BCL1 lymphoma protection induced by idiotype DNA vaccination is entirely dependent on anti-idiotypic antibodies

Abstract

DNA vaccination with the idiotype (Id) of tumour B-cell membrane immunoglobulins (Ig) is a validated strategy to induce tumour protection to several mouse lymphomas. The relative contribution of anti-Id antibodies and T lymphocytes to tumour rejection is still debated. Previous studies in the BCL1 lymphoma model showed that scFv DNA immunisation induces a polyclonal antibody response restricted to conformational epitopes formed by the parental V_L/V_H association. We implemented a system based on this specificity to investigate the mechanism of BCL1 lymphoma protection induced by DNA immunisation. Antibody response and survival of mice immunised with the tumour Id scFv were compared with those of mice immunised simultaneously with two chimeric scFvs, containing either the tumour-derived V_L or V_H paired to an irrelevant V_H or V_L domain, respectively. Animals vaccinated with one or both chimeric constructs were not protected, despite the exposure to all putative tumour Id-derived MHC class I and class II T-cell epitopes. In addition, conformational antibodies induced by DNA vaccination caused tumour cells apoptosis and cell cycle arrest in vitro and transferred protection in vivo. Therefore, lymphoma rejection appears to be completely dependent on the induction of anti-Id antibodies.

Introduction

The antibody idiotype (Id) is formed by the two variable immunoglobulin (Ig) domains (V_L and V_H), whose association generates a collection of unique antigenic determinants. Since the idiotype of B-cell membrane Ig is a clonal marker potentially able to stimulate antibody production as well as CD4⁺ and CD8⁺ T cells [7], it represents a suitable target for vaccination against B-cell lymphomas. Genetic vaccination with constructs encoding the Id sequences can be used to induce an immune response against the Id, providing a convenient alternative to immunisation with the Id protein [5]. DNA vaccines consisting of either a chimeric mouse/human Ig [23, 24] or an scFv encoding a lymphoma-derived Id [4, 6, 10, 12, 17] were able to induce specific anti-Id immune responses and tumour protection in different lymphoma models [5]. Upon DNA immunisation, the in vivo antigen synthesis and processing should favour the loading of antigenic peptides onto MHC class I molecules and the induction of CTL responses. However, studies in the A31 lymphoma and in the 5T33 myeloma models failed to detect proliferation of Id-specific CTLs after immunisation with a DNA-scFv vaccine [12]. Further work using a chimeric mouse/human IgG DNA vaccine in the 38C13 lymphoma model showed that protection was not impaired by CD4⁺/CD8⁺ T-cell depletion [22].

On the contrary, in the same 38C13 model, as well as in the A20 lymphoma, protection was found to be dependent on CD4⁺ and CD8⁺ T cells, when animals were vaccinated with a DNA construct encoding an scFv fused to two different chemokines [6]. In addition, cytotoxic cell lines were efficiently generated from mice vaccinated with an scFv adenoviral vaccine [1]. These lymphocytes displayed specific proliferation and cytotoxic activity against the original A20 lymphoma cell line. In the same model, an Id-derived MHC class I epitope has been identified [1]. Thus, the role played by anti-Id antibodies in lymphoma protection upon DNA immunisation remains an open question.

We have previously shown in the BCL1 model that DNA vaccination with an scFv, covalently linked to the CH3 domain of human IgG1 H chain (γ1-CH3), efficiently broke T-cell tolerance and induced a highly specific anti-Id immune response [4]. In contrast to protein immunisation, antibodies raised with chimeric constructs, engineered by replacing either the V_L or the V_H BCL1 regions with an irrelevant V domain, failed to recognise the BCL1 idiotype (Id^BCL1) and were directed exclusively against the original immunising V_L/V_H associations [3].

In this report, antibody and T-lymphocyte contribution to lymphoma rejection upon DNA vaccination was investigated for the first time in the BCL1 lymphoma model. Thanks to the specificity of anti-idiotypic antibodies induced by DNA vaccination, we could deliver all putative BCL1 V-region peptides for T-cell priming while avoiding induction of anti-Id^BCL1 antibodies. This strategy allowed us to assess the role of cellular mediated protection, without the need to deplete CD4⁺ and/or CD8⁺ T cells. We show that the induction of anti-Id^BCL1 antibodies is a necessary condition for lymphoma protection induced by DNA vaccination. In addition, we demonstrate that, despite their limited specificity to combined V_L/V_H epitopes, DNA-induced conformational anti-Id antibodies are able to stimulate in vitro BCL1 tumour cell death and cell cycle arrest.

