SEARCH

SEARCH BY CITATION

Keywords:

  • Human embryonic stem cells;
  • Glycosphingolipids;
  • Matrix-assisted laser desorption ionization mass spectrometry;
  • Differentiation;
  • Glycosyltransferases;
  • Surface markers

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Glycosphingolipids (GSLs) are ubiquitous components of cell membranes that can act as mediators of cell adhesion and signal transduction and can possibly be used as cell type-specific markers. Our previous study indicated that there was a striking switch in the core structures of GSLs during differentiation of human embryonic stem cells (hESCs) into embryoid body (EB), suggesting a close association of GSLs with cell differentiation. In this study, to further clarify if alterations in GSL patterns are correlated with lineage-specific differentiation of hESCs, we analyzed changes in GSLs as hESCs were differentiated into neural progenitors or endodermal cells by matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) and tandem mass spectrometry (MS/MS) analyses. During hESC differentiation into neural progenitor cells, we found that the core structures of GSLs switched from globo- and lacto- to mostly ganglio-series dominated by GD3. On the other hand, when hESCs were differentiated into endodermal cells, patterns of GSLs totally differed from those observed in EB outgrowth and neural progenitors. The most prominent GSL identified by the MALDI-MS and MS/MS analysis was Gb4Ceramide, with no appreciable amount of stage-specific embryonic antigens 3 or 4, or GD3, in endodermal cells. These changes in GSL profiling were accompanied by alterations in the biosynthetic pathways of expressions of key glycosyltransferases. Our findings suggest that changes in GSLs are closely associated with lineage specificity and differentiation of hESCs. STEM CELLS 2011;29:1995–2004.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Human embryonic stem cells (hESCs) are pluripotent stem cells derived from the inner cell mass of blastocysts; they possess the capacity for extensive proliferation in vitro and are capable of differentiating into a variety of specialized cell types [1, 2]. These features allow hESCs to be used as useful tools for regenerative medicine and research [3, 4]. However, the mechanisms of the growth and differentiation of hESCs are poorly understood.

Numerous studies confirmed that glycosphingolipids (GSLs) are key molecules required for regulating cellular processes [5, 6] and their expressions are developmentally controlled and cell type-specific [7, 8]. For example, stage-specific embryonic antigens (SSEA)-1, -3, and -4 were identified at defined stages of embryonic development [9–11]. They were identified as GSLs or as glycan epitopes with uncharacterized carriers. They are highly expressed in mouse and human ESCs and are considered to be specific cell markers [12]. Conversely, certain cells possess some defined GSL antigens and can be detected by specific antibodies. For example, LeX (CD15), which was found on neutrophils, is thought to mediate phagocytosis and chemotaxis [13]. The human natural killer cell antigen-1 (sulfo-3GlcAβ13Galβ14GlcNAcCer), expressed at the cell surface, is uniquely enriched in neural cells and natural killer cells, and is thought to play important roles in cell-ell interactions [14].

Recently, functional studies of GSLs revealed that specific GSLs form clusters at the cell surface and interact with membrane proteins such as caveolin-1, integrins, growth factor receptors, and tetraspanins. Such interactions contribute to modulating cell adhesion, growth, and motility [15, 16]. Conversely, GSLs were also found to be involved in the process of the epithelial mesenchymal transition (EMT) [17, 18]. The EMT process is generally recognized as an important phenomenon that occurs during embryonic development [19, 20] and is also implicated in the progression of primary tumors toward metastasis [21–23]. These findings suggest that specific GSL patterns or their clustered microdomains on the cell surface can serve as cell-specific markers and also functionally as complexes to initiate intracellular signaling through interactions with other functional membrane components. In our previous studies [24, 25], a systematic survey of changes in the expression profile of GSLs in undifferentiated hESCs and differentiated embryoid bodies (EBs) further suggested that specific hESC markers may be expressed in cancers, and specific biomarkers for cancer may also be potential candidates for hESC markers.

