DNA vaccines, electroporation and their applications in cancer treatment

Abstract

Numerous animal studies and recent clinical studies have shown that electroporation-delivered DNA vaccines can elicit robust Ag-specific CTL responses and reduce disease severity. However, cancer antigens are generally poorly immunogenic, requiring special conditions for immune response induction. To date, many different approaches have been used to elicit Ag-specific CTL and anti-neoplastic responses to DNA vaccines against cancer. In vivo electroporation is one example, whereas others include DNA manipulation, xenogeneic antigen use, immune stimulatory molecule and immune response regulator application, DNA prime-boost immunization strategy use and different DNA delivery methods. These strategies likely increase the immunogenicity of cancer DNA vaccines, thereby contributing to cancer eradication. However, cancer cells are heterogeneous and might become CTL-resistant. Thus, understanding the CTL resistance mechanism(s) employed by cancer cells is critical to develop counter-measures for this immune escape. In this review, the use of electroporation as a DNA delivery method, the strategies used to enhance the immune responses, the cancer antigens that have been tested, and the escape mechanism(s) used by tumor cells are discussed, with a focus on the progress of clinical trials using cancer DNA vaccines.

Keywords:

• DNA vaccine
• cancer
• anti-tumor immunity
• tumor immune evasion

Abbreviations

AFP	=	α-fetoprotein
APCs	=	antigen presenting cells
CEA	=	carcinoembryonic antigen
CTLA-4	=	cytotoxic T lymphocyte-associated antigen-4
DCs	=	dendritic cells
EP	=	electroporation
GITR	=	glucocorticoid-induced tumor necrosis factor receptor family-related gene
hTERT	=	human telomerase reverse transcriptase
HPV	=	human papillomavirus
HSP	=	heat shock protein
HSV	=	herpes simplex virus
ID	=	intradermal
IM	=	intramuscular
MAGE	=	melanoma-associated antigen
MART	=	melanoma antigen recognized by T cells
PAP	=	prostatic acid phosphatase
PD	=	programmed death
PRAME	=	preferentially expressed antigen in melanoma
PSA	=	prostate-specific antigen
PSMA	=	prostate-specific membrane antigen
WT1	=	Wilm's tumor

Introduction

DNA vaccines are not replicating and the vaccine products are expressed within the host cells. They can be constructed to mimic the specificity and safety of subunit vaccines. Due to the production and processing of immunogenic proteins within host cells, DNA vaccines are likely to induce immune responses in a manner similar to live-attenuated vaccine types while causing no pathogenic infection in vivo. By directly administering DNA vaccines into the host, the host cells express the antigenic protein. This process has been known to induce both Ag-specific antibody and cellular responses.^1-3 DNA vaccines are also expected to be safe, stable and cost-effective in their applications, and are easily and rapidly engineered by recent recombinant DNA technologies.

For delivery of DNA vaccines, the majority of DNA vaccine studies have utilized muscle or skin as an immunization target. In intramuscular (IM) injection, DNA vaccines are taken up and expressed by muscle cells and local antigen presenting cells (APCs). The local APCs then migrate to the draining lymph nodes for the induction of adaptive immune responses.^4,5 The APCs, not muscle cells are believed to present antigen to CD4+ and CD8+ T cells by cross-presentation of secreted antigen or by direct transfection of DNA vaccines.^3,6,7 In this case, APCs can provide co-stimulatory signals and cytokines necessary for stimulation of naïve T cells. In addition, antigenic proteins expressed and/or secreted by APCs or non-hematogenous target tissues (e.g., muscle cells in the case of IM injection and keratinocytes in the case of intradermal injection) are presumably recognized by B cells for subsequent antibody production in association with helper functions of Ag-specific CD4+ T cells. In intradermal (ID) delivery, DNA can be taken up by Langerhans cells and/or dermal dendritic cells, which migrate to the draining lymph nodes for induction of adaptive immune responses.⁸ In particular, DNA vaccines, which are generated from bacteria, contain un-methylated CpG motifs and stimulate the innate immune responses by interacting with Toll-like receptor 9 expressed on the surface of APCs (reviewed in Ref. 9). This non-specific activation of APCs likely influences Ag-specific immune responses to DNA vaccines.

For increasing the translation efficacy of DNA vaccines, the genes that code for antigens are codon-optimized and are even linked with specific DNA sequences encoding products that behave as immune-stimulatory or intracellular antigen-targeting molecules. Moreover, for the augmentation of adaptive immune responses, cancer DNA vaccines are co-delivered with cytokines and/or immune response regulators as adjuvants. The DNA prime-boost approach and xenogeneic antigen use have also been proven to be useful strategies to enhance adaptive immune responses. These molecular and biological strategies have contributed to upgrading the immunogenic quality of DNA vaccines.

Cancer antigens are often thought to be poorly immunogenic due to their self-tolerance state. Thus, overcoming immunological tolerance to cancer antigens is a pre-requisite for the induction of potent adaptive immune responses in cancer DNA vaccination. Moreover, cancer antigens are expressed in normal tissues. Thus, cancer-reactive immune cells might attack the normal tissues, possibly causing autoimmune diseases. To date, a number of cancer antigens have been reported. However, continuous work is still needed to identify more potential cancer antigens from malignancies.

