We have previously identified the cell adhesion protein podocalyxin expressed in a human pluripotent stem cell, embryonal carcinoma (EC), which is a malignant germ cell. Podocalyxin is a heavily glycosylated membrane protein with amino acid sequence homology to the hematopoietic stem cell marker CD34. Since the initial discovery of podocalyxin in a cancerous stem cell, numerous new studies have identified podocalyxin in many different human cancers and in embryonic stem cells lines (ES) derived from human embryos. Embryonal carcinoma, as do all human pluripotent stem cells, expresses TRA-1-60 and TRA-1-81 antigens, and although their molecular identities are unknown, they are commonly used as markers of undifferentiated pluripotent human stem cells. We report here that purified podocalyxin from embryonal carcinoma has binding activity with the TRA-1-60 and TRA-1-81 antibodies. Embryonal carcinoma cells treated with retinoic acid undergo differentiation and lose the TRA-1-60/TRA-1-81 markers from their plasma membrane surface. We show that podocalyxin is modified in the retinoic acid-treated cells and has an apparent molecular mass of 170 kDa on protein blots as compared with the apparent 200-kDa molecular weight form of podocalyxin expressed in untreated cells. Furthermore, the modified form of podocalyxin no longer reacts with the TRA-1-60/TRA-1-81 antibodies. Thus, embryonal carcinoma expresses two distinct forms of podocalyxin, and the larger version is a molecular carrier of the human stem cell-defining antigens TRA-1-60 and TRA-1-81.
Podocalyxin, a member of the CD34-related family of sialomucins, is a transmembrane glycoprotein originally identified and cloned from human kidney as a component of the podocyte glycocalyx [1, –3]. Subsequently, podocalyxin was identified in subsets of human hematopoietic cells and shown, like CD34, to interact specifically with the endothelial cell adhesion receptor l-selectin [4, , , –8]. We have previously identified podocalyxin as a testis tumor marker expressed on human embryonal carcinoma (EC) cell lines . Human EC cell lines are established malignant pluripotent stem cell lines derived from human germ cell tumors . Currently, EC is one of three sources of human pluripotent stem cells, with the other two being embryonic stem cells (ES) derived from the inner cell mass of the blastocyst and embryonic germ cells from the fetus .
Since the initial discovery that podocalyxin is expressed in human EC stem cells, several new studies have shown podocalyxin expression in other human and mouse stem cells. Podocalyxin has been identified by transcriptome profiles in two human ES cell lines studied , and another recent study reported the identification by microarray analysis of podocalyxin expression in six of six ES cell lines studied . Remarkably, this report identified podocalyxin as one of only six proteins as markers of undifferentiated human pluripotent ES cells in all ES lines studied . Another recent report characterized the gene expression of two human ES lines and showed podocalyxin highly expressed compared with pooled human RNA . Also, podocalyxin has been identified on human and murine progenitor blood cells, but not on mature blood cells, suggesting a role for podocalyxin in multipotent blood stem cells [15, 16]. Although originally identified as a structural, extracellular matrix component of the kidney, podocalyxin is rapidly acquiring a role as a marker of pluripotent and multipotent stem cells. Intriguingly, complementing its new role as a stem cell marker, podocalyxin is emerging as a marker of human malignancy. In addition to being a testis tumor marker, within the last year it has been identified as a marker in several different human cancers including breast, liver, kidney, and blood cell cancers [17, , –20].
We have previously shown that podocalyxin is expressed in EC cells as a heavily glycosylated transmembrane protein with an apparent molecular weight (MW) of 200 kDa on protein gels . We originally discovered podocalyxin in EC cell lines because of its strong reactivity to the plant lectin peanut agglutinin (PNA), a protein with high binding affinity for terminal d-galactose residues found on glycoproteins . Podocalyxin is biochemically similar to another EC tumor marker that has been characterized with (and named after) several different monoclonal antibodies, including GCTM2, TRA-1-60, and TRA-1-81 antibodies [22, , –25]. It has been proposed that the different antibodies recognize distinct and unique epitopes on the same large glycoprotein, reported to be a large keratan sulfated proteoglycan expressed on human EC cells; however, final confirmation of a single carrier protein awaits its molecular identity . Human ES cells derived from inner cells of the blastocyst also express GCTM2, TRA-1-60, and TRA-1-81 antigens, and, as seen with differentiation of embryonal carcinoma, these stem cell-specific markers are lost during ES cell differentiation . These antigens are widely used in the stem cell research community as markers of undifferentiated pluripotent ES cells, yet their molecular identity has remained elusive for decades, and there is a need to determine the molecular nature of these and other markers of human stem cells . We have previously shown that the GCTM2 monoclonal antibodies are reactive to podocalyxin from embryonal carcinoma and have proposed that podocalyxin may be the common molecular carrier of GCTM2, TRA-1-60, and TRA-1-81 stem cell antigens .
