Interleukin 4 content in chronic lymphocytic leukaemia (CLL) B cells and blood CD8+ T cells from B-CLL patients: impact on clonal B-cell apoptosis

Authors


Neil E. Kay, M.D., Hematology Consultant, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA. E-mail: kay.neil@mayo.edu

Abstract

B-chronic lymphocytic leukaemia (CLL) clonal B cells are characterized by resistance to apoptosis. We evaluated clonal B cells and blood T cells for interleukin 4 (IL-4) content as IL-4 is able to increase CLL cell resistance to apoptosis. The content of IL-4 in CD8+ T cells of CLL patients (n = 9) ranged from 37% to 63% of the total CD8+ T cells (mean level of 49% ± 3·4) compared with a range of 5–10% for control CD8+ T cells. Clonal B cells positive for cytoplasmic IL-4 ranged from 1% to 97% (mean value 57·8 ± 6·9%). CD8+ T cells and clonal B cells secreted detectable levels of IL-4, but only clonal CLL B cells (n = 4) secreted IL−4 in association with increasing cell numbers. Fludarabine (F-ara-AMP, 0·1–100 μmol/ml) was able to downregulate the IL-4 content of CD8+ T cells, but not clonal B-cell IL-4. Culture supernatant from CLL CD8+ T cells decreased the spontaneous apoptotic rate of clonal B cells that was reversed with anti-IL-4 and soluble IL-4 receptor. These findings show that IL-4 is present in the microenvironment of B-CLL. In addition, use of agents that can interfere with IL-4 presentation to clonal B cells can be effective in increasing clonal B-cell apoptosis.

B-chronic lymphocytic leukaemia (B-CLL) is a disease primarily characterized by resistance to apoptosis. Thus, the clonal B cells in this disease accumulate because of a very prolonged survival time and not an increased rate of cell division. The exact aetiology of the dysregulated apoptosis or programmed cell death (PCD) is not known but may reflect multiple mechanisms. These mechanisms could at least include critical changes in the expression of genes (Thomas et al, 1996; Pepper et al, 1997; Kitada et al, 1998; Osorio et al, 1998) that regulate apoptosis (i.e. p53, Rb, Bcl-2 gene family) and/or cytokines that may have an impact on PCD, such as interferon alpha (IF-α) or β, basic fibroblast growth factor, interleukin 13 (IL-13) and IL-4 (Dancescu et al, 1992; Buschle et al, 1993; Panayiotidis et al, 1993; Chaouchi et al, 1994; Jewell et al, 1994; Jewell, 1996). The effects of one or more of these factors with the clonal B cell in B-CLL may be necessary in generating the apoptotic resistance seen in B-CLL patients. As the regulation of apoptosis may also be related to the emergence of chemoresistance in B-CLL (Kitada et al, 1998; Pepper et al, 1998, 1999), further analysis of the complex factors that could have an impact on B-CLL apoptosis is vital to the understanding of both disease pathogenesis and treatment efficacy.

CLL B cells, when placed in culture, will gradually become apoptotic (Collins et al, 1989; Jurlander, 1998). Deprivation of one or more cytokines that regulate apoptosis is probably critical for the onset of apoptosis of the CLL B cell. Cytokines that inhibit apoptosis in vitro include interferon alpha (IF-α), IF-β, IL-2, IL-4, IL-5, IL-10 and IL-13 (Dancescu et al, 1992; Buschle et al, 1993; Fluckiger et al, 1993, 1994; Panayiotidis et al, 1993; Chaouchi et al, 1994, 1996; Jewell et al, 1994; Mainou-Fowler et al, 1994; Huang et al, 1995; Jewell, 1996; Mainou-Fowler & Prentice et al, 1996; Bartik et al, 1998). There have been numerous investigations which show that IL-4 consistently induces resistance to in vitro apoptosis of CLL B cells. The exact mechanism for IL-4-induced resistance of clonal B cells to apoptosis is unknown, but IL-4 can either stabilize or induce augmented levels of Bcl-2 in the clonal CLL B cells (Dancescu et al, 1992; Panayiotidis et al, 1993). This finding may also be relevant to the emergence of drug resistance in the clonal B cells of B-CLL. However, the relevance of this cytokine to the in vivo apoptosis of clonal B cells in B-CLL depends in part on confirmation there is an in vivo source of IL-4 in B-CLL.

