New T cell epitopes identified from an anti-idiotypic antibody mimicking ovarian cancer associated antigen

Abstract

Anti-idiotype (Id) antibodies can be used to induce specific cellular immune responses against tumor antigens, but the mechanism of antigenicity is not always clear. We previously reported an anti-Id antibody, 6B11, which mimics human ovarian cancer associated antigen OC166-9. To explore the molecular basis of cellular immune response induced by 6B11, a panel of peptides derived from complementarity determining region (CDR) of 6B11 were synthesized. After a series of immunologic experiments, we found that the light chain CDR3 peptide and heavy chain CDR3 peptide were the MHC class I and class II epitopes of 6B11, respectively. The combination of MHC class I and class II epitopes is more effective than 6B11 in inducing specific cellular immune response against ovarian cancer. Our study provided the structural basis of antigenicity of 6B11. The identification of antigen-specific T cell eptitopes in 6B11 should facilitate the design of epitope-based vaccine against human ovarian cancer.

Introduction

Based on Jerne’s Network Theory, anti-idiotype (Id) antibodies can mimic the original antigens as surrogate antigen. Anti-Id antibodies could therefore be used as the source of antigen for active immunotherapy. In the last decades, anti-Id antibodies that mimic distinct tumor associated antigens (TAAs) of different histology or distinct determinants of a TAA have been used for active specific immunotherapy against human malignancies. Many of these anti-Id antibodies were able to induce measurable active immune responses in patients [1–7].

The active immune responses induced by anti-Id antibodies are mainly anti-anti-Id antibody response and delayed-type hypersensitivity (DTH) reaction, which contributes to the better prognosis of tumors. Immunization of naïve syngeneic mice with idiotypic antibody or anti-Id antibody could induce the antigen-specific Th and CTL cells against cancer, even though mice had never been exposed to the true tumor antigen [8]. Most likely, Id determinants in antibody trigger Id-specific B and T cell responses; the anti-Id antibody can then be presented further to activate anti-anti-Id specific immune responses. The antibody-based network might be important in maintaining T cells memory, especially in the absence of continuous antigen stimulation. Consistent with this hypothesis, a humoral and/or a cellular immune response to antigen have been induced in patients with melanoma immunized with anti-id monoclonal antibody (mAb), which mimic the HMW-MAA [9–14]. This immune response has been found to be associated with regression of metastases in some patients [10] and with disease-free interval and/or survival prolongation [9–12]. Similarly, a monoclonal anti-Id antibody ACA125 mimicking CA125 showed interestingly not only specific anti-anti-Id immune response in advanced ovarian cancer patients, which can specifically bind to CA125-positive ovarian cancer cells and inhibit the binding of the Ab1 murine mAb to CA125-positive cells, but also inducing of lysing CA125-positive ovarian cancer cells by peripheral blood lymphocytes from ovarian cancer patients immunized with ACA125 [15]. As previously reported, we prepared monoclonal anti-Id antibody 6B11 against a mAb, COC166-9, which recognizes the protective determinants on ovarian cancer associated antigen OC166-9. 6B11 inhibited the binding of OC166-9 to COC166-9, and it was classified as internal image antibody mimicking determinants of OC166-9 [16]. In further studies, 6B11 was found to induce specific antibody response and DTH reaction against OC166-9 in mice [17], which was associated with the prolonged life span of immune reconstituted SCID mice loaded with human ovarian cancer [18]. Furthermore, 6B11-primed HLA-A*2 T cells from healthy donors or ovarian cancer patients could kill OC166-9 positive ovarian cancer cells of the corresponding HLA subtype [19].

The basis for structural mimicry between antigen and the corresponding anti-Id antibody is unclear. One possibility is represented by the interactions of a TAA and corresponding anti-Id antibody with the same area of the antigen combining site of the anti-TAA antibody. These interactions are likely to involve side chains of amino acids, which can form strong H-bonds or even stronger electrostatic interactions between positively and negatively charged terminal groups of the side chains and reversely charged groups in antigen binding regions of antibodies and of T cell receptors.

Peptides derived from TAA may be presented by HLA class I molecules, but they are mainly weak inducer of CTLs [20]. However, peptides in anti-Id antibody vaccine induce CTLs, which can cross-recognize target antigen positive tumor cells. CTL precursors of unknown specificity are present during the induction of Id/anti-Id cascade [21]. The mechanisms by which anti-Id antibodies induce specific cellular immune response against tumor have not been fully investigated. Whether anti-Id antibodies express epitopes recognized by T cells remains to be determined.