After the identification of podocalyxin in EC, we looked at the fate of podocalyxin after treatment with retinoic acid (RA) in two separate EC cell lines; NTERA2/D.l (NT2.D1) and NCCIT are characterized and documented EC cell lines that differentiate into neuroectodermal lineages when exposed to retinoic acid [29, 30, 32, 33]. We report here that podocalyxin expression in undifferentiated and retinoic acid-treated EC cells are distinct by showing that purified podocalyxin from untreated EC cells reacts positively with the stem cell-specific antibodies TRA-1-60 and TRA-1-81, whereas the purified form of podocalyxin from the retinoic acid-treated EC cells is modified and is no longer reactive to the TRA-1-60 and TRA-1-81 antibodies. Because the TRA-1-60 and TRA-1-81 antigens are carbohydrate epitopes , post-translational glycosylation changes must occur on podocalyxin during the retinoic acid treatment of EC cells, and these changes result in the loss or modification of the TRA-1-60 and TRA-1-81 stem cell antigens. This evidence is further proof that podocalyxin is a carrier of the TRA-1-60 and TRA-1-81 human stem cell antigens found on the membrane surface of human stem cells, and that podocalyxin is a leading candidate as the protein which expresses the distinct markers—TRA-1-60, TRA-1-81, and GCTM2—that are extensively used by the research community to define human pluripotent stem cells.
TERA-1 EC cell line (ATCC number HTB-105; ATCC, Manassas, VA, http://www.atcc.org/) was grown in McCoy's 5a modified medium. NTERA-2.D1 cell line (ATCC number CRL-1973) was grown in Dulbecco's modified Eagle's medium. NCCIT (ATCC number CRL-2073), HEK293, COS-1, and LNCaP cell lines were grown in RPMI 1640. All media were supplemented with 10% fetal bovine serum and 100 μg/ml of streptomycin sulfate. NT2.D1 cells were treated with 10 μM of transretinoic acid (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com/) made in 10 mM stock solutions of dimethyl sulfoxide (DMSO) and added to the media. All tissue culture reagents were obtained from Invitrogen (Carlsbad, CA, http://www.invitrogen.com/). EC cells were scraped and split 1–3, whereas RA-treated EC cells were removed with trypsin before and during RA treatment to pass the cells until cells stop dividing. RA was replaced daily during EC RA treatment.
Transient Transfection of Podocalyxin
The human podocalyxin gene inserted into the mammalian expression vector pcDNA3 (Invitrogen) was used to transfect the HEK293 human kidney and the COS-1 monkey kidney cell lines using Lipofectamine transfection reagent (Invitrogen). Approximately 40 μg of plasmid DNA was used for each 60-cm2 culture dish. Cells were harvested 48 hours after transfection.
Protein lysates were extracted from cells by homogenizing the cells in phosphate/saline buffer (PBS; pH 7.0) with 1% Chaps (Sigma-Aldrich) detergent and protease inhibitors including 0.1 mM phenylmethylsulfonyl fluoride, l μg/ml leupeptin, 1 μg/ml aprotinin, l μg/ml soybean trypsin inhibitor and l mM EDTA (Sigma-Aldrich). Insoluble cellular debris was removed by centrifugation, and podocalyxin was purified from the soluble protein lysate with PNA lectin coupled to CNBr-Sepharose-4B beads (Sigma-Aldrich) at 1 mg/ml concentration according to the manufacturer's instructions. Immunopurification was accomplished using mouse anti-podocalyxin monoclonal antibodies (3D3 antibodies) coupled to CNBr-Sepharose beads (0.6 ml of mouse ascites fluid with 0.5 ml of beads). TRA-1-60 immunopurification was done by adding 10 μg of TRA-1-60 antibody (Chemicon, Billerica, MA, http://www.chemicon.com/) and 200 μl of anti-mouse IgM-agarose beads (Sigma-Aldrich) to l ml of cell lysates. All protein lysates were at concentrations of 4 mg/ml. To isolate podocalyxin, the beads were centrifuged, washed six times with buffer, and mixed with SDS-polyacrylamide gel electrophoresis (PAGE) protein-loading buffer. Mock purifications were preformed identically except that glycine-coupled Sepharose beads were used instead of PNA-Sepharose beads, and mouse IgG was coupled to beads for 3D3 immunoprecipitation control or added to protein lysates for TRA-1-60 immunoprecipitation control. Podocalyxin was immunopurified with goat anti-human podocalyxin antibodies (R&D Systems, Minneapolis, MN, http://www.rndsystems.com/; catalog number AF1658) at 20 μl to 0.5 ml EC lysate and mock-controlled immunopurified experiments were done using goat anti-rabbit IgG antibodies followed by 200 μl of protein A/G-Sepharose beads (Sigma-Aldrich) and processed as describe with PNA purifications.
