Rituxan (anti-CD20 antibody)-induced translocation of CD20 into lipid rafts is crucial for calcium influx and apoptosis

Abstract

Rituxan, a chimeric anti-CD20 antibody, is the first antibody approved for immunotherapy in non-Hodgkin's B-cell lymphoma and other B-cell lymphoproliferative disorders. Additionally, efficacy of Rituxan treatment has been reported in nonmalignant autoimmune diseases such as rheumatoid arthritis. Crosslinking of CD20 molecules by Rituxan induces therapeutic B-cell depletion. CD20 is a B-lymphocyte specific integral membrane protein, proposed to function as a store-operated calcium channel, which is activated upon receptor-stimulated calcium depletion of intracellular stores. Crosslinking of CD20 by antibodies has been reported to induce a redistribution of CD20 molecules to specialized microdomains at the plasma membrane known as lipid rafts. Here, we report that in the absence of Rituxan, CD20 exhibits a low affinity to lipid rafts. However, binding of Rituxan significantly increases the affinity of CD20 for lipid rafts resulting in its redistribution to a fraction resistant to Triton X-100 solubilization. Furthermore, we demonstrate that disturbing the raft integrity by cholesterol extraction results in dissociation of CD20 from a Triton X-100 resistant fraction followed by complete inhibition of Rituxan-induced calcium entry and apoptosis. The integrity of lipid rafts seems to play a crucial role for CD20-induced caspase activation. These data show, for the first time, that Rituxan-induced translocation of CD20 to lipid rafts is important for increased intracellular Ca²⁺ levels and downstream apoptotic signalling.

Introduction

Rituxan, a chimeric anti-CD20 antibody (Rituximab, Mabthera, C2B8), is in clinical use for non-Hodgkin's B-cell lymphoma and has also shown excellent efficacy in inducing clinical improvement and remission in rheumatoid arthritis patients [1–4]. The effectiveness of Rituxan-based therapy is achieved by B-cell depletion. Several mechanisms have been proposed to be responsible for the therapeutic activity of Rituxan, including antibody-dependent cell cytotoxicity (ADCC), activation of the complement system, and CD20-mediated regulation of the cell cycle and apoptosis [5–7]. Crosslinking CD20 with anti-CD20 monoclonal antibodies like Rituxan, 2H7 and 1F5 triggers cell-cycle block at the G1 phase and inhibits in vitro B-cell differentiation and EBV or pokeweed mitogen-induced Ig secretion [8].

CD20 belongs to the MS4A gene family, which consists of at least 25 members clustered on human chromosome 11q12–13·1 [9,10]. The MS4A family has a predicted tetraspanning membrane topology with an N- and C-terminal cytoplasmic domain. CD20 is the best studied member of this family and is specifically expressed on the surface of B-cells and cells from most B-cell lymphoproliferative disorders [11]. Different isoforms of CD20 (33, 35, 37 kD) result from multiple phosphorylation of serine and threonine residues in the cytoplasmic domains, implying that CD20 is highly regulated by phosphorylation.

