Methods developed over the last several years have allowed tumor antigens that are recognized by MHC class I- and class II-restricted T cells to be readily identified. The screening of patient sera against tumor cell cDNA expression libraries has also resulted in the identification of a large number of antigens, some of which were also found to be recognized by tumor-reactive T cells. These antigens, as well as proteins that appear to be overexpressed in tumors, represent targets that can potentially be used for the development of cancer vaccines.
The availability of a wide variety of targets has led to a number of clinical vaccine trials. Melanoma patients have been immunized with peptides that appear to represent dominant T cell epitopes from a number of melanoma antigens, including MAGE-3, MART-1, and gp100. Additional targets now being evaluated in clinical trials include peptides derived from proteins that are overexpressed in particular tumors, such as HER-2/neu, as well as peptides from oncogenes such as ras that are frequently mutated in a variety of tumor types. The human papilloma virus (HPV) is associated with the development of cervical carcinoma, and peptide epitopes identified from viral proteins represent additional targets for vaccine therapies. A variety of methods have been used for peptide delivery; peptides have been administered in soluble form, emulsified in adjuvants, and pulsed on antigen-presenting cells (APC) such as dendritic cells (DC). A variety of cytokines, including interleukin (IL)-2, granulocyte/monocyte colony-stimulating factor (GM-CSF), and IL-12, have also been administered in attempts to enhance immune responses to peptide immunization.
In these primarily phase I trials, while the lack of toxicity of these treatments has been demonstrated, substantial tumor regressions have only been observed in a limited number of patients. Nevertheless, a critical feature of these trials, as well as future clinical trials, is the development of effective strategies for monitoring immune responses in vaccinated patients, which include in vivo as well as in vitro assays. It is hoped that the careful evaluation of immune responses to vaccination, in conjunction with animal model studies, will result in the development of effective vaccines for the treatment of cancer patients.
I. MAJOR HISTOCOMPATIBILITY COMPLEX CLASS I- AND CLASS IIRESTRICTED T-CELL RECOGNITION
Cytotoxic T lymphocytes (CTL) recognize short peptides in association with class I MHC molecules. Peptide titration studies, as well as studies carried out with peptides of varying lengths, have demonstrated that the optimal peptides generally have ranged between 8 and 10 amino acids. Studies carried out in the early 1990s by Rammensee and others demonstrated that peptides eluted from class I MHC molecules, in addition to ranging between 8 and 10 amino acids in length, contained characteristic amino acids at particular positions in the sequence. For example, peptides that bound to the human HLA-A2 class I gene product generally possessed an aliphatic residue such as L or M at position 2 (P2) and a V or L residue at the carboxy terminus of the peptide (P9 or P10) (Table I). The amino acids at these positions, which have been termed primary anchor residues, appear to be critical residues for binding to class I MHC molecules, and the solution of the crystal structure of MHC class I gene products containing bound peptides has confirmed these findings. As a result of these studies, binding motifs have now been defined for multiple HLA class I alleles.
TABLE I Melanoma Antigen T-Cell Epitopes: Comparison with the HLA-A2-Binding MotifPosition 1 2 3 4 5 6 7 8 9 Anchors or auxiliary anchorsa L V V M L Preferred residues E K K Other residues I A G I I A E L Y P K L Y S F F D Y T H K P T N M M G Y S F V R V H HLA-A2-restricted melanoma antigen epitopes MART-1 27-35 A A G I G I L T V gp100 154-162 K T W G Q Y W Q V gp100 209-217 I T D Q V P F S V gp100 280-288 Y L E P G P V T A gp100 619-627 R L M K Q D F S V gp100 639-647 R L P R I F C S C gp100 457-466 L L D G T A T L R L gp100 476-485 V L Y R Y G S F S V Tyrosinase 1-9 M L L A V L Y C L Tyrosinase 369-377 Y M D G T M S Q V TRP-2180-188 S V Y D F F V W L
aData indicating primary and secondary anchor residues and residues and preferred residues, as well as other residues commonly observed in HLA-A2-binding peptides, were obtained from the internet site www.uni-tuebingen.de/uni/kxi/database.html.
