Phospho-specific flow cytometry identifies aberrant signaling in indolent B-cell lymphoma
© Blix et al.; licensee BioMed Central Ltd. 2012
Received: 23 April 2012
Accepted: 12 October 2012
Published: 16 October 2012
Skip to main content
© Blix et al.; licensee BioMed Central Ltd. 2012
Received: 23 April 2012
Accepted: 12 October 2012
Published: 16 October 2012
Knowledge about signaling pathways in malignant cells may provide prognostic and diagnostic information in addition to identify potential molecular targets for therapy. B-cell receptor (BCR) and co-receptor CD40 signaling is essential for normal B cells, and there is increasing evidence that signaling via BCR and CD40 plays an important role in the pathogenesis of B-cell lymphoma. The aim of this study was to investigate basal and induced signaling in lymphoma B cells and infiltrating T cells in single-cell suspensions of biopsies from small cell lymphocytic lymphoma/chronic lymphocytic leukemia (SLL/CLL) and marginal zone lymphoma (MZL) patients.
Samples from untreated SLL/CLL and MZL patients were examined for basal and activation induced signaling by phospho-specific flow cytometry. A panel of 9 stimulation conditions targeting B and T cells, including crosslinking of the B cell receptor (BCR), CD40 ligand and interleukins in combination with 12 matching phospho-protein readouts was used to study signaling.
Malignant B cells from SLL/CLL patients had higher basal levels of phosphorylated (p)-SFKs, p-PLCγ, p-ERK, p-p38, p-p65 (NF-κB), p-STAT5 and p-STAT6, compared to healthy donor B cells. In contrast, anti-BCR induced signaling was highly impaired in SLL/CLL and MZL B cells as determined by low p-SFK, p-SYK and p-PLCγ levels. Impaired anti-BCR-induced p-PLCγ was associated with reduced surface expression of IgM and CD79b. Similarly, CD40L-induced p-ERK and p-p38 were also significantly reduced in lymphoma B cells, whereas p-p65 (NF-κB) was equal to that of normal B cells. In contrast, IL-2, IL-7 and IL-15 induced p-STAT5 in tumor-infiltrating T cells were not different from normal T cells.
BCR signaling and CD40L-induced p-p38 was suppressed in malignant B cells from SLL/CLL and MZL patients. Single-cell phospho-specific flow cytometry for detection of basal as well as activation-induced phosphorylation of signaling proteins in distinct cell populations can be used to identify aberrant signaling pathways.
Small lymphocytic lymphoma/chronic lymphocytic leukemia (SLL/CLL) and marginal zone lymphoma (MZL) are indolent lymphoid malignancies that arise from mature B cells. The exact cellular origin of SLL/CLL is still controversial, although there is evidence that the lymphoma arises from memory B cells . By contrast, splenic MZL possibly arises from naïve B cells whereas nodal MZL cells are presumed to develop from normal marginal zone B cells . Therefore, signaling pathways that are required for normal B cell maturation and function are likely disturbed in SLL/CLL and MZL. These signaling pathways include the signals propagated through the B-cell receptor (BCR), CD40 and cytokine receptors.
BCR is important for the proliferation, differentiation and apoptosis of B cells . Antigen stimulation via BCR in normal B cells initiates phosphorylation of the immunoreceptor tyrosine-based activation motifs (ITAMs) in the cytoplasmic tails of CD79a and CD79b. Phosphorylation of ITAMs is mediated by different Src family kinases (SFKs) including FYN, BLK, HCK, FGR, LCK and LYN . The phosphorylated ITAMs serve as docking sites for SYK which is then phosphorylated at conserved tyrosine residues by SFKs. This activation initiates the coordinate assembly of the signalosome, composed of a variety of intracellular signaling molecules and includes BTK, phosphatidylinositol 3-kinase (PI3K), Vav and PLCγ. PLCγ activates PKC via DAG, and this further phosphorylates downstream signaling proteins like ERK, p38 and ultimately leads to activation of the pleiotropic transcription factor NF-κB (p65) . The balance of these signals determines the B-cell fate . There is now increasing evidence that signaling via BCR plays an important role in the pathogenesis of CLL . However, there are conflicting results whether exposure to anti-IgM in vitro promotes or suppresses apoptosis in CLL cells  and other signals provided by the tumor microenvironment likely determines the outcome .
