The identification of cancer antigens for immunotherapy represents a crucial step towards clinical immunotherapy. Tumor antigens are divided into tumor-unique antigens, whose expression is restricted to tumor tissue- or tumor-associated antigens which are usually overexpressed in tumors, but may show low levels of expression in normal tissues. In functional terms, TAs can be broadly divided into antigens that are required for tumor development and progression (indispensable antigens) and those antigens that are nonessential. Immunotherapy approaches targeting antigens that are not crucial for tumor development may eventually result in antigen-loss variants arising within the tumor, eventually leading to tumor escape. Therefore an ideal TA candidate would be quintessential for tumor development and would also be expressed in a wide variety of tumors making it a “Universal TA” [4]. The National Cancer Institute recently conducted a program to prioritize cancer antigens to establish a list of “well-vetted,” priority ranked TA targets based on predefined unprejudiced criteria [5]. Adopting a pairwise approach, the criteria weighting for antigens, in descending order, was as follows: (a) therapeutic function, (b) immunogenicity, (c) role of the antigen in oncogenicity, (d) specificity, (e) expression level and the percentage of antigen-positive cells, (f) stem cell expression, (g) number of patients with antigen-positive cancers, (h) number of antigenic epitopes, and (i) cellular location of antigen expression [5].

The initial step in reverse immunology is the search for protein/gene expression patterns selectively observed in tumor cells. Protein overexpression could be detected by using techniques such as immunofluorescence, western blotting, flow cytometry, etc. The shortcoming of these techniques is that they are time consuming and depend on the availability of antibodies with high sensitivity and specificity. Also, with this method it is difficult to estimate protein turnover [6]. New large-scale gene expression assays such as cDNA microarrays, oligonucleotide chips, cDNA library sequencing, serial analysis of gene expression (SAGE), massively parallel signature sequencing (MPSS), subtractive hybridization, differential display PCR, and representational difference analysis (RDA) are a few of the techniques widely used to decipher complex expression patterns and identify new candidate antigens. SEREX (serological analysis of recombinantly expressed clones) has been a widely used technique that relies on the use of cancer patients’ sera to screen selected tumor cDNA libraries [7, 8]. Tumor antigens identified via this method usually contain CD4⁺ T-helper lymphocyte epitopes. The first tumor antigen, MAGE-1, was discovered through autologous typing and application of a newly developed DNA-cloning technique for defining the targets of T-cell recognition [9]. A melanoma patient with unusually favorable clinical course was identified to have CTLs that recognized autologous tumor cells. With antigen-specific T cells as a reagent and through the use of cosmid gene libraries, it was possible to identify and clone the MAGE-1 gene. Studies on MAGE-1 showed for the first time that the human immune system can respond to TA, and the findings transformed the study of tumor antigens and stirred a dynamic effort to discover tumor antigens, which has resulted in a long and still-growing list of antigens from a variety of tumors which potentially serve as targets for immunotherapy.

The use of antigenic epitopes/peptides to promote antitumor immunity represents one of the simplest and most applicable ways of targeting cancer cells expressing the respective protein. Recent studies have shown that immunotherapy approaches targeting multiple epitopes at the same time [10] or use of long synthetic peptides that would comprise multiple epitopes [11] can result in delivering clinical benefits to a greater number of patients. This strategy would broaden the clinical response in several ways:

A number of computer-aided tools have been developed for the prediction of T-cell epitopes. These algorithms are based on the natural processing and presentation of proteins; the number of computer-aided algorithms available for the prediction of better T-cell epitopes reflects the existing level of understanding antigen processing. Thus, the tools designed for T helper epitope prediction are comparatively less advanced than for CTL epitope prediction.

For a target protein to be successfully presented on an MHC class I or class II molecule, it must undergo a number of processing steps resulting in the transport and cleavage of the peptide. Epitopes presented on MHC class I molecules are 8–11 amino acid length chains and are predominantly derived from intracellular proteins. A cytosolic multi-subunit proteolytic complex, known as the proteasome, degrades proteins to peptides which are later transported into the endoplasmic reticulum (ER) by the adenosine triphosphate (ATP)-dependent transporters associated with antigen processing (TAPs) [12]. The alpha chain of MHC I binds to the beta-2 microglobulin unit with the help of calnexin. Once stable, calnexin is replaced with calreticulin and tapasin [13]. Within the ER, peptides undergo further N-terminal trimming before their subsequent loading into the empty MHC-binding cleft. The epitope binds tightly to the epitope-binding cleft stabilizing the trimeric complex transported to the cell surface via the ER and Golgi network [13]. Exogenous proteins are processed mainly by the MHC class II pathway. The alpha and beta subunits of the MHC class II molecule are preoccupied by a non-polymorphic invariant chain (Ii) which acts as a chaperone for class II folding and prohibits binding of intracellular proteins to MHC II [14]. The extracellular proteins are engulfed by endosomes which transport them to the Golgi apparatus. The inactive MHC II molecule is also transported to the Golgi where proteolytic degradation of Ii occurs, leaving the class II-associated invariant peptide (CLIP) in the peptide-binding cleft. The MHC II–CLIP complex can then interact with human leucocyte antigen (HLA)-DM (H-2M in mouse) which activates the dissociation of CLIP, allowing the loading of peptides into the empty MHC class II cleft [14]. The development of software prediction algorithms is based on peptide–MHC interactions or on proteasomal degradation.