Extensive studies have also been carried out to define the nature of CD4+ helper T-cell epitopes, which are recognized in association with MHC class II molecules. The length of MHC class II-binding peptides varies more than class I-binding peptides, and it has been more difficult to establish binding motifs for class II alleles. Nevertheless, minimal epitopes of 9 or 10 amino acids have been identified for a number of MHC class II antigens. Compilations of these sequences have also resulted in the identification of two or three primary anchor residue positions that appear to be involved in binding to particular HLA alleles, and determination of the crystal structure of class II MHC molecules bound to peptides has confirmed these results.
II. IDENTIFICATION OF ANTIGENS RECOGNIZED BY CLASS I-RESTRICTED, TUMORREACTIVE T CELLS
In the early 1990s, efficient methods were developed for the isolation of genes that encode antigens recognized by tumor reactive T cells. Tumor-specific T-cell lines and clones were initially identified from tumor-infiltrating lymphocytes (TIL), as well as peripheral blood mononuclear cells (PBMC) that had been stimulated in vitro sensitization with tumor cells. The genes that encode tumor antigens were primarily identified through the screening of cDNA libraries generated from tumor cell mRNA. This was carried out by transfecting DNA that was isolated from pools of cDNA clones into antigen-negative target cells that expressed the appropriate MHC restriction element.
The transfected target cells were then assayed for their ability to stimulate the release of cytokines such as interferon-γ (IFN-γ), tumor necrosis factor α (TNF-α), or GM-CSF from tumor-specific T cells. Positive pools were then subdivided and screened using an iterative process until an individual positive cDNA clone encoding the tumor antigen was identified. The majority of antigens identified were isolated following screening with melanoma reactive T cells; however, T cells isolated from other tumor types, such as renal and squamous carcinomas, have also been utilized to identify antigens expressed in these tumors.
These antigens have been sorted into several general categories on the basis of their expression patterns. Melanocyte differentiation antigens represent a family of proteins that are limited in their expression to normal melanocytes and melanomas. T cells have been shown to recognize nonmutated epitopes derived from several members of this family, including MART-1, gp100, TRP-1, TRP-2, and tyrosinase. One epitope appears to predominate in the HLA-A2- restricted T-cell response to MART-1, but for certain gene products, such as gp100, several epitopes have been identified. Based on the frequency of responsiveness of HLA-A2-restricted TIL, as well as sensitized PBMC, three peptides appear to predominate in the response to this antigen: gp100:209-217 (ITDQVPFSY), gp100:280-288 (YLEPGPVTA), and gp100:154-162 (KTWGQYWQV).
Members of a second family of antigens, the cancer-testis antigens, are limited in their expression in normal tissues to the testis. These include members of the MAGE, BAGE, GAGE, and NY-ESO-1 protein families. These gene products, in contrast to melanocyte differentiation antigens, are expressed in a variety of tumor types, including breast, lung, bladder, and prostate cancers. The NY-ESO-1 gene product was initially identified by screening the reactivity of patient sera against a phage cDNA expression library. The NY-ESO-1 product is expressed in about 30% of patients with these cancers, and subsequent studies have shown that the majority of these patients possess antibody directed against this gene product.
Mutated antigens have been identified in melanoma as well as a number of additional tumor types. Examples of mutated epitopes that are expressed in more than a single tumor have been described; however, only a small minority of all the tumors of a particular type express a particular mutation. These antigens, which represent foreign epitopes, thus may be more immunogenic than normal self-peptides identified from antigens such as MART-1 and gp100; however, only a small percentage of patients can be treated with these vaccines.
Unique mechanisms appear to be involved in maintaining self-tolerance toward certain antigens that are expressed in a variety of normal tissues. A T-cell clone directed against the PRAME melanoma antigen expressed an NK inhibitory receptor that recognized a product of the HLA-C gene products. In addition, evidence shows that professional antigen-presenting cells may fail to efficiently present certain melanoma T-cell epitopes as a result of altered processing. Additional studies have focused on the identification of analogs of tumor antigen epitopes with enhanced immunogenicity. Results of in vitro-binding assays indicate that the immunodominant HLA-A2 MART-1 epitope, as well as a majority of the gp100 epitopes recognized in the context of HLA-A2, possesses a relatively low affinity for this HLA allele. In addition, these peptides contain residues that fail to conform to the binding motif at one of the primary anchor residue positions. As described further later, peptides with enhanced binding to HLA-A2, as well as with enhanced immunogenicity, can be derived by substituting an optimal for a nonoptimal anchor residue.
