Cancer Gene Therapy (2006)
13, 725–731. doi:10.1038/sj.cgt.7700950; published online 10 March 2006
The rationale for prophylactic cancer vaccines and need for a paradigm shift
R E Sobol
1
1Sidney Kimmel Cancer Center, Introgen Therapeutics Inc., Houston, TX, USA
Correspondence: Dr RE Sobol, Sidney Kimmel Cancer Center, Introgen Therapeutics Inc., 2250 Holcombe Blvd., Houston, TX 77030, USA. E-mail:
r.sobol@introgen.com
Received 3 July 2005; Revised 20 January 2006; Accepted 22 January 2006; Published online 10 March 2006.
Top of page Abstract
This review summarizes clinical experience with infectious disease vaccines and data from animal tumor models that support a paradigm shift for cancer vaccines from therapeutic to prevention applications.
Keywords:
cancer prevention, cancer vaccines
I shall endeavour still further to prosecute this inquiry, an inquiry I trust not merely speculative, but of sufficient moment to inspire the pleasing hope of its becoming essentially beneficial to mankind – Edward Jenner
An ounce of prevention is worth a pound of cure...
Top of page Introduction
While the history of medicine is replete with many remarkable achievements, perhaps none are as great or underappreciated as the development of vaccines to prevent disease. Scourges of the past that claimed countless lives and produced immeasurable suffering have been totally eliminated by successful vaccines. These diseases and the medical achievements that lead to their eradication relegated to the archives of history. Unfortunately, the larger size, protracted nature and liabilities of prophylactic clinical trials have combined to impede the development of prophylactic agents. Recent experience with tamoxifen, finasteride and cyclooxygenase-2 (COX-2) inhibitors to prevent breast, prostate and colon cancers, respectively, has shown both the promise and risks of pharmacological cancer prevention and importantly demonstrated the feasibility of performing prophylactic clinical trials.
1,
2,
3,
4 It is the purpose of this review to summarize previous studies that provide the rationale for a paradigm shift in the development of cancer vaccines to emphasize prevention applications.
Top of page The past: lessons from clinical experience, animal tumor models and infectious diseases
Clinical experience
The field of tumor immunology was born over a century ago when astute clinicians observed cancer remissions in patients with concurrent infections.
5 Since that time, tumor immunologists have continually applied the latest technologies and advances in understanding of the immune system and tumor biology to develop novel tumor immunotherapies. Efforts to induce infections at tumor sites were among the initial immunotherapeutic approaches that have culminated in the currently approved use of intratumoral administration of the bacillus-Calmette-Guerin (BCG) bacteria as an approved therapy for superficial bladder cancer.
6 Subsequently, increased knowledge of the specific cytokines mediating immune reactions combined with recombinant DNA technology permitted purified cytokine production in large quantities and led to approved applications of interferon(IFN)-alpha and interleukin(IL)-2 for the treatment of selected cancers.
7,
8 The discovery of antibodies and technology breakthroughs permitting the production of large quantities of pure, monoclonal antibodies eventually led after many decades of experimental trials to approved antibody therapies directed against cell surface antigens for the treatment of B-cell malignancies, breast and colon cancers.
9,
10,
11,
12,
13
Similarly, therapeutic tumor vaccines have been extensively studied in animal models and in clinical trials.
5,
14,
15,
16,
17,
18,
19,
20,
21,
22,
23,
24,
25,
26,
27,
28,
29,
30,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42,
43,
44,
45 As with cytokine and antibody therapies, tumor vaccines have evolved and incorporated advances in knowledge and technologies. Crude tumor lysates, purified tumor antigens, whole tumor cells, tumor cells genetically engineered to secrete immunostimulatory cytokines and recombinant DNA antigens in a wide variety of delivery vehicles and formulations have been evaluated as therapeutic vaccines to treat established tumors.
5,
14,
15,
16,
17,
18,
19,
20,
21,
22,
23,
24,
25,
26,
27,
28,
29,
30,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42,
43,
44,
45 Despite the hopes born of ever more powerful technologies and deeper understanding, these approaches have been relatively disappointing in providing significant clinical benefit in the therapeutic setting.
5,
14,
15,
16,
17,
18,
19,
20,
21,
22,
23,
24,
25,
26,
27,
28,
29,
30,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42,
43,
44,
45,
46
Animal models
Evaluations of experimental vaccines in animal tumor models have revealed several important principles that have largely been ignored in the development of tumor vaccines. Numerous studies with a wide variety of transplantable tumors have indicated that vaccine preparations are extraordinarily effective at preventing tumor development but are largely ineffective at treating established malignancies.