We have previously shown that circulating CD8+ T cells in B-CLL contain excessive amounts of IL-4 compared with CD8+ T cells obtained from healthy volunteers (Mu et al, 1997). In this report, we provide additional information regarding the presence of IL-4-containing CD8+ T cells in B-CLL. We have also detected IL-4 in the cytoplasm of clonal CLL B cells. As both cell types also secrete IL-4, we investigated whether the culture medium (CM) could modulate B-cell apoptosis. In addition, we wished to determine whether fludarabine could alter the IL-4 content or if neutralization of IL-4 in CM could alter CLL B-cell apoptosis.

Patients and methods

Patients We studied patients diagnosed with B-CLL. All patients had a CD5+/CD19+ membrane phenotype upon evaluation using flow cytometry. The range of clinical Rai stages for these patients was 0–4 and no patient had therapy within 3–6 weeks of the blood collection. No patient had been treated previously with fludarabine at the time of our laboratory studies.

Lymphocyte populations Peripheral blood mononuclear cells (PBMCs) were isolated from the blood of healthy normal donors and B-CLL patients by differential centrifugation using a ficoll-hypaque density gradient. Monocytes were removed by allowing PBMCs to adhere to a plastic surface for 1 h while being cultured in Roswell Park Memorial Institute (RPMI)-1640 medium containing 10% fetal calf serum. Then, the non-adherent lymphocytes were harvested in a sterile fashion and typically consisted of greater than 90% peripheral blood lymphocytes (PBLs). In some cases, further purification of CD8+ T cells was accomplished via magnetic bead isolation. In brief, PBMCs were suspended with magnetic beads (Miltenyi Biotech, Auburn, CA, USA) coated with mouse anti-human CD14, CD16, CD19 and CD4 monoclonal antibodies. Non-adherent CD8+ T-cell populations were obtained following passage through the magnetic bead column. The negatively selected CD8+ T cells were assessed for purity by flow cytometer (FACScan, Becton-Dickinson, San Jose, CA, USA) using an anti-CD8 monoclonal antibody reagent (Becton-Dickinson). The purity was typically greater than 92%. An alternative source of T cells was obtained by performing sheep erythrocyte (E) rosettes of the PBLs and harvesting the E rosette-positive cells from a ficoll-hypaque monolayer. The purity of T cells with this approach was usually greater than 95%. B-CLL B cells were purified through E-rosette formation with CLL PBLs, followed by ficoll-hypaque centrifugation to remove T cells. The B-cell population at the interface was always greater than 95% CD19+/CD5+ when determined using flow cytometry.

Intracellular cytokine detection PBLs or CD8+ T cells or CLL B cells, after initial isolation or following different in vitro culture time periods, were fixed with 4% paraformaldehyde at 4°C for 20 min, washed with phosphate-buffered saline (PBS) and resuspended in 50 ml of 1% saponin buffer. The PBLs, CD8+ T or CLL B cells were incubated with anti-human IL-4 (1 μg/ml, Pharmingen, San Diego, CA, USA) conjugated with phycoerythrin (PE) and/or anti-human interferon-γ conjugated with fluorescein at 4°C for 30 min. After the cells were washed, they were resuspended in 4% paraformaldehyde and analysed immediately by flow cytometry using a FACScan (Becton-Dickinson). To ensure that there was little or no non-specific binding of anti-IL-4, we preincubated non-labelled anti-IL-4 antibody (10 μg/ml) with the cells for 30 min, then washed with PBS and added the labelled (PE-conjugated) anti-IL-4 antibody to the cells.

Enzyme-linked immunosorbent assay (ELISA) determinations for IL-4. Clonal B cells or CD8+ T cells were placed in culture at various concentrations for 20–24 h in tissue culture medium at 37°C in 5% CO2. Culture supernatants were then harvested and tested for IL-4 content using an ELISA kit provided by Pharmingen (San Diego, CA, USA). The assay wells used to construct a standard curve and IL-4 content (pg/ml) in the culture supernatants were determined according to the manufacturer's specifications.