Activation of CD40, expressed by normal as well as malignant B cells, is an important co-stimulatory signal that enhances cell viability and promotes isotype class switching . Furthermore, activation of CD40 on B cells induces expression of the co-stimulatory molecule B7 (CD80), improves presentation of alloantigen , and has been shown to activate NF-κB [10, 11]. CD40-induced signaling in CLL cells results in up-regulation of NF-κB and activation of anti-apoptotic pathways [12–14], and induces drug resistance . CD40 stimulation can also activate p38 in B-cell lymphoma cell lines . Cytokine signaling in indolent B-cell lymphoma might also be important for lymphomagenesis, since IL-2 stimulation of CLL cells down-regulates p27 and forces the cells to traverse cell cycle . However, conflicting results have been reported regarding IL-2 production in T cells from CLL patients [18, 19].
Apart from providing information on potential molecular targets for therapy, the study of signaling pathways may provide prognostic and diagnostic information. Different properties of BCR signaling have been identified in normal B cells and in lymphoma B cells from follicular lymphoma patients [20, 21], and impaired BCR signaling identified a subset of follicular lymphoma tumor cells with negative prognostic impact on patients overall survival . Therefore, measuring phospho-proteins by flow cytometry to study signaling networks in cancer cells as well as in infiltrating immune cells at the single cell level [23, 24] is feasible and relatively easy to introduce into clinical practice.
The aim of this study was to use phospho-specific flow cytometry to investigate basal and induced signaling in lymphoma B cells and infiltrating T cells in single-cell suspensions of biopsies from SLL/CLL and MZL patients. The results were compared with those of peripheral blood B cells and T cells from healthy donors (PBMC). We used 9 different stimuli targeting B- and T cells including CD40 ligand (CD40L), BCR engagement by F(ab’)2 (anti-BCR), interleukin 2 (IL-2), IL-7, and IL-15. The phospho-proteins studied included SFKs SYK, PLCγ, AKT, S6, ERK, p38, STAT1, STAT3, STAT5, STAT6 and NF-κB p65. Together, this yielded a comprehensive view of signaling networks in SLL/CLL and MZL lymphoma B cells and in tumor-infiltrating T cells.
The study was approved by Regional Committee for Medical Research Ethics (REK 2.2007.2949). Tumor biopsies from previously untreated patients with SLL/CLL (n = 11) and MZL (n = 5) were collected for diagnostic purposes. Left-over samples were used for this study after informed consent from the patients. Single-cell suspensions were prepared from tumor biopsies and stored in liquid nitrogen in cryotubes until used. All biopsies were reviewed and subtyped according to the WHO classification  by a hematopathologist.in 2010. Peripheral blood was drawn from consenting healthy blood donors (n=9) and PBMC were isolated by density gradient centrifugation (Lymphoprep Axis-Shield, Dundee, United Kingdom). Cells were stored in liquid nitrogen in cryotubes until used.
Anti-BCR antibodies were a mixture of goat polyclonal anti-human IgM F(ab’)2 and goat polyclonal anti-human IgG F(ab’)2 (Invitrogen Carlsbad, CA, USA); each at a final concentration of 10 μg/ml. H2O2 (Sigma-Aldrich, Oslo, Norway) was used at a final concentration of 3.3 mM. The cells were stimulated with IL-2 at 500 U/ml, IL-7 at 20 ng/ml, IL-15 at 20 ng/ml, and soluble CD40 ligand at 200 ng/ml (all from Peprotech, Rocky Hill, NJ, USA). Phorbol 12-Myristate 13-Acetate (PMA) and ionomycin (both from Sigma-Aldrich, Saint Louis, Missouri, USA) were used at 1 μg/ml each.