III. IDENTIFICATION OF ANTIGENS RECOGNIZED BY TUMOR-REACTIVE, CLASS IIRESTRICTED T CELLS
Antigens that are recognized by class II-restricted, tumor reactive T cells have been identified using a number of methods, including the screening of cDNA libraries using techniques similar to those used to identify genes that encode class I-restricted antigens. Class II-restricted, melanoma-reactive T cells have also been examined for their ability to recognize target cells that have been transfected with genes that encode known class I-restricted melanoma antigens. Using this approach, class II-restricted epitopes from tyrosinase, gp100, TRP-1, and TRP-2 have been identified. Multiple HLA class II-restricted T-cell epitopes have also been identified from the MAGE-3 protein by carrying out in vitro sensitizations using antigen-presenting cells that had been pulsed with candidate peptides.
Antigens that are recognized in the context of HLA class II alleles can be processed through either an endogenous pathway or an exogenous pathway resulting from endocytosis in antigen-presenting cells such as monocytes, dendritic cells, and B cells. Epitopes from the NY-ESO-1 and MAGE-3 cancer-testis antigens have been identified by stimulating with APC that have been pulsed with purified recombinant proteins, followed by screening either candidate peptides identified using HLA class II-binding motifs or a library of overlapping peptides.
IV. IDENTIFICATION OF SHARED EPITHELIAL TUMOR ANTIGENS
Several attempts have been made to generate T-cell responses against candidate antigens that are expressed in common tumor types, such as breast and ovarian cancer, as many of the antigens recognized by melanoma reactive T cells are limited in their expression to cells of this tissue type. The HER-2/neu oncogene, a member of the epidermal growth factor receptor family, has been shown to be widely overexpressed in breast, colorectal, and ovarian adenocarcinomas, and overexpression appears to be correlated with poor prognosis in patients. The peptide KIFGSLAFL, representing amino acids 369-377 (p369), has been shown by a number of groups to be recognized by HLA-A2-restricted T-cells clones and lines isolated from patients with breast and ovarian cancer; however, additional epitopes have also been identified from this molecule.
Viral antigens, which are expressed in certain tumor types, represent attractive targets for vaccine therapies. The HPV is expressed in over 90% of cervical carcinomas, and the HPV 16 genotype is found in 50% of squamous carcinomas of the cervix. HPV E6 and E7 proteins appear to be involved in tumorigenesis, and selective pressure to maintain expression of these products may prevent the generation of antigen loss variants. Studies carried out in HLA-A2 transgenic mice resulted in the identification of several candidate epitopes from these gene products. The peptide corresponding to amino acids 86-93 of the HPV E7 protein (TLGIVCPI) has been shown to bind with high affinity to HLA-A2, and the stimulation of human lymphocytes with this peptide appears to be capable of raising tumor-reactive CTL.
Antigens that are overexpressed in particular tumor types represent potential targets for the development of anticancer vaccines. The carcinoembryonic antigen (CEA) is a tissue-specific gene product that is expressed in a high percentage of colon tumors as well as breast carcinomas. Although a low level of CEA expression has been observed in normal colon tissue, this gene product appears to be overexpressed in tumor cells, and thus may represent a target for vaccine therapies. Several closely related gene products, which include normal cross-reacting antigen (NCA) and biliary glycoprotein (BGP), are expressed in a variety of normal tissues. Studies described in more detail later have resulted in the identification of an HLA-A2-restricted T-cell epitope that is expressed in the CEA, but not in the NCA or BGP gene products.
Differentiation antigens whose expression is restricted to prostate tissue also represent candidate antigens for tumors of this histology. The prostate-specific antigen (PSA), as well as prostate-specific membrane antigen (PSMA), is highly restricted in its expression to normal prostate and appears to be highly expressed by the majority of tumors of this histology. Increased expression of these markers in patient serum is associated with progressive disease, and clinical trials utilizing peptides derived from these proteins are detailed in the following section.
