Generation of human tumor-specific CTLs in HLA-A2.1–transgenic mice using unfractionated peptides from eluates of human primary breast and ovarian tumors

Abstract

HER-2/neu oncoprotein is overexpressed in a variety of human tumors and is associated with malignant transformation and aggressive disease. Due to its overexpression in tumor cells and because it has been shown to be immunogenic, this protein represents an excellent target for T-cell immunotherapy. Peptide extracts derived from primary HLA-A*0201-positive (+) HER-2/neu ⁺ human tumors by acid elution (acid cell extracts (ACEs)) were tested for their capacity to elicit in HLA-A*0201 transgenic mice, cytotoxic T lymphocytes (CTLs) lysing HLA-A*0201⁺ HER-2/neu ⁺ tumor cells. Injections of ACE in transgenic mice induced CTLs capable of specifically lysing HER-2/neu ⁺ tumor cell lines (also including the original HER-2/neu ⁺ primary tumor cells from which the ACEs were derived) in an HLA-A*0201–restricted fashion. Adoptive transfer of ACE-induced CTLs was sufficient to significantly prolong survival of SCID mice inoculated with HLA-A*0201⁺ HER-2/neu ⁺ human tumor cell lines. Cytotoxicity of such ACE-induced CTL lines was directed, at least as detected herein, also against the HER-2/neu peptides HER-2 (9₃₆₉) and HER-2 (9₄₃₅) demonstrating the immunodominance of these epitopes. HER-2 peptide–specific CTLs generated in the HLA-A*0201–transgenic mice, upon peptide immunization, lysed in vitro HER-2/neu ⁺ human tumor cell lines in an HLA-A*0201–restricted manner and, when adoptively transferred, conferred sufficient protection in SCID mice inoculated with the same human tumor cell lines as above. However, CTLs induced by ACEs displayed enhanced efficacy in the therapy of xenografted SCID mice compared with the HER-2 peptide–specific CTLs (i.e., HER-2 [9₃₆₉] or HER-2 [9₄₃₅]). Even by administering mixtures of CTLs specific for each of these peptides, the prolongation of survival achieved was still inferior compared with that obtained with ACE-induced CTLs. This suggested that additional epitopes may contribute to the immunogenicity of such tumor-derived ACEs. Thus, immunization with ACEs from HER-2/neu ⁺ primary tumor cells appears to be an effective approach to generate multiple and potent CTL-mediated immune responses against HER-2/neu ⁺ tumors expressing the appropriate HLA allele(s). By screening ACE-induced CTL lines with synthetic peptides encompassing the HER-2/neu sequence, it is feasible to identify immunodominant epitopes which may be used in mixtures as vaccines with enhanced efficacy in both the prevention and therapy of HER-2/neu ⁺ malignancies.

Introduction

Cytotoxic T lymphocytes (CTLs) play an important role in antitumor immune responses [1]. These cells recognize eight to ten amino acid–long peptides derived from tumor antigens presented by MHC class I molecules on the surface of tumor cells [2]. Such CTLs are capable of lysing tumor cells, possess memory, and can respond upon rechallenge with the same antigenic peptide [3]. Tumor-specific CTLs have been useful for the identification of several tumor antigens, which can be classified as follows: (a) tissue-specific shared differentiation antigens expressed in melanomas and melanocytes, (b) tumor-specific shared antigens expressed in several tumors of distinct histology, (c) unique/mutated tumor antigens, and (d) translocated and overexpressed oncoproteins [4].

The main advantage in using defined tumor antigens in peptide-based vaccination strategies for cancer immunotherapy is that the purity of the antigenic preparation is likely to enhance the effectiveness of the vaccine and, at the same time, minimizes the possibility of autoimmune reactions due to the absence of additional tumor material (i.e., self-proteins shared between tumor cells and healthy tissue). On the other hand, the use of defined CTL epitopes in peptide-based vaccines suffers from major drawbacks. First, it is not clear which of the identified tumor-specific antigens can mount in vivo an efficient antitumor response. To this end, Anichini et al. [5] demonstrated that CTLs in HLA-A2.1 melanoma patients did not recognize any of the known melanoma-specific Melan-A/MART-1, tyrosinase or MAGE antigens. Second, although immunization with a tumor antigen–derived peptide increases the frequency of CTLs specific for this particular peptide [6], this does not necessarily correlate with recognition of tumor cells presenting naturally processed epitopes of the relevant tumor antigen [1, 7]. Another serious problem that may arise with the use of vaccines consisting of single-peptide antigen is the risk of generating tumor escape variants [8].

The use of tumor cell lysates or peptide extracts isolated from tumor cells as a possible source of tumor antigen(s) for vaccination studies represents a strategy that is not affected by these limitations. The advantage of using tumor-derived peptides as a source of tumor antigen is that the presence of multiple immunogenic peptides will induce several CTL clones of different peptide specificity which will mount an effective antitumor response. The plethora of such tumor peptide-specific CTL clones will reduce the risk of generation of escape variants since tumor cells will be targeted by multiple antigens restricted by multiple HLA alleles. In addition, peptide mixtures isolated from tumor cells bypass the need for characterizing tumor-specific antigens and allow the application of vaccination protocols to several types of cancer even in cases where tumor-specific peptides have not yet been identified. The potential drawback of vaccinating cancer patients with tumor lysates or unfractionated peptides thereof is the induction of autoimmune reactions directed against self-antigens expressed on healthy tissue(s).

