Following the identification of tumor-associated antigens recognized by T cells on human tumor cells, numerous clinical vaccination studies have been performed in patients with metastatic melanoma. Accompanied by an improvement of vaccine approaches including new adjuvants such as CpG oligonucleotides and further development in the field of immuno-monitoring, the induction of antigen-specific T cell responses in the blood of tumor patients has been demonstrated in a number of studies. However, significant tumor regressions, especially complete responses in patients with visceral metastases have been observed infrequently, and in most studies overall response rates did not exceed 10%.1 One obstacle might be immune escape by downregulation or complete loss of antigen expression as many antigens identified so far do not play an essential functional role for the tumor.2–4 To overcome this it would be particularly important to choose a suitable target antigen for vaccination therapy, which should be broadly expressed and functionally relevant for the tumor. On our search for suitable target antigens for the active-specific immunotherapy of melanoma, we focused on the human melanoma-associated chondroitin sulfate proteoglycan (MCSP), also known as high molecular weight-melanoma-associated antigen, which has been shown to be expressed in the majority of human melanoma lesions and cultured cells with a limited expression in normal tissues.5, 6 In addition, MCSP expression was found in uveal melanoma, which is refractory to standard chemotherapy.7
MCSP represents an unique glycoprotein–proteoglycan complex, with a 250 kDa core glycoprotein to which, via serine residues, the larger than 450 kDa proteoglycan component is attached.8 MCSP has been implicated for some time in numerous aspects of melanoma cell biology, including adhesion, spreading and migration. In fact, melanoma cell adhesion, chemotactic responses to fibronectin and cytoplasmatic spreading on extracellular matrix proteins have been shown to be inhibited by MCSP-specific antibodies in vitro.9–11 Furthermore, MCSP forms a complex with membrane-Type 3 matrix metalloproteinase (MT3-MMP) resulting in increased proteolysis of the extracellular matrix and enhanced tumor cell invasion.12
NG2, the rat homologue of MCSP, is expressed by nascent pericytes during early stages of angiogenesis and plays an important role in neovascularization.13 In line with these findings, NG2 knock-out mice show a significantly reduced proliferation of pericytes and endothelial cells within the retina and cornea.13 Glioblastoma multiforme tumors transfected with NG2, showed a significantly increased neovascularization rate with a higher vascular density than control tumors in a rat model. This accelerated angiogenesis was accompanied by a dramatic increase in tumor growth.14 Similar results have also been described for prostate cancer and uveal melanoma underlining the significance of NG2/MCSP as an anti-tumor target antigen.15, 16
Taken together, targeting MCSP should lead to reduced migration, invasion and metastasis of melanoma cells and prevent tumor outgrowth by inhibition of angiogenesis.
In the present study, we identified a HLA-DRB1*1101-restricted T cell epitope within the extracellular domain of the MCSP core protein (aa 1285–1296). ELISPOT analysis revealed T cell reactivity against the identified peptide in both, healthy donors and melanoma patients although this reactivity was weaker as it has been demonstrated for a previously identified MCSP peptide. Importantly, peptide-specific T cells could be amplified in vitro and directly recognized HLA-DR and MCSP expressing melanoma cells, indicating that the antigenic peptide is naturally processed by tumor cells.
Material and methods
Experiments were performed with blood cells from human healthy donors after having obtained informed consent and approval by the ethics committees of the Medical Faculty of the University of Erlangen–Nuremberg and the Medical Faculty of the University of Marburg.