Stimulation of the B-cell receptor induces depletion of intracellular calcium stores which in turn results in the activation of store-operated calcium channels at the plasma membrane. A sustained influx of extracellular calcium ensures the progression of calcium-dependent signalling processes such as transcriptional control, cell cycle progression or apoptosis. The induction of apoptosis is blocked by chelating intracellular or extracellular calcium [12,13]. Studies using cell lines transfected with CD20 show an increased calcium conductance across the plasma membrane, strongly suggesting that CD20 functions as a calcium channel important for regulating cell cycle progression and calcium homeostasis [14,15]. Furthermore, it was reported that reduced expression levels of CD20 in B-cell lines, achieved by antisense CD20 sequence, result in a significantly decreased calcium entry across the plasma membrane [15,16]. These results provide the first evidence that CD20 functions as a store-operated calcium channel [17]. However, the mechanism of how the decrease in luminal calcium concentration causes an activation of store-operated calcium entry at the plasma membrane is still not understood. Hypercrosslinking of CD20 antibodies bound to the cell surface results in an increase in calcium conductivity without preceding depletion of intracellular calcium stores, uncoupling the store-operated channel activity from regulation via intracellular calcium levels [14]. Binding of antibodies to CD20 is also reported to cause a rapid redistribution of CD20 molecules to lipid rafts, which represent specialized microdomains of the plasma membrane, highly enriched in sphingolipids and cholesterol [18]. Lipid rafts are implicated in the organization of numerous membrane-associated signalling pathways providing a platform for the scaffolding of messenger molecules [19,20]. Truncation of the CD20 cytoplasmic domain (Δ219-225) abolished CD20 lipid raft association and significantly decreased the calcium influx downstream of B-cell receptor-stimulated calcium mobilization from intracellular stores [15].

The current study was initiated to investigate the role of CD20 lipid raft localization for CD20 calcium channel activity by directly crosslinking CD20 by Rituxan. The data described here provides evidence that agents disturbing the raft integrity inhibit Rituxan-induced translocation of CD20 into lipid rafts as well as Rituxan-induced calcium influx and subsequent caspase-mediated apoptosis.

Materials and methods

Cells, antibodies and reagents

Ramos B cells were maintained in culture in RPMI 1640 supplemented with foetal calf serum (10%), HEPES (100 m m), sodium pyruvate (1 m m), and l-glutamine (2 m m). Rituxan (chimeric anti-CD20 IgG1) and chimeric irrelevant antibody were purchased from Adallen Pharma Ltd, hypercrosslinking goat anti-human IgG F(ab′)₂ from Sigma-Aldrich. Anti-human IgM F(ab′)₂ was purchased from Biosource. For immunoblotting, polyclonal anti-CD20 was kindly provided by Julie Deans et al. (University of Calgary) polyclonal anti-Lyn was purchased form Santa Cruz Biotechnology, Horseradish peroxidase (HRP)-conjugated Cholera toxin subunit B (CTB) from Sigma-Aldrich, and goat HRP-conjugated anti-rabbit IgG from Beckman Coulter. AnnexinV-Fluos was purchased from Roche, propidium iodide and Fluo-3 AM from Molecular Probes. Triton X-100 and Brij 58 were purchased from Pierce. Methyl-β-cyclodextrin (MCD) and cholesterol-loaded MCD were purchased from Sigma-Aldrich. Poly D-lysine coated black 384 microtiter plates were purchased from Perbio. Apo-ONE caspase 3/7 assay was from Promega.

Cell stimulation and lipid raft isolation

Ramos cells (5 × 10⁸ cells per gradient) were incubated with Rituxan (70 µg per gradient) in RPMI medium at 37°C for 10 min and lysed for 1 h either by 1% Brij 58 at room temperature, or by 1% Triton X-100 on ice in Tris-buffer (20 m m Tris-HCl pH 7·4, 150 m m NaCl, 1 µg/ml aprotinin, 1 µg/ml leupeptin, 1 m m sodium vanadate, 2 m m glycerolphosphate). To isolate lipid rafts, lysates were mixed with an equal volume of 80% sucrose in Tris-buffer supplemented with 0·1% detergent and overlaid with 35% sucrose and 5% sucrose in Tris-buffer supplemented with 0·1% detergent. After centrifugation at 200 000× g for 15 h, 1 ml fractions were collected and analysed by immunoblotting according to standard procedure. For GM1 dot blot, aliquots from each fraction were spotted onto nitrocellulose (Schleicher & Schuell) and Cholera toxin subunit B was used for detection. The signal was detected by enhanced chemiluminiscence (ECL reagents, Santa Cruz). To deplete cholesterol, cells were treated with 1% MCD for 1 h at 37°C prior to antibody incubation. To re-establish lipid raft integrity, cells were treated with cholesterol-loaded MCD (MCD to cholesterol 25 : 1 weight ratio) for 1 h.