Co-culture methods Fludarabine (9-beta-d-arabinfuranosyl-2-flouroadenine-monophosphate; F-ara-AMP) was obtained from Berlex (Berlex Laboratories, Richmond, CA, USA). F-ara-AMP was used in co-culture with PBMCs or E rosette-purified cells after dilution in PBS from a fresh, sterile powder form. In order to determine what the effect of F-ara-AMP was on cytokine synthesis in control and B-CLL T cells, we incubated PBMCs or E rosette-purified T cells with various concentrations of F-ara-AMP (0·1–100 μmol/l) for 24–72 h prior to testing for cytoplasmic IL-4 or IF-γ using flow cytometry. Exposure of CLL T cells to F-ara-AMP at 0·5–100 μmol/l/ml for 24 h was not toxic as the level of non-viable cells did not exceed 5–7%.

Apoptosis measurements Membrane permeability, as an indicator of late apoptosis and cell death, was evaluated by staining with the DNA fluorochrome, propidium iodide (PI). In brief, cells (1–5 × 105) were washed in PBS and resuspended in 1 ml of medium containing 0·1% sodium citrate, 0·1% Triton X-100 and 50 μg/ml PI (Sigma Chemical, St. Louis, MO, USA). Following a 30-min incubation at 4°C in the dark, cell nuclei were analysed by flow cytometry using a Becton-Dickinson FACScan. Cellular debris was excluded from the analysis by raising the forward-scatter threshold and the DNA content of the nuclei was registered on a logarithmic scale. The percentage of cells in the hypodiploid region was then calculated. The percentage of apoptotic cells was determined from DNA histograms as a ratio of cells with hypodiploid DNA content to total number of cells.

In some experiments, either anti-IL-4 or soluble IL-4 receptor (sIL-4R, kind gift of Immunex, Seattle, WA, USA) was added to a population of magnetic bead-purified CLL B cells (5 × 105). The anti-IL-4 was used at a final concentration of 10 μg/ml while the sIL-4R was used at final concentrations of 100 and 500 ng/ml. The level of apoptosis in the anti-IL-4- or sIL-4R-exposed CLL B cells was compared to CLL B cells incubated in culture medium alone.

Statistical analyses This was carried out for all data using a paired Student's t-test. P-values less than 0·5 were considered significant. Most of the data is presented as mean values ± one standard error of the mean (SEM).

Results

Cytoplasmic content of IL-4 in T cells of B-CLL patients and controls

We have previously determined that blood CD8+ T cells from B-CLL patients contain a skewed distribution of cytoplasmic IL-4 compared with control CD8+ T cells following exposure to phorbol myriacetate (PMA) and ionomycin (Mu et al, 1997). In this current analysis, we studied several additional B-CLL patients (n = 9) in order to determine the levels of IL-4-containing CD8+ T cells with no in vitro stimulation of magnetic bead-purified CD8+ T cells. The content of IL-4+ CD8+ T cells in these patients ranged from 37% to 63% of the total CD8+ T cells with a mean level of 49% ± 3·4 (mean ± one standard error). Detection of IL-4 in isolated CD8+ T cells or purified T cells obtained from age-matched, healthy controls was usually less than 5–10%. Levels of IL-4 in the B-CLL CD8+ T cells were measured over 24–72 h of in vitro culture. In general, the levels of IL-4-positive CD8+ T cells were maintained or rose over the 24–48 h culture periods when cultured in media alone (data not shown). The flow cytometric analysis of B-CLL CD8+ T cells revealed two populations of CD8+ cells, similar to those previously described in the blood of healthy normal donors (Prince et al, 1993). These were designated as dim and bright CD8+ cells, based on the staining brightness for CD8 antigen (Fig 1). All B-CLL patients evaluated in this study had T-cell populations containing the two subgroups of CD8 cells. B-CLL CD8+ T cells typically have a majority of bright CD8+ T cells compared with dim CD8+ T cells (data not shown). The analysis of IL-4 in PBMCs from B-CLL patients revealed that the CLL B cells also contained cytoplasmic IL-4 (see upper left of Fig 1A). We conducted further studies to document the prevalence of clonal B cells containing IL-4.

Figure 1.

(A) Two-colour histogram demonstrating the presence of both dim and bright CD8+ T cells with IL-4 present in the cytoplasm of a B-CLL patient. (B) Two-colour histogram demonstrating the presence of IL-4 in magnetic bead-purified CLL B cells.