In this study, we have examined the structural basis of 6B11 to induce ovarian cancer-specific immune response. We found that 6B11 could induce T cell immune response in HLA-A*2 ovarian cancer patients. The potential HLA-A*0201 ligands within the CDR region of both light and heavy chain of 6B11 were predicted by using SYFPEITHI algorithm and tested in T2 binding assay for HLA-A*2 binding. Then we investigated whether these epitopes on 6B11 were indeed effective in the induction of anti-tumor cellular immune responses. The results show that the peptides in heavy chain CDR3 (VH CDR3) and light chain CDR3 (VL CDR3) of 6B11 are associated with the mimicry of antigen to induce specific cellular immune responses. We provide a possible explanation for the phenomenon that certain antibodies, especially anti-Id antibodies, have the potential to induce specific cellular immune response against original antigen.

Materials and methods

In vitro stimuli and cells

6B11 is a murine anti-Id antibody of IgG1 subclass that mimics protective determinants of ovarian cancer associated antigen OC166-9. 6B11 was prepared as previously reported [22]. Synthesized by solid-phase method, peptides of 6B11 CDRs were purified by high-performance liquid chromatography, and their sequences were confirmed by mass spectrometry. HOC1A cells [23] established in our laboratory is human HLA-A*2 and OC166-9 positive ovarian adenocarcinoma cell line. K562 is human chronic myelogenous leukemia cell line. HLE deriving from ATCC is human HLA-A*2 positive and OC166-9 negative hepatoma cell line. T2 cell is human T-B hybrid which only expresses low levels of HLA-A*2 molecule. HOC1A was grown in MCDB culture medium, and other cell lines were grown in RPMI-1640 culture medium supplemented with 10% heat inactivated fetal calf serum (FCS), 2 Mm glutamin, penicillin (100 U/ml) and streptomycin (100 μg/ml).

Epitope prediction

Prediction of potential HLA-A*0201 ligands was carried out as described [24]. Briefly, CDRs peptides of 6B11 were screened against a matrix pattern, which evaluated every amino acid within nonamer peptides fitting the HLA-A*0201 motif. Anchor residues were given values of 1; other residues 0–1, reflecting amino acid preferences for certain positions within the peptide. Such motif predictions are available using the database SYFPEITHI (http://www.uni-tuebingen.de/uni/kxi/).

Peptides binding assay

Peptides were incubated with T2 cells in RPMI-1640 culture medium supplemented with 5% FCS overnight. The final concentration of peptide was 0.3, 3 or 30 μg/ml. The surface expression of HLA molecule on T2 cells was monitored after staining with the primary W6/32 monoclonal antibody (SEROTEC, Oxford, England) and a FITC-coupled goat anti-mouse IgG antibody (Sigma) on confocal microscopy (Leica, Heidelberg, Germany). The results were presented as follow: – means no binding detectable at any peptide concentration, + means binding detectable at only the highest peptide concentration (30 μg/ml), ++ means binding detectable starting at the 3 μg/ml peptide concentration and +++ means binding detectable starting at the 0.3 μg/ml peptide concentration. The ability of peptides to stabilize HLA-A*2 on the surface of T2 cells was assessed by fluorescent activated cell sorter (FACS) (Becton Dickinson, Mountain View, CA) after staining peptide-treated cells with anti-HLA-A*2 antibody. Five T2 binding positive peptides, NNKYYNTAL (NNK), LITTKIAWY (LIT), IALITTKIA (IAL), VHSNGNTYL (VHS), SQSTHFPYT (SQS) were evaluated for binding to HLA-A*2. A synthetic peptide capable of causing a >150% increase in Mean Fluorescence Intensity (MFI) values over the negative control was considered as a “high binder”. A well-known HLA-A2*1 binder peptide, GILGFVFTL derived from influenza matrix antigen and a nonbinding peptide, IAGNSAYEY were used as positive and negative controls, respectively. Each peptide was used at 50 μM concentration.

Monocyte isolation

Fresh, whole blood was drawn with informed consent from HLA-A*2 healthy donors into vacutainer tubes (Becton Dickinson, Mountain View, CA) containing EDTA. Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll–Paque (Amersham Biosciences, Uppsala, Sweden), as previously described [25]. For monocyte isolation by plastic adherence, 1 × 10⁷ PBMCs in 1 ml AIM-V culture medium (Invitrogen, San Diego, CA) per well were distributed into 6-well plates (CorningCostar, Corning, NY) and allowed to adhere at 37°C for 2 h. The AIM-V culture medium contained 5% human AB serum (heat inactivated, Krackeler Scientific, Albany, NY) (referred to as complete medium, 5% CM). Non-adherent cells were removed as a source of lymphocytes and adherent cells were washed carefully, twice, with prewarmed 5% CM as monocytes.

