Anti-idiotypic antibodies bind to unique regions on other antibody molecules. One of the central hypotheses of modern immunology is that the immune system is regulated through a network of antibody-anti-idiotypic antibody interactions. One consequence of this theory is the idea that a specific B- or T-cell clone can be specifically activated by the appropriate anti-idiotypic antibody. In this way, antiidiotypic antibodies may be used as vaccines to induce active immunity against a foreign antigen. Using animal models, it has been possible to demonstrate that anti-idiotypic antibody vaccines can induce specific protective immunity against a variety of infectious agents and tumor antigens. Based on these types of studies, anti-idiotypic vaccines are being tested in patient trials in an attempt to induce specific immunity against human cancers.
The last two decades have seen a frenzy of activity in the search for tumor-specific antigens. With few exceptions, all tumor antigens defined to date are differentiation antigens, molecules expressed on selected normal cell populations during differentiation, or molecules found on several types of tumors as well as normal testes. These antigens may be abundantly expressed by certain tumor cells but they are also found on some normal tissues, although perhaps at a much lower level of expression. This difficulty in identifying tumor-restricted antigens has led to two approaches toward immunizing patients against tumors.
The first approach uses whole cell or cell extract vaccines consisting of mixtures of undefined antigens with the theory being that, even if we cannot identify them, tumor-specific antigens probably exist and can be recognized by the immune system if administered correctly. The second approach attempts to immunize patients against defined tumor antigens that have been identified as potential targets due to the presence of either naturally occurring antibodies or T lymphocytes that recognize these molecules. Among these immune targets are both carbohydrate antigens (e.g., gangliosides, TF, Tn) and proteins. The majority of antigens identified so far are autoantigens, i.e., they are nonmutated molecules present on certain normal cells, as well as on tumor cells. As a result, the host is generally tolerant toward these antigens and it is often difficult to induce a specific immune response. Several strategies designed to break tolerance against these antigens are being pursued, including chemical modification of the antigen, addition of carrier proteins with T-cell epitopes, and addition of potent immune adjuvants.
This article focuses on a different strategy to immunize against defined tumor antigens--the use of anti-idiotypic antibody vaccines. Anti-idiotypic antibodies can be used as surrogate antigens and can induce immune responses against the native antigen. Further, anti-idiotypic antibodies offer certain potential advantages over vaccines using native antigen.
II. WHAT ARE ANTI-IDIOTYPIC ANTIBODIES?
The basic antibody (immunoglobulin) structure consists of two heavy chains and two light chains covalently linked by disulfide bonds (Fig. 1). Each heavy and light chain is made up of a variable domain and one (light chain) or three (heavy chains) constant domains. Within a given species, the amino acid sequences of the constant domains are largely conserved between antibodies of the same isotype and are not directly involved in antibody binding to antigen. In contrast, three-dimensional interactions of the variable domains of the heavy and light chains form the antigen-binding site of the antibody.
FIGURE 1 Structure of an IgG antibody molecule. The IgG molecule consists of two heavy chains and two light chains covalently linked through disulfide bonds. The variable domains (shaded), which contain regions unique to the individual heavy and light chains, form the antigen-binding site of the antibody. The constant domains (open) determine antibody functions such as complement fixation. From Chapman, P. B., and Houghton, A. N. (1992). Anti-idiotypic vaccines. Biol. Ther. Cancer Updates 2(5).
Although a detailed description of the basis of antibody diversity is beyond the scope of this article, it is useful to remember that the DNA encoding the variable domain of a heavy chain (VH) is assembled from three genomic segments; variable (V), diversity (D), and joining (J). For human heavy chains, it is estimated that there are 100-200 V segments, at least 4 D segments, and at least 6 J segments. DNA encoding light chains is assembled from a large number of V and 5 J segments (there is no D segment for light chains). In addition to this recombinatorial diversity for both heavy and light chains, somatic hypermutation also occurs within the VH and VL DNA.
It has been estimated that, in mice, this diversity would generate approximately 2.7 × 108 different antibody molecules. Thus, the potential number of unique antibody molecules is enormous. Within each VH and VL domain are three hypervariable regions, also called complementarity-determining regions (CDRs). Association of the heavy and light chains in three dimensions forms the antigen-binding site, and the amino acid residues of the CDRs are usually critical in forming hydrogen bonds, van der Waal forces, and salt bridges with which the antibody binds antigen.