Materials and methods

Mice

Female BALB/c mice (6–7 weeks old) were obtained from Harlan (Milan, Italy) and housed at the ICGEB animal house.

Cell lines

BCL1 is a spontaneous B-cell leukaemia/lymphoma of BALB/c origin expressing high levels of membrane IgM/λ [20]. The cell line exists in two variants. The original BCL1 clone was maintained through serial passages in syngeneic BALB/c mice. The BCL1.3B3 variant was maintained in tissue culture. This in vitro cell line can be activated by LPS to secrete IgM. BCL1 cells express both MHC I and MHC II antigens, as reported by Sugie et al. [21]. Both cell lines were kindly provided by Dr E. Vitetta (University of Texas Southwestern Medical Center, Dallas, Tex., USA) and cultured in RPMI 1640 supplemented with 10% FCS, 50 μM 2-mercaptoethanol. The hybridoma 123bcl1 secreting BCL1 IgM was kindly provided by Dr F. Stevenson (Molecular Immunology Group, Tenovus Laboratory, Southampton University Hospitals Trust, Southampton, UK) and cultured in RPMI 1640 supplemented with 10% FCS.

Cell lines mɛ-V^BCL1_L/V^6C6_H and mɛ-V^6C6_L/V^BCL1_H derive from the stable transfection of Sp2/0 cells with an ɛ-isotype membrane version of the scFvs [3]. These molecules contain an scFv fused to the C-terminal region of the human membrane IgE H chain, from ɛ-CH4 down to the cytoplasmic tail [2].

DNA vaccines

Cloning of pBCL1, p6C6 and chimeric constructs has been described in detail previously [3, 4]. Briefly, BCL1 V_L and V_H domains were amplified by RT-PCR from total RNA of BCL1.3B3 cell line. Both domains were cloned in pcDNA3 plasmid downstream to the genomic secretion signal sequence of human IgG1 and upstream to the third constant domain of human IgG1 (γ1-CH3) [13].

p6C6 and chimeric constructs derive from replacement of V_L and/or V_H domains with the corresponding domains of the 6C6 murine monoclonal antibody [13]. To obtain pCH3 vector, γ-CH3 was amplified from pBCL1 with primer ApaLI (5′-GGAGGCTCTGTGCACTCCGAAGGGCAGCCCCGAGAA-3′) and KpnI (5′-GGGGCCTCTGGTACCGGGAGGCGTGGTCTTGTAGTT-3′). Upon ApaLI/KpnI digestion, PCR product was cloned in the corresponding sites of pUT-SEC vector, downstream to the genomic secretion signal sequence of human IgG1 [13]. pCH3 was obtained by inserting the fragment HindIII/BsrGI from the modified pUT-SEC in the corresponding sites of pBCL1.

Immunisation, tumour challenge and immune sera transfer

BALB/c mice were immunised intradermally four times at 2-week intervals with the different plasmid constructs. The mice’s abdominal areas were shaved, and 1-μm gold particles carrying 1–3 μg DNA were injected at 400 psi using the BioRad gene delivery device (BioRad, Hercules, Calif., USA). Sera were collected 2 weeks after the third boost. Two weeks after the last boost, mice were challenged intraperitoneally with 5×10⁴ viable BCL1 tumour cells. For the sera transfer experiments, mice were injected simultaneously with 5×10⁴ viable BCL1 tumour cells and 10 or 100 μl of nonimmune and pBCL1-induced sera.

Antibody response

Reactivity against human γ1-CH3 was determined by ELISA using plates coated with 3 μg ml⁻¹ of human IgG (Sigma Immunochemicals, St Louis, Mo., USA). Immune response to murine Id^BCL1 was detected in ELISA, using plates coated with around 1 μg ml⁻¹ of murine BCL1 IgM purified from hybridoma 123bcl1 supernatant, as well as by flow cytometry on BCL1.3B3 cells using a FACScalibur (Becton Dickinson Immunocytometry, San José, Calif., USA). Immune sera were diluted 1:200 and immunocomplexes detected with HRP or FITC-conjugated goat antimouse IgG antibodies (Pierce, Rockford, Ill., USA; Kirkengaard and Perry, Gaithersburg, Md., USA).