NCCIT cells were grown on LAB-TEK II plastic chamber slides (Nalge Nunc International, Naperville, IL, http://www.nalgenunc.com) for 12 days with and without retinoic acid. Cells were washed PBS, fixed with 4% formaldehyde, blocked with 10% calf serum, and incubated with TRA-1-60 or TRA-1-81 antibodies at 1 μl/ml concentration in 0.1% Tween-20 PBS buffer, washed with PBS, and incubated for 1 hour with 0.5 μg/ml fluorescein isothiocyanate-conjugated goat anti-mouse IgM antibody (Vector Labs, Burlingame, CA, http://www.vectorlabs.com/) and visualized with a 60× lens on a Zeiss microscope (Carl Zeiss, Jena, Germany, http://www.zeiss.com/). Similar controls were done on treated and untreated cells without the addition of the primary TRA antibodies.
Protein samples were separated by 5%, 7.5% or 10% SDS-PAGE protein gels and either stain in Coomassie blue dye in 50% methanol, 5% acetic acid or transferred for 1 hour at 300 mV to 0.45 μm nitrocellulose paper. PNA blots were blocked in 3% bovine serum albumin for 1 hour, incubated with peroxidase-conjugated PNA (0.1 μg/ml; Sigma-Aldrich), washed with PBS for 1 hour, and developed with ECL (Amersham, Buckinghamshire, United Kingdom, http://www.amersham.com/). All monoclonal antibody blots were blocked in 5% nonfat milk/PBS. The 4F10 and 3D3 were used at 5 μl/ml and 1 μl/ml ascites fluid, the goat anti-podocalyxin antibodies were used at 2 μl/ml. TRA-1-60 and TRA-1-81 (Chemicon) were used at 1 μg/ml in PBS with 0.1% Tween-20 and 3% nonfat milk protein. Washed blots were incubated with secondary goat anti-mouse IgG peroxidase-conjugated antibodies (Vector Labs) at 1 μg/ml in PBS with 0.1% Tween-20 and 3% nonfat milk buffer for 3D3 and 4F10 podocalyxin blots, secondary goat anti-mouse IgM peroxidase-conjugated antibodies (Vector Labs) for the TRA blots, and sheep anti-goat IgG peroxidase-conjugated antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com/) for the goat anti-podocalyxin polyclonal antibody (Podo) blots. The blots were developed with ECL. Protein concentrations of cell and tissue lysates were determined with the BCA Protein Assay (Pierce, Rockford, IL, http://www.piercenet.com/).
