Introduction

The slow, smouldering disease presented initially in many of B-chronic lymphocytic leukaemia (B-CLL) patients makes this disease an ideal candidate for immune therapeutic intervention. Active immunotherapy against leukaemia-associated antigens requires the transfer of tumour antigens from malignant cells to ‘professional’ antigen-presenting cells (APC) to initiate a clinically relevant antitumour response. Presentation of antigens by APC is necessary as tumour cells themselves usually lack the accessory molecules required for the induction and expansion of cytolytic T-cells that recognise tumour antigens in MHC class I- and II-restricted manner.1

Dendritic cells (DC) possess the unique ability to efficiently present antigens to naïve T-cells and are a key player in the primary immune response.2,3 DC loaded with immunoglobulin VH-CDR3 (complementarity determining region III) peptides could elicit MHC class I- and II dependent immune response in B-CLL.4 In animal models, fusion of DC with tumour cells induced a polyclonal antitumour immune response leading to the elimination of metastatic tumour.5

Another approach to induce immune response is to use DC pulsed with apoptotic bodies derived from tumour cells. MHC class I and II molecules of DC present proteins better from phagocytosed cellular fragments compared to preprocessed peptides.6 DC that had acquired antigens from apoptotic bodies induced MHC class I-restricted CTL and antitumour immunity in vivo.7,8

In this study, we compare the efficacy of DC that have endocytosed apoptotic bodies or have been fused with tumour cells as APC to stimulate autologous T-cell responses against CLL.

Materials and methods

Patients

All research samples were collected in accordance with the Helsinki declaration on human research and as per the protocol approved by the institutional ethics committee. Six B-CLL patients (three males and three females) with a mean age of 68.5 (range 50–80 years) were studied (Table 1). Three healthy control donors (one male and two females with a mean age of 53 years (range 42–60) were also included. The diagnosis and staging (Rai) as well as the criteria for progressive and nonprogressive disease have been described earlier.9,10 Four patients had stage 0, one patient stage 3 and one patient stage 4 disease. All patients had nonprogressive disease and had not previously received any therapy.

Table 1 Clinical characteristics of B−CLL patients
Full size table

Cell separation

Peripheral blood was collected by venipuncture in sterile heparinised tubes. Peripheral blood mononuclear cells (PBMC) were obtained by centrifugation on a Ficoll-paque gradient (Amersham Pharmacia Biotech, Uppsala, Sweden) as described.11

B-cell purification

PBMC were washed three times with phosphate-buffered saline (PBS), placed on nylon wool columns (Biotest, Breiech, Germany) and B cells were eluted from the column.12 Effluent B cells were collected and cell purity was determined by flow cytometry.

T-cell purification

PBMC of patients were washed three times with PBS and passed through a nylon wool column (Biotest). The effluent cells were further enriched immunomagnetically with anti-CD19 MidiMACS beads (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's specification. The purity of T-cells was >95% as determined by flow cytometry after anti-CD3 staining.

Flow cytometry and immunophenotyping

Cells were analysed by flow cytometry (FACScan Becton-Dickinson (BD), Mountain View, CA, USA) using conjugated CD3, CD5, CD14, CD19 and CD83 monoclonal antibodies (MAb), and relevant isotype controls CD19 and CD83 antibodies were obtained from DAKO A/S (Glostrup, Denmark).

Surface marker staining was performed as described earlier.13 Intracellular staining was carried out according to a modified method as described by Mendes et al.14 Briefly, DC were fixed in 0.5 ml of fixing solution (2% paraformaldehyde in PBS). After washing, the cells were permeabilised with 0.5 ml of 1 × permeabilising solution (BD) and incubated for 15 min followed by washing in buffer (0.1% saponin, 0.2% bovine serum albumin and 0.1% sodium azide). Anti-CD19 (10 μl) or a relevant isotype control antibody was added. The cells were incubated for an additional 30 min in dark and then washed twice with washing buffer and once with PBS. The mean fluorescence intensity and the relative cell frequency expressing the respective surface markers were analysed using the CELLQuest software (BD).