Cell lines, reagents and antibodies
EBV-transformed B (EBV-B) cell lines and tumor cell lines were cultured in RPMI 1640 medium (Cambrex, Verviers, Belgium) supplemented with 10% FCS (PAA Laboratories, Coelbe, Germany), 20 μg/ml gentamicin (PAA Laboratories), 2 mM L-glutamine (PAA Laboratories), 10 mM Hepes (PAA Laboratories) and 10 mM sodium pyruvate (PAA Laboratories). DCs and CD4+ T cells were cultured in the same medium supplemented with 1% autologous plasma or 10% human serum, hereafter referred to as DC or T cell medium, respectively. Human IL-2 was purchased from Roche Diagnostics GmbH (Mannheim, Germany), IL-4, IL-1β, IL-6 and GM-CSF from Strathmann Biotech (Hamburg, Germany), IL-7 from Biomol GmbH (Hamburg, Germany), TNF-α from Bender (Vienna, Austria) and PGE2 from Pfizer (Puurs, Belgium). The following mouse anti-human mAbs were used for blocking experiments: anti-HLA-DQ (SPVL3) from Immunotech (Marseille, France), anti-HLA-DP (B7/21) and anti-HLA-DR (L234) from BD Biosciences (Heidelberg, Germany). For flow cytometric analysis of melanoma cells, antibodies against HLA-DR (1E5) from Immunotools (Friesoyte, Germany) and against MCSP (LHM2) from Abcam (Cambridge, UK) were used.
Flow cytometric analysis
For immunofluorescence, staining tumor cells were washed and stained for 20 min at 4°C with optimal dilution of each antibody. Cells were washed again and analyzed by flow cytometry (FACSScan™ and CELLQuest™ software, Becton Dickinson, Heidelberg, Germany).
Peptide PPADIVFSVKSPPSAGYLVMVSRGALADEPP = MCSP1270–1300 was selected by using a database allowing the prediction of T cell epitopes for a given HLA Class I or II molecule (17; SYFPEITHI-Database: www.uni-tuebingen.de/uni/kxi/). The chosen peptide of 31 amino acids length (aa 1270–1300) covers many of the predicted HLA-DR binding motifs within the sequence of the MCSP core protein. To potentially achieve an additional HLA Class I presentation of putative CD8+ T cell epitopes enclosed within this long sequence a modified peptide derived from the protein transduction domain of HIV TAT protein was added at the NH2 terminus. Peptide YARAAARQARA had previously been shown to efficiently enter the cytosol of human Jurkat T cells.18 Finally, the resulting peptide was YARAAARQARAPPADIVFSVKSPPSAGYLVMVSRGALADEPP (42 aa) and was synthesized by conventional solid-phase peptide synthesis using F-moc for transient NH2-terminal protection and were characterized using mass spectrometry.
DCs and CD4+ responder cells
DCs were generated from leukapheresis products from donor 11325 as described previously.19 Briefly, PBMC were obtained by Ficoll density gradient centrifugation and monocytes were isolated by plastic adherence and cultured in the presence of 800 IU/ml GM-CSF and 250 IU/ml IL-4 for 6 days to generate immature, monocyte-derived DCs. Maturation was induced by adding a cytokine mixture consisting of 10 ng/ml IL-1β, 1,000 U/ml IL-6, 10 ng/ml TNF-α and 1 μg/ml PGE2.20 Mature DCs were harvested on day 7. CD4+ T lymphocytes were isolated from PBMC by negative selection using anti-CD8+, anti-CD14+, and anti-CD19+ mAbs coupled to magnetic microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany).
Generation of peptide-specific CD4+ T cell clones
Autologous immature DCs were loaded overnight with 20 μg/ml of the modified MCSP peptide and matured after 6 hr by adding a cytokine mixture, as mentioned above. CD4+ T cells (1 × 105) were co-cultured with 1 × 104 peptide-loaded DCs in 100 μl T cell medium/round-bottomed microwell, supplemented with IL-6 (1,000 U/ml), IL-12 (20 ng/ml) and TNF-α (5 ng/ml). The CD4+ T lymphocytes were restimulated on days 7, 14 and 21 with autologous DCs freshly loaded with peptide overnight (20 μg/ml) and matured by the cytokine cocktail, adding IL-2 (10 U/ml) and IL-7 (5 ng/ml). Aliquots of each microculture (∼4,000 cells) were assessed on day 30 for their capacity to produce IFN-γ when stimulated with ∼15.000 autologous EBV-B cells, which were loaded overnight with 10 μg/ml of peptide or with a control peptide derived from the MAGE-A1 protein. After 20 hr of co-culture in round-bottom microwells in T cell medium supplemented with IL-2 (25 U/ml), IFN-γ released in the supernatant was measured by ELISA using reagents from Medgenix Diagnostics-Biosource (Fleurus, Belgium). Microcultures, which were tested positive, were then cloned by limiting dilution, using irradiated, autologous, peptide-loaded EBV-B cells (104 cells/round-bottomed microwell) as stimulator cells and irradiated allogeneic LG2-EBV-B cells (104 cells/well) as feeder cells in the presence of IL-2 (50 U/ml), IL-4 (5 U/ml), IL-7 (5 ng/ml) and PHA (125 ng/ml) (Sigma-Aldrich, Munich, Germany).