Cytoplasmic content of IL-4 in CLL B cells

To confirm that clonal B cells contained IL-4, we first purified clonal B cells using magnetic bead methodology and then assessed them for cytoplasmic IL-4. Figure 1B illustrates that, in this patient, almost all clonal B cells were positive for IL-4 when tested immediately after isolation. We have analysed the IL-4 content of freshly isolated clonal B cells in 23 additional B-CLL patients. The numbers of clonal B cells positive for cytoplasmic IL-4 ranged from 1% to 97%, with a mean value of 57·8 ± 6·9% (standard error). The mean value for IL-4 in these B-CLL patients (n = 19) after 24 h of culture was 51 ± 8·1% with a range of 1·4 −99·5%. In 6 out of the 19 CLL patients, the numbers of clonal B cells positive for IL-4 declined significantly from a mean of 77·4 ± 8·1% (range of 47·5–97·4%) to 37·8 ± 9·9% (range of 7·2–67·5%) after a 24-h culture. The remainder of the CLL patients (n = 13) B cells exhibited little or no (< 10%) decline in IL-4-positive B cells after 24 h of in vitro culture.

To confirm that CLL B cells secrete IL-4, we measured IL-4 in the culture medium of clonal B cells after 24-h incubations. Each purified B-cell population (n = 14) was cultured for 24 h at 37°C in tissue culture medium and the culture supernatants harvested for ELISA determination of IL-4 content. Each CLL clone secreted detectable levels of IL-4 with a positive dose–response relationship of the cell concentration to levels of IL-4. Thus, the mean levels ± SEM were 120 ± 56·4, 305 ± 168·9, 1020 ± 770·1 and 5628 ± 3202 pg/ml of IL-4 for cell concentrations of 1 × 106, 1 × 107, 3 × 107 and 5 × 107 cells per ml respectively (Fig 2). In a separate set of experiments, purified CLL CD8+ T cells (n = 3) were studied for secretion of IL-4 over a 24-h culture period. We were not able to demonstrate a dose–response relationship between CLL CD8+ T cells and IL-4 secretion levels (data not shown). However, in all three cases, we detected IL-4 in the culture supernatant ranging from 2·5 to 37·6 pg/ml (cell concentration of 1 × 106 CD8+ T cells per ml).

Figure 2.

Relationship of IL-4 secreted by clonal B cells in vitro to the cell concentration of these cells. Data for IL-4 is represented as the mean ± one standard error of the mean.

Modulation of IL-4 by F-ara-AMP in the CD8+ T cells from B-CLL patients

The ability of F-ara-AMP to modulate IL-4 levels was studied because of the previous reports showing inhibition of total RNA synthesis in intact cells by this drug (Huang & Plunkett et al, 1991) and the possibility that manipulation of IL-4 content in CLL T cells could be relevant to CLL apoptotic status. We incubated either magnetic bead-purified CD8+ T cells or E rosette-positive cells from eight B-CLL patients with F-ara-AMP for 24–48 h and then reanalysed their IL-4 cytoplasmic content using flow cytometry. There was a decrease of IL-4 compared with control cells (media only) for the T cells cultured with F-ara-AMP (concentrations ranging from 0·1 to 100 μmol/l) (Fig 3). The maximal decrease in IL-4 cytoplasmic content occurred at the 1 μmol/l concentration for 24 and 48 h of co-culture with F-ara-AMP (Fig 3A and B). Thus, the mean value for IL-4+ CD8+ cells exposed to 1 μmol/l during the 24-h culture period was 51 ± 12·5%, representing a 49% decrease from control levels (P < 0·05). The 1 μmol/l F-ara-AMP concentration was more effective in decreasing IL-4 content in CD8 cells at 48 h (mean value of 66·8 ± 13% IL-4+ CD8+ cells), but was not statistically different from the other F-ara-AMP concentrations (Fig 3B).

Figure 3.

Fludarabine (F-ara-AMP)-induced reduction of IL-4 in B-CLL CD8+ T cells after a 24-h (A) and 48-h (B) co-culture with CLL T cells. Final concentrations of F-ara-AMP were from 100 to 0·1 μmol/l/ml. Data is presented as mean percentage reduction plus or minus one standard error of the mean. Asterisks indicate significant difference (*P < 0·05, **P < 0·10) between F-ara-AMP-treated T cells and control T cells (non-exposed).