Anti-CD20-PerCPCy5.5 (clone SK7), anti-CD5-PE-Cy7 (clone L17F12), anti-p-PLCγ2-Alexa488 (Y759), anti-p-ERK1/2-Alexa488 (T202/Y204), anti-p-SYK / Zap70-Alexa647 (Y352/Y319), anti-p-STAT1-Alexa488 (Y701), anti-p-STAT6-Alexa647 (Y641), anti-p-p38-Alexa488 (T180/Y182), anti-p-STAT3-Alexa647 (Y705), anti-p-LCK-Alexa488 (Y505), anti-p-STAT5-Alexa647 (Y694) and anti-p-65(NF-κB)-Alexa488 (Ser529) were all from BD Biosciences (San Jose, CA, USA). The antibodies anti-p-S6-Alexa647 (Ser235/236) and anti-p-AKT-Alexa647 (Ser473) were from Cell Signaling Technology (Danvers, MA, USA). The following Abs were used for immune phenotypic analysis: CD3 PacBlue (UCHT1), CD5 PE-Cy7 (L17F12), CD20 APC-H7 (L27), CD79a APC (HM47), CD79b PE (SN8) and CD23 PerCPCY5.5 (M-L233) were from BD, and IgM FITC (AHI1608), IgG FITC (AHI1308), IgL PE (AHI1907) and IgK APC (MH10515) were from Invitrogen.
Individual cryotubes were thawed, cells were washed in RPMI, counted and incubated for 30 minutes in a tissue culture incubator at 37°C with 5% CO2. After this initial incubation, 200 μl of the cell suspension was aliquoted into a 96-well plate, and the cells were incubated an additional 45 minutes at 37°C. Cells were then left unstimulated or activated with anti-BCR, H2O2, IL-2, IL-7, IL-15, soluble CD40 ligand or with PMA and ionomycin for 4, 15 or 45 minutes.
Signaling was stopped by fixation in paraformaldehyde (Electron Microscopy Service, Hatfield, PA, USA) for 5 minutes at a final concentration of 1.5% at room temperature. The cells were pelleted by high speed centrifugation at 800 g, resuspended in PBS and permeabilized in 90% methanol at −20°C for at least 10 minutes. The cells where then pelleted by high speed centrifugation at 800 g and resuspended in PBS.
Fluorescent cell bar coding were then performed as previously described . Briefly, cells in each well (=one stimulation condition) were stained with a unique combination of two different fluorescent esters; Pacific Blue and Pacific Orange (Invitrogen AS, Oslo, Norway), each used at 3 different concentration level. This bar coding made it possible to identify 3x3 different cell populations (i.e. all the different stimulation conditions given to one patient sample). Pacific Blue was used at a final concentrations of 0.000780, 0.00702 or 0.0498, ng/μL and Pacific Orange was used at 0.00870, 0.0870 or 0.522 ng/μL. Labeling was stopped by adding PBS w/ 1% BSA and then pelleted by high speed centrifugation (800 g), resuspended in PBS with 1% BSA (Sigma-Aldrich, Oslo, Norway) and combined in one tube per individual patient sample.
The barcoded cells were aliquoted into six tubes for staining with different antibody panels. Each panel contained a backbone of the antibodies anti-CD20 and anti-CD5 in addition to two different phospho-antibodies. The cells were stained for 30 minutes in the dark at 4°C, pelleted by centrifugation at 350 g and resuspended in PBS.
For phenotypic analysis, freshly thawed patient samples were stained with various antibodies for 30 minutes in the dark at 4°C, pelleted by centrifugation at 350 g and resuspended in PBS. For staining with CD79a (intracellular epitope), the cells were fixed and permeabilized using paraformaldehyde and 90% methanol as described above, prior to Ab staining.
Acquisition was performed with a three-laser flow cytometer (FACSAria or LSR II, Becton Dickinson, Franklin Lakes, NJ, USA). Data were collected and analyzed using BD FACSDiva software and Cytobank Software ( http://www.Cytobank.org), respectively. Only data from samples of which at least 50% of cells responded to any of the stimulation condition were included.
GraphPad Software (La Jolla, CA, USA) was used to determine statistical significance of difference between groups by applying statistical tests as specified in results.