Ca²⁺-influx

Ramos cells (1·25 × 10⁶ cells/ml) were loaded with Fluo-3AM (4 µm) in assay buffer (10 m m HEPES, pH 7·4, 150 m m NaCl, 2·5 m m KCl, 10 m m glucose, 1·2 m m MgCl₂, 1·5 m m CaCl₂, 250 µm sulfinpyrazone) and seeded into a Poly D-lysine coated black microtiter plate at a density of 5 × 10⁴ cells per well. After 1 h incubation at 37°C the cells were stimulated with Rituxan (0–25 µg/ml) for 10 min at 37°C, washed twice in assay buffer including brilliant black (0·04%) and treated with anti-human IgG F(ab′)₂ (30 µg/ml). Immediately after the addition of the secondary antibody, quantitative changes in the intracellular Ca²⁺ were monitored by FLIPR (Molecular Devices, λ_Ex: 488 nm, λ_EM: 540 nm) [21]. To deplete cholesterol, cells were preincubated 1 h at 37°C in the presence of different concentrations of MCD (0–2%) or for control in the presence of cholesterol-loaded MCD (0–2%). For data analysis, Ca²⁺ response was measured as peak fluorescence intensity minus basal fluorescence intensity. All data are expressed as means of at least n = 5.

Apoptosis assays

Ramos cells (1·25 × 10⁶ cells/ml) were stimulated with Rituxan (25 µg/ml) in RPMI 1640 at 37°C for 10 min, washed twice in medium in order to remove excess of unbound Rituxan, and treated with anti-human IgG F(ab′)₂ (30 µg/ml) in the presence of MCD (0; 0·12%; 0·25%) and/or EGTA (5 m m). Cells (1·25 × 10⁵ cells per well) were seeded into uncoated black microtiter plates and after 22 h incubation at 37°C activity of caspase 3 and 7 was assayed by cleavage of a rhodamine 110 labelled peptide (Z-DEVD-R110) using the Apo-ONE homogenous caspase 3/7 assay kit according to the manufacturer's instructions (Promega) (λ_Ex: 498 nm, λ_EM: 521 nm). Alternatively, the incidence of apoptotic cells was assessed by staining with FITC-labelled annexin V. Antibody-stimulated cells (1·25 × 10⁶ cells per well) were incubated at 37°C for 7 h, washed twice in assay buffer (10 m m HEPES pH 7·4, 150 m m NaCl, 5 m m CaCl₂) and incubated in 100 µl labelling solution (assay buffer with annexin V-Fluos according to the manufacturer's instructions and propidium iodide (10 µg/ml)) for 10 min on ice. The samples were diluted with 400 µl assay buffer and analysed by flow cytometry. Single Annexin V stained cells were scored as apoptotic cells, double stained cells as necrotic.

Results

A high-affinity raft association of CD20 is dependent on binding of Rituxan and cholesterol presence

