Repositioning Therapeutic Cancer Vaccines in the Dawning Era of Potent Immune Interventions

Adrian Bot; Francesco Marincola; Kent A Smith


Expert Rev Vaccines. 2013;12(10):1219-1234. 

In This Article

Cancer Vaccines: What Have We Learned?

The Concept of Immunization Against Cancer Is Deeply Rooted in Classical Immunology & Anti-microbial Vaccination

The principle of therapeutic vaccination against cancer is based on three key pillars of modern experimental medicine and immunology: i) the discovery that microbial infection can result in tumor regression,[1] ii) the observation that one can induce immune responses against tumor cells,[2] and iii) the cloning of the first tumor-associated antigens.[3] As Burnet and Thomas advanced the hypothesis of cancer immune surveillance in the 1950s,[4] subsequent utilization of immunodeficient animals revealed a key role of the immune system in the monitoring and removal of emerging tumors and shed light on a rather dynamic interaction between tumors and immunity.[5] These findings along with the simplicity of the therapeutic vaccination, a concept borrowed and modified from antimicrobial vaccination, fueled a cornucopia of investigational cancer vaccines.[6]

Faulty Premises Directed Cancer Vaccine Development on a Wrong Path

One catalyst for the proliferation of clinical trials with cancer vaccines has been the promising results in many preclinical studies with peptides,[7] recombinant proteins,[8] dendritic cells (DCs),[9] allogeneic tumor cells[10] or cell lysates.[11] We now know that most of the preclinical models utilized to date lack key features present in the clinical setting and, therefore, lack direct translational value.[12]

In addition, the conservative regulatory strategy to first test vaccines in advanced cancer, along with the desire of many companies to seek out shorter development paths, positioned investigational cancer vaccines in unfavorable indications such as metastatic disease. That was coupled with a decisional process based on either an objective response in advanced cancer,[13] a goal that proved to be rather out of touch with the science, or the immune response to the vaccine. The latter also proved to be often misleading, as immune responses against vaccines, easily measurable in accessible specimens, proved in many circumstances to be a 'red herring', not reflective of an adequately potent antitumoral immune response.[14,15] Based on faulty decisional process, certain cancer vaccines were prematurely terminated, while others advanced inadequately to later stage clinical trials without appropriate optimization and eventually lead to their failure. In a few cases where investigational vaccines were deemed successful in Phase I and II trials based on seemingly positive immune responses against the vaccine rather than objective clinical responses, they perished in Phase III when confronted with stringent end points (survival) but only after consuming vast R&D expenditures.

In hindsight, we realized that these factors were a substantial setback to the field. They pigeonholed the therapeutic vaccines onto a faulty path, thus explaining why the numerous vaccines that showed early preclinical and clinical promise failed to deliver in pivotal trials.

Success Has Been Scarce & Came at a Considerable Development Cost With Limited Clinical Benefit

Provenge® was a notable exception[16] to all cancer vaccines that failed late stage clinical development. As the first therapeutic vaccine to be approved for a cancer indication in the United States, it impacted the field at three levels. First, it formally demonstrated through randomized and adequately designed trials that immune stimulation through vaccination can have an impact on cancer in humans. Second, it showed that an antigen (prostate acid phosphatase)-targeted intervention could mobilize the immune response in a clinically meaningful fashion. Third, Provenge validated a path for developing an autologous cell-based immune intervention which has a significant impact beyond vaccine development. Since this vaccine was developed for castrate-resistant prostate cancer, a relatively advanced disease setting, it required a sizable clinical program to demonstrate a subtle, yet statistically significant, clinical benefit. This was reflected in a survival benefit of several months without objective responses or progression-free survival advantages. The outcome of the primary mechanism of action of the vaccine could be a slow down of disease progression. However, there have been some questions in regard to the proof of efficacy of this vaccine, related to the validity of the control group and the potentially deleterious impact of the leukapheresis component of the clinical protocol.[17] To date, some of these questions have been partially addressed, although Provenge still faces challenges in terms of utilization within the clinical armamentarium for prostate carcinoma due to cost and advancement of other therapies.[18,19]

Thus, while Provenge blazed the path toward approval of therapeutic vaccines in cancer, its impact on clinical outcome evident at cohort but not individual patient level, consumed significant resources and increased the risk in late development stage.