In cancer immunotherapy, CTLs are thought to be a major effector cell population targeting neoplastic cells for their lysis. Cancer-reactive CTLs kill neoplastic cells through 2 major killing pathways, the perforin/granzyme B-mediated and Fas/Fas ligand-mediated pathways.^10,11 Despite the induction and potentiation of cancer Ag-specific CTL activity, cancer cells may become immune-resistant cells, no longer responding to the CTL recognition or attacks. This immune-resistant state of cancer cells is often observed after long-term immunotherapy or under conditions of immune selection. It has been reported that cancer cells develop many different mechanism(s) to resist host immune surveillance (reviewed in Ref. 12). The unresponsiveness of cancer cells to cancer-reactive CTLs undermines the ultimate usefulness of cancer DNA vaccines in treating cancer patients. Therefore, it is likely that blocking the immune escape processes of cancer cells is important for achieving the ultimate goal of curing cancer.

As the cancer-killing mechanism(s) of DNA vaccination and conventional therapy modalities such as surgery, radiation, and chemotherapy are all variable, combined therapy using these different modalities might be an option to limit the chance of cancer cells' becoming immune-resistant, thereby achieving better cancer control. In this context, DNA vaccines coding for numerous cancer-specific and cancer-associated antigens have been tested for their safety and effectiveness pre-clinically, as well as in patients. In this article, delivery methods of DNA vaccines, the manipulation of antigens and the use of xenogeneic antigens, approaches to enhance the immunogenicity of DNA vaccines, the types of cancer antigens tested, as well as the immune evasion mechanism(s) utilized by cancer cells and the possible counter-measures for this immune escape will be discussed.

Electroporation as a DNA delivery method

Because the construction and mass production of DNA vaccines are far easier than those of other conventional vaccine types (live attenuated or killed vaccines, and subunit vaccines), as well as their advantages in bio-stability and biosafety, DNA vaccines are thought of as a second-generation vaccine. Historically, striated muscles were first demonstrated to be a site of gene expression following naked DNA injection.¹³ Subsequently, needle injection of DNAs encoding viral antigens into the muscle sites was reported to induce Ag-specific antibody and cellular responses (CTL responses, in particular) in influenza and HIV-1 animal models.^2,14 At the same time, ID delivery of cDNA coding for antigens was also shown to induce Ag-specific antibody and CTL responses.¹⁵ These early animal studies showed that Ag-encoding DNA molecules could be taken up and expressed by striated muscle or skin cells for the induction of adaptive immune responses. Since then, human studies have been initiated to evaluate the efficacy of DNA vaccines in treating various diseases. Contrary to the data on small animal models, the immunogenicity of DNA vaccines in large animals and humans has been far below what is expected,^16-18 possibly due to a low level of protein expression by DNA molecules in relation to the body size and/or a lack of APCs at the DNA injection sites. To overcome this hurdle, many different approaches such as co-injection of DNA vaccines with various immune stimulatory molecules, DNA codon optimization and the targeting of antigens to different cellular compartments, ligation of immune stimulatory molecules to antigens, DNA prime-boost injection schemes and direct injection of DNA vaccines into the lymph nodes have been taken,^19-21 (reviewed in Ref. 22,23). In addition, different DNA delivery methods including gene gun delivery, jet injection, electroporation, tattooing, use of liposome, polymers and nanoparticles have also been tested,^24,25 (reviewed in Ref. 22). Among these methods, gene gun delivery and electroporation methods have been most intensively studied. For instance, a DNA/gold particle bombardment strategy using a gene gun was able to induce potent Ag-specific CTL responses even at a lower dose of DNA than is used for IM injection.²⁶ This gene gun delivery technology utilized the bombardment of DNA/gold particles on the skin surface, where the DNA molecules were taken up and expressed by local dendritic cells (DCs) for antigen processing and presentation.²⁷ Later, in vivo electroporation (EP) was employed as a DNA delivery method, resulting in a dramatic effect on adaptive immune response induction. When compared with gene gun delivery, in vivo EP of DNA plasmid vectors into the muscles was shown to be more effective for protein production and Ag-specific CTL response induction.²⁸ IM-EP is known to increase muscle cell permeability, thus permitting more DNA molecules to enter into the cells for subsequent protein expression.^29-31 IM-EP also increases distribution and cellular uptake of DNA molecules at the injection sites.³² In addition, the injection of needle probes and the application of subsequent electric pulses act to attract immune cells, including DCs, to the DNA injection sites, thus increasing the magnitude and duration of the Ag-specific immune response.³³ In a phase I clinical trial, IM-EP of human papillomavirus (HPV) E6 and E7 DNA vaccines induced robust antibody and CTL responses.³⁴ In a phase II trial, the same DNA vaccine regimens displayed clinical benefits by decreasing the HPV viral load and the HPV-associated disease severity (personal communication with Bagarazzi M). Moreover, ID-EP of DNA vaccines has also been shown to induce both robust and durable immune responses in primates.³⁵ Simultaneously, intratumoral EP of either DNA coding for cytokines or chemotherapeutic drugs has also been shown to have a positive effect on treating skin tumors in animals and humans.^36-40 EP of cDNA expressing antibodies has also been tested as a passive immune therapy strategy. In particular, our un-reported data demonstrated that a serum anti-HER2 antibody was detectable for over 90 days after a single IM-EP of 50 μg of anti-HER2 IgG heavy chain cDNA and 50 μg of anti-HER2 IgG light chain cDNA in mice, suggesting the potential benefit of EP in the passive delivery of antibodies in the form of DNA for cancer treatment. Taken together, EP is thought to be the most potent DNA delivery method for protein expression and subsequent immune induction in small animals and humans. However, more clinical trials using EP are needed to verify the utility of EP-delivered DNA vaccines in treating cancer patients.

Strategies to improve immunogenicity

As mentioned before, cancer antigens are generally poorly immunogenic and require special strategies to increase their immunogenicity. To date, various approaches such as DNA manipulation and the use of xenogeneic antigens, the use of immune response regulators, the co-injection of cytokines and other molecules, the use of DNA prime-boost immunization strategies, and different DNA delivery methods have been mainly taken to achieve this goal. Table 1 summarizes the 5 major strategies that have been employed to enhance the immunogenicity of cancer DNA vaccines.

