The antibody-based therapies we have been discussing are a form of passive immunotherapy. That is, the molecules or substances are introduced into the body, rather than the body creating its own immune response. Vaccines, on the other hand, are considered active immunotherapy because they generate an intrinsic immune response. They are also considered a form of specific immunotherapy because they attempt to stimulate an immune response that can directly target the tumor antigens, in contrast to non-specific approaches such as cytokines that broadly stimulate the immune system. Video Download Free RealPlayer
Efforts to treat cancer with vaccines date back to the origins of immunology.1 Patients have been injected with autologous and allogeneic malignant cells, usually irradiated to prevent further growth. However, measuring immune response was problematic. Now that we have identified several tumor antigens and the immune response they provoke, we have made progress in developing cancer vaccines.
As Dr. Disis explained, tumor cells express specific antigens on the cell surface, usually within the MHC molecules. The problem with tumor cells is that they cannot stimulate a T cell response by na´ve T cells, in part because they lack necessary co-stimulatory molecules. However, dendritic cells – a type of antigen-presenting cell – can provide them. [Figure 9. Processing and Presentation of Class I Antigens] We can take a dendritic cell and import tumor antigens into it by a variety of mechanisms. The dendritic cell can then present the tumor antigens on their surface within MHC molecules, ready to activate T cells. Once the T cells are activated, they are capable of recognizing and destroying antigen-expressing tumors.
Peptide, autologous, and viral vector vaccines
Numerous approaches to stimulate the immune system to recognize tumors have been tried over the years. Vaccines consisting of peptide or protein administered with an adjuvant have been the most frequently used. These adjuvants might be compounds such as bacterial cell wall components that incite an inflammatory response or cytokines such as IL-12 or GM-CSF, as Dr. Weber described. Monocytes, neutrophils, eosinophils, and T cells are all recruited to the site of the inflammation where adjuvant is used, but it is believed that in situ dendritic cells ultimately take up the proteins and peptides, process them if necessary, and present them on the cell surface as peptides capable of binding to the MHC molecules. Dendritic cells can then stimulate T cells that have the receptors to recognize those particular peptides. Video
Table 2 [Table 2. Peptide and Protein Vaccines] summarizes some of the studies published in the past year or two using peptide or protein vaccines.62-67 These studies are primarily in melanoma. The targeted antigens are gp100, tyrosinase, and MAGE, which are found in melanoma. Dr. Disis' group worked with HER-2/neu peptides. Lymphomas may be targeted by using idiotype protein. Adjuvants include IL-12, QS-21, and GM-CSF. Immunologic response was monitored by various assays—ELISPOT and microcytotoxicity in particular, but also by delayed-type hypersensitivity (DTH) skin testing. All of these studies report at least a few patients developing some type of antigen-specific immune response, to varying degrees. This response will certainly depend on the antigen, the tumor system, the patient, and disease stage. As we look across Table 2 to clinical response, we see they have been modest, although there are some instances of complete or partial remissions, or prolongation of survival.
An interesting approach to detecting a clinical response was used in lymphoma patients, in whom circulating tumor cells could be detected by polymerase chain reaction (PCR) prior to immunization.67 PCR enables us to actually search for lymphoma cells by identifying the particular chromosomal translocation abnormality they are known to possess. In this study, tumor cells could not be detected in peripheral blood of some patients following immunization. This method of measuring clinical responses is becoming increasingly important, particularly for diseases that cannot be monitored by a CT scan.
Tumor cell-based vaccines are another vital area of research. Some studies have used autologous tumor, in which tumor cells are extracted from surgical resection or biopsy specimens. Allogeneic cell lines have also been developed for tumors such as melanoma that likely encompass many of the tumor-associated antigens expressed by the melanomas of most affected individuals. Tumor cells can also be modified to make them more immunogenic. To accomplish that, tumor cells may be infected with various types of viruses so that viral proteins are expressed on the surface, and transduced with genes expressing cytokines such as IL-2 and GM-CSF, or genes for HLA molecules or co-stimulatory molecules. The idea is irradiate the cells so they can no longer proliferate, then inject the tumor cells back into the patient. What we hope to see is the immune system activated by either the tumor cell or the inflammatory response that includes recruitment of dendritic cells. As the injected tumor cells undergo apoptosis or are destroyed by the inflammatory reaction, antigens are picked up by the dendritic cells and represented to the T cells.
Table 3 [Table 3. Tumor Cell Vaccines] illustrates results of some of the current research on tumor vaccines.68-72 These studies employed several strategies. One of the most popular is to transduce with a vector containing GM-CSF, so that the tumor secretes GM-CSF and sets up an inflammatory response.68,69 In animal studies, this strategy has been the most promising in inducing a protective immune response. Newcastle virus is used to infect the cells in another study72 to some effect, as shown by median survival of 46 weeks. Use of Bacille Calmette-Guerin (BCG) as an inflammation inducing adjuvant along with autologous colorectal cancer cells71 showed increased DTH but no survival benefit. In fact, most of these tumor cell approaches show an immune response, but again limited clinical response. Nonetheless, CancerVax70, which has been heavily tested in melanoma, has been suggested in nonrandomized studies to provide a survival benefit.
Viral vectors and indeed, naked DNA in the form of plasmids encoding tumor antigens, can be used to immunize people. Initially these vaccines were administered to muscle cells, but it is likely that dendritic cells were ultimately the targets infected by the virus, or were picking up the antigen released by apoptotic muscle cells.73-76 [Table 4. Viral Vector and Plasmid Vaccines] Poxviruses are another popular way to apply this approach, and a considerable amount of work has been done with vaccinia and avian and fowl pox vectors. One of the most interesting strategies is the prime boost approach. An example of this is when a patient is first immunized with vaccinia virus encoding the gene for CEA. Vaccinia is very immunogenic, so it can only be used once or twice – after the first injection patients develop high neutralizing antibody titers. In subsequent immunizations the antibody immediately binds to the virus, making it difficult to get true immunization against the encoded tumor antigen. However if you follow with an avian vector expressing CEA, we see better immunologic responses.74 Clinical responses have remained limited, but various modifications of these strategies continue to be explored.
Dendritic cell vaccines
Dendritic cells (DC), as mentioned earlier, are in the final common pathway for activating na´ve T cells by many of these vaccine strategies. Remember these are the professional antigen-presenting cells, and can be found in most areas of the body. They circulate in peripheral blood, and are present as Langerhans cells in the epidermis. Video
Variables associated with employing dendritic cell vaccines are numerous. First we must consider what is the best source or lineage of dendritic cell to use. Should we use DC directly isolated from the peripheral blood or generate them ex vivo from precursors? Next, how do we load the antigen? Maturation and/or activation is another factor to consider, as data suggest immature dendritic cells may give a diminished immune response. Route of administration is always a question with tumor vaccines, as there are advantages and disadvantages to all of the available routes. Perhaps most importantly, we need to understand how best to evaluate the immune and clinical response to dendritic cell vaccines to permit efficient development of this strategy.
The clinical utility of dendritic cell vaccines was initially limited by the fact that they are not numerous, and they are difficult to obtain. It is now possible to obtain them in large numbers by generating them in vitro. We can directly isolate them from peripheral blood, although that is still difficult because there are so few unless they are mobilized by cytokines such as Flt3-ligand. CD34+ cells from bone marrow or peripheral blood can be stimulated with a cocktail of cytokines to generate dendritic cells. Monocytes in the peripheral blood, either adherent to plastic or selected for CD14, and cultured in media containing GM-CSF and IL-4, or GM-CSF and IL-13, can generate dendritic cells as well. We can also convert non-dendritic cells to dendritic cells. In this process, cells from patients with chronic myelogenous leukemia, for example, may be differentiated into dendritic cells. These cells contain the bcr-abl translocation and thus present bcr-abl peptides on the surface of the cell.
However they are generated, it is important that dendritic cells are mature. Mature cells have upregulated MHC molecules and certain chemokines and chemokine receptors, and downregulated others. The changes in the chemokine receptors allow the dendritic cells to migrate from their peripheral site into the lymph nodes. Mature dendritic cells produce IL-12, essential not just for attraction of NK cells but also cytolytic T cells, and for shifting helper T cell responses in a Th1 direction. We have shown in vitro that matured dendritic cells, loaded with antigens, seem to stimulate greater cytolytic T cell activity against tumor antigens – CEA in this case.77 [Figure 10. Maturation of DC: Effects on Phenotype and Antigen Loading and Processing] The sequence of loading and maturation is important as well. For example, if we use a protein or messenger RNA to load a dendritic cell, this processing is something only immature cells are good at so the cells should be matured after they are loaded. If we load with peptide, which requires no processing, we mature first and then load to optimize the number of MHC molecules on the surface.
IL-12 is a good indicator of functional maturity, although many of the maturation strategies do not actually result in IL-12 production. For example, maturation with CD40 ligand alone results in CD80 and CD83 upregulation, but only small amounts of IL-12.78 What we have shown to be most effective for maturation is either CD40 ligand plus interferon gamma or adding lipopolysaccharide (LPS).
The route of administration of dendritic cells continues to be controversial. Intradermal, subcutaneous, or in some cases intralymphatic are some of the more commonly used methods. We looked at the ability of dendritic cells to migrate to areas with high lymphocyte concentrations – after all, to boost immunity, the dendritic cells must meet the T cells somewhere. We labeled the dendritic cells with indium-111 so we could measure them with a gamma camera. When we injected the cells intravenously, they migrated to the lungs first and then redistributed to liver and spleen over the next 6 to 24 hours.79 We saw no dendritic cells in the lymph nodes. In fact, only intradermal administration demonstrated dendritic cells in the lymph nodes, and that is about 1%. It is possible that greater migration to critical areas may be achieved through manipulating chemokines receptors.
As Dr. Disis has mentioned, a certain threshold of antigen-specific T cells must be achieved to have a clinical response. Depending on the type of assay used we believe that threshold needs to be 1% or more of circulating cells. How to apply in vivo indicators of immune response such as DTH reactivity is unclear. We demonstrated in a phase I study that it was safe and feasible to administer the dendritic cell vaccine, and found a single minor response in about 25 patients. We did see T cell infiltration at the injection sites where the vaccine was administered, although there was little pure DTH reactivity in terms of erythema and induration.80
A variety of in vitro assays have also been used to measure T cell responses from the peripheral blood. The enzyme-linked immunoSPOT (ELISPOT) assay is one used in many investigations. This assay involves coating plates with antibodies that can recognize a particular cytokine, such as interferon gamma production by T cells in response to antigen. T cells and antigen are added and T cell stimulation is permitted for several hours. An antibody that can recognize the cytokine at a different site is then introduced. Binding by this antibody is demonstrated by an enzymatic reaction as in a typical enzyme-linked immunoassay (ELISA). Ultimately, we see a spot at each site where there is a cytokine-secreting T cell. This assay can be difficult to read, particularly when there are a large number of spots. Automated readers are available. Figure 11 is an example of what data from an ELISPOT assay looks like. [Figure 11. Normal Donor PBMC (EBV+) ELISPOT Analysis] This data is from a healthy volunteer who is EBV-positive, which means the volunteer had been exposed to EBV and had a specific response. This assay shows approximately 75 to 100 cells per 200,000 peripheral blood mononuclear cells recognized EBV.
Other assays used include the tetramer assay, which is essentially a phenotypic assay, and intracellular cytokine assay. The intracellular cytokine assay is a functional assay that can measure multiple cytokines, use defined antigens, and measure class I or II response. However, it is technically demanding to perform and requires 5 hours of in vitro stimulation. Regardless of the assay used, it is important to look at a variety of time points after immunization to quantify results.
In order to stimulate an immune response that can be measured by an assay, the immune system first must recognize it. Dr. Disis has introduced the importance of tumor antigens. Most of the studies conducted to date use a target cell mixed with a peptide or protein, so that the target cell expresses the antigen we want the immune system to attack. This approach permits us to determine if the T cells recognizing the antigen of interest have been stimulated, but it does not tell us whether the T cells can destroy actual patient tumor cells. One study found that autologous dendritic cells loaded with idiotype (ID) protein could stimulate cytolytic T cells capable of recognizing patient myeloma cells.81
HER-2/neu is another antigen used in dendritic cell vaccines. Currently it is under investigation in patients with no evidence of disease, as they may be more likely to mount an immune response and receive clinical benefit. Patients received dendritic cells loaded with a protein fragment of HER-2/neu or the control antigen, KLH, every 3 weeks. In vitro, we were able to demonstrate that this approach did indeed induce T cells that lysed HER-2/neu-expressing target cells. In the first two patients we have detected induction of HER-2/neu and KLH-specific T cell responses by ELISPOT. The response to KLH demonstrates a primary immune response can be elicited, since most people have not been exposed to it.
Table 5 [Table 5. Dendritic Cell Vaccines] shows the experience with dendritic cell vaccines in the past 2 years.82-94 Several generalizations can be made from these studies. First, dendritic cell-based immunizations are feasible and safe. Clinically significant adverse events are exceedingly rare. Second, low levels of antigen-specific immune responses can be induced in some patients in most of the studies. Third, clinical responses, including complete responses, can be evoked in a minority of patients and these responses are more common in studies targeting melanoma. Fourth, no one strategy appears to be significantly more effective. This suggests the continued need to use immune response assays to help guide further development of more potent vaccines. If it is true that antigen-specific T cells must account for 0.1% to 1% of circulating T cells in order for there to be clinical activity, we will need to significantly improve these vaccines before being able to make definitive statements about their clinical studies.