Each mature B cell makes an antibody with a unique antigen-binding specificity determined by a unique set of VH and VL domains. This means that each antibody contains unique regions within its variable domains. These unique regions are termed "idiotopes" and the set of idiotopes expressed by a given antibody defines the "idiotype" of the antibody. Because idiotopes are unique epitopes, it is possible to raise antibodies against idiotopes; in fact, idiotopes are defined by the antibodies raised against them. These antibodies are "anti-idiotypic" antibodies because they bind to an idiotope expressed by another antibody molecule. Anti-idiotypic antibodies can recognize idiotopes expressed entirely on the variable domain of either the heavy or the light chain, or anti-idiotypic antibodies can recognize idiotopes defined by amino acids from both heavy and light chain variable domains (Fig. 2). Some anti-idiotypic antibodies bind to the actual antigen-binding site of the target antibody, as discussed later.
FIGURE 2 Interactions between an antibody and an anti-idiotypic antibody. An antibody (Ab1) recognizing an antigen (solid triangle) is itself recognized by Ab2 antibodies. Because the antigen-binding sites of antibodies are also unique epitopes (i.e., idiotopes), the Ab1-Ab2 relationship is reciprocal, i.e., the Ab1 is also an antiidiotypic antibody with relation to the Ab2. From Chapman, P. B., and Houghton, A. N. (1992). Anti-idiotypic vaccines. Biol. Ther. Cancer Updates 2(5).
III. MODULATION OF THE IMMUNE RESPONSE BY ANTI-IDIOTYPIC ANTIBODIES
A. The Jerne Hypothesis
Anti-idiotypic antibodies were initially described in 1963 by Henry Kunkel and Jacques Oudin working independently. They observed that an animal immunized with an antigen produced an antibody response (Ab1). If this Ab1 were isolated and injected into a second, naive animal, antibodies (Ab2) were produced that bound specifically to the Ab1. These Ab2 antibodies were termed anti-idiotypic antibodies because they recognized an epitope on the Ab1 that was unique to the Ab1. Oudin also observed that the antigenic specificity of the Ab1 tended to correlate with the idiotopes expressed by the Ab1. In other words, different Ab1 antibodies recognizing the same antigen tended to express the same idiotopes. Antiidiotypic antibodies were defined as antibodies with specificity for an Ab1, or group of Ab1 antibodies, elicited by a specific antigen. Later, Alfred Nisonoff and others described two types of Ab2 antibodies-- Ab2α and Ab2β--based on whether the Ab2 could bind to Ab1 in the presence of antigen. Ab2β antiidiotypic antibodies do not bind to Ab1 in the presence of excess antigen, presumably because Ab2β antibodies bind to the antigen-binding site of the Ab1. Ab2α anti-idiotypic antibodies, however, bind to idiotopes distinct from the antigen-binding site of the Ab1.
In 1974, Niels Jerne hypothesized that the immune system (both B and T cells) was regulated by a network of idiotype-anti-idiotype interactions. Jerne envisioned that for every antibody, there were corresponding Ab2α and Ab2β anti-idiotypic antibodies encoded by the genome (Fig. 3). Binding of soluble Ab2α to Ab1 on the surface of a B cell resulted in inhibition of the B-cell clone, whereas binding by soluble Ab2β resulted in stimulation. A curious aspect of idiotype--anti-idiotype interactions as envisioned by Jerne is that every Ab1-Ab2α interaction is reciprocally an Ab1-Ab2β interaction. For example, in Fig. 3, if the Ab2α antibody were considered the Ab1, then the antibody labeled Ab1 (solid) becomes an Ab2β because its idiotope (solid rectangle) is recognized by the antigen-binding site of the other antibody.
Jerne proposed that introduction of a foreign antigen resulted in an initial elimination of circulating Ab1. This would release the inhibitory effect on Ab2β B cells, leading to the production of Ab2β and stimulation of the Ab1 B cell. Elimination of Ab1 by the foreign antigen would also eliminate the stimulatory effect of Ab1 on B cells, making Ab2α antiidiotypic antibodies, resulting in less of the inhibitory Ab2α antibodies in the circulation. Therefore, the overall effect would be more Ab1 production. As the antigen is cleared, the network would return to its previous homeostatic steady state. Through this interconnecting, dynamic network of idiotypic-antiidiotypic interactions, the immune system was hypothesized to be self-regulating. It is important to remember that Jerne hypothesized that this idiotype network would also regulate T-cell reactivities.