Table 1. Immune-enhancing strategies for cancer vaccines

DNA manipulation and the use of xenogeneic antigens
Codon optimization	^41,43
CD4+ Th cell epitope domain conjugation	^44,45
Heat shock protein conjugation	^42
Intracellular or extracellular antigen targeting	^41,43, reviewed in^23
Use of xenogeneic antigens	^46–48
Use of regulators of immune responses
Anti-OX-40, anti-4-1BB, anti-GITR Abs	^51–54
Anti-CTLA-4 Abs	^57,58
Co-injection of cytokines and other molecules
IL-2, IL-12 & GM-CSF	^52,60–62,66,67,98
Regulators of DC functions	^64,65
DNA prime-boost immunization strategies
DNA prime-adenovirus boost	^19,68–70
DNA prime-peptide boost	^21
Use of different DNA delivery methods
Gene gun delivery	^26,27
Jet injection	^25
Electroporation	^28,33, reviewed in^23,34
Tattooing	^24
Liposome, nanoparticles	^71,72
Direct injection to the lymph nodes	^20,21

DNA manipulation and the use of xenogeneic antigens

Advances in molecular biological technology have made DNA construction and manipulation far easier than before. Recently, most DNA vaccines are engineered to have codon-optimized antigen genes, which allow for more antigenic protein production. In addition, the strategy of linking antigen-coding genes with other DNA sequences, the products of which are associated with immune activation or the targeting of the antigens to intracellular compartments, has also been evaluated. For instance, the endolysosomal targeting of codon-optimized E7 genes resulted in more antigenic protein production, leading to enhanced Ag-specific CTL and antitumor responses to E6/E7-expressing tumor cell (TC-1) challenges.⁴¹ Conjugation of heat shock protein to HPV 16 E7 antigens also improved E7 DNA vaccine potency against TC-1 cells through augmenting Ag-specific CTL responses.⁴² Conjugation of the IgE leader sequence to codon-optimized E6 and E7 genes also resulted in the induction of much stronger CTL responses and antitumor activity against TC-1 cell challenges.⁴³ Conjugation of the CD4+ T helper cell epitope domain to CTL epitope sequences has also been tested in humans for increasing CTL responses.^44,45 Linking E7 DNA sequences to the genes coding for lysosome-associated membrane protein 1-targeting peptides (for endolysosomal targeting), bacterial toxin (for cellular targeting), calreticulin (for endoplasmic reticulum targeting), viral protein 22 (for intracellular antigen spreading), γ-tubulin (for centrosomal compartment targeting), and heat shock protein (HSP)70 have been demonstrated to be effective at inducing greater antitumor CTL and antitumor responses (reviewed in Ref. 23). These preclinical studies demonstrated that antigen codon optimization, CD4+ T helper cell epitope domain conjugation, and antigen targeting to specific intracellular compartments might be promising strategies to augment cancer Ag-specific CTL responses and antitumor activity.

In the case of non-mutated self-antigens, use of a xenogeneic antigen has been thought to be another promising approach to overcome immunological tolerance to self-antigens. Because xenogeneic antigens are an altered self-protein bearing amino acid substitutions in one or more epitopes, these antigens might overcome immune tolerance and be recognized by T cells, which can then be further activated for the induction of immune responses to both the xenogeneic antigens and the native antigens. The use of homologous proteins from a different species, called xenogeneic antigens, as an immunogen was well demonstrated with melanoma differentiation antigens, such as gp100 and tyrosinase.^46-48 For example, DNA vaccines expressing human gp100 but not murine gp100 induced cross-reactive and anti-melanoma protective CD8+ T cell responses in a murine model.⁴⁶ In a canine melanoma treatment, DNA vaccines encoding human tyrosinase showed a positive effect on the induction of cross-reactive antibodies against canine and human tyrosinases and increased the mean survival time.^47,48 These animal studies suggest that use of a xenogeneic antigen may be an option to overcome natural immunological tolerance to cancer-associated self-antigens.

Use of regulators of immune responses

The fate of T cells is determined by 2 types of molecules expressed on T cell surface, which act as regulators of immune responses.^49,50 One type comprises co-stimulatory molecules, including 4-1BB, GITR [glucocorticoid-induced tumor necrosis factor receptor family-related gene], OX40, ICOS [inducible T cell co-stimulator], CD27, and CD30, whereas the other comprises co-inhibitory molecules including CTLA-4 [cytotoxic T lymphocyte-associated antigen-4], PD [programmed death], and B and T lymphocyte attenuator. To date, numerous animal and clinical studies have been performed to test the ability of these molecules to regulate the efficacy of cancer therapy. For example, we reported using 2 separate animal tumor models that the combined delivery of tumor vaccines with agonistic anti-4-1BB Abs results in the dramatic enhancement of antitumor therapeutic activity by increasing Ag-specific CTL activity.^51,52 DC-based tumor vaccines, in combination with both anti-OX-40 and anti-4-1BB Abs, were also effective at enhancing antitumor activity in a HER2/neu transgenic animal model.⁵³ The combined injection of DNA vaccines coding for gp100 and Trp2 with anti-GITR Abs increased protection against a lethal challenge with B16 melanoma and induced memory CD8+ T cell responses, even in the absence of CD4+ T cells.⁵⁴ In contrast, blocking CTLA-4 using antagonistic anti-CTLA-4 Abs is associated with increased antitumor T cell responses in mice.^55,56 Trp2- and gp100-encoding DNA vaccines in combination with anti-CTLA-4 Abs also displayed enhanced Ag-specific T cell responses and anti-melanoma activity in a B16 model.⁵⁷ In a clinical trial, anti-CTLA-4 Abs improved overall survival in patients with previously treated metastatic melanoma.⁵⁸ It has recently been reported that anti-CTLA-4 Abs play a more important role in broadening tumor-reactive CD8+ T cell responses rather than increasing tumor-specific T cell reactivity,⁵⁹ suggesting that anti-CTLA-4 Abs are mainly associated with Ag-specific T cell activation. Taken together, these animal and clinical results strongly support the notion that the immunogenicity of cancer DNA vaccines could be modulated by combining them with the regulators of immune responses. Presently, the anti-CTLA-4 blocking Ab, Ipilimumab has been approved by the US FDA for the treatment of late stage melanoma.

