From a paper presented at the International Symposium on Embryonic Stem Cells – Prospects for Human Health, at the University of Sheffield, UK, 10 September 2001.
Biochemical properties of a keratan sulphate/chondroitin sulphate proteoglycan expressed in primate pluripotent stem cells*
Version of Record online: 25 APR 2002
Journal of Anatomy
Volume 200, Issue 3, pages 259–265, March 2002
How to Cite
Cooper, S., Bennett, W., Andrade, J., Reubinoff, B. E., Thomson, J. and Pera, M. F. (2002), Biochemical properties of a keratan sulphate/chondroitin sulphate proteoglycan expressed in primate pluripotent stem cells. Journal of Anatomy, 200: 259–265. doi: 10.1046/j.1469-7580.2002.00034.x
- Issue online: 25 APR 2002
- Version of Record online: 25 APR 2002
- Accepted for publication 8 February 2002
- chondroitin sulphate;
- embryonal carcinoma;
- embryonic stem cell;
- keratan sulphate;
We previously identified a pericellular matrix keratan sulphate/chondroitin sulphate proteoglycan present on the surface of human embryonal carcinoma stem cells, cells whose differentiation mimics early development. Antibodies reactive with various epitopes on this molecule define a cluster of differentiation markers for primate pluripotent stem cells. We describe the purification of a form of this molecule which is secreted or shed into the culture medium. Biochemical analysis of the secreted form of this molecule shows that the monomeric form, whilst containing keratan sulphate, resembles mucins in its structure and its modification with O-linked carbohydrate. Immunofluorescence and immunoblotting data show that monkey and human pluripotent stem cells react with antibodies directed against epitopes on either carbohydrate side chains or the protein core of the molecule.
Primate pluripotent stem cell lines have now been developed from human embryonal carcinomas (ECs), monkey and human blastocysts, and human embryonic and fetal gonads (Thomson et al. 1995; Shamblott et al. 1998; Thomson et al. 1998; Pera et al. 2000; Reubinoff et al. 2000). The establishment of diploid human embryonic stem (ES) or embryonic germ cell lines has the potential to revolutionize biomedical research and to provide new opportunities for cell-based therapy. Although human ES cell cultures have been characterized using antibodies against cell surface antigens, the cultures are in fact heterogenous both in morphology and in terms of their surface antigen profile (unpublished observations). The biological significance of this heterogeneity is not yet understood, but it may be that differences in reactivity with various markers reflect differences in developmental potential amongst subpopulations within the cultures. A more precise definition of stem cell phenotype may assist in the development of more effective culture methodology for ES cells. The power of precise immunological analysis of cell differentiation lineage is clearly illustrated in the case of haematopoietic or lymphoid cells, where the structure of the stem cell populations at various levels of differentiation is well defined by expression of surface markers, most of which are characterized at the molecular level.
Previously, we identified a keratan sulphate proteoglycan present on the cell surface of human EC cells (Pera et al. 1988), tumour cells whose differentiation mimics early steps in mammalian development (Andrews, 1988). The proteoglycan was localized to the pericellular matrix, and its expression was differentiation dependent: surface expression of the molecule on stem cells was lost following spontaneous or induced differentiation in vitro (Pera et al. 1988, 1989; Roach et al. 1994). This proteoglycan represents a surface antigen recognized by a cluster of monoclonal antibodies raised in different laboratories against human EC cells (Badcock et al. 1999). Using a monoclonal antibody reactive with the core protein, we found widespread expression in various epithelia of mid-trimester human fetal tissues (Mason & Pera, 1992), reminiscent of previous findings with monoclonal antibodies against keratan sulphate glycosaminoglycans (reviewed in Funderburgh, 2000).
In an earlier study, we reported the purification of the matrix-associated proteoglycan from human EC cells (Cooper et al. 1992). Much of the material so isolated was in an aggregated form. While keratan sulphate and chondroitin sulphate accounted for all of the glycosaminoglycan content of the pericellular matrix form, only chemical deglycosylation achieved complete removal of sugar residues, to reveal core protein bands of Mr 55 and 48 kDa. We have observed previously that the proteoglycan could be detected in culture medium by immunoassay (Pera et al. 1988). Consequently, a new purification protocol was developed to study the secreted form of the molecule. The purified material was used as an immunogen in the production of a second monoclonal antibody, and the expression of the molecule on human ES cells and rhesus monkey ES cells was examined.