Lipid rafts can be distinguished from the bulky membrane by their insolubility in different detergents. However, the amount of proteins embedded into lipid rafts and the size of lipid rafts purified on sucrose gradients can vary considerably dependent on the detergent used and solubilization temperature [22]. Here, we compared two different detergents for their ability to solubilize CD20. Using cold extraction by Triton X-100 for membrane solubilization, no significant amount of CD20 was detected in the raft fractions. However, binding of Rituxan to CD20 significantly increased the affinity of CD20 to lipid rafts as shown by the acquired resistance towards solubilization by Triton X-100 (Fig. 1a). Hyper-crosslinking of Rituxan by anti-human IgG F(ab′)₂ was not required and did not further increase CD20 raft association (results not shown). After solubilization at room temperature with Brij 58, a CD20 subpopulation was obtained which was already associated with the low density sucrose fractions even in the absence of Rituxan and crosslinking by Rituxan increased the amount of CD20 concentration in rafts (Fig. 1a). Comparing both extraction procedures, we observed that not all CD20 molecules moved into Brij 58 resistant lipid rafts after Rituxan treatment. Performing the Brij 58 extraction at room temperature allows increased mobility of proteins shuttling between raft and nonraft area, which might explain why in the Brij 58 raft fraction a subpopulation of CD20 molecules are found outside rafts, which was not observed with the cold Triton X-100 extraction. We further demonstrated that CD20 raft association is dependent on cholesterol presence as extraction of cholesterol from the plasma membrane by MCD resulted in complete dissociation of CD20 molecules from the Triton X-100 resistant fractions at a MCD concentration that did not cause cell cytotoxicity (Fig. 1b). The association of Lyn (or GM1) with rafts in the presence of 1% MCD was only slightly affected, indicating that Lyn associates with a higher affinity to lipid rafts than CD20. Subsequent treatment of the cells with cholesterol-loaded MDC re-established CD20 raft localization, demonstrating that the presence of CD20 in lipid rafts is cholesterol-dependent.

Fig. 1.

(a) CD20 has a high affinity for lipid rafts after Rituxan binding. Lipid rafts were isolated from Ramos cells after treatment either by Rituxan or by an irrelevant antibody (– RTX). After the cells were treated with detergent (Triton X-100 at 4°C or Brij 58 at room temperature), lysates were fractionated by sucrose density gradient centrifugation and fractions were analysed by immunoblot with anti-CD20, anti-Lyn, and anti-GM1 (cholera subunit B). Data are representative of 3 experiments. (b) Rituxan–stimulated association of CD20 to lipid rafts is dependent on raft integrity. To deplete cholesterol, cells were incubated with 1% MCD and treated as above. The effect of MCD was reversed by cholesterol loading.

Rituxan-induced association of CD20 to lipid rafts is crucial for Ca²⁺-influx and downstream signalling

Previous studies showed that Rituxan in combination with hypercrosslinking antibody induces Ca²⁺-influx across the plasma membrane bypassing CD20 activation through store-depletion. In this context, we set up a FLIPR-based technique to follow changes in intracellular calcium concentrations in Ramos cells upon antibody stimulation. As the affinity of CD20 to lipid rafts was increased upon Rituxan stimulation, we investigated whether CD20 raft association is a prerequisite for CD20-mediated calcium influx. Cholesterol was extracted from the plasma membrane by MCD treatment prior to antibody incubation. Incubating the cells for 1 h with 0·12% MCD reduced Rituxan-induced calcium influx to 50% as compared to the untreated control (Fig. 2a). Rituxan-induced calcium-influx was reduced to a background level when increasing MCD concentration to 0·25% and above. The effect of MCD was reversed when cells were treated with cholesterol-loaded MCD (Fig. 2b). Furthermore, we observed that in the presence of cholesterol-loaded MCD the Ca²⁺-response to Rituxan was above the signal of untreated cells. This suggests that increasing concentrations of exogenous cholesterol might induce the formation of additional lipid rafts inducing further clustering of CD20. Hypercrosslinking of surface-bound Rituxan was essential for this experimental setup as cells treated with Rituxan alone showed very small Ca²⁺ influx. The crosslinking antibody anti-human IgG F(ab′)₂ was tested for Ca²⁺-stimulation and was shown to have no effect (Fig. 2c). For the positive control, Ca²⁺ signal was monitored after IgM crosslinking. Crosslinking of IgM consists of a biphasic Ca²⁺-signal, an initial peak fluorescence caused by the release of Ca²⁺ from intracellular stores followed by a more sustained Ca²⁺ signal mediated through store-operated calcium channels at the plasma membrane. When compared to IgM-stimulated Ramos cells, the Rituxan-induced peak fluorescence (exemplified for 12·5 µg/ml RTX and 30 µg/ml hypercrosslinking antibody) was reproducibly in the range of 40% to 50% of the IgM (12·5 µg/ml) induced Ca²⁺-response. The rise and decline of the initial phase of the Ca²⁺-signal induced by crosslinking the B-cell receptor was more rapid and steeper as compared to the Ca²⁺-influx achieved by RTX/IgG (Fig. 2c). In the presence of cholesterol the Ca²⁺ signal induced by RTX/IgG was two-fold elevated and more sustained than in the absence of cholesterol (Fig. 2c). MCD at the concentration used here (up to 1%) did not induce cell lysis or damage as investigated by microscopy and flow cytometry of propidium iodide stained cells (data not shown). Furthermore, the cell number in all samples was counted and was not influenced by MCD. MCD treatment had no inhibitory effect on Fluo-3 AM uptake. For detergent-based lipid raft isolation, a much higher cell number (5 × 10⁸ per gradient) was used, requiring higher concentration of MCD (1%) to disturb CD20 lipid raft association. Cell integrity was confirmed as above.

