A Unique Interplay Between Rap1 and E-Cadherin in the Endocytic Pathway Regulates Self-Renewal of Human Embryonic Stem Cells
Key Words. Rap1 • E-cadherin • Human embryonic stem cells • Endocytic pathway • Self-renewal • Differentiation
ABSTRACT
Regulatory mechanisms pertaining to the self-renewal of stem cells remain incompletely understood. Here, we show that functional interactions between small GTPase Rap1 and the adhesion molecule E-cadherin uniquely regulate the self-renewal of human embryonic stem cells (hESCs). Inhibition of Rap1 suppresses colony formation and self-renewal of hESCs, whereas overexpression of Rap1 augments hESC clonogenicity. Rap1 does not directly influence the expression of the pluripotency genes Oct4 and Nanog. Instead, it affects the endocytic recycling pathway involved in the formation and mainte- nance of E-cadherin-mediated cell–cell cohesion, which is essential for the colony formation and self-renewal of hESCs. Conversely, distinct from epithelial cells, disrup- tion of E-cadherin mediated cell–cell adhesions induces lysosome delivery and degradation of Rap1. This in turn leads to a further downregulation of E-cadherin function and a subsequent reduction in hESC clonogenic capacity. These findings provide the first demonstration that the interplay between Rap1 and E-cadherin along the endo- cytic recycling pathway serves as a timely and efficient mechanism to regulate hESC self-renewal. Given the availability of specific activators for Rap1, this work pro- vides a new perspective to enable better maintenance of human pluripotent stem cells.
Stem cells offer a unique renewable resource for cell replace- ment therapies [1]. However, the mechanisms underlying the molecular regulators that couple extracellular signals to vari- ous intracellular responses to coordinate self-renewal of stem cell are not fully understood [2–4]. Of many pivotal regulators, Rap1 seems to fulfill the evolutionarily conserved function that enables cells to coordinate their responses to microenvir- onmental changes [5–7]. Rap1 is a member of the Ras family of small G proteins [8] and was originally discovered in the 1980s as a Ras homolog and as a suppressor of Ras induced signaling [8, 9]. Rap1 can be activated by various extracellular stimuli, which induces the conversion of the inactive GDP- bound form into the active GTP-bound form. Rap1 activation allows the recruitment of a variety of intracellular effectors, resulting in a host of distinctive cellular functions [6, 10].
Recent studies suggest that Rap1 plays a pivotal role in the formation and maintenance of cadherin-mediated cell–cell junctions [11–16]. The initial evidence came from a genetic study in Drosophila melanogaster in 2002 [11]. This study demonstrated that Rap1 regulates morphogenetic processes through the control of Drosophila Epithelial-cadherin (DE- cadherin) repositioning for proper adherens junction formation and distribution following cytokinesis [11]. Consistent with this original finding, additional lines of evidence suggest that Rap1 plays a major role in the inside-out [12, 13] and out- side-in [14] regulation of E-cadherin functions in human adult epithelial cells.
E-cadherin is a classic member of the cadherin family and was originally identified as a cell surface glycoprotein respon- sible for Ca2+-dependent homophilic cell–cell adhesion during morula compaction in the preimplantation mouse embryo and during chick embryo development [17, 18]. Three decades since its discovery, the molecular mechanisms whereby E-cadherin regulates various cellular processes and functions remain incompletely understood [19–21]. Studies during the past 10 years have begun to reveal the importance of intracel- lular trafficking as a means of regulating the functions of E-cadherin [20, 21] and the possible role of Rap1 in the pro- cess of E-cadherin trafficking [12–14, 21]. However, the interplay between Rap1 and E-cadherin in stem cells and the impact of the two molecules on cell functions remain unclear. This is further complicated by the well-established cell-type specificity of many Rap1 functions [15, 16, 22–25]. Indeed, polarized epithelial cells [14], transformed epithelial cells [12], carcinoma cells [13, 26], and other cells may not share identical mechanisms for regulating cell–cell junction forma- tion, maintenance, and disassembly [5, 26–28].
