B-cell chronic lymphocytic leukaemia (B-CLL) is characterized by an accumulation of clonal malignant B cells. The intrinsic characteristics that permit this accumulation have been extensively studied and described. However, it is possible that proliferation and survival of this malignant clone is facilitated by a disruption in the interaction between B and T cells that normally regulate the immune system. In this study, using flow cytometry and cell culture techniques, marked abnormalities of the expression of certain key activation and interaction molecules on the peripheral blood T cells of patients with B-CLL were demonstrated. In particular, on comparison with normal controls, there was a marked reduction in the number of circulating T cells expressing CD25 (interleukin 2 receptor) (P = 0·007), CD28 (P = 0·01) and CD152 (CTLA-4) (P = 0·001). There was also a reduction in the number of circulating T cells expressing CD4 (P = 0·03), CD5 (P = 0·05) and CD11a (P = 0·01). There was no difference in the number expressing T-cell receptor αβ (P = 0·1), CD8 (P = 0·4), CD54 (P = 0·4) and CD154 (P = 0·5), and the only marker expressed on a greater number of circulating T cells in B-CLL patients was HLA-DR (P = 0·05). These results suggest that there is a profound T-cell dysregulation that may contribute to the survival of the malignant B cells in patients with B-CLL and to the related autoimmune phenomena of the disease.
In the normal immune response, T-cell activation is mediated by interactions between antigen presenting cells (APCs) such as B-cells and dendritic cells (DCs). These interactions involve a pathway of highly regulated events critical for specific activation and control of both B and T cells (Fig 1). A recent study has suggested that surface expression of CD154, the ligand for CD40, is reduced in the ‘normal’ T-cell compartment of patients with B-CLL (Cantwell et al, 1997). This finding, if confirmed, would lend support to a role for impaired T-cell function in the pathogenesis of B-CLL proliferation and associated autoimmunity. However, CD40 binding to CD154 is only one in a complex series of events in T-cell activation and interaction with B-cells.
When T cells encounter antigen in conjunction with APCs, cell–cell contact is first established with these APCs via leucocyte function associated molecule-1 (LFA-1) (CD11a) and intercellular adhesion molecule-1 (ICAM-1) (CD54) interaction, when cell binding is still non-specific and of low affinity as no antigen recognition has taken place. The LFA-1/ICAM-1 interaction is a major contributor to adhesion between T cells and other lymphoid cells (Makgoba et al, 1989; Figdo et al, 1990; Lub et al, 1995). The binding of LFA-1 (CD11a)/ICAM-1 (CD54) allows the cells to be brought into close enough contact for antigen recognition to take place via the T-cell receptor (TCR)/CD3 complex.
The immune response continues via signalling between HLA antigens and the TCR, a process that is facilitated by the CD3 complex and strengthened by CD4 (MHC class II) or CD8 (MHC class I) binding, thus triggering the T-cell activation cascade (Benjamini et al, 1996). Signalling through the TCR/CD3 complex leads to the initiation of key T-cell immune responses, including cytokine production and surface marker upregulation. The cell surface marker CD28 on the T cells interacts with the CD80/CD86 receptors on the B cells and upregulation of CD154 on the T cell leads to binding with its appropriate ligand, CD40, on the B-cell surface (Durie et al, 1994; Clark et al, 1996; Grewal & Flavell, 1996, 1997; Lenschow et al, 1996). Production of interleukin 2 (IL-2) is initiated and receptors for this cytokine (CD25/IL-2R) are constitutively expressed approximately 48 h post activation, facilitating recruitment of T cells and their continued activation and clonal expansion (Waldmann, 1986, 1991; Taniguchi & Minami, 1993). The final key stage in this cascade is the expression of CTLA-4 (CD152), believed to send a negative ‘off’ signal to the T cell and to control the immune response, either by terminating T-cell proliferation or by inducing apoptosis (Walunas et al, 1994; Krummel & Allison, 1995; Tivol et al, 1995; Schweitzer & Sharpe, 1998).