Figure 4A and B summarizes the levels of cytoplasmic IL-4 in dim and bright subgroups of the CD8+ T cells, respectively, following co-culture with F-ara-AMP for 24 h. The level of cytoplasmic IL-4 was decreased after incubation with F-ara-AMP for 24 h for both the dim and bright CD8+ T-cell populations, with the greatest mean decrease were seen at the 1 μmol/l F-ara-AMP concentration for both CD8+ T-cell subgroups (Figs 4A and B). The bright CD8+ T cells had greater mean decreases than the dim CD8+ T cells for cytoplasmic IL-4 at the 100 and 10 μmol/l concentrations (Fig 4). We also tested whether F-ara-AMP was able to modify the IL-4 levels in CLL B cells using co-culture experiments (n = 7) of purified clonal B cells and F-ara-AMP. In contrast to CLL T cells, F-ara-AMP was not able to significantly modulate the CLL IL-4 content after co-culture with F-ara-AMP from 100 to 0·1 μm concentrations (data not shown).

Figure 4.

F-ara-AMP-induced reduction of cytoplasmic IL-4 in dim and bright B-CLL CD8+ T cells after a 24-h co-culture with F-ara-AMP (100–0·1 μmol/l/ml). (A) shows the mean percentage reduction plus or minus one standard error of the mean for dim CD8+ T cells, (B) shows the data for the bright CD8+ T cells. Asterisks indicate significant difference (*P < 0·05, **P < 0·10) between F-ara-AMP-treated T cells and control T cells (non-exposed).

Modulation of B-CLL apoptosis

A: with supernatants derived from IL-4-containing T cells . Because IL-4 is a cytokine that can increase B-CLL clonal B-cell resistance to apoptosis (Dancescu et al, 1992; Panayiotidis et al, 1993; Huang et al, 1995), we wished to determine if IL-4+ CD8+ T cells from the B-CLL patients were able to modulate the level of in vitro apoptosis in B-CLL B cells. To test this, we harvested culture supernatants from isolated CLL T cells. These were obtained after 24 h of in vitro culture. The T-cell culture supernatants (TCS) were immediately added to isolated, autologous B-CLL B cells and their apoptotic status determined using flow cytometry over 24–72 h of co-culture. Figure 5 shows a representative experiment (one of three). The level of apoptosis for B cells cultured in TCS was comparable to cells cultured in culture medium alone at the 24-h and 48-h culture time-points. However, the degree of apoptosis was clearly reduced from culture medium alone at 72-h and 96-h culture time-points for the clonal B cells cultured with the B-CLL T-cell supernatants.

Figure 5.

Apoptosis (%) of clonal CLL B cells with and without exposure to culture supernatants from CLL T cells. Data is shown as mean percentage of apoptosis after 24–96 h of in vitro culture of clonal B cells. This figure shows representative resultf of one of three separate experiments.

We added anti-IL-4 to the T-cell culture supernatants at the start of the in vitro culture in order to determine whether the decrease in apoptosis was related to the IL-4 in the B-CLL T-cell culture supernatants. Figure 6 shows that addition of monoclonal anti-IL-4 (10 μg/ml) to the B-CLL T-cell culture supernatant reverses the decrease in clonal B-cell apoptosis seen with CLL T-cell supernatants at the 48-h and 72-h time-points. However, reversal of clonal B-cell apoptosis when anti-IL-4 was added to the cultures was greater than the apoptosis levels for CLL B cells with media alone for the 24-h, 48-h and 72-h time-points of culture.

Figure 6.

Apoptosis (mean percentage) of clonal CLL B cells after exposure to culture supernatants from CLL T cells (TCS). The T-cell culture supernatants were added to CLL clonal B cells with or without the simultaneous addition of anti-IL-4 (10 μg/ml). This figure shows representative results of one of three separate experiments.