The use of tumor lysates or total peptide extracts as global antitumor vaccines pulsed on dendritic cells (DCs) represents a widely applied immunization strategy in various preclinical models [4, 9–16]. In clinical settings, vaccination of patients with advanced melanoma [17] or renal cell carcinoma [13] with tumor lysate-pulsed autologous DCs induced antitumor immunity in vivo, which was associated with measurable DTH reactivity and clinical responses.

In our recent report [18], we have demonstrated that acid cell extracts (ACEs) from patients’ tumor cells also contain peptides recognizable by autologous T_H cells and CTLs in the context of MHC class II and class I molecules, respectively, and that this CD4⁺ recognition results in enhanced killing of the autologous tumor by the CTLs. In a later study [19], by extracting peptides from HER-2/neu–expressing tumor cells we could induce, in vitro, several HER-2/neu peptide-specific CTL lines and clones capable of lysing their autologous HER-2/neu ⁺ tumor cells.

In the present study, we immunized HLA-A*0201–transgenic HHD mice with ACEs from HER-2/neu–overexpressing tumors isolated from pleural effusions from HLA-A*0201⁺ patients with metastatic disease. By repetitive in vitro stimulations with ACE-pulsed DCs, we generated CTL lines recognizing HER-2/neu peptides in an HLA-A*0201–restricted fashion. Such CTL lines exhibited strong in vitro cytotoxicity against human HER-2/neu ⁺ HLA-A*0201⁺ tumor cell lines and significantly inhibited their growth in SCID mice. These data demonstrate the immunodominance of HER-2/neu peptides in ACEs isolated from HER-2/neu–overexpressing human tumors. They also suggest that such ACEs may be used as polyepitope vaccines in the immunotherapy of HER-2/neu ⁺ tumors.

Materials and methods

Patients

Five HLA-A*0201⁺ patients with ovarian (n=2) and breast (n=3) adenocarcinoma (clinical stage IV, tumor grade III) were included in this study. HER-2/neu expression was determined both on solid tumors by immunohistochemistry, when these patients had undergone surgery in the past, and on single tumor cells by flow cytometry, in the current stage of disease, when patients developed ascites. In the former case, expression of HER-2/neu was performed by estimating the number and intensity of stained tumor cells per section of tumor specimen as previously reported [20] and using DAKO’s 0–3 scoring system. When using flow cytometry, the expression of HER-2/neu was determined by comparing the mean fluorescence intensity (MFI) of the primary tumor cells with the MFI of tumor cell lines expressing HER-2/neu at different levels (i.e., HER-2/neu expression of the MDA-231 cell line is scored as 1, of MCF-7 as 2, and of SKBR-3 as 3) [21]. In all five patients examined, HER-2/neu expression on primary tumor cells in the peritoneal effusions was scored as 2 (Ova 1, Br 2, Br 3) or 3 (Ova 2, Br 1) which correlated well with the immunohistohemistry analyses. HLA-A2.1 expression on primary tumor cells was determined by flow cytometry on a FACSCalibur (Becton Dickinson, Mountain View, CA, USA) using the anti-HLA-A2.1 BB7.2 MoAb (provided by Prof. H-G. Rammensee, Department of Immunology, University of Tubingen) and an antimouse IgG antibody conjugated with FITC (BD Biosciences Pharmingen, San Diego, CA, USA). All patients fulfilled the following criteria: Karnofsky performance status >80%; bilirubin and creatinine levels <1.7 ng/dl and <2.2 ng/dl, respectively; leukocyte count >3,000/μl; and platelet count >100,000/μl. Patients had not received any antineoplastic therapy during the 3 weeks preceding the onset of the study. All patients were apprised of the study and consent was obtained, consistent with the policies of St Savas Cancer Hospital.

Animals

HHD mice are β_2m ^−/−, D^b−/− and express an HLA-A*0201 monochain composed of a chimeric heavy chain (α1 and α2 domains of HLA-A*0201 and the α3 and intracellular domains of D^b) linked by its N-terminus to the C-terminus of the human β_2m [22]. These mice were generated by crossing HHD transgenic C57BL/6×SJL mice with H-2b^−/−β2m^−/−C57BL/6 double knockout mice. HHD mice were kindly provided by Prof. Francois Lemonnier at the Unite d’Immunite Cellulaire Antivirale, Institut Pasteur, Paris, France.

Synthetic peptides

Peptides were synthesized by standard solid-phase chemistry on a multiple peptide synthesizer and analyzed by mass spectrometry (MS). All peptides were >90% pure, as indicated by analytical HPLC. Lyophilized peptides were stored at −20°C.

The following HER-2/neu–derived peptides were synthesized: HER-2/neu (9₆₆₅), HER-2 (9₆₈₉), HER-2 (9₃₆₉), HER-2 (9₄₃₅), HER-2 (10₉₅₂), HER-2 (9₈₅₁), and HER-2 (9₄₀₂). These are high-binding affinity peptides for HLA-A2.1, eliciting strong CTL activity in vitro [23, 24]. The gp100-derived peptide gp (9₁₅₄) and the Melan-A/MART-1–derived peptide Melan (9₂₇) were used as controls. The last peptides have been demonstrated to elicit in vitro HLA-A2.1 CTL activity [25].

