherin+/Pdgfra + cells), then to DE cells (defined as E-cadherin+/Cxcr4+ cells) on day 5 (Fig. 1). A plot of our previous microarray analysis results of the APS and DE cells [23,24] shows the time dependent expression of Foxa2, Sox17 and Daf1 (Fig. 1a). Foxa2 was highly expressed in the APS and DE. Sox17 was lowly expressed in the APS and highly expressed in the DE. Daf1 expression was absent in the APS and present in the DE (Fig. 1a). We then analyzed Daf1+ cells for the expression of Nanog, a pluripotent marker, using a mouse Nanog-iPS cell line bearing a green fluorescent protein (GFP) reporter driven under Nanog promoter. Daf1-positive cells turned out to be Nanog/GFP negative (Additional file 1). Taken together, Daf1 marks a subpopulation of DE and is first expressed in the DE, but not earlier in the APS.
To identify the differences between Daf1 + DE and Daf1-DE cells, we purified Daf1+/-DE cells and compared their properties. Real time PCR analysis of the sorted Daf1+/-DE cells confirmed Daf1 expression is enriched in Daf1 + DE (Fig. 1b, c). Expressions of an M phase marker, phosphorylated histone H3 (pH3) or Proliferating Cell Nuclear Antigen (PCNA) that marks proliferating cells at every phase of the cell except G0 were enriched in Daf1-DE than in Daf1 + DE cells, revealed by western blot analysis (Fig. 1d). We analyzed the cell cycle phases and found that Daf1 + DE cells showed a longer Go/G1 phase and shorter S, M/G2 phase compared to Daf1-DE cells (Fig. 1e, Additional file 2). This suggested that cell proliferation was decreased in Daf1 + DE.

Daf1 + DE are low adhesive cells
Next, we performed flow cytometry using antibody against Daf1and plated at a same cell number of the sorted DE cells (Daf1-DE or Daf1 + DE) onto matrigel-pre-coated dishes (Fig. 2a). After 24 h culture, the Daf1+ DE gave rise to fewer cell number compared to that of Daf1-DE (Fig. 2b,  c). The results suggested a higher level of cell death and/ or lower plating efficiency of Daf1 + DE cells. We then examined cell-matrix adhesion and the extent of apoptosis, in which cells were allowed to adhere to the plate for 90 mins [26,27] [28]. Our results revealed that Daf1 + DE cells showed a lower matrix adhesion ratio than Daf1-DE (Fig. 2d), and that Daf1 + DE cells were more apoptotic and showed a higher Cleaved Caspase3/7+ ratio (Fig. 2e) than Daf1-DE. Cells bind to extracellular matrix (ECM) through the ECM receptor, integrin [29,30]. We found that expression of several subtypes of the Integrin genes, such as Itgα1, Itgα3, and Itgα8, was down-regulated in Daf1 + DE cells (Fig. 2f). Our results suggest that Daf1 + DE cells have lower adhesion capacity than Daf1-DE cells, which led to a lower plating efficiency and triggered cell death in the attached cells, compared to Daf1-DE cells.

Daf1 is a marker of late stage DE
To characterize the adherent cells onto Matrigel-precoated dish, we assayed for the expression of the DE markers, Foxa2 and Sox17, immediately after the adhesion assay (Fig. 3). Interestingly, whereas both Daf1+/-DE populations showed similar capacity for giving rise to Foxa2expressing cells, Daf1 + DE cells showed a higher capacity for giving rise to Sox17-expressing cells (Fig. 3a, b). Foxa2 is expressed in both the APS and DE, whereas Sox17 is expressed at a higher level in the DE. We then confirmed Sox17 protein expression of the sorted Daf1+/-DE cells directly (without plating) by immunocytochemical analysis (Fig. 3c). Sox17 expression was found in a higher proportion and seemingly higher level in the sorted Daf1 + DE cells compared to Daf1-DE cells (Fig. 3c). Western blot analysis confirmed a higher level of Sox17 expression in the Daf1 + DE cells (Fig. 3d) (Fig. 3f).

