Comparison of the mRNA expression profile of B‐cell receptor components in normal CD5‐high B‐lymphocytes and chronic lymphocytic leukemia: a key role of ZAP70

Abstract The B‐cell receptor (BCR) signaling pathway is of great importance for B‐cell survival and proliferation. The BCR expressed on malignant B‐CLL cells contributes to the disease pathogenesis, and its signaling pathway is currently the target of several therapeutic strategies. Although various BCR alterations have been described in B‐CLL at the protein level, the mRNA expression levels of tyrosine kinases in B‐CLL compared to that in normal CD5‐high and CD5‐low B‐lymphocytes remain unknown. In the current study, we measured the mRNA expression levels of CD79A, CD79B, LYN, SYK, SHP1, and ZAP70 in purified populations of CD5‐high B‐CLL cells, CD5‐low B‐cells from the peripheral blood of healthy donors, and CD5‐high B‐cells from human tonsils. Here, we report a clear separation in the B‐CLL dataset between the ZAP70‐high and ZAP70‐low subgroups. Each subgroup has a unique expression profile of BCR signaling components that might reflect the functional status of the BCR signaling pathway. Moreover, the ZAP70‐low subgroup does not resemble either CD5‐high B‐lymphocytes from the tonsils or CD5‐low lymphocytes from PBMC (P < 0.05). We also show that ZAP70 is the only gene that is differentially expressed in CD5‐high and CD5‐low normal B‐lymphocytes, confirming the key role of Zap‐70 tyrosine kinase in BCR signaling alterations in B‐CLL.


Introduction
B-cell chronic lymphocytic leukemia (B-CLL) is the most common type of leukemia in Western countries, and the majority of patients are >60 years of age at diagnosis [1]. Over the last decades, the understanding of B-CLL biology has advanced considerably with the discovery of alterations in the B-cell signaling transduction pathways that contribute to the disease heterogeneity and help predict its clinical course.
All malignant B-CLL cells share several unique features, including the absence of a common molecular abnormality, randomly balanced chromosomal translocations, and the inability of malignant B-lymphocytes to exist as a transformed cell line [2,3]. Most intriguingly, however, nearly all B-CLL cells express CD5 [4], which is a common T-cell surface antigen, and it remains unclear whether this feature is a fingerprint of the normal cellular counterpart or a malignancy-related property. After the discovery of the CD5-high B-CLL phenotype, Hayakawa et al. [5] demonstrated the presence of CD5-high B-cells in fetal livers and their preferential abundance in young mice. Similarly, CD5-high B-cells are more frequently observed in human cord blood than in adult peripheral blood [6], and their percentage decreases with age. In adult human lymphoid tissue, CD5-high B-cells are located in the follicle mantle zone and comprise a large subpopulation of tonsillar B-cells. In a mouse model, it has recently been shown that a subset of CD5-high cells has a potential to give rise to B-CLL-like cells in adult animals [7].
Despite the shared pattern of surface antigens, subpopulations of B-CLL cells demonstrate a striking heterogeneity in their molecular and functional features. Thus, B-CLL cells in the peripheral blood have a genetic profile and phenotype that are similar to those of memory B-cells, while B-CLL cells in the lymph node and bone marrow exhibit the characteristics of activated B-cells with an increased proliferative potential [8][9][10]. At the molecular level, the diversity of B-CLL clones, which are known to contribute to disease pathogenesis, affect the BCR signaling pathway.
The B-cell receptor is a transmembrane immunoglobulin complex that is associated with an Igα (CD79a) and Igβ (CD79b) heterodimer. In normal B-cells, tyrosine kinases, such as Lyn and Syk, phosphorylate the ITAM motifs in the CD79α and CD79β receptor subunits, resulting in the downstream activation of BTK, PI3K, and PLCγ and further signal propagation [11]. BCR abnormalities in B-CLL cells include low to undetectable levels of monoclonal surface immunoglobulins, a reduced expression of CD79b, and a malfunction in the downstream pathway, which is predicated by the constitutive activation of both the Lyn and Syk kinases [12,13]. The constitutive activation of Lyn leads to the phosphorylation of the immunoreceptor tyrosine inhibitory motifs (ITIMs) in CD5 inhibitory coreceptors, which are aberrantly expressed on B-CLL cells. Thus, CD5 provides an anchoring site for Src homology 2 domaincontaining phosphatase 1 (Shp-1), thereby triggering negative feedback signaling. In addition, compared to normal B-cells, Syk tyrosine kinase has been reported to be overexpressed in B-CLL cells at both the mRNA and protein levels [14]. However, the most obscure feature of B-CLL signaling is the expression of Zap-70 tyrosine kinase in malignant lymphocytes. Zap-70 is normally present in T-cells and B-CLL cells and is thought to reflect the BCR activation status, which, in turn, correlates with increased tumor proliferation and a shorter time to disease progression [13,15]. Altogether, these findings implicate the antigen-dependent BCR activation as a major pathway of B-CLL progression and pathogenesis [16].
Although it is known that Zap-70 can be expressed in a subpopulation of normal CD5-high tonsillar B-cells depending on the state of their activation, the BCR status and signaling transduction pathways in these unconventional B-lymphocytes remain to be elucidated. In this work, we describe for the first time the transcriptional profiles of BCR signaling components in CD5-high and CD5-low normal B-cells, compare normal B-cells to malignant B-CLL lymphocytes, and confirm the role of ZAP70 as a unique kinase gene that allows for the distinction among different normal and tumor B-cell subpopulations.