Preparation of primary tumor cells

Tumor cells were prepared as described previously [26]. Specimens of peritoneal effusions (1–2) collected from the patients during routine aspirations were centrifuged at 400 g for 5 min to sediment cells, which were placed on top of a 75% Ficoll-Hypaque gradient, overlaid on 100% Ficoll-Hypaque, and spun at 700 g for 25 min. Tumor cells were collected from the top of the 75% Ficoll-Hypaque.

Preparation of ACEs

An estimated 1–2×10⁸ tumor cells were washed in HBSS (Life Technologies, Gaithersburg, MD, USA), followed by homogenization in 1-ml citrate-phosphate buffer (0.131-M citric acid, 0.066-M Na₂HPO₄, 0.15-M NaCl pH 3.0). The homogenization product was titrated with 10% trifluoroacetic acid and clarified by two successive centrifugations at 2,500 g and 80,000 g for 30 min and 5 h, respectively [27]. Peptides were processed immediately on a Sep-Pak C18 cartridge (Waters, Bedford, MA, USA) equilibrated prior to use with 3-ml acetonitrile, followed by 3-ml deionized water [18]. The eluate was allowed to flow through the cartridge by gravity, the column was washed with deionized water, and bound material was finally eluted with 2 ml of 60% acetonitrile in deionized water and lyophilized in a Speed-Vac (Heto Lab Equipment, Allerod, Denmark). Dry product was reconstituted in approximately 1-ml HBSS and further processed on a Centricon centrifuge concentrator (Amicon, Beverly, MA, USA) with a cutoff of 10 kDa by centrifugation at 2,500 g at 4 °C for 2–3 h. The filtrate was aliquoted and stored at −20°C.

Generation of CTLs in HHD mice

The 5–7 HHD mice were injected subcutaneously (s.c.) at the basis of the tail with ACEs extracted from 3–5×10⁷ primary HLA-A*0201⁺ HER-2/neu ⁺ tumor cells in 250-μl HBSS. Injections were performed in 250 μl of incomplete Freund’s adjuvant (IFA) (500 μl total injected volume). Generation of HER-2/neu peptide–specific CTLs in HHD mice was performed upon s.c. injections with 100 μg of the synthetic peptide in 250-μl HBSS at the basis of the tail in 250-μl IFA [28]. Spleen cells were harvested on day 10 and restimulated in vitro as described below.

T₂ binding assays

ACE preparations were tested for binding to T₂ cells (provided by Prof. H.-G. Rammensee) in the HLA-A2.1 stabilization assay [1]. T₂ cells were grown for 2–3 days in RPMI 1640 medium supplemented with 10% fetal calf serum, 2 mM L-glutamine, and 50 μg/ml gentamycin (all purchased from Life Technologies, Gaithersburg, MD, USA) before use in the experiments. ACE preparations were loaded overnight onto T₂ cells (1×10⁶ cells/ml in serum-free RPMI 1640 medium) at a dose equivalent to 5×10⁶ primary tumor cells. The HLA-A*0201 strongly binding HER-2/neu peptides used herein served as positive controls whereas the HLA-DR4–binding peptide HER-2 (16₈₈₄) [29] served as negative control. All peptides were tested at 20 μg/ml final concentration. Pulsing of T₂ cells with these peptides was performed after overnight incubation. Cells were then washed and analyzed for HLA-A2.1 expression as above. Results are given as mean fluorescence intensity (MFI).

Pulsing with ACEs or HER-2/neu–derived peptides

The amount of ACEs (prepared from HLA-A*0201⁺, HER-2/neu ⁺ primary ovarian or breast tumor cells) or HER-2/neu synthetic peptides for pulsing HHD splenocytes was estimated based on its capacity to induce stabilization of HLA-A2.1 expression on T₂ cells [23]. As shown in Table 1, ACEs extracted from 5×10⁶ HER-2/neu ⁺ HLA-A2.1⁺ primary tumor cells or 20 μg/ml of each HER-2/neu peptide induced high levels of HLA-A2.1 expression on 1×10⁶ T₂ cells. Therefore we kept these concentrations for in vitro pulsing of HHD splenocytes or loading of T₂ targets. Incubations with ACEs or peptides were performed overnight in CO₂ incubators.

Table 1.

Increase of HLA-A*0201 molecule expression on T₂ cells upon preincubation with ACE from HLA-A*0201⁺ primary tumor cells

Preincubation of T2 cells with	MFI
Nonea	704
ACE Br1	2,090
ACE Br2	1,720
ACE Br3	2,190
ACE Ova 1	1,670
ACE Ova 2	2,840
HER-2 (9369)	2,805
HER-2 (9435)	2,380
HER-2 (9665)	2,955
HER-2 (9689)	2,560
HER-2 (9402)	1,860
HER-2 (9851)	1,750
HER-2 (10952)	1,890
HER-2 (16884)	715

^aT₂ cell mice preincubated overnight either in plain medium or with ACE preparations from HER-2/neu ⁺ tumor cells collected from malignant effusions of HLA-A*0201⁺ patients with breast (n=3) or ovarian (n=2) cancer (each ACE is equivalent to 5×10⁶ tumor cells). Peptide HER-2 (16₈₈₄) was used as negative control. All other HER-2 peptides bind to HLA-A*0201 molecules and therefore represent positive controls. T₂ cells were loaded using 20 μg/ml of each peptide