Daf1+/-DE cells can give rise to pancreatic and intestinal fates
We identified that Daf1-DE cells acquired Daf1 expression and turned into Daf1 + DE. If Daf1-DE represents early DE and Daf1 + DE represents late DE, both DE cells could give rise to regional endoderm derivatives of the gut. We next examined the abilities of the Daf1+/-DE cells to differentiate into pancreatic or intestinal cells. To visualize pancreatic differentiation, we used Pdx1/GFP mouse ESCs, in which GFP expression is driven by the Pdx1 promoter [25]. We isolated Daf1+/-DE cells by flow cytometry (purity > 98 %), seeded the cells on MEF feeders, and cultured in the presence of Activin and FGF2 (Fig. 4a) [25]. We found that both Daf1-DE and Daf1+ cells gave rise to Pdx1/GFP+ cells, after plated for 3 days (day 8). The Daf1 + DE-derived cells yielded less Pdx1/ GFP cells compared to that of the Daf1-DE cells (Fig. 4b,  c). This discrepancy was probably due to a lower proliferation of the Daf1 + DE cells compared to Daf1-DE cells.
Since the pancreas derives from the anterior DE, we then asked if the Daf1+/-DE cells could also differentiate into the posterior DE derivative of the intestine. DE derived from R1 ESCs were sorted and seeded onto MEF feeders and cultured at the presence of BIO and DAPT [16,31]. After five days, both Daf1-DE and Daf1 + DE cells differentiated into Cdx2+ intestinal cells. Similar to pancreatic differentiation, Daf1 + DE cells are less proliferative and gave rise to small Cdx2+ colonies than Daf1-DE cells (Fig. 4c). Our results therefore suggest that both Daf1+/-DEs are capable of differentiating into the endodermal derivatives of the pancreas and intestine.

Discussion
Previously, we identified that Daf1 is expressed in the DE. Here, we identified Daf1 as a late DE marker. Daf1 is an inhibitor of complement activation [20]. Daf1 is expressed in the kidney, spleen, testis, intestine, and bronchi of the adult mouse [21]. Daf1 deficiency is reported in autoimmune hemolytic anemia patients [32]. In Daf1 knockout mice, IFN-γ expression increases, resulting in enhanced T cell response autoimmunity [33]. However, gastrointestinal-tract develops normally in Daf1 knockout mice. Here, we examined the detailed expression patterns during DE differentiation using ESCs.
DE cells are defined as E-cadherin+/Cxcr4+ cells [12]. However, both E-cadherin and Cxcr4 are also expressed in the primitive streak [34,35] colonies, due to their less proliferative and low adhesive characteristics than that of Daf1-DE cells (Fig. 5). A slight decrease in S, M/G2 phase and increase in G0/G1 phase in Daf1 + DE cells might reflect their property as more differentiated cells. Daf1 + DE cells seem to differentiate as efficiently into Pdx1-expressing cells, but not as efficiently into Cdx2-expressing cells, compared to Daf1-DE cells. This might due to a partial loss in differentiation potency of Daf1+ DE cells into the intestinal fate. We previously reported that regional-specific endodermal fates are determined sequentially in the order of stomach, intestine and pancreas, in the chick embryos [36]. It is possible that Daf1 + DE gradually lose potency to differentiate into intestinal lineages, but retains differentiation potency into pancreatic lineages, compared to that of Daf1-DE. Cell-matrix adhesion is also necessary for cell differentiation [37,38]. Integrin expression promotes DE differentiation from human pluripotent stem cells [39]. Integrin is a receptor of extracellular matrix expressed in the cells, which enables binding of the cells to the extracellular matrix. Itgα5 and ItgαV are DE-specific Integrins. Knockdown of either Itgα5 or ItgαV inhibits DE differentiation [39]. Both Daf1-and Daf1 + DE cells express Itgα5 and ItgαV. We found that the expressions of Itgα1, Itgα3, and Itgα8 decreased in Daf1 + DE cells. Itgα1 is an attachment molecule of the DE [40]. Itgα3 is expressed in the DE and expression decreases in the Foxa2 null mouse embryo [41]. Itgα8 null mice have abnormal lung morphogenesis [42]. These Daf1-DE specific integrins could regulate DE differentiation and modulate their behavior. The integrins are known to show distinct ligand binding specificities among the superfamily members. Itgα1β1, Itgα3β1 bind specifically to laminin and Itgα8β1 binds specifically to fibronectin [43]. The lowered expression of Itgα1, Itgα3 might explain the decreased adhesion of Daf1 + DE cells to Matrigel, which composed mainly of laminin. The decrease in adhesion to matrix might reflect the developmental transition from early to late DE.  We found that Daf1-DE could turn into Daf1 + DE cells. Sox17 expression was higher in Daf1 + DE than in Daf1-DE cells. Furthermore, expression of a primitive streak marker, Brachyury [44], was higher in Daf1-DE cells (SO unpublished). Therefore, E-cadherin+/Cxcr4+ DE could be a mixed population of both primitive streak and DE cells.