Samples
The B-CLL specimens were obtained from untreated patients undergoing lymphoma diagnosis verification at the National Research Centre for Haematology (Moscow) or "GeneTechnology" Diagnostic Centre (Moscow). The samples were immunophenotyped by flow cytometry for each patient. Peripheral blood from healthy donors (n = 10) was used for the normal T-and B-lymphocyte sorting. Additionally, 10 specimens of nonmalignant tonsils from children undergoing routine tonsillectomies were obtained from St. Vladimir's Children Hospital (Moscow, Russia). All tissue specimens were evaluated by immunohistochemistry; and single cell suspensions were prepared using a MediMachine (BD, USA) for the flow cytometry and CD5+CD19+ B-cell sorting. All human material was obtained with informed consent. All procedures were in accordance with the ethical standards of the responsible local committee (at the National Research Centre for Haematology) and the Helsinki Declaration of 1975 as revised in 2008.

Flow cytometry
PBMCs of B-CLL patients and healthy donors were derived from whole blood by Ficoll gradient centrifugation. Single cell suspensions were prepared from fresh human tonsil specimens using a MediMachine (BD, USA) and filtered through a 40-μm cell filter. Surface staining was performed with the monoclonal antibodies anti-CD5 FITC (clone DK23, DAKO, USA), anti-CD19 PE-Cy5 (clone HIB19, eBioscience, USA), anti-CD23 PE (clone MHM6, DAKO, USA), FMC7 FITC (clone MCA792F, Serotec, USA), anti-CD20 PE-Cy5 (clone 2H7, eBioscience, USA), and anti-κ light Ig chain and anti-λ light Ig chain (κ-FITC/λ-PE, AHM4907, Invitrogen, USA). The cells were incubated with the antibodies at +4°C for 20 min, washed with PBS, and analyzed on a FACSAria SORP cell sorter (BD Biosciences, USA). Diva software (BD Biosciences, USA) was employed for the data analysis. All samples were kept at +4°C during staining and prior to FACS analysis. Primary FACS profiles were obtained for each specimen on the day of retrieval typically in 4 h after extraction. Aliquots of cells were frozen in FBS with 10% DMSO in liquid nitrogen for further FACS sorting.

Cell sorting
Sorting was performed separately on thawed specimens using a 5-laser FACSAria SORP instrument (BD Biosciences, USA) with a 70-μm nozzle and the corresponding system pressure parameters. Over 10,000 cells were sorted in each sample. The purity of the sorted populations was assessed by flow cytometry and exceeded 98%.

RNA and cDNA
RNA was extracted from thawed suspensions of the sorted and unsorted cells using the RNeasy Mini Kit (Qiagen, USA) according to the manufacturer's instructions. The RNA concentration was measured using a NanoPhotometer (Implen, Germany), and its purity was assessed according to the A260/A280 and A260/A230 ratios. cDNA was transcribed using the ImProm-II AMV-Reverse Transcription Kit (Promega, USA) according to the manufacturer's instructions.

Primers and real-time PCR
Real-time qPCR was further performed on Quantica (Barlow Scientific, UK) and StepOne (Applied Biosystems, USA) cyclers using Taq-polymerase in SYBR Green I buffer (Syntol, Russia). The reaction protocol included denaturation (95°C, 10 min), followed by 40 amplification cycles (95°C, 15 sec; 60°C, 30 sec; and 72°C, 60 sec). All samples were processed in triplicate. All primers were synthesized and HPLC-purified by Syntol (Russia). A control cDNA sample was included in each PCR run and served as an inter-run calibrator (IRC) to standardize the data. The primer sequences are listed in Table S2.