Induction of ACE- or peptide-specific CTL lines in vitro

Splenocytes (2×10⁶ cells/ml) from immunized HHD transgenic mice were restimulated in vitro with syngeneic irradiated (3,000 rads) splenocytes (6×10⁶ cells/ml) pulsed either with ACE or with the synthetic HER-2/neu peptide. Cultures were seeded in 24-well plates (Costar, Cambridge, MA, USA) in a total of 2 ml X-Vivo 15 medium (BioWhittaker Europe, Belgium) supplemented with 1% syngeneic heat-inactivated serum (complete medium). Five days later, half of the medium was replenished with fresh medium containing 100 U/ml rIL-2 and 20 ng/ml murine rIL-15 (R&D Systems, Abington, UK). On day 10, recovered responder splenocytes were washed and restimulated with fresh irradiated syngeneic splenocytes (at a cell ratio of 1:3) pulsed with ACE or HER-2/neu peptide. Fresh rIL-2 and rIL-15 (at 50 U/ml and 10 ng/ml, respectively, final concentration) were added to the medium. Depending on the quality of culture (i.e., when nonviable cells exceeded 40%) viable responders were isolated over Ficoll-Hypaque. Five days later (i.e., day 15), the cytotoxicity of splenocytes was measured by a standard ⁵¹Cr-release assay.

CTL expansion procedures

HHD CTLs were expanded in tissue cultures following a method described by Riddell et al. [30] with some modifications. A total of 1×10⁵ CTLs were resuspended in 25 ml of complete medium with 25×10⁶ ACE- or peptide-pulsed syngeneic irradiated (3,000 rads) splenocytes. One day after inititating the cultures, 100 IU/ml of rIL-2 was added to the cultures. The cultures were fed with fresh medium containing 50 IU/ml rIL-2 every 3–4 days. Usually on day 10–12, cultures were split because T-cell concentration reached numbers >1.0–1.5×10⁶. On average, approximately 4–6×10⁷ CTLs were obtained by day 15–18. In some cases, after day 10–12, cultures were stimulated with IL-2 and 30 ng/ml anti-CD3 MoAb (CD3 ε-chain, clone 145–2C11; BD Pharmingen) to achieve even higher expansion. Before transferring into SCID mice, CTLs were tested for specificity against ACE- or peptide-pulsed T₂ targets (see below).

Tumor cell lines and transfections

The ovarian tumor cell line SKOV-3 was maintained in culture in L-15 medium (Life Technologies, Gaithersburg, MD, USA) supplemented with 10% fetal calf serum (FCS) (Life Technologies) and 20 μg/ml gentamycin. The parental SKOV-3 cell line and SKOV-3 transfected with the HLA-A2 expression vector containing full-length HLA-A2.1 cDNA (SKOV-3.A2) were kindly donated by Dr. C.G. Ioannides (Department of Gynecologic Oncology and Immunology, University of Texas MD, TX, USA, and Anderson Cancer Center, Houston, TX, USA). Transfectants were selected in culture with 800 μg/ml G418 (Sigma, St Louis, MO, USA). The MCF-7 breast cancer cell line was maintained in RPMI 1640 culture medium (Life Technologies) supplemented with 10% FCS (Life Technologies). This cell line was purchased from the American Type Culture Collection (Manassas, VA, USA).

Cytotoxicity assay

The cytotoxicity assay was performed as described previously [26]. Briefly, effector CTLs (1×10⁶/ml) were placed in 100-μl aliquots into 96-well V-bottom plates (Costar). As targets, primary tumor cells or tumor cell lines were labeled with sodium [⁵¹Cr] chromate (Radiochemical Centre, Amersham, UK; 100–200 μCi isotope per 1–2×10⁶ target cells) and added to effectors at the indicated E/T ratios. For peptide recognition, T₂ cells were incubated overnight at 26°C together with 20 μg/ml peptide (or ACE), washed, and then labeled. Incubation was performed for 6 h in CO₂ incubators. In some experiments, blocking of cytotoxicity was performed by preincubating the target cells with 10 μg/ml of the BB7.2 MoAb. Supernatants (100 μl) were collected from each well (all determinations performed in triplicate), and radioactivity was measured in a γ-counter (Packard, Downers Grove, IL, USA). The percentage of cytotoxicity was calculated according to the following formula: lysis = 100 × [(test ⁵¹Cr release) – (spontaneous ⁵¹Cr release)] / [(maximum ⁵¹Cr release) – (spontaneous ⁵¹Cr release)]. Cytotoxicity values were considered to indicate significant recognition of a target when the differences between mean values (from triplicate analyses) for percent lysis of the particular target (e.g., pulsed T₂ cells, primary tumor cells, or transfected tumor cell lines) and unloaded T2 cells were ≥10% at an E/T ratio of 40:1, and statistically significant at p<0.05.