T-cell proliferation assay

The Id^BCL1 and Id^BCL1/CH3 idiotypic proteins were either produced from mammalian cells or expressed in bacteria. In this latter case, V^BCL1_L and V^BCL1_H domains were amplified from pBCL1 with primer BssHII (5′-GCGCGCATGCCCAGGCTGTTGTGACTCAGGAA-3′) and BspEI (5′-GCTAGCCCCAGAGCCTCCGGAGGAGACTGTGAGAGTGGTG-3′). Upon BssHII/BspeI digestion, PCR product was cloned in pBCL1. CH3 domain was amplified from pBCL1, in order to remove the stop codon and insert a NheI site, with primers BspeI (5′-ACAGTCTCCGGAGGCTCTGGGGGGCAGCCC-3′) and NheI (5′-GAATTCATTGCTAGCTTTACCCGGAGACAGGGAGAG-3′). The amplified fragment was inserted in the modified pBCL1 vector. A BssHII/NheI fragment was cloned in pDAN3 vector (Invitrogen, Leek, The Netherlands), downstream to the pelB signal for the secretion in the periplasm, to generate pDAN3BCL1/CH3. pDAN3BCL1 was obtained from pDANBCL1/CH3 by removing the CH3 domain and inserting a BspeI/NheI minilinker (5′-CCGGAGGCTCTGGGG-3′). pDAN3BCL1/CH3 and pDAN3BCL1 were used to transform the HB2151 E.coli strain. Bacteria were grown in 2YT ampicilin medium at 37°C to 0.7 A_600 nm, induced with 0.5 mM IPTG and incubated at 30°C o/n. The periplasmic scFv fraction was prepared as previously described [19]. Upon purification using the Ni-NTA kit (Qiagen, Valencia, Calif., USA), proteins were dialysed o/n against PBS. To express Id^BCL1 protein in mammalian cells, the pBCL1 construct was modified by replacing the CH3 domain with an SV5 tag [11] and used to stably transfect Sp2/0 cells. Id^BCL1/CH3 and Id^BCL1 were purified from Sp2/0 culture supernatants by affinity chromatography with either an anti-SV5 or an anti-hIgG antibody (Pierce) coupled to Sepharose (Amersham Biosciences, Little Chalfont, UK). CD4⁺ T lymphocytes were purified from splenocytes of immunised mice using antimouse CD4⁺ magnetic beads (Miltenyi Biotech, Bergisch Gladbach, Germany). CD11c⁺ dendritic cells were purified from spleen of BALB/c control mice with anti-CD11c⁺ coupled magnetic beads (Miltenyi Biotec). Purity of cell populations was assessed by FACS analysis and was >80%. For the in vitro stimulation assay, 4×10⁴ splenic DCs were plated in 96-well plates and incubated with either the idiotypic proteins or human IgG for 3 h at 37°C (100 μl volume). After incubation, 1×10⁵ CD4⁺ T cells from immunised mice were added to wells (in 100 μl volume). After 72 h, ³H-thymidine (0.5 μCi/well) was added for 18 h, and incorporation was measured by liquid scintillation counting after collection of cells on a glass fibre filter with an automatic cell harvester (Tomtec, Calif., USA).

In vitro antibodies mediated cell death and cell cycle arrest

Lymphocytes BCL1 of tumour-bearing mice were purified from splenocytes by Ficoll-Plaque (Amersham) treatment. One hundred thousand BCL1 lymphocytes, A20 cells and BCL1.3B3 cells were incubated o/n in a 96-well microplate with mice sera and anti-Fas (Pharmingen) (100 ng) antibody. Cells were washed with 2 ml of PBS and stained with annexin-V-fluos (Roche, Basel, Switzerland) and propidium iodide (PI, 1 μg ml⁻¹; Molecular Probes, Eugene, Ore., USA) in 500 μl of binding buffer (10 mM HEPES/NaOH, pH 7.4, 140 mM NaCl, 5 mM CaCl₂). Samples were incubated 20 min in the dark at RT and analysed by flow cytometry. For cell cycle analysis, BCL1 cells were cross-linked, stained with annexin-V-fluos as above and washed with PBS. Cells were then fixed with 70% ethanol 30 min at 4°C and washed two times with PBS. Samples were resuspended in a solution of PI (50 μg μl⁻¹) containing RNase (Amersham) (6 μg ml⁻¹) 30 min in the dark at RT. Viable cells were gated and analysed for cell cycle status by flow cytometry.

