CD24 Induces Apoptosis in Human B Cells Via the Glycolipid-Enriched Membrane Domains/Rafts-Mediated Signaling System

Cells and reagents

CD24⁺ BL-derived cell lines P32/SH and Namalwa were obtained from the Japanese Cancer Research Resources Bank (Tokyo, Japan) and Institute for Fermentation (Osaka, Japan), respectively. The CD24⁻ BL line Daudi from the Japanese Cancer Research Resources Bank was also used. Cells were cultured in RPMI 1640 supplemented with 10% FCS at 37°C in a humidified 5% CO₂ atmosphere.

The mouse mAbs used were anti-CD24 (ALB-9) and anti-CD48 (J4.57) from Coulter/Immunotech (Westbrook, MA), anti-CD48 (5-4.8) from Santa Cruz Biotechnology (Santa Cruz, CA), anti-CD55 (IA10) and anti-CD59 (p282H19) from PharMingen (San Diego, CA), anti-phosphotyrosine (4G10) from Upstate Biotechnology (Lake Placid, NY), anti-Lyn (clone 42) and anti-ERK1 (MK12) from Transduction Laboratories (Lexington, KY), anti-GM3 (M2590) from Wako Pure Chemical Industries (Osaka, Japan), and anti-β-actin from Seikagaku (Tokyo, Japan). Anti-CD3 (OKT-3), anti-CD45 (9.4), and anti-μ (DA4.4) from American Type Culture Collection (Manassas, VA) and anti-CD24 (L30) (22) were also used. The rabbit polyclonal Abs used were anti-Lyn from Serotec (Oxford, U.K.); anti-CD55, anti-Csk, and anti-Blk from Santa Cruz Biotechnology; anti-phospho-specific mitogen-activated protein (MAP) kinase, anti-phospho-specific stress-activated protein (SAP) kinase, anti-SAP kinase, anti-phospho-specific p38 MAP kinase, and anti-p38 MAP kinase from New England Biolabs (Beverly, MA); anti-G_iα from Calbiochem-Novabiochem (San Diego, CA); and anti-μ from Jackson ImmunoResearch (West Grove, PA). The goat anti-cholera toxin B subunit (CT-B) polyclonal Ab from Calbiochem was also used. Secondary Abs, including fluorescein-conjugated and enzyme-conjugated Abs, and fluorescein-conjugated avidin were purchased from either Jackson ImmunoResearch or Molecular Probes (Eugene, OR). Biotinylation of mAb was performed as described previously (23). Conjugation of Abs to polystyrene beads was performed as described previously (39). To cross-link CD24, cells were cultured in the presence of Abs as indicated. In these cases, Abs were dialyzed in PBS before addition to culture to remove additives. The recombinant CT-B was obtained from Calbiochem. All chemical reagents were obtained from Sigma-Aldrich (St. Louis, MO) unless otherwise indicated.

Immunofluorescence study

The cells were stained with FITC-labeled mAbs and analyzed by flow cytometry (EPICS-XL, Coulter) as described previously (23). In some experiments cells were stained with a combination of Alexa Fluor 488-labeled anti-CD24 mAb and Alexa Fluor 543-labeled either anti-Lyn, anti-Blk, or anti-Csk polyclonal Ab, then examined by confocal microscopy (GB-200, Olympus, Tokyo, Japan). Briefly, after a 30-min incubation with anti-CD24 mAb (L30) on ice, cells were washed intensively with ice-cold PBS and incubated with Alexa Fluor 488-conjugated goat anti-mouse polyclonal Ab for 30 min on ice. After intensive washing with ice-cold PBS, cells were either left on ice as an unpatched control or incubated for 30 min at 37°C to induce patching as described previously (10, 40). After intensive washing and fixation with 3% paraformaldehyde in PBS, cells were further stained with a combination of rabbit anti-Lyn, anti-Blk, or anti-Csk polyclonal Ab followed by Alexa Fluor 543-conjugated goat anti-rabbit polyclonal Ab and examined using the appropriate filter set. Because Alexa Fluor 488-conjugated goat anti mouse Ab absorbed any cross-reactivity, no cross-reaction against rabbit Ab was detected (data not shown). Similarly, Alexa Fluor 543-conjugated goat anti-rabbit Ab showed no cross-reactivity against mouse Ab (data not shown). No cross-talk was detected between Alexa Fluor 488 and Alexa Fluor 543 channels (data not shown).