DC generation and 6B11 or peptide loading

Monocytes were cultured in 6-well plates containing 2 ml 5% CM per well supplemented with 1000 U/ml recombinant human (rh) granulocyte-macrophage colony-stimulating factor (GM-CSF) and rh IL-4 (R&D Systems, Minneapolis, MN) for the generation of immature DCs. Every 2 days 500 μl of the medium was exchanged with 5% CM containing above cytokines. Day 6 immature DCs were matured using 1000 U/ml of rh TNF-α, rh IFN-α, rh IL-6 and 1 μg/ml of PGE2 (R&D Systems) for 24 h. The surface expression of CD80, CD86, CD83, CD1a and HLA-DR molecules on mature DCs were analysed by FACS. All mAbs and their isotype-matched negative controls were purchased from BD. DCs were harvested, washed, pulsed with 30 μg/ml peptides or 40 μg/ml 6B11 and irradiated (5000 rad) prior to being used as stimulators.

Generation of 6B11-CTLs or peptide-CTLs

The protocol used here is a modification of the method described by Plebanski [26]. 6B11-CTLs were generated by one or three rounds of in vitro stimulation (IVS) with 6B11-pulsed DCs. Non-adherent lymphocytes from above were added in bulk (CD4 T, CD8 T, NK, etc.) to 6B11-pulsed DCs in 5% CM. Cultures were re-stimulated with 6B11-pulsed DCs every 7 days. Peptide-CTLs were generated by three rounds of IVS with peptide or KLH-pulsed DCs. Non-adherent lymphotytes were added in bulk to DCs loaded with peptide or 5 μg/ml of KLH in 5% CM. Cultures were re-stimulated every 7 days. For each stimulation T:DC ratio was set at 10. On day 3 of each IVS cycle, 10 U/ml of rh IL-2 (R&D Systems) was added.

After one round of IVS, the 6B11-CTLs were harvested for the proliferation assay and IFN-γ ELISPOT assay. After three rounds of IVS, 6B11-CTLs and peptide-CTLs were harvested as effector cells and used in ⁵¹Cr-release assay. The CD4 components of 6B11-CTLs, the CD4 or CD8 components of peptide-CTLs were separated by negative magnetic beads sorting, then used in cell proliferation and IFN-γ ELISPOT assay.

Bromodeoxyuridine (BrdU) ELISA cell proliferation assay

The proliferation of 6B11-CTLs and 6B11 specific CD4 T cells (1 × 10⁵/well) was measured by an ELISA-based BrdU incorporation assay after stimulation with peptide-pulsed DCs at 37°C in 5% CO₂ incubator for 120 h. The T:DC ratio was set at 10. Cultured cells were incubated with 10 μl/well BrdU labeling solution for the last 2 h of the stimulation and removed the labeling medium to dry cells at 60°C for 1 h. Dry cells were added 200 μl/well FixDenat at 37°C for 30 min and incubated with 100 μl/well mAb against BrdU for 90 min. The BrdU incorporation was measured directly using substrate colorimetrics by an automatic microplate reader (Hercules, CA) at a test wavelength of 450 nm and a reference wavelength of 490 nm. Unpulsed DCs and PHA were used as negative and positive control of the experiments, respectively.

The MHC restriction of the proliferative response was demonstrated by preincubating the stimulators with anti-human MHC class II antibody (SEROTEC) at a final concentration of 1:50. Experiments were performed at least three times in triplicate for all samples.

⁵¹Cr-release cytotoxicity assay

The standard 4 h ⁵¹Cr-release assay was performed in 96-well V-bottomed microplates. Target cells in suspension (T2 and K562 cells) were labeled with 100 μCi Na⁵¹₂ CrO₄ (Amersham Life Science, Amersham, UK) per 1 × 10⁶ cells either overnight in 0.5 ml RPMI 1640 culture medium containing 5% FCS or for 90 min at 37°C directly with the cell pellets in the case of adherent cells (HOC1A and HLE cells). Labeling was terminated by washing the targets with cold media containing 5% FCS for a total of three washes. To each well containing effector CTLs in same volume seeded at different E:T ratios, 5 × 10³ targets in 100 μl culture medium were delivered. Spontaneous release wells contained targets in media alone, while maximal release wells contained targets in medium of 1% Triton X-100 detergent. The MHC restriction of CTLs mediated killing was demonstrated by preincubating the targets with anti-human MHC class I antibody at a final concentration of 1:40 (SEROTEC).