Materials and methods
Enzyme-linked immunosorbent assay (ELISA)
Enzyme-linked immunosorbent assay (ELISA) with the GCTM-2 antibody, reactive with an epitope on the proteoglycan core protein, and a monoclonal antibody against fibronectin (Sigma Chemical Co.) was carried out as described previously. The titre of proteoglycan immunoreactivity was estimated at various stages of the purification as described (Cooper et al. 1992).
Production of GCT 27 C-4 cell conditioned medium
The cell line GCT 27 C-4, a nullipotent clone of human EC cells (Pera et al. 1989), was subcultured at a 1 : 2 split ratio and grown overnight in a mixture of Minimal Essential Medium-Alpha and Ham’s F12 medium (1 : 1 v/v) supplemented with 10% fetal calf serum, 1 mm glutamine and 1 µg mL−1 hydrocortisone. The cells were then washed twice with Iscove’s Modified Dulbecco’s medium supplemented with 35 µg mL−1 human transferrin, 5 µg mL−1 bovine serum albumin, 2.5 µg mL−1 human insulin, 1 mm glutamine and 1 µg mL−1 hydrocortisone, and were grown in this medium for 2–3 days. The conditioned medium was harvested, fresh medium was added and another harvest was carried out 2–3 days later.
Purification of immunoreactive proteoglycan from culture medium
Conditioned medium was filtered through Millex AP50 prefilters (Millipore Corporation) to remove cell debris. It was then passed through a 2–5 mL peanut lectin affinity column (Vector Laboratories) overnight at 40 mL h−1 at 4 °C. The column was washed with 10 volumes of 10 mm Tris/HCL, 150 mm NaCl, 0.1 mm CaCl2, 0.01 mm MnCl2 pH 7.4; subsequent procedures were carried out at room temperature.
Bound proteins were eluted with 0.6 m galactose in the wash buffer, collecting 0.5-mL fractions. Immunopositive fractions, as determined by ELISA, were pooled and diluted 1 : 1 with 50 mm Tris/HCL pH 6.8 (loading buffer). They were then loaded onto a 1-mL MonoQ anion exchange column on an FPLC apparatus (Pharmacia), and washed with 10 volumes of loading buffer. Proteins were eluted with a 0–1 m NaCl gradient. Strongly bound material was eluded with 2 m NaCl at the end of the run. The immunopositive material that eluted between 0.4 m and 0.6 m NaCl was pooled and concentrated using MicroSep concentrators (Flowgen) with a nominal Mr cut-off of 30 kDa. This concentrated material was separated on a Superose 6 HR 10/10 column (Pharmacia), and 24 mL was equilibrated with 50 mm Tris/HCl, 0.5 m NaCl pH 6.8. Molecular mass markers used to calibrate the column were thyroglobulin, 669 kDa; ferritin, 440 kDa; and catalase, 232 kDa. The majority of immunoreactive material (eluting between 700 and 400 kDa) was pooled and diluted four-fold with 50 mm Tris/HCl pH 7.4 (final salt concentration 0.12 m NaCl), prior to affinity chromatography on a 1-mL heparin-Sepharose column (Pharmacia). The flow through was collected and subjected to further analysis. Bound human fibronectin was eluted from the heparin column with 2 m NaCl in 50 mm Tris/HCl, pH 7.4.
Purified proteoglycan was run on 7.5% sodium dodecyl suplhate polyacrylamide gel electrophoresis (SDS-PAGE) and then silver-stained and immunoblotted (Cooper et al. 1992).
Enzyme digestion with glycosidases
Purified proteoglycan was subjected to keratanase (keratan sulphate 1,4-β-d-galactohydrolase; EC 22.214.171.124; Pseudomonas sp., ICN Biomedicals Ltd) or chondroitinase (chondroitinase ABC lyase; EC 126.96.36.199; Proteus vulgaris, also from ICN) digestion followed by SDS/PAGE and immunoblotting as described previously (Cooper et al. 1992). Sequential digestion with sialidase (EC 188.8.131.52; Arthrobacter ureafaciens) and O-glycosidase (Endo-α-N-acetylgalactosaminidase; EC 184.108.40.206/220.127.116.11: Streptococcus pneumoniae), both from Oxford Glycosystems, was carried out as follows: 0.2 units of sialidase was reconstituted in 10 µL of reaction buffer (100 mm Na acetate pH 5.0). 5× reaction buffer was added to the proteoglycan followed by the sialidase and incubated at 37 °C for 2 h. O-glycosidase reaction buffer (5×, 100 mm Na citrate pH 6.0, 100 mg mL−1 bovine serum albumin, 0.02% sodium azide) was added to the mixture, followed by 6 mUnits of O-glycosidase, and incubated at 37 °C for 2 h. Finally, reducing SDS sample buffer was added and the samples were subjected to SDS/PAGE followed by immunoblotting as above.