Podocalyxin Expression in Human Embryonal Carcinoma
Figure 1 shows podocalyxin expressed in four different human EC cell lines as a large heavily glycosylated membrane protein with an approximate MW of 200 kDa on protein gels. Using the 3D3 mouse anti-podocalyxin monoclonal antibody that was generated against a peptide corresponding to a 15-amino acid sequence in podocalyxin's extracellular domain , podocalyxin is detected in the protein lysates of TERA-1, NTERA2.Dl (NT2.D1) human EC cells (Fig. 1, lanes 1, 2), HET-833K EC cells (HT-E; Fig. 1, lane 5), and NCCIT EC cells (Fig. 1, lane 8). The calculated MW of podocalyxin from the amino acid sequence is 55 kDa; however, podocalyxin is observed on blots as a 200-kDa MW diffuse protein band. Podocalyxin is a sialomucin glycoprotein, and the apparent 150-kDa MW difference between the calculated and observed MW is attributed to glycosylation [2, 3]. Podocalyxin transiently expressed in the human embryonic kidney cell line (HEK293) (Fig. 1, lanes 3, 10) or COS-1 monkey kidney epithelial cells (Fig. 1, lane 4) is detected on the 3D3 Western blots as a diffuse protein band with an approximate MW of 170 kDa. The MW of podocalyxin expressed in transfected HEK293 and COS-l cells is similar to the reported MW of podocalyxin in human kidney and the observed MW we detect in human kidney tissue protein lysate samples (Fig. 1, lane 6). As has been previously reported, HEK293 cells endogenously express podocalyxin (Fig. 1, lane 11) as a protein with an apparent MW of approximately 170 kDa, but there is considerably less expressed compared with the EC cells . Mock-transfected COS-1 cells (Fig. 1A, lane 7) do not express detectable levels of podocalyxin. Finally, EC cells treated with the differentiation reagent retinoic acid express a smaller form of podocalyxin (Fig. 1, lane 9), and its appearance on 3D3 blots looks remarkably similar to podocalyxin expressed in HEK293 cells (Fig. 1, lanes 10, 11).
Podocalyxin Is Modified in Retinoic Acid-Treated Embryonal Carcinoma Cells
Our discovery of podocalyxin expressed on human EC cells and subsequent studies by several other laboratories reporting podocalyxin expression on human and murine stem cells has implications that podocalyxin is a marker of pluripotent stem cells [9, , , –13]. To further investigate the fate of podocalyxin on EC cells, we have looked at the expression of podocalyxin on EC cells treated with the differentiation agent, retinoic acid, a derivative of vitamin A, which induces differentiation of embryonic and malignant cells . The NT2.D1 cell line will undergo extensive differentiation into several different morphologically distinct somatic cell types when exposed for several weeks to RA, and is an established model for studying differentiation [29, 30, 33]. Molecular changes can be detected in the pluripotent NT2.D1 cells as early as 7 days of exposure with retinoic acid, and include the loss of the pluripotent cell surface stem cell markers TRA-1-60, TRA-1-81, GCTM2, and SSEA-3 . We examined the expression of podocalyxin in RA-treated NT2.Dl cells; after 7 days of treatment, the NT2.Dl cells appeared larger and flatter, and had less of a tendency to form aggregates, consistent with published reports . Podocalyxin could still be detected in the cell lysates of these treated cells; however, podocalyxin was detected on 3D3 Western blots as a smaller-size protein with an approximate molecular weight of 170 kDa (Fig. 2A, lane 2). Notably, the modified form of podocalyxin was similar to the size of podocalyxin transiently expressed in the HEK293 and COS-1 cell lines (Fig. 1, lanes 9, 10) and to the reported molecular weight of podocalyxin isolated from human kidney (Fig. 1A, lane 6) . This result was also repeated in a second EC cell line, NCCIT, and similar modifications of podocalyxin were detected after 12 days of RA treatment (see Fig. 4). Furthermore, after several weeks of continuous RA exposure, both EC cell lines will stop growing, with most of the cells will having a neuronal or a large and flat looking appearance. During the long exposure to RA, podocalyxin expression is markedly reduced in the cells and eventually becomes undetectable in the cell lysates of these senescence cells. The eventual loss of podocalyxin expression as the EC cells move down differentiation pathways during the continuous exposure to RA is in agreement with the transcription profiling studies and other studies that indicate podocalyxin as a marker of pluripotent stem cells [9, , , –13].
Podocalyxin Is a PNA-Binding Glycoprotein
Podocalyxin binds to PNA , a plant lectin specific for terminal d-galactose residues . Podocalyxin is detected on PNA blots of NT2.Dl cell lysates as a 200-kDa diffuse protein band (Fig. 2B, lane 1) and has a similar electrophoretic mobility on protein gels as does the band detected with the 3D3 podocalyxin antibodies (Fig. 2A, lane 1). Likewise, the modified form of podocalyxin from RA-treated NT2.Dl cells is detected with PNA lectin as a 170-kDa diffuse protein band (Fig. 2B, lane 2). Podocalyxin is the only major PNA-binding protein in EC cell protein lysates (Fig. 2B, lane 1), and is amenable to isolation by PNA-affinity purification . Purified podocalyxin preps from NT2.Dl cell lysates (Fig. 2C, lane 3) and RA-NT2.D1 cell lysates (Fig. 2C, lane 4) show no protein bands on Coomassie-stained gels compared with the total cell lysates (Fig. 2C, lanes 1, 2). Purified podocalyxin is also not detected on the stained gels, which is presumably due to the extensive glycosylation of the podocalyxin protein backbone. However, podocalyxin is readily detected with 3D3 podocalyxin antibodies (Fig. 2A, lanes 3, 4) and PNA (Fig. 2B, lanes 3, 4) in purified podocalyxin fractions. Podocalyxin is not detected on 3D3 blots (Fig. 2A, lane 5) or PNA blots (Fig. 2B, lane 5) in control mock-purified fractions from NT2.Dl cells. Podocalyxin is also not detected in a control cell line LNCaP, a malignant prostate cell line (Fig. 2A, lane 6; Fig. 2B, lane 6).