In stem cells, the role and regulatory mechanism of Rap1 in self-renewal and differentiation are not known. A few lines of evidence suggest that Rap1 is essential during early embryo development [29–31]. For instance, loss-of-function mutations in Rap1 resulted in Drosophila embryonic lethality [29–31], whereas knockout of C3G (a Rap1 exchange factor) caused early embryonic lethality in mice due to aberrant epi- thelial cell adhesion in developing tissues [30].
Notably, the outcomes of Rap1 loss-of-function mutations in embryos [29, 30] are similar to those of E-cadherin knock- out [32, 33]. This suggests a functional association between these two molecules in early embryologic development and raises the possibility that the interplay between Rap1 and E-cadherin may influence stem cell functions.
Using human embryonic stem cells (hESCs) as a model, we demonstrate here that a functional crosstalk between Rap1 and E-cadherin along the endocytic recycling pathway uniquely and rapidly regulates hESC self-renewal. Rap1 impacts hESC clonogenicity via regulation of E-cadherin- mediated cell–cell adhesions. As opposed to epithelial cells, hESCs lack a negative feedback-like mechanism between Rap1 and E-cadherin. On disassembly of adherens junctions, Rap1 is delivered to lysosomes and degraded, which further hampers hESC clonogenicity. This work provides new insights into the molecular mechanisms governing the self- renewal of hESCs.
Culture of hESCs and Assessment of Oct4, Nanog, and AP
H1 and H9 hESC lines were obtained from Wicell and the CA1 cell line was a gift from Dr. Nagy (Mount Sinai Hospi- tal). All cell lines were approved for use by the local ethics board and the Stem Cell Oversight Committee of the Cana- dian Institutes of Health Research. Undifferentiated hESCs were maintained under feeder-free conditions and Oct4 and Nanog were examined using specific antibodies. Alkaline phosphatase (AP) activity was determined as described previ- ously [34–36] (Supporting Information Methods).
Clonogenic and Self-Renewal Assays of hESCs
Single hESCs were dissociated with collagenase IV for 5–10 minutes, followed by 4 mM ethylenediaminetetraacetic acid (EDTA)-PBS or phosphate-buffered saline (PBS)-based cell dissociation buffer (Invitrogen, Carlsbad, CA, http://www.in- vitrogen.com) and passed through a 40-lm cell strainer BD (Becton, Dickinson and Company, Franklin Lakes, NJ, http// www.bd.com). Dissociated single cells were seeded on Matri- gel and cultured in mouse embryonic fibroblast-conditioned medium (MEFCM) supplemented with 12 ng/ml basic fibro- blast growth factor (bFGF). To evaluate clonogenic capacity, hESC colony numbers were monitored every 2–3 days after seeding. After 11 days, growing colonies were dissociated, and single cells were replated in the same conditions at low dilution (1,500 cells per cm2 or 500 cells per well of 96-well plate), and colonies were counted 3 days later to assess the self-renewal capacity of individual hESC clones [37].
Rap1 Activation Assays
Rap1 activation was examined using a well-established pull- down method based on the specific binding of a GST fusion protein containing the Rap-binding domain of RalGDS (RalGDS-RBD/GST) to the active GTP-bound form of Rap1 [14]. Experiments were performed according to the manufac- ture’s protocol using the Active Rap1 Pull-Down and Detec- tion kit (Pierce, Rockford, IL, http://www.piercenet.com; Cat. 89872) (Supporting Information Methods).
Plasmids and Transfection
The plasmid encoding Rap1a-green fluorescent protein (GFP) was from Dr. S. F. Retta (University of Torino, Italy). The Yellow Fluorescent Protein (YFP)-tagged E-cadherin con- struct was a kind gift from Dr. Yap (the University of Queensland, Australia). DsRed-Rap1a was subcloned by digesting the Rap1a segment from pGFP-Rap1a with BamH1/ XhoI and inserting it into the pDsRed2-C1 vector (Clontech, Palo Alto, CA, http://www.clontech.com). hESCs were trans- fected with 1 lg of plasmids using Lipofectamine2000 (Invi- trogen) at approximately 50% confluence as described elsewhere [38] (Supporting Information Methods).