RT-PCR

RT-PCR was performed on total RNA extracted from 10⁴ splenocytes of control and surviving mice and from 10² and 10³ splenocytes of tumour-bearing mice using BCL1-specific primers 117–138 (5′-ATTGCCATGGGTTGGAGCTGTA-3′) and 549–573 (5′-AAAGTAGTTACCATAGTATCTTGCA-3′), as previously described [15]. The amplified fragment of 375 bp, from the leader sequence to the CDR3 region of BCL1-V_H domain, were separated in a 2% agarose gel and visualised by ethidium bromide staining.

Statistical analysis

Survival analysis was performed using the Kaplan-Meier product-limit method. The log-rank test was used to compare survival curves.

Results

T-cell responses

It has been previously demonstrated that a DNA vaccine (pBCL1) encoding the BCL1 scFv covalently linked to the human γ1-CH3 domain (Id^BCL1/CH3) efficiently evoked anti-Id^BCL1 antibodies [4]. We investigated whether this vaccine formulation also induces anti-Id T-cell responses. For this purpose, proliferation of splenic CD4⁺ and CD8⁺ T cells of immunised mice was analysed in vitro upon antigen exposure. CD11c⁺ dendritic cells were purified from the spleens of BALB/c control mice and pulsed with hIgG, Id^BCL1/CH3 or Id^BCL1 proteins. DCs were then incubated with CD4⁺ or CD8⁺ T cells isolated from the spleens of mice immunised with pBCL1 or pCH3, a control plasmid containing only the γ1-CH3 domain. CD4⁺ T cells of mice immunised with pBCL1 as well as with pCH3 proliferated upon stimulation with hIgG and Id^BCL1/CH3-loaded DCs (Fig. 1). Conversely, we could not detect CD4⁺ lymphocytes from pBCL1 immunised mice that responded to DCs pulsed with the Id^BCL1 protein (Fig. 1). The same assay, performed on CD8⁺ purified T cells, did not reveal any proliferation, either with hIgG- or Id^BCL1/CH3-loaded DCs (data not shown).

Fig. 1.

Proliferation of splenic CD4⁺ T cells. DCs were purified from spleens of BALB/c control mice and pulsed with 1.5 μg/ml of hIgG, Id^BCL1/CH3 and Id^BCL1. DCs were then incubated with CD4⁺ T cells extracted from mice immunised with pCH3 or pBCL1. T-cell proliferation was detected by ³H-thymidine incorporation.

Immunisation strategy and antibody responses

Although a specific activation of anti-idiotypic T lymphocytes cells was not detected in vitro, the involvement of anti-idiotypic CD4⁺ and/or CD8⁺ T cells in tumour protection in vivo can not be excluded. To address this issue, we implemented a vaccination system, based on the humoral response specificity previously reported in our model. Indeed, anti-Id antibodies raised by DNA vaccination have been found to be highly specific, i.e. directed exclusively to the combined V_L/V_H associations used for immunisation [3]. Taking advantage of this specificity, we reasoned that mice immunised with either one or two chimeric constructs each containing a single BCL1-derived V domain should allow the selective stimulation of anti-BCL1–specific T lymphocytes without inducing anti-Id^BCL1 antibodies. Thus, by comparing the antitumour effect induced in these animals with that in mice treated with the complete BCL1 idiotype, we were able to evaluate in vivo the relative contribution of T lymphocytes and antibodies in the lymphoma rejection.

Figure 2 shows the different construct combinations used to immunise five groups of mice, as well as the delivered putative BCL1-derived T-cell epitopes. pBCL1 is the plasmid containing the BCL1 scFv, while plasmid p6C6 contains an irrelevant scFv Id derived from a murine mAb [3, 4]. The two chimeric constructs, pchB and pchC, contain only one of the BCL1-derived V regions (V^BCL1_L and V^BCL1_H, respectively) paired to the corresponding 6C6-derived partners [3].