In this study, we further investigated the GSL profile of hESCs which had differentiated toward neural progenitor cells and definitive endodermal lineages and studied the role of specific glycosyltransferases during the lineage-specific differentiation of hESCs.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Differentiation of hESCs into Neural Progenitor Cells

The protocol for differentiation of hESCs into neural progenitor cells was modified from the procedure described by Zhang and coworkers [26] and Cho et al. [27]. hESC lines were expanded on irradiated mouse embryonic fibroblast layers in hESC growth medium as previously described [24]. For differentiation, hESC colonies were detached from the mouse embryonic fibroblasts by treatment of cultures with 1 mg/ml of dispase (BD Biosciences, San Jose, CA) and grown as floating cell aggregates of EBs for 4 days with a daily change of hESC medium without basic fibroblast growth factor (bFGF). Floating cell aggregates were switched to a neural medium consisting of Dulbecco/s Modified Eagle Medium/F12 (2:1), supplemented with N2 (Gibco-Invitrogen, Carlsbad, CA), 0.1 mM nonessential amino acids, 2 μg/ml heparin, and 10 ng/ml bFGF for another 3 days. On day 7, floating cell aggregates were then grown on an adherent substrate in the same neural medium with a change of medium every other day. ESC aggregates attached and formed individual colonies by approximately days 7–8. Neuroepithelial cells, manifesting as columnar cells organized into neural tube-like rosettes, developed on approximately day 12–14. The neural rosettes were isolated and expanded in neural medium for another 7 days. Immunostaining and flow cytometric analysis with the markers nestin, Sox-2, Pax-6, and Sox-1 were performed as previously described [24].

Differentiation of hESCs into Definitive Endodermal Lineages

The procedure for differentiation of definitive endoderm from hESCs was modified from that described by Cai et al. [28]. hESC lines were expanded on irradiated mouse embryonic fibroblasts in hESC growth medium as previously described [24]. For differentiation, culture media of hESCs were replaced by Roswell Park Memorial Institute medium 1640 (Gibco-Invitrogen, Carlsbad, CA) supplemented with 0.5 mg/ml bovine serum albumin fraction V (Gibco-Invitrogen, Carlsbad, CA) and 100 ng/ml human Activin A (PeproTech, Rocky Hill, NJ) for 1 day. The next day, 0.1% insulin-transferrin-selenine (Gibco-Invitrogen, Carlsbad, CA) was added to this medium. On days 3–7, the concentration of insulin-transferrin-selenine was increased to 1% in the same medium.

Antibodies Used in the Flow Cytometric Analysis or Immunofluorescence Analysis

For marker analysis of neural progenitor cells, primary antibodies used were anti-nestin (mAb clone 196908, R&D, Minneapolis, MN), Sox-2 (AB5603, Millipore-Chemicon, Billerica, MA), Pax-6 (mAb clone AD2.38, Millipore-Chemicon, Billerica, MA), and Sox-1 (AB15760, Millipore-Chemicon, Billerica, MA). For analysis of definitive endoderm markers, primary antibodies used were anti-Sox-17 (purified goat immunoglobulin G, R&D, Minneapolis, MN) and FOXA2 (ab40874, Abcam, Cambridge, MA). For the hESC-specific marker analysis, primary antibodies used were anti-Oct3/4 (mAb clone C-10, Santa Cruz Biotechnology, Santa Cruz, CA) and SSEA-4 (mAb clone MC-813-70, Millipore-Chemicon, Billerica, MA). For the cell-surface GSL analysis, primary antibodies used were anti-SSEA-3 (mAb clone MC-631, Millipore-Chemicon, Billerica, MA), Globo H (mAb prepared from VK9 hybridoma provided by Dr. Alice Yu, Academia Sinica, Taipei, Taiwan), Lc4Ceramide (Lc4Cer [ms IgM, clone K21, GeneTex, Irvine, CA]), H type 1 (mAb clone 17-206, Abcam, Cambridge, MA), GM1 (ab31119, Abcam, Cambridge, MA), GM3 (mAb clone GMR6, Seikagaku, Tokyo, Japan), and GD3 (mAb clone MB3.6, BD Pharmingen, Franklin Lakes, NJ). The secondary antibodies used were Alexa Fluor 488-conjugated goat anti-mouse IgG, goat anti-mouse IgM, goat anti-human IgG, and goat anti-human IgM (Gibco-Invitrogen, Carlsbad, CA). The negative isotype controls, depending on the species and subclasses of the primary antibodies used, were rat IgM (for SSEA-3), mouse IgG3 (for SSEA-4, VK9, H type 1, and GD3), IgG2b (for Oct3/4), and mouse IgM (for GM3), all from eBioscience (San Diego, CA).