Generation of DCs

DC were generated from CD14+ cells as previously described.15 Briefly, PBMC incubated with anti-CD14+-coated MiniMACS beads (Miltenyi) were passed through MidiMACS columns according to the manufacturer's instruction. CD14+ cells were resuspended in the medium and purity was determined by flow cytometry (83.2±3.8%; mean±s.e.m.). The isolated CD14+ cells were cultured with recombinant (r) GM-CSF (50 ng/ml) (Schering-Plough Research Institute (SPRI), Kenilworth, NJ, USA) and rIL-4 (SPRI; 10 ng/ml) for 5 days. For maturation of DC, rTNF-α (Pharmingen, San Diego, CA, USA) (20 ng/ml) was added at day 5 and the cells were cultured for further 2 days. The harvested cells had the typical morphology of DC with CD3−, CD14−, CD19−, CD83+, CD80+.13

Fusion of tumour B cells and DC (hybrids)

Fusion was performed according to a modified method by Gong et al.16 To optimise the time and ratio of tumour cells:DC, different ratios and times were tested. The ratio of 10:1 tumour and DC fused for 5 min was determined to be optimum. Autologous DC and B cells were washed twice with serum-free medium and incubated for 5 min at a ratio of 1:10 in 50% polyethylene glycol (PEG) solution (Sigma-Aldrich, UK). Prewarmed medium (37°C) was added slowly to the fused cells to dilute the PEG, followed by centrifugation. After washing, the cells were resuspended in the medium and analysed by flow cytometry. DC with CD83+ and CD19+ surface markers were considered as hybrid cells.

Generation of Apo-DC

After purification of B cells, induction of apoptosis was performed according to a modified method by Comby et al.17 Briefly, cells were subjected to 5 Gy irradiation from a Cobalt source (Alcyon, CGR, Paris, France). Irradiated cells were resuspended in the medium and incubated at 37°C with 5% CO2 for 24 h. The rate of apoptosis was analysed using annexin-V FITC apoptosis detection kit (Annexins Research, PharMingen, BD, USA) and 7AAD (PharMingen, BD) using flow cytometry.1 At the end of 24 h of culture, apoptotic cells were spun down, washed and added at a ratio of 1:2 to immature DC. The cells were cocultured for an additional 24 h. Following the initial culture to allow endocytosis of apoptotic bodies, TNFα (20 ng/ml) was added and cultured for an additional 36 h for maturation of DC. After maturation of DC the engulfment of apoptotic cells was measured by intracellular staining of DC for B-cell marker (CD19) along with surface staining for DC marker (CD83). Cells with CD83 and CD19 markers as gated for DC were considered as DC with apoptotic bodies designated Apo-DC.

T-cell proliferation assay

The ability of hybrid cells and Apo-DC to induce an autologous T-cell response was determined in an in vitro proliferation assay. Fusion hybrids or Apo-DC (2 × 104) were added to 2 × 105 autologous T-cells in a 96-well U-bottom plate and incubated at 37°C for 5 days in RPMI 1640 medium (Gibco, BRL, UK) with 10% heat-inactivated pooled human AB+ serum, 1 mM L-glutamine, 100 IU/ml penicillin and 100 μg/ml streptomycin (referred to as medium). Cultures were incubated for a total of 5 days, with 1 μCi [3H]thymidine (Amersham Pharmacia Biotech, Uppsala, Sweden) added to each well for the final 18 h. Cells were harvested and the incorporated radioactivity was measured in a beta-scintillation counter (Microbeta 1450, Wallac, Turku, Finland). Proliferation of T-cells alone or cocultured with autologous DC in the absence of tumour antigen was examined as controls. Baseline proliferation values were determined by assaying T-cell stimulation by Apo-DC and B-cell–DC fusion hybrids of three healthy donors.

Enzyme-linked immunospot assay (ELISPOT)

Identification of IFN-γ secreting T-cells was performed as described earlier.18 The 96-well flat-bottomed nitrocellulose microtiter plates (Millipore AB, Stockholm, Sweden) were coated with anti-human IFN-γ MAb (clone 1-D1 K MABTECH Sweden) at 4°C overnight. Fused cells or Apo-DC (2 × 104) were added to 2 × 105 autologous T-cells and incubated at 37°C overnight. They were then transferred to the washed microtiter plates and incubated at 37°C for 1 day. Biotinylated anti-IFN-γ avidin-conjugated alkaline phosphatase and BCIP® (Sigma Diagnostic, St Louis, MO, USA) were used for the colorimetric reaction. Spots were quantitated using an automatic ELISPOT reader (Axioplan2 Microscope reader, Carl Zeiss, Jena, Germany).

Values obtained with autologous T-cells from three nonleukaemic healthy donors cocultured with fusion or Apo-DC were used to define the baseline value for IFN-γ secreting T-cells.