Tumor rejection models

SCID mice were inoculated s.c. with 5×10⁵ tumor cells, each in 0.5-ml PBS. Injections with ACE- or peptide-specific CTLs (2×10⁷ cells in 0.5-ml PBS) were administered intraperitoneally (i.p.) at the time when the tumor was palpable (ca. 12–16 days after inoculation of mice with the tumor cell lines). Tumor size was monitored regularly every 4 days and was expressed as the product of the perpendicular diameters of individual tumors. Each animal experiment was repeated twice. The observation was terminated with the euthanasia of mice when the tumor mass grew up to 200–250 mm² mean diameter. A nonparametric Wilcoxon rank test was used in the statistical analysis of the size of tumor in individual groups. The difference was considered statistically significant when p<0.05.

Results

ACE binding to HLA-A2.1

To establish the ability of our ACE preparations to bind to HLA-A2.1, the T₂ MHC class–peptide stabilization assay was performed as described previously [1]. As shown in Table 1, all ACE preparations induced a twofold to threefold increase in MFI for HLA-A2.1 expression which was almost comparable to that induced by the HER-2 peptides used as positive control. HLA-A2.1 expression with the negative control peptide HER-2 (16₈₈₄) was comparable to that induced in plain medium.

Induction of CTL responses specific for ACE from HER-2/neu⁺ HLA-A*0201⁺ primary tumor cells

The capacity of ACE from HER-2/neu ⁺ HLA-A*0201⁺ primary tumor cells to prime in vivo CTLs was evaluated using HLA-A*0201–transgenic HHD mice. Splenocytes from mice immunized with ACEs derived from each of the five individual cases, in IFA, were stimulated and restimulated once in vitro with the relevant ACE-pulsed syngeneic splenocytes in the presence of IL-2 and IL-15. Such splenocytes were capable of exerting cytotoxic activity against ACE-pulsed, but not unpulsed, T₂ cells (Fig. 1; left panel). The same CTL effectors also lysed intact primary tumor cells from which the ACEs were prepared (Fig. 1; left panel). CTL responses were HLA-A2.1–restricted because they could be highly reduced when an anti-HLA-A2.1 MoAb was added during the cytotoxicity assay. This antibody-mediated inhibition was specific because an isotype-matched anti-TNP MoAb had no effect (data not shown).

Fig. 1.

Cytotoxicity mediated by HHD splenocytes after in vivo stimulation and in vitro restimulation with ACE derived from primary HER-2/neu⁺ HLA-A*0201⁺ tumor cells. HHD mice were immunized with ACEs in IFA and 10 days later their splenocytes were restimulated in vitro with syngeneic irradiated splenocytes pulsed with ACEs as described in “Materials and methods.” Cytotoxicity was tested against T₂ cells pulsed with ACEs on unpulsed as well as against the primary tumor cells (Pri. Tumor) from which the ACE was extracted (i.e., primary tumor cells from patients with ovarian cancer [Ova 1 and Ova 2] or breast cancer [Br1, Br2 and Br3]) in the absence or presence of anti-HLA-A2.1 MoAb. The same effectors were also tested against the HER-2/neu ⁺ HLA-A*0201⁺ ovarian and breast cancer cell lines SKOV3.A2 and MCF-7, respectively, as well as against the HER-2/neu ⁺ HLA-A2⁻ SKOV3 cell line. One experiment of three performed is shown. Mean values ± SD, from triplicate cultures, are shown

ACE-induced HHD cytotoxic effectors recognize HER-2/neu synthetic peptides

We then asked whether the ACE-induced HLA-A2.1 restricted cytotoxic HHD splenocytes could recognize HER-2/neu peptides. To address this, we performed cytotoxicity experiments with HLA-A2⁺ HER-2/neu ⁺ tumor targets. Thus, as shown in Fig. 1 (middle and right panels), ACE-induced HHD CTLs could lyse the HLA-A2.1–transfected HER-2/neu ⁺ ovarian tumor cell line SKOV3.A2 (but not the parental SKOV3 targets which do not express HLA-A2.1) as well as the HER-2/neu ⁺ HLA-A*0201⁺ breast cancer cell line MCF-7. The cytotoxic response against this latter tumor target was largely abrogated in the presence of anti-HLA-A2.1 MoAb (Fig. 1; right panel). Control splenocytes from nonimmunized HHD mice, stimulated and restimulated as above, lacked specificity against these HER-2/neu ⁺ HLA-A2*0201⁺ targets exhibiting only weak nonspecific cytotoxicity (Fig. 2).

Fig. 2.

Cytotoxicity mediated by splenocytes from untreated control HHD mice. These HHD splenocytes were stimulated only in vitro with syngeneic irradiated splenocytes pulsed with ACEs as described in “Materials and methods” for splenocytes from treated HHD mice. See also legend to Fig. 1

To further analyze the ACE-induced cytotoxic responses, we examined the cytotoxicity profile of the HHD bulk splenocytes against T₂ cells pulsed with known synthetic HER-2/neu CTL epitopes and control peptides (i.e., HLA-A2.1–restricted CTL epitopes from proteins apparently not expressed in breast and ovarian cancer). When T₂ cells were loaded with the CTL epitope HER-2 (9₃₆₉) these were lysed by each of the five ACE-induced HHD effector bulk cultures (Fig. 3). HER-2 (9₄₃₅) was also recognized by four of five bulk cultures, whereas peptides HER-2 (9₆₆₅) and HER-2 (10₉₅₂) were recognized by two of five cultures. HER-2 (9₄₀₂), HER-2 (9₈₅₁) and HER-2 (9₆₈₉) were not recognized by any of the ACE-induced HHD CTL effectors (Fig. 3). The same effectors recognizing synthetic peptides derived from the HER-2/neu sequence did not lyse T₂ cells pulsed with (a) Melan-A/MART-1 CTL peptide (9₂₇) or gp100 CTL peptide (9₁₅₄) and (b) ACEs from an HLA-A2⁺ HER-2/neu ⁻ EBV-transformed B cell line (Fig. 3).