The plates were gently spun for 2 min and incubated at 37°C for 4 h. For harvesting assay, 100 μl of supernatants from each well was transferred to corresponding counting tubes and γ counts were determined by a γ counter (Nuclear instruments factory, XiAn, China). Cytolytic activity of CTLs was expressed in the percentage of specific lysis as follows: specific lysis = [(experimental release – spontaneous release) / (maximal release – spontaneous release)] × 100. The ratio of spontaneous release to maximal release was <0.2 in all experiments. All cytolytic analyses described in this study were performed in triplicate and repeated at least three times in separated experiments.

Cytokine assay

6B11-CTLs or peptide-CTLs were incubated with HOC1A or HLE cells at a ratio of 40:1 in 37°C for 48 h. Supernatants of the cultures were harvested, centrifuged at 3000g and stored at −20° until required. Levels of IL-2, IL-4 and IFN-γ were assayed with sandwich ELISA kits (Peprotech, London, UK) according to the manufacturer’s instruction. Biotinylated antibody was added and detected with avidin-peroxidase plus 2,2-azino-di-[3-ethyl-genzthiazoline sulfonate] substrate containing H₂O₂. The colorimetric reaction was read at a test wavelength of 450 nm using microplate reader. The concentrations of the cytokines were calculated from the standard curve of the corresponding recombinant cytokine. All tests were performed in duplicate and repeated at least three times.

IFN-γ ELISPOT assay

The protocol for IFN-γ ELISPOT assay (Diaclone, Besancon, France) was used with some modifications, as described by the manufacturer. Briefly, 96 polyvinylidene difluoride-bottom-well filtration plates were coated overnight with anti-IFN-γ capturing antibody at 4°C. HOC1A or HLE cells as stimulators were incubated with different effector cells at a ratio of 1:20. The captured IFN-γ was then revealed by a secondary anti-IFN-γ antibody conjugated to alkaline phosphatase which was, in turn, detected by BCIP/NBT substrate and spots were counted with an automatic reader (CTL, Cleveland, OH).

Results

Screening HLA-A*0201 binding peptide

Using the SYFPEITHI algorithm, we predicted HLA-A*0201 binding peptides from CDRs of 6B11 (Table 1). Peptides with a score >6.0 were synthesized, which included 4 peptides in VH CDR2, 4 peptides in VH CDR3, 4 peptides in VL CDR1 and 1 peptide in VL CDR3 (Table 1). Because different MHC class I molecules prefer a different length of ligands, SYFPEITHI algorithm predicts that either nonamer or decamer peptides will bind to HLA-A*0201. The octamer peptides of VH CDR1 and VL CDR2 were also synthesized for additional experiments.

Table 1.

Screening HLA-A*0201 binding peptides from CDRs of 6B11

CDRs	CDRs sequences	Predicted peptides	SYFPEITHI score	T2 binding
VH CDR 1	PTYGIGVG	P T Y G I G V G	0	−
VH CDR2	HIWWNNNKYYNTALKS	H I W W N N N K Y	12	−
		N N K Y Y N T A L	10	+++
		N N N K Y Y N T A	9	−
		W N N N K Y Y N T	7	−
VH CDR 3	PIALITTKIAWYFDV	L I T T K I A W Y	16	+
		A L I T T K I A W	15	−
		I A L I T T K I A	13	+
		T K I A W Y F D V	12	−
VL CDR1	RSSQNLVHSNGNTYLH	N L V H S N G N T	14	−
		V H S N G N T Y L	13	+
		S S Q N L V H S N	10	−
		L V H S N G N T Y	9	−
VL CDR2	PIVSNRFS	P I V S N R F S	0	−
VL CDR3	SQSTHFPYT	S Q S T H F P Y T	7	++