Immunoblotting, immunoprecipitation, and immunocomplex kinase assay

Immunoblotting was performed as described previously (41). Briefly, cell lysates were prepared by solubilizing the cells in lysis buffer (containing 20 mM Na₂PO₄ (pH 7.4), 150 mM NaCl, 1% Triton X-100, 10 trypsin inhibitory units/ml aprotinin, 5 mM PMSF, 100 mM NaF, and 2 mM Na₃VO₄). After centrifugation, supernatants were obtained, and the protein concentration of each cell lysate was determined with a Bio-Rad protein assay kit (Bio-Rad, Hercules, CA). Fifty micrograms of each cell lysate was electrophoretically separated on an SDS-polyacrylamide gel and transferred to a nitrocellulose membrane using a semidry Transblot system (Bio-Rad). After blocking, the membranes were incubated with the appropriate combination of primary and secondary Abs as indicated, washed intensively, then examined using the enhanced chemiluminescence reagent system (ECL, Amersham, Aylesbury, U.K.). The results obtained from a 1-min exposure of the ECL-treated membrane to film are presented.

For immunoprecipitation, 500 μg of the cell lysates were incubated with 1 μg of the indicated Ab and 30 μl of 50% protein G-agarose (Roche, Mannheim, Germany) for 1 h. After intensive washing, the immunoprecipitates were separated by electrophoresis and analyzed as described above. Immunocomplex kinase assay was also performed using enolase, as exogenous substrate, as described previously (41).

Detection of apoptosis

To detect apoptosis, cells were assayed for DNA ladder formation by gel electrophoresis. After treatment of cells as indicated, DNA was extracted from the cells, separated by 1% agarose gel electrophoresis, and examined under UV light as described previously (37). To quantitate the incidence of apoptotic cells, cells were stained with FITC-labeled annexin V using the MEBCYTO Apoptosis Kit (Medical & Biological Laboratories, Nagoya, Japan) and then analyzed by flow cytometry according to the manufacturer’s protocol.

Sedimentation of the GEM fraction and lipid analysis

The GEM fractions were prepared by linear sucrose gradient centrifugation as described previously (40, 42). Briefly, 1 × 10⁸ cells were lysed and homogenized in Tris-buffered saline (25 mM Tris (pH 7.5) and 150 mM NaCl) containing 1% Triton X-100, 1 mM PMSF, 1 mM Na₃Vo₄, and aprotinin (10 trypsin inhibitory units/ml). Cell extracts were adjusted to 40% sucrose in Tris-buffered saline. A linear gradient (5–30% sucrose in Tris-buffered saline) was formed above the lysate and centrifuged at 39,000 rpm for 18 h at 4°C in a Beckman SW40Ti rotor (Palo Alto, CA). In some experiments cells were biotinylated before homogenization as described previously (42). The lipids extracted from the gradient fractions as described previously (40, 42) were separated on a high performance TLC plate (Merck, Darmstadt, Germany), followed by blotting to polyvinylidene difluoride membrane (Millipore, Bedford, MA) and immunostained with anti-GM3 mAb as described previously (40, 42).

CD24 expression in BL-derived cell lines

We first examined whether BL cell lines express CD24. Among eight BL lines tested, including P32/SH, Namalwa, Ramos, Daudi, CA-46, Jiyoyie, EB-3, and SCC-3, only the first two lines expressed CD24. As shown in Fig. 1⇓, both P32/SH and Namalwa lines expressed high levels of CD24 by flow cytometry. In addition, other GPI-anchored proteins, including CD48, CD55, and CD59 (Fig. 1⇓), but not CD90 (data not shown), were expressed on both lines. Flow cytometric analysis also revealed that both lines showed a similar pattern of expression of B cell differentiation Ags, including surface IgM (sIgM; Fig. 1⇓), CD19, CD20, CD72, CD80, CD81, and HLA-DR (data not shown). A GC B cell-specific marker, CD10 (20, 21, 23, 36), and a protein tyrosine phosphatase, CD45, were also expressed on both lines (Fig. 1⇓). We used these two CD24⁺ BL lines in the following experiments.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Flow cytometric analysis of surface Ags expressed on BL cell lines. P32/SH and Namalwa cells were stained with specific mAbs against GPI-anchored proteins and other molecules, as indicated in the figure, and analyzed by flow cytometry. The histograms obtained were displayed. x-axis, Fluorescence intensity; y-axis, relative cell number; CNT, isotype-matched mouse Ig.