Fig. 3.

ACE-induced HHD CTLs recognize HER-2/neu peptides. HHD splenocytes stimulated in vivo and restimulated in vitro with ACEs from Br1, Br2, Br3, Ova1, and Ova2 primary tumor cells were examined for cytotoxicity against T₂ cells (E/T ratio 20:1) pulsed with the same ACEs or with various HLA-A*0201–restricted HER-2/neu peptides or control gp100 and Melan-A/MART-1 peptides. ACE-EBV indicates control extracts from an HLA-A2*0201⁺, HER-2/neu ⁻ Epstein-Barr virus–transformed B cell line. One representative experiment of three conducted is shown. See also legend to Fig. 1

Adoptively transferred ACE-induced HHD CTLs protect SCID mice against the growth of human cell lines

To determine whether ACE-induced HHD splenic CTLs were also effective in mediating antitumor responses in vivo, groups of ten mice were injected with 2×10⁷ ACE-specific CTLs or control splenocytes from either nontreated HHD mice or from HHD mice immunized with irrelevant ACEs from an HLA-A2⁺ HER-2/neu ⁻ EBV-B cell line. Prior to this, mice were inoculated with 5×10⁵ cells of the human ovarian cell line SKOV3 either nontransfected (SKOV3: HER-2/neu ⁺, HLA-A2⁻) or transfected to express the HLA-A2 molecule (SKOV3.A2). Adoptive transfer of HHD splenocytes was performed when tumor growth was palpable (i.e., between days 12 and 16). Transfer of HHD CTLs induced by ACEs from patients’ Br1 tumor cells (ACE–CTL Br1) exerted high levels of antitumor activity against the growth of transplanted SKOV3.A2 tumor cells: Solid tumor formation induced upon inoculation of SKOV3.A2 cells reached an area of 205 mm² on day 124 (Fig. 4B) as compared to a similar tumor size reached already on day 40 (Fig. 4B) when mice were treated with control splenocytes (i.e., CTLs from nontreated HHD animals or from HHD mice immunized with ACEs from an HLA-A2⁺, HER-2/neu ⁻ EBV-transformed B cell line) (Fig. 4B). ACE-CTL-Br1 had no effect on the growth of the parental SKOV3 cells which lacked the expression of HLA-A2 allele (Fig. 4A). Similar profiles of in vivo antitumor activity were also observed with HHD splenic CTLs induced by ACEs of tumor cells from patients Ova1, Ova2, Br2, and Br3: tumors induced by the inoculation of SKOV3.A2 ovarian cells had a significant delay in their growth and reached a size of 200–215 mm² between days 96 and 124 (Fig. 4b). The same CTLs remained without any effect when SCID mice were xenografted with parental SKOV3 tumor cells (Fig. 4A).

Fig. 4A, B.

Adoptive transfer of ACE-induced HHD CTLs delays the growth of human ovarian tumor SKOV3.A2 cell line in SCID mice. CTLs (ACE-CTL Br 1, ACE-CTL Br 2, ACE-CTL Br3, ACE-CTL Ova 1, ACE-CTL Ova 2) were induced in HHD mice upon injection with each of the five different ACEs extracted from patients’ Br1, Br2, Br3, Ova1 and Ova2 primary tumor cells and restimulated in vitro with the same ACE. CTLs (2×10⁷) were injected i.p. once in SCID mice with s.c. growing tumors induced upon inoculation, 12–16 days before, with the SKOV3.A2 (HER-2/neu ⁺ HLA-A*0201⁺) or the parental SKOV3 (HER-2/neu ⁺ HLA-A*0201⁻) cell lines. Control CTLs consisted of splenocytes from nontreated HHD mice. Splenocytes from HHD mice immunized with ACEs from an HLA-A0201⁺ HER-2/neu ⁻ EBV-B cell line (and cultured in vitro as the relevant ACE-CTL) were also used as irrelevant CTLs. Such CTLs were capable of lysing the original EBV-B line (data not shown). Tumor growth was monitored in groups of ten mice every 4 days throughout the experiments. Mice were sacrificed by euthanasia when tumor size reached an area bigger than 200  mm². Arrows indicate the day of CTL transfer. Data are reported as the average tumor area ± SD of ten mice per group. One representative experiment of two performed is shown