Although the predictive algorithm shows some degrees of successes in identifying immune stimulating sequences, there may be poor correlation between predicted and actual binding activities to MHC class I molecule, and even worse in stimulating an effective CTL response [27]. The T2 cell-binding assay [28] was employed to validate the actual binding affinity of the above peptides to HLA-A*2 molecules. The accumulation of HLA-A*0201 molecules at the T2 cell surface recorded by using indirect antibody staining and Confocal microscopy analysis reflects the capacity of exogenous peptides to bind and stabilize this MHC class I molecule. In Table 1, “T2 binding” indicates the affinity of each peptide for HLA-A*2 in T2 binding assay. High, intermediate and low affinity means that the surface expression of HLA-A*0201 molecules is stable at a peptide concentration of 0.3, 3, and 30 μg/ml, which are expressed as +++, ++, and + respectively. Absence of surface accumulation of HLA-A*0201 molecules at a peptide concentration of 30 μg/ml suggests no binding, which is expressed as –. At the end, one peptide with high affinity, one peptide with intermediate affinity, and three peptides with low affinity were found, while the rest of the peptides had no binding activity. In FACS assay, an increase in MFI values > 200% was noted with all these peptides. This increase was similar to that observed using the positive control peptide. All results are consistent with those of T2 binding assay except the VHS peptide (Fig. 1).

Fig. 1.

Confocal laser scanning microscopy and FACS detects the MHC class I molecule surface expression of T2 cells. T2 cells were incubated with 30 μg/ml of each peptide for 24 h respectively. The surface expressions of HLA-A*0201 molecules at T2 cells were recorded using indirect antibody staining and confocal analysis. a high affinity peptide, b intermediate affinity peptide, c low affinity peptide, d zero affinity peptide, e single T2 cell control. The ability of peptides to stabilize HLA-A*2 on the surface of T2 cells was assessed by FACS after staining peptide-treated cells with anti-HLA-A*2 antibody. Data are showed as MFI f

Only peptides with a score > 6.0 are listed in column “Predicted peptides”. The peptides presented in the figure were synthesized and tested in a T2 binding assay. Column “T2 binding” indicates the affinity of the peptides.

Proliferative response of 6B11-primed T cells induced by 6B11 CDRs peptides

DCs were generated, and the phenotypes were analyzed by FACS to test the maturation status of DCs. High levels of CD80, CD86, CD83 and HLA-DR molecules were found to be expressed on these DCs (Fig. 2a–d), indicating that these DCs were in the matured state [29].

Fig. 2.

Phenotypic analysis of mature DCs and detection of the 6B11 CDRs peptides responsible for inducing proliferative response of 6B11-primed T cell. a–d Immature DCs were matured with rh IFN-α , rh TNF-α, rh IL-6 and PGE2 for 24 h. Expression levels of several cell surface markers on mature DCs were determined by FACS analysis. Data are representative of one experiment out of four performed. e 6B11-CTLs were in vitro stimulated with PHA, 6B11 or peptide-pulsed DCs and unpulsed DCs preincubated with anti-MHC class II antibody or not at a ratio of 10:1 for 120 h, and the proliferative response was measured using BrdU ELISA. f CD4 T cells of 6B11-CTLs were separated by magnetic beads sorting and used as responder cells for cell proliferation assay

To explore the possible Th epitope in CDRs of 6B11, we generated T cells (6B11-CTLs) in vitro by priming peripheral blood mononuclear cells with 6B11 and investigated the ability of DCs pulsed by each original CDRs peptide to induce proliferation of 6B11-CTLs (Fig. 2e). Phytohemagglutinine (PHA), as positive control, induced high levels of proliferative response of 6B11-CTLs. The 6B11 or each original CDRs peptide-pulsed DCs were used as stimulating cells. Similar to the 6B11-pulsed DCs, both VH and VL CDR3 peptides-pulsed DCs, but not DCs pulsed with other peptides, induced significant proliferation response of 6B11-CTLs; the VH CDR3 peptide induced much higher proliferation level than that of the VL CDR3 peptide. Furthermore, the VH CDR3 peptide-induced proliferative response could be blocked by anti-MHC class II-antibody, indicating that the proliferation was mainly CD4 T cell dependent; The VL CDR3 peptide related proliferative response could not be blocked by anti MHC class II antibody, which suggests that the proliferative response was independent of CD4 T cell. Other peptide-pulsed DCs induced negligible proliferative response similar to that of unpulsed DCs. Then we separated the CD4 T cells of 6B11-CTLs and used them as responder cells in proliferation assay to test the above results further (Fig. 2f). We found that the VH CDR3 peptide-pulsed DCs could duplicate the ability of 6B11-pulsed DCs to induce a higher percentage of proliferative response, while the VL CDR3 peptide-pulsed DCs lost the ability to induce the proliferative response of the primed CD4 T cells. These results confirmed once again that the VH CDR3 peptide related proliferative response was CD4-dependent, while the VL CDR3 peptide related proliferative response was independent of CD4 T cells.