Co-injection of cytokines and other molecules

The co-injection of DNA vaccines with molecular adjuvants encoding cytokines, chemokines and other immune stimulatory molecules has been known to influence the immunogenicity of DNA vaccines. Previously, our group demonstrated using a herpes simplex virus (HSV)-2 gD DNA vaccine model that the naked injection of gD DNA vaccines along with plasmid DNA encoding Th0-, Th1-, and Th2-type cytokines, chemokines and co-stimulatory molecules can modulate Ag-specific antibody and cellular responses, and enhance protective immunity against HSV-2 challenges (reviewed in Ref. 23). These data suggest that these immune stimulatory molecules could be useful as cancer vaccine adjuvants for enhancing Ag-specific CTL responses, which are required for cancer eradication. In this context, the co-injection of prostate-specific antigen (PSA) DNA vaccines with IL-2 and GM-CSF cDNA augments Ag-specific CTL responses and antitumor protection from a challenge with PSA-expressing tumor cells.⁶⁰ Our data has also shown that IM-EP of HPV E7 DNA vaccines plus IL-2 cDNA is able to enhance antitumor therapeutic activity by increasing the Ag-specific CTL activity in a TC-1 model.⁵² IM injection of carcinoembryonic antigen (CEA) DNA vaccines plus IL-12 expression plasmids resulted in enhanced antitumor effectiveness in an animal tumor model.⁶¹ Contrary to this, we observed that IM-EP of E7 DNA vaccines plus IL-12 plasmid suppressed Ag-specific CTL responses and antitumor activity against TC-1 cells, suggesting that a high level of IL-12 expressed by EP might be detrimental to immune induction.⁶² However, direct EP of IL-12 cDNA alone into tumor sites is also capable of eradicating established B16 melanoma and MC32 colon cancer by inducing tumor Ag-specific CTL responses.^39,63 In a phase I clinical trial including 24 metastatic melanoma patients, intratumoral EP of IL-12 cDNA also showed minimal systemic toxicity.³⁸ In that study, 2 (10%) of the 19 patients with non-electroporated distant lesions and no other systemic therapy showed complete regression of all metastases, whereas 8 patients (42%) displayed disease stabilization or a partial response, suggesting that IL-12 alone can also be useful for achieving cancer control. In addition, the co-injection of HPV E7 DNA vaccines with plasmid DNA encoding anti-apoptotic molecules and serine protease inhibitors increased Ag-specific CTL responses and antitumor activity against TC-1 cells through increased DC functions.^64,65 In a phase I/II clinical study, a different dose of GM-CSF DNA was used as the adjuvant of a multipeptide vaccine (gp100 and tyrosinase) in the treatment of 19 melanoma patients (stage III/IV), and 8 patients (42%) developed T cell responses without any GM-CSF-related side effects.⁶⁶ In a phase I/II clinical trial of 12 patients with follicular B-cell lymphoma in remission after chemotherapy, the delivery of naked DNA vaccines expressing chimeric immunoglobulin plus GM-CSF also demonstrated both the induction of Ag-specific T cell responses and the safety of GM-CSF as an adjuvant.⁶⁷ Taken together, co-injection approaches using immune-stimulatory molecules and regulators of DC functions likely influence the magnitude and the duration of cancer DNA vaccines in their induction of cancer-reactive CTL responses.

DNA prime-boost immunization strategies

DNA prime-boost immunization appears to be a promising strategy as a DNA vaccine regimen. This prime-boost approach has been shown to be promising in generating greater immune responses compared with DNA vaccination alone. For instance, IM-EP of CEA DNA vaccines followed by boosting with adenovirus encoding CEA induced the greatest level of Ag-specific CD4+ and CD8 + T cell responses in wild type mice, and the repeated injection of this prime-boost scheme induced Ag-specific CD8+ T cell responses in CEA transgenic mice,⁶⁸ suggesting that a prime-boost strategy might break through self-tolerance to engender robust immune responses. A prime-boost strategy using either a CEA DNA vaccine fused at the C-terminal end to tetanus toxin fragment C or an adenovirus expressing the fused CEA/toxin C was also more immunogenic than that using the normal CEA counterparts.⁶⁹ Furthermore, CEA DNA vaccine/adenovirus prime-boost schemes were safe and immunogenic in monkeys.¹⁹ However, a recent clinical trial using HER2/CEA DNA vaccine/adenovirus prime-boosting strategy failed to show any measurable T cell responses to the 2 tested antigens.⁷⁰

Use of different DNA delivery methods

To improve the potency of cancer DNA vaccines, many different delivery methods besides aforementioned gen gun delivery and electroporation have also been tested. For instance, jet injector delivery method using HER2/neu DNA vaccines was compared with gene gun delivery, demonstrating that gene gun was superior to jet injector in eliciting Ag-specific CTL responses and antitumor protective immunity.²⁵ Similarly, tattooing delivery method using tetanus toxin fragment C-conjugated MART-1 DNA vaccines has been found to be highly immunogenic in humans.²⁴ Furthermore, direct intranodal delivery of DNA vaccines was also tested for its efficacy.^20,21 DNA vaccines were also formulated with chemical compounds (such as polymers and nanoparticles) and then tested for their utility. For instance, liposome-DNA vaccine complex was more effective at increasing CD8+ T cell numbers and controlling established tumors through the production of IL-12 from dendritic cells and of IFN-γ from NK cells, as compared with DNA vaccines alone.⁷¹ Similarly, higher anti-melanoma immune responses were elicited by formulation of DNA vaccines with nanoparticle, as compared with DNA vaccines alone.⁷² Thus, the use of these formulations and newer delivery strategies likely hold promise for augmenting the potency of cancer DNA vaccines.