Short Interfering RNA Targeted Rap1 and E-Cadherin Knockdown
hESCs were seeded in 24-well plates and transfected with validated short interfering RNA (siRNA) with a specific target sequence for human Rap1a, E-cadherin, or with nontargeting control oligonucleotides (Ambion, Austin, TX, http:// www.ambion.com). The target and nontarget sequences are described in the Supporting Information Methods. hESCs were transfected using Lipofectamine2000 with 2.5 pmol of Rap1a or E-cadherin siRNA. One day after siRNA knock- down, cells were subjected to RNA isolation for real time quantitative PCR (Q-PCR) or to confocal microscopy. Two days after transfection, the cells were dissociated and reseeded in the feeder-free culture condition to examine clonogenic capacity.
Inhibition of Rap1 Endocytosis and Recycling Inhibition of E-cadherin endocytosis and recycling were per- formed as described previously [14]. Briefly, hESCs were pre- incubated at 37◦C in MEFCM containing either 1 lM Cyto- chalasin D (CytD; Sigma, St. Louis, MD, http:// www.sigmaaldrich.com/) for 30 minutes, 1 lM Bafilomycin A1 (BAF; Sigma), or 1 mM N-ethylmaleimide (NEM; Sigma) for 90 minutes, followed by [ethylenebis(oxyethylenenitrilo)]- tetraacetic acid (EGTA) or EDTA treatment (4 mM) at 37◦C for 30 minutes.
Real Time Q-PCR
Q-PCR was performed as described previously [34] (Support- ing Information Methods).
Western Blot
hESCs were lysed with Nonidet P40 buffer. Whole cell lysate supernatants were separated by electrophoresis on 12% SDS- polyacrylamide gel and transferred to a polyvinylidine difluo- ride membrane (Bio-Rad, Hercules, CA, http://www.bio
-rad.com). Antibodies used for Western blotting were rabbit anti-human Rap1 (1:1,000; Pierce), mouse anti-E-cadherin (1:1000; clone HECD-1; Abcam, Cambridge, U.K., http:// www.abcam.com), mouse anti-human a-tubulin (1:5,000; Bio- Rad), and anti-rabbit or mouse IgG-HRP (1:2,000; Bio-Rad). The detailed procedures are described in the Supporting Infor- mation Methods.
Immunofluorescence Staining and Fluorescent Labeling
Immunofluorescence staining and LysoTracker fluorescent labeling were conducted as described previously [34, 35] and in the Supporting Information Methods.
Fluorescence microscopy, Image Analysis, and Time- Lapse Live-Cell Imaging
Fluorescence images were obtained using either an Olympus FluoView FV1000 laser-scanning confocal microscope (Olympus, Tokyo, http://www.olympus-global.com) or a Leica AF6000 microscope system equipped with a deconvolution module (Leica Microsystems, Heidelberg, Germany, http:// www.leica-microsystems.com). Olympuse FluoView, Leica LAS AF6000 and Volocity 4.1 (Improvision) softwares were used for image processing, deconvolution, and quantitative imaging analyses wherever appropriate. Confocal images acquired under the identical exposure time and instrument set- tings among different groups were used for colocalization and quantitative fluorescence intensity analyses. Time-lapse live- cell imaging is described in the Supporting Information Methods.