Cytotoxic activity of 6B11 or 6B11 CDRs peptide-specific CTL lines

To identify the potential peptides contributing to the induction of cytotoxic CTLs, we primed autologous lymphocytes using 6B11 or peptide-pulsed DCs for three cycles to generate 6B11-CTL lines and peptide-primed CTLs lines (peptide-CTLs), then they were tested for cytotoxicity using standard ⁵¹Cr-release cytotoxicity assay (Fig. 3). Peptide-pulsed T2 cells, HOC1A (HLA-A*2 + , OC166-9 +  ovarian cancer cell line) and HLE cells (HLA-A*2 + , OC166-9- hepatoma cell line), preincubated with anti-MHC class I antibody or not, were used as target cells (Fig. 3a–j). We found that VH CDR3 peptide-CTLs could kill T2 cells pulsed with VH CDR3 peptide and HOC1A targets, and the cytotoxic activity was independent of MHC class I. VL CDR3 peptide-CTLs elicited cytotoxicity against T2 cells pulsed with VL CDR3 peptide and HOC1A targets. Such a cytotoxic activity was MHC class I molecule dependent as it was blocked totally by anti–MHC class I antibody. Consistent with the observed results, 6B11-CTLs could kill both VH or VL CDR3 peptide-pulsed T2 cells and HOC1A targets, which were dependent partially on MHC class I molecule. There was no significant cytotoxicity against HOC1A targets with T cells induced by other CDR-derived peptides (Fig. 3).

Fig. 3.

Cytotoxicity of 6B11 CDRs deriving peptides-specific CTLs. Lymphocytes from HLA-A*2 donors were cocultured with 6B11 or peptide and KLH-pulsed DCs at a ratio of 10:1 for 3 cycles. a–j The cytotoxic activities of each peptide and KLH-primed CTLs against this peptide loaded T2 cells (column-dotted), HOC1A (column-grey shading), HLE (column-no fill) or corresponding targets preincubated with anti-MHC class I antibody (column-diagonal striped, column-horizontal striped, column-netted designed) were tested at different E:T ratios. The cytotoxic activities of each peptide and KLH-primed CTLs against unloaded T2 cells (column-grey shading) and unpulsed DCs-primed T cells (column striped) or naïve T cells (column black fill) against each peptide loaded T2 cells were also tested as negative controls. k The cytotoxic activities of VH or (and) VL CDR3 peptide-CTLs and 6B11-CTLs against HOC1A, HLE or K562 were tested at different E:T ratios. The cytotoxic activities of unpulsed DCs-primed T cells (column-grey shading)and naïve T cells (column diamond designed) against HOC1A were also tested as negative controls

In the next experiments, we tested the specificity and effectiveness of cytotoxic activities induced by VH or VL CDR3 peptide-CTLs (Fig. 3k). VH CDR3 peptide-CTLs killed both HOC1A and NK sensitive K562 cells, which showed that part of the cytotoxic activity was nonspecific and might reflect the activity of NK cells. In contrast, the VL CDR3 peptide-CTLs could only kill HOC1A targets, but not K562 cells, demonstrating that the activity is highly antigen specific and unlikely to be mediated by NK cells.

Cytokines profile produced by 6B11-CTLs and peptide-CTLs

6B11 or CDR peptide-CTLs were incubated with HOC1A or HLE cells for 48 h, followed by detection of cytokines released in the supernatant (Fig. 4). 6B11-CTLs, VH CDR3 peptide, VL CDR3 peptide or VH and VL CDR3 peptide-CTLs all secreted high levels of IL-2 and IFN-γ in the presence of HOC1A targets, while the levels of IL-2 and IFN-γ were negligible in the presence HLE targets. The difference was statistically significant (P < 0.01). The levels of IL-4 in different groups of T cells were similar when stimulated with either HOC1A or HLE targets.

Fig. 4.