Data normalization
qPCR data were normalized according to the method proposed by Vandesompele et al. [17]. The following three reference genes were used for the normalization: YWHAZ, UBC, and HPRT1. In short, the relative expression value of each gene of interest was divided by the geometric mean of the reference genes' relative expression values. Three suitable reference genes were established previously [18] and used in the current work.

Statistical analysis
The statistical analysis and data plotting were performed using GraphPad Prism (GraphPad Software, version 5, San Diego, CA, USA). The differences in the mRNA expression levels between groups of malignant and normal lymphocytes were considered significant based on a nonparametric Mann-Whitney test with Bonferroni correction for multiple testing.
All sorted populations were reanalyzed by flow cytometry to confirm a purity >98%, and 10 4 -10 5 cells were collected from each sample. All samples had a similar CD5low/CD19+ phenotype. Unlike the B-CLL lymphocytes, these cells did not express CD23, CD43, CD25, or CD27 on their surface.

Cell sorting of B-CLL samples is a prerequisite for a gene expression analysis
The differential diagnosis of B-CLL was established according to the WHO classification as follows: the clonality of the neoplastic lymphocytes was confirmed by both surface and intracellular kappa/lambda immunoglobulin light chain expression, and all cells had the following The CD38 expression level in the B-CLL lymphocytes varied substantially; however, the majority of the cases were CD38-negative (the data regarding the population percentages and CD38 levels are summarized in Table S1). B-CLL cells were sorted from PBMC samples (n = 20) according to the CD3−/ CD5+/CD19+/CD23+ immunophenotype. An example of the sorting strategy of the B-CLL population is presented in Figure 2.
In order to assess whether the sorting procedure changed the mRNA expression levels of the major signaling BCRassociated proteins, 26 unsorted B-CLL specimens were compared to 20 sorted B-CLL specimens (Fig. 3). The differences in the mRNA expression levels between the sorted and unsorted samples were statistically significant for ZAP70 (P < 0.0001), LYN (P < 0.0001), and SHP1 (P < 0.001), but not for SYK. The most obvious difference was observed in ZAP70 as its mRNA expression level decreased by two orders of magnitude in the sorted B-CLL group (median 0.047 in sorted B-CLL vs. 1.481 in unsorted B-CLL, P < 0.001), which is very important considering its prognostic value. In the purified B-CLL subpopulation, the mRNA expression levels of SHP1 and LYN were decreased, and the expression levels of CD79A and CD79B were elevated (both 10-fold) compared to the levels in the unsorted samples. Altogether, these observations highlight the problem of sample heterogeneity in mRNA expression measurements. While constitutive B-cell mRNAs (such as CD79A, CD79B, and SYK) exhibit similar or decreased expression levels when normalized to housekeeping genes in the unsorted samples, the expression levels of other mRNAs (ZAP70, SHP1, and LYN) may be biased and, thus, "upregulated" due to the presence of T-cells in the unsorted samples. Here, we demonstrate the statistical error that is introduced by sample heterogeneity.
Then, the mRNA expression levels of signal molecules in the sorted normal T-cells, different populations of normal B-cells, and malignant B-CLL lymphocytes were assessed.

Expression of signal molecules in normal T-cells
T-cells have the lowest mRNA expression levels of LYN (median 0.01), SYK (median 0.06), CD79A (median 0.13), and CD79B (median 0.37) (Table 1, Fig. 4). Because Zap-70 tyrosine kinase is a normal counterpart of the T-cell receptor, ZAP70 was upregulated in the T-cells compared to all other groups (median 1.3). Simultaneously, we observed lower levels of SHP1 (P < 0.05) mRNA, even though it is a conventional tyrosine phosphatase that blocks downstream TCR signaling.