Statistical Analysis
Results are expressed as mean 6 Standard Deviation (SD). Statistical significance was determined using a Student’s t test, ANOVA, or v2 wherever appropriate. Results were con- sidered significant when p < .05. Suppression of Rap1 Hampers hESC Self-Renewal We began by examining the Rap1 protein expression and its subcellular localization in hESCs to assess the function of Rap1 in regulating the self-renewal of hESCs. Using a well- established pull-down method [14], we identified the expres- sion of both total Rap1 and active Rap1 (GTP-bound Rap1) in undifferentiated hESCs (Fig. 1A). These undifferentiated hESCs were characterized by the expression of Nanog, Oct4, SSEA4, Tra-1-60, Tra-1-81, and AP, and the capacity to form teratoma (Fig. 1B and data not shown). Consistent with previ- ous findings in epithelial cells [13, 14], we observed that Rap1 in undifferentiated hESCs was localized to the plasma membrane and the perinuclear region (Fig. 1C, GFP-Rap1- transfected hESCs; Supporting Information Fig. 1, immunocy- tochemical staining). The presence of Rap1 protein in hESCs prompted us to explore whether Rap1 plays an important role in the self- renewal of hESCs. We suppressed Rap1 function by treating hESCs with GGTI-298. GGTI-298 specifically inhibits Rap1 but not Ras and has been well-characterized and used in numerous studies of Rap1 signaling [39–43]. Treatment of hESCs with GGTI-298 led to a significant reduction in Rap1 protein levels (Fig. 1D), and resulted in a decrease in the number of undifferentiated colonies as indicated by AP activ- ity (Fig. 1E). AP activity was closely associated with the expression of Nanog and Oct4 (Fig. 1B), and has been used as a sensitive and reliable marker for undifferentiated hESCs [44–46]. Collectively, these results suggest that Rap1 is required for the maintenance of hESCs in an undifferentiated state. Using a novel clonogenic assay based on a primary and secondary plating strategy [37, 44, 47], we further examined the role of Rap1 in the self-renewal of hESCs. In the assay, hESCs treated with the Rap1 inhibitor exhibited a marked decrease in clonogenic capacity and self-renewal potential (Fig. 2A) and an increase in cell deaths (unpublished observa- tions). In contrast, vehicle-treated hESCs formed typical colo- nies and maintained self-renewal capacity (Fig. 2A). These results demonstrate that Rap1 plays an important role in hESC self-renewal. Rap1 Knockdown and Overexpression Further Con- firm the Crucial Role of Rap1 in hESC Self-Renewal To validate the findings from GGTI-298 Rap1 inhibitor, we an- alyzed the effects of Rap1 siRNA knockdown on hESC self- renewal. Rap1 exists in two closely related isoforms, Rap1a and Rap1b, which differ by only a few residues [7]. As most studies of Rap1 function in epithelial cells focus on Rap1a [10, 12–14, 48], we used Rap1a (Rap1a siRNA oligonucleotides and Rap1a overexpression) throughout our experiments. In contrast to the scrambled siRNA control, siRNA-mediated knockdown of Rap1 resulted in a significant downregulation of Rap1 mRNA (>80% reduction at 24 hours; Fig. 2B) and of Rap1 protein (>5-fold reduction at 48 hours; Fig. 2C). A clon- ogenic assay performed in parallel revealed that Rap1a knock- down resulted in a sixfold reduction in secondary colony numbers as compared with the scrambled control (Fig. 2D), suggesting a reduced capacity for self-renewal. Moreover, cells in the Rap1a knockdown group exhibited a spindle-like mor- phology and were negative for AP staining (Fig. 2D), indica- tive of a loss of undifferentiated characteristics. Thus, the out- comes of siRNA knockdown experiments confirm that Rap1 plays an essential role in hESC self-renewal.
To test whether elevated levels of Rap1 could enhance hESC self-renewal, we performed clonogenic assays after transfection of hESCs with Rap1a. As shown in Figure 2E, hESCs overexpressing Rap1a exhibited a twofold increase in the total number of colonies compared to control cells (Fig. 2E). These results provide further evidence in support of the important role Rap1 plays in hESC self-renewal, prompting us to investigate the underlying mechanism.
Rap1 Influences hESC Clonogenicity via E-Cadherin Nanog and Oct4 are pluripotency genes used for characteriz- ing undifferentiated hESCs [49]. Transient downregulation of Nanog has been shown to predispose ESCs toward differentia- tion [50], whereas overexpression of these genes can reprogram adult cells into induced pluripotent stem cells [51, 52], indicating their pivotal roles in ESC self-renewal and pluripo- tency. Therefore, we examined whether Rap1 directly impacts these two genes in hESCs. As determined by Q-PCR, siRNA- mediated silencing of Rap1 for 24–48 hours did not suppress Nanog and Oct4 gene expressions (Fig. 2B) despite its marked inhibition of the clonogenic capacity of hESCs (Fig. 2D). This finding suggests that the observed suppressive effect of Rap1 on hESC self-renewal may not require a primary downregulation of Oct4 and Nanog pluripotency genes.