Fig. 2.

a DNA-immunising constructs. The construct pBCL1 contains both BCL1-derived V_L and V_H domains, while the p6C6 vector contains the Id of an irrelevant murine monoclonal antibody (mAb 6C6). The chimeric constructs pchB and pchC contain the V^BCL1_L/V^6C6_H and V^6C6_L/V^BCL1_H domains, respectively. b Strategy of immunisation. Eight mice/group were immunised four times at 0, 14, 28, and 42 days with the indicated constructs. Each group was exposed to the indicated combinations of putative MHC class I and II T-cell epitopes derived from the V^BCL1_L and/or V^BCL1_H domains.

Following immunisation, antibody responses were determined by ELISA and by flow cytometry on BCL1.3B3 cells, which is a variant of the original tumour that can grow only in vitro. Sera of mice immunised with the scFv encoding both tumour-derived V_L and V_H domains (pBCL1) reacted specifically with BCL1-IgM (Fig. 3a, b). As expected, immunisation with the chimeric constructs delivered separately did not induce production of anti-Id^BCL1 antibodies. Strikingly, mice immunised simultaneously with both chimeric constructs failed to develop anti-Id^BCL1 humoral responses, despite the delivery of each BCL1-derived V domain, albeit in different Id context. However, these sera specifically recognised the γ1-CH3 domain (Fig. 3a) and the original immunising V_L/V_H associations (Fig. 3c), as shown by flow cytometry analysis on Sp2/0 cells displaying the chimeric Ids. These cells were obtained by transfection with constructs encoding ɛ-isotype membrane version of the chimeric scFvs, as previously described [3].

Fig. 3.

Antibody responses induced by DNA vaccination. a Ab levels were measured by ELISA against γ1-CH3 or Id^BCL1 on blood samples collected after three DNA shots at 1:200 dilutions. Reactivity is plotted as mean for each mice group and the standard deviations (SDs) among mice sera are indicated. b Mice sera were diluted 1:200 and analysed by flow cytometry on BCL1.3B3 cells. The reactivity of one mouse per group is shown as a representative response. c Ab reactivities of sera from mice immunised with one or both chimeric constructs were analysed by flow cytometry. Sera were incubated at 1:200 dilution with Sp2/0 cells displaying on their surface the indicated Ids. Sera of nonimmunised mice were used as negative controls. The reactivity of one mouse per group is shown as a representative response.

Tumour survival

The five groups of immunised mice were next challenged with 5×10⁴ tumour cells. Splenomegaly was monitored to measure lymphoma growth (not shown) and mice were followed up for survival (Fig. 4). Control mice (p6C6) started to die by day 30. No difference was reported between control mice and mice immunised with each of the chimeric constructs (p=0.4503 for pchB and p=0.9052 for pchC). Interestingly, the simultaneous vaccination with both chimeric constructs did not prevent tumour development, and the death of mice was not delayed as compared to negative controls (p=0.6163). Immunisation with pBCL1 resulted in a significantly longer survival with respect to controls (p<0.001), with 25% of animals alive after more than 18 months. Comparable results were obtained in two additional experiments (data not shown). Therefore, the induction of anti-Id^BCL1 antibodies strictly correlates with tumour survival. Mice vaccinated simultaneously with the two chimeras were not protected, despite the exposure to all putative Id^BCL1-derived MHC I– and MHC II–restricted T-cell epitopes.

Fig. 4.

Survival of immunised mice following challenge with BCL1 lymphoma. Two weeks after the fourth shot, mice received 5×10⁴ viable BCL1 lymphoma cells intraperitoneally. Animals were then followed up until death. The p value refers to the pBCL1 group compared with the p6C6 control group.

To confirm the essential role of anti-Id antibodies in tumour rejection, two groups of mice were injected with 5×10⁴ BCL1 tumour cells and 10 or 100 μl of nonimmune or anti-Id^BCL1 sera (Fig. 5). The injection of 10 μl resulted in a small but significant delay of tumour development and in the complete protection of one animal (p=0.0362). Treatment of mice with 100 μl of anti-Id^BCL1 serum improved survival significantly, resulting in protection of two animals after 5 months (p=0.0006).