Cross-linking of CD24 by specific Abs induces apoptosis in BL-derived cell lines

Next we tested whether cross-linking of CD24 mediated by specific Abs induces any physiological responses in BL cells. When P32/SH cells treated with a combination of anti-CD24 mAb, L30, and rabbit secondary polyclonal anti-mouse Ig Ab for 24 h were examined morphologically by light microscopy, a portion of cells exhibited cleaved nuclei, a finding typical of apoptosis (Fig. 2⇓A). No such figures were observed in untreated cells (data not shown). The following results clearly show that cross-linking of CD24 indeed induced apoptosis in BL cells. First, DNA prepared from P32/SH cells on which CD24 is cross-linked showed oligonucleosomal ladder fragmentation on agarose gel electrophoresis (Fig. 2⇓B). Second, cleavage of the nuclear DNA was also confirmed by detection of subploid cells with propidium iodide staining (data not shown). Third, there was a significant increase in cells binding to annexin V after CD24 cross-linking by immobilized anti-CD24 mAb on polystyrene beads (Fig. 2⇓C).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Cross-linking of CD24 induces apoptosis in P32/SH BL cells. A, P32/SH cells were cultured in the presence of anti-CD24 mAb L30 (5 μg/ml) and rabbit anti-mouse Ig (5 μg/ml) for 24 h, then cytocentrifuged and Giemsa-stained, and morphological appearance was examined by light microscopy. Typical apoptotic cells showing characteristic cleaved nuclei are indicated by arrowheads. Magnification, ×400. B, DNA ladder formation after cross-linking of CD24 was examined. P32/SH cells were treated with (+; lane 2) and without (−; lane 1) anti-CD24 mAb (α CD24) as in A. The extracted DNA (1.5 μg/lane) from each sample was separated by 1% agarose gel electrophoresis. C, Detection of annexin V-binding cells. After 24-h culture in the presence of 5 μg/ml of either anti-CD24 mAb (α CD24, L30) or isotype-matched mouse Ig (CNT) immobilized on the polystyrene beads, P32/SH cells were incubated with FITC-conjugated annexin V, then analyzed by flow cytometry. A typical histogram obtained from each experiment was displayed. Experiments were performed in triplicate, and the mean ± SD percentages of annexin V-bound cells are indicated. x-axis, Fluorescence intensity; y-axis, relative cell number.

As presented in Fig. 3⇓A, when P32/SH cells were treated with the anti-CD24 mAb L30 alone for 24 h, <10% of the cells underwent apoptosis as assessed by annexin V binding. However, in the presence of the secondary rabbit polyclonal anti-mouse Ig Ab L30 induced the appearance of a significant number of annexin V-bound cells (Fig. 3⇓A), suggesting that strong cross-linkage of CD24 is required to induce apoptosis in these cells. The fact that either a combination of biotinylated L30 and streptavidin (Fig. 4⇓B) or L30 immobilized on polystyrene beads (Fig. 2⇑C) also induced a comparable level of apoptosis further supports the above idea. We also examined the effect of ALB-9, another anti-CD24 mAb, and observed identical results (Fig. 3⇓A). These phenomena were specific to CD24, because anti-CD45, which binds to BL cells, did not induce apoptosis (Fig. 3⇓A). Isotype-matched control mouse Ig (Fig. 2⇑C) or anti-CD3 did not induce apoptosis (Fig. 3⇓A). We similarly examined Namalwa cells and observed mostly identical results (Fig. 3⇓B). By contrast, when CD24⁻ Daudi BL cells, in which signaling for apoptosis can occur by BCR cross-linking, were similarly examined, treatment with a combination of L30 and rabbit anti-mouse Ig Ab failed to induce apoptosis (Fig. 3⇓C). Thus, these data suggest that sufficient CD24 cross-linking specifically induces apoptosis in CD24⁺ BL cells, and that this is not due to the nonspecific binding of either mouse Ig or secondary rabbit polyclonal anti-mouse Ig Ab.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Induction of apoptosis in BL cells by cross-linking of CD24. A, CD24⁺ P32/SH cells (2.5 × 10⁵ cells in 500 μl of medium) were treated or not treated (columns 1 and 2) for 24 h with 5 μg/ml of specific mAb against either CD3 (α CD3, OKT-3, does not bind to P32/SH cells, columns 3 and 4), CD45 (α CD45, 9.4, binds to P32/SH cells, columns 5 and 6), or CD24 (L30, columns 7 and 8; ALB-9, columns 9 and 10) in the presence (columns 2, 4, 6, 8, and 10) and absence (columns 1, 3, 5, 7, and 9) of 5 μg/ml of secondary rabbit anti-mouse Ig Ab (α MsIg). After staining with FITC-conjugated annexin V, the incidence of apoptotic cells was examined on flow cytometry. Experiments were performed in triplicate, and the mean ± SD numbers of cells that bound annexin V, i.e., those undergoing apoptosis, are presented. Identical results were obtained from at least three independent experiments. B, CD24⁺ Namalwa cells were also examined in the same manner as in A. C, As a negative control, CD24⁻ Daudi cells were examined as described in A (columns 1–4). To confirm that signaling for apoptosis can occur in the cells through another pathway, Daudi cells were treated with anti-μ heavy chain mAb (α μ, DA4.4, 5 μg/ml) in the presence (column 6) and the absence (column 5) of α MsIg (5 μg/ml).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