5,
14,
15 Consistent with these observations, therapeutic tumor vaccines in animal models are most effective when the tumor inoculum is small and lose efficacy when the tumor burden is increased.
5,
14,
15 These fundamental findings have been observed for decades and indicate that tumor vaccines are far more likely to be effective in preventing cancers that are known to originate from a single cell than in curing clinically apparent malignancies comprised of billions of cells that have evolved to escape immune defenses. By way of analogies, a rifle is useful to kill a single tiger but would be of little help to thwart an attack by 100 000 bengals. Similarly, an extinguisher may stop a small kitchen fire but could not save a house in flames. While the development of therapeutic vaccines to treat established tumors should continue, decades of results in experimental animal tumor models clearly indicate that tumor vaccines are far more likely to be effective in the prophylactic setting to prevent cancer development than they will be to treat established cancers with larger tumor burdens.
Infectious disease vaccines
There is also much to be learned from the parallel development of vaccines for infectious diseases that have enjoyed great success in disease prevention. How many of these extraordinary agents would be available today if their development was predicated (as it is for tumor vaccines) upon the demonstration of efficacy in the terminal stages of infections? It is well known that most vaccines against infectious diseases are not effective post-infection as therapeutic agents. For example, polio and pneumoccocal vaccines have no meaningful clinical activity when administered after infections have become established yet, they are effective in disease prevention. Neither of these vaccines would be considered valuable if judged on their ability to provide clinical benefit in end-stage polio or pneumonia patients. A few infectious disease vaccines, rabies is an example, are effective post-infection.
47 However, their efficacy is greatest when administered soon after exposure, and clinical benefit declines when they are given later in the course of disease progression. These findings indicate that vaccine preparations with limited or no benefit in established disease may have extraordinary activity in disease prevention. The origin of infectious disease vaccines – the cow pox/small pox vaccine by Jenner
48 – focused this field on prophylactic applications resulting in a storied history of successes that are among the greatest medical achievements. How would this field have faired if it persisted in developing therapeutic vaccines to treat established disease? It would likely be considered as we currently view the field of tumor vaccines – a series of promising conceptions that continually disappoint in the clinical setting requiring further development and breakthroughs. The data from animal tumor and infectious disease models suggest that an important 'breakthrough' may be to apply tumor vaccines in a more appropriate clinical setting – to prevent rather than to treat cancer.
Top of page The present: target antigens, transgenic tumor models, candidate prophylactic vaccines and target populations
Target antigens
In fairness to tumor immunologists, they have had a more challenging road to travel in the development of vaccines compared to their counterparts in infectious diseases. The causative agents of most infectious diseases have been known and isolated for decades facilitating vaccine development. In addition, these 'foreign' pathogens are viewed as non-self antigens by the immune system and have the propensity to stimulate the most efficacious immune responses. As tumors arise from host tissues, it is believed that they largely express 'self' antigens to which the immune system has been tolerized making the development of vaccines more problematic. However, progress in tumor biology over the past decade has identified several tumor-associated antigens against which tolerance does not develop, is limited or may be overcome. These tumor associated antigens may be suitable for prophylactic vaccine development and include tumor-glycolipids and glycoproteins (e.g. gangliosides); antigens formed by chromosome translocations or oncogene/tumor suppressor gene mutations (e.g.
bcr/abl,
ras); developmental antigens (e.g. MAGE, tyrosinase, melan-A and gp75); antigens upregulated in malignant transformation (oncofetal antigens–carcinoembryonic antigen (CEA), alphafetoprotein (AFP), growth factor receptors-Her2/neu, telomerase and p53) and viral antigens associated with tumor pathogenesis (hepatitis, papilloma and Epstein–Barr viruses).
The hepatitis, papilloma and Epstein–Barr viruses are associated with the pathogenesis of hepatomas, cervical cancers and lymphomas/nasopharyngeal carcinomas, respectively. The complete gene sequences of these viruses are known, facilitating their incorporation into vaccine preparations. These viral antigens should share the advantages of vaccines that have been successful in the prevention of other diseases caused by 'foreign' pathogens. In this regard, trials of papilloma and hepatitis vaccines have shown promise in the successful elimination of the causative agents of cervical and hepatocellular carcinomas and should decrease the incidence of these malignancies.