The 200-kDa MW Form of Purified Podocalyxin Has Binding Activity with the TRA-1-60/TRA-1-81 Antibodies
Human EC cell lines express the stem cell-defining antigens TRA-1-60 and TRA-1-81 that are detected on EC cells lysate blots as diffuse 200-kDa MW proteins . The TRA-1-60 and TRA-1-81 monoclonal antibodies bind to unknown carbohydrate-based epitopes expressed in EC on a large 200-kDa MW proteoglycan . Using the TRA-1-60 and TRA-1-81 antibodies, we looked at TRA-1-60 and TRA-1-81 expression in RA-treated NT2.Dl cells. A 200-kDa MW protein band is detected by TRA-1-60 (Fig. 4A, lane 1) and TRA-1-81 (Fig. 4B, lane 1) antibodies in NT2.Dl cell lysates, and is similar to the protein detected on blots of NT2.Dl protein lysates with 3D3 podocalyxin antibodies (Fig. 4D) or PNA lectin (Fig. 4C). However, as had been observed and reported previously [23, 26], the TRA-1-60 and TRA-1-81 200-kDa band is not detected in RA-treated NT2.Dl protein lysates (Fig. 4A, 4B, lane 2). The TRA-1-60 and TRA-1-81 antibodies do not detect any specific protein bands in the control LNCaP cell line protein lysates (Fig. 4A, 4B, lane 3).
We have previously shown that GCTM2 antibodies will bind to purified podocalyxin and have hypothesized that the reported unknown common proteoglycan, which expresses the TRA-1-60, TRA-1-81, and GCTM-2 antigens, is podocalyxin . To further confirm this hypothesis, we tested the reactivity of the TRA antibodies to PNA-purified preps of podocalyxin from RA-treated and untreated NT2.Dl cells. Using a second distinct podocalyxin monoclonal antibody 4F10, which was generated against a peptide corresponding to a 15-amino acid sequence in the cytoplasmic domain of podocalyxin , we detect the purified 200-kDa MW podocalyxin band in untreated NT2.Dl cells (Fig. 4E, lane 1) and the 170-kDa MW form in the RA-treated cells (Fig. 4E, lane 2). These results are similar to the results obtained with the 3D3 podocalyxin antibodies (Fig. 2A). However, only the purified 200-kDa form of podocalyxin is detected with TRA-1-60 (Fig. 4F, lane 1) and TRA-1-81 antibodies (Fig. 4G, lane 1), whereas the purified 170-kDa MW podocalyxin is not reactive to either antibody (Fig. 4F, 4G, lane 2). These findings are similar to what is seen in the NT2.D1 protein lysate preps (Fig. 2A), and are evidence that podocalyxin expresses the TRA-1-60/TRA-1-81 epitopes, whereas the modified smaller form of podocalyxin has lost or modified these epitopes and no longer reacts with the TRA antibodies.
To further confirm these results, we repeated the RA experiments with a second EC cell line, NCCIT. Furthermore, in an effort to rule out the possibility of contaminating proteins in the PNA-purified podocalyxin preps, we used new anti-podocalyxin polyclonal antibodies to immunopurify podocalyxin. The podocalyxin-specific polyclonal antibodies detect the 200-kDa MW form of podocalyxin in untreated NCCIT cell lysates (Fig. 3A1, lane 1) and the 170-kDa MW form in NCCIT cell lysates from cells treated with RA for 12 days (Fig. 3A1, lane 2) Also, podocalyxin expression is reduced in NCCIT protein lysates from NCCIT cells exposed to RA for approximately 30 days (Fig. 3A1, lane 3). Blots with the TRA antibodies of the same lysates as Figure 3A1 show that only protein lysates from the untreated NCCIT cells have TRA-1-60 and TRA-1-81 reactivity (Fig. 3A2, lane 1), which is similar to the results from the NT2.Dl cells (Fig. 4). Actin blots of the same lysates as in Figure 3A1 show relatively equal protein loading of the three protein preps from the treated and untreated NCCIT cells (Fig. 3B, lanes 1, 2, 3).