A synergism between the cross-linking of CD24 and that of BCR on apoptosis-inducing effect in BL cells. A, P32/SH cells treated (columns 2–4 and 6–8) and not treated (columns 1 and 5) with suboptimal doses of anti-CD24 mAb (α CD24, L30, 0.5 μg/ml) and anti-μ heavy chain mAb (α μ, DA4.4, 0.1 μg/ml) as indicated in the presence (columns 5–8) or the absence (columns 1–4) of 5 μg/ml of secondary rabbit anti-mouse Ig Ab (α MsIg) for 24 h were examined as described in Fig. 3⇑. To confirm that the cross-linking of BCR with a full dose (F) of anti-μ heavy chain mAb can induce apoptosis, P32/SH cells were treated with 5 μg/ml of DA4.4 (columns 9 and 10) in the presence (column 10) or the absence (column 9) of α MsIg. B, P32/SH cells were treated (columns 3, 4, and 6) or not treated (columns 1, 2, and 5) with a suboptimal dose of biotin-conjugated anti-CD24 mAb L30 (5.0 μg/ml) in the presence (columns 2, 4, and 6) or the absence (columns 1, 3, and 5) of 1 μg/ml of avidin. To test the effect of a suboptimal dose of anti-μ heavy chain mAb, DA4.4 (0.1 μg/ml) was also added (columns 5 and 6). After 24-h incubation, cells were examined as described in Fig. 3⇑. To confirm that a full dose of biotinylated L30 can sufficiently cross-link CD24 to induce apoptosis, P32/SH cells were also treated with 20 μg/ml of biotinylated L30 in the presence (column 8) or the absence (column 7) of avidin.

Effect of cross-linking of another GPI-anchored protein on BL cells

Next we examined whether the induction of apoptosis is specific for CD24 or is a common feature of GPI-anchored proteins expressed on these BL cells. As shown in Fig. 5⇓, when P32/SH cells were treated with a combination of anti-CD48 mAb and secondary rabbit anti-mouse Ig, a number of annexin V-binding cells comparable to that obtained with anti-CD24 mAb was induced. In contrast, the mAbs against either CD55 or CD59 could only induce a low number of annexin V-binding cells, even in the presence of secondary rabbit anti-mouse Ig Ab (Fig. 5⇓). These data suggest that induction of apoptosis upon cross-linking is a common feature of GPI-anchored proteins expressed on these BL cells, while the magnitude of cell death is variable in each case.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Induction of apoptosis in BL cells by cross-linking of other GPI-anchored proteins. P32/SH cells treated or not treated (columns 1 and 2) with 5 μg/ml of specific mAbs against GPI-anchored proteins (CD48, columns 3 and 4; CD55, columns 5 and 6; CD59, columns 7 and 8) in the presence (columns 2, 4, 6, and 8) or the absence (columns 1, 3, 5, and 7) of 5 μg/ml of secondary rabbit anti-mouse Ig Ab (α MsIg) for 24 h were examined as described in Fig. 3⇑.