This study was therefore designed to examine, in vitro, T cells from the peripheral blood of patients with B-CLL, in five stages. The first was to confirm a previous report of reduced T-cell CD154 expression (Cantwell et al, 1997). The second was to examine expression of other key molecules involved in antigen recognition and T-cell activation: CD25 (IL-2R), CD28, CD152 (CTLA-4), TCRαβ, CD4, CD5, CD8 and HLA-DR. The third was to examine the expression of the adhesion molecules critical for cell to cell contact: LFA-1 (CD11a) and ICAM-1 (CD54). The fourth stage involved the removal of the malignant B-cell clone using magnetic bead separation and the subsequent stimulation of the remaining T cells for analysis of two key surface antigens, CD28 and CD152. The final stage involved depleting the B cells, stimulating with OKT3 and permeabilizing for analysis of internal CD25, CD28 and CD152.
Patients and methods
This study was approved by the local research ethics committee.
Patients Between September 1997 and November 1999, 33 blood samples were taken with informed consent from 27 patients with classic B-CLL who were either untreated or not treated for the preceding 6 months (19 men and eight women, median age 72 years, range 49–86 years). None of the patients had received purine analogues prior to this study. In addition, 26 control samples were taken from 13 age-matched normal subjects.
Cell isolation Peripheral blood mononuclear cells (PBMCs) were isolated from sodium heparin anticoagulated blood by density gradient centrifugation with the lymphocyte separation medium Lymphoprep (Nycomed Amersham, Amersham, UK). Thereafter, PBMCs were washed twice in Roswell Park Memorial Institute (RPMI)-1640 medium (Gibco Life Technologies, Paisley, UK), counted, resuspended in RPMI-1640 medium supplemented with 10% human AB serum (National Blood Transfusion Service, Bristol, UK).
Cell culture A controlled comparison between the mitogenic effects of phytohaemagglutinin (PHA) and OKT3 (anti-CD3) was carried out for expression of each of the cell surface markers. In these experiments, the most effective method of upregulating each marker was determined prior to data collection. PBMCs for CD28 and CD154 analysis were cultured for 4 h in RPMI-1640 medium (Gibco) with 10% human AB serum at 37°C in 5% CO2. The culture plates (Nycomed) were coated with OKT3 (Janssen-Cilag, High Wycombe, UK) for 2 h at room temperature at a concentration of 1 mg/ml in 500 ml of phosphate-buffered saline (PBS) per well and then washed five times in excess PBS prior to the addition of the cells at a concentration 1 × 106 per well.
PBMCs for CD25, CD152 and HLA-DR analysis were cultured for 48 h in RPMI-1640 medium with 10% AB serum at 37°C in 5% CO2. PHA (5 μl) (Murex, Maidenhead, UK) was added to 1 × 106 cells at a final concentration of 1 mg/ml for CD25 activation. OKT3 was used as described above for CD152 and HLA-DR analysis.
B-cell depletion Sodium heparin anticoagulated blood was added 1:1 with cold PBS and layered onto Lymphoprep as described above. Cells were resuspended in buffer (PBS + 2 mmol/l EDTA; Sigma-Aldrich, Poole, UK) and 0·5% bovine serum albumin (BSA; Lorne Laboratories, Reading, UK). CD19 microbeads (Miltenyi-Biotec, Bergisch Gladbach, Germany) were added at a concentration of 30 μl per 1 × 107 total cells for normal controls, and between 55 μl and 70 μl per 1 × 107 total cells for CLL patients, depending on the white cell count of the patient. After 15 min at 4°C, cells were then washed in buffer and passed over two LS+ separation columns (Miltenyi Biotec). After washing, cells were plated for culture with RPMI + 10% human AB serum on 24-well plates precoated with OKT3 for 4 h and 48 h, as described above.
Permeabilization B cells were depleted as described. After 48 h activation with OKT3 as described, cells were harvested and washed in PBS. The cell pellet was resuspended and antibodies for the cell surface marker added and incubated in the dark at room temperature for 25 min. DAKO Intrastain Reagent A (100 μl) (Dako, Ely, UK) was added to each tube, vortexed gently and incubated for 15 min at room temperature. Cells were then washed in 1 ml PBS. DAKO Intrastain Reagent B (100 μl) was added to each tube followed immediately by 10 μl of the relevant antibody for intracellular staining and incubated for 15 min in the dark at room temperature. Cells were then washed, resuspended in 1 ml PBS and analysed.