Cancer antigens

To date, scientists have identified a number of cancer antigens. In general, cancer antigens are classified as altered self-antigens, cancer/testis antigens, cancer over-expressed antigens, cancer differentiation antigens, and cancer-specific antigens. The typical examples of these antigens are as follows: cancer overexpressed antigens: survivin, human telomerase reverse transcriptase (hTERT), CEA, HER2/neu, α-fetoprotein (AFP), Wilm's tumor (WT1) antigens, PSA, prostate-specific membrane antigen (PSMA), prostatic acid phosphatase (PAP); cancer/testis antigens: melanoma-associated antigen (MAGE), NY-ESO-1, preferentially expressed antigen in melanoma (PRAME); differentiation antigens: melanoma antigen recognized by T cells (MART)-1, gp100, tyrosinase, idiotype immunoglobulin; altered self-antigens: mutated p53; cancer-specific antigens: HPV E6 and E7. As most of these cancer antigens are derived from self-proteins, they are often weak antigens. However, these antigens are recognized by T cells, such that Ag-specific T cells are likely induced and further activated by applying various strategies. A list of cancer DNA vaccines tested in humans (based on our PubMed search) is shown in Table 2.

Table 2. DNA vaccines tested in clinical trials and the results

Tumor antigens	Delivery methods	Disease targets	Results
PSA(GM-CSF & IL-2)	IM & ID injection	Advanced prostate cancer	Ag-specific humoral and cellular responses detected at 900 μg PSA DNA vaccine doses^97
PAP+GM-CSF	ID injection	Recurrent prostate cancer	No adverse effects observed. Ag-specific IFN-γ secreting CD8+ T cell responses detected in 14% patients, Ag-specific CD4+/CD8+ T cell proliferative responses detected in 41% patients. No patients showed Ab responses. PSA doubling time was increased^98 Booster DNA injection induced cytolyic T cell responses in 58% patients^99
PSMA epitope (Tetanus toxin fragment C DOMain-conjugated)	IM injection & IM-EP	Recurrent prostate cancer	Anti-DOM IgG detected^44 DOM-specific CD4+ T cells responses detected in 97% patients. PSMA-specific CD8+ T cell responses detected in 53% patients. PSA doubling time was increased^45
CEA/HER2	DNA prime-Ad boost	Solid tumors	No Ag-specific cellular responses detected^70
AFP	DNA prime-Ad boost	Hepatocellular cancer	Ag-specific T cell responses detected. More favorable progression-free survival in one patient having higher T cell responses^86
HER2 (GM-CSF/IL-2)	IM and IC injection	Breast cancer	Safe and tolerable, long-lasting antibody and cellular responses to HER2 detected^144
Tyrosinase epitopes	Intranodal injection	Melanoma	Practical and tolerable, 42% patients showed immune responses. No clinical responses^110
gp100	IM & ID injection	Melanoma	Tolerable but no clinical and immune responses detected^108
gp100+GM-CSF	Particle-mediated epidermal delivery	Melanoma	Safe and tolerable (mild to no toxicity), modest activation of anti-melanoma responses detected in some patients^109
Melan-A/MART-1/Tyrosinase	Intranodal injection	Melanoma	Safe and tolerable, no clinical responses. Pre-existing/immune responses to Melan-A/MART-1. No immune responses to tyrosinase^20
PRAME+PSMA (Antigen fragments)	DNA prime-peptide boost (Intranodal injection)	Ag-expressing metastatic solid tumors	63% patients showed Ag-specific T cells responses. No partial or complete responses, Stable disease in 7 patients/PSA level decline in some patients^21
Chimeric idiotypes	IM injection	Follicular B-cell lymphoma	Modest levels of immune responses detected^67
E7	IM injection	CIN2/3	Safe and tolerable, no significant immune response detected, little effect on CIN regression^18
E6/E7	IM-EP	CIN2/3	Safe and tolerable, robust Ag-specific Ab and cellular, in particular CTL responses detected^34
E6/E7	IM-EP	CIN2/3	Significant reduction in HPV viral load and the lesion severity (personal communication with Bagarazzi M)

Ad, adenovirus; AFP, α-fetoprotein; CIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; IC, intracutaneous; ID, intradermal; IM, intramuscular; EP, electroporation; PAP, prostatic acid phosphatase; PRAME, preferentially expressed antigen in melanoma; PSA, prostate-specific antigen; PSMA, prostate-specific membrane antigen

Cancer over-expressed antigens

Normal cellular proteins such as survivin (a member of the inhibitor of apoptosis protein family) and hTERT are overexpressed in some major types of cancer.^73-75 In a murine melanoma model, priming with survivin DNA vaccines plus IL-2 cDNA and boosting with rAd were effective at inducing antitumor effects.⁷⁶ hTERT DNA vaccines were also able to induce Ag-specific CD8+ T cell and protective antitumor responses to Ag-expressing tumor cells.^77,78 These animal data show that survivin and hTERT might be used to induce Ag-specific CTL responses for the eradication of malignancies expressing survivin and hTERT.