MALDI-MS Profiling and MALDI Collision-Induced Dissociation MS/MS Analysis

For each cell line, GSLs were purified and permethylated as previously described [24]. MALDI-MS profiling of permethylated GSLs was performed as described below [24]. Briefly, MALDI-MS profiling of permethylated GSLs was carried out on an ABI 4700 Proteomics Analyzer (Applied Biosystems, Foster City, CA) using a 2,5-dihydroxybenzoic acid matrix (10 mg/ml in water). Low- and high-energy collision-induced dissociation MALDI-tandem mass spectrometry (MS/MS) sequencing was performed on a Quadrupole/Time-of-Flight Ultima MALDI (Waters Micromass, Milford, MA) with α-cyano-4-hydrocinnamic acid and a 4700 Proteomics Analyzer using the 2,5-dihydroxybenzoic acid matrix, respectively.

Quantitative Reverse-Transcription Polymerase Chain Reaction

Total RNA samples were isolated from each cell line using the Trizol reagent (Gibco-Invitrogen). Complementary cDNAs were synthesized from total RNAs using Superscript III reverse transcriptase (Gibco-Invitrogen) with random hexamer primers. A TaqMan real-time polymerase chain reaction (PCR) assay was performed on an ABI HT7900 Fast Real-Time PCR System (Applied Biosystems). Five nanograms of a cDNA sample was used for the quantitative PCR reaction. The threshold cycle numbers (Ct value) and quantity of each analyzed sample were determined using the comparative Ct method.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Immunofluorescence and Flow Cytometric Analysis of GSL Expression Patterns

In our previous study [24], changes in GSL expression in undifferentiated hESCs and in differentiated EB outgrowth after 16 days of induction were examined. With EB outgrowth, three germ layer markers were detected in various ratios, for which ectoderm cells seemed to be dominant in the samples (68%), then mesoderm cells (32%), and only a few endoderm cells were detected (4%). Therefore, the GSL pattern observed in EB outgrowth might reflect the complexity of the composition of EB outgrowth cells. To clarify the GSL expression pattern after differentiation of hESCs, we analyzed GSL patterns as ESCs differentiated into specific cell lineages.

To examine changes in GSLs in specific cell lineages, hESCs were, respectively, induced to differentiate in vitro into neural progenitor cells and definitive endoderm, as described in Materials and Methods section. Once ESCs had differentiated into neural progenitor cells, hESCs aggregated and developed a columnar morphology. After selection and expansion, cells gradually organized into typical neural tube-like rosettes from central columnar cells (Fig. 1A). After 21 days of differentiation, neuroectodermal markers such as nestin (97%), Sox2 (96%), Pax6 (99%), and Sox1 (75%) were expressed based on immunofluorescence and flow cytometric analyses (Fig. 1B). Conversely, to specifically differentiate hESCs in vitro into definitive endoderm as described in Materials and Methods section, hESCs were found to have gradually transformed into less-dense, flattened cells containing prominent nuclei on day 7 (Fig. 1A). The appearance of endodermal cell markers such as Sox17 (16%) and FoxA2 (95%) showed that these cells had differentiated into definitive endodermal lineages (Fig. 1C). In contrast, markers for neuroectodermal or definitive endoderm cells were not detected in undifferentiated hESCs (Fig. 1D). In addition, Oct3/4 and SSEA-4 were analyzed by immunofluorescence as controls to monitor the undifferentiated status of hESCs (Fig. 1D).

thumbnail image

Figure 1. Characterization of hESCs, neural progenitor cells, and definitive endoderm with immunofluorescence and flow cytometry. (A): Phase-contrast microscopy of hESCs, neural progenitor cells, and definitive endoderm was carried out as described in Materials and Methods section. (B): Immunofluorescence and flow cytometric analyses of neural progenitor cells with antibodies directed against nestin, Sox-2, Pax-6, and Sox-1. (C): Immunofluorescence and flow cytometric analyses of definitive endoderm cells with antibodies against Sox-17 and FOXA2. (D): Lack of expression of neural progenitor cell markers (anti-nestin and Pax-6) and definitive endoderm markers (anti-Sox-17) in undifferentiated hESCs by immunofluorescence. In addition, Oct3/4 and SSEA-4 were analyzed as the controls to monitor the undifferentiated status of hESCs. In the immunofluorescence assay, cell marker-specific antibodies were stained in green, and nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (blue). In the flow cytometric analysis (B and C), viable cells were gated; cells stained with specific antibodies are shown with a bold black line and isotype antibodies with a thin gray line. Values represent the mean of three experiments. Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole; hESC, human embryonic stem cells; SSEA, stage-specific embryonic antigen.