BCR signaling is controlled by protein tyrosine phosphatases (PTPs) that dephosphorylate signaling molecules after activation in order to terminate signaling. Hydrogen peroxide (H2O2) regulates the quantity and length of signaling by inhibiting BCR-induced PTP activity . When H2O2 was added immediately after BCR cross-linking, BCR-induced signaling was restored in lymphoma B cells, as BCR and H2O2-induced p-PLCγ, p-SFKs and p-ERK were no longer significantly different from healthy donor B cells (Figure 3A). Thus, these results suggest that lymphoma B cells have impaired BCR-induced signaling, but that inhibition of phosphatases can restore signaling in these cells. Furthermore, in normal B cells, all investigated phospho-proteins except p-S6, showed higher expression levels at 4 minutes compared to 45 minutes after BCR activation (Figure 3B). Delayed S6 phosphorylation with strongest activation 45 minutes after BCR cross-linking was also evident in SLL/CLL and MZL cells. Normal B cells had a significant decrease in levels of phosphorylated PLCγ, SYK, SFKs and STAT5 from 4 to 45 minutes, in contrast to SLL/CLL and MZL malignant B cells which showed no significant decrease (Figure 3B). Thus, malignant B cells from SLL/CLL and MZL had low, but sustained BCR-induced signaling. Based on the observed results, a model for BCR-induced signaling in lymphoma B cells from SLL/CLL and MZL patients was constructed (Figure 3C).
Immunophenotyping data in SLL/CLL and MZL
Mean MFI relative to normal B cells (SEM)
In this study we used phospho-specific flow cytometry to map differences in signaling properties within the B- and T-cell subsets from SLL/CLL and MZL patient samples. We found increased basal levels of several phospho-proteins in lymphoma B cells, whereas they overall had impaired, but sustained anti-BCR-induced p-PLCγ, p-SYK/Zap70, p-SFKs and p-ERK, compared to healthy donor B cells. Importantly, impaired BCR-induced p-PLCγ was associated with reduced surface expression of IgM and CD79b. Additional signaling aberrations in lymphoma B cells included CD40L-induced p-p38 and p-ERK.
Overall, malignant B cells from SLL/CLL patients showed significant higher basal levels of several phospho-proteins, including p-SFKs, p-PLCγ, p-ERK, p-p38, p-p65 (NF-κB), p-STAT5 and p-STAT6, but the levels varied considerably between different patients. Constitutive active STATs in SLL/CLL and MZL potentially have biological significance, as targeting JAK/STAT pathways had therapeutic benefits in relapsed lymphomas . JAK2 inhibition by SB1518 prevented tyrosine phosphorylation of STAT proteins, leading to cell cycle arrest and induction of apoptosis . Higher levels of basal p-ERK and p-p38 in SLL/CLL lymphoma cells are also in agreement with previous reports [32, 33]. Earlier work with primary CLL samples have shown constitutive phosphorylation of the SFK LYN, relative to normal B cells  and constitutive phosphorylation of SYK relative to cell lines [35, 36]. We also observed higher basal levels of p-SFKs, whereas basal p-SYK levels did not reach statistical significance, possibly due to low number of patients. The biological significance of basal p-SYK in CLL has clinical relevance, since SLL/CLL patients treated with the SYK inhibitor R406 has shown promising response rates . Whether the lymphoma patients whose lymphoma B cells’ have high basal levels of signaling proteins such as p-SFK and p-SYK, also are the ones with the greatest clinical responses upon specific kinase inhibitor therapy, should be the focus of future studies.
We found that BCR-induced p-SFK, p-SYK/p-Zap70, p-PLCγ and p-ERK were highly impaired in SLL/CLL and MZL lymphoma B cells, compared to normal B cells. Furthermore, we found reduced levels of surface IgM and CD79b in CLL/SLL lymphoma cells, and this correlated with impaired anti-BCR-induced p-PLCγ. Low expression of CD79b in CLL cells has also been reported previously . Furthermore, CLL tumor cells which are unresponsive to anti-IgM, can respond to anti-CD79a treatment, indicating a deficit in signal transmission from the BCR to CD79a/b . However, since a subgroup of CLL samples was unresponsive to activation with anti-CD79a, a potential defect further downstream in the BCR signaling pathway is also possible . Surface IgM expression varies considerably among primary CLL samples, with a subset of patients having markedly decreased IgM expression on the malignant cells [39–41]. Anti-IgM stimulation in primary CLL samples results in global tyrosine phosphorylation mainly in unmutated CLL, but not in mutated CLL samples . The differential response to BCR stimulation in unmutated vs. mutated CLL has been confirmed by other groups [42, 43]. We did not find significant different BCR-induced phosphorylation of target proteins (SFKs, SYK, PLCγ, ERK and STAT5) between unmutated and mutated SLL/CLL, possibly due to small sample size (data not shown). Further, ZAP-70 expression can enhance BCR signaling after anti-IgM treatment, independent of its kinase activity , and CLL cells that expressed ZAP-70 had significantly higher levels of phosphorylated CD79b compared to CLL lacking ZAP-70.