Expression of B-cell receptor components in CD5-high and CD5-low normal B-cells
One of the aims of our study was to describe the expression profiles of the BCR signaling components in nonmalignant CD5-high B-cells and compare these profiles to those of normal B-cells from peripheral blood. The expression level of SYK mRNA in the CD5-high B-cells from human tonsils was similar to that in the normal CD5-low B-cells (Table 1), although the variation coefficient in this group was several fold higher (CV = 31.7% for CD5-low B-cells vs. 119.4% for CD5-high B-cells). The differences in the mRNA expression levels of all other genes, except for ZAP70, in the CD5-high and CD5-low B-cells were statistically insignificant. The expression of ZAP70 in the CD5-high B-cells was more than 10-fold higher than that in the CD5-low B-lymphocytes (median 1.3 vs. 0.08, respectively, P = 0.04). Fifteen samples of human tonsils were immunofluorescently stained for intracellular Zap-70, and eight samples contained Zap-70positive B-cells (data not shown).

Expression of signal molecules in sorted B-CLL lymphocytes
CD79A was expressed at nearly the same level (range 4.83-11.04) in all types of B-cells, including the malignant B-CLL lymphocytes. As expected, the mRNA levels of CD79B and SYK in the B-CLL group was compared to that in the normal CD5-high and CD5-low B-cells (P < 0.05, see Table 1). SHP1 mRNA was decreased in the B-CLL samples to the same extent as that in the T-cells compared to that in the normal B-cells. The range of the expression of ZAP70 in the B-CLL group varied dramatically -from 1 × 10 −5 to 0.78 (median 0.07, Table 2). Interestingly, a clear division of the B-CLL group into ZAP70-high and ZAP70-low subgroups (Fig. 4) was observed.
Thus, we separated the ZAP70-low (below 0.02) and ZAP70-high (range 0.07-0.78) subgroups and compared their levels of mRNAs (Table 2). In the B-CLL ZAP70low subgroup, a statistically significant downregulation of SYK, SHP1, and CD79B, and an upregulation of CD79A mRNA levels (P < 0.05 for all datasets) were observed. Thus, the ZAP70-high and ZAP70-low subgroups of B-CLL have different transcription profiles for all BCR signaling molecules, except for LYN tyrosine kinase gene, which tends to be downregulated in all B-CLL lymphocytes compared to normal CD5-high and CD5-low B-cells. The ZAP70-low subgroup appears to have a unique expression profile because it does not resemble either the CD5-high B-lymphocytes from the tonsils or the CD5-low lymphocytes from the peripheral blood (the differences were statistically significant for all mRNAs, except for CD79A).

Differences in protein expression of signal tyrosine kinases in sorted B-CLL cells
Finally, we have tested whether differences in mRNA expression can be confirmed on protein level and analyzed protein expression of BCR tyrosine kinases in ZAP70-high and ZAP70-low B-CLL subsets compared to T-cells as a control. In ZAP70-low group, we confirmed a significant downregulation of Zap-70 protein both by flow cytometry and immunoblotting. We also confirmed the downregulation of Syk tyrosine kinase for ZAP70-low group on a protein level, and no differences in Lyn expression, the only discordance in transcriptional and translational levels was for SHP-1 phosphatase that was uniformly expressed in all tested groups (Fig. 5).