FIGURE 3 Possible mechanism of immune regulation through a network of idiotype-anti-idiotype interactions. A B cell expresses surface antibody (Ab1) with specificity for a particular antigen. This Ab1 also expresses an idiotope outside of its antigen-binding region (depicted as rectangles). An anti-idiotypic antibody (Ab2α) can recognize this idiotope, and the interaction results in the suppression of the Ab1 B-cell clone. At the same time, this interaction also occurs between soluble Ab1 and cell surface Ab2α (not shown), resulting in the activation of the Ab2α B cell and further secretion of the Ab2α antibody. The Ab2β anti-idiotypic antibody expresses an idiotope that binds to the antigen-binding domain of Ab1 and mimics the antigen. Interactions between soluble Ab2β and cell surface Ab1 result in activation of the Ab1 B cell and enhanced secretion of the Ab1 antibody. Reciprocally, Ab1 binds to an idiotope.
A special feature of the Ab2β anti-idiotypic antibodies was that they could functionally mimic the foreign antigen. This was first suggested by Alfred Nisonoff and Edmundo Lamoyi and independently by Ivan Roitt. Because of this, it should be possible to induce immunity against a foreign antigen by using an appropriate anti-idiotypic antibody as a surrogate immunogen. This was confirmed in the early 1980s when several investigators demonstrated that it was possible to induce specific and protective immunity in rodents against a variety of infectious agents using antiidiotypic monoclonal antibodies (MAb). Not only could antibodies be induced against protein antigens, but anti-idiotypic antibodies mimicking nonprotein antigens were effective immunogens. As predicted by the Jerne hypothesis, it was possible to demonstrate the induction of T-cell immunity as well as humoral immunity.
B. Structural Basis for Antigen Mimicry
There are at least three potential mechanisms by which anti-idiotypic antibodies can induce antibody against a foreign antigen. Some anti-idiotypic antibodies appear to reproduce the antigenic epitope within one of its CDRs. Anti-idiotypic antibodies that induced antibodies against the GAT antigen (a random synthetic terpolymer made up of glutamic acid60 alanine30 tyrosine10) contain "GAT-like" sequences within the VH CDR3 domains (corresponding to the D segment at the DNA level). These antibodies contained either Glu-Glu-Tyr or Tyr-Tyr-Glu sequences, and it has been hypothesized that this is the basis for their ability to mimic the GAT antigen.
In the hepatitis B surface antigen (HBsAg) system, an anti-idiotypic MAb, designated 2F10, can induce antibodies against HBsAg and can prime HBsAgspecific helper T cells (TH) in mice. A 15 amino acid sequence within the VH CDR3 was identified that was homologous with a region on HBsAg representing a common immunogenic determinant. In fact, CD4 T cells from patients immunized against hepatitis B could be stimulated using only this 15 amino acid peptide. Mark Greene's laboratory showed that a 16 amino acid epitope of reovirus type 3 hemagglutinin is mimicked by an anti-idiotypic MAb. CDR2 regions from both VH and VL contribute to form a homologous epitope. These examples demonstrate one mechanism of antigenic mimicry by antiidiotypic MAb, which is expression of the original antigenic epitope within the variable domains of the anti-idiotypic MAb. This is not the only mechanism of antigenic mimicry and certainly cannot explain antiidiotypic antibody mimicry of nonprotein antigens.
Some anti-idiotypic antibodies may recapitulate the conformation and charge of the antigen sufficiently to activate the appropriate B-cell clone. Perhaps the best studied example is MAb E225, an antiidiotypic MAb that mimics an epitope on lysozyme recognized by the Ab1 MAb D1.3. Bentley and colleagues at the Institut Pasteur were able to determine the crystal structure of E225-D1.3 complexes and compare how D1.3 binds to E225 with how D1.3 binds to lysozyme. This analysis showed that E225 does not form an image of the lysozyme epitope on the atomic level. Thirteen amino acids within the D1.3 variable domains directly contacted E225. Of these 13 amino acids, only 7 also served as contact residues in D1.3 binding to lysozyme. Further, these 7 amino acids interacted with E225 differently than with lysozyme (Table I). These data suggest that an anti-idiotypic MAb does not need to mimic the antigen at the atomic level in order to induce antibodies that cross-react with the antigen. This may be how an anti-idiotypic MAb (which is a protein) can mimic a nonprotein antigen such as a carbohydrate.
