| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
REVIEWS |
a Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA
ABSTRACT
Genetic immunization, also known as DNA or polynucleotide immunization, is a novel strategy for vaccine development in which plasmid DNA encoding either individual or a collection of antigens is directly administered to a host. Such immunization leads to host expression of the delivered foreign gene, resulting in the induction of a specific immune response against the in vivo produced antigen. DNA immunization has been shown to induce protective immune responses in several infectious disease and cancer experimental model systems. Furthermore, DNA vaccines have recently entered the clinic for analysis as both prophylactic and therapeutic agents. Although the mechanisms of immunity to DNA have not yet been fully elucidated, it has become apparent that the immune response achieved by DNA vaccination is quite malleable, and can be manipulated by altering the conditions under which the vaccine is administered. Either through changing the method or location of immunization, altering the number of immunostimulatory sequences in the plasmid, altering the immunization regimen, or coadministering genes for cytokines or costimulatory molecules, one can modulate both the magnitude and orientation of the subsequent immune response. Through maximization of this feature of DNA immunization, we will likely be able to design vaccines and immunotherapeutic agents that are tailored to the correlates of protection for a particular disease, resulting in a new generation of more focused and effective immune stimulating agents.—Cohen, A. D., Boyer, J. D., Weiner, D. B. Modulating the immune response to genetic immunization. FASEB J. 12, 1611–1626 (1998)
Key Words: immune modulation • plasmid • DNA vaccine • CpG motifs • cytokines
BACKGROUND
Despite the marked advances in public health measures and antimicrobial medications over the last half-century, infectious diseases remain one of the leading causes of morbidity and mortality worldwide. Age-old diseases such as tuberculosis and malaria still afflict millions each year, while newer pathogens like human immunodeficiency virus (HIV)2 continue to spread in epidemic proportions in many parts of the world. Even though effective therapies for many of these diseases now exist, in the majority of cases their prohibitive cost and relative lack of availability in developing countries makes their widespread application untenable. For these reasons, the most powerful and cost-effective way to control infectious disease remains prophylactic vaccination. Traditional vaccination strategies such as live, attenuated or whole, inactivated agents have been very successful in the past. However, for many microorganisms that still lack an effective vaccine, these strategies may not be appropriate, either due to safety issues or lack of immune potency. Therefore, it is imperative that new approaches to vaccine development continue to be explored.
Genetic immunization, also known as DNA or polynucleotide immunization, represents a novel approach for achieving specific immune activation. It has been known for decades that delivery of naked DNA into an animal could lead to in vivo gene expression (reviewed in ref 1). It was not until 1992–1993, however, that the first publications on DNA immunization (2–6) focused new attention on the importance of this approach for vaccine and immune therapeutic development. The concept behind genetic immunization is a simple one: genes encoding an antigen (or antigens) specific to a particular pathogen are cloned into a plasmid with an appropriate promotor, and the plasmid DNA is administered to the vaccine recipient. The DNA is taken up by host cells and the gene is expressed. The resultant foreign protein is produced within the host cell and then processed and presented appropriately to the immune system, inducing a specific immune response. Immunization with DNA thus mimics aspects of live infection, with pathogen protein(s) synthesized endogenously by host cells. This synthesis leads to the induction of a cytotoxic T lymphocyte (CTL) response via the major histocompatibility complex (MHC) class I-restricted pathway. Concurrently, protein(s) are released extracellularly. It is believed that this exogenously released antigen primes the induction of a humoral response, as well as a helper T lymphocyte (Th) response via MHC class II-restricted antigen presentation by antigen-presenting cells (APCs) that have taken up the foreign antigen. In this way genetic immunization confers the same broad immunological advantages as immunization with live, attenuated vectors without the accompanying safety concerns associated with live infection, concerns that include vaccine reversion to more virulent pathogen forms in immunocompromised patients and spreading of infectious vaccine to unintended populations. DNA vaccines are also likely to be attractive from a health economics perspective: they are relatively easy to manufacture in large quantities and do not require any special transportation or storage conditions that could hinder their widespread distribution.
The first reports of DNA immunization that served to focus the field were made in late 1992 and early 1993. Tang et al. (2) reported that inoculation of mice with gene gun-delivered plasmid DNA induced an antibody response to the encoded protein. Almost concurrently, it was demonstrated that DNA vaccines could generate immune responses against influenza (3, 4), HIV-1 (5), and hepatitis B (6). These early studies had several important features: they showed that DNA vaccines could induce immunity in several different disease models; that antibody, Th, and CTL responses could all be generated; that a response could be induced by different routes of immunization (e.g., intramuscular, epidermal, or mucosal); and, in the case of influenza, that these vaccines could protect animals from subsequent challenge with pathogenic virus. Since then, genetic immunization has been investigated in numerous infectious disease models in addition to those mentioned above, with vaccines currently in various stages of development for HIV-2 (7, 8), herpes simplex virus 1 (HSV-1) (9), HSV-2 (10), rabies (11), hepatitis C (12), tuberculosis (13), malaria (14), mycoplasma (15), Leishmania major (16), cytomegalovirus (17), Toxoplasma gondii (18), rotavirus (19), and, most recently, Ebola (20). In addition to infectious diseases applications, genetic immunization may have potential as a means of cancer immunotherapy. Injection of plasmid DNA encoding tumor-associated antigens has been shown to cause a tumor antigen-specific immune response, and can protect mice from lethal tumor challenge (21, 22).