Download figure to PowerPoint

In addition, flow cytometric analysis was used to characterize GSLs of the undifferentiated hESCs, differentiated neural progenitor cells, and definitive endoderm cells using GSL-specific antibodies (Fig. 2). All experiments were performed three times, and data were presented as mean ± SE, with symbols * and ** to indicate statistical significant differences of p values <.05 and <.001, respectively, in Figure 2. It was found that the well-known marker of hESCs, SSEA-3, was highly expressed in undifferentiated hESCs (94%) and was reduced to 3% in neural progenitor cells. In definitive endoderm cells, the expression of SSEA-3 had also decreased to 63%. Conversely, SSEA-4 was also highly expressed in hESCs but decreased to only 86% after differentiation into neural progenitor cells. However, 96% of definitive endoderm cells expressed SSEA-4 (Fig. 2A). In these studies, specific monoclonal antibodies (mAbs), MC631 (anti-SSEA-3) and MC813-70 (anti-SSEA-4) were used, respectively. Furthermore, Globo H was also highly expressed in undifferentiated hESCs (∼99%), but it almost completely disappeared from neural progenitor cells (2%; Fig. 2A). In contrast, Globo H was slightly downregulated to 85% after differentiation into definitive endoderm cells.

thumbnail image

Figure 2. Flow cytometric analyses of glycosphingolipids (GSLs) in undifferentiated hESCs, neural progenitor cells, and endodermal cells. (A): SSEA-3, SSEA-4, Globo H, Lc4Cer, and fucosyl Lc4Cer (using the H type 1 antibody) were analyzed using GSL-specific antibodies versus the isotype controls. (B): GM3, GD3, and GM1 were analyzed using GSL-specific antibodies versus the isotype controls. Cells stained with specific antibodies are shown by green, while isotype antibodies are shown by black. Statistical data are presented as mean ± SE from three experiments. The symbols * and ** depict statistical significance with p values <.05 and <.001, respectively, by t test. Abbreviations: hESCs, human embryonic stem cells; SSEA, stage-specific embryonic antigen.

Download figure to PowerPoint

In addition to SSEA-3, -4, and Globo H, the lactose-series of GSLs, such as Lc4 and fucosyl Lc4Cer (H type 1 antigenic determinant), were examined in hESCs and definitive endoderm cells. It was found that their expressions only decreased to a small extent in definitive endoderm cells (Fig. 2A). Conversely, ganglio-series GSLs, such as GM3, GM1, and GD3 were not detected in definitive endoderm cells. These findings distinctly differed from their high expression in neural progenitor cells (Fig. 2B).

MALDI-MS and MS/MS Analyses of GSL Expression Patterns

To decipher precise profiles of GSLs in differentiated neural progenitors and endodermal cells, we used MALDI-MS and MS/MS sequencing to systematically analyze GSLs in these cells. Total crude extracts after Folch partitioning were permethylated and subjected to MALDI-MS and MS/MS analyses as previously described [24]. The annotations of GSL profiles in Figure 3 were assigned based on m/z values of each of the major molecular ion signals. m/z values of the respective GSLs were fitted to the expected core structures of the three GSL series (globo-, ganglio-, and lacto-series) along with the usual ranges of the most common permutation of sphingosine and fatty acyl chains, as previously described [24].

thumbnail image

Figure 3. Matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) profiles of permethylated glycosphingolipids (GSLs) from undifferentiated hESCs, EB outgrowth, neural progenitor cells, and endodermal cells. MALDI-MS profiles from hESCs, EB outgrowth, neural progenitor cells, and definitive endoderm are shown in parallel. Identification of individual GSLs was based primarily on their determined molecular masses, facilitated by the knowledge of the core sequences of ganglio-, lacto-, and globo-series, and the usual range of fatty acyl heterogeneity associated with the ceramide moiety. Fatty acyl heterogeneities associated with GSLs from hESCs were inferred to be primarily C16:0 and C18:0, with a mass difference of 28 U but several of the GSLs also additionally carried longer fatty acyl chains of up to C24:0/C24:1 (at 112/110 mass units higher than those with C16:0), as bracketed and annotated accordingly on the Figure. Abbreviations: EB, embryoid body; Fuc, fucose; Gal, galactose; Glc, glucose; GlcNAC, N-acetylglucosamine; GAlNAc, N-acetylgalactosamine; hESCs, human embryonic stem cells; NeuAc, N-acetylneuraminic acid.