Next, we did immunopurification and control mock purifications of podocalyxin using the anti-podocalyxin polyclonal antibodies and tested the purified podocalyxin preps with the TRA-1-60 and TRA-1-81 antibodies. Immunopurified podocalyxin from the untreated NCCIT cell lysates reacted positively with TRA-1-60 (Fig. 3, C2, lane 2) and TRA-1-60 antibodies (Fig. 3, C3, lane 2), but control purification preps from untreated NCCIT lysates show no reactivity to the antibodies (Fig. 3, C2, C3, lane 1). TRA-1-60 or TRA-1-81 reactivity could not be detected in either immunopurified podocalyxin (Fig. 3, C2, C3, lane 4) or control purified preps (Fig. 3, C2, C3, lane 3) from protein lysates of NCCIT cells treated with RA for 12 days before harvest. Both immunopurified forms of podocalyxin react with PNA (Fig. 3, C1, lanes 2, 4), but no PNA reactivity is seen in the control purification preps. (Fig. 3, C1, lanes 1, 3).
As an additional confirmation of the loss of TRA epitopes on the RA-treated NCCIT cells, we looked at the TRA antigens expression on the surface of the cells with immunofluorescence studies. TRA-1-60 expression was seen in approximately 50%–80% of the untreated NCCIT cells, depending on levels EC cell aggregation (Fig. 3D). However, less than 1% of the NCCIT cells treated with RA for 12 days were positive for TRA-1-60 reactivity, indicating that, as seen with the TRA blots, these treated EC cells loss or modified the TRA-1-60 antigen on the surface of their cells. Similar results were seen with the TRA-1-81 antibodies, and control experiments show no fluorescence in the untreated NCCIT cells (data not shown).
Podocalyxin Expresses the TRA-1-60 and TRA-1-81 Pluripotent-Defining Antigens in Embryonal Carcinoma
Unlike the NCCIT and NT2.D1 EC lines, the TERA-1 cells will not differentiate in the presence of RA. To determine whether podocalyxin is expressing the TRA antigens in TERA-1 cells, we performed several different purification schemes of podocalyxin from protein lysates of TERA-1 cells. We tested PNA-purified preps from TERA-l cells and found that the TRA-1-81 antibodies detected a 200-kDa protein band in purified podocalyxin preps (Fig. 5A, lane 2) that was similar to the TRA-1-81 specific band in the TERA-1 lysate (Fig. 5A, lane 1), but the band was not detected in mock-purified podocalyxin preps (Fig. 5A, lane 3). Next, we purified the TRA-1-60 antigen from TERA-1 protein lysates with TRA-1-60 antibodies and identified a 200-kDa PNA-binding protein in these purified fractions (Fig. 5B, lane 1) that was not detected in mock-purified TRA-1-60 antigen preps (Fig. 5B, lane 2). Podocalyxin 3D3 antibodies covalently coupled to Sepharose beads were used to immunopurify podocalyxin from TERA-1 lysates. TRA-1-60 antibodies detected a 200-kDa protein band in 3D3 immunopurified podocalyxin preps (Fig. 5C, lane 1) but did not detect a similar band in mock-purified fractions of podocalyxin (Fig. 5C, lane 2). Finally, we used the goat anti-podocalyxin polyclonal antibodies to immunopurify podocalyxin, which reacted positively on blots with the monoclonal podocalyxin antibody 3D3 (Fig. 5D, lane 1), PNA lectin (Fig. 5E, lane 1), TRA-1-81 antibodies (Fig. 5F, lane 1), and TRA-1-60 antibodies (Fig. 5G, lane 1). All four blots show a similar diffuse 200-kDa MW band that is not detected in any of the control purification fractions (Fig. 5D–5G, lane 2). Thus, the 200-kDa MW form of podocalyxin from three distinct EC cell lines—TERA-1, NT2.Dl, and NCCIT—has binding activity with the TRA-1-60 and TRA-1-81 antibodies.