Conclusions
Our data indicate that DE can be divided into two stages: early and late DEs. Early DE consists of E-cadherin + Cxcr4 + Daf1-Foxa2 + Sox17 low cells that show higher proliferative activity and higher cell-matrix adhesive capacity. Later on, these DE cells differentiate into E-cadherin + Cxcr4 + Daf1+ Foxa2 + Sox17 high late DE cells that show a decreased proliferation and low cell-matrix adhesion capacity.
Our findings would contribute to the understandings of the differentiation of the primitive streak and DE during gastrulation.

Microarray analysis
Our previously described results were used in the present study [22,23].

Flow cytometry analysis
Cells were dissociated with Cell Dissociation Buffer (Invitrogen) and stained with the appropriate antibodies. The stained cells were recovered using FACS Aria II (BD Biosciences). Data were recorded using the BD FACS Diva Software program (BD Biosciences) and analyzed using the FlowJo program (Tree Star).

Western blot analysis
Cells were homogenized in SDS sample buffer (62.5 mM Tris-HCl, 10 % glycerol, 2 % SDS, pH 6.8). After centrifugation, the supernatants were collected and used as total protein extracts. Total proteins were subjected to 8 % SDS-PAGE and transferred to polyvinylidene difluoride membranes (Immobilon-P Transfer Membrane, Millipore). The membranes were incubated with antibodies listed above. Horseradish peroxidase (HRP) conjugated antirabbit IgG (1:20000, CST) was used as the secondary antibody. Chemiluminescence signals were detected with ECL Plus Western Blotting Detection Reagents (GE Healthcare, Japan).

RT-PCR analysis
RNA was extracted from the cells using the RNeasy Micro-Kit (QIAGEN) and then 1 μg of RNA was reverse transcribed using ReverTra Ace (TOYOBO), ribonuclease inhibitor, recombinant (TOYOBO), and Oligo dT primers (TOYOBO). Primer sequences are shown in Additional file 3.

Cell cycle analysis
Cells were dissociated with Cell Dissociation Buffer (Gibco). Dissociated cells were washed with PBS and treated with Vybrant DyeCycle Violet Stain (Life Technologies) for 30 min at 37°C. Cells were analyzed by FACS Canto (BD Biosciences).

Cell-matrix adhesion analysis
The sorted cells were plated onto matrigel-precoated dishes with serum free medium for 90 min. The attached cells were fixed with 4 % PFA for 5 min, then stained with DAPI (1:2000, Roche). Cell counts were performed as previously described [46].

Additional files
Additional file 1: Daf1-positive cells are negative for Nanog expression. Mouse Nanog-iPS cells, in which GFP expression is driven by Nanog promoter [45], are differentiated into DE. Cxcr4+/E-cadherin + cells were sorted and analyzed for Daf1 and Nanog-GFP expression. Daf1