Fig. 5.

Immune sera transfer. Mice were injected intraperitoneally with 5×10⁴ viable BCL1 lymphoma cells and 10 or 100 μl of control or pBCL1-induced sera. Animals were followed up for survival. p Values refer to pBCL1 groups compared to control group.

In vitro effect of anti-Id antibodies

To characterise the capacity of DNA-induced antibodies to trigger tumour cell death, BCL1.3B3 cells were incubated with different amounts of vaccinated or nonimmune mice sera and analysed by annexin/propidium iodide assay. Anti-Id^BCL1 antibodies induced a dose-dependent effect on BCL1.3B3 cells, with ~93% cell death at the higher dose of anti-Id^BCL1 serum (Fig. 6a). The same serum on A20 cells, expressing an irrelevant Id, did not induce cell death.

Fig. 6.

a Dose/effect response on BCL1.3B3 cells treated with different amounts of nonimmune or pBCL1-induced sera. Cell death was detected using annexin V and PI. Percentage of viable cells is indicated. A20 B-cell line was used as negative control. Data are normalised to the negative control (nonimmune/BCL1.3B3), plotted as 100%. b Induction of BCL1 tumour cell death following o/n incubation with pBCL1 and pchB+pchC induced sera or anti-FAS (100 ng) antibody. Data are normalised to the negative control (untreated cells), plotted as 100%. SDs of three different experiments are indicated. c Tumour cells were treated with a nonimmune and a pBCL1-induced serum and stained with annexin V. Viable cells were gated and analysed for cell cycle status with PI. The percentages of viable cells on S/G2/M phases (M1 region) are indicated. d Amplification of BCL1 V_H in surviving mice. RT-PCR was performed on 10⁴ splenocytes of control (lane 1) and surviving (lanes 2, 3 and 4) mice, and on 10² (lane 5) and 10³ (lane 6) splenocytes from tumour-bearing mice. Amplification products were separated on a 2% agarose gel.

Similar results were found with the original BCL1 tumour cells. These cells were obtained from the spleens of animals at the terminal stage of the disease. Treatment of these cells with sera of mice immunised with pBCL1 induced death of ~64% of cells (Fig. 6b). This response was comparable to the level of apoptosis induced by an anti-FAS antibody (~87% cell death). As expected, sera of mice immunised with the two chimeric constructs failed to induce cell death (Fig. 6b).

We also investigated whether anti-Id antibodies, induced by DNA vaccination, were able to cause tumour cell cycle arrest, as demonstrated for antibodies raised by protein immunisation [16, 18]. To address this question, the cell cycle distribution of BCL1 cells surviving the anti-Id sera treatment was analysed. A clear reduction in the number of viable cycling cells was observed, from 53% of control cells in S/G2/M phases to ~38% upon incubation with anti-Id^BCL1 antibodies (Fig. 6c). These results indicate that conformational DNA-induced anti-Id antibodies are able to stimulate both BCL1 tumour cell death and cell cycle arrest.

Detection of BCL1 cells in surviving mice

The observation that cell cycle arrest can take place in vitro in a fraction of BCL1 cells, suggests that dormant tumour cells could be present in surviving mice, as reported upon protein vaccination [8, 16]. To address this issue, total RNA from splenocytes of control, surviving (more than 1 year) and tumour-bearing mice was extracted in order to amplify the BCL1 V_H region, as previously described [15]. A fragment of 375 bp, from the leader sequence to the CDR3 region of BCL1 V_H, was amplified from all surviving (Fig. 6d, lanes 2, 3, and 4) and tumour-bearing mice (lanes 5 and 6), but not from control mice (lane 1), suggesting the presence of dormant IgM⁺ cells in the spleens of long-lived animals. By comparing with the positive controls (lanes 5 and 6), we established that these cells account for around 1% of the total splenocytes.

Discussion

The mechanism of lymphoma protection induced by anti-Id DNA vaccines is not yet well characterised. Despite different indications on the important role of anti-Id antibodies in tumour rejection [12, 22], there is also evidence of specific proliferation of CTLs generated from Id-vaccinated animals [1]. In addition, in some animal models, T-cell depletion experiments revealed evidence of the essential role played by CD4⁺/CD8⁺ T cells in lymphoma eradication [6].