Synergism between cross-linking of CD24 and that of BCR on the effect of apoptosis induction in BL cells

It was reported that sufficiently strong cross-linking of BCR also induces apoptosis in BL cells (37, 38). Therefore, we tested the synergism between cross-linking of BCR and that of CD24 on the induction of apoptosis. Although the DA4.4 mAb against human μ heavy chain was able to induce marked apoptosis in P32/SH cells at full dose (5 μg/ml), it did not induce apoptosis at a suboptimal dose (0.1 μg/ml) even in the presence of secondary rabbit anti-mouse Ig Ab (Fig. 4⇑A, compare columns 3, 7, 9, and 10). Similarly, a suboptimal dose (0.5 μg/ml) of anti-CD24 mAb, L30, also failed to induce significant apoptosis in P32/SH cells (Fig. 4⇑A, columns 2 and 6). However, when suboptimal doses of both DA4.4 and L30 were mixed, a significant level of apoptosis was induced in P32/SH cells with the help of secondary rabbit anti-mouse Ig Ab (Fig. 4⇑A, columns 6–8). The data indicate a synergism between cross-linking of CD24 and that of BCR on the effect of apoptosis induction in BL cells.

To further examine whether cross-linking between CD24 and BCR is necessary for the synergistic effect on apoptosis induction, we used the biotin-avidin system. As presented in Fig. 4⇑B, a combination of a full dose of biotinylated L30 (20 μg/ml) and avidin (1 μg/ml) is sufficient to induce apoptosis in P32/SH cells (columns 7 and 8). Although a suboptimal dose of biotinylated L30 (5 μg/ml) induced a low level of apoptosis in the presence of avidin (Fig. 4⇑B, columns 3 and 4), simultaneous addition of a suboptimal dose (0.1 μg/ml) of DA4.4 remarkably enhanced the induction of apoptosis (Fig. 4⇑B, columns 5 and 6). Based on these results, we concluded that independent cross-linking of each molecule is sufficient to induce synergistic induction of apoptosis and that cross-linking between CD24 and BCR is unnecessary.

Distribution of CD24 and other molecules in linear sucrose gradient centrifugation

GPI-anchored proteins are thought to be localized in the GEM domain. To investigate the participation of GEM in CD24-mediated apoptosis, we examined the distribution of CD24 in linear sucrose gradient centrifugation. When cell lysates prepared from biotinylated P32/SH cells were fractionated, most of the biotinylated proteins were liberated with 1% Triton X-100 from the plasma membrane and thus recovered in fractions 11–13 as high density detergent-soluble fractions (DS) (Fig. 6⇓A). Some were not solubilized from plasma membrane and thus were recovered in buoyant, low density fractions 5–9, which include the GEM (Fig. 6⇓A). By immunoprecipitation, most CD24 molecules were recovered in buoyant, low density fractions 6 and 7 (Fig. 6⇓B). Consistent with previous reports, immunoprecipitated CD24 molecules appeared as a broad band on SDS-PAGE gels (5, 16, 43), because of their extensive glycosylation (13, 18). By contrast, when CD24⁻ Daudi BL cells were similarly examined, anti-CD24 mAb L30 failed to precipitate any biotinylated proteins from either the GEM or DS (data not shown), suggesting the specificity of this mAb. In contrast, CD45 molecules, which have been reported to be found predominantly in the DS (10, 44), were immunoprecipitated only from the DS in both P32/SH cells (Fig. 6⇓B) and Daudi cells (data not shown). These findings indicate that the failure to immunoprecipitate CD24 from the DS is not due to disruption of immunoprecipitation in the presence of high sucrose concentrations and that CD24 molecules are indeed distributed only in the GEM fractions.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Sucrose density gradient analysis of CD24 and other proteins from Triton X-100 lysates. A, Triton X-100 lysates obtained from biotinylated P32/SH cells were centrifuged on linear sucrose density gradients. After centrifugation, 1-ml gradient fractions were collected from the top to the bottom to yield a total of 13 fractions (lanes 1–13 are compatible to fractions 1–13). Gradient fractions were separated on 10% SDS-polyacrylamide gels, transferred to nitrocellulose membranes, and stained with HRP-conjugated avidin. B, P32/SH cells were treated (Cross-link, α CD24, middle panel) and not treated (−, top and bottom panels) with anti-CD24 mAb (L30) immobilized on polystyrene beads for 5 min. After biotinylation, Triton X-100 lysates were obtained and fractionated as described in A. The proteins from each fraction were immunoprecipitated with anti-CD24 mAb (IP, α CD24, L30, top and middle panels). After separation by 10% SDS-PAGE gel electrophoresis, the proteins were detected as described in A. As a control for the protein predominantly found in the dense fraction, immunoprecipitation of each fraction with anti-CD45 mAb (IP, α CD45, bottom panel) was performed in the same manner. C, Gradient fractions obtained from the P32/SH cell lysates in B were separated on 10% SDS-PAGE gels, and Lyn, Blk, and Csk proteins were detected by immunoblotting with rabbit polyclonal Ab for each kinase, as indicated. D, In the first and second panels from the top, lipids extracted from gradient fractions obtained in B were separated on a TLC plate and were then blotted onto a polyvinylidene difluoride membrane. Immunostaining with anti-GM3 mAb (GM3) was performed. In the other panels, gradient fractions obtained from the P32/SH cell lysates in B were separated on 10% SDS-PAGE gels, and G_iα (α-Gαi), CD48, and CD55 proteins were detected by immunoblotting as indicated.