Results

Immunophenotyping

The process of immunomagnetic selection allowed us to obtain cell populations of high purity at the start of the culture. The purity of CD19+ B-cells in the six B-CLL patients ranged from 90 to 98% (95.3±1.3; mean±s.e.m.). The purity of monocytes (CD14+) ranged from 70 to 99% (83% ±3.8; mean±s.e.m.).

DC generated by the method described in the Materials and methods section were nonadherent and displayed the characteristic morphology described for DC. They also displayed the typical phenotype with high CD83 expression and lack of CD14 (<1% were CD14+) in CLL patients as well as in healthy control donors.

Quantitation of DC–tumour hybrids was performed by calculating the percentage of cells showing dual positivity for DC marker (CD83) and B-cell marker (CD19). By this method, an average hybridisation frequency of 10.5±2.6% (mean±s.e.m.) was estimated over five experiments (Table 2). We speculate that the actual percentage of hybrids generated may have been slightly higher, since dilution of the relevant surface markers over two cells following fusion of their membranes may have resulted in lowered detection sensitivity.

Table 2 Percentage hybridisation and endocytosis of apoptotic B−CLL cells
Full size table

Following irradiation, 49.2±9.8% (mean±s.e.m.) of the B-CLL cells were determined to be apoptotic as measured by Annexin-V positivity. Following coincubation with DC, 22.6 ±6.2% (mean±s.e.m.) of CD83+ cells was determined to have endocytosed CD19+ apoptotic bodies (Figure 1). Thus, in a comparison between fusion hybrids and apoptotic B-CLL cells, Apo-DC had a greater uptake of CLL antigens and thus represented a more efficient antigen delivery system than DC–tumour fusions. In preliminary experiments, irradiated or unirradiated CLL cells by themselves did not demonstrate any T-cell stimulatory activity (data not shown).

Figure 1
figure 1

Phenotypic markers expressed by DC following endocytosis of apoptotic B-CLL cells. The histogram shows fluorescence value of cells, gated for the forward and side scatter characteristic of DC. The fraction of cells, dual positive for CD83-FTIC and CD19-PE, represents DC that had endocytosed apoptotic B-CLL cells (Apo-DC). Data shown were obtained with one patient and similar results were noted with cells of the other patients.

Full size image

Proliferative T-cell response to tumour hybrids and Apo-DC

The capacity to stimulate proliferation of unprimed T-cells is an essential criterion for an APC to generate antileukaemic immune response. The ability of fused and Apo-DC to induce an autologous T-cell response was measured in an in vitro proliferation assay. Average proliferation value, obtained with Apo-DC from three healthy nonleukaemic volunteers, has been depicted as the broken line. Proliferation of autologous T-cells by fusion hybrids from normal nonleukaemic donors was lower than the value represented by the broken line (data not shown). No significant proliferative response was observed in patients or in control donors when DC fusion hybrids were used as stimulators. However, autologous T-cells from all five patients showed a proliferative response with Apo-DC (Figure 2).

Figure 2
figure 2

Proliferation of autologous T-cells following stimulation with DC fusion hybrids or Apo-DC. Fusion hybrids and Apo-DC of B-CLL patients listed in Table 1 were cocultured with purified autologous T-cells as described in Materials and methods. [3H]thymidine was added for the final 18 h of culture. Results represent mean±s.d. c.p.m. of triplicate cultures. Broken horizontal line represents the baseline proliferation obtained with Apo-DC and autologous T-cells of three nonleukaemic donors (enriched normal B-cells). Comparable or lower value than that depicted by the baseline was obtained when fusion hybrids and autologous T-cells of three nonleukaemic donors were tested (data not shown).

Full size image

Production of IFN-γ by autologous T-cells in response to stimulation by DC–tumour hybrids and Apo-DC

Autologous T-cells stimulated with DC–tumour fusion hybrids or Apo-DC were analysed for IFN-γ production by the ELISPOT assay. DC–tumour fusion hybrids and Apo-DC elicited IFN-γ responses in three of five and four of four patients respectively (Figure 3). IFN-γ secretion was statistically significantly higher (P<0.05) using Apo-DC as compared to hybrid tumour cells. T-cells from nonleukaemic donors stimulated with autologous Apo-DC yielded <10 spots/106cells (n=3), which was considered as the baseline value (broken line). Comparable or lower values were also obtained when autologous fusion hybrids of nonleukaemic donors were used as APC (data not shown).