Download figure to PowerPoint

As shown in Figure 3, MS analysis confirmed the presence of SSEA-3, -4, Globo H, Lc4Cer, and H-type 1 (fucosyl Lc4Cer) in undifferentiated hESCs. In addition, when hESCs differentiated into neural progenitor cells, the GSL expression pattern changed and became dominated by ganglio-series GSLs. GD3 and GM3 were the most abundant GSLs, but other ganglio-series GSLs including GM1 and GD1, were also detected in these neural progenitor cells. Thus the GSL profile of neural progenitor cells was similar to that for EB outgrowth [24], because ectoderm cells are known to be the most dominant cell type among EB outgrowth cells (Supporting Information Fig. S1 of [24]).

However, when hESCs differentiated into definitive endodermal lineages, GSL patterns were quite distinct from those observed in neural progenitor cells (Fig. 3). As shown in Figure 3, the most prominent GSL signal endodermal at m/z 1,460 corresponded to a mixture of Lc4 and Gb4Cer, with the latter being more abundant according to the MS/MS analysis (Fig. 4). Minor signals, which could be assigned as GM3, GM1, GD1, Globo H, and fucosyl Lc4/Gb4Cer, were also detected (Fig. 3).

thumbnail image

Figure 4. Matrix-assisted laser desorption collision-induced dissociation ionization mass spectrometry (MALDI CID MS/MS) analyses of parent ions at m/z 1,460 from undifferentiated hESCs and hESC-derived endodermal cells. (A): MALDI CID MS/MS analysis was performed as described in Materials and Methods section and [24]. A majority of the afforded fragment ions corresponded to Y, B, and C ions derived from single or double glycosidic cleavages. C” ions, as annotated, were commonly observed along with or in place of C ions and occurred at two mass units lower than C ions. Assignments of all major fragment ions are shown in the embedded illustrations. Only the O atoms at preferred cleavage sites were drawn out to distinguish between B and C or C” ions. As shown in the panel of undifferentiated hESCs, the intensity ratios of the Y ion at m/z 1,200 and 996 and the B ions at m/z 282 and 486 indicated that Lc4Cer was more abundant than Gb4Cer in undifferentiated hESCs. In contrast, in the definitive endoderm, the intensity ratios of the Y ion at m/z 1,200 and 996 and the B ions at m/z 282 and 486 indicated that the Gb4Cer was the more abundant in definitive endoderm. Abbreviations: hESCs, human embryonic stem cells.

Download figure to PowerPoint

Interestingly, the relative amount of Lc4Cer and Gb4Cer that share the same molecular weight and hence giving a common molecular ion at m/z 1,460 (Fig. 4) by MS analysis was shown by further MS/MS analysis to differ significantly in undifferentiated hESCs and definitive endoderm. Although most of the afforded fragment ions were commonly present in the two MS/MS spectra, the Y ion at m/z 1,200 deriving from neutral loss of a terminal N-acetyl-hexosamines (HexNAc) was clearly more abundant relative to the Y ion at m/z 996 resulting from loss of a terminal Hex-HexNAc in the case of definitive endoderm compared to that of hESCs. This was further corroborated by the relative abundance of the two B ions at m/z 282 and 486, corresponding to terminal HexNAc and Hex-HexNAc, respectively. Collectively, the MS/MS data strongly indicated that the Gb4 structure, namely HexNAc-Hex-Lac, was more abundant in definitive endoderm, whereas the Lc4 structure, Hex-HexNAc-Lac, was the more abundant in hESC. Interestingly, according to our MS/MS analysis, there were significant differences in the compositions of the spectrum of m/z 1,460 between undifferentiated hESCs and definitive endoderm (Fig. 4).

Alterations of the Expressions of GSL-Related GTs

Expressions of glycosyltransferases (GTs) involved in GSL biosynthetic pathways were analyzed by a quantitative reverse-transcription PCR to elucidate the mechanism underlying the alterations in GSLs during hESC differentiation into neural progenitors and definitive endoderm cells (Fig. 5). First, B3GALT5, which was responsible for SSEA-3 (Gb5Cer) and Lc4Cer synthesis in neural progenitor cells, was downregulated to 6% of the level in undifferentiated hESCs, thus contributing to decreases in SSEA-3 and Lc4Cer expressions in those cells (Fig. 5A). Furthermore, two fucosyltransferases, FUT1 and FUT2, which catalyze the synthesis of Globo H and fucosyl Lc4Cer, also decreased during hESC differentiation into neural progenitor cells (10% and 2%, respectively; Fig. 5A). These changes may have led to the downregulation of globo- and lacto-series GSL synthesis during hESC differentiation. In contrast to GTs, which are involved in the biosynthesis of globo- and lacto-series GSLs, GTs involved in the biosynthesis of ganglio-series GSLs, such as ST3GAL5(GM3 synthases), ST8SIA1 (GD3 and GT3 synthases), and ST3GAL1 (sialyltransferase 4), increased by 2.7-, 10.1-, and 1.5-fold, respectively (Fig. 5A). These changes may have accounted for the increased expressions of gangliosides in differentiated neural progenitor cells.