Fig. 2.

Lipid raft integrity is crucial for Ca²⁺-influx mediated by CD20. Rituxan-induced store-operated Ca²⁺ entry in Fluo-3 AM loaded Ramos cells was monitored by FLIPR. Ramos cells were incubated 1 h at 37°C either with Fluo-3AM alone (untreated control), or in combination with (a) MCD or (b) cholesterol-loaded MCD (MDC, C). The cells were then stimulated with different concentrations of Rituxan (0–25 µg/ml), washed twice (to remove the excess of unbound Rituxan in order to avoid neutralization of the hypercrosslinking antibody) and activated by hypercrosslinking antibody (30 µg/ml). The concentration of the hypercrosslinking antibody had been optimized in previous experiments (data not shown). Representative traces are shown as quantitative changes in intracellular Ca²⁺ concentrations expressed as peak fluorescence minus basal fluorescence. Data shown are representative of 3 independent experiments with n = 5 each. (c) The kinetics of the Ca²⁺-influx traces are shown. The addition of antibody is indicated by an arrow (time of addition = 20 s). For the positive control the cells were stimulated with IgM (12·5 µg/ml). Representative kinetic traces in the presence of RTX (12·5 µg/ml) and hypercrosslinking IgG (30 µg/ml) are depicted for 3 different conditions (no MCD, 0·5% MCD, 0·5% MCD, cholesterol). For the negative control, cells were treated with IgG alone (30 µg/ml).

As Rituxan is known to induce apoptosis, our next step was to determine the effect of CD20 raft association on apoptosis. When Ramos cells were exposed to Rituxan in the presence of crosslinking antibody, more than 60% of the cells were annexin V positive stained, indicating significant apoptotic progression (Fig. 3a). The apoptotic induction was dependent on calcium as chelating calcium by EGTA completely inhibited this process. Disturbing the raft integrity by cholesterol depletion resulted in a significant reduction of annexin V positive cells, indicating that CD20 raft association is important for the initiation of CD20-mediated apoptotic signalling (Fig. 3b). In all samples the amount of necrotic cells, as assayed by double staining with propidium iodide and annexin V, was below 7%. The incubation with MCD alone did not induce apoptotic event per se, demonstrating that the MCD concentrations used did not have a cell damaging effect. As caspases have been reported to be essential effector molecules in the apoptotic process upon Rituxan treatment [7,12,13], we investigated whether their activation is dependent on raft integrity. Rituxan-stimulated Ramos cells showed an increased activity of caspase 3 and 7 which peaked after 22 h incubation. The caspase activity was reduced to background levels in the presence of MCD (0·25%), indicating that raft integrity is crucial for Rituxan-stimulated apoptosis (Fig. 3c). MCD alone did not influence caspase 3/7 activities. As previously reported, chelating calcium also reduced caspase activity [12]. We could not detect a physical association of caspase 3 with lipid raft fractions by Western blot analysis of sucrose fractions in contrast to previous data [23]. This suggests that downstream effector molecules such as caspases are activated by second messengers. Incubating the cells in the presence of MCD did not induce spontaneous apoptosis. Cholesterol add back experiments were not carried out in the long term incubation assays (22 h), as the presence of foetal calf serum in the medium interferes with cholesterol.