CEA is over-expressed in colorectal cancer,^79,80 whereas HER2/neu is highly expressed in breast cancer.⁸¹ In an animal study, the co-injection of CEA DNA vaccines with IL-12 cDNA augmented antitumor resistance to CEA-expressing tumor cells in a CD8+ T cell- and perforin-dependent manner.^61,82 A CEA DNA priming and adenovirus boosting strategy also increased antitumor protective responses to CEA-expressing tumors.⁶⁸ DNA vaccines targeting 2 antigens, CEA and HER2/neu were able to induce Ag-specific immune responses by breaking immunological tolerance in CEA/HER2 double-transgenic mice and NOD/scid-DR1 mice, leading to the inhibition of CEA- and HER2/neu-expressing tumor growth.⁸³ However, prime-boost immunization strategies using HER2/CEA DNA vaccines and HER2/CEA-expressing adenovirus failed to induce measurable cellular responses to the 2 tested antigens in humans,⁷⁰ underscoring the importance of breaking immune tolerance prior to inducing Ag-specific immune responses.

AFP is an oncofetal protein which is overexpressed by many hepatocellular cancers.⁸⁴ AFP+GM-CSF DNA prime-adenovirus (expressing AFP) boost vaccines were able to induce Ag-specific Th1 type and antitumor protective responses in mice.⁸⁵ In a clinical trial, AFP DNA prime-adenovirus boost vaccines also induced Ag-specific T cell responses in 2 hepatocellular patients, one of which showed higher Ag-specific T cell responses and more favorable progression-free survival,⁸⁶ suggesting that AFP might be an antigen with potential for hepatocellular cancer immunotherapy.

The WT1 antigen is a transcription factor that is overexpressed both in leukemia cells and in some solid tumors, including lung cancer and colon cancer.^87–89 In humanized transgenic mice, DNA vaccines encoding WT1-derived epitopes, in particular WT1_37–45, were able to elicit Ag-specific CTL responses that recognized leukemia cells. The expanded human CTLs were also able to kill Ag-expressing human target cells in vitro,⁹⁰ suggesting a potential use for WT1 in therapeutic immunization against leukemia and other cancers expressing WT1.

PSA is a serum marker that is widely used in the detection and monitoring of prostate cancer.^91,92 PSMA and PAP are also known to be over-expressed in prostate cancer.^93,94 In a murine model, ID injection of PSA DNA vaccines resulted in the induction of potent Ag-specific antibody and cellular responses, as well as of antitumor resistance to Ag-expressing tumors.⁹⁵ Furthermore, IM-EP of DNA vaccines expressing PSA and PSMA as a dual antigen approach induced potent Ag-specific antibody and cellular responses to the 2 tested antigens in animals.⁹⁶ In patients with advanced prostate cancer, IM and ID delivery of 900 μg PSA DNA vaccines in combination with subcutaneous delivery of GM-CSF and IL-2 induced Ag-specific humoral and cellular responses.⁹⁷ In 41% of the treated prostate cancer patients, the ID injection of PAP DNA vaccines plus GM-CSF cDNA displayed Ag-specific CD4+/CD8+ T cell proliferative responses and the increased PSA doubling time.⁹⁸ Moreover, the booster ID injections of PAP DNA vaccines plus GM-CSF cDNA to the immunized patients have been shown to induce cytolytic T cell responses in 58% patients.⁹⁹ Furthermore, in patients with recurrent prostate cancer, IM-EP of DNA vaccines expressing the tetanus toxin DOMmain-conjugated PSMA CTL epitopes showed anti-DOM IgG production and CD4+ T cell responses, as well as PSMA-specific CD8+ T cell responses with the increased PSA doubling time.^44,45 In patients with Ag-positive solid tumors, PRAME+PSMA fragment DNA prime-peptide boost vaccines (by intranodal delivery) induced Ag-specific T cell responses in 63 melanoma patients and showed stable disease in 7 patients, but no partial or complete responses.²¹ In that study, some patients showed decreased PSA levels. These promising data warrant further evaluation of IM-EP and intranodal delivery of PSA, PSMA and PAP DNA vaccines in treating patients with prostate cancer.

Cancer/testis tumor-associated antigens

Cancer/testis antigens, such as MAGE and NY-ESO-1 are expressed in many cancer cells but, among the normal tissues, are solely expressed in the testes.¹⁰⁰ In a murine model, DNA vaccines encoding HSP70/MAGE-3 fusion proteins were immunogenic and conferred potent antitumor immunity against MAGE-3-expressing tumors.¹⁰¹ Particle-mediated ID delivery of NY-ESO-1 DNA vaccines into 16 patients with advanced cancer [diagnosed as prostate adenocarcinoma (10 patients), non-small cell lung cancer (NSCLC; 5 patients) or esophageal carcinoma (1 patient)] also led to the induction of Ag-specific CD4 and/or CD8 T cell responses without any clinical benefits in most of the tested patients.¹⁰² In view of this finding, an association between NY-ESO-1-specific immune responses and clinical responses remains to be determined.