CD40L-induced signaling was also impaired in SLL/CLL and MZL lymphoma B cells compared to normal B cells with significant lower phosphorylation of p38, ERK and S6 in SLL/CLL and lower p38 and ERK in MZL. This finding is in line with previous observations were CD40L stimulation resulted in diminished protein tyrosine kinase phosphorylation in CLL B cells compared to normal B cells, despite similar expression levels of CD40 . The reason for diminished p-p38 expression in SLL/CLL and MZL lymphoma B cells is unclear. Activation of p38 has a pro-apoptotic function in CLL cells, and earlier work has shown that rituximab-induced apoptosis is dependent on phosphorylation of p38 . Furthermore, recent work in primary CLL cells illustrates that chemotherapy-induced up-regulation of the pro-apoptotic protein NOXA is at least partly dependent on p38 . Thus, loss of p38 function is likely to give the tumor cells a survival advantage. Therefore, the role of p-p38 in B-cell malignancies warrants further investigation. In contrast to the overall weak CD40L induced phosphorylation of p38 in malignant B cells, p-p65 (NF-κB) expression was markedly heterogeneous in SLL/CLL and MZL. This finding is in accordance with earlier work where lymphoma B cells from CLL patients were heterogeneous in basal- as well as activation-induced NF-κB . This has potentially clinical implications as a correlation between the NF-κB subunit Rel-A (p65) DNA binding in CLL cells and lymphocyte doubling time was identified, and Rel-A DNA binding was positively correlated with in vitro resistance to fludarabine .
CD40 stimulation resulted in strong phosphorylation of S6 in MZL cells, in a subset of SLL/CLL samples, and in normal B cells. Interestingly, induction of p-S6 seemed to be partially independent of PI3K since we did not find an increased p-AKT level after stimulation with CD40L. Similarly, in SLL/CLL B-lymphoma cells BCR activation resulted in high levels of p-S6, with only small increase in p-AKT. PI3K-independent activation of mTor has previously been described in transformed B cells . Furthermore, it was previously shown that marginal zone B cells expressed more PI3K independent p-S6 after BCR stimulation than follicular B cells . Inhibition of mTor is a potential target in cancer therapy, and clinical trials with its inhibitor everolimus have shown promising results in CLL . Thus, the marked, but variable levels of p-S6 in both unstimulated (data not shown) and CD40L activated cells suggests that detection of p-S6 could be tested as a biomarker in clinical trials with mTor inhibitors.
Tumor-infiltrating T cells in SLL/CLL and MZL showed no significant differences in cytokine induced STAT5 phosphorylation, compared to normal T cells. This was in contrast to ligand independent stimulation which resulted in decreased phosphorylation of p38 and NF-κB p65 in the former. Earlier work with tumor-infiltrating T cells from CLL patients demonstrates defective immunological synapse formation with antigen presenting cells together with reduced tyrosine-phosphorylated proteins at the synapse . Of interest, we found a marked variability in cytokine-induced signaling between SLL/CLL and MZL cell samples. Whether this has prognostic significance for these diseases, remains to be demonstrated.
In conclusion, this work has provided a better overview of basal, BCR- and CD40L-induced signaling in lymphoma B cells as well as cytokine-induced signaling responses in infiltrating T cells in patients with SLL/CLL and MZL. Although we did not identify unique signaling profiles that could distinguish SLL/CLL from MZL, we identified contrasting signaling abnormalities in the lymphoma B cells compared to normal B cells. Further studies using single-cell phospho-specific flow cytometry to obtain patient specific signaling aberrations could provide an opportunity to personalize inhibitor treatment in B cell lymphoma patients.
The authors thank Ron Levy for supporting this work and reading the manuscript and Tor B. Stuge for technical counseling. E.S.B was supported by grant from Helse Nord. J.H.M was supported by the Norwegian Cancer Society and the Research Council of Norway.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.