V. CLINICAL VACCINE TRIALS
A. Melanoma Antigens
Peptide vaccine trials were initially carried out in melanoma patients utilizing epitopes from MAGE-1 and MAGE-3 antigens that are recognized in the context of the HLA-A1 class I allele. Patients that expressed this HLA allele and whose tumors were shown to express MAGE-3 by RT-PCR were immunized with the MAGE-3 peptide EVDPIGHLY in soluble form. In one of the initial clinical trials, 12 melanoma patients were treated with this peptide, which was administered subcutaneously as well as intradermally in saline. This peptide was injected monthly, but 6 of the 12 patients were withdrawn after one or two injections because of the rapid progression of the disease. Partial regressions were reported in 3 of the 6 remaining patients, but cells in the peripheral blood of immunized patients did not show evidence for increased reactivity to the MAGE-3 peptide.
Another melanoma clinical trial has involved intradermal injection of a total of six HLA-A2-restricted peptides, two derived from each of the melanocyte differentiation antigens MART-1, gp100, and tyrosinase. All of the peptides utilized in this trial had previously been shown to be recognized by melanomareactive CTL. Six patients with metastatic melanoma were immunized with 100 μg of each of these peptides weekly for 4 weeks, and delayed type hypersensitivity (DTH) responses were assessed by examining skin induration 24-48 h following peptide injection. In five out of the six patients, some evidence for a specific DTH response was seen with one of the tyrosinase peptides, but not with the other peptides.
Some evidence was found for increased reactivity against the MART-1 peptides and one of the tyrosinase peptides following vaccination; however, no tumor regression was observed in any of the immunized patients. In an attempt to further boost tumor-specific immune responses, GM-CSF was administered to patients that were vaccinated with five peptides derived from MART-1, tyrosinase, and gp100. Injection of GM-CSF appeared to increase the CTL activity directed against tyrosinase peptides in the three patients that were examined and appeared to enhance reactivity against MART-1 in one out of three patients; however, no significant clinical responses were observed in this trial.
Vaccinations have also been carried out using APC that have been pulsed with peptides identified from tumor antigens. A clinical trial carried out in melanoma patients involved intradermal injections of cultured monocytes that had been pulsed with the MAGE-1 HLA-A1 peptide. The results of patient monitoring, which was carried out by testing skin reactions to intradermal peptide inject, indicated that MAGE-1-specific immune responses were not induced in vaccinated patients. When assays were carried out on T cells that had been expanded in vitro using peptide-pulsed target cells, a low level of peptide reactivity, as well as tumor reactivity, appeared to be present in postimmunization, but not preimmunization samples from some patients. No specific therapeutic responses were observed in this trial. In a separate clinical trial, melanoma patients were immunized with APC that had been pulsed with MAGE-1:161-169 and MAGE-3:168-176 HLA-A1- restricted epitopes, along with MART-1:27-35, tyrosinase 1-9, and gp100:154-162 HLA-A2-restricted peptides.
The pulsed target cells were injected into an uninvolved inguinal lymph node at weekly intervals, and patients were evaluated for clinical responses as well as DTH responses to the peptides. One complete and one partial response were observed in the six HLA-A2+ patients that were immunized with the HLA-A2 restricted peptides, and one out of the six HLA-A1+ patients that had been immunized with the MAGE peptides demonstrated a partial response. The results of skin testing indicated that the majority of patients had been primed by immunization with the peptide-pulsed dendritic cells.
A recent report has detailed the results of a clinical trial in which melanoma patients received intradermal as well as subcutaneous injections of dendritic cells that had been pulsed with the MAGE-3 HLAA1- restricted epitope. Vaccination appeared to result in the expansion of MAGE-3 peptide reactive T cells in 8 out of 11 patients, as determined by semiquantitative precursor frequency analysis. Complete regression of individual metastases was observed in 6 out of the 11 patients; however, disease progression was seen in all of the patients.
Studies carried out in the Surgery Branch of the National Cancer Institute focused initially on immunization with the MART-1:27-35 peptide (AAGIGILTV). Immunization of 18 patients with metastatic melanoma with this peptide was carried out subcutaneously in incomplete Freund’s adjuvant (IFA) at 3-week intervals. In 15 of the patients, evidence for an increase in the precursor frequency of T cells reactive with the MART-1:27-35 peptide was seen following immunization. No evidence was obtained for a difference in the response of patients to doses that varied from 0.1 to 10 mg per injection. In 1 of the patients that appeared to respond to immunization, a partial response was observed that is ongoing for 4 years, but significant tumor regression was not observed in the other patients. A minimum of three in vitro stimulations appeared to be required to elicit significant responses from both immunized and nonimmunized PBMC samples, indicating that the precursor frequency of MART-1-reactive T cells was similar in both nonimmunized and immunized patients.