thumbnail image

Figure 5. Expression of glycosphingolipid (GSL)-related GTs during differentiation of hESCs into neural progenitor cells or definitive endodermal lineages. Expression levels of globo- and lacto-series GSL-related GT genes (B3GALT5, FUTI, and FUT2) and ganglio-series-related GT genes (ST3GAL5, ST8SIA1, and ST3GAL1) were analyzed by a reverse-transcription polymerase chain reaction. (A): Changes in selected GT gene expressions during differentiation of hESCs into neural progenitor cells. (B): Changes in selected GT gene expressions during differentiation of hESCs into definitive endoderm. Relative quantities were present as the ratios of neural progenitor cells/hESCs or definitive endoderm/hESCs. Values of the Y-axis represent log2 relative quantities; therefore, negative values indicate downregulation and positive values indicate upregulation. Error bars represent one standard deviation from the mean of relative quantities. Multiples of change of GT genes during differentiation of hESCs are summarized and shown in the diagrams. Abbreviations: GTs, glycosyltransferases; hESCs, human embryonic stem cells.

Download figure to PowerPoint

Conversely, when hESCs were differentiated into the cell lineages of definitive endoderm, expressions of GTs were downregulated to some extent. GTs involved in the biosynthesis of globo- and lacto-series GSLs, such as B3GALT5, FUT1, and FUT2, respectively, decreased by 0.2-, 0.6-, and 0.9-fold (Fig. 5B). GTs involved in the biosynthesis of ganglio-series GSLs, such as ST3GAL5, ST8SIA1, and ST3GAL1, respectively, decreased to 70%, 30%, and 80% (Fig. 5B). These findings explained the decreases in globo- and lacto-series GSLs and the diminished expressions of ganglio-series GSLs in definitive endoderm cells. GSL biosynthetic pathways and multiples of changes in expressions of key GTs during hESC differentiation are summarized in Figure 5A and 5B.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Our previous results showed that the GSL profiles drastically changed during differentiation of hESCs to EB outgrowth cells to form three germ layers [24]. To clarify changes in GSLs after lineage-specific differentiation of hESCs in this study, we induced hESCs to specifically differentiate into neural progenitor cells (97% nestin+, 99% Pax6+, and 96% Sox2+ cells) or definitive endodermal lineage (16% SOX17+, 95% FOXA2+) and assessed the expressions of GSLs. When hESCs were differentiated into neural progenitor cells, the GSL expression pattern was similar to that for EB outgrowth, because ectoderm cells were dominant in the EB outgrowth sample [24]. However, when hESCs were differentiated into definitive endodermal lineages, GSL patterns totally differed from those observed in EB outgrowth or in neural progenitors (Fig. 3). The most prominent GSL in the MALDI-MS spectrum was Gb4Cer in definitive endoderm, based on our MS/MS analysis.

SSEA-3 and -4 are well known markers of hESCs and the epitopes defined by their specific mAbs MC631and MC813-70 were, respectively, delineated as GalNAcβ1-3Galα1-4Gal and NeuAcα2-3Galβ1-3GalNAc [10, 11]. In addition, Globo H was shown to be highly uniquely expressed in hESCs [24] and its polyclonal antibodies reacted with both Globo H and SSEA-3 [25]. Previous studies in stem cells often relied on the use of these mAbs which recognize glycan epitopes, but these epitopes were now shown to be present in various glycoconjugates ([10] and this study). For example, anti-SSEA-3 MC631 was reactive with essentially all undifferentiated hESCs and became mostly undetectable in neural progenitor cells. These results were confirmed by the absence of a MALDI-MS signal at m/z 1,664 assigned to Gb5Cer (SSEA-3). In definitive endoderm cells, however, the appearance of apparent SSEA-3+ cells detected with its mAb also decreased but was still present in as much as 62% of immunostained cells. But, the MALDI-MS analysis indicated only a small amount, if any, of m/z 1,664.2 in the spectrum for this cell lineage. It was reported that mAb MC631 reacted against SSEA-3 (Gb5Cer), Gb4, SSEA-4, Globo H, and Forssman glycolipids [10]. Therefore, immunostained MC631+ cells of definitive endoderm cells can be attributed to cross-reactivities with the mAb against glycoconjugates which carry the MC631 epitope. Indeed, our MS analysis of definitive endoderm cells confirmed the presence of a signal corresponding to Gb4Cer GSL at m/z 1,460. However, the possibility of a cross-reaction with other uncharacterized entities cannot be excluded.