Induction of CTL responses specific for HER-2/neu peptides in HHD mice

The data so far demonstrated that HHD splenocytes induced in vivo by ACEs from primary HLA-A*0201⁺ HER-2/neu ⁺ tumors and restimulated in vitro with the same antigenic material could recognize in all cases (five of five) peptide HER-2 (9₃₆₉) and, in 80% of cases, peptide HER-2 (9₄₃₅). Moreover, these ACE-induced HHD CTL effectors conferred significant protection of SCID mice against the growth of HER-2/neu ⁺ HLA-A2⁺ human ovarian SKOV3.A2 tumor cells. Thus, it was reasonable to assume that such bulk CTLs contain clones predominantly recognizing HER-2/neu peptides in the context of HLA-A*0201 which should carry most of their in vitro and in vivo antitumor activity. To test this, we assessed the capacity of peptides HER-2 (9₃₆₉) and HER-2 (9₄₃₅) to prime CTL responses in transgenic HHD mice. Mice immunized with each of these peptides in IFA developed CTLs that upon restimulation in vitro lysed T₂ targets pulsed with the same peptides used for vaccination in vivo (Fig. 5). The same peptide-specific CTLs also lysed T₂ cells loaded with ACEs demonstrating the presence of the relevant peptide within such preparations. That ACE-loaded T₂ cells were lysed at significantly lower levels compared to the cytotoxicity levels achieved with peptide-loaded T₂ cells (Fig. 5) can be attributed to the apparently much lower concentration of the peptide within ACE compared to the dose of 20 μg/ml that was used for loading. Most important, the HER-2 peptide-specific HHD CTLs were capable of lysing both SKOV3.A2 (but not SKOV3) and MCF-7 targets demonstrating that the recognized HER-2 CTL epitopes (i.e., [9₃₆₉] and HER-2 [9₄₃₅]) are naturally processed and expressed on the surface of tumor cells in the context of HLA-A2.1 molecules (lysis of MCF-7 targets was almost totally abrogated by the addition of the HLA-A2.1 MoAb) (Fig. 5; lower panel).

Fig. 5.

Cytotoxicity mediated by HHD splenocytes after in vivo stimulation and in vitro restimulation with HER-2/neu–derived synthetic peptides. HHD mice were injected with 100 μg of each peptide in IFA, and 10 days later their splenocytes were restimulated in vitro with syngeneic irradiated splenocytes pulsed with the peptide used for injection as described in “Materials and methods.” Cytotoxicity was tested against various targets as indicated. One experiment of three conducted is shown. See also legend to Fig. 1.

Adoptively transferred HER-2/neu peptide–induced CTLs protect SCID mice from tumor growth

To gain further insights into the antitumor activity of the HER-2 peptide–induced HHD splenic CTLs, these were adoptively transferred in SCID mice, and their potential to confer protection from the growth of SKOV3.A2 and MCF-7 tumor cell lines was compared with that mediated by the ACE-induced CTLs. As expected, there was no therapeutic effect against the SKOV3-induced tumors not expressing HLA-A*0201 (Fig. 6A, B). HER-2 (9₃₆₉)-specific CTLs (HER [9₃₆₉] CTLs) protected SCID mice inoculated with SKOV3.A2 (Fig. 6C) or MCF-7 (Fig. 6E) at almost similar levels to those of HER (9₄₃₅) CTLs (Fig.6C, E) (p<0.05, compared to irrelevant CTLs or CTLs from nontreated mice; Fig. 6C, E). However, each of the ACE-induced CTLs were more effective than HER (9₃₆₉) CTLs or HER (9₄₃₅) CTLs in the immunotherapy of SKOV3.A2 and MCF-7 tumors (p<0.01, for Fig. 6C, E, compared with Fig. 6D, F, respectively). When administered in mixtures, the two peptide-specific CTLs (Fig. 6C, E) induced a more profound delay of tumor growth which, however, was still significantly (p<0.05) less intense compared with that induced by ACE-specific CTLs.

Fig. 6A–F.

Therapeutic capacity of HER-2/neu peptide-specific HHD CTLs in SCID mice xenografted with HER-2/neu ⁺ HLA-A0201⁺ human ovarian and breast tumor cell lines. CTLs (2×10⁷) were i.p. injected per mouse in groups of ten SCID mice which were previously (12–16 days) s.c. inoculated with the indicated tumor cell lines. For nontreated and irrelevant CTLs, see legend to Fig. 4. Injection of CTL mixtures consisted of 1×10⁷ CTLs specific for each of the HER-2/neu peptides. As a comparison, groups of SCID mice were therapeutically treated with ACE-induced HHD CTLs, as shown in Fig. 4. Tumor growth was monitored in groups of ten mice, until tumor area had reached a volume >200 mm². Data are reported as the average tumor area ± SD of ten mice per group. One representative experiment of two performed is shown

Discussion

This study represents the first description of the use of HLA-A*0201–transgenic mice for quantitating the in vivo immunogenicity of ACEs prepared from HER-2/neu ⁺ HLA-A*0201⁺ primary human tumor cells. In addition to validating the transgenic animal system as a convenient and rapid way to assay the suitability of HER-2/neu ⁺ tumor cell–derived ACEs as a source of immunogenic epitopes generating in vivo HER-2 peptide-specific CTLs, we have used SCID/human tumor models to address the in vivo antitumor activity of these CTLs. In the first step of our experimental approach, we succeeded in generating and expanding ACE-specific CTLs, lysing in vitro HER-2/neu ⁺ human tumor cell lines in an HLA-A*0201–restricted manner and suppressing, upon adoptive transfer, in SCID mice the growth of transplanted HER-2/neu ⁺ HLA-A*0201⁺ tumor cells. The ACE-specific HHD CTLs were tested in a second step, for reactivity against synthetic HLA-A*0201–restricted HER-2/neu epitopes. In this respect, we could identify two immunodominant nanomers (i.e., HER-2 [9₃₆₉]and HER-2 [9₄₃₅]) that were recognized by such CTLs. These findings suggested that among the ACE-induced CTLs, there are also CTL clones specific for HER-2/neu epitopes and that these particular epitopes are, in agreement with previous reports [19, 21, 23, 24, 31], naturally expressed by tumor cell lines. Finally, in a third step, we injected HHD mice with either of the HER-2/neu CTL peptides and in this way could directly demonstrate their capacity to generate in vivo specific CTLs which, upon restimulation in vitro, recognized both SKOV3.A2 and MCF-7 cell lines and also provided SCID mice with protection against the growth of the same tumor lines.