Peptides from the gp100 antigen have also been tested for their ability to stimulate immune responses in HLA-A2 melanoma patients. Patients were immunized with gp100:209-217 (ITDQVPFSY), gp100:280-288 (YLEPGPVTA), or gp100:154-162 (KTWGQYWQV), and in vitro responses to the peptides were analyzed following several in vitro stimulations. The patients were immunized subcutaneously with antigen emulsified in IFA at 3-week intervals. One of the patients that received the gp100:209-217 peptide had a complete response that lasted 4 months, but not other significant clinical responses were seen. Evidence for in vivo priming was found in patients that had received the gp100:209-217 and 280-288 peptides, but not in patients that had received the gp100:154-162 peptide.
Several studies have focused on the identification of modified T-cell epitopes with enhanced immunogenicity. The immunodominant MART-1 epitope, as well as the three immunodominant HLA-A2-binding epitopes from gp100, does not appear to conform to the optimal HLA-A2-binding motif. Immunization with the modified gp100:209-217(2M) peptide, containing a substitution of an optimal M residue for T at the P2 position, appeared to significantly enhance HLA-A2 binding, as well as the in vitro immunogenicity of this peptide. Modification of the gp100:280-288 peptide by the substitution of V, which represent the optimal C-terminal anchor residue, for A at the P9 position similarly enhanced peptide binding to HLA-A2 and in vitro immunogenicity.
These results led to clinical trials involving vaccination with modified tumor antigen peptides. In one trial carried out in the Surgery Branch of the National Cancer Institute, melanoma patients were immunized with the modified gp100:209-217(2M) peptide in IFA. The peptide-specific, as well as tumor-specific, responses of PBMC isolated from immunized patients were assessed following a single in vitro stimulation with peptide. Specific responses were elicited in 10 out of 11 patients immunized with the gp100:209-217(2M) peptide, but only 2 out of 8 patients immunized with the parental peptide. Results indicate that PBMC obtained before the in vivo immunization failed to respond to a single in vitro peptide stimulation, in agreement with previous results (Table II). In contrast, PBMC from the majority of the patients that were obtained following the in vivo immunizations demonstrated a vigorous in vitro response to peptide-pulsed targets, as well as HLA-A2+ melanoma cells (Table II). Significant clinical responses were not, however, seen in patients that received either the native or the modified gp100 peptide alone in IFA. An additional group of 31 patients were immunized with the modified gp100:209-217 peptide in IFA along with high-dose IL-2, and in this group, 42% demonstrated a partial or complete response. Although the response rate to high-dose IL-2 seen in previous clinical trials was only 17%, a larger randomized trial will be needed to determine the significance of these results.
TABLE II Reactivity of PBMC from Patients Immunized with the 209-2M PeptidePatient No. of immunizations T2 -- T2 (280) T2 (209) Stimulator a 501 mel (A2+) (pg/ml IFN-γ) SK23 mel (A2+) 888 mel (A2) 624.28 mel (A2) 7 0 169 175 220 28 72 84 51 2 209 243 2,445 1,211 2,037 98 60 8 0 528 691 729 70 640 933 806 2 202 284 13,600 11,580 14,720 408 489 9 0 13 13 10 NDb ND ND ND 2 229 590 3,987 67 889 291 235 10 0 117 147 150 19 90 39 42 2 15 18 24,040 23,860 21,580 2 4 11 0 46 50 47 11 39 14 17 2 29 30 106 5 43 4 10
aPBMC were incubated with 1 μM gp100:2092M peptide for 13 days and tested for their response to T2 cells that were pulsed with the native gp100:209 peptide, the irrelevant gp100:280 peptide, and melanoma cell lines that either did or did not express HLA-A2. Data taken from Rosenberg et al. (1999). J. Immunol. 163, 1690.