Lesson 1: Predictive Biomarkers to Direct Cancer Vaccines to Responsive Patients

Developers of new cancer vaccines have attempted to build on the successes and failures of their predecessors, ascribable to the hurdles described in the Figure 1A & 1B. A notable example is the recombinant MAGE-A3 protein vaccine developed by GlaxoSmithKline for intermediate-stage non-small-cell lung cancer (NSCLC) and melanoma.[20] The vaccine boasts a potent adjuvant combination of Toll-like receptor (TLR9 and TLR4) ligands, together with a more conventional depot-type adjuvant formulation, designed to harness the immunogenic potential of MAGE-A3.[21] Importantly, the MAGE-A3 vaccine program utilizes a patient stratification approach based on an immune gene signature in the tumor tissue at baseline. This screening technique was developed during a Phase II trial utilizing retrospective correlational analysis of gene signatures in tumor tissue with clinical outcome of immunization. This technology is currently being prospectively tested in a randomized Phase III trial, the largest clinical endeavor of its kind to date.[22] The complex but informative design of the trial also attempts to build on the potential compatibility of chemotherapy and vaccination, tested in a stage of non-small-cell lung cancer where there is typically a relatively confined tumor burden but no option for tumor resection.

Figure 1.

Repositioning cancer vaccines. (A) Learning from cancer vaccines experience. Therapeutic vaccination encounters several categories of roadblocks in advanced cancer, as represented in the outer part of the diagram, in red: gaps within the immune repertoire, an immunosuppressed state associated with cancer, and various immune-inhibiting mechanisms within the tumor environment. Detailed characterization of these mechanisms explain why therapeutic vaccines failed or showed modest results in advanced disease setting, what type of immune interventions need to be developed for successful therapy of cancer (as represented within the inner part of the diagram) and how vaccines can be adequately positioned or integrated with other therapies. (B) Therapeutic vaccines could be more optimally positioned in minimal residual disease when the immune repertoire is relatively competent and the tumor microenvironment-associated inhibiting mechanisms are limited or inexistent. (C) An integrated approach for immune intervention. First, repertoire generation can be achieved by adoptively transferring T cells that express endogenous T-cell receptor (tumor-infiltrating lymphocytes), or genetically engineered T-cell receptor, or chimeric antigen receptors. Second, to activate, amplify and maintain this repertoire, one could utilize preparative regimens based on lymphodepletion, T-cell stimulating cytokines such as IL-2 or integration of costimulatory domains into antigen receptors borne by T cells and vaccination. Finally, blocking co-inhibitory molecules expressed within the tumor environment, or engineering T cells that are more resilient to the endogenous immune inhibiting mechanisms, could fully leverage the potential of the effector T cells.

In summary, discovery and utilization of predictive biomarkers such as immune gene signatures, could direct the utilization of cancer vaccines to patients who have a higher likelihood to benefit, a 'win–win' for both the vaccine and patients.

Lesson 2: Rational Combination of Vaccines With Standard of Care

In terms of vaccines utilized as combination with standard of care, one of the most promising current approaches is that of Immatics. The company is conducting a Phase III trial in advanced renal cell carcinoma (RCC) utilizing a collection of peptides discovered via peptide elution from tumor cells. This approach combining a vaccine with sunitinib® (tyrosine kinase inhibitor) and immune-modulating doses of cyclophosphamide attempts to rationally and synergistically position vaccination as an integral part of standard of care.[23] A detailed understanding of the immune response associated with favorable clinical outcome is warranted for a rationale optimization of a combinatorial protocol. For example, this program has been among the first to offer a solid scientific rationale for co-targeting immune suppressor cells in the course of vaccination.[24]

In all, an elegant and feasible strategy to progress cancer vaccines is to define standard of care that is compatible with vaccination, leading to improved clinical outcome on combinatorial treatment.

Lesson 3: Utilization of More Immunogenic Platform Technologies

Another vaccine in late stage clinical trials for the treatment of prostate cancer is the engineered pox virus PROSTVAC® vaccine developed by Bavarian Nordic and designed at the National Cancer Institute.[25] The vaccine, consisting of recombinant vaccinia virus, encodes a triad of human costimulatory molecules (B7-1, ICAM-1 and LFA-3) together with the prostate-specific antigen. In addition to this enhanced vaccination platform, the vaccine also targets a relevant antigen also found in early, hormone-naïve cancer, and MRD associated with rising prostate-specific antigen levels.