B: modulation of apoptosis of clonal B cells by anti-IL-4 and sIL-4R The ability of anti-IL-4 to reverse the apoptosis of CLL B cells to levels beyond that seen in media alone suggested a direct effect on B cells. Thus, we tested if anti-IL-4 (10 μg/ml) alone could increase spontaneous B-CLL apoptosis. These experiments indicated that anti-IL-4 added to CLL B cells was able to significantly enhance cell death above the spontaneous levels detected for B cells in medium alone. This data is summarized in Table I (n = 5). Thus, the mean level of apoptosis was always significantly greater for CLL B cells cultured with anti-IL-4 than CLL B cells in media alone (Table I). In two experiments, we compared the apoptotic level of CLL B cells exposed to sIL-4R (100 and 500 ng) for 24 h with CLL B cells cultured in medium alone. The level of apoptosis was greater for the 100 and 500 ng sIL-4R-exposed B cells (65·8 ± 17·9 and 71·5 ± 3·9% respectively) than the cells in culture medium alone (51·8 ± 4·9%).

Table I.  Comparison of clonal B-cell apoptotic levels with and without anti-IL-4*.
Culture conditionsLevel of apoptosisP-value
  • *

    Purified CLL B cells were cultured with and without anti-IL4 (10 μg/ml) for 24–72 h prior to apoptotic levels being determined, as described in Patients and methods.

  •   †Data is mean ± one standard error of the mean for five separate experiments.

  •   P-values less than 0·5 were considered significant.

Medium (24 h)15·3 ± 4·1 
Medium (24 h) with Anti-IL-425·5 ± 6·3P < 0·04
Medium (48 h)27·8 ± 8·1 
Medium (48 h) with anti-IL-445·5 ± 8·3P < 0·03
Medium (72 h)42·4 ± 12·8 
Medium (72 h) with anti-IL-472·1 ± 11·6P = 0·03

Discussion

In this study, we have presented additional evidence showing that CLL blood CD8+ T cells contain IL-4 and also show that CLL B cells contain cytoplasmic IL-4. In addition, both of these cell populations are able to secrete IL-4. Importantly, CLL B cells exhibit a dose–response relationship to IL-4 secretion levels. These findings are relevant to the apoptotic status of CLL B cells as IL-4 is able to modulate the apoptotic status of clonal B cells. Indeed the use of anti-IL-4 and recombinant soluble IL-4R cultured with CLL B cells resulted in a consistent increase in the apoptotic levels of these cells. Soluble IL-4R is a molecule that can block B-cell binding of IL-4 and inhibit IL-4 induction of B-cell function (Maliszewski et al, 1990). As CLL B cells also express surface IL-4 receptors (Douglas et al, 1997; unpublished observations), the IL-4 produced and secreted by CLL B cells or T cells could influence the apoptotic status of CLL B cells via an autocrine or paracrine mechanism.

Our results confirm our earlier findings and those of others regarding the presence of excessive IL-4-containing CD8+ T cells in B-CLL (Mu et al, 1997; de Totero et al, 1999). We also found IL-4 in both dim and high-density CD8+ T cells. The supernatants from these T cells were able to increase the resistance of clonal CLL B cells to in vitro apoptosis. The exposure of the IL-4+ CD8+ T cells to fludarabine resulted in downregulation but not complete abrogation of IL-4 synthesis in these T cells, based on cytoplasmic flow cytometry analysis. The maximum inhibitory activity of fludarabine was around the 1 μmol/l concentration of fludarabine. In relation to this latter data, mean plasma levels of fludarabine of around 1–2 μmol/l were detected on d 1 and d 5 after intravenous administration of once daily fludarabine phosphate doses of 25 mg/m2 (Hersh et al, 1986). It would have been of interest to determine whether the supernatants obtained from the fludarabine-exposed T cells were less able to support low levels of CLL B-cell apoptosis. However, the presence of fludarabine in the T-cell culture supernatants made these experiments difficult to interpret. In addition, as anti-IL-4 alone is able to increase the apoptotic levels of CLL B cells in vitro, it is difficult to state that the IL-4 in the T-cell culture supernatant was the active moiety in the reduction of CLL B-cell apoptosis by CLL T-cell culture supernatants.