TABLE I Atomic Interactions Involving Amino Acid Residues of MAb D1.3 That Contact Both Antigen (Lysozyme) and Anti-idiotypic MAb (E225)aAmino acid within the D1.3 idiotope Intactions with lysozyme Interactions with anti-idiotypic MAb E225 Tyr32 van der Waal H bond VL Tyr50 H bond H bond Trp92 van der Waal H bond Trp52 van der Waal van der Wall VH Asp54 H bond One salt bridge, one H bond Asp100 Five H bonds van der Waal Tyr101 H bond van der Waal
aSeven amino acid residues within the variable domains of MAb D1.3 are involved in binding to both antigen (lysozyme) and anti-idiotype MAb E225. For five of the seven amino acids, the nature of the atomic interaction with antigen differs from the interactions with anti-idiotypic MAb. Adapted from Bentley, G.A., Boulot G., Riottot, M. M., and Poljak, R.J. (1990) Three-dimensional structure of an idiotope-anti-idiotope complex. Nature 348; 254-257.
Although not entirely consistent with Jerne's first approximation of his network theory, an antiidiotypic antibody might be able to activate a specific B-cell clone without providing a true internal image of the antigen. One example is an Ab2 antibody raised by Oosterlaken and colleagues in the Netherlands against an Ab1 that neutralizes Semliki forest virus (SFV). Although this Ab2 could immunize BALB/c mice, the same mouse strain from which the Ab2 was made, mice from other strains did not respond. This suggests that the Ab2 does not provide a pure internal image of the SFV antigen, as internal image antiidiotypic antibodies should immunize across strain and species barriers. It is conceivable that this Ab2 activates a B-cell clone in BALB/c mice that produces an antibody with significant affinity for SFV without providing a true internal image.
Despite these complexities, it is apparent that antiidiotypic antibodies can induce immunity against protein and nonprotein antigens and, in certain circumstances, can induce T-cell immunity. Further, studies in infectious disease animal models have shown that this immunity can be protective. These beneficial effects are not limited to rodents; Ronald Kennedy and colleagues have demonstrated that an anti-idiotypic antibody vaccine can be used to induce protective immunity against hepatitis B in chimpanzees. Because of these encouraging results, investigators have applied the concept of anti-idiotypic vaccines to the problem of cancer.
IV. USE OF ANTI-IDIOTYPIC ANTIBODIES AS TUMOR VACCINES
A. Potential Advantages of Anti-idiotypic Antibody Tumor Vaccines
A practical question arises: Why use an anti-idiotypic MAb vaccine? Given that they can be quite difficult to generate, clone, and characterize, it is reasonable to consider that there are at least two potential advantages of an anti-idiotypic MAb vaccine. The first is that the antigen of interest may be difficult to obtain in sufficient quantity or may be dangerous to handle (infectious, tumorigenic). An anti-idiotypic MAb mimicking the antigen circumvents these problems.
The second advantage is that anti-idiotypic MAb can be more effective immunogens than some native antigens. In some cases, it has been possible to demonstrate that anti-idiotypic MAb vaccines can break immunological tolerance against the native antigen. One example, reported by Stein and Soderstrom, used neonatal mice incapable of mounting an immune response to bacterial capsular polysaccharide. If mice were primed with anti-idiotypic antibody, they were able to respond to antigen and develop protective immunity. Because most tumor antigens are differentiation antigens and are expressed on selected normal tissues, the host is usually tolerant to these molecules. Therefore, the ability to break tolerance may be an important feature of anti-idiotypic MAb vaccines.