Previous studies on epithelial cells have demonstrated that Rap1 regulates E-cadherin-mediated (cell–cell) adhesions in- dependent of integrin-mediated (cell–matrix) adhesions [12– 14]. Current studies on hESCs suggest that E-cadherin is expressed in undifferentiated hESCs, whereas integrins are expressed in both differentiated and undifferentiated hESCs [49, 53–55]. We have consistently observed that E-cadherin expression was restricted to undifferentiated hESCs that coex- pressed Oct4 (Fig. 3A). We speculated that Rap1 may influ- ence hESC clonogenic capacity by affecting E-cadherin func- tion. To test this possibility, we examined E-cadherin expression during clonogenic assays in the absence or pres- ence of the Rap1 inhibitor GGTI-298. In the absence of the Rap1 inhibitor, the total number of adherent cells and the expression levels of E-cadherin were similar between test and control groups at 24 hours after reseeding (Fig. 3B and data not shown). However, exposing hESCs to the Rap1 inhibitor resulted in a marked reduction in E-cadherin protein levels (Fig. 3C) and hESC clonogenic capacity (Fig. 3D), suggesting that Rap1 may influence hESC self-renewal by affecting E-cadherin protein levels.
Rap1 Regulates hESC Self-Renewal by Coordinating the Spatio-Temporal Turnover of E-Cadherin
The functional interaction between Rap1 and E-cadherin could occur at the transcriptional and/or posttranscriptional levels [19–21, 56]. To determine whether the interplay between Rap1 and E-cadherin influences the gene expression of one another, we silenced either Rap1 or E-cadherin expres- sion using siRNA and examined the reciprocal effect on the mRNA expression by Q-PCR. The Rap1 siRNA-mediated knockdown did not result in a significant change in steady- state levels of E-cadherin mRNA and vice versa (Fig. 3E), suggesting that the functional interaction between Rap1 and E-cadherin in hESCs does not occur at the transcriptional level. We reasoned that Rap1 might affect E-cadherin protein turnover by regulating E-cadherin’s delivery to the plasma membrane, as observed in epithelial cells [12–14]. To test this possibility, we cotransfected hESCs with YFP-E-cadherin and DsRed-Rap1a, and performed confocal imaging. We observed that newly synthesized Rap1 and E-cadherin appeared in the cytoplasmic regions as early as 5 hours, showing limited spa- tial overlap (Fig. 4A). However, more than 16 hours, both proteins were recruited and overlapped at cell–cell contact sites (Fig. 4B, 4C), suggesting that Rap1 does not join E-cad- herin until it reaches regions close to the plasma membrane. In particular, Rap1 was highly accumulated and overlaid with E-cadherin at zipper-like structures of the nascent cell–cell contact sites (Fig. 4D, top panel on the right). Conversely, Rap1 was not detected at membrane regions lacking E-cad- herin-mediated cell–cell junctions (Fig. 4D, bottom panel on the right). These observations suggest that Rap1 plays a role in the formation of E-cadherin mediated cell–cell junctions in hESCs.
To determine whether Rap1 was crucial for the aforementioned process, we silenced Rap1 with siRNA 16 hours post cotransfection of hESCs with YFP-E-cadherin and DsRed- Rap1a, followed by confocal microscopy examination. In cells treated with the scrambled control siRNA, E-cadherin and Rap1 predominantly localized and overlapped at cell–cell boun- daries (Fig. 4E). In contrast, Rap1 knockdown resulted in a marked increase in the cytoplasmic pool of E-cadherin in 82 6 4% of cotransfected cells (Fig. 4E) and a significant reduction in pixel fluorescence intensity of E-cadherin at the plasma membrane (Fig. 4F). Furthermore, quantitative imaging analysis revealed that the ratio of total cytoplasmic to membrane E-cad- herin fluorescence in hESCs was increased fivefold in Rap1 knockdown versus control group (Fig. 4G, 4H). These results suggest that Rap1 is required for the proper delivery of E-cad- herin to the nascent cell–cell contact sites in hESCs.
