The lipid-modulating effects of a CD4-specific recombinant antibody correlate with ZAP-70 segregation outside membrane rafts

Myriam Chentouf^a,†, Maxime Rigo^a,†, Soufiane Ghannam^a, Isabelle Navarro-Teulon^a, Sébastien Mongrand^b, André Pèlegrin^a and Thierry Chardès^a,*

Read Also

The classical concept of plasma membrane, proposed by Singer and Nicolson [1], wherein proteins diffuse freely in two-dimensional homogeneous bi-layers, has been drastically modified during the last decade. Indeed, membrane rafts, which are “discrete” domains with a mean diameter of 10–200 nm, can be distinguished from the rest of the membrane due to their protein and lipid composition [2]. The ability of membrane rafts to segregate proteins in a defined lipid environment provides a mechanism for signaling compartmentalization in the plasma membrane by concentrating some components in membrane rafts and excluding others. Cholesterol/sphingolipids confer organization of membrane rafts through self-assembly but could also be incorporated in the quaternary structure of raft-located protein complexes (“lubrication” concept) [2], thus favoring their inclusion into and the assembly of functionalized membrane rafts. They play a major role in the modulation of apoptosis [3] and cell growth [4], and their targeting represents an attractive strategy of raft-based therapeutics [5, 6]. Indeed, cholesterol sequesters and inhibitors of cholesterol synthesis as well as sphingolipid-modulating drugs, which mainly act on enzymes involved in ceramide metabolism [5], regulate cell growth, differentiation, stress response and apoptosis [7]. Ceramide accumulation occurs through two main pathways: hydrolysis from sphingomyelin through sphingomyelinase (SMase) stimulation and de novo synthesis by ceramide synthase activation. Interestingly, the anti-CD20 antibody rituximab [4] as well as CD40- [8] and CD95-specific [3, 9, 10] antibodies increase in vitro ceramide release, thus rendering cells sensitive to apoptosis or inhibition of proliferation. In vivo therapeutic synergy between rituximab and lipids modulators, such as fenretinide [11, 12] and aplidine [13], has been demonstrated, thus leading to the idea of combining lipid modulators and antibodies, as proposed in a combinatorial phase I/II clinical trial in B cell lymphoma (trial NCT00288067). Altogether, these results clearly emphasized a dynamic crosstalk between sphingolipids/cholesterol and proteins (lipid-protein “rheostat”) in membrane rafts that can be modulated by antibodies [5].

We found that, in Jurkat T cells, rIgG₁ 13B8.2, a baculovirus-expressed recombinant IgG₁ (rIgG₁) anti-CD4 antibody [14, 15], induced accumulation/retention of CD4 inside membrane rafts, recruitment of TCR, CD3z, kinases, adaptor proteins and PKC-q, but excluded ZAP-70 and its downstream targets SLP-76, PLCg1, and Vav-1 [16]. Analysis of key upstream events showed that modulation of ZAP-70 Tyr²⁹² and Tyr³¹⁹ phosphorylation occurred concomitantly with rIgG₁ 13B8.2-induced ZAP-70 exclusion from the membrane rafts [16]. Such antibody-induced modulation of membrane raft signaling, which leads to inhibition both of NF-kB nuclear translocation [17] and of binding to the IL2 gene promoter [18], partly explains the anti-proliferative effect of rIgG₁ 13B8.2 in T cell lymphomas [19]; however the effects of rIgG₁ 13B8.2 on lipid dynamics in membrane rafts remain unknown.  Therefore, we decided to examine the effects of the anti-CD4 rIgG₁ 13B8.2 antibody on the lipid composition of membrane rafts in a T lymphoma cell line.  Here we report that, besides CD4/ZAP-70 protein reorganization, rIgG₁ 13B8.2 affected the lipid rheostat by increasing ceramide release through acid SMase activation and decreasing phosphatidylserine synthesis without modifying the cholesterol content of GM1-positive membrane rafts. Finally, incubation of Jurkat T cells with exogenous SMase not only increased ceramide release, but also segregated ZAP-70 kinase outside GM1⁺ membrane rafts like following treatment with rIgG₁ 13B8.2.