49,
50,
51
Chromosome translocations and mutated oncogenes like
bcr/
abl and
ras contain mutated sequences that may serve as neoantigens against which immune tolerance has not developed and should not induce autoimmunity. Whereas some translocations and oncogene mutations occur in 'hot spots' and are shared between individuals, more frequently they are numerous, highly varied and this class of antigens may be too impractical for broad prophylactic vaccine development.
Some degree of tolerance may have developed against other types of tumor-associated antigens that contain epitopes shared by a large number of tumors that are overexpressed in malignancies. Targets with these characteristics include other oncogenes (
Her-2/neu), developmental (CEA, AFP), differentiation (MUC-1, Melan-A), tumor suppressor (p53) and telomerase antigens. It should be noted that features desirable for therapeutic applications such as the generation of high-avidity immune effectors or oncogenic functional dependence of target antigens may not be necessary for effective prophylaxis. The lower avidity of immune effectors against tolerized antigens may be sufficient to prevent tumor development,
30,
31 particularly if the tumor burden is small. Target antigens associated with tumor induction or maintenance, while desirable for therapeutic vaccines to avoid escape variants, may not be required for prophylaxis when destruction of a single aberrant cell may theoretically be sufficient for efficacy. Hence, overexpressed, tolerized antigens that would be less desirable candidates for immunotherapy may possibly be very effective targets for prophylactic applications. Vaccines incorporating these antigens have been efficacious in transgenic animal models of cancer prevention, as summarized in the next section.
Transgenic tumor models and candidate vaccines
The recent generation of transgenic animals provides heretofore unavailable tumor models approximating clinical circumstances to evaluate the feasibility and utility of cancer prophylaxis with vaccines containing tolerized tumor antigens.
52 In some of these systems, transplantable mouse tumor models have 'knocked in' human tumor antigens to permit the assessment of vaccine approaches against tolerized antigens. In other models, oncogenes or tumor suppressor genes are either 'knocked in' or 'knocked out', respectively, to generate spontaneously arising tumors. To develop models of specific tumor histological types, tissue-specific promoters are utilized to drive the expression of oncogenes or conditional deletion of tumor suppressor genes or activation of oncogenes is achieved via cre-lox technology. In addition, 'knockout' mice with deletions of immunologically relevant molecules have been generated to assess functional mechanisms. Combinations of these approaches have resulted in the production of animal tumor models that are much more relevant to the evaluation of prophylactic vaccines employing tolerized antigens than previously available tumor transplantation models.
52 The results of immune prophylaxis in these models have been encouraging and key findings from these studies are briefly summarized below.
A vaccine combining IL-12 and allogeneic mammary carcinoma cells expressing p185(neu) completely prevented tumor onset in HER-2/neu transgenic BALB/c mice (NeuT mice).
53 The immune protection elicited was independent of CTL activity and the IL-12-engineered cell vaccine elicited a high production of IFN-gamma and IL-4 and a strong anti-HER-2/neu antibody response. Immune protection was lost or markedly impaired in BALB-neuT mice lacking IFN-gamma or antibody production, respectively. In the these studies, NeuT mice were crossed with knockout mice lacking IFN-gamma production (IFN-gamma(-/-)) or with B-cell-deficient mice (microMT). Vaccination did not protect NeuT-IFN-gamma(-/-) mice, indicating a central role of IFN-gamma. The block of Ab production in NeuT-microMT mice was incomplete. About one-third of NeuT-microMT mice failed to produce Abs and displayed a rapid tumor onset. By contrast, those NeuT-microMT mice that responded to the vaccine with a robust production of anti-p185(neu) Ab displayed a markedly delayed tumor onset. These findings show that inhibition of HER-2/neu carcinogenesis depends on cytokines and specific antibody responses. Similar associations with antibody and IFN-gamma responses were also observed in an animal model of acute promyelocytic leukemia evaluating a DNA-based vaccine fusing the human promyelocytic leukemia-retinoic acid receptor-alpha (PML-RAR
) oncogene to tetanus fragment C (FrC) sequences. The DNA vaccine had a pronounced effect on survival, both alone and when combined with all-
trans retinoic acid (ATRA). The survival advantage was concomitant with time-dependent antibody production and an increase in IFN-gamma. When DNA vaccination and conventional ATRA therapy were combined, they induce protective immune responses against leukemia progression.
54
Several encouraging studies with CEA vaccines have been performed in CEA transgenic mice.