The data reported here further support our hypothesis that podocalyxin is a human stem cell marker on human EC that carries the TRA-1-60 and TRA-1-81 pluripotent stem cell antigens. Podocalyxin is a member of a family of three related sialomucins (membrane proteins with large amounts of O-linked glycosylation within their mucin domains and containing large quantities of negatively charged sialic acid residues) that include CD34 and endoglycan . CD34 is a differentiation marker of hematopoietic cells. It is expressed on immature progenitor blood cells but is lost during blood cell differentiation . CD34 is also a ligand for the leukocyte adhesion molecule L-selectin, and has been shown to have a role in leukocyte extravasation . Likewise, podocalyxin is expressed on vascular endothelium and hematopoietic cells, and also binds to l-selectin [4, , –7]. In the high endothelial venules, podocalyxin is postulated to function as an adhesion molecule for leukocyte extravasation .
Podocalyxin was originally identified as a major structural extracellular matrix sialoglycoprotein of the glomerular podocytes [1, 2]. These highly differentiated epithelial cells have interdigitating foot processes that form the filtration slits over the glomerular basement membrane. The integrity of the slits is crucial for proper blood filtering, and is maintained, in part, by podocalyxin. It is postulated that the negatively charged podocalyxin protein functions in the extracellular glycocalyx of the glomerulus as an antiadhesion molecule that provides structural support to the podocyte slits [34, , –37]. The putative topology of podocalyxin is an integral membrane protein with one transmembrane domain, a small cytoplasmic domain, and a large extracellular domain containing five putative N-linked glycosylation sites, three putative glycosaminoglycan sites, and a 270-amino acid mucin domain. The mucin domain has high serine/threonine content for putative O-glycosylation modifications. The calculated molecular weight of podocalyxin—a 528-amino acid protein—is 55 kDa, but the apparent molecular weight of the human kidney version of podocalyxin is 165–180 kDa, and the difference in molecular weight is due to post-translational glycosylation .
The results presented herein indicate that there are at least two versions of podocalyxin expressed in human cells. The first is a 170-kDa MW form expressed in human kidney tissue, the human embryonic kidney cell line HEK293, mammalian cells transfected with the podocalyxin gene, and NT2.D1 and NCCIT EC cells treated with RA for 7–12 days. The second podocalyxin version is an embryonic or stem cell form with a higher molecular weight of 200 kDa on protein gels and blots and expressed in human EC pluripotent cells. The 200-kDa MW form of podocalyxin expressed in the NT2.Dl and NCCIT EC cells is modified (as defined by size and TRA-1-60/TRA-1-81 reactivity) with exposure with retinoic acid to the EC cells. The size differences between the two forms of podocalyxin must be due, in part, to glycosylation differences indicated by the change in reactivity to the carbohydrate-binding antibodies TRA-1-60 and TRA-1-81. It is possible that the glycosylation differences may be due to glycosaminoglycan modifications of the EC form of podocalyxin. Podocalyxin expressed in the kidney is reported to be heavily glycosylated, but contains no glycosaminoglycan modifications; however, the glycoprotein that has been identified on EC as the common carrier of the TRA-1-60/TRA-1-81 antigens is reported to be modified with keratan sulfate glycosaminoglycan groups . This could explain the increased molecular weight seen with the 200-kDa MW form of podocalyxin. Also, although CD34 does not have glycosaminoglycan modifications, endoglycan has been reported as a proteoglycan, so glycosaminoglycans modifications do occur in this family of sialomucins . Nevertheless, little is actually known about the glycosylation of this core protein. The identification of podocalyxin as a carrier molecule of the TRA-1-60 and TRA-1-81 epitopes will allow for further studies into the molecular structure of these carbohydrate-based stem cell markers.
In addition to its role as a human stem cell marker, podocalyxin is a marker of malignancy. We have previously identified podocalyxin as a testis tumor marker . Since this finding, several other research groups have documented a role for podocalyxin in cancer development. Podocalyxin has been found to be an independent marker for human breast cancer progression and is hypothesized to contribute through its adhesion/antiadhesion properties to a breast cell metastasis phenotype . Given this finding, it is worth noting that studies have shown that PNA reactivity on surface glycoproteins is associated with a high metastasis potential . It will be of interest to determine whether PNA reactivity of podocalyxin expressed on breast tumors has a correlation with the malignancy potential of the tumor. Another study shows that podocalyxin is a target gene of the Wilms' tumor suppressor and is repressed by tumor suppressor protein p53, thus implying a role for podocalyxin in Wilms' tumors . Podocalyxin has also been identified as a marker of hepatocellular carcinoma and a marker of blasts in acute leukemia [19, 20].