Discussion
FACS purification of cell populations for further qPCR analysis provides a powerful combination of methods that allows for gene expression profiling of both cancer cells and cancer-associated cells. In this study, FACS sorting was performed to evaluate the quality of the B-CLL gene expression data. We revealed statistically significant differences in the mRNAs of the BCR components in the sorted and unsorted B-CLL samples, except for SYK tyrosine kinase. Thus, comparisons of unsorted B-CLL samples and sorted normal B-cells can be incorrect in signal transduction pathways' studies because the admixture of T-cells can alter the quantity of the tyrosine kinase cDNA. Using the FACS sorting method, we isolated the nonmalignant CD5-high B-cells and, for the first time, described the transcriptional profile of the BCR components.
CD5-high normal B-cells were first described in the fetal spleen, where they comprised 40-60% of the splenic B-cells, but were infrequent in the adult spleen and fetal liver [19]. In adult tonsils, CD5-high B-cells bear CD23 (low), IgM, and IgD [20]. These B-cells seed in the mantle zone of the lymphoid follicles, respond to polyclonal activation, have unmutated V-kappa genes, and are not capable of producing natural polyspecific antibodies [21][22][23]. It was shown that Zap-70 was present in normal CD5-high B-cells from human tonsils but not in B-cells from peripheral blood [24]. Here, we show that CD5-high and CD5low B-cells have similar mRNA expression levels of SYK, LYN, CD79A, CD79B, and SHP1 and confirm that ZAP70 is the only upstream tyrosine kinase gene that allows for the discrimination among these cells based on the mRNA level. The Zap-70 protein is a 70-kDa member of the Syk family of protein tyrosine kinases and was first identified as a crucial element for the downstream signaling from the T-cell receptor [25]. Similarly to Syk tyrosine kinase after B-cell receptor (BCR) engagement, Zap-70 is recruited to the phosphorylated immunoreceptor tyrosine activation motifs of the CD3 chains that are present in the T-cell receptor, where it subsequently becomes phosphorylated and initiates several signaling cascades [26]. The expression of the Zap-70 protein and its correlation with the adverse prognosis in certain subsets of B-lymphocyte-derived malignancies prompted the analysis of its functions in a B-lymphocyte setting. In B-lymphocytes, Zap-70 binds to the parts of the heterodimer CD79a/CD79b [27] and delays BCR internalization after IgM stimulation [28]. In 2003, Weistner et al. found that Zap-70 expression was a valuable prognostic marker in B-CLL and that its expression is mainly correlated with known prognostic factors, such as the IGHV mutation status and percentage of CD38positive B-cells [29]. The low levels of Zap-70 corresponded to indolent B-CLL, and high Zap-70 levels corresponded to a more aggressive disease or disease progression. Further studies confirmed the predictive value of this marker [30][31][32]; however, the flow cytometry measurement of the Zap-70 status is often inaccurate because of small differences between positive and negative populations [33] that can also be affected by the choice of antibody clone and gating procedures [34]. Consequently, mRNA quantification for the determination of ZAP70 expression was proposed [35]. There were several attempts to measure ZAP70 mRNA expression in normal and malignant tissues [36][37][38][39]. In those studies, the aim was to determine the correlation between ZAP70 expression and other predictors of a pure outcome, such as the IGHV mutation status and CD38 expression. To normalize the ZAP70 mRNA expression levels in tumor lymphocytes, GUS, GAPDH, and PP1A were proposed; however, each of these genes was used as a sole reference gene. To date, there are no studies including several housekeeping genes simultaneously for the accurate normalization of the ZAP70 mRNA level. The  best single-gene SD-based approach assumes that the gene expression levels are normally distributed throughout the sample set [40]; however, the normality of the distribution in a dataset is rarely addressed in clinical studies. Accurate normalization with several reference genes increases the method's resolution and allows for resolving groups with only a 10-fold difference in gene expression level as demonstrated in our previous work [18,41]. In current study, we used three housekeeping genes, namely, YWHAZ, UBC, and HPRT1, that have been shown to be optimal for PCR data normalization in different types of lymphoid tissue [18], Regarding ZAP70 expression, the B-CLL samples in our dataset can be divided into the following two groups ( Table 2): low ZAP70 mRNA expression (8 samples, normalized cDNA quantity lower than 0.04, median 0.01) and high ZAP70 mRNA expression (12 samples, normalized cDNA quantity higher than 0.15, median 0.31). The ZAP70 mRNA level in the first group corresponds to that in the CD5-low normal B-cells, and the ZAP70 mRNA level in the second group corresponds to that in the CD5-high B-cells from the tonsils. Overall, in the ZAP70-low group, we observed a downregulated mRNA expression of SYK, CD79B, and SHP1 compared to that in the ZAP70-high group. Moreover, the ZAP70-low B-CLL cells demonstrated a lower mRNA expression of SHP1, CD79B, SYK, and LYN compared to the normal CD5-high and CD5-low B-cells (P < 0.05 for all groups), thus demonstrating a unique mRNA profile that differs from any other type of B-cells. This is the first report of the downregulation of a signal tyrosine kinase at the mRNA level. In 2009, Buchner et al. [14] showed an elevated mRNA expression level of SYK tyrosine kinase in B-CLL cells, but these authors used the βglucuronidase gene as the sole reference gene in the ΔΔC t method, making their results questionable since nothing is known about the expression stability of this gene in normal and malignant lymphoid tissue.