Fig. 3.

Lipid raft integrity is crucial for Rituxan-induced apoptosis. Ramos cells were preincubated with Rituxan (25 µg/ml), washed and cultured with hypercrosslinking antibody (30 µg/ml) in the presence of MCD (0·12% or 0·25%) with and without EGTA (5 m m). (a) After 7 h of incubation, apoptotic cells were scored by single staining of Annexin V-Fluos on a flow cytometer. Propidium iodide and Annexin V double stained cells were interpreted as necrotic and were below 7% under experimental conditions. (b) Summary of flow cytometry data. (c) Alternatively, after 22 h cells were assayed for caspase-3 and -7 activity by cleavage of fluorescent labelled substrate peptide (Z-DEVD-R110). Activation of caspases correlates with an increase in fluorescent units (λ_Ex: 498 nm, λ_EM: 521 nm). Data are representative of 2 independent experiments with n = 3 each.

In summary, the data described here correlate Rituxan-stimulated moblilization of CD20 into lipid rafts with an increased calcium influx across the plasma membrane and subsequent apoptosis.

Discussion

Although antibody-induced CD20 lipid raft association is well documented, only a few studies are available investigating the impact of CD20 lipid raft localization on CD20-dependent signalling [24]. Recent data strongly suggest that CD20 functions as a store-operated calcium channel, regulated by the status of the intracellular calcium stores [15]. We investigated calcium influx across the plasma membrane after direct crosslinking of CD20 molecules at the cell surface by Rituxan, thereby uncoupling the store-operated channel activity from regulation via intracellular calcium levels.

In the absence of anti-CD20 antibody, CD20 molecules are constitutively associated with a low-affinity to lipid rafts. This association is sensitive to solubilization by Triton X-100, but not by Brij 58. These results are consistent with the observation made by Li and coworkers comparing the solubilization of CD20 in different detergents [15,24]. Interestingly, binding of Rituxan to CD20 results in a high affinity of CD20 to lipid rafts as shown by an acquired insolubility of CD20 to Triton X-100. Taking the model of Pizzo & Viola [22] into account, we conclude that after antibody binding, CD20 molecules preferentially migrate from the border of a raft area (which is resistant to Brij 58 solubilization and sensitive to Triton X-100) to a nucleus raft area (which is resistant to Triton X-100 solubilization). Therefore, the local concentration of CD20 molecules increases in a restricted area of the plasma membrane, thus facilitating oligomerization of CD20 to form a functional channel complex. Alternatively, Rituxan binding to CD20 could result in smaller lipid rafts’ merging to form larger ones harbouring a higher concentration of CD20 molecules. Monovalent Fab fragments of Rituxan are sufficient to induce high affinity association of CD20 to lipid raft (data not shown) [24]. This observation emphasizes that crosslinking of two adjacent epitopes by Rituxan is not essential for clustering of CD20 molecules inside rafts, but that instead the Fab fragments stabilize a conformation of CD20 which exhibits a higher affinity for rafts. The higher affinity association of CD20 to rafts might create a suitable environment for CD20 activation providing close proximity to raft-localized kinases [25,26].