In terms of cell-based immunotherapy, DCs are still being pursued in late phase clinical development but they also incorporate additional features aimed to enhance efficacy. For example, Argos Therapeutics is currently testing a DC vaccine that encompasses cells genetically engineered to produce IL-12 on deployment of the CD40/CD40L pathway in vivo.[26] These cells are also transduced with whole tumor mRNA in an attempt to elicit as broad an immune response as possible. Further, this program also attempts to build on the standard of care Sunitinib as a combination chemoimmunotherapy directed against advanced RCC. In addition, Northwest Therapeutics is also developing a DC vaccine platform, optimized by introducing steps to initiate DC maturation during the last stage of the manufacturing process to maximize the immunogenicity.[27]

GM-CSF is a potent immune-modulating cytokine.[28] Another vaccine platform advanced by Kite Pharma, utilizes autologous DCs that are genetically engineered to coexpress high levels of GM-CSF and a self tumor-associated antigen (carbonic anhydrase IX) intimately involved in the biology of RCC.[29] These engineered DCs can directly prime antigen-specific T cells and they also have, in principle, the capability to recruit additional professional antigen-presenting cells (APCs) at the site of vaccination, possibly leading to an amplification of the immune response through cross-processing.

Another vaccine platform makes use of allogeneic or autologous cancer cells genetically engineered to express GM-CSF thereby utilizing a cross-priming mechanism for induction of immunity and have been extensively reviewed elsewhere.[30]

Another category of improved vaccines are DCs engineered with 'supra-physiologic' properties, such as the lack of expression of inhibiting receptors, by genetic manipulation of key master switches such as FOXO3.[31] Another approach pursued by Celldex aims at directly targeting DCs in vivo by utilizing engineered antibodies capable of shuttling tumor-associated antigens via specific receptors, such as DEC-205, which in the context of TLR-stimulation could lead to robust cross-priming of T-cell immunity.[32] CureVacc is also advancing an all RNA-based antigen-specific vaccine formulation that builds on the robust and intrinsic TLR-mediated adjuvant activity of formulated and stabilized RNA organized as secondary and tertiary structures.[33]

Thus, a major area of activity is the design and development of vaccine platforms and candidates that are endowed with an increased immunogenicity, primarily utilizing microbial vectors, TLR ligands or genetically manipulated DCs.

Lesson 4: Redirect Vaccination to the Most Immunogenic Tumor-Associated Antigens

A complementary avenue is redirecting vaccines to more immunobiologically relevant targets.

One category of such targets is expressed on tumor-initiating cells, providing in principle an opportunity to prevent disease relapse.[34]

Another path is represented by designing patient-customized vaccines based on epitope discovery from individual patients, catalyzed by whole cancer exome analysis. This approach is based on the finding that the majority of tumor-infiltrating lymphocytes (TILs) recognize such 'neo-epitopes' in cancer patients[35] and from recent progress in bioinformatics.[36] However, there are still major feasibility hurdles to overcome in this area. Similarly, another appealing category of new targets consists of proteins expressed as alternate reading frames in cancerous cells carrying indel mutations or other gene disregulations.[37] These targets represent non-self-antigens from an immunological standpoint but are shared across patient populations.

Last but not least, directing vaccines to oncogenic non-self proteins represents a most appealing strategy in certain tumors, as dramatically reflected by the rate of immune and clinical response in patients with intravulvar neoplasia immunized against HPV-16 oncoproteins.[38]

In essence, there are still underexplored categories of vaccine targets that could provide an increased opportunity due to their non-self and thus more immunogenic features.

Lesson 5: More Reliable, Earlier Measurements of Vaccine's Success Are Needed

Attention is also being given to improved methods of early prediction of the clinical activity of vaccines and immune interventions in general, especially in early development when companies choose to allocate measured resources. For example, utilization of a gene signature, which represents the immune constant for tumor rejection[39] and indicates whether a tumor mass is under immune-mediated attack, could be a powerful and objective way to evaluate the activity of an investigational vaccine. This approach could also circumvent many of the drawbacks of conventional immune monitoring in the peripheral blood.[40] Most likely, conventional immune monitoring will still be necessary to indicate whether a vaccine candidate is biologically active or not, but should be used cautiously for decision making. Immune monitoring should utilize standardized specimen collection, analysis and data reporting.[41]

In conclusion, there are several strategies aimed at enhancing the clinical effectiveness of cancer vaccines that are currently being tested: patient selection based on gene signature or preexisting immunity;[42] vaccine administration as combination therapy with standard of care or new agents that disrupt immune-inhibiting pathways[43]; utilization of more immunogenic vaccine platforms and of target antigens ideally associated with tumor-initiating cells and that escape the process of central and peripheral immune tolerance.