Flow cytometry All cells stained for antibody were pretreated with 20 μl of human IgG (Sigma–Aldrich) at a final concentration of 250 μg/ml to prevent non-specific binding. Monoclonal antibody was used at 10 μl per 1 × 106 total cells and incubated for 15 min in the dark at 22°C. All cells were analysed on a Coulter Epics Elite flow cytometer. PBMCs were stained with the pan T-cell marker fluorescein isothiocyanate (FITC)-conjugated CD2 (Serotec, Kidlington, Oxford, UK) for the analysis of phycoerythrin (PE)-conjugated CD28 (Beckman Coulter, High Wycombe, UK), PE-conjugated CD154 (Ancell Corporation, Bingham, Nottingham, UK) or PE-conjugated CD152 (Beckman Coulter). PE-conjugated CD154 was also analysed in conjunction with the pan T-cell marker FITC-conjugated CD3 (Serotec), following activation with the pan T-cell mitogen OKT3 (anti-CD3 antibody). PBMCs were stained using the pan T-cell marker FITC-conjugated CD3 (Serotec) for analysis of PE-conjugated CD25, HLA-DR, CD4, CD5, CD8, LFA-1, ICAM-1 and TCRαβ (all Serotec).
The Mann–Whitney (Wilcoxon) comparison of medians was used for statistical analysis.
CD154 expression on CD2+ T cells
A greater number of CD2+ T cells from patients with B-CLL expressed the CD40 ligand CD154 than normals after stimulation (16·6% ± 14 vs. 11·1% ± 4, P = 0·5), although this difference was not significant (Table I). The use of an anti-CD3 mitogen followed by an anti-CD3 monoclonal antibody may effect results previously described by others (Cantwell et al, 1997). OKT3 triggers stimulation of T cells by binding to and occupying the CD3 receptor (Sgro, 1995; Bonnefoy-Berard & Revillard, 1996; Reinke et al, 1997; Brusa et al, 1998). We have shown that if a CD3 monoclonal antibody is subsequently used to identify the T-cell population, a false low number is obtained. OKT3 receptor occupation will prevent and/or reduce CD3 monoclonal antibody (mAb) binding. We therefore used CD2 as our pan T-cell marker and subsequently showed that we were able to identify a greater number of T cells than with CD3 (81·2% ± 3·1 CD2 vs. 71·3% ± 5·6 CD3, P = 0·04).
Table I. Expression of cell surface markers that showed no significant difference between patients and normals.
B-CLL (mean ± SD)
Normal (mean ± SD)
77·6% ± 24·4
95·7% ± 2·2
34·8% ± 8·8
37·4% ± 4·3
41·2% ± 15·1
40·3% ± 11·9
16·6% ± 14·1
11·1% ± 4·5
CD25, CD28, CD152, TCRαβ, CD4, CD5, CD8 and HLA-DR expression on CD2+ T cells
Significantly fewer T cells from B-CLL patients than normals expressed the markers CD25 (IL-2R) (30·7% ± 27 vs. 76·5% ± 18, P = 0·007), CD28 (52·8% ± 23 vs. 79·9% ± 12, P = 0·01) and CD152 (CTLA-4) (1·9% ± 2 vs. 15·5% ± 7, P = 0·001) after stimulation. Fifty percent of B-CLL patients failed to express any surface CD152 following OKT3 activation (range 0–7·5%), while the percentages of CD2+ T cells expressing CD25 after activation varied greatly between patients (range 1·8–91·2%) (Table II).
Table II. Expression of cell surface markers that were significantly different between patients and normals.