The HER-2/neu peptides recognized by ACE-induced CTLs have been demonstrated in previous and recent reports to bind with high affinities to HLA-A2.1 molecules and to elicit CTLs from tumor-associated lymphocytes of patients with breast and ovarian cancer [19, 23, 24]. In our most recent reports [21, 31], we have shown that the same peptides are also recognized by CTLs from patients with colorectal, lung, and prostate cancer. Moreover, we [19, 21, 31] and others [23, 32] have demonstrated that CTLs specific to these peptides specifically lysed HER-2/neu ⁺ HLA-A*0201⁺ tumor cell lines and autologous tumor cells. By using a protocol similar to ours, Lustgarten et al. [33] generated HER-2 (9₃₆₉)–specific CTLs in double transgenic mice expressing HLA-A2.1 and human CD8 which could recognize HER-2/neu ⁺ tumor cell lines in vitro expressing HLA-A2.1. Thus, although our ACE-specific HHD CTLs, at least those described herein, recognize already known HER-2/neu CTL peptides, it is still fair enough to assume that by screening additional peptides from the HER-2/neu sequence with HLA-A2.1–binding motifs it may be possible to also detect novel CTL epitopes. Therefore, we postulate that ACEs from primary tumors may act in a way similar to polyepitope constructs which contain multiple individually immunogenic epitopes.

A variety of delivery modalities have been used so far for human polyepitope vaccines, including attenuated poxvirus vectors [34], adenovirus [35], naked DNA [36], or transfected dendritic cells [37]. ACEs derived from primary tumor cells may have an advantage over such polyepitope cancer vaccines, not solely due to the fact that these apparently contain a plethora of CTL epitopes but also due to the simplicity of both the method used for their preparation and administration. Another potentially important question for polyepitope cancer vaccines is the source of CD4 T cell help. Reports so far [38, 39] are not conclusive as to whether the immunogenicity of CTL polypeptides constructs is dependent on the presence of T_Hepitope(s). We have recently reported [18] that ACEs prepared from primary tumor cells of patients with metastatic lung, breast, and ovarian cancer are potent stimulators of MHC class II–restricted and autologous tumor–specific CD4⁺ T cells which potentiate the cytotoxic activity of autologous MHC class I–restricted ACE-specific CD8⁺ T cells against the primary autologous tumor cells. These data indicated that, most probably, our ACE preparations also contain T_H epitopes. Injection of ACE in HLA-DR transgenic mice should be an appropriate approach for detecting such HER-2/neu T_H epitopes capable of stimulating CD4⁺ T cell responses.

For the first time, our data also demonstrate therapy of human HER-2/neu ⁺ tumors in preclinical models through passive immunization based on the adoptive transfer of ACE-specific or human HER-2/neu peptide–specific CTLs. Only very recently, Ercolini et al. [40] reported that CTLs specific for a rat HER-2/neu epitope HER-2 (10₄₂₀) were able to cure FVB/N mice of transplanted rat neu-expressing murine tumor cells. We have shown that immune splenocytes induced in HHD mice with five different ACE preparations (each one derived from HER-2/neu ⁺ HLA-A2⁺ primary tumor cells) when adoptively transferred, significantly prolonged the survival of SCID mice inoculated with human tumor cell lines expressing both HER-2/neu and HLA-A2 molecules. Our finding that such effector splenocytes, at least as far as detected herein, could recognize peptides HER-2 (9₃₆₉) (five of five) and HER-2 (9₄₃₅) (four of five) indicate that these particular CTL epitopes belong to the immunodominant MHC class I epitopes within such preparations. Of course, the possibility exists that additional HER-2/neu–derived peptides might also be recognized by the ACE-specific splenocytes. Alternatively, ACE-induced CTLs may exhibit HLA-A*0201–restricted specificity for peptides derived from other than HER-2/neu proteins. Such possibilities are supported by the fact that adoptive transfer of mixtures of CTLs specific for HER-2 (9₃₆₉) and HER-2 (9₄₃₅) induced less protection as compared to that achieved by ACE-specific CTLs. By testing cytotoxicity levels against T₂ cells pulsed with various HER-2/neu–derived peptides having HLA-A2–binding motifs, we may be able to characterize additional HER-2/neu epitopes recognized by the ACE-specific CTLs. Protein profile analysis of the primary tumor cells used for ACE preparation and transplantable human cell lines may lead to the identification of common proteins capable of generating, in an HLA-A*0201–restricted fashion, CTLs with antitumor properties. Immunogenic peptides from such proteins will then be useful in vaccination studies for cancer immunotherapy.