Anti-idiotypic MAbs offer a similar advantage in immunizing against nonprotein tumor antigens. B-cell responses to nonprotein antigens (T-cell independent) are generally low-titer, transient IgM responses. Anti-idiotypic MAbs that mimic nonprotein antigens represent a xenogeneic protein mimic that can be more immunogenic than the original nonprotein antigen. Presumably, one of the ways this could occur is that the anti-idiotypic MAb can provide TH cell epitopes, which the B cell can present to TH cells through its MHC class II molecules. Indeed, in the reovirus system, the VH domain of an anti-idiotypic MAb was demonstrated to contain a TH determinant. As noted earlier, evidence from animal studies shows that anti-idiotypic MAb mimicking carbohydrate bacterial antigens can be more effective than the antigens themselves in inducing immunity.
These theoretical and actual advantages of antiidiotypic antibody vaccines have provided a rationale for developing antitumor vaccines using antiidiotypic antibodies.
B. Anti-idiotypic Monoclonal Antibodies Can Be Used to Immunize Animals against Tumors
Anti-idiotypic antibodies have been used to induce immune responses in rodents against a variety of antigens expressed by tumor cells. In all cases, the antitumor immune response was characterized by the induction of antibodies against the tumor antigen. In a few cases, it was possible to demonstrate a delayedtype hypersensitivity (DTH) reaction against tumor cells, suggesting recruitment of TH cells, although more detailed T-cell studies have not been reported. Only a very few studies have assessed whether the antitumor immune response induced by the antiidiotypic antibody provided any protective effect; in the few studies reporting tumor protection, the effect was incomplete.
Studies using anti-idiotypic MAb raised against Ab1 recognizing nonprotein tumor antigens have confirmed previous observations in other systems that immune responses against nonprotein antigens can be induced by anti-idiotypic MAb. Tumor responses have been induced against both carbohydrate and ganglioside (acidic glycolipids) tumor antigens. This is of particular importance for several reasons. First, it is clear that carbohydrates and gangliosides represent potentially important antigenic targets for the immunotherapy of tumors. Second, for nonprotein antigens, the anti-idiotypic vaccine approach holds special potential advantages over immunization with native antigen, as discussed earlier.
C. Anti-idiotypic Antibody Vaccines in Cancer Trials
Several clinical trials have been completed in which anti-idiotypic MAbs were used to immunize cancer patients against tumor antigens (Table II). In these early trials, the primary goal was to see if immunization resulted in a detectable immune response against the antigen expressed by the tumor cell. As a result, investigators have explored different doses and schedules of anti-idiotypic MAb, as well as trying to increase their immunogenicity by adding immune adjuvants. Triggering antitumor effects has generally not been an end point of these early trials and, in fact, many of the trials (although not all) were conducted in patients who were free of disease after surgery or chemotherapy but who were at high risk for recurrence.
TABLE II Clinical Trails Using Anti-idiotypic Antibodies to Immunize Patients against Tumor AntigensaAntigen Tumor Anti-idiotypic vaccines Results HMW-MAA Melanoma MF2-23 conjugated to KLH, BCG adjuvant; MF11-30; Melimmune-1 and -2 14/23 patients immunized with MF2-23 vaccine developed antibodies to antigen-positive melanoma. Three partial tumor responses. No anti-HMW-MAA responses induced by the other anti-id vaccines GD3 Melanoma, SCLC BEC2 + BCG adjuvant In a series of trials, 20-30% of patients developed detectable anti-GD3 antibodies. In SCLC after a major response to chemotherapy, vaccination associated with prolonged survival compared to historical experience GD2 Melanoma, stage IV 1A7 + QS21 adjuvant Anti-GD2 and antibodies detectable in 40/47 patients. One CR noted CEA Colon carcinoma (advanced), lung adenocarcinoma 3H1 + alum adjuvant Approximately 50% of patients developed detectable anti-CEA antibodies. Also, 50% developed proliferative T-cell responses against CEA. No clinical responses observed 791 Typ72 Colon carcinoma (advanced) 105AD7 After immunization with 105AD7, it was possible to detect enhanced IL-2 release in response to antigenexpressing cells and cytotoxicity against autologous tumor cells. Anti-791Typ72 antibodies have not been detected. No clinical responses observed gp37 T-cell lymphoma One of four patients developed anti-gp37 antibodies. One patient had a dramatic clinical response lasting 11 months CA125 Ovarian carcinoma ACA125 9/16 patients developed anti-CA125 antibodies. 9/16 also developed CA125-specific PBMC responses GA733-2 Colon Anti-idiotypic MAb against MAb 17-1A Induced a proliferative T-cell response against antigen. Three different peptides recognized.
aHMW-MAA, high molecular weight melanoma-associated antigen; PBMC, peripheral blood mononuclear cells; KLH, keyhole limpet hemocyanin; BCG, bacille Calmette-Guérin; HMFG, human milk fat globulin; CEA, carcinoembryonic antigen; SCLC, small cell lung carcinoma; CR, complete response.