B-CLL (mean ± SD)
Normal (mean ± SD)
90·7% ± 18·5
65·9% ± 18·7
47·4% ± 27·4
66·2% ± 4·9
85·0% ± 10·4
99·2% ± 1·4
30·7% ± 27·9
76·5% ± 18·4
77·9% ± 22·5
99·9% ± 0·08
52·8% ± 23·2
79·9% ± 12·1
1·9% ± 2·7
15·5% ± 7·6
In unstimulated T cells from patients with B-CLL compared with those from normals, there was a significant reduction of the percentages of cells expressing the cell surface markers CD4 (47·4% ± 27 vs. 66·2% ± 4, P = 0·03) and CD5 (85·0% ± 10 vs. 99·2% ± 1, P = 0·05), with a wide range of expression between patients (Table II).
Although percentages of unstimulated CD3+ cells from B-CLL patients expressing CD8 and TCRαβ were lower, these were not significant differences (Table I).
In contrast, significantly greater numbers of stimulated CD3+ cells from B-CLL patients expressed HLA-DR compared with normals (90·7% ± 18 vs. 65·9% ± 18, P = 0·05) (Table II).
Expression of adhesion molecules LFA-1 (CD11a) and ICAM-1 (CD54)
Significantly fewer T cells of patients with B-CLL expressed the adhesion molecule LFA-1 (CD11a) than normals (77·9% ± 22 vs. 99·9% ± 0·08, P = 0·01) (Table II) in unstimulated cells. Expression of ICAM-1 (CD54) was not significantly different between patients and normals (Table I).
Expression of CD28 and CD152 on stimulated CD2+ T cells after depletion of B cells
T cells from patients with B-CLL continued to show a reduction in the expression of the two surface antigens CD28 (45·0% ± 19·3 vs. 88·3% ± 8·6, P = 0·01) and CD152 (6·8% ± 5·1 vs. 15·7% ± 8·1, P = 0·05) compared with normal controls after depletion of the malignant CD19+ B-cell population and stimulation of the remaining cell population (Table III). CD19 cells were depleted from PBMCs and any remaining B cells were counted using CD20. The mean CD20 count after B-cell depletion was 3·2% for B-CLL patients and 2·3% for normal controls.
Expression of intracellular CD25, CD28 and CD152 on stimulated and unstimulated CD2+ T cells after B-cell depletion
After permeabilization of stimulated CD2+ cells, intracellular CD28 was expressed in significantly fewer cells in B-CLL patients than in normals (46·2% ± 18·0 vs. 65·9% ± 6·7, P = 0·02) (Table IV). CD25 and CD152 positivity were also reduced on stimulated T cells, but not significantly. In contrast, significantly more unstimulated CD2+ cells from CLL patients expressed intracellular CD25 (7·3% ± 3·3 vs. 3·3% ± 2·2, P = 0·01) and intracellular CD152 (9·2% ± 4·1 vs. 5·4% ± 2·6, P = 0·03) than normals. Intracellular CD28 positivity was reduced in unstimulated CD2+ cells from CLL patients, but this difference was not significant.
Table IV. Expression of intracellular markers following B-cell depletion and permeabilization. A. Stimulated cells
B-CLL (mean ± SD)
Normal (mean ± SD)
46·8% ± 11·1
55·2% ± 17·9
46·2% ± 18·0
65·9% ± 6·7
56·1% ± 14·4
67·3% ± 6·9
B. Unstimulated cells
B-CLL (mean ± SD)
Normal (mean ± SD)
7·3% ± 3·3
3·3% ± 2·2
44·4% ± 14·2
54·6% ± 18·5
9·2% ± 4·1
5·4% ± 2·6
When levels of intracellular CD25, CD28 and CD152 were compared in activated and unactivated cells in normals and patients, there was a significant difference seen in the expression of CD25 and CD152 after activation. In contrast, levels of CD28 were not significantly different in either patients or controls after activation.
These results suggest that the circulating T-cell compartment in the peripheral blood is profoundly dysregulated in patients with B-CLL. Expression of key cell surface activation and interaction markers is markedly reduced after activation, with the exception of HLA-DR and CD154, the ligand for B-cell CD40. Each of the key markers discussed here plays a major role in stimulating not only the T cell on which it is expressed, but also in the activation of and interaction with many other key immune responders. The complete lack of expression of surface CD152 (CTLA-4) on the circulating T cells of half the patients with B-CLL, the persistence of its reduced expression after B-cell depletion and the increased intracellular expression in unstimulated cells are of interest. CD152 plays a pivotal role in immune regulation by effectively providing a negative feedback (‘switch-off’) signal to the T cell once an immune response has been initiated and completed (Walunas et al, 1994; Krummel et al, 1995; Tivol et al, 1995; Schweitzer & Sharpe, 1998).