We also examined the distribution of cellular proteins after sucrose density gradient centrifugation by immunoblotting; the majority of cellular proteins, including Blk, Csk (Fig. 6⇑C), and Syk (data not shown), were only recovered in DS. In contrast, approximately half of all Lyn protein, an Src family PTK, was recovered in the GEM (Fig. 6⇑C). Furthermore, immunoprecipitation revealed that a portion of Lyn was detected in anti-CD24 immunoprecipitates from the GEM fractions (Fig. 7⇓). Because no Lyn was detected in control mouse Ig immunoprecipitates from the same fractions (data not shown), the results indicate that CD24 and Lyn are specifically associated in the GEM.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Coimmunoprecipitation of Lyn with CD24 from the GEM fractions. The immunoprecipitates with anti-CD24 mAb (α CD24, L30) were prepared from five to nine fractions separated from P32/SH cell lysates as described in Fig. 6⇑B. The Lyn (A) and CD24 (B) proteins in each immunoprecipitate were detected as described in Fig. 6⇑, C and B, respectively. Immunoprecipitation was also performed with anti-Lyn mAb in duplicate (C and D). Detection of Lyn proteins in immunoprecipitates and assessment of the enolase transphosphorylation activity of Lyn were performed by immunoblotting (C, Blot, α Lyn) and immunocomplex kinase assay (D, ICK), respectively.

Next we examined whether the distribution of various proteins in linear sucrose gradient centrifugation may change after CD24 cross-linking. As shown in Fig. 6⇑B, when cell lysates prepared from P32/SH cells treated with the anti-CD24 mAb L30 immobilized on polystyrene beads were fractionated, the distribution of CD24 shifted to a higher density site, and it was recovered mainly in fractions 7 and 8. It is noteworthy that the distribution of Lyn in sucrose gradient centrifugation also shifted similarly (Fig. 6⇑C). In contrast, cross-linking of CD24 did not affect the other proteins, including Blk, Csk (Fig. 6⇑C), Syk, or CD45 (data not shown), all of which remained in the DS. No such shift was induced by anti-CD3 or anti-CD45 mAb immobilized on polystyrene beads (data not shown). From these results, mobility shift of CD24 and Lyn by CD24 cross-linking was regarded to be a phenomenon specific to CD24 and was not due to nonspecific binding of Ab.

To assess the biochemical basis of the difference between GEM from untreated and CD24 cross-linked cells, we examined the distribution of the other presumably GEM-associated molecules in sucrose gradient centrifugation before and after treatment with anti-CD24 mAb. GSL is a major constituent of the GEM domain. In P32/SH cells, GM3 was found to be the most abundantly expressed GSL as assessed by lipid analysis (data not shown). As shown in Fig. 6⇑D, TLC blotting analysis showed that the majority of GM3 was recovered in the low density GEM fractions. A dually acylated heterodimeric G protein α subunit, G_iα, is known to reside in GEM (45, 46, 47), and indeed, immunoblotting analysis revealed that G_iα was recovered only in low density fractions 5–7 after sucrose gradient centrifugation (Fig. 6⇑D). Other GPI-anchored proteins, CD48 and CD55, were also recovered only in GEM (Fig. 6⇑D). However, when the cells were CD24 cross-linked the distribution of all these molecules shifted to a higher density site, and after sucrose gradient centrifugation they were recovered in fractions 6–8 (Fig. 6⇑D), the same as CD24 (Fig. 6⇑B). The results indicate that all GEM-associated molecules show the same distribution shift after sucrose gradient centrifugation following CD24 cross-linking.