Figure 3
figure 3

Gamma interferon release of autologous T-cells after stimulation with fusion hybrids or Apo-DC. T-cells cocultured with fusion hybrids or Apo-DC were assayed for IFN-γ release by the ELISPOT assay. Data represent mean±s.d. from triplicate wells. Baseline value represented by the broken line was determined using T-cells cocultured with autologous Apo-DC from three nonleukaemic donors (enriched normal B cells). Comparable or lower value than that depicted by the baseline was obtained when fusion hybrids and autologous T-cells of three nonleukaemic donors were tested (data not shown). Purified T-cells from the same patients, polyclonally stimulated with phytohaemagglutinin (PHA) was included as a positive control (data not shown). ND indicates that Apo-DC were not tested for patient 1.

Full size image

Discussion

The use of DC as cellular adjuvants to elicit antitumour immune responses is an important development in the area of cancer vaccination. In this study we observed that DC primed with apoptotic tumour cells and, to a lesser extent, fusion hybrids of DC and B-CLL cells stimulate autologous T-cell responses. Our observations confirm the existence of immunodominant epitopes on tumour cells which are capable of eliciting autologous immune responses when presented in conjunction with potent APC. A comparison of DC–tumour hybrids and Apo-DC revealed that both IFN-γ production and autologous T-cell proliferative responses are induced when apoptotic bodies were used, whereas only IFN-γ production was noted with DC–tumour cell hybrids.

The existence of naturally occurring leukaemic cell and tumour IgVH-CDR3-specific T-cells in B-CLL4,15,19,20 as well as in multiple myeloma and B-cell lymphoma21,22,23 has been reported earlier. Thus, tumour cells exhibit antigenic epitopes capable of inducing a tumour-specific cellular response. Such antigens may potentially be used in a therapeutic approach to induce a T-cell-specific immunity. Multiple myeloma patients, vaccinated with the complete tumour derived idiotypic immunoglobulin, mounted MHC class I- as well as MHC class II- restricted18 T-cell response directed against VH-CDR1-3 regions.

Apoptotic tumour cells have also been described as a suitable vehicle to deliver antigens for processing and presentation.24 An advantage offered by apoptotic bodies is that presentation of intracellular proteins released into tumour lysates or necrotic cells that are not relevant to induction of antitumour responses can be avoided.25 Apo-DC are an attractive alternative since phagocytosed cellular fragments are 300 times more efficient in forming MHC–peptide complexes than preprocessed peptides.26 Considering the above reports as well as the observations of the present study, apoptotic bodies from leukaemic B-cells engulfed by DC (Apo-DC) may be a good antigen presentation system for the induction of immune response in B-CLL patients.

The internalisation rate of apoptotic cells in our study was 22.6±6.2% (mean±s.e.m.). Hoffmann et al24 reported 27% internalisation after 2 h coincubation and 48% after overnight incubation. An autologous T-cell immune response was observed in five out of six CLL patients using Apo-DC (Figure 3). This is similar to a report showing that phagocytosed apoptotic tumour cells induced T-cell proliferation and IFN-γ production in a mouse melanoma model.27 When the fusion DC–normal B-cells were used as APC, no significant proliferation response was observed in autologous normal donor T-cells. Our observations contrast previous reports that demonstrate autologous T-cell stimulation by DC fusion hybrids in solid tumours.28 The disparate observations may be because of differences in solid tumours and B-CLL cells. In four out of five CLL patients, IFN-γ production was observed but in a lower amount than that induced by Apo-DC (Figure 3). Our findings are similar to those reported with pancreatic tumour cells that demonstrate activation of cytotoxic T-cells following cross priming by apoptotic bodies.29 The reasons for the weak proliferation response in T-cells cocultured with DC fusion hybrids are not clear, but it has been shown that purified hybrid cells are more effective in inducing an immune response (IFN-γ production) as compared to fusion mixture.30 The higher proliferative response that we obtained using Apo-DC may be attributed to the fact that particulate antigens prime T-cell more effectively.31 Furthermore, residual nonhybridised B-CLL cells may serve to anergise T-cells.32 All these parameters may affect the proliferation response when using DC fusion hybrids to stimulate autologous antileukaemic T-cells.

In conclusion, this study provides preclinical data on developing a DC-based vaccination strategy in CLL. Our results demonstrate that it is feasible to produce Apo-DC with relative ease. These cells can be efficiently utilised for generating an antileukaemic immune response in B-CLL patients.