Comparison of the cytokines profile induced by 6B11 and identified peptides. 6B11, VH CDR3 peptide, VL CDR3 peptide or VH and VL CDR3 peptides-specific CTLs were cocultured with HOC1A and HLE cells for 48 h. The unpulsed DCs-primed T cells and naïve T cells were used as negative controls. The supernatants were harvested to detect the production of IL-2, IFN-γ and IL-4 with sandwich ELISA

Functional IFN-γ producing cells in both 6B11-CTLs and CDR3 peptide-CTLs

6B11-CTLs, VH and VL CDR3 peptide-CTLs, CD4 or CD8 components of VH and VL CDR3 peptide-CTLs were incubated with different targets for 24 h, the percentage of functional IFN-γ producing cells were measured using an IFN-γ enzyme linked immunosorbent spot-forming cell assay (IFN-γ ELISPOT assay) (Fig. 5). As shown, for HOC1A targets the percentage of IFN-γ producing cells in VH and VL CDR3 peptide-CTLs, CD4 or CD8 CTLs components were significantly higher than that in T cells primed by unpulsed DCs or culture medium. The frequency of IFN-γ producing cells in CDR3 peptide-CTLs was similar to that in 6B11-CTLs. The background responses against HLE targets was low in all groups (all P > 0.05).

Fig. 5.

Similar frequency of IFN-γ producing cells presents in 6B11-CTLs and CDR3 peptides-primed CTLs. VH and VL CDR3 peptides specific CTLs, CD4 or CD8 CTLs components were cocultured with HOC1A and HLE cells at a ratio of 20:1 for 24 h, then all cells were discarded. The forming spots representing the IFN-γ producing cells were counted. 6B11-CTLs and unpulsed DCs-primed T cells were used as positive and negative controls respectively

Discussion

When 6B11 is injected into host or tested in vitro, the exogenous protein is likely to be internalized and degraded into peptides by antigen-presenting cells. The degraded peptides bound to proper MHC molecules are presented to T cell by the antigen-presenting cells. T cells with appropriate receptors are expanded and expected to constitute the cytotoxic, helper and memory T cells against 6B11 or original antigen. Because T cells primed by 6B11 can recognize OC166-9 positive ovarian cancer cells, it is necessary to present the peptides in 6B11, which are responsible for the induction of specific T cells against OC166-9. CDRs are the docking sites of Id or anti-Id, so they most likely contain the functional epitopes of antibody. Our previous studies showed that 6B11-ScFv, an anti-Id antibody containing VH, VL CDRs of 6B11 and a linker, could prime T cells to kill OC166-9 positive ovarian cancer cells [30]. The 6B11 CDRs were likely to possess the epitopes responsible for the induction of cellular immune responses against ovarian cancer. In this study, the peptides in CDRs of 6B11 were synthesized and screened for the T cell epitopes against COC166-9 (Id) which specifically recognizes an ovarian cancer antigen OC166-9.

We find that VH and VL CDR3 peptides are the Th and CTL epitopes, respectively, mimicking protective determinants of OC166-9. VH CDR3 peptide as a Th epitope of 6B11 could induce the proliferation of 6B11 or VH CDR3 peptide-primed CD4 T cells, and the proliferative response could be blocked significantly by anti-MHC class II antibody, but not by anti-MHC class I antibody. These results are consistent with the previous studies. 2F10 is an anti-Id antibody capable of mimicking determinants of hepatitis B surface antigen. A 15-amino acid sequence within the VH CDR3 of 2F10 was identified as B and Th-epitope which had significant homology to a determinant of hepatitis B surface antigen [31]. In the CEA system, 3H1, an ati-Id antibody mimicking CEA, and CEA were found to have homologous sequences in both direct and reverse orientations. Peptides based on the homologous sequences were synthesized for analysis. A peptide within VL CDR2 of 3H1 could induce the proliferative responses of PBMCs from a group of 3H1-immunized CEA-positive cancer patients, and as a Th epitope it mainly acted on CD4 T cells [32]. Since antibody response based on anti-Id antibody is T cell dependent, it is reasonable that anti-Id antibody contains Th epitope. Nevertheless, we also found that VL CDR3 peptide was a functional CTL epitope, which is the first ovarian cancer related CTL epitope identified from an anti-Id antibody to our knowledge. VL CDR3 peptide-CTLs could bind to HLA-A*2 positive T2 cells and kill OC166-9 positive ovarian cancer cells of HLA-A*2 subtype, but they lacked cytotoxic activity to OC166-9 negative cells of HLA-A*2 subtype. Thus, 6B11, as an anti-Id antibody, contains both Th and CTL epitopes that mimic a yet unidentified ovarian cancer antigen. Similarly, a CTL epitope against melanoma was identified recently from MF-11-30, an anti-Id antibody mimicking high molecular weight melanoma associated antigen [33]. These results challenge the traditional theory that anti-Id antibody as anti tumor vaccine is based on mimicking B cell epitope of the original antigen.