Monoclonal anti-idiotypic antibodies were first used to immunize melanoma patients against the high molecular weight proteoglycan-melanoma-associated antigen (HMW-MAA). Several anti-idiotypic MAbs have been produced against HMW-MAA, but only one particular one so far--MK2-23--appears to induce antibodies against the HMW-MAA itself. Mittelman immunized patients with MK2-23 conjugated to KLH as a carrier protein and mixed with BCG adjuvant and found that 14/23 patients developed antibodies to HMW-MAA and 3 patients actually had partial shrinkage of their tumors.
Anti-idiotypic MAb vaccines are also being used to immunize cancer patients against nonprotein tumor antigens. 1A7 is an anti-idiotypic MAb that mimics the ganglioside GD2. In stage IV melanoma patients immunized with 1A7 + alum, it was possible to detect anti-GD2 antibodies in 40/47 patients, although it required that the sera be partially purified and concentrated. One of the patients had complete shrinkage of the melanoma tumors as a result of immunization.
BEC2, another example of an anti-idiotypic mouse MAb that mimics a ganglioside antigen, in this case GD3, is being used to immunize both melanoma and small cell lung cancer (SCLC) patients. Immunizing with BEC2 mixed with BCG, it is possible to detect the induction of anti-GD3 antibodies in the serum of 20-30% of patients. An intriguing observation was made in a pilot trial in which BEC2 was used to immunize patients with SCLC who had had a partial or complete response to chemotherapy. Among the seven patients with limited stage disease, only one patient relapsed with SCLC. This has led to a multicenter randomized phase III trial testing BEC2 in limited stage SCLC.
Another phase III randomized trial conducted to test an anti-idiotypic MAb vaccine was conducted in advanced colon carcinoma patients to test 105AD7, a human anti-idiotypic MAb that mimics a protein designated 791Typ72, also known as decayaccelerating factor. Initial nonrandomized results were promising, but the randomized trial did not show a survival advantage in patients receiving 105AD7. The trial design may have been too stringent to detect a modest effect of the vaccine, however.
These studies support previous animal studies that showed that anti-idiotypic MAb can be used to immunize against both protein and nonprotein antigens. Also, it appears that the use of immune adjuvants (such as BCG) will be needed to optimize immunogenicity. To date, significant toxicity has not been observed in patients immunized with anti-idiotypic MAb. One randomized phase III trial is underway and others are planned in order to test whether these vaccines can make an impact on the natural history of cancer.
V. FUTURE DIRECTIONS OF ANTI-IDIOTYPIC ANTIBODY CANCER VACCINES
There have now been several pilot trials and one phase III trial testing anti-idiotypic antibody vaccines in cancer patients. The studies so far have shown that it is possible to induce antibodies against specific selfantigens, although the antibody titers are low. The success of anti-idiotypic vaccine therapy depends not only on the identification of appropriate tumor antigen targets, but also on whether relevant titers of antibodies can be induced. It is not surprising that it is more difficult to induce an immune response against an autoantigen, such as a tumor antigen, than against a foreign antigen, such as a viral antigen. For this reason, parameters such as vaccine dose, route, schedule, and use of immune adjuvants will be critical for optimization of anti-idiotypic antibody vaccines. It is also possible that modification of the anti-idiotypic antibody molecule can improve its immunogenicity. Such modifications may include removal of the Fc region by creating Fab or F(ab)2 fragments, chimerization of the molecule to replacing the mouse constant regions with homologous human regions, and attaching carrier molecules with known T-cell epitopes. Once optimized, it could turn out that anti-idiotypic antibody vaccines are best used to prime patients for an immune response against the native antigen.