55,
56 Immune prophylaxis was evaluated utilizing recombinant poxviral vectors (recombinant vaccinia, recombinant fowlpox) encoding the CEA transgene as well as a triad of costimulatory molecules (B7-1, ICAM-1 and LFA-3 'TRICOM'). These studies were performed in transgenic mice that express the human CEA gene bred with mice bearing a mutation in the Apc (Delta850) gene (multiple intestinal neoplasia mice), where the progeny (CEA transgenic/multiple intestinal neoplasia) spontaneously develop multiple intestinal neoplasms that overexpress CEA and COX-2. Beginning at 30 days of age, the administration of a prime/boost recombinant CEA-poxvirus-based vaccine regimen or celecoxib (1000 p.p.m.) reduced the number of intestinal neoplasms by 54 and 65%, respectively. Combining the CEA-based vaccine with the celecoxib treatment reduced tumor burden by 95% and significantly improved overall long-term survival. Both tumor reduction and improved overall survival were achieved without any evidence of autoimmunity directed at CEA-expressing or other normal tissues. Celecoxib is prescribed for the treatment of familial adenomatous polyposis in humans, and the CEA-based vaccines have been well tolerated and capable of eliciting anti-CEA host immune responses in early clinical studies. The results suggest that the administration of a recombinant poxvirus-based vaccine is compatible with celecoxib, and this combined chemoimmuno-based approach might lead to an additive therapeutic antitumor benefit not only in patients diagnosed with familial adenomatous polyposis but, perhaps, in other preventive settings in which COX-2 overexpression is associated with progression from premalignancy to neoplasia. However, recent concerns regarding the cardiac safety of currently available COX-2 inhibitor therapy may preclude their use in cancer prophylaxis.
4
A carcinoembryonic Ag (CEA)-based DNA vaccine encoding both CEA and CD40 ligand (CD40L) trimer achieved effective tumor-protective immunity against murine colon carcinoma in CEA-A2Kb double transgenic mice by activating both naive T cells and dendritic cells.
57 Peripheral T-cell tolerance to CEA was broken by this dual-function DNA vaccine, whose efficacy was further enhanced by treatments with a recombinant Ab-IL-2 fusion protein (huKS1/4-IL-2). A lethal challenge of MC38-CEA-KS Ag murine colon carcinoma cells was completely rejected in 100% of experimental animals treated by oral administration of this DNA vaccine carried by attenuated
Salmonella typhimurium, followed by huKS1/4-IL-2 treatment. The oral DNA minigene vaccine induced effective HLA-A2-restricted, CEA-specific antitumor CTL responses. Importantly, peripheral T-cell tolerance against CEA in CEA-A2Kb double transgenic mice was broken by the CEA CD40L vaccine.
57
Another CD40L directed vaccine vector has demonstrated efficacy in an animal model transgenic for the human tumor antigen MUC-1.
58 This tumor antigen is an epithelial mucin glycoprotein that is overexpressed in 90% of all adenocarcinomas including breast, lung, pancreas, prostate, stomach, colon and ovary. An adenoviral vector vaccine was constructed encoding a fusion protein containing an amino-terminal tumor-associated antigen fragment fused to the CD40L. When this vector was injected into hMUC-1.Tg mice, which are transgenic for the hMUC-1 antigen, the growth of syngeneic hMUC-1-positive LL1/LL2hMUC-1 mouse cancer cells was suppressed in 100% of the injected animals. The hMUC-1.Tg mice are anergic to the hMUC-1 antigen before the injection of the vector. In related studies, subcutaneous injection of an adenoviral vector encoding a fusion protein of the human papillomavirus E7 foreign antigen linked to the CD40L generates CD8+ T cell-dependent immunity that suppresses the growth of the E7-positive syngeneic TC-1 tumors in C57BL/6 mice for up to 1 year. These experimental results show that it is possible to employ CD40L directed vaccines to activate a long-lasting cellular immune response against self-antigens in anergic animals. The vector-mediated
in vivo activation, and tumor-associated antigen loading of dendritic cells, does not appear to require additional cytokine boosting to induce effective antitumor immunity. This vector design may therefore be of use in the development of immunoprohylaxis for the many carcinomas in which the hMUC-1 antigen is overexpressed.
58
Target populations
Progress in epidemiology and cancer genetics has identified populations at increased risk for developing particular types of cancer. These individuals may have a genetic predisposition to develop cancer (BRCA1/BRCA2 mutations and breast cancer) and/or have pre-malignant lesions (colonic adenomatous polyps and colon cancer) or engage in activities that increase the risk for developing malignancies (smokers and lung cancer). These populations include those in whom cancer chemoprevention trials have previously been performed.