Most of the recent BCR studies in B-CLL cells focused on measuring the protein expression and phosphorylation level of these molecules, namely, Lyn and Syk kinases, which were reported to be constitutively active and upregulated at the protein level [12,42]. While Contri et al. [12] showed that Lyn is overexpressed at protein level, they reported no differences in mRNA expression of LYN measured by RT qPCR (data were not shown). Gobessi et al. [13] reported the substantial quantities of Y352phosphorylated SYK in most of B-CLL samples, but there were no data on the overall quantity of SYK protein in B-CLL cells compared to normal B-lymphocytes. To date, there are no studies that directly compare mRNA and protein expression of BCR tyrosine kinases in B-CLL cells and all the clinical studies that investigate the impact of BCR signaling pathway alterations in overall survival and clinical prognosis measure mRNA expression without confirmation of alterations at protein level [43]. The possible reason is the difference in the detection limit of real-time PCR as a semiquantitave method compared to qualitative measurement of protein expression by Western blotting.
The only exception is Zap-70 molecule, that is routinely measured both on protein and mRNA levels with flow cytometry and real-time PCR, respectively. For instance, assessing the samples from E2997 multicenter phase III trial, Kaplan et al. [44] showed that B-CLL samples can be clearly distinguished into two subpopulations based on ZAP70 expression. They have also showed that ZAP70-high group samples demonstrate higher levels of antiapoptotic proteins (Bcl-2, Puma, survivin), which is in line with our hypothesis on the defective apoptosis as a main result of BCR hyperactivation. Using high-resolution immunophenotyping technology, this group also reported that increased expression of ZAP70 is correlated with low phosphorylation levels of Syk and Zap-70 and that more aggressive clinical course of B-CLL is related to higher levels of Zap-70 more than BCR activation pathway [45], thus justifying the importance of functional measurements of tyrosine kinases both in fundamental and clinical research.
We have observed two distinct groups of B-CLLs based on differential ZAP70 mRNA expression. However, none of these groups fully correspond to any of the observed normal B-cell populations in terms of investigated mRNA expression profiles. For instance, ZAP70-low B-CLLs have similarly decreased mRNAs of SYK and other BCRassociated molecules. Such generally "low-expressing" profile is highly uncommon for normal B lymphocytes and may correspond to "signaling-deficient" B-CLL cases, cited in literature previously [46]. At the same time, ZAP70-high B-CLLs do not exhibit such obvious abnormalities but differ from normal B-cells in expression of CD79B (which is common for all CLL samples) together with differences in SYK and LYN mRNA, which expression is similar to CD5-high B-cells and slightly downregulated compared to normal CD5-low B-lymphocytes. In our research, we have also confirmed the downregulation of Syk tyrosine kinase in ZAP70-low group, and for the first time describe the correlation between the mRNA and protein expression of signaling tyrosine kinases in different subsets of B-CLL samples, however, a quantitative measurement of phosphorylation pattern of signaling tyrosine kinases as a hallmark of their functionality is an important issue for further investigation.
Thus, in the ZAP70-low group, we observed BCR signaling incompetence, which suggests that other mechanisms may override the inhibition of the B-cell receptor pathway and eventually promote B-cell proliferation. The evidence supporting this hypothesis is that a subset of B-CLL cells with unresponsiveness to sIg ligation display constitutively activated extracellular signal-regulated kinase ERK1/2 in the absence of Akt activation and the presence of increased NFAT transactivation [47]. In contrast, the ZAP70-high group displays an overexpression of all kinases involved in BCR signalosome formation together with the downregulation of SHP1 inhibiting phosphatase gene (Fig. 6), which leads to the activation of the NFkB and AKT survival pathways and the expression of antiapoptotic proteins, such as mcl-1 and bcl-2, which have been shown to act as crucial players in defective apoptosis in B-CLL cells [48,49]. Altogether, these data are relevant for differential therapeutic strategies and suggest a new approach of selecting therapeutic kinase inhibitors for the two B-CLL subsets that could be easily distinguished by the ZAP70 mRNA expression.
In conclusion, our observations suggest that it is necessary to perform a cell sorting prior to an analysis of gene expression to provide new insight into the regulation of the transcriptional levels of PTKs of B-cell receptor in different B-CLL subsets with high and low ZAP70 expression. Our data are consistent with a model in which Zap-70 is a key regulatory molecule, and its expression allows for the discrimination among different types of normal and malignant lymphocytes.

Supporting Information
Additional supporting information may be found in the online version of this article: Figure S1. Immunophenotype of B-CLL cells. The upper and middle panels show extracellular staining, the lower panel shows intracellular staining. Lymphocytes are gated according to forward and side light scatter characteristics; the malignant population defined by CD19+CD5+ cell phenotype is highlighted in blue. Kappa clone is confirmed on both extra-and intracellular staining. Table S1. B-CLL activation status based on surface CD38 expression. Table S2. Primer sequences used in the study.