SHIP‐1 protein level and phosphorylation status differs between CLL cells segregated by ZAP‐70 expression

The clinical course of B-cell chronic lymphocytic leukaemia (CLL) is very heterogeneous, with some patients that present an indolent disease and others that succumb rapidly despite therapy. Leukaemic cells from aggressive CLL patients typically display B-cell receptors (BCR) encoded by unmutated immunoglobulin variable heavy-chain genes (IGHV) and express the protein tyrosine kinase ZAP-70 (Crespo et al, 2003). In contrast, mutated IGHV genes and the absence of ZAP-70 are mostly found in patients with indolent disease (Crespo et al, 2003). It has been proposed that survival differences between these two subgroups are related to the differential ability of the BCR to respond to stimulation (Chen et al, 2002; Stevenson & Caligaris-Cappio, 2004). Thus, ZAP-70+ patients display a more effective BCR signal transduction that could contribute to their relatively aggressive clinical behaviour (Chen et al, 2002). Inhibitory phosphatases SHP-1 (SH2 domain-containing tyrosine phosphatase-1), SHIP-1 (SH2 domain-containing phosphatidylinositol 5-phosphatase-1) and SHIP-2 are involved in the complex and organized machinery aimed at counterbalancing B-cell activation upon BCR engagement by restricting the duration and/or intensity of the signalling (Zhang et al, 2000; March & Ravichandran, 2002). As ZAP-70− patients present an impaired BCR signalling capacity, we have explored the possibility that these phosphatases were preferentially expressed in CLL cells from this subgroup. Following informed consent and ethical approval, Western blotting analysis was performed on 49 purified CLL samples. SHP-1 and SHIP-1 proteins were found to be expressed in B cells from ZAP-70− and ZAP-70+ CLL patients, while SHIP-2 could not be detected in any of the samples analysed (Fig 1A), even when the amount of protein loaded in each lane was duplicated (data not shown). Normal tonsillar B lymphocytes (TBL), used as controls, showed specific bands at 68, 145 and 150 kDa corresponding to SHP-1, SHIP-1 and SHIP-2 respectively (Fig 1A). There were no significant differences in SHP-1 expression between CLL subgroups (data not shown). By contrast, ZAP-70− CLL cells not only displayed a higher SHIP-1 protein expression compared with ZAP-70+ subgroup (Fig 1B), but also it was constitutively tyrosine phosphorylated to a greater extent (Fig 1C and D). As SHIP-1 phosphorylation may correlate with its inhibitory capacity by enhancing the interaction with other adapter proteins (March & Ravichandran, 2002), ZAP-70− CLL cells might be more prone to inhibition by SHIP-1 than ZAP-70+ ones. SHIP-1 expression and phosphorylation in chronic lymphocytic leukaemia (CLL) cells from ZAP-70− and ZAP-70+ patients. The study included 49 typical B-CLL patients, 35 males and 14 females, with a median age of 70 years. ZAP-70 expression was evaluated by flow cytometry as previously described (Crespo et al, 2003) and was negative in 49% of the patients evaluated. Patients were considered ZAP-70+ when 20% or more B cells expressed ZAP-70 (Crespo et al, 2003). CLL cells purification was performed by mouse immunoglobulin (Ig)G specific for CD3, CD14, CD16 and CD56 and magnetic beads coated with anti-mouse IgG antibodies (Dynabeads M450; Dynal, Oslo, Norway), according to the manufacturer’s instructions. The purity of B-cell population was always >96% by flow cytometric analysis using anti-CD19 monoclonal antibody (MoAb). (A) Immunoblotting analysis of 20 μg of whole cell lysates from purified CLL samples (n = 49) and tonsillar B lymphocytes (TBL) were performed using anti-SHP-1, anti-SHIP-1 (BD Transduction Laboratories, Franklin Lakes, CA, USA) and anti-SHIP-2 antibodies (Santa Cruz Biotechologies, Santa Cruz, CA, USA). The same membrane was blotted with MoAb anti-β-actin (BD Transduction Laboratories) to compare the total amount of protein in each sample. The results obtained for twelve representative patients are shown. (B) The figure shows the densitometric measurements of SHIP-1-specific bands from Fig 1A normalized with β-actin expression. *P < 0·01 (Mann–Whitney non-parametric test). It should be noted that differences in SHIP-1 