Rotary shadowing electron microscopy
Purified intact proteoglycan samples were prepared for rotary shadowing as described previously (Cooper et al. 1992) and examined under a Philips EX301 electron microscope at a nominal magnification of ×45 000.
Production of a new monoclonal antibody
Except where noted, the standard immunological protocols described by (Harlow & Lane, 1996) were followed.
The secreted proteoglycan was purified as described above. The crude lectin-bound glycoprotein fraction from conditioned medium was used as the immunogen. Conditioned medium glycoproteins (6.5 mg) were mixed with the synthetic adjuvant Adju-Prime (Pierce Chemical Co.) in accordance with the manufacturer’s instructions and injected intraperitoneally into female Balb/C mice approximately 6 weeks old. The mice were boosted twice with the same dose of immunogen alone administered intraperitoneally at biweekly intervals. Test bleeds, obtained after the second immunization, were screened by antibody capture ELISA against serial dilutions of the second major peak of proteoglycan immunoreactivity eluded from a Mono-Q Sepharose column as described above.
Spleen cells from the mouse whose serum displayed the highest titre of antibody against the semipurified proteoglycan were fused to NS-1 myeloma cells and hybrids selected using standard protocols. Supernatants from 472 wells containing approximately 1000 hybridoma clones were screened by ELISA against semipurified proteoglycan as described above; 48 positive wells were rescreened by ELISA and indirect immunofluorescence on methanol : acetone-fixed monolayers of GCT 27 C-4 cells (Pera et al. 1988). One hybridoma supernatant (TG 343) showed strong reactions in both of these assays.
Following subcloning by limiting dilution, the reactivity of TG 343 supernatant against semipurified proteoglycan and cultured cells was confirmed. TG 343 supernatant was next tested in an immunoblot assay against purified proteoglycan either in native form, or following treatment with keratanase or chondroitinase ABC as described above. The class and subclass of the antibody were determined using a commercially available kit (Amersham International).
Staining of rhesus monkey and human embryonic stem cells
Rhesus monkey embryonic stem cell line 278.5, and human embryonic stem cell lines HES-1 and HES-2 were cultured as described earlier (Thomson et al. 1995; Reubinoff et al. 2000). Human ES cell lines HES-3 and HES-4 were established from blastocysts as described (Reubinoff et al. 2000) and show expected properties of human ES cells including a diploid karyotype and formation of teratomas containing derivatives of all three germ layers in animal hosts; these cell lines were grown under the same conditions as HES-1 and HES-2. Immunostaining was performed as described previously. In some experiments the proportion of cells reactive with GCTM-2 or TRA-1-60 was estimated by harvesting cells using dispase and trypsin, staining in suspension and counting the number of positive cells under the fluorescence microscope. Immunoblotting was carried out on extracts prepared from cell line HES-2 by extraction with 1.0% NP-40, 150 mm NaCl and 10 mm Tris-HCl, pH 7.5, followed by low-speed centrifugation to remove nuclei and debris. Proteins were separated on SDS-PAGE reducing gels, and transferred to membranes which were probed as described above.
Purification of soluble proteoglycan
A combination of lectin affinity, anion exchange and gel filtration chromatography typically resulted in a purification of the proteoglycan from serum-free conditioned medium of approximately 5000-fold, with a yield of several per cent, as estimated from ELISA results (typical result for a 2.5-L preparation, Table 1). Losses were due in large part to heterogeneity in charge and size of the immunoreactive material bound to the lectin column. The preparation was subjected to SDS/PAGE followed by silver staining or immunoblotting (Figs 1 and 2 lane A, Fig. 3 lane A) and was shown to correspond to a single broad band of Mr approximately 200 kDa with very little immunoreactive material in the stacking gel. Fibronectin present in the immunoreactive peak from gel filtration was removed by heparin affinity chromatography with a loss of proteoglycan of about 15%, the amount that remained bound to the column in the presence of 1 m NaCl (not shown).