Here, the contribution provided by T cells and antibodies to BCL1 tumour protection was dissected in vivo, taking advantage of the humoral response specificity previously reported in this model after DNA vaccination [3]. We immunised mice with two chimeric constructs containing either the BCL1 V_L or V_H domain, each one assembled with an irrelevant partner. Animals vaccinated with the two chimeric vaccines did not develop specific anti-Id^BCL1 antibodies. However, these mice were exposed to all putative Id-derived MHC class I and class II T-cell epitopes and developed specific anti-chimeric Id humoral responses.

Interestingly, only mice immunised with the complete Id^BCL1 were protected, whereas the delivery of either one or two chimeric constructs was not beneficial. These results demonstrated that the induction of anti-Id antibodies is a necessary condition for tumour protection and are in agreement with previous observations reported in this model upon protein immunisation [9, 16]. Indeed, passive immunisation of SCID mice with protein-induced anti-Id antibodies and depletion of natural killer (NK) cells was sufficient to induce BCL1 tumour protection [16]. These antibodies were found to be able to trigger tumour cell death and cell cycle arrest upon IgM cross-linking [16]. We investigated whether DNA-induced antibodies retain the same ability, despite their exclusive reactivity to V_L/V_H combined epitopes [3]. The ex vivo results indicated that antibodies elicited by DNA vaccination induce both tumour cell death and cell cycle arrest. These data suggested that the functionality of anti-Id antibodies is not impaired by DNA vaccination. In this view, it should be considered that DNA vaccination was found to be as efficient as protein immunisation in inducing protection, although antibodies titres were lower (data not shown). Recently, a role for idiotype-specific CD4⁺ T cells in tumour rejection has been reported in a transgenic model of B-cell lymphoma [14]. In our system, exclusive involvement of anti-Id CD4⁺ T cells in lymphoma protection can be excluded, since these lymphocytes should be activated by the immunisation with the chimeric constructs. Accordingly, while priming of CH3-specific CD4⁺ T cells was clearly evidenced, proliferation of CD4⁺ lymphocytes was not observed upon stimulation with the Id^BCL1 protein.

It has been proposed that CTLs may have a regulatory effect in maintaining the BCL1 tumour dormant state following protein immunisation with the tumoral IgM [8, 16]. Indeed, we detected residual tumour IgM⁺ cells in surviving mice, suggesting that DNA vaccination does not result in complete tumour rejection. Nevertheless, in vitro proliferation of CTLs specific for Id-derived epitopes was not evident in our model, and we did not observe regrowth of tumour in surviving mice. Interestingly, all these animals had high anti-Id antibody titres (not shown). Even though the activation of CTLs upon DNA immunisation cannot be ruled out, these lymphocytes would appear not to be effective in inducing tumour protection in the absence of anti-Id antibodies.

The fact that in this and other cases [12, 22, 24] anti-idiotype DNA vaccines failed to stimulate CD8⁺ lymphocytes could be dependent on the vaccine design or on a relative paucity of MHC class I epitopes in the idiotype sequences. Accordingly, it should be noted that in the IgM⁺ lymphomas in which V regions are in germline configuration, strong tolerance is expected for most of the peptides derived from the frameworks, CDR1 and CDR2 regions of Ig V domains.

Overall, this report demonstrates the role of DNA-induced anti-Id antibodies in tumour protection in animals with a fully functional immune system. In fact, depletion of CD4⁺ and/or CD8⁺ T cells was avoided. Hence, this is the first study on the mechanism of tumor protection performed with functionally intact cellular and humoral immune response components.

As the induction of conformational antibodies through DNA vaccination could be of general relevance, the same strategy could be used with any idiotype, regardless of the sequences of the V domains. On the contrary, the activation of T-cell responses requires the presence of T-cell epitopes in the lymphoma V domains compatible for presentation by the MHC I and MHC II alleles of the patient.

In this view, since anti-Id antibodies should be easily elicited for all lymphomas, strategies that favour the induction of humoral responses are likely to be broadly efficient and extensible to all surface Ig-positive B-cell tumours.

Abbreviations

CTL

Cytotoxic lymphocyte

DC

Dendritic cell

Id

Idiotype

Ig

Immunoglobulin

MHC

Major histocompatibility complex