Changes in lipid metabolism, which are closely connected to changes in lipid membrane composition, can dramatically affect the localization and function of membrane raft resident proteins [34]. For example, modifications of fatty acids in membrane rafts modulate ZAP-70 phosphorylation and consequently suppress signal transduction in T cells [35]. In a previous work we showed that, in Jurkat T cells, the anti-CD4 antibody rIgG₁ 13B8.2 induces accumulation of CD4 inside membrane rafts and exclusion of the ZAP-70 kinase and its downstream targets SLP-76, PLCg1 and Vav-1 from the raft machinery [16]. We now report that rIgG₁ 13B8.2 affects the lipid rheostat in GM1⁺ membrane rafts of Jurkat T cells by increasing ceramide release through acid sphingomyelinase activation and by decreasing the level of phosphatidylserine without modifying the cholesterol content. These effects are correlated with ZAP-70 exclusion from membrane rafts.

            Cells undergo physiological turnover through induction of apoptosis and phagocytic removal, partly through externalization of phosphatidylserines from the cytosolic leaflet to the outer leaflet of the membrane. Phosphatidylserine externalization in cells undergoing death receptor-mediated apoptosis seems to be Ca⁺⁺-dependent [36]. Our finding that rIgG₁ 13B8.2 lowers phosphatidylserine level in Jurkat T cells explains why treatment with rIgG₁ 13B8.2  did not induce phosphatidylserine-dependent apoptosis in T cell lymphomas [19] and blocked CD3-induced Ca⁺⁺ increase and the subsequent signaling pathways [33]. Similarly, other CD4-specific antibodies have been shown to modulate phosphatidylserine level in vitro [37].

            We also report that rIgG₁ 13B8.2 treatment induces a time-dependent increase in ceramide production through acid SMase activation. Activation of the sphingomyelin/SMase/ceramide axis can be induced by radiations, chemotherapeutic agents or receptor ligands [3, 4, 38] and, in some cases, a biphasic ceramide release was observed with a first short-term (1-10 minutes) phase [3, 4, 39, 40] followed by a later one [39]. Such biphasic cycle was also reported following neutral SMase activation [41]. We thus hypothesize that rIgG₁ 13B8.2, which triggers CD4/ZAP-70 raft reorganization mainly during the first 30-60 seconds post-treatment [16], induces ceramide release first from the membrane sphingomyelin pool through direct acid SMase activation in the outer leaflet of the membrane, and then stimulates a second ceramide burst either through de novo ceramide synthesis [42], or through PKC-mediated phosphorylation of lysosomal acid SMase leading to ceramide release [43]. The inhibition observed with the SMase inhibitor imipramine argues in favor of the implication of the sphingomyelin/SMase/ceramide axis, but additional experiments are needed to assess whether the imipramine-induced inhibition affects cell signaling. The inability of the myriocin and fumonisin B1 inhibitors to block ceramide release suggests that de novo ceramide synthesis is probably not activated by rIgG₁ 13B8.2 treatment, as previously reported for other antibodies [4, 38]. The role of antibody-mediated activation of intracellular lysosomal acid SMase remains to be clarified. Interestingly, antibody-triggered CD40 raft clustering [8] and ultra-violet radiations [40] also induce translocation of acid SMase from the intracellular pool leading to ceramide release in membrane rafts [8]. Ultra-violet radiations as well as chemotherapeutic agents [44] stimulate PKC-d-mediated phosphorylation of intracellular acid SMase leading to remodeling of the cellular cytoskeleton and further acid SMase translocation to the membrane. Since rIgG₁ 13B8.2 modulates PKC-q distribution together with CD4 [16], PKC-q could participate in the activation/phosphorylation of lysosomal acid SMase in Jurkat T cells following treatment with rIgG₁ 13B8.2.