Curiously, the frequency of T cells reactive with the gp100:209-217 peptide appeared to be lower in patients that received injections of the modified peptide plus IL-2 than in patients that received injections of the peptide alone. This may have resulted from the redistribution of specific T cells from the blood to the site of tumor, as well as from apoptosis of the injected T cells resulting from stimulation by tumor cells. As discussed earlier, the majority of the clinical responses were actually observed in patients that received injections of peptide plus IL-2. Additional studies are needed to help resolve this issue.
Peptide variants of the MART-1 epitope with enhanced immunogenicity than the native peptide have also been described. Modified peptides containing a substitution of leucine or methionine for the alanine residue at the P2 anchor residue position in the MART-1 10-mer EAAGIGILTV, as well as a peptide containing a substitution of methionine for alanine at the P2 position in the MART-1 9-mer AAGIGILTV, possessed a higher HLA-A2 binding affinity than the parental peptides. The results of in vitro sensitization studies indicated that the modified MART-1 9-mer and 10-mer peptides could generate tumor-reactive T cells more efficiently than the parental peptides. Clinical trials are now being carried out to evaluate the in vivo efficacy of these modifications.
B. Epithelial Tumor Cell Antigens
Previous observations have indicated that stimulation with the HLA-A2-restricted p369 peptide derived from HER/2-neu resulted in the generation of tumor reactive T cells. In a trial carried out in the Surgery Branch of the National Cancer Institute, four HLAA2 + patients with metastatic breast, ovarian, or colorectal adenocarcinomas were immunized with the p369 peptide in IFA. Peptide-reactive T cells were generated following a single in vitro stimulation of PBMC from three of the four immunized patients, and this response appeared to be dependent on in vivo priming. These T cells recognized antigen-negative HLA-A2+ target cells that were pulsed with as little as 1 ng/ml of the specific peptide; however, multiple HLA-A2+ tumor cells that expressed high levels of HER-2/neu were not recognized. In addition, HLAA2 + target cells that were infected with a vaccinia virus construct encoding HER-2/neu, and which consequently expressed high levels of this gene product, were not recognized by the peptide-reactive T cells. It is possible that T cells with a relatively low affinity for this peptide-MHC complex were generated by in vivo immunization and that only low levels of the p369 epitope are naturally processed and presented on the tumor cell surface. The multiple in vitro stimulations carried out by previous investigators may have resulted in the stimulation of T cells with a higher affinity than those generated from immunized patients. Further investigation is needed to resolve the discrepancies between these findings, but these results indicate potential difficulties that may be associated with the use of candidate peptides.
A clinical trial carried out in 51 prostate cancer patients involved immunizations with two HLA-A2- binding peptides from PSMA. Immunizations were carried out by an intravenous injection of either the peptides alone or peptide-pulsed autologous dendritic cells. Injections of between 106 and 2 × 107 peptidepulsed autologous dendritic cells were carried out at 6- to 8-week intervals, with patients receiving between four and five cell injections. An increase in the in vitro response to PMSA was reported in immunized patients, and the circulating levels of PSA were reported to decrease; however, only limited evidence for objective clinical responses has been presented. Candidate epitopes have also been identified in the PSA antigen. A single 30-mer peptide comprising two HLA-A2-binding peptides, as well as an HLAA3- binding peptide, all derived from PSA, has been tested in vitro as a multivalent vaccine candidate. Following in vitro stimulation of PBMC with the 30-mer peptide, CTL were generated that appeared to recognize targets pulsed with each of the individual peptides as well as tumor cells.
In another clinical trial, multiple injections of a recombinant vaccinia virus encoding CEA were administered to colon cancer patients. A peptide from CEA termed CAP-1 (YLSGANLNL), which appeared to possess a relatively high binding affinity for HLA-A2, was then used to stimulate PBMC from HLA-A2+-vaccinated patients. Peptides derived from a similar region of the closely related NCA and BGP proteins, which are expressed in a variety of normal tissues, contained several substitutions from the CAP-1 peptide. Following several rounds of in vitro stimulation with the CAP-1 peptide, CTL lines generated from vaccinated patients appeared to recognize peptide-pulsed targets, as well as HLA-A2+, CEA-expressing tumors. Significantly weaker responses were elicited when PBMC obtained prior to vaccination were stimulated with the CAP-1 peptide. Additional studies have indicated that a substitution of D for N at position 6 of this peptide resulted in the generation of a peptide with enhanced in vitro immunogenicity, which will be tested in future clinical trials.