In this paper we provide, for the first time, evidence that Rituxan-induced ‘high–affinity’ raft association of CD20 is a prerequisite for calcium entry and downstream apoptotic signalling. By disturbing the raft integrity Rituxan-induced calcium influx was reduced to background levels. Replenishing the pool of plasma membrane cholesterol re-established both calcium influx and CD20 association to rafts. In line with this observation, Li and coworkers recently made the observation that calcium influx in BCR-stimulated cells was significantly reduced by cholesterol depletion, suggesting that the lipid raft environment of CD20 is important for channel activity [15]. Furthermore, deletion of a membrane-proximal cytoplasmic sequence of CD20, which was shown to control CD20 raft association, abolished calcium entry downstream of BCR. Interestingly, we observed with increasing concentrations of cholesterol an elevation in the Ca²⁺-response, indicating that exogenous cholesterol might induce the formation of additional lipid rafts. Although soluble Rituxan was sufficient to induce ‘high affinity’ association of CD20 with lipid rafts (as shown by Western blot analysis), the calcium influx induced by Rituxan alone was too small to be measured and to ensure apoptotic onset. Extensive crosslinking of surface-bound Rituxan most likely exceeded the cytosolic calcium threshold required for the progression of apoptosis. Hypercrosslinking of Rituxan might also induce a merger of neighbouring rafts creating a CD20 signalling platform. This implies that CD20 raft localization is a crucial step in CD20 activation, but other factors like src-family kinases might be involved in modification of CD20 activity as, e.g. phosphorylation of CD20 might stabilize an active CD20 channel conformation. The phosphorylation status of CD20 has been shown to be dependent on stimulation of B-cells [25]. Interestingly, during the submission process of this manuscript, a Rituxan-induced mobilization of Erk and its subsequent phospohorylation has been reported [27].

Numerous studies have described the involvement of calcium homeostasis in apoptosis and an increased intracellular calcium concentration has been correlated with the induction of caspase 3 activity [13,28]. CD20-mediated apoptosis is reported to be inhibited by EGTA or BAPTA, which blocks changes in cytosolic calcium, or by selective inhibitors of src-family kinases [13,29]. In our studies progression of CD20-mediated apoptosis, as assayed by annexin V positive cells and activity of caspase 3 and 7, was significantly reduced when the integrity of lipid rafts was disturbed.

In contrast to our studies, Chan et al. [30] reported that CD20-induced cell death is independent of both caspases and CD20 redistribution to lipid rafts. The authors described that CD20 can evoke apoptosis without involvement of caspase 3, which is in disagreement to previous data [12,31]. The authors also excluded the relevance of lipid rafts to CD20 signalling which is in contrast to our own and other groups’ studies [15,24]. Chan et al. [30] based their conclusion on the observations with B1, an anti-CD20 antibody with an overlapping binding epitope to Rituxan. In contrast to Rituxan, B1 is unable to mobilize CD20 to lipid rafts [6,24], but is still able to induce apoptosis. However, recent data from the same group described that B1-mediated cell death lacks TUNEL positivity and DNA fragmentation and so does not reflect classical apoptosis but instead is more reminiscent of necrosis [32]. In our studies B1 is unable either to mobilize CD20 to lipid rafts or to stimulate a calcium influx or to induce apoptosis (data not shown). These data support the notion that Rituxan and B1 induce intracellular signalling in a different way. Not all anti-CD20 induced cellular responses can be attributed to lipid rafts. For instance, homotypic aggregation of cells which is strongly induced by B1, but not by Rituxan is not dependent on lipid raft integrity.

A role for lipid rafts in the regulation of other ion channels has been reported. For instance, in platelet cells expressing the store-operated Ca²⁺ channel TRPC1, treatment with MCD inhibits Ca²⁺-conductance implying a correlation between raft recruitment of TRPC1 and its activity [33]. The integrity of lipid rafts is also required for the Kv 1·3 channel activity, since a change in the composition of rafts by cholesterol extraction inhibited channel activity [34,35].

In this report, we focused our studies on Rituxan, the therapeutically most relevant anti-CD20 antibody. On stimulation with Rituxan, CD20 molecules migrate to the nucleus of rafts, which leads to a conformational changes in the CD20 complex and/or induces a close proximity with other raft-localized molecules such as src-family kinases, resulting in an increased calcium conductivity and induction of apoptosis.