Recognition assays with peptides
CD4+ T cells (4 × 103/ microwell) were co-cultured with 15 × 103 peptide-loaded EBV-B cells from donors with different HLA Class II typing. Supernatants were harvested after 20 hr and IFN-γ production was measured by ELISA. To screen a set of truncated peptides, autologous EBV-B cells were incubated for 1 hr in the presence of different peptides. To identify the HLA restriction of the T cell clone, blocking of the Ag-induced production of INF-γ was investigated using mAbs against HLA-DR (L234), HLA-DQ (SPVL3) and HLA-DP (B7/21). All mAbs were used at a final concentration of 5 μg/ml.
Measurement of MCSP T cell reactivity by ELISPOT analysis
PBMC (2.5 × 106 per 24 well) from healthy blood donors and melanoma patients (after having obtained informed consent) were stimulated once with peptide GYLVMVSRGALA (MCSP1285–1296) (10 μg/ml) in the presence of IL-2 (5 U/ml) and IL-7 (10 ng/ml). The ELISPOT assay was performed on day 7 using triplicates at 3 × 105/flat bottomed 96 well in medium containing 10% heat inactivated human serum and stimulated with 10 μg/ml of peptide. After 20 hr, wells were washed and incubated with biotinylated mAb to IFN-γ (7-B6-1, Mabtech, Hamburg, Germany) for 2 hr. After washing, 100 μl of HRP-conjugated avidin (Vector Laboratories, Burlingham) were added for 45 min at room temperature. The plates were again washed and the spots were developed using 3-amino-9-ethylcarbazole (Sigma-Aldrich, Munich, Germany), diluted 1 ml into 35 ml of 0.1 mol/l sodium acetate buffer, filtered and mixed with 35 μl of H2O2. Spots were counted with an Automated Elisa-Spot Assay Video Analysis System (A-EL-VIS, Hannover, Germany). Background without antigen was subtracted and responses were considered significant if a minimum of 10 spot forming cells per well were detected, and additionally, this number was at least twice that in negative control wells.
Recognition of tumor cells
PBMC from patient K30 who remained free of tumor after surgery for macroscopic regional lymph node metastases were stimulated weekly with 10 μg/ml of peptide GYLVMVSRGALA in the presence of IL-2 (10 U/ml) and IL-7 (10 ng/ml). On day 21, IFN-γ production upon stimulation with different melanoma cell lines (10.000 per 96 microwell) was analyzed by ELISPOT. As a positive control PBMC were stimulated with 10 μg/ml phytohaemagglutinin (Sigma-Aldrich, Munich, Germany).
To isolate MCSP-reactive CD4+ T cells, CD4+ T cells of a healthy donor were stimulated with autologous DCs loaded with a 31 amino acid long MCSP peptide (aa 1270–1300) to which a modified peptide derived from the protein transduction domain of HIV TAT protein was added at the NH2 terminus. Peptide YARAAARQARA had previously been shown to efficiently cross cell membranes and enter the cytosol of human Jurkat cells.18 This approach should therefore allow the processing and presentation of both, HLA Class I and II-restricted epitopes potentially incorporated in the chosen MCSP fragment.