C. Mutated Oncogenes and Viral Antigens
Tumors have been shown to express a wide variety of mutated gene products, and although many of these appear to be restricted in their expression to either one or a relatively small number of tumors, a number of common mutations have also been identified. Common point mutations in the ras and p53 oncogenes have been identified in a variety of tumor types, and the identification of mutated T-cell epitopes derived from these gene products could lead to the development of widely applicable vaccines. Codon 12 of the ras oncogene is frequently mutated in epithelial tumors, and over 90% of the mutated ras gene products contain a substitution of alanine, valine, or cysteine for the normal glycine residue at this position. In one study, ras mutations were identified in tumors obtained from patients with pancreatic adenocarcinomas.
Synthetic peptides comprising amino acids 5-21 of the mutated ras gene product were used to pulse autologous PBMC, which were then injected intravenously into patients with pancreatic adenocarcinomas on days 14 and 35 and subsequently at an interval of 4 to 6 weeks. In 2 of the 5 patients injected with the mutated ras peptide, transient peptidespecific proliferative responses were observed. Peptide-specific class I as well as class II-restricted T cells appeared to be elicited by in vitro stimulation of PBMC obtained from one of the patients vaccinated with the mutated ras peptide. In another study, ras mutations were initially identified in tumors isolated from patients with colon, lung, or pancreatic adenocarcinomas.
Patients were immunized with mutated ras peptides whose sequences corresponded to those found in autologous tumor cells. Peptides were administered in Detox, an adjuvant composed of an oil in water emulsion with an added bacterial cell wall and lipid A components. In 3 out of the 8 vaccinated patients, T-cell lines were generated following in vitro stimulation of PBMC. These T-cell lines recognized the mutated ras peptide, but failed to respond to the nonmutated peptide. In a second study carried out by the same group, 15 patients were immunized with mutated ras peptides that had been emulsified in Detox adjuvant, and immune responses were evaluated in 10 patients that received a full course of three immunizations.
Peptide-specific responses, which consisted of either CD4+ or CD8+ T cells, could be generated by in vitro stimulation of PBMC obtained from 3 of the 10 immunized patients. Clinical antitumor responses were noted in these trials, which may be a result of the advanced state of disease in these patients. The association of HPV 16 and 18 with cervical cancer has led to clinical trials involving vaccination with candidate epitopes derived from the viral E6 and E7 sequences. Initial studies suggested that peptide as well as tumor-reactive T cells could be elicited by in vitro stimulation of PBMC from HLAA2 cervical carcinoma patients with a peptide consisting of amino acids 11 to 20 of the HPV16-E7 protein (E711-20:YMLDLQPETT). In another report, an in vitro culture of PBMC with DC that had been pulsed with recombinant E7 protein appeared to result in the generation of both class I- and class II-restricted T cells reactive with this antigen. Patients with cervical carcinoma were immunized with two HPV16 CTL epitopes, the E7:11-20 peptide discussed earlier and a second peptide comprising amino acids 86 to 93 of the E7 protein. These peptides were administered in combination with a pan-class II peptide, termed PADRE, that is capable of eliciting T-cell help in patients that express a variety of HLA haplotypes. Strong responses to the PADRE peptide could be generated in 4 out of the12 immunized patients, indicating that at least some of the patients were immunocompetent. Decreased responses were also observed with PBMC from the majority of the cervical carcinoma patients following in vitro challenge with an HLA-A2-restricted influenza peptide, as well as with the other recall antigens. A number of the patients in this study had received prior treatments, such as chemotherapy, which may have had an impact on their response to these antigens.
Evidence for depressed immune responses that are not related to therapy has been obtained in studies of patients with certain cancers, as well as in some mouse model systems. As cited earlier, however, immunization of cancer patients with peptides from the gp100 and HER-2/neu antigens appears to result in the generation of brisk peptide-specific responses. Evaluation of the role of general as well as specific suppression in modulating antitumor responses may lead to the development of strategies that result in enhanced tumor regression.