To address whether spatio-temporal delivery of E-cadherin to the plasma membrane influences the self-renewal of hESCs, we examined E-cadherin fate after hESC dissociation and its impact on hESC clonogenicity. Upon dissociation, E-cadherin was rapidly internalized into cytoplasmic regions (Fig. 5A, 5B, nontransfected hESCs), in a pattern similar to that observed in Rap1 knockdown cells (Fig. 4E–4H, hESCs transfected with YFP-E-cadherin and DsRed-Rap1a). Notably, approximately 98–99% of single hESCs that failed to relocal- ize cytoplasmic E-cadherin to the plasma membrane eventu- ally lost both E-cadherin and AP expression, resulting in increased cell deaths (Fig. 5C, 5D). In contrast, timely rees- tablishment of E-cadherin-mediated cell–cell contacts among neighboring cells within 2 hours after reseeding (i.e., relocalization to cell membrane) allowed hESCs to develop into typi- cal hESC colonies with positive AP expression (Fig. 5C, 5E). These results suggest that failure to reestablish E-cadherin mediated cell–cell contacts within a short time window results in a reduction in E-cadherin protein levels and hESC clonoge- nicity (Fig. 5C–5E). Consistent with the above findings, siRNA knockdown of E-cadherin significantly downregulated apoptotic inhibitory gene Bcl-XL and upregulated pro-apopto- tic gene Caspase-3 (Fig. 5F), suggesting an important role of E-cadherin in hESC survival.
Taken together with our earlier observations (Fig. 3B– 3E), these results suggest that Rap1 regulates hESC self- renewal by coordinating the spatio-temporal turnover of E- cadherin.
A Unique Positive Feedback-Like Interplay Between Rap1 and E-Cadherin Takes Place in the Endocytic Pathway upon hESC Dissociation
A previous study in epithelial cells showed that disruption of E-cadherin mediated cell–cell cohesion did not influence total Rap1 expression [14]. Instead, it activated Rap1, which in turn facilitated the reestablishment of E-cadherin mediated cell–cell cohesion after reseeding, showing a negative feed- back reaction [14]. Strikingly, using the same experimental procedure previously applied for epithelial cells [14], we found that, in contrast to epithelial cells, both total Rap1 and active GTP-Rap1 were markedly reduced upon hESC dissociation into single cells (Fig. 6A, 6B). In particular, the loss of Rap1 occurred in a timely manner during cell dissociation (Fig. 6A, 6B) and was unrelated to cell death since cell viability remained the same in each group, as determined by flow cytometry and trypan blue exclusion (>85%). These results suggest that hESCs do not have a negative feedback reaction to maintain or upregulate Rap1 after cell dissociation as do epithelial cells [14]. Instead, maintenance of total Rap1 in hESCs is associated with the persistence of adherens junctions.
To examine whether endocytosis and endocytic trafficking contribute to the loss of Rap1 upon hESC dissociation, we inhibited these processes using three agents that act at defined points along endocytosis and endosome trafficking pathways. We first disrupted the actin cytoskeleton [14, 57] by pretreat- ment of hESCs with Cytochalasin D (CytD) before the disrup- tion of intercellular junctions with the calcium chelating agent EGTA. In epithelial cells, this treatment has been shown to inhibit E-cadherin endocytosis despite the loss of cell–cell adhesion [14]. In hESCs, CytD pretreatment led to the persist- ence of Rap1 after exposure to EGTA (Fig. 6C, 6D), suggest- ing that endocytosis is required for the diminution of Rap1 levels. Next, to determine whether the reduction of Rap1 is associated with vesicular traffic events, we blocked endosome acidification and trafficking with the vacuolar H(+)-ATPase proton pump inhibitor Bafilomycin A1 [58]. Exposure of hESCs to Bafilomycin A1 before EGTA treatment prevented Rap1 loss and sustained colony integrity (Fig. 6C, 6D), imply- ing that vesicular trafficking is also required for the downreg- ulation of Rap1 upon disruption of intercellular junctions in hESCs. Consistently, using NEM, a known agent blocking vesicular fusion events [59, 60], we obtained similar results (Fig. 6C, 6D). Taken together, these results suggest that the downregulation of Rap1 triggered by hESC dissociation requires endocytosis and endosome trafficking.