CD8+ T cells obtained from healthy controls usually express low levels of IL-4 (Kaminski et al, 1998). However, more recently, CD8+ T-cell subsets have been described in humans (Salgame et al, 1991; Seder & Le Gros, 1995) with cytokine profiles that resemble T-helper type 2 cells (i.e. contain IL-4, IL-5, IL-10 and IL-13). Thus, patients who have dermatitis, acquired immunodeficiency syndrome (AIDS) or leprosy have blood CD8+ T cells that synthesize IL-4 as a major component of their cytokine profile (Salgame et al, 1992; Maggi et al, 1994; Secrist et al, 1995). The induction of IL-4 in human T cells is complex. For example, IL-4 synthesis in human CD4+ T cells is stimulated by co-culture with IL-4, low levels of antigen and by peptides with low affinity for the T-cell receptor (TCR) (Seder et al, 1992; Pfeiffer et al, 1995; Secrist et al, 1995). Of relevance to our observations is the fact that B cells have been shown to induce IL-4 synthesis by T cells in the context of antigen-presenting cells (Gajewski et al, 1991). While the mechanism for this is not yet clear, we have conducted preliminary experiments that suggest that B-CLL cells can induce normal T cells to generated IL-4 (preliminary observations and data not shown). Further work will be carried out to determine if the clonal B cells are a major factor in the induction of IL-4 synthesis of CD8+ T cells in B-CLL.

Our study also found that CLL B cells both contain and secrete IL-4 in a dose–response relationship. This data strongly suggests that the CLL B cell can be a major source of IL-4 at the microenvironmental level. Normal B cells and CD5+ B cells have been shown to possess either mRNA transcripts or IL-4 protein (Plate et al, 1993; Pistoia, 1997). We have also determined that anti-IL-4 and soluble IL-4R can increase levels of apoptosis of clonal B cells when added to purified clonal B cells cultured in media alone. Thus, it is probable that the IL-4 secreted by CLL B cells is an important regulator of CLL B-cell apoptotic status in vivo. Several different reports have demonstrated that IL-4 is able to regulate the apoptotic status of clonal B cells obtained from B-CLL patients (Dancescu et al, 1992; Panayiotidis et al, 1993; Huang et al, 1995; Mainou-Fowler et al, 1995; Pu & Bezwoda et al 1997). These studies indicate that IL-4 impact on clonal B cells leads to an increased resistance to apoptosis. The mechanism for IL-4-induced apoptosis resistance is probably related in part to the alteration of Bcl-2 levels either through stabilization of Bcl-2 (Dancescu et al, 1992) or via increased levels of Bcl-2 (Panayiotidis et al, 1993). Enhanced expression of Bcl-2 is found in most if not all clonal B cells freshly isolated from patients with B-CLL (Reed, 1998). The increased expression of Bcl-2 in the clonal B cells is positively associated with apoptosis resistance whether apoptosis is mediated by DNA damage-dependent or DNA-independent inducers of apoptosis (Kitada et al, 1998). The mechanism for both the increase in Bcl-2 content and resistance to apoptosis of clonal CLL B cells in vivo is unknown. Our study, showing increased numbers of IL-4-containing CD8+ T cells and clonal B cells in B-CLL that secrete IL-4, indicates that these cells could be the source of a cytokine that confers apoptotic resistance to the clonal B cells.

The in vitro suppression of IL-4 in blood T cells by fludarabine suggests a unique mechanism of action for this drug in B-CLL, that is, the alteration of clonal B-cell apoptotic status by manipulating the presence of IL-4 in the microenvironment. While we did not observe that fludarabine could alter the IL-4 status of clonal CLL B cells in vitro, we did note that addition of anti-IL-4 could increase the apoptotic rate of these cells. Thus, combinations of fludarabine with anti-IL-4 may be of clinical interest. Previous work has also documented that IL-4 can manipulate the impact of chemotherapeutic drugs on B-CLL apoptosis (Mainou-Fowler et al, 1995; Mentz et al, 1996; Frankfurt et al, 1997). One study detected that the combination of chlorambucil- and theophylline-induced apoptosis of clonal B cells in B-CLL was at least partially inhibited by IL-4 (Mentz et al, 1996). In addition, another investigation showed that IL-4 could significantly impair the impact of fludarabine toxicity for malignant CLL B cells (Byrd et al, 1998). These in vitro observations also indicate that therapeutic alterations directed at the downregulation of IL-4 production or its availability in the microenvironment of B-CLL may be of therapeutic benefit.

Acknowledgments

This work is supported in part through funds supplied by Berlex Laboratories. We are indebted to Ms. Susan C. McLean for her energy and excellent technical assistance in the preparation of this manuscript.

Ancillary