1,
2,
3,
59 Over time, advances in genomics research will identify individuals at risk for developing particular cancers with ever greater precision. These populations are most likely to benefit from prophylactic interventions and define appropriate groups to target for vaccine development and testing.
In view of the growing research efforts in chemoprevention, the NCI has developed a Prevention Trials Decision Network (PTDN) to formalize the evaluation and approval process for large-scale chemoprevention trials. The PTDN addresses large trial prioritization and the associated issues of minority recruitment and retention; identification and validation of biomarkers as intermediate end points for cancer; and chemopreventive agent selection and development.
60 A comprehensive database is being established to support the PTDN's decision-making process and will help to determine which agents investigated in preclinical and early-phase clinical trials should move to large-scale testing. Cohorts for large-scale chemoprevention trials include individuals who are determined to be at high risk as a result of genetic predisposition, carcinogenic exposure or the presence of biomarkers indicative of increased risk. Current large-scale trials in well-defined, high-risk populations include the Breast Cancer Prevention Trial (tamoxifen), the Prostate Cancer Prevention Trial (finasteride) and the
N-(4-hydroxyphenyl) retinamide (4-HPR) breast cancer prevention study being conducted in Milan. Biomarker studies will provide valuable information for refining the design and facilitating the implementation of future large-scale trials. The validation of biomarkers as surrogate end points for cancer incidence in high-risk cohorts will allow more agents to be evaluated in shorter studies that use fewer subjects to achieve the desired statistical power. It would seem logical for the PTDN to develop analogous algorithms for the evaluation and testing of immunoprohylactic agents.
Top of page The future: development algorithms and considerations to optimize prophylactic vaccines
The ideal prophylactic vaccine would prevent the development of cancer and generate long-lasting immunity without causing deleterious autoimmune responses following practical means of administration. Theoretically, these properties of candidate vaccines could be defined in transgenic animal tumor models and phase I clinical trials in cancer patients who have failed standard treatments. This general approach could identify vaccines suitable for further clinical development either as prophylactic or therapeutic agents. Importantly, investigations correlating success in transgenic animal models with clinical results will be necessary to identify the types of immune responses that predict success for prophylactic and therapeutic applications. This knowledge will facilitate selection of candidate agents for development and potentially define useful surrogate end points to shorten the duration of clinical trials. Initiatives to identify these fundamental principles should become a higher priority for both public and private sector funding.
It would seem clear that the first efforts to develop prophylactic cancer vaccines would be focused upon those malignancies where infectious agents contribute to tumor pathogenesis. Papilloma, hepatitis and Epstein–Barr virus vaccines are currently under development with encouraging results observed in clinical trials. Preliminary results suggest that effective immunity against these pathogens reduces the incidence of infectious disease and secondarily reduces the incidence of malignancy associated with these pathogens.
49,
50,
51 The hepatitis B virus vaccine is the paradigm for this class of cancer prevention agents. Effective hepatitis B vaccines have been shown to decrease the incidence of hepatomas associated with this viral infection.
51
It may be reasonable to demonstrate efficacy and safety in transgenic animal tumor models as an initial component of a general algorithm to develop vaccines targeting cancers that are not associated with viral pathogens. Subsequently, phase I clinical trials could be performed in patients who have failed standard cancer therapies. Demonstration of safety, immunological and/or antitumor activity would identify those agents most appropriate for future development as therapeutic or prophylactic agents. Vaccines chosen for prophylactic development would be employed in prevention trials with end points similar to those of other previously performed prophylactic studies.
1,
2,
3 As an important stimulus, it may be very useful for government funding and regulatory agencies to establish standard algorithms that will facilitate vaccine development.
In conclusion, it is possible that our efforts in the war against cancer have been misplaced. On the battlefield of cancer therapeutics, we must defeat a billion enemies. In prophylaxis, we need to conquer just a single opponent – the one transformed cell that may give rise to malignancy. While current vaccine technologies and formulations may already lead to effective prophylaxis, there are several areas of investigation that are likely to enhance vaccine efficacy. Clearly, there is much to be learned regarding the optimum types of immunity, target antigens, routes and schedules of administration to optimize prophylactic cancer vaccines. Initiatives to address these fundamental issues in transgenic animal models and clinical studies must be initiated. These efforts may yield unprecedented results and potentially guide the development of powerful agents to eradicate cancer.
Linear Mode