In the validation phase of reverse immunology, the natural presentation and immunogenicity of the selected epitopes should be corroborated. Using cell lines or tumor tissues expressing the appropriate antigen and HLA allele, biochemical methods can be used to elute peptides from the cell surface or from isolated HLA antigen. The purified products are then analyzed by mass spectrometry [26] to derive sequence information and identify the target peptide (as discussed in detail below). Though this technique confirms the expression of HLA-bound ligands, it does not allow the assessment of peptides’ immunogenicity. Assessing the immunogenicity of predicted peptides relies on demonstrating their ability to stimulate MHC class I- or class II-restricted T-lymphocyte responses. These may be either a primary response, where naïve T cells respond to antigens in culture or secondary, where, for example, patient CD8⁺ or CD4⁺ T cells, already exposed to antigen in vivo, demonstrate a secondary response. However, patient response to self-(tumor) antigens is generally quite weak, and T lymphocytes may become tolerant towards this antigen. Tolerance may be overcome by exposure of patient lymphocytes to a combination of interleukin (IL)-2 and IL-12 in vitro, which enhance the tumor-specific CD8⁺ T-cell response and additionally prevent overgrowth of nonspecific, less-effective lymphokine-activated killer cells [27]. IL-12 is a potent inducer of tumor-specific CTLs and promotes the production of Th1 cytokines [28].

The immunospot assay, which is based on the detection of cytokine secretion in response to antigen, is used to detect antigen-reactive T-cell responses. Most current assays for measuring T-cell cytotoxicity are based on alterations in plasma membrane permeability and the subsequent release (leakage) of components into the supernatant (51Cr, lactate dehydrogenase assays) or the uptake of dyes (CFSE), which are normally excluded by viable cells. Another alternative is the use of flow cytometry to detect the expression of CD107 in the membrane, which is transiently expressed during the process of cell killing [29]. Use of cytokine-secretion assays, intracellular cytokine assays, HLA class I multimer staining (e.g., peptide-specific tetramers), etc. are among the techniques that have been thoroughly validated and established recently [30]. The use of tetramers has proved to be especially successful for the identification of peptide-specific CD8⁺ T lymphocytes and to a lesser extent for CD4⁺ T helper cells.

The techniques commonly used for isolation of MHC-associated peptides usually involve immunoaffinity chromatography and acid elution. Immuno-affinity chromatography (IAC) combines the use of LC with the specific binding of HLA antigens to antibodies or related agents [31]. The source material for this approach is usually frozen tissue, blood cells, or cultured cell lines. Solid tissue is first mechanically dissociated in the presence of protease inhibitors (to avoid any cleavage of MHC complexes) at a stable pH value (usually between pH 7.0 and 8.0). After washing (by centrifugation and filtration) the lysate is passed over MHC-specific monoclonal antibodies bound to sepharose beads. The beads are then washed to remove excess detergents, and MHC complexes are released from the antibodies by acid treatment. The peptides can be separated from the proteins by ultrafiltration, and the flowthrough is usually lyophilized before fractionation and sequence analysis. Though the technique provides highly pure isolates, it suffers disadvantages such as high cost due to the requirement of large amounts of antibodies (10–30 mg per isolation), complexity of the protocol, and inability to distinguish intracellular and extracellular MHC complexes [32]. Recent studies that tried to include desalting and inclusion of specific ions in desalting buffer have been shown to enhance peptide yield [33].

The acid elution technique is based on the release of MHC–peptide complexes from the cell surface by short acid treatment at pH 3.3 [34]. The source materials for this technique are cells from dissociated tissue or adherent or suspension cell cultures. Having intact cells is a prerequisite for the technique since cell damage will lead to the release of proteases generating peptide fragments from highly abundant cell proteins. The major advantage of the technique is that it is cost effective, simple, and it could differentiate intracellular and extracellular MHC complexes. This method has been successfully employed to identify T-cell epitopes from melanoma cells [35], as well as from the bcr–abl fusion protein expressed at the cell surface [36].

HPLC fractionation can be performed prior to tandem mass spectrometry, which in combination allows high-resolution separation and sequencing of single peptides from complex samples [37]. Mass spectrometry (MS) is based on precise determination of molecular masses of analyte molecules. Following determination of the molecular mass of the analyte by various means (depending on MS instrumentation), the peptide sequence can be derived by fragmentation of the analyte ion. Hence MSMS analysis allows the detection of a single peptide from a complex mixture of peptide pools. Figure 4.3 illustrates an example of MSMS spectrum of peptides.