One question that arises is: Do anti-idiotypic MAb vaccines offer any real advantage over immunizing with antigen? With regard to protein antigens, the answer may be "no." As we come to a better understanding of how to activate T cells against tumor antigens, and as we explore newer immunization techniques, such as peptide constructs or DNA immunization, it seems that immunizing with some form of the antigen itself will be more likely to result in an effective immune response than immunizing with an anti-idiotypic MAb. However, the situation may be quite different in the case of a nonprotein antigen. These antigens also tend to be poorly immunogenic but we cannot easily manipulate these molecules to circumvent this poor immunogenicity the way we can manipulate proteins. In this case, anti-idiotypic MAb vaccines represent a strategy to induce an immune response against a poorly immunogenic antigen.
Once an anti-idiotypic MAb vaccine is shown to induce a specific immune response against an appropriate tumor antigen convincingly and reproducibly, it is necessary to test the real question in a welldesigned phase III trial: Can an effective antitumor immune response be induced in patients?
Paul B. Chapman
Memorial Sloan-Kettering Cancer Center
ANTIBODIES IN THE GENE THERAPY OF CANCER ; CANCER VACCINES: GENE THERAPY AND DENDRITIC CELL BASED VACCINES ; CANCER VACCINES: PEPTIDE-AND PROTEIN-BASED VACCINES ; CARBOHYDRATE-BASED VACCINES ; DNABASED CANCER VACCINES ; TARGETED TOXINS ; TUMOR ANTIGENS
adjuvant A substance used to boost an immune response against a specific antigen.
complementarity-determining regions (CDRs) Hypervariable regions within the heavy and light chain variable domains of an immunoglobulin molecule. CDRs are generally involved in contacting antigen and so are critical in determining the specificity of an antibody.
differentiation antigen An antigen expressed on a normal cell as part of its differentiation program. The antigen may be lineage specific and may be expressed only at a certain stage of cellular differentiation.
epitope The region on an antigen molecule recognized by a specific antibody or T-cell clone.
idiotope A unique epitope on an immunoglobulin molecule. By definition, idiotopes lie within the variable domains of the immunoglobulin molecule.
idiotype The set of idiotopes expressed by a given immunoglobulin molecule.
isotype Refers to the type of heavy chain in an antibody molecule. There are nine antibody isotypes identified in humans (IgM, IgD, IgG1, IgG2, IgG3, IgG4, IgA1, IgA2, IgE), each of which is encoded by a separate gene. The isotype can be identified serologically, as isotype-specific antibodies can be raised.
tolerance A state of immunological unresponsiveness toward a given antigen.
Bhattacharya-Chatterjee, M., and Foon, K. A. (1998). Antiidiotype antibody vaccine therapies of cancer. Cancer Treat. Res. 94, 51-68.
Chapman, P. B., and Houghton, A. N. (1991). Induction of IgG antibodies against GD3 in rabbits by an anti-idiotypic monoclonal antibody. J. Clin. Invest. 88, 186-192.
Foon, K. A., John, W. J., Chakraborty, M., et al. (1997). Clinical and immune responses in advanced colorectal cancer patients treated with anti-idiotype monoclonal antibody vaccine that mimics the carcinoembryonic antigen. Clin. Cancer Res. 3, 1267-1276.
Foon, K. A., Lutzky, J., Baral, R. N., et al. (2000). Clinical and immune responses in advanced melanoma patients immunized with an anti-idiotype antibody mimicking disialoganglioside GD2. J. Clin. Oncol. 18, 376-384.
Grant, S. C., Kris, M. G., Houghton, A. N., and Chapman, P. B. (1999). Long survival of patients with small cell lung cancer after adjuvant treatment with the anti-idiotypic antibody BEC2 plus BCG. Clin. Cancer Res. 5, 1319-1324.
Greenspan, N. S., and Bona, C. A. (1993). Idiotypes: Structure and immunogenicity. FASEB J. 7, 437-444.
Jerne, N. K. (1974). Towards a network theory of the immune system. Ann. Immunol. (Inst. Pasteur) 125, 373-389.
Kennedy, R. C., Melnick, J. L., and Dreesman, G. R. (1986) Anti-idiotypies and immunity. Sci. Am. 255, 48-56.