Disruption of hESC Adherens Junctions Causes Rap1 Delivery to Lysosomes
To determine whether the endocytosis-dependent diminution of Rap1 protein triggered by hESC dissociation is involved in Rap1 delivery to lysosomes, we transfected hESCs with GFP-Rap1 and then treated the cells with the calcium chelating agent EGTA in the presence of LysoTracker (a fluo- rescent probe for lysosomes; Fig. 7A, 7B). By quantitative colocalization analyses, we found that colocalization of GFP- Rap1 with LysoTracker appeared 15 minutes post EGTA treatment, and increased to approximately 25% and 70% by 30 and 90 minutes, respectively (Fig. 7C, 7D). The identity of Rap1-positive lysosomes was further confirmed by Lysoso- mal-Associated Membrane Protein one (LAMP-1) staining (data not shown). These findings indicate that in undifferenti- ated hESCs, Rap1 gradually transits from early/late endo- somes to lysosomes upon disruption of E-cadherin mediated cell–cell adhesions. Analogously, the loss of Rap1 reduces the localization of E-cadherin at cell–cell junctions and decreases the expression of E-cadherin protein in hESCs (Figs. 3–5). Collectively, these results suggest that a positive feedback- like interplay between Rap1 and E-cadherin takes place upon hESC dissociation, which leads to endosome-to-lysosome sorting and degradation of both molecules and impairs the self-renewal capacity of hESCs.
This study provides the first demonstration that the temporal interplay between the small G protein Rap1 and E-cadherin along the endocytic recycling pathway participates in the self- renewal of hESCs. Rap1 impacts hESC clonogenicity (Figs. 1–3) via affecting the initiation and formation of E-cadherin- mediated cell–cell contacts (Figs. 3–5). In turn, the inadequate formation of E-cadherin-mediated cell–cell adhesion in hESCs exacerbates the diminution of E-cadherin by accelerating Rap1 degradation, further hampering the clonogenic capacity and self-renewal of hESCs (Figs. 5–7).
The important roles of cadherins in the maintenance and function of adult stem cells have been reported [62–65]. For instance, Drosophila Epithelial-cadherin-mediated adhesion is crucial for germline stem cell maintenance and function in the Drosophila ovary [64]; N-cadherin+CD45— cells lining the bone surface support the self-renewal of mouse hemato- poietic stem cells [62, 63]; E-cadherin expressed on mouse neural stem cells regulates neural stem cell self-renewal [65]. Our study further suggests that E-cadherin plays an important role in the clonogenicity of hESCs (Fig. 5). Although E-cad- herin seems to be dispensable for mouse ESC self-renewal [66], mouse ESCs cultured on E-cadherin-coated plates show a higher proliferative capacity and lower dependence on leu- kemia inhibitory factor [67]. Furthermore, E-cadherin has been recently shown to be a critical regulator for pluripotency of mouse epiblast stem cells. The maintenance of these cells is different from mouse ESCs but similar to hESCs [68]. Col- lectively, these observations suggest that E-cadherin plays an important role in the self-renewal of stem cells.
E-cadherin is a transmembrane glycoprotein that mediates calcium-dependent, homophilic cell–cell adhesion, constitut- ing an essential component of adherens junctions in all epithe- lial tissues. In addition, E-cadherin may influence signaling pathways mediated by growth factor receptors and transduce adhesion-related signaling via its cytoplasmic domain-associ- ated proteins such as b-catenin [20, 21, 69]. Although insufficient evidence exists at the moment, E-cadherin may contribute to microenvironmental cues required for hESC self-renewal via adhesive and signaling functions. Unlike mouse ESCs, dissociated hESCs are very susceptible to apo- ptosis and undergo massive cell death after seeding as single cells [46, 70]. One hypothetical ‘‘parts list’’ for the stem cell microenvironment is the stem cell itself [71]. Our findings seem to be in favor of this hypothesis. Single hESCs incapa- ble of reestablishing E-cadherin mediated cell–cell
interactions in a timely manner display a decrease in self- renewal and an increase in cell deaths (Fig. 5). In addition, E- cadherin may influence downstream signals via its intracellu- lar adaptor protein b-catenin by serving as a ‘‘sink’’ during Wnt/b-catenin signaling [72, 73]. b-catenin is a key factor for the canonical Wnt signaling pathway [73] and has been impli- cated in the up-regulation of Nanog and Oct4 genes via nu- clear translocation and association with the transcription fac- tor T-cell factor (TCF) [74]. Notably, TCF binding sites are colocalized with many Oct4 and Nanog binding sites in the genome, indicating that the Wnt/b-catenin signaling pathway is integral to the pluripotency circuitry mediated by the TCF transcription factor [75].