We further examined whether the association between CD24 and Lyn in the GEM fractions was affected by CD24 cross-linking. As shown in Fig. 7⇑C, the total amount of Lyn protein precipitated with anti-Lyn mAb did not change significantly after CD24 cross-linking. However, the recovery of Lyn in anti-CD24 immunoprecipitates was markedly increased after CD24 cross-linking (Fig. 7⇑A). Moreover, CD24 cross-linking resulted in the enhanced kinase activity of Lyn in GEM fractions as revealed by increased transphosphorylation of enolase, an exogenous substrate, by in vitro kinase assay (Fig. 7⇑D).

Colocalization of CD24 and Lyn, but not Blk, in cell membrane after cross-linking of CD24

Because biochemical data have indicated that CD24 is closely correlated with Lyn, but not with Blk or Csk, in the GEM domain, we next examined the correlation between CD24 and those PTKs in the cell membrane by using immunostaining and confocal microscopy. In cells left on ice throughout the staining process, CD24 and those PTKs were evenly distributed (Fig. 8⇓, B, D, and F). However, when the cells were incubated at 37°C, CD24 became concentrated to distinct patches within the membrane (Fig. 8⇓, A, C, and E). Significantly, Lyn was found to clearly colocalize with the CD24-patches as presented in a overlay picture in Fig. 8⇓A. In contrast, Blk and Csk did not colocalize with the CD24 patches (Fig. 8⇓, C and E). These results are consistent with the biochemical data and indicate that Lyn, but not Blk and Csk, is preferentially colocalized with CD24 patches after Ab-mediated CD24 cross-linking.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

Cocapping of Lyn, but not Blk and Csk, with CD24 on P32/SH cells after treatment with anti-CD24 mAb. P32/SH cells stained with Alexa 488-labeled anti-CD24 mAb L30 were either left on ice as an unpatched control (Unpatched; B, D, and F) or incubated for 30 min at 37°C to induce patching as described in Materials and Methods (CD24-patched; A, C, and E). After fixation, cells were incubated with rabbit anti-PTK polyclonal Abs as indicated and further labeled with Alexa 543-conjugated goat anti-rabbit Ab. Each fluorescence was visualized by confocal microscopy using an appropriate filter set. Typical results from each treatment were presented. For each experiment, the pictures obtained by staining with either Alexa 488 or Alexa 543 were superimposed (Overlay).

Translocation of BCR in the GEM fraction after CD24 cross-linking

Upon cross-linking the BCR has been shown to move into GEM in mouse and human B cells (44, 47, 48, 49). Therefore, we examined whether the same was true in the P32/SH cells used in the present study. As shown in Fig. 9⇓, when cell lysates prepared from P32/SH cells were fractionated, μ heavy chain was recovered only in DS after sucrose gradient centrifugation. However, after BCR cross-linking by treating cells with the anti-μ heavy chain mAb immobilized on polystyrene beads, a small portion of μ heavy chain was translocated in low density fractions 5 and 6 (Fig. 9⇓). G_iα, a marker protein for GEM, was also recovered in fractions 5–7 regardless of BCR cross-linking (Figs. 6⇑D and 9). The results suggest that BCR cross-linking induces recruitment of a fraction of BCR itself in the GEM fraction, but that it does not affect the distribution of GEM on sucrose gradient centrifugation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 9.

Sucrose density gradient analysis of μ heavy chain from Triton X-100 lysates. Triton X-100 lysates obtained from P32/SH cells treated and not treated (Cross-link, −, top panel) with anti-CD24 mAb (α CD24, L30, second panel from the top) or anti-μ heavy chain mAb (α μ, DA4.4, third and fourth panels from the top) immobilized on polystyrene beads for 5 min were fractionated as described in Fig. 6⇑A. Gradient fractions were separated on 10% SDS-polyacrylamide gels, and μ heavy chain (Blot, α μ) and G_iα (Blot, α Gαi) proteins were detected by immunoblotting.