expression were statistically significant (P < 0·01) even when data from patient 6, who displayed a high SHIP-1:β-actin ratio (ratio: 2·2) was not included in the statistical analysis (data not shown). (C) Immunoblotting analysis of 20 μg of whole cell lysates from purified CLL samples (n = 24) using an antibody that specifically recognizes total SHIP-1 or the tyrosine phosphorylated form (Tyr1021) of the phosphatase (P-SHIP-1) (cell signalling). The same membrane was blotted with MoAb anti-β-actin to compare the total amount of protein in each sample. The results obtained for twelve representative patients are shown. (D) The figure shows the densitometric measurements of P-SHIP-1 specific bands from Fig 1A normalized with SHIP-1 expression. *P < 0·01 (Mann–Whitney non-parametric test). It should be noted that differences in SHIP-1 phosphorylation were statistically significant (P < 0·05) even when data from patient 3, who displayed an extremely high-ratio P-SHIP-1:SHIP-1 (ratio: 3·8) was not included in the statistical analysis (data not shown). (E) 5 × 106 B cells from CLL patients (n = 8) were incubated with goat anti-μ IgG (anti-μ) and RAG (Jackson ImmunoResearch, Bar Harbor, ME, USA) to achieve B-cell receptor (BCR) homoaggregation. Considering that CLL cells express high levels of FcγRIIB, F(ab)′2 fragments of both antibodies were used to avoid their interaction through the Fc portion. Stimulation was stopped after 5, 15 and 30 min and protein extracts were obtained and separated by 8% sodium dodecyl sulphate polyacrylamide gel electrophoresis. P-SHIP-1 expression was evaluated by Western blotting and the same membrane was reprobed with anti-SHIP-1 antibodies. CLL cells were incubated with pervanadate (0·1 mmol/l orthovanadate and 0·3 mmol/l H2O2) as a positive control of phosphorylation (P). (F) Densitometric measurements of P-SHIP-1 specific bands from Fig 1E were normalized with SHIP-1 expression and the increment in SHIP-1 phosphorylation upon BCR engagement is expressed as a percentage relative to unstimulated cells. Values are the mean ± SEM (four ZAP-70− and four ZAP-70+ CLL samples), #P < 0·05 stimulated versus unstimulated cells (Student’s paired t-test), *P < 0·05 ZAP-70−versus ZAP-70+ cells (Mann–Whitney non-parametric test). Finally, we evaluated SHIP-1 phosphorylation status upon BCR cross-linking with goat anti-human IgM (anti-μ) and rabbit anti-goat IgG (RAG). We found that, in ZAP-70− patients, who present an impaired IgM signalling capacity (Chen et al, 2002), BCR cross-linking led to a time-dependent increase in reactivity of the anti-P-SHIP-1 antibody, which was maximal at 15 min and then decreased back towards baseline by 30 min (Fig 1E and F). By contrast, ZAP-70+ samples did not modify SHIP-1 phosphorylation status upon BCR engagement (Fig 1E and F) although they expressed similar levels of surface IgM compared with ZAP-70− samples (data not shown). In conclusion, we found that the inhibitory phosphatase SHIP-1 exclusively participated in BCR signal transduction in ZAP-70− CLL cells, wherein it is expressed and constitutively tyrosine phosphorylated to a greater extent compared with ZAP-70+ samples. Taken together, our data suggest that SHIP-1 might be involved in the impaired BCR signalling commonly found in the former subgroup of patients. Moreover, given that SHIP-1 can negatively modulate not only BCR but also cytokine and chemokine receptor signalling (March & Ravichandran, 2002), the possibility exist that ZAP-70− CLL cells, by expressing higher SHIP-1 levels, hold higher signalling thresholds to different microenvironment stimuli. Experiments are in progress to determine whether SHIP-1 can regulate CLL cell responsiveness. We acknowledge Fundación de la Hemofilia for the use of the FACScan cytometer. We would like to thank Selma Tolosa and Evelia López for their excellent technical assistance and Federico Ramírez and Fernanda Palacios for their assistance with the FACScan. We are also grateful to Servicio de Otorrinolaringología (Hospital de Clínicas José de San Martín, UBA) for providing tonsil samples. This work was supported by grants from: Consejo Nacional de Investigaciones Científicas y Técnicas, Agencia-FonCyT and Lady Tata Memorial Trust.