The finding of increased expression intracellularly in unstimulated cells suggests that these T cells may be in a partial state of activation, yet are unable to effectively express the antigen or signal externally. Without the external expression, failure to ‘switch off’ the T-cell compartment could lead to the survival of a clone of T cells in this partial state of activation, able to make weak immune responses, possibly against self-antigen, yet unable to mount an effective response to known T-cell mitogens. CD152 knock-out mice display a clinical and pathological syndrome that is similar to that seen in the B-CLL patient (Tivol et al, 1995). This includes spontaneous lymphoproliferative disease with lymphocytic infiltrates in many tissues, splenomegaly, lymphadenopathy and higher rates of autoimmune phenomena than in CD152+ mice. Failure to complete a T-cell response or a continued partial response to malignant B-cell antigens could facilitate their proliferation in these mice and in human subjects. The lack of expression of CD152 may result in a failure to delete autoreactive T cells or prevent antigen-specific apoptosis of activated T cells. Anderson et al (2000) have recently shown that, if T cells are not fully activated or the TCR signal is weak, blockade of CD152 paradoxically inhibits immune responses. The lack of expression of CD152 demonstrated in the CLL T cells may equate to blockade in this respect.
Reduced expression of surface and intracellular CD28 after activation suggests a relative inability to interact effectively with any cell that expresses the CD28 ligands CD80 and CD86, including B cells, and thus an impaired ability to promote antigen presentation and processing. Cell–cell adhesion may also be impaired because of reduced expression of LFA-1 (CD11a). A reduced expression of surface CD25 after activation suggests a relative inability to respond to IL-2 and, thus, an impaired ability to control T-cell activation. However, an increased expression of intracellular CD25 in unstimulated T cells correlates with the finding of increased internal CD152. Both results imply that the T cells are in a partially activated state. Reduced expression of CD4 implies that the initial activation signal generated by MHC class II antigen presentation through the CD3/TCR/CD4 complex on the T-cell surface may be weakened, although reduction in expression of TCRαβ is not statistically significant. CD4 plays an important role in both adhesion between B and T cells and also in generating unique and rapid signals to the cell nucleus for activation (Benjamini et al, 1996). A reduction or absence of this antigen may impair or prevent transmission of signals for T-cell activation.
In contrast to our findings of reduced expression of these markers and to the findings of others (Cantwell et al, 1997), we have shown that surface CD154 (CD40L) is expressed equally on T cells in normal subjects and B-CLL patients. By using CD2 as the pan T-cell marker after activation instead of CD3, we may have included T cells that would not have been detected owing to CD3 receptor occupation by OKT3 or other anti-CD3 activators. OKT3 is believed to occupy and modulate the CD3 antigen and to either become endocytosed into the cell or block the receptor, preventing antibody binding and thus reducing the total number of T cells that will stain positively with CD3 (Sgro, 1995; Bonnefoy-Berard & Revillard, 1996; Reinke et al, 1997; Brusa et al, 1998).
We have also shown that HLA-DR is expressed on significantly more T cells in B-CLL patients than in normal controls. This anomaly of normal or increased expression of CD154 and HLA-DR compared with reduced expression of the other antigens in the activation pathway requires further investigation.
These abnormalities of cell surface antigen expression on B-CLL T cells are not altered by depletion of the malignant B-cell clone. This implies that either the effect of the malignant cells is long-lived or that the T-cell abnormality is a primary one. The former explanation is the more probable, suggesting a chronic but not necessarily irreversible dysfunction of the T-cell compartment. Studies are currently underway to address this question.
This work was supported by grants from the Cornwall Leukaemia Trust and the Plymouth and District Leukaemia Fund. We thank Dr Anton Kruger for his encouragement and support, Nick Hurlock for excellent technical assistance and Dr Adrian Copplestone for patient selection.