|Step||Volume (mL)||Protein* (µg mL−1)||Relative titre†||Yield (% of initial)||Purification|
Deglycosylation of soluble proteoglycan
Both chondroitinase (data not shown) and keratanase (Fig. 2 lane B) partially deglycosylated and enhanced the immunoreactivity of the proteoglycan, evidenced by an increase in electrophoretic mobility on SDS-PAGE and a stronger reaction in immunoblotting. Results with chondroitinase were more variable than with keratanase. A combination of sialidase followed by O-glycosidase similarly enhanced the immunoreactivity of the purified material and increased its electrophoretic mobility, and converted a proportion of the purified proteoglycan to two immunoreactive bands of Mr 55 and 48 kDa (Fig. 3 lane B). Two bands of similar size were obtained previously following chemical deglycosylation of the cell surface proteoglycan (Cooper et al. 1992).
Electron microscopy of soluble proteoglycan
The purified proteoglycan molecules that ran on Superose 6 gel filtration at ~700 kDa appeared fibrillar with a length of about 80 nm and thickness of 3.5 nm (Fig. 4). This is in contrast to the proteoglycan purified from the cell surface (Cooper et al. 1992) the majority of which formed long filamentous aggregates several micrometres in length.
Production of a new monoclonal antibody
The availability of the purified proteoglycan enabled us to produce a new mouse monoclonal antibody for further immunocytochemical studies, Clone TG343. Both antibodies reacted with the native proteoglycan in this assay (Fig. 5). Pretreatment of the proteoglycan with chondroitinase ABC, which sometimes produces a modest degree of enhancement of immunoreactivity with GCTM-2, had no effect on its reaction with either antibody in this experiment, indicating that the new antibody was probably not directed against chondroitin sulphate chains. Pretreatment with keratanase enhanced the immunoreactivity of the proteoglycan with both antibodies, and resulted in visualization of antigen with greater electrophoretic mobility; in some blots, enzymatic treatment degraded material reactive with either antibody down to the size of the protein core. However, ELISAs in which GCTM-2 and TG 343 were allowed to complete for binding to immobilized semipurified proteoglycan showed that TG 343 did not interfere with GCTM-2 binding (not shown). Thus it is unlikely that the epitopes identified by the two antibodies are identical.
Reactivity of primate pluripotential cell lines with antibodies to the proteoglycan
Feeder-dependent human embryonal carcinoma cell line GCT-27X-1, rhesus monkey ES cell line 278.5 and human ES cell lines HES1, 2, 3 and 4 were examined by indirect immunofluorescence microscopy for reactivity with antibodies against several monoclonal antibodies reactive with the proteoglycan. The majority of cells in cultures fixed 5–7 days after subculture under conditions optimal for stem cell renewal were stained with all antibodies and the staining had a similar appearance in all cell lines: a granular staining outlining the cell body, often with a dark area in the nucleus, and with some deposition onto the culture surface. In the case of the human ES cell lines HES-1 and HES-2 the proportion of cells positive with TRA1-60 was consistently higher (70% vs. 50%) than that stained with GCTM-2. The proportion of cells staining with TG 343 was similar to that stained with GCTM-2. Immunoblotting of detergent-soluble extract from cell line HES-2 revealed that GCTM-2 reacted with a broad band of around 200 kDa (Fig. 6).
Biochemical properties of the secreted proteoglycan
The proteoglycan studied here exits in several forms, depending on the source from which it is isolated. The properties of the soluble form of the molecule differ from those of the previously described pericellular matrix form in several respects. In preparations from conditioned medium, very little immunoreactive material remained in the stacking gel during SDS-PAGE, the majority migrating as a broad band of approximately 200 kDa, in contrast to the pericellular matrix form. The soluble molecules are less highly charged than those on the cell surface, most eluting at a lower ionic strength from anion exchange columns. We noted previously that if subconfluent monolayers were extracted with 1 m guanidinium hydrochloride, some of the immunoreactivity eluted at lower ionic strength during ion exchange chromatography, and was smaller than the bulk of the insoluble material. It is possible that upon reaching confluence cells begin to assemble the proteoglycan monomers into a large pericellular matrix complex.