Cancer differentiation antigens

The proteins MART-1 (also called Melan-A), gp100 and Trp2 are differentiation antigens that are mainly expressed in melanoma cells and normal melanocytes.^103-105 DNA vaccines expressing human gp100 and Trp2 have shown antitumor protective and therapeutic activities against B16 melanoma in mice,¹⁰⁶ suggesting that a xenogeneic antigen might be effective at inducing cross-reactive antitumor responses to self-antigen-expressing tumor cells. Consistent with this finding, human Trp2-encoding DNA vaccines provided significant protection against the metastatic growth of B16 melanoma in mice through the induction of cross-reactive antibody and CD8+ T cell responses.¹⁰⁷ Intradermal and intramuscular delivery of gp100 failed to show any clinical and immune responses in melanoma patients.¹⁰⁸ Particle-mediated epidermal delivery of cDNA coding for gp100 and GM-CSG to uninvolved skin of melanoma patients resulted in modest activation of anti-melanoma responses while showing its safety and tolerability.¹⁰⁹ However, the intranodal delivery of tyrosinase epitope DNA vaccines induced Ag-specific immune responses in 42% patients without any clinical responses while overall survival was in favor of immune responders.¹¹⁰ In patients with stage IV melanoma, moreover, the intranodal delivery of MART-1-expressing DNA vaccines resulted in induction of Ag-specific immune responses without any clinical benefits.²⁰

In the treatment of hematological malignancies, the induction of long-term memory immune responses is likely critical for preventing cancer relapse. Idiotype immunoglobulin has also been identified to have unique protein sequences in the variable regions of B cell immunoglobulin.^111,112 Thus, the idiotype of B-cell malignancy might serve as a tumor-specific antigen. Immunization with DNA vaccines encoding chemokine-fused short idiotype single-chain (sFv) polypeptides containing only the antigenic variable regions of light and heavy chains resulted in the induction of Ag-specific immune responses and antitumor protective effects in a T cell-dependent manner in mice.¹¹³ When intramuscularly injected into patients with follicular B-cell lymphoma, DNA vaccines encoding chimeric idiotypes were also immunogenic.⁶⁷ In this context, differentiation antigens represent promising targets that can be employed to induce long-term tumor-specific CTL immunity for controlling the relapse of hematological malignancy.

A mutated antigen, p53

The p53 protein plays a role in cell cycle regulation, genomic stability, angiogenesis and apoptosis.^114,115 In cancer cells, p53 is often mutated,¹¹⁶ and is thus thought to be a potential target for cancer immunotherapy. In an animal model, IM-EP of a DNA vaccine encoding mutated human p53 displayed mouse p53 cross-reactive CD8+ T cell responses and antitumor protective and therapeutic activities in a CD8+ T cell-dependent manner,¹¹⁷ suggesting that mutated p53 might also represent a useful target for cancer immune therapy.

Cancer-specific antigens (HPV E6 and E7)

HPV infection is a major cause of oral cancer, as well as anogenital cancers, including cervical cancer (reviewed in Ref. 118). HPV E6 and E7 proteins are known to be associated with oncogenesis through the inhibition of the functions of p53 and Rb, which are cell cycle regulatory proteins.^119,120 In this context, HPV E6 and E7 proteins have been tested as immunization targets for HPV-associated cancers. In numerous animal studies, DNA vaccines encoding HPV E6 and E7 cDNAs have shown to induce robust antigen-specific antibody and cellular responses.^{41-43,121,122} A clinical trial using IM-EP of HPV E6/E7 DNA vaccines showed that these vaccines were able to induce robust Ag-specific humoral and cellular responses (CTL, in particular).³⁴ A recent clinical trial further demonstrated the clinical efficacy of IM-EP of HPV E6/E7 DNA vaccines in reducing the viral load and precancerous disease severity (personal communication with Bagarazzi M). Therefore, it is likely that HPV E6 and E7 DNA vaccines might be applicable in treating HPV-associated oral and anogenital cancers, as well as in treating precancerous diseases.

Tumor immune evasion

Because cancer cells are somewhat different from their normal counterparts, these cells can possibly be recognized and eradicated by the immune system. However, cancer cells are heterogeneous and employ many different mechanisms to evade host immune surveillance (reviewed in¹²). Clarification of the immune escape mechanism(s) utilized by cancer cells appears to be important for developing counter-measures for this immune escape. Table 3 summarizes the immune evasion mechanism(s) employed by cancer cells and the cancer microenvironment.

Table 3. Immune evasion mechanisms employed by tumor cells and the tumor microenvironment

Tumor cell-mediated mechanism(s)
TGF-β production	Inhibition of effector T cell functions	^123
Indoleamine 2,3-dioxygenase expression	Depletion of tryptophan pool against T cells	^124,125
PD-L1 expression	Suppression of effector T cells	^126,127
p53 mutation	Deregulation of cell cycles and survival	^130
Serine protease inhibitor PI-9/SPI-6 expression	Inhibition of granzyme B function	^131
cFLIP over-expression	Blocking of death receptor-dependent tumor cell lysis pathway	^132
Antigen epitope alteration	Escape from effector T cell recognition	^63,133
Antigen loss	Escape from effector T cell recognition	^134,135
Downregulation of MHC class I	Escape from effector T cell recognition	^136
Overexpression of anti-apoptotic molecules	Survival from CTL-mediated apoptosis	^137
Immune cell-mediated mechanism(s)
Tumor-associated macrophages	Suppression of effector T cells	Reviewed in^138
Regulatory T cells	Suppression of effector T cells	Reviewed in^139
Myeloid-derived suppressor cells	Suppression of effector T cells	Reviewed in^140