An effective vaccine generally includes a means to induce helper responses, assisting CTL activation and supporting memory responses essential to overcome minimal residual disease in tumor patients [34]. It is often difficult to generate helper T cell responses against self antigens. Vaccines based on self antigens are thus best supported by including xenogeneic helper epitopes derived, for example, from green fluorescent protein (GFP), KLH or tetanus toxoid (TT) [35–37]. Xenogeneic help may not be required for antigens expressed in the context of malignant transformation, as shown in clinical trials for vaccination against tumor antigen MAGE, where KLH-supplemented vaccination made no difference for clinical outcome [38]. There are examples for tumor derived or related helper peptides [39]. In this regard, some helper peptides have been identified from prostatic acid phosphatase and 1 from NY-ESO, expressed in several cancers [40, 41]. Our experiments showed that lymphocytes primed by VL CDR3 peptide collaborating with VH CDR3 peptide or KLH could kill corresponding peptide-pulsed T2 or target antigen positive cells, and the former collaboration is more potent in inducing cytotoxicity. Although VH and VL CDR3 peptides are Th and CTL epitopes derived from an anti-Id antibody respectively, they may be considered as a set of related epitopes naturally generated in tumor. These results should facilitate 6B11-based vaccine design as the related helper peptide is more effective vaccine because T cell help targeting the same TAA may be more specific or more effective than help against foreign antigens.

According to the cytokines profile, CD4 T cells have been classified into Th1 and Th2 subsets [42]. These subsets have distinctly functional characteristics-Th1 cells producing IL-2 and IFN-γ, and are associated with cell mediated immunity and inhibit Th2 cells, while Th2 cells producing IL-4, assisting B cells to make antibody and inhibiting Th1 cells. CD8 T cells consisting of comparable populations, Tc1 and Tc2, respectively secret IL-2/IFN-γ and IL-4 [43, 44]. We exposed VH and (or) VL CDR3 peptides-CTLs to different targets, and then investigated the condition of cytokines profile. For HOC1A targets, 6B11-CTLs,VH CDR3 peptide, VLCDR3 peptide or VH and VL CDR3 peptides-CTLs all secreted high levels of IL-2 and IFN-γ, while the levels of IL-4 were undetectable. The immune responses skewed towards Th1 and Tc1 polarization. For HLE targets the levels of IL-2, IFN-γ and IL-4 were similar among different groups of T cells. In further experiments, IFN-γ ELISPOT assay confirmed that both 6B11-CTLs and VH and VL CDR3 peptides-CTLs contained specific responding cells which secreted IFN-γ as an effector cytokine. These results show that VH and VL CDR3 peptides stimulate the same kind of CTLs as 6B11 and induce the Th1 and Tc1 polarization. The VH and VL CDR3 peptides could totally duplicate the biological activity of intact 6B11 to induce specific cellular immune responses against ovarian cancer, which forms the basis of cellular immune responses induced by 6B11.

Anti-Id antibody [45] or xenogeneic protein [46] as anti-tumor vaccine could induce stronger immune response against original TAA itself. This difference may be inherent, for example, as an anti-Id antibody is an “internal image antigen”, which is expressed in a different molecular environment and may thus overcome the immunosuppression in the host by stimulating silent clones and/or by allowing T cell help to become active, making the overall immune response stronger [47]. Another reason for the increased anti-tumor activity by anti-Id antibody relies on the predictability of immune response against tumor. Antigen is an extremely heterogeneous molecule, thus antigen based vaccine presenting multiple epitopes could potentially involve a mixture of counter immune responses. Similar phenomenon was observed when intact HER-2/neu antigen was used as the immunogen. Furthermore, our previous studies show that 6B11 as anti-Id antibody can induce both humoral and cellular immune responses, which are expected to be more effective in tumor therapy.

In summary, our studies confirm that 6B11 is the first ovarian cancer anti-Id antibody, which contains both Th and CTL epitopes. This is different than the traditional theory that anti-Id antibody as anti-tumor vaccine is based on mimicry of B cell epitope of original antigen and provides a alternative theory for anti-tumor vaccine development. The studies may also help us to better understand the anti-Id antibody induced anti-tumor cellular immune responses. The acquired peptides as epitope based vaccine could be used in immunotherapy of ovarian cancer in the future.

Abbreviations

anti-Id

Anti-idiotype

CDR

Complementarity determining region

TAA

Tumor associated antigen

IVS

In vitro stimulating

KLH

Keyhole limpet hemocyanin

BrdU

5-Bromo-2′-deoxyuridine

ELISPOT

Enzyme-linked immunosorbent spot-forming cell assay