Generation of MCSP-specific CD4+ T cell clones from the blood of a healthy donor
A total of 96 microcultures were set up, each containing CD4+ T cells from donor 11325 and autologous stimulator DCs loaded with the MCSP peptide stimulator cells. Responder cells were restimulated 3 times with DCs loaded with peptide and tested 10 days after the last restimulation for IFN-γ production after contact with autologous EBV-B cells loaded either with the MCSP peptide or with a control peptide. Five out of 96 microcultures showed a significant peptide-specific T cell reactivity and 2 of them were subsequently cloned by limiting dilution. Stably growing peptide-specific T cell clones (n = 6) could only be established from microculture F1 and further experiments were performed with Clone 3 generated from this microculture. Clone 3 specifically recognized autologous EBV-B cells loaded with the MCSP peptide (Fig. 1).
Identification of the 16-mer antigenic peptide
To determine the core epitope recognized by Clone 3 within the 42-mer, a set of 16-mer peptides, overlapping each other by 12 amino acids and spanning the whole MCSP fragment, was tested for recognition. In addition, the modified HIV TAT peptide was tested for recognition by the T cells. Clone 3 recognized 2 overlapping 16-mer peptides PPSAGYLVMVSRGALA (MCSP1281–1296) and GYLVMVSRGALADEPP (MCSP1285–1300), but not the modified TAT peptide YARAAARQARA (Fig. 2). In a further experiment, a set of truncated peptides derived from the sequence of the 16-mer peptide PPSAGYLVMVSRGALA recognized best by Clone 3 was tested to define the fine specificity of Clone 3. The 12-mer GYLVMVSRGALA (MCSP1285–1296) was the shortest peptide efficiently recognized by the T cell clone (data not shown).
Determination of the HLA-restriction of the CD4+ T cell clone
To analyze the HLA-restriction of Clone 3, we tested whether monoclonal anti-DR, anti-DQ, or anti-DP antibodies would inhibit the recognition of antigen-presenting cells by the CD4+ T cells. Peptide recognition was significantly impaired in the presence of anti-DR antibodies, but not in the presence of anti-DQ or anti-DP antibodies (Fig. 3). Donor 11325 was typed HLA-DRB1*1101/1301. To further determine the peptide presenting HLA-DR allele several EBV-B cell lines with known HLA-DR typing were tested for their ability to present the antigenic peptide. All of the EBV-B cell lines expressing HLA-DRB1*1101 were able to present the antigenic peptide to Clone 3 (Fig. 4). HLA-DRB1*1101 is expressed by ∼17% of Caucasians.
T cell reactivity of healthy donors and tumor patients
To study the immunogenicity of the identified MCSP, epitope ELISPOT analysis was performed with PBMC from the blood of healthy donors and patients with Stage III/IV melanoma. After 1 in vitro stimulation, 2 out of 9 healthy donors (Fig. 5a) and 3 out of 16 patients (Fig. 5b) showed significant IFN-γ secretion upon stimulation with peptide MCSP1285–1296. Patient 30K who showed MCSP T cell reactivity in her blood had a history of macroscopic lymph node metastasis from an ulcerated, nodular melanoma on the lower leg. Following complete lymph node dissection of the groin in 2004, the patient remained tumor free until today. To further investigate the significance of the MCSP, T cell reactivity PBMC from patient 30K were restimulated weekly to enrich the MCSP specific T cells. After 2 restimulations, recognition of melanoma cell lines was tested by ELISPOT analysis. As shown in Figure 6, 3 out of 5 tested cell lines were strongly recognized by the T cells from patient 30K. Flow cytometric analysis revealed that all melanoma cell lines strongly expressed MCSP on their cell surface whereas HLA-DR molecules were only expressed by 3 cell lines, but not by ER-MEL-4 and ER-MEL-6 suggesting that these cell lines could not be recognized because of the lack of HLA expression (Fig. 7). It should be mentioned that MZ-MEL.43 and SK-MEL 28 were clearly recognized by the T cells although they do not express HLA-DR11 molecules indicating that the antigenic peptide can be presented by more than 1 HLA-DR molecule as it has already been shown for other T cell epitopes.21