            Exogeneous SMase was reported to induce ceramide release in CD3-stimulated T cells concomitantly with inhibition of Ca⁺⁺ flux [32] and inhibition of IL2 production through inhibition of NF-kB activity [45]. Ceramide also blocked IL2 production in T lymphoma cells through PKC-q-mediated, but not TNF-a-induced, NF-kB activation [45]. Sphingomyelin appears to be a critical raft constituent that enables translocation of signaling molecules, such as ZAP-70 and PKC-q, as demonstrated in sphingomyelin synthase 1 knockdown cells [46].  Here we demonstrate that ZAP-70 is excluded from membrane rafts of Jurkat T cells following treatment with bacterial SMase, probably through sphingomyelin consumption. The use of sphingomyelin-deficient cells could probably clarify this point. Similarly, treatment with rIgG₁ 13B8.2 induces ceramide release (this work) and inhibits CD3-induced T cell activation and proliferation through blockade of intracellular Ca⁺⁺ flux, NF-kB activation and IL2 secretion [15, 33]. These events have been associated with CD4/PKC-q translocation inside rafts and ZAP-70 exclusion from membrane rafts [16]. All these findings indicate that the lipid-protein rheostat in membrane rafts can be modulated by therapeutic antibodies in order to physically structure signaling platforms through sphingomyelin consumption/ceramide-induced aggregation and to maximize and synergize their anti-tumoral effects through dynamic lipid/protein partitioning inside/outside rafts.

The signaling pathways, through which CD4/ZAP-70/PKC-q modulation and SMase-induced ceramide release could synergize, remain to be elucidated in rIgG₁ 13B8.2-treated cells. Two upstream PKC-q effectors (PDK1 [47] and ADAP [48]) and one downstream effector (the adapter protein Carma1 [49]) could play a role because they are involved in the PKC-q-mediated NF-kB pathway [47―49] and ceramide inhibits PKC-q-mediated NF-kB activity [45]. PDK1 recruits both PKC-q and Carma1 to membrane rafts, phosphorylates PKC-q, thereby regulating NF-kB activation [47]. PDK1 also phosphorylates PKC-d [50], which in turn activates lysosomal acid SMase through phosphorylation [43], leading to ceramide release. Moreover, sphingosine, a hydrolysis product of ceramide, can activate PDK1 by autophosphorylation [51]. ADAP binds to Carma1 and regulates NF-kB activation [48] but also binds to PI3K-generated phosphoinositides [52] which activate PDK1 activity [50]. Finally Carma1, upon phosphorylation mediated by PKC-q, bridges membrane-proximal events and nuclear signaling leading to NF-kB activation, and also organizes protein translocation in membrane rafts [53]. We can thus hypothesize that, upon CD4 targeting by rIgG₁ 13B8.2, ZAP-70 is physically excluded from membrane rafts, this exclusion being correlated with acid SMase-mediated sphingomyelin consumption for ceramide synthesis, thus inhibiting NF-kB activity. ADAP/PDK1 proteins are concomitantly recruited to membrane rafts together with CD4 and PKC-q [16], enabling PI3K activation, thereby producing phosphoinositides and further activating PDK1 and PKC-q. PKC-q-mediated phosphorylation of acid SMase increases ceramide release which acts as a negative feedback to control PKC activation, thus strengthening the inhibition of NF-kB activity. These antibody-induced events lead to inhibition of cell proliferation [19]. In conclusion, in this work we show that besides modulating CD4/ZAP-70 in membrane rafts, rIgG₁ 13B8.2 activates the acid sphingomyelinase/ceramide pathway, an important event for structuring raft platforms and transducing CD4-related intracellular signals, which can fine-tune the rIgG₁ 13B8.2-triggered tumoral effects. These findings indicate that the assessment of the lipid-protein rheostat in membrane rafts following treatment with biotechnological drugs could open new avenues for raft-based therapeutics based on the combination of lipid modulators, such as fenretinide [11, 12] and aplidine [13] or the fatty acid synthase inhibitor C75 [54], and therapeutic antibodies.