Because we observed synergism between cross-linking of CD24 and BCR on the effect of apoptosis induction, next we examined whether CD24 cross-linking also affects the distribution of the μ heavy chain by sucrose gradient centrifugation. As shown in Fig. 9⇑, CD24 cross-linking also resulted in translocation of a small portion of the μ heavy chain in the low density fractions. However, in contrast to BCR cross-linking the translocated μ heavy chain was recovered in fractions 6 and 7, where the other GEM-associated molecules were located after CD24 cross-linking (Fig. 6⇑, B and D).

Treatment with CT-B induces apoptosis in BL cells

As demonstrated in this study, cross-linking of GPI-anchored proteins that reside in GEM induces apoptosis in BL cells. We also investigated whether the induction of apoptosis is specific to GPI-anchored proteins or is a common feature of GEM-resident molecules. To test this, we selected CT-B. CT-B has been well characterized as a specific ligand for GSL GM1 that resides in the GEM domain (10, 39). Cross-linking of GEM with CT-B has been reported to induce intracellular signaling events in T cells via a GEM-mediated mechanism (10, 39). Flow cytometric analysis revealed binding of CT-B to P32/SH cells (Fig. 10⇓A), indicating that this cell line expresses GM1. As shown in Fig. 10⇓B, when P32/SH cells were treated with CT-B alone for 24 h, the percent increase in apoptotic cells, assessed by annexin V binding, was not significant. However, in the presence of the goat anti-CT-B Ab, CT-B induced a significant increase in number of annexin V-bound cells (Fig. 10⇓B), indicating that sufficient cross-linking of GM1 also can induce apoptosis in BL cells. We also attempted to determine whether GM3 cross-linking with specific Ab also induces apoptosis of BL cells, but flow cytometric analysis revealed that the anti-GM3 mAb we used for TLC blotting in this study does not recognize the extracellular portion of GM3 and thus is not useful for cross-linking (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 10.

Induction of apoptosis in BL cells by treatment with CT-B. A, P32/SH cells were reacted with biotinylated CT-B, incubated with FITC-conjugated avidin, and analyzed by flow cytometry. The histograms obtained (CT-B, dark line) were superimposed on that of the negative control (CNT, cells stained with FITC-avidin alone, light line) and displayed. x-axis, fluorescence intensity; y-axis, relative cell number. B, P32/SH cells (2.5 × 10⁵ cells in 500 μl of medium) were treated (columns 3 and 4) or not treated (columns 1 and 2) with 5 μg/ml of CT-B in the presence (columns 2 and 4) and the absence (columns 1 and 3) of 5 μg/ml of polyclonal goat anti-CT-B Ab (α CT-B) for 24 h. After staining with FITC-conjugated annexin V, the proportion of apoptotic cells was determined by flow cytometry. Experiments were performed in triplicate, and the mean ± SD of the cells that bound annexin V, i.e., those undergoing apoptosis, are shown. Identical results were obtained from at least three independent experiments.

Intracellular signaling mediated by cross-linking of CD24

In the cases of other cell types, it has been reported that cross-linking of CD24 as well as other GPI-anchored proteins results in intracellular signaling events such as an activation of protein kinases accompanied by an increased level of tyrosine phosphorylation of intracellular proteins (1, 2, 4, 6, 12). Thus, we attempted to observe similar changes in BL cells expressing CD24. Inconsistent with the previous reports, no significant change in tyrosine phosphorylation level of total cellular proteins by CD24 cross-linking detected in immunoblotting analysis using anti-phosphorylated tyrosine Ab (data not shown). However, when individual molecules were examined by immunoblotting using phospho-specific Abs, clear elevations of phosphorylations of ERK1, SAP kinase, and p38 MAP kinase were detected (Fig. 11⇓), suggesting that CD24 cross-linking results in the activation of these kinases.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 11.

Early phosphorylation of MAP kinases in P32/SH cells mediated by CD24 cross-linking. Protein lysates were prepared from P32/SH cells treated (lanes 2–5) or not treated (lane 1) with anti-CD24 mAb (α CD24, immobilized L30 on polystyrene beads, 5 μg/ml). A 50-μg sample of each cell lysate was electrophoretically separated on 10% SDS-PAGE gels and immunoblotted with specific Abs against MAP kinases as indicated (A–C). As a quantitative control for amounts of protein in samples loaded, anti-β-actin (α β-actin) mAb was also examined (D). P-MAPK, phospho-MAP kinase; P-SAPK, phospho-SAP kinase; P-p38, Phospho-p38 MAP kinase.