MCSP is strongly expressed in a high percentage of melanoma lesions with low intralesional and interlesional heterogeneity and appears to be functionally relevant for adhesion, migration, invasion and proliferation of melanoma cells.6 These properties and the availability of MCSP-specific mAb have provided the rationale to use MCSP as a target of passive antibody-based as well as active-specific immunotherapy. Several in vitro as well as animal studies have demonstrated that targeting MCSP with antibodies alone or conjugated to toxins effectively inhibits tumor growth.22–24 On the basis of these findings, MCSP-specific mAbs alone or in conjugation with toxins have been used to treat patients with advanced melanoma.25–27 Finally, immunizing Stage IV melanoma patients with an anti-idiotypic mAb mimicking the determinant recognized by the anti-MCSP mAb 763.74 resulted in prolonged survival.28
On the basis of these studies, we hypothesized whether MCSP might be a target for cellular immune responses. In the present study, we could indeed identify a CD4+ T helper cell epitope located in the extracellular region of the MCSP core protein. The MCSP specific CD4+ T cells directly recognized HLA-matched MCSP expressing melanoma cells demonstrating that the peptide seems to be naturally processed by the tumor cells. ELISPOT analysis revealed T cell reactivity against the identified peptide in both, healthy donors and melanoma patients but this reactivity was much weaker as it has been demonstrated for a previously identified MCSP peptide.29 This finding corroborates the existence of a diverse MCSP-specific T cell reactivity directed against different CD4+ T cell epitopes with varying immunogenicity.
MCSP has been shown to be also expressed in a number of normal tissues including basal cells of the epidermis, cells within hair follicles, chondrocytes, smooth muscle cells, pericytes and others.6 These findings obviously raise concerns about the induction of autoimmunity when immunizing with this antigen.
However, it should be mentioned that targeting MCSP in melanoma patients with unlabeled or radiolabeled anti-MCSP antibody 9.2.27 has not been associated with significant normal-organ-related accumulation or toxicity.30–32 Furthermore, we could detect significant MCSP T cell reactivity in the blood of both, healthy donors and melanoma patients without any clinical signs of autoimmunity.
In this context, it should be mentioned that most somatic cells do not express HLA Class II molecules under normal conditions and thus cannot be recognized by CD4+ T cells. In addition, we know from quantitative real-time PCR data that MCSP is 10–1,000-fold overexpressed in melanoma cells when compared with normal tissues,29 a situation which we know from several other tumor antigens which are successfully targeted in antibody-based cancer therapy such as Her2, CD20 or VEGF.33–35 Meanwhile, we have started to systematically study the protein expression in tumors and normal tissues by immunohistochemistry and our preliminary data confirm that MCSP seems to be significantly overexpressed in melanoma lesions when compared with normal tissues (data not shown). Therefore, we feel that it should be possible to induce or amplify MCSP specific CD4+ T cell responses either alone or in combination with antibody responses to attack melanoma cells in vivo without major collateral damages.
The immunogenicity of the newly identified peptide seems to be lower when compared with the previously identified MCSP693–708 peptide29 because spontaneously occurring immune responses occur less frequently. Thus, we could only detect T cell reactivity against peptide MCSP1281–1296 in 2 out of 9 healthy donors and 3 out of 16 melanoma patients, whereas in our previous study there was a significant T cell reactivity against peptide MCSP693–708 in 12 out of 14 donors and 11 out of 42 patients.
Nevertheless, we could demonstrate that PBMC from a melanoma patient, which were prestimulated with peptide, strongly recognized several MCSP and HLA-DR expressing melanoma cell lines. These data suggest that vaccination against MCSP CD4+ T cell epitopes might lead to the induction of tumor-reactive CD4+ T cells which only can recognize target cells expressing both the antigen and HLA Class II molecules. Therefore, the risk of autoimmunity induction seems to be minimal as most somatic cells do not express HLA Class II molecules.