Tumor cells release TGF-β, which acts on tumor Ag-specific CTLs and inactivates the functions of perforin, granzyme B, Fas ligand, and interferon-γ, which are the effector molecules of CTLs for tumor cell killing.¹²³ Tumor cells also express indoleamine 2,3-dioxygenase, which exploits the tryptophan pool and induces tryptophan deficiency to prevent T cell proliferation.^124,125 In addition to these immune inhibitory molecules, tumor cells utilize immune-inhibitory ligands on the cell surface as a mechanism to resist CTL-mediated tumor cell destruction. Cancer cell-associated B7-H1 (also called PD-L1) induces the apoptosis of cancer-specific T cells in vitro and promotes the growth of tumor cells in animals.¹²⁶ The infusion of anti-B7-H1 Abs to block B7-H1 resulted in the increased efficacy of adoptive T cell therapy for curing mice with tumors.¹²⁷ Furthermore, in PD-1 deficient mice, tumor growth was suppressed in a manner similar to that of the application of anti-PD-L1 Abs,¹²⁸ supporting a dominant role for PD-1/PD-L1 in local immune suppression within the tumor microenvironment. These data suggest that PD-L1 expressed on tumor cells might contribute to the suppression of tumor-reactive CTLs. In patients with advanced cancers, including non-small-cell lung cancer, melanoma, and renal-cell cancer, blocking PD-L1 using anti-PD-L1 Abs induced sustained tumor regression and prolonged disease stabilization.¹²⁹

The mutation of p53 was suggested as a possible mechanism for the insensitivity of tumor cells to tumor-reactive CTLs.¹³⁰ In that study, the restoration of mutant p53 function by infecting the cells with adenovirus expressing wild type p53 resulted in increased sensitivity to CTL-mediated lysis by the Fas-mediated pathway. The serine protease inhibitor PI-9/SPI-6 (inactivator of granzyme B) is also associated with tumor cell insensitivity to Ag-specific CTL-mediated killing by blocking granzyme B proteins.¹³¹ cFLIP over-expression is also responsible for tumor cell resistance to CTL-mediated lysis by blocking the death receptor-dependent tumor cell lysis pathway.¹³² Moreover, tumor cells do not present antigens on the surface of tumor cells in the context of MHC class I antigens and thus remaining unrecognizable by tumor-reactive CTLs.¹³³ Consistent with this, we recently reported that MC32 colon cancer cells could evade CEA-specific CTLs by the loss of antigen expression on the tumor cell surface in an MC32 colon cancer model.⁶³ In addition, tumor cells lose an antigen for CTL recognition.¹³⁴ We also found that MC32 cells remain without antigen expression under severe immune selection conditions, thereby evading Ag-specific CTLs.¹³⁵ The down-regulation of MHC class I molecules is also a frequently observed mechanism for immune escape in cancer cells.¹³⁶ Furthermore, the overexpression of anti-apoptotic molecules causes tumor cells to become insensitive to CTL-mediated tumor cell killing.¹³⁷ Thus, reversing tumor cell-driven immune dysfunction is required to achieve CTL-mediated tumor destruction.

The tumor microenvironment often contains immunosuppressive cells that secrete immunosuppressive molecules and limit the functions of CTLs. For instance, tumor-associated macrophages promote tumor cell growth and metastasis and inhibit effector T cell activity, whereas removing this cell population enhances anti-cancer therapy (reviewed in Ref. 138). Regulatory T cells are also involved in inhibiting T cell function by many different strategies, but the suppression of these cell populations results in improved vaccine immunogenicity (reviewed in Ref. 139). Myeloid-derived stem cells are also known to suppress T cell responses and to induce tolerance to cancer-associated antigens (reviewed in Ref. 140). Blocking or removing these negative regulatory immune cells is one additional potential method to improve the immunogenicity of cancer DNA vaccines and the reactivity of Ag-specific CTLs to cancer cells. Moreover, if one antigen is lost or an antigen epitope presentation is altered during long-term immune therapy, it might be useful to use more than one cancer antigen as a pre-measure to the loss of one antigen. Combined therapy using cancer DNA vaccines and other therapy protocols with different antitumor killing mechanisms might also be required for obtaining better clinical efficacy of cancer DNA vaccines. This is based upon our previous findings that cisplatin-based chemotherapy and E7 vaccination synergized antitumor therapeutic activity by rendering tumor cells more sensitive to Ag-specific CTL-mediated killing and by eliciting long-term memory responses for preventing tumor recurrence in an animal model.¹⁴¹ A similar finding was obtained with radiotherapy plus E7 cancer vaccination,¹⁴² as well as chemotherapy plus adoptive therapy using Ag-specific CTL effector cells.¹⁴³ In addition, during treatment with cancer DNA vaccines, blocking or inhibiting the negative molecules such as TGF-β, indoleamine 2,3-dioxygenase, PD-L1, SPI-6, cFLIP, anti-apoptotic molecules and others involved in cancer cell immune escape likely plays an important role in augmenting cancer antigen-specific CTL and anti-neoplastic responses. Taken together, these multimodal treatment approaches are promising to limit the escape of tumor cells and simultaneously achieve cancer eradication.

Conclusion

During the last 2 decades, the possibility of using cancer DNA vaccines as a means to induce therapeutically robust cancer-specific CTL responses has been intensively evaluated. However, the clinical benefits of such vaccines in treating cancer patients have yet to be demonstrated. This lack of clinical benefits is likely attributed to a number of biological factors. For instance, cancer antigens are generally of low immunogenicity due to self-tolerance, and cancer cells employ different mechanisms to escape host immune surveillance. The applications of immune-enhancing strategies in cancer DNA vaccination might help to overcome immune tolerance and to induce robust cancer-reactive immune responses, which are required for cancer eradication. At the same time, blocking or removing negative immune cells and molecules likely leads to the development of more effective anticancer DNA vaccines for treating patients with cancers. Moreover, cancer DNA vaccines have been tested on subjects with very late stage cancer; this might be the reason why the vaccines have not exhibited any cancer-eradicating effects.

Combined therapy using cancer DNA vaccines and other strategies that limit the emergence of immune-evading cancer cells and simultaneously induce robust cancer-specific CTL responses is expected to be of benefit to cancer patients.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Funding

This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2012R1A1A2038430).