Conversely, by immunostaining with its mAb, MC813-70, SSEA-4 (Neu5Ac-Gb5Cer) was also highly expressed in hESCs, but positive cells decreased to 86% of cells after differentiation into neural progenitor cells. In addition, 96% of definitive endoderm cells seemed to show positive immunostaining with this mAb. However, the MS analysis failed to detect the presence of SSEA-4 (Neu5Ac-Gb5Cer) in either neural progenitor or endodermal cells. Instead, GD1a and GM1b which cross-react with the SSEA-4 mAb were detected in both neural progenitor and definitive endoderm cells. In other words, the MS analysis failed to detect SSEA-4 in cells committed toward ectoderm or endoderm lineages, while a specific mAb used for immunostaining seemed to reveal the presence of SSEA-4+ cells after lineage differentiation. This discrepancy, however, can be reconciled in view of cross-reactivities of mAbs used with various glycoconjugates in cells containing specific glycan epitopes. While it is possible that the MS profile may be biased against the SSEA-4 epitope, the lack of immunostaining with mAb MC631 (against SSEA-3) in neural progenitor cells, which is known to recognize SSEA-4 (Neu5Ac-Gb5Cer) equally well [10, 11], attests to the absence of SSEA-4 in neural ectoderm cells. Furthermore, the anti-SSEA-4 mAb, MC813-70, is known to cross-react with GD1a and GM1b, which carry the NeuAcα2-3Galβ1-3GalNAc epitope for this mAb [10, 11]. The MS analysis of neural progenitor and definitive endoderm cells, however, confirmed the presence of these gangliosides. Therefore, immunostained SSEA-4+ cells can be attributed to cross-reactivities of mAb MC813-70 with GM1b, GD1a, or even extended core 1 o-glycan glycoproteins [10, 11, 29].

Similarly, there were essentially no immunostained Globo H+ cells in neural progenitors and significant decreases in definitive endoderm in flow cytometric analysis. MS analyses confirmed the absence of Globo H in neural progenitors, while small amount of Globo H was demonstrated in definitive endodermal cells. This discrepancy between the results from immunostaining and MS analyses could be accounted for by the findings that the antibodies against Globo H cross-react with SSEA-3 and its precursor Gb4Cer [25], which is the major GSL in this endodermal cell.

It was also noted in immunostaining and the MS analysis that lactose-series of GSLs, such as Lc4Cer and fucosyl Lc4Cer, were found in both hESCs and definitive endoderm cells (albeit in small amounts) and their expressions apparently disappeared in neural ectoderm cells. Further, the MS/MS analysis indicated that Lc4Cer contributed more to the signal attributed to Gb4/Lc4Cer at m/z 1,460 in hESCs, but in definitive endoderm, Gb4Cer was the more-abundant component instead. Therefore, these results showed that the relative abundance of Lc4Cer versus Gb4Cer differed in undifferentiated hESCs and definitive endoderm. Epitopes for mAbs against Lc4Cer and fucosyl Lc4Cer were separately, shown to be Lec and H type 1 [30], which represent two members of blood group antigens; these glycan determinants are also known to be present on both glycolipids and glycoproteins [31]. In the definitive endoderm, the MALDI-MS analysis indeed indicated the presence of Lc4Cer and fucosyl Lc4Cer, which were not found in neural progenitor cells. Conversely, GD3 was not detectable in hESCs or definitive endoderm cells but were highly expressed in neural progenitor cells, thus qualifying it as a distinctive marker, whereas the expression of other ganglio-series GSLs, such as GM3, GM1, and GD1, were more universal.