Rotary shadowing confirms the difference between the two forms. The majority of the proteoglycan purified from the cell surface formed long filamentous aggregates several micrometres in length although a few putative monomers approximately 80 nm long and 3.5 nm wide were observed. Most of the material in the soluble form was monomeric. Formation of large aggregates is also observed for some secreted mucins (Strous & Dekker, 1992), which the proteoglycan resembles in appearance, but mucin aggregates are generally formed by intermolecular disulphide bonds. The aggregation observed in this proteoglycan is not due to disulphide linkages, or to hydrophobic interactions, since neither reducing conditions nor detergents can disaggregate the molecules.
Both chondroitinase (data not shown) and keratanase (Fig. 4B) partially deglycosylated and enhanced the immunoreactivity of the proteoglycan. A combination of sialidase followed by O-glycosidase similarly enhanced the immunoreactivity of the purified material and increased its electrophoretic mobility, and converted a proportion of the purified proteoglycan to two immunoreactive bands identical in size to those obtained by chemical deglycosylation of the cell surface proteoglycan (Cooper et al. 1992).
This development of this purification methodology, and the production of a new monoclonal against the core protein should facilitate molecular cloning of the protein core and analysis of its function in pluripotent stem cells.
Expression in primate pluripotent stem cells
These studies extend and confirm previous results which demonstrate that this proteoglycan is a major surface antigen of primate pluripotent stem cells. Reactivity with antibodies recognizing both the core protein and carbohydrate epitopes, plus the immunoblotting data, strongly indicate that normal primate pluripotent cells express a molecular form of this molecule similar to that found on embryonal carcinoma cells. The proportion of cells stained with antibody TRA1-60 was somewhat higher than that seen with antibody GCTM-2; neither stained all cells in ES cultures maintained under optimal conditions. It is possible that the differences in reactivity with the two antibodies reflect some type of epitope masking in different cell types. The biological significance of this antigen heterogeneity remains to be determined, but the antibodies described herein will help determine the developmental potential of the subpopulations of cells within the cultures. Further study of this proteoglycan may also elucidate its function on the surface of pluripotent cells.
This work was begun with support from the Cancer Research Campaign, UK, and has been further supported by the National Health and Medical Research Council.
- 1988) Human teratocarcinomas. Biochim. Biophys. Acta 200, 17–36. (
- 1999) The human embryonal carcinoma marker antigen TRA-1-60 is a sialylated keratan sulfate proteoglycan. Cancer Res. 59, 4715–4719. , , , (
- 1992) A novel keratan sulphate proteoglycan from a human embryonal carcinoma cell line. Biochem. J. 286, 959–966. , , , (
- 2000) Keratan sulfate: structure, biosynthesis, and function. Glycobiology 10, 951–958. (
- 1996) Antibodies: a Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Habor Laboratory. , (
- 1992) Immunohistochemical and biochemical characterisation of the expression of a human embryonal carcinoma cell proteoglycan antigen in human germ cell tumours and other tissues. Eur. J. Cancer xx, 1090–1098. , (
- 1988) Analysis of cell-differentiation lineage in human teratomas using new monoclonal antibodies to cytostructural antigens of embryonal carcinoma cells. Differentiation 39, 139–149. , , , , , (
- 1989) Isolation and characterization of a multipotent clone of human embryonal carcinoma cells. Differentiation 42, 10–23. , , , (
- 2000) Human embryonic stem cells. J. Cell Sci. 113, 5–10. , , (
- 2000) Embryonic stem cell lines from human blastocysts: somatic differentiation in vitro. Nat. Biotechnol. 18, 399–404. , , , , (
- 1994) Hepatocytic transcription factor expression in human embryonal carcinoma and yolk sac carcinoma cell lines: expression of HNF-3 alpha in models of early endodermal cell differentiation. Exp. Cell Res. 215, 189–198. , , (
- 1998) Derivation of pluripotent stem cells from cultured human primordial germ cells. Proc. Natl Acad. Sci. USA 95, 13726–13731. , , , , , , et al. (
- 1992) Mucin-type glycoproteins. Crit Rev. Biochem. Mol Biol. 27, 57–92. , (
- 1995) Isolation of a primate embryonic stem cell line. Proc. Natl Acad. Sci. USA 92, 7844–7848. , , , , , , (
- 1998) Embryonic stem cell lines derived from human blastocysts. Science 282, 1145–1147. , , , , , , et al. (