VI. FUTURE DIRECTIONS
Use of the vaccine strategies described earlier have resulted in clinical responses in only a small minority of the treated patients. These clinical trials are predominantly in phase I testing, however, and the lack of response may have been due to the advanced stage of cancer present in these patients. Future trials directed toward prevention of recurrence in patients with no or minimal disease may demonstrate the efficacy of these approaches. Further evaluation of a variety of additional approaches to tumor vaccine therapy approaches is also needed. Vaccination with multiple class I- and class II-restricted tumor peptides may be more effective than treatment with individual epitopes. In addition, immunization with purified recombinant tumor antigen proteins may elicit cell-mediated as well as humoral responses against multiple epitopes and may be more effective at mediating tumor regression than immunization with peptides from these antigens.
Paul F. Robbins
National Institutes of Health, Bethesda, Maryland
ANTI-IDIOTYPIC ANTIBODY VACCINES ; CANCER VACCINES: GENE THERAPY AND DENDRITIC CELL BASED VACCINES ; CARBOHYDRATE-BASED VACCINES ; DNA-BASED CANCER VACCINES ; TARGETED TOXINS ; TUMOR ANTIGENS
adjuvant Material used to enhance the immunogenicity of a protein or peptide.
cytokines Secreted proteins involved in the activation and proliferation of immune cells.
cytotoxic T lymphocytes Immune cells capable of lysing targets expressing the appropriate antigen.
epitope Short peptide sequence bound to MHC gene product that serves as the target recognized by the antigenspecific T-cell receptor.
immunogenicity Ability of a particular molecule to elicit a specific immune response.
major histocompatibility complex (MHC) Locus encoding polymorphic genes utilized by the immune systems for antigen presentation and for self/nonself recognition.
melanocyte Skin cell responsible for the production of the pigment melanin.
transfection Introduction of DNA into a cell using physical manipulation.
Boon, T., and Van den Eynde, B. J. (2000). Shared tumorspecific antigens. In "Biologic Therapy of Cancer" (S. A. Rosenberg, ed.), 3rd Ed., p. 493.
Lippincott Williams & Wilkins. Coulie, P. G., Brichard, V., Van Pel, A., Wolfel, T., Schneider, J., Traversari, C., Mattei, S., De Plaen, E., Lurquin, C., Szikora, J. P., Renauld, J.-C., and Boon, T. (1994). A new gene coding for a differentiation antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 melanomas. J. Exp. Med. 180, 35.
Kawakami, Y., Eliyahu, S., Delgado, C. H., Robbins, P. F., Rivoltini, L., Topalian, S. L., Miki, T., and Rosenberg, S. A. (1994). Cloning of the gene coding for a shared human melamona antigen recognized by autologous T cells infiltrating into tumor. Proc. Natl. Acad. Sci. USA 91, 3515.
Nestle, F. O., Alijagic, S., Gilliet, M., Sun, Y., Grabbe, S., Dummer, R., Burg, G., and Schedendorf, D. (1998). Vaccination of melanoma patients with peptide- or tumor lysatepulsed dendritic cells. Nature Med. 4, 328.
Robbins, P. F. (2000). Differentiation antigens. In "Biologic Therapy of Cancer" (S. A. Rosenberg, ed.), 3rd Ed., p. 504. Lippincott Williams & Wilkins.
Rosenberg, S. A., Yang, J. C., Schwartzentruber, D. J., Hw, P., Marincola, F. M., Topalian, S. L., Restifo, N. P., Dudley, M. E., Schwarz, S. L., Spiess, P. J., Wunderlich, J. R., Parkhurst, M. R., Kawakami, Y., Seipp, C. A., Einhorn, J. H., and White, D. E. (1998). Immunologic and therapeutic evaluation of a synthetic vaccine for the treatment of patients with metastatic melanoma. Nature Med. 4, 321.
van der Bruggen, P., Traversari, C., Chomez, P., Lurquin, C., De Plaen, E., Van den Eynde, B., Knuth, A., and Boon, T. (1991). A gene encoding a tumor antigen recognized by cytolytic T lymphocytes on human melanomas. Science 254, 1643.
van der Bruggen, P., Traversari, C., Chomez, P., Lurquin, C., De Plaen, E., Van den Eynde, B., Knuth, A., and Boon, T. (1996). Generation of cytotoxic T-cell responses with synthetic melanoma-associated peptides in vivo: Implications for tumor vaccines with melanoma-associated antigens Int. J. Can. 66, 162.