It remains to be determined whether the 200-kDa MW form of podocalyxin expressed in testis cancer EC cells is also present in other types of cancers in which podocalyxin has been detected. However, the finding that podocalyxin carries the TRA-1-60 antigen has important implications in human testis cancer. There have been several reports showing compelling data that TRA-1-60 is a potentially clinically useful serum marker [38, –40]. The identification of TRA-1-60 on podocalyxin provides two very important pieces of information to further study the TRA-1-60 serum tumor marker. First, this finding provides an explanation as to why TRA-1-60 is found on the surface of testis cancer and as a soluble antigen in the serum of patients with testis cancer. Podocalyxin is in a family of membrane proteins that are known to be shed from the surface of the cell by proteolysis into the extracellular milieu, and CD34 is a well-studied example of this process . Second, this finding provides a putative molecular identity of the testis cancer serum marker, which will allow further studies of podocalyxin as a clinical marker of testis cancer. Along these lines, it will be of interest to look at the serum levels of podocalyxin in the other nontestis cancers in which podocalyxin has been discovered.
The findings presented herein support our hypothesis that the 200-kDa MW form of podocalyxin is a carrier of the pluripotent stem cell makers, TRA-1-60 and TRA-1-81. We herein refer to this form of podocalyxin as SC-podocalyxin—for stem cell podocalyxin. At this time, we can not rule out the possibility that, in the three EC cell lines we examined, podocalyxin is specifically interacting within a protein complex with another glycoconjugate that carries the TRA-1-60/TRA-1-81 epitopes and, thus, copurifies with podocalyxin and is detected on blots with an apparent 200-kDa MW similar to podocalyxin. Further studies to determine what other proteins podocalyxin is interacting with in pluripotent stem cells, structural studies of the carbohydrate modifications and the extracellular domain of podocalyxin, and identification of the cellular machinery responsible for the synthesis of the TRA epitopes should provide more insight into this protein complex possibility, and also shed more light onto the role podocalyxin has in the biology of pluripotent stem cells. Also, it is possible that the TRA-1-60 and TRA-1-81 epitopes are expressed on other glycoconjugates within EC—we have not eliminated this possibility, nor has it yet been determined whether podocalyxin identified in the established human ES cell lines is SC-podocalyxin carrying TRA-1-60 and TRA-1-81 antigens. If SC-podocalyxin is expressed in ES cell lines, this would be further evidence supporting the emerging theory that ES cell lines are derived from early germ cells within the inner cells of the blastocysts; if this theory is correct, it would link EC and ES stem cell lines as being derived from the same source—germ cells—and expressing the same stem cell marker—SC-podocalyxin .
We believe that SC-podocalyxin is a true marker of human pluripotent stem cells. Several lines of evidence support this statement. First, podocalyxin is closely related in protein sequence and structure to a bona fide stem cell marker, CD34. Second, SC-podocalyxin is expressed in all human pluripotent EC cell lines and carries the TRA-1-60 and TRA-1-81 stem cell markers, which are widely used in human stem cell research as positive indicators of a true pluripotent human stem cell. Third, the TRA-1-60 and TRA-1-81 epitopes are lost from podocalyxin once the EC cells are treated with RA, and the loss of TRA-1-60/TRA-181 reactivity on the surface of the stem cell is one of the established properties that define stem cell differentiation. Fourth, podocalyxin has been detected in eight of eight human ES cell lines studied, with the conclusion that it is not only highly expressed in the ES lines, but also one of a handful of proteins that define undifferentiated ES cells. Fifth, podocalyxin is a marker of blood stem cells. Finally, the SC-podocalyxin form has not been reported in any differentiated or somatic tissues or cell lines to date.
The authors indicate no potential conflicts of interest.
We thank Dr. David Kershaw (University of Michigan, Ann Arbor, MI) for podocalyxin antibodies (3D3 and 4F10) and the podocalyxin expression vector. We thank Justine Curley (Beth Israel Deaconess Medical Center, Boston, MA) for excellent technical assistance and advice with the immunofluorescence studies.