Expression levels of key glycosyltransferases in these samples were also investigated. When hESCs were differentiated into neural progenitor cells, the RNA expressions of glycosyltransferases responsible for the biosyntheses of globo-series GSLs, such as FUT1, FUT2, and B3GALT5, were downregulated, compared to undifferentiated hESCs. Furthermore, in definitive endoderm cells, the MS analyses indicated that Gb4Cer was highly expressed without the expression of SSEA-3 (Gb5Cer) or SSEA-4 (sialyl-Gb5Cer). These results can be explained by a decrease in the expression of Gb5 synthase (B3GALT5) during hESC differentiation into specific endodermal lineage. In addition, the expressions of glycosyltransferases of ganglio-series GSLs, such as ST3GAL1, ST3GAL5, and ST8SIA1, increased by 1.5–10.1-fold. Conversely, when hESCs were differentiated into definitive endoderm cells, expressions of these glycosyltransferases were downregulated. Therefore, when hESCs were differentiated into different cell lineages, there were substantial differences in GSL profiles and also in the expressions of key glycosyltransferases. Although we did not analyze the expression of glycosidase in this study, it was reported that neuraminidase 3 (NEU3), a plasma membrane-associated ganglioside sialidase was markedly upregulated in human colon and renal carcinomas and involved in apoptosis suppression [32, 33]. Whether NEU3 could also affect the expression of gangliosides to modulate the proliferation and differentiation of hESC awaits further studies.

Functional roles of GSLs were recently investigated. For example, GM3 was found to be present together with caveolin-1 and Src in a low-density membrane fraction resistant to Triton X-100, and the interaction of GM3 and caveolin-1 greatly reduced the motility of ovarian cancer A2980 cells through inactivation of Src [34]. Another motility inhibition effect of gangliosides was identified in the interaction of either GM3 or GM2 with the tetraspanins, CD9 and CD82 [35–37]. Furthermore, complexes of GM3 with tetraspanin CD9 and GM2 with tetraspanin CD82 in microdomains of cell surface inhibited cell motility and growth by blocking the associations of integrin α-3 with FGFR and integrin α-3 with cMet [38, 39]. In addition, based on the finding of a metabolic shift from oxidative to anaerobic glycolysis, Warburg [40] had claimed an association between hypoxia and oncogenic transformation. Recently, it was shown that gangliotetraosylceramide and ganglioside GM2 did play functional role(s) in EMT process induced by hypoxia and transforming growth factor-β [17, 18]. All these results suggest that the expression and organization status of GSLs play a central role in the cellular phenotype and also in controlling biological functions and tumor progression.

Our finding of a switch in the core structures of GSLs from globo- and lacto- to ganglio-series during differentiation of hESC to EB outgrowth [24] is consistent with the observed changes in expression patterns of GSLs during embryonic development of mice [5]. It was reported that globo-series SSEA3, SSEA4, and Globo H were expressed at high levels in four cell stages, and then later decline during mouse embryogenesis [5]. Conversely, SSEA1 [III3Fuc-nLc4Cer; Galβ1-4(Fucα1-3)GlcNAcβ1-3Galβ1-4Glcβ1-1′Cer] was not expressed until the morula stage of mouse embryos [5]. In addition, ganglio-series GM3, GD3, GT3, GM2, and GD2 were expressed in later stages on neural crest formation in mice [7, 41] and GD3, GD1a, and GT1b were also expressed after somite formation [7, 42]. In addition, in Ugcg-suppressed mouse ESCs, it was found that there were significant defects in neural differentiation in vitro, especially neural maturation related to glial fibrillary acidic protein and microtubule-associated protein-2 expressions [43]. Therefore, present observations concerning GSL profile in hESCs and the differentiated ectodermal lineage seemed to agree with these previous reports in embryonic development of mice. However, data on GSL profile in early human embryos and their lineage-specific differentiation are lacking.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

In this study, we assessed the expressions of GSLs when hESCs were differentiated into neural progenitor cells or definitive endoderm. The GSL patterns from each cell lineage were totally different from each other. Moreover, the changes of expression levels of key glycosyltransferases in these samples were coincident with the GSL expression patterns in substance. This study provides valuable GSL profile characterization for the lineage commitment of in vitro hESC differentiation and should help to elucidate molecular mechanisms underlying differentiation of hESCs and facilitate the identification of in vitro differentiation for clinical applications. In addition, because of cross-reactivities of exiting mAbs with various GSLs containing the epitopes, the expressions of GSLs in hESCs and their derivatives will rely on a detailed mass spectrometry analyses.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

This study was supported by grants from the Genomics Research Center, Academia Sinica and the National Science Council, Taiwan (NSC99-3111-B-001-006). The Resource Core Facilities for hESCs was supported by an NSC grant (NSC99-3111-B-001-003). MALDI-MS and MS/MS data were acquired at the NRPGM Core Facilities for Proteomics and Glycomics by NSC grants NSC 98-3112-B-001-025.

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

The authors indicate no potential conflicts of interest.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES