Chapter 27: Vaccine Therapy

Immunity elicited by therapeutic cancer vaccines offers several advantages over passive immunotherapy using monoclonal antibodies. In active immune therapy, all components of the effector immune response are host derived without murine or xenogeneic components that could cause indirect toxicity. The lack of foreign components also allows the host response to be sustained. Also, if the vaccine contains more than a single determinant of the target antigen, the immune response could be broad in scope, recognizing more than a single epitope in the antigen (polyclonal). This feature might be of particular importance for cancer immunotherapy, as mutation of individual peptide epitopes is a possible mechanism of immune evasion by tumors. In addition to inducing antibodies, which can recognize intact proteins on the surface of tumor cells, vaccines activate T cells that can recognize peptide fragments derived from proteins that may be endogenously processed and presented on the surface of tumor cells. Such T cells have various effector mechanisms capable of neutralizing tumor cells, including lysis of the tumor cell by cell-to-cell contact and the local production of cytokines that might directly neutralize tumor cells (e.g., interferon-γ).

Most therapeutic cancer vaccines that are being tested in clinical trials have at least three components: antigenic material derived from the tumor, a carrier, and an adjuvant. The antigenic material is usually a protein or peptide derived from the tumor that is either uniquely expressed or is overexpressed in the tumor, compared with normal tissues. A unique tumor antigen or the overexpression of the antigen to prevent the tumor is necessary to prevent the induction of an unwanted autoimmune response against normal tissues following vaccination. The carrier is necessary for delivery of the tumor antigen to antigen-presenting cells, such as dendritic cells, so as to induce the immune response against the tumor antigen. The third component of a cancer vaccine, the adjuvant, is usually a cytokine or other nonspecific immune stimulant to facilitate an enhanced immune response against the tumor antigen.

ANTIGEN DISCOVERY

Both conventional and novel technologies used to define cancer-associated antigens, such as serologic analysis by recombinant expression cloning (SEREX), serial analysis of gene expression (SAGE), screening tumor complementary DNA (cDNA) libraries with tumor-reactive T cells, and characterization of peptides eluted from tumor-derived human leukocyte antigen (HLA) molecules, have resulted in a rapidly growing list of candidate tumor antigens for various hematologic malignancies (Table 27–1). The majority of these candidate antigens have been identified since 1998. Furthermore, the application of genomic and proteomic techniques, combined with the feasibility of isolating sufficient quantities of clonogenic tumor cells from individual patients, should identify additional targets that are differentially expressed in tumors as compared with normal tissues.

Desirable characteristics for candidate target antigens for immune therapy include antigens that are selectively or aberrantly expressed by the tumor or that are required to maintain the malignant cell phenotype or cell survival. Even though the adaptive immune response is comprised of both humoral and cellular components, most efforts at tumor vaccination have focused on eliciting T-cell responses because of their central role in regulating and mediating the overall adaptive response (Chap. 76). Host T-cell recognition of such antigens requires that they are naturally processed and presented by tumor cells into peptides that bind host HLA molecules. Optimally, the candidate antigen should contain both CD4+ and CD8+ T-cell epitopes. Antigens recognized by humoral immune responses must be expressed on the tumor cell surface, and the relevant epitopes must be accessible to antibody molecules. Immunogenic tumor antigens should provoke a strong effector response and not induce tolerance.

Vaccine therapy does not necessarily require a completely defined tumor antigen. Vaccines can consist of whole tumor cells or subcellular components containing putative antigens. For example, autologous tumor cells engineered to overexpress cytokines such as granulocyte-macrophage colony-stimulating factor (GM-CSF),^15,16 or activated ex vivo by CD40 receptor engagement,¹⁷ can be effective at inducing tumor-specific T cells with as yet undefined antigen specificity. Similarly, transfer of the gene encoding CD40-ligand into chronic lymphocytic leukemia cells induced CD4+ and CD8+ T-cell responses in human patients.¹⁸ Vaccination with membrane proteins extracted from tumor cells and incorporated into liposomes along with interleukin (IL)-2 is another strategy that is in clinical testing.¹⁹

VACCINE DELIVERY

Effective delivery of the target antigen to the immune system is critical for the successful induction of immunity. For most tumor antigens, this is a daunting challenge, as most antigens (with the exception of viral antigens associated with cancers) are weakly immunogenic, self, or tissue-differentiation antigens.

Many vaccine-delivery strategies use dendritic cells. These key antigen-presenting cells are principally responsible for initiating a host immune response.²⁰ The cells, represented in minute quantities, have the powerful capacity to take up antigens, and once activated, to present processed peptides to T cells. Accordingly, optimizing the delivery of tumor antigens to specialized antigen-presenting cells is critical. Such efforts have included isolation of dendritic cells from blood, followed by physical loading with protein or peptide antigens, or introducing the genes for candidate antigens by transfection with cDNA or messenger RNA, or by fusion with whole tumor cells. Loaded dendritic cells have been administered to patients as vaccines.²¹

An alternative strategy is to target the delivery of antigens to dendritic cells in vivo. Traditional approaches focused on attempts to make the antigen look foreign to the host immune system; for example, by chemical linkage to larger, highly immunogenic proteins (carriers) or incorporation into liposomes. Rational approaches to increase the efficiency of antigen delivery to dendritic cells have included genetic fusion of the gene encoding the antigen to one encoding biologically active molecules that has the ability to target cell-surface receptors on antigen-presenting cells. Such targeting molecules have included cytokines, chemokines, antibody Fc or Fab fragments, transferrin, CD40, and mannose, which serve as ligands for specific receptors on antigen-presenting cells.²² Such molecular vaccines can be administered as naked DNA or as fusion proteins. Other promising approaches to target dendritic cells in vivo are represented by recombinant viral or bacterial vectors or virus-like particles.^23,24

IMMUNOSTIMULANTS TO ENHANCE VACCINE EFFICACY

Traditionally, immunologic adjuvants, described by the late Charles Janeway as “immunology’s dirty little secret,” such as alum and oil-in-water emulsions (e.g., incomplete Freund adjuvant), provide a physical depot for slow release of antigen. Adjuvants also serve as general immune stimulants by providing a danger signal to activate antigen-presenting cells. This feature describes classical adjuvant components, such as bacterial cell wall extracts, as well as unmethylated CpG DNA sequences, which deliver maturation signals to dendritic cells through toll-like receptors (Chap. 20).²⁵ The incorporation of either recombinant cytokines or their genes into vaccine formulations may increase vaccine potency by broadly enhancing the function of either antigen-presenting cells or T cells. Consequently, cytokines, such as interferon-γ, IL-2, and IL-15, may be useful as components of vaccines.²⁶ Such cytokines also can help direct the type of immune response elicited. For example, IL-12 elicits primarily T-helper (Th) type 1 cell responses, whereas the inclusion of IL-4 or IL-10 generally induces predominantly Th2 cell responses (Chap. 76). Some cytokines, such as GM-CSF, which can induce dendritic cell differentiation, can also function as an adjuvant by recruiting antigen-presenting cells to local vaccination sites.²⁷

Cancer vaccine trials might not fit into the paradigm developed for chemotherapeutic agents, which have direct effects on tumor and normal host cells. For example, studies in heavily pretreated patients with terminal disease might be inappropriate for vaccines, which generally require an intact host immune system. For this reason, even safety cannot be evaluated completely in patients who cannot make an immune response, because any toxicity will likely be indirect, resulting from the immune response elicited. In addition, animal models show that the immune system may be more effective at clearing minimal residual disease than at clearing advanced tumor cell burdens. Accordingly, several late-stage clinical trials of cancer vaccines are testing this approach in the setting of clinical remission, after primary surgery or chemotherapy.

Although conventional clinical trials generally test one experimental agent at a time, vaccine formulations may contain several components. The simultaneous optimization of multiple variables (e.g., vaccine and adjuvant dose and schedule, and routes of administration) in a single clinical study often requires the application of novel, more flexible clinical trial design.²⁸

ASSAYS OF VACCINE EFFICACY

The development of surrogate measures of vaccine efficacy has potential value for answering the scientific question of whether it is even possible to vaccinate human patients against a candidate antigen. Traditional assays of immune response, including simple lymphoproliferation and cytotoxicity assays, requiring prolonged periods of prior stimulation, are being replaced by quantitative assays that can measure effector function of T cells directly sampled from blood (e.g., enzyme-linked immunospot assay) and by sensitive tetramer binding assays (Table 27–2).²⁹ In some cases, tetramer-binding assays have been combined with intracellular cytokine production to provide both quantitative and functional analyses of antigen-specific T cells.³⁰ An important aim of clinical trials is to determine which, if any, of these measures of immune response are valid surrogates for vaccine efficacy.

CLINICAL TRIALS OF PEPTIDE VACCINATION IN MYELOID LEUKEMIAS

Peptide vaccines derived from primary granule proteins, including proteinase 3, the Wilms tumor 1 protein (WT1), and the fusion sequence of the BCR-ABL protein of chronic myelogenous leukemia, show promising results in early trials of patients with myeloid leukemias.^{27,28,29,30,31} Both WT1 and PR1 vaccines (a 9 amino acid peptide derived from proteinase 3) induce peptide-specific cytotoxic CD8+ T cells, which are associated with a fall in WT1 expression, and have induced complete and partial remissions in patients with myeloid leukemia who have relapsed. Other vaccines, such as BCR-ABL peptides and heat-shock protein 70 peptide complexes, have also been administered to patients with chronic myelogenous leukemia on conventional treatment with imatinib or interferon-α. Vaccination was associated with a reduction in BCR-ABL transcripts and cytogenetic or molecular responses in some patients.³¹ These preliminary trials show that peptide vaccines that target patients with myeloid malignancies are safe and can induce responses in a subset of patients.

B-CELL ANTIGEN-RECEPTOR VACCINES AS SCIENTIFIC PROOF OF PRINCIPLE

B cells are clonally restricted to express surface immunoglobulin receptors that have unique epitopes present in the antibody variable region termed idiotypes (Chap. 75). Idiotypes expressed by B-cell malignancies are clonally distributed and thus can serve as tumor-specific target antigen for specific immunotherapy. Idiotypes were initially validated as tumor-rejection antigens in mouse models of myeloma and lymphoma,^31,32 and the first clinical trial testing this approach in human patients with lymphoma was reported in 1992.^33,34 Customized idiotype proteins were isolated by heterohybridoma fusion, conjugated chemically to keyhole limpet hemocyanin (KLH), which functioned as a carrier, and emulsified in a simple oil-in-water emulsion. These vaccines elicited predominantly antibody responses.

Subsequently, guided by additional data from murine lymphoma models (Fig. 27–1), recombinant GM-CSF protein was substituted as the immunologic adjuvant. Soluble GM-CSF, initially mixed with the vaccine and then administered for three additional daily doses subcutaneously as close as possible to the original site of immunization, significantly enhanced vaccine potency, consistent with previous gene therapy studies.³⁵ The cellular mechanism of this effect required CD8+ and CD4+ T cells.³⁶

A phase II study was designed to test these vaccines in the setting of minimal residual disease, defined as first remission after chemotherapy in follicular lymphoma patients.⁴ Previously untreated patients first received treatment with uniform chemotherapy to achieve complete remission. After a 6-month break to allow for immune reconstitution, idiotype proteins conjugated with KLH plus GM-CSF vaccines were administered in five monthly doses. Surrogate assays for vaccine efficacy were developed that used autologous lymphoma cells as targets for both B- and T-cell responses. In 19 patients (86 percent), vaccination elicited CD8+ cytotoxic T-lymphocyte cells reactive with the lymphoma cell (Fig. 27–2). More than half of the patients remain in continuous first complete remission, even after a median followup of more than 7 years. A randomized, controlled phase III trial testing this vaccine formulation in follicular lymphoma patients in first remission was reported to improve disease-free survival as compared with controls, suggesting that therapeutic cancer vaccines could induce meaningful clinical benefit in cancer patients.³⁷

IMPEDIMENTS TO VACCINE THERAPY

Despite the success of the idiotype vaccine phase III trial in follicular lymphoma, most other phase III trials of cancer vaccines have been disappointing and objective clinical response rates have been low. Potential reasons for the failure despite the high immunogenicity of vaccines³⁸ may be categorized into factors affecting the afferent or priming phase of the immune response and factors influencing the efferent or effector phase of the immune response. For instance, during the afferent phase of the immune response, it is possible that the magnitude of the T-cell response or the avidity of the induced T-cells was not high enough following vaccination. During the effector phase of the immune response, it is possible that the antitumor T-cells may not have trafficked to the tumor site or, if they trafficked, they may not have been able to overcome newly recognized immunosuppressive mechanisms present in the tumor microenvironment. Several immunosuppressive mechanisms were shown to impair the function of tumor-specific effector T cells in the tumor microenvironment in several animal models and in some human cancers.³⁹ Important negative regulatory pathways that inhibit T-cell function include extrinsic suppression by regulatory T-cells; direct inhibition through inhibitory ligands such as cytotoxic T-lymphocyte antigen-4 (CTLA-4), programmed death-ligand (PD-L) 1, PD-L2, and B7-H4; soluble factors such as transforming growth factor β and IL-10; and metabolic dysregulation of essential amino acids such as tryptophan.³⁹ Therapeutic monoclonal antibodies such as ipilimumab and pembrolizumab that block the immune checkpoint pathways mediated by CTLA-4 and programmed cell death protein-1 (PD-1), respectively, prevent downregulation of T-cell responses and enhance anti-tumor immunity⁴⁰ and are effective in the treatment of nonhematopoietic tumors. Such immune checkpoint inhibitors offer a novel therapeutic approach that might be highly relevant for hematopoietic tumors.

Advances in understanding of immune tolerance and tumor-induced immunosuppression have now provided several novel agents to augment both the afferent and efferent phases of the immune responses in combination strategies with therapeutic vaccines. It is, therefore, an appropriate time to consider the best way in which leukemia vaccines might be incorporated into the overall treatment strategy for that cancer. One strategy is to vaccinate patients after chemotherapy, when a state of minimal residual disease has been achieved. Moreover, lymphopenia induced by chemotherapy can also enhance antitumor immunity by promoting expansion of vaccine-induced effector T cells and minimizing tumor-induced immune suppression.³³ In this setting, the profound lymphopenia from chemotherapy results in release of cytokines, such as IL-15, providing a powerful proliferative drive to lymphocytes, which could be exploited to increase specific lymphocyte responses to the vaccine. Moreover, this strategy can eradicate cells that suppress antitumor responses such as regulatory T cells, and may overcome inherent defects in T-cell signaling, processing, or presentation.³³ In addition, the afferent phase of the immune response could be enhanced by using more potent vaccines with novel adjuvants such as toll-like receptor ligands, or by vaccinating donors of transplant recipients who have a healthy immune system as opposed to patients who may be immunocompromised either from the cancer or from therapy.⁴¹ Alternatively, vaccines could be used in combination with agents that inhibit immunosuppressive mechanisms, such as immune checkpoint receptors/ligands⁴² and/or deplete regulatory T cells,⁴³ to augment the afferent and/or effector phase of the immune response. The use of these agents in combination with therapeutic cancer vaccines may lead to enhanced antitumor immunity and improved clinical outcome.