In the past years, endocytic trafficking and recycling has been demonstrated to play a major role in outside-in and inside-out signaling pathways that modulate cell adhesion and function [19–21, 56]. Our data suggest that the interplay between Rap1 and E-cadherin along the endocytic recycling pathway serves as a timely and efficient means to regulate hESC self-renewal. Inhibition of Rap1 disrupts the mainte- nance and formation of the circumferential distribution of E-cadherin and reduces the recycling of E-cadherin back to the plasma membrane in hESCs (Figs. 4, 5). Such a regula- tory mechanism along the endocytic pathway may allow for a precise and rapid modulation of E-cadherin function [19, 56] and of hESC self-renewal (Fig. 5). In addition to influencing the expression of E-cadherin protein, Rap1 knockdown also led to a decrease in the expression of Nanog protein in com- parison with the control group (Supporting Information Fig. 2). Nanog has been shown as the gateway to the pluripotent ground state of stem cells [76, 77]. Further studies will help to elucidate whether Rap1 regulates the expression of Nanog protein via affecting E-cadherin expression.
Notably, distinct from adult epithelial cells, a unique positive feedback-like interaction between Rap1 and E-cadherin along the endocytic recycling pathway takes place during hESC self-renewal. In epithelial cells, adherens junction disrup- tion and E-cadherin internalization trigger a strong activation of Rap1, which in turn promotes E-cadherin recycling [14]. Such a negative feedback-like mechanism will help stabilize the system for normal tissue turnover [10, 14]. As opposed to epithelial cells, a rapid lysosomal translocation and degradation of Rap1 in hESCs occurs after adherens junction disruption (Figs. 6 and 7), which further reduces the recycling of endocy- tosed E-cadherin back to the plasma membrane. Such a positive feedback-like mechanism via endocytic trafficking and lyso- some degradation pathways seems to provide a rapid response to changes in culture conditions during hESC maintenance.
The above observations may partially explain why disso- ciated single hESCs survive poorly and hardly reaggregate to form embryoid bodies in suspension culture conditions [46, 70]. It seems that loss of Rap1 on hESC dissociation leads to E-cadherin downregulation, thereby hindering the reestablish- ment of E-cadherin mediated cell–cell contacts. The downreg- ulation of E-cadherin stimulates Caspase-3 (a pro-apoptotic gene) and suppresses Bcl-XL (an apoptotic inhibitory gene; Fig. 5F), resulting in an increase in cell deaths (Fig. 5).
Our findings have important application value to surmount several technical difficulties encountered with hESCs. For instance, to augment hESC survival, the time allotted for hESC dissociation should be controlled to less than 15 minutes and reseeding after dissociation should be completed within 30 minutes to limit Rap1 delivery to the lysosome (Fig. 7). In addition, we observed that the ROCK inhibitor Y- 27632, which has been reported to augment hESC mainte- nance [70], reduces Rap1 degradation and E-cadherin diminu- tion after hESC dissociation (unpublished observations). This observation leads to a potential connection between ROCK in- hibition and Rap1 and E-cadherin functions in the mainte- nance of hESCs. Furthermore, we observed that Bombesin, a small molecule that reduces Rap1 degradation in epithelial cells [78], also reduces Rap1 and E-cadherin loss upon GGTI 298 hESC dissociation and increases hESC colonogenic capacity (unpub- lished observations).