PRIMATE STUDIES
The generation of immune responses through genetic immunization has not been limited solely to rodents, however. DNA vaccines have induced antigen-specific immunity in monkeys, chimpanzees, and humans. African green monkeys immunized with plasmids encoding the influenza virus hemagglutinin and nucleoprotein (NP) proteins developed hemagglutination-inhibiting antibody titers equal to or greater than those induced by commercially available whole inactivated influenza virus vaccines. Subsequent boosting with the same DNA constructs drove titers even higher, against both homologous and drifted strains of virus. These vaccines, however, were not protective, though they did alter the time course of infection (23, 24). DNA vaccines encoding simian immunodeficiency virus (SIV) proteins have elicited neutralizing antibodies and CTL responses in rhesus macaques, although these vaccines also did not protect animals from pathogenic viral challenge (25, 26). Immune responses to HIV-1 antigens have been generated in macaques by using genetic immunization (27, 28), and multicomponent HIV-1 DNA vaccines have successfully protected macaques from challenge with chimeric SHIV virus (29). Protection from SHIV challenge has also been achieved by an immunization regimen that primed with an HIV-1 env DNA vaccine and boosted with recombinant gp160 protein (30). In addition, immune responses to HIV-1 DNA vaccines have begun to be investigated in the chimpanzee model. Immunization with DNA vaccines encoding HIV-1 env, rev, and gag/pol induced antigen-specific humoral and cellular immune responses and prevented the establishment of infection in two of two vaccinated animals challenged with a high-dose heterologous strain of HIV-1 (31). The potential role of genetic immunization as immunotherapy has also been examined. Immunization of HIV-infected chimpanzees with an HIV-1 env/rev construct led to significant boosting of env-specific humoral responses and a decline in viral load in immunized animals (32).
DNA vaccines have recently entered the clinic for initial safety and immunogenicity testing in humans. Currently, phase I/II trials are under way using DNA immunogens as potential immunotherapies for cancer, including carcinoembryonic antigen (CEA) for colon cancer and a T cell receptor Vß for cutaneous T cell lymphoma. Clinical trials using DNA vaccines are also under way for herpes, influenza, hepatitis B, HIV, and malaria. In a trial of 15 previously untreated HIV-infected volunteers receiving three intramuscular doses of 30, 100, or 300 µg of an HIV-1 env/rev DNA construct, vaccine administration was well tolerated, without significant local or systemic reactions or any significant laboratory abnormalities. Boosting of env-specific antibodies was observed in the 100 and 300 µg dose group, though no consistent effect on cellular responses to HIV was noted. CD4/CD8 lymphocyte levels and plasma HIV concentration remained relatively unchanged throughout the study (33). Another trial in nine asymptomatic HIV-infected humans used plasmids encoding the HIV-1 regulatory genes rev, nef, and tat. Enhanced HIV-specific cellular responses were seen in several patients without consistent changes in lymphocyte subsets or viral load. Again, no local or systemic toxicity was noted (34). Similarly, a trial of DNA vaccination in HIV-infected volunteers on highly active antiretroviral therapy has commenced in an effort to examine whether better immune responses to HIV-1 DNA vaccines could be achieved in infected humans if viral replication is well controlled. In addition, over 80 HIV-seronegative volunteers have been enrolled to date in clinical trials testing both env/rev and gag/pol DNA vaccines (
Table 1).
Malaria trials are progressing as well: a DNA vaccine for malaria has recently been reported that was well tolerated and induced cellular immune responses in human volunteers [Third National Symposium on Basic Aspects of Vaccines, 1998]. The next year or two should reveal in fine detail the potency of genetic immunization for inducing immune responses in humans.
|
The evidence supporting DNA vaccines as a conceptually safe and effective means of generating protective immune responses in animals has been reviewed extensively in several recent articles (1, 35–39) and will not be further discussed in detail here. The remainder of this review will focus on a feature of genetic immunization that is currently undergoing a great deal of investigation: namely, how the immune response to a DNA vaccine is generated, and how this response can be manipulated by altering the conditions under which the vaccine is administered. These conditions include the presence or absence of immunostimulatory sequences (ISS) in the plasmid, the method and location of immunization, the form of the immunogen, the immunization regimen, and the presence or absence of coadministered cytokines or costimulatory molecules. As discussed below, by modifying one or more of these conditions investigators have been able to alter both the magnitude and orientation of the immune response, selectively enhancing antibody or cell-mediated responses by steering the Th response toward a particular subtype. This susceptibility of DNA vaccine-induced immunity to manipulation is likely to play an important role in the rational design of DNA-based prophylactic vaccines and immunotherapeutic agents.
MECHANISM OF IMMUNITY
Although the ability of genetic immunization to induce immune responses has been well documented, the mechanism by which these responses arise is less clear. Each of the three arms of the immune system needs to encounter antigen in a different context: antibodies usually bind soluble antigen; CD4 T cells primarily recognize peptide-MHC class II complexes on the surface of APCs that have endocytosed and processed exogenous antigens; and CD8 T cells are generally restricted to peptide-MHC class I complexes derived from endogenously produced protein that has undergone proteosome-dependent intracellular processing. How then are DNA vaccines able to activate all three arms of the immune system? Given the reports of long-term gene expression and protein production in transfected myocytes after intramuscular inoculation (40), it was originally thought that these cells had been engineered to become antigen-presenting cells. Myocytes, which constitutively express class I MHC molecules (41), could induce CD8+ CTL responses through the continuous production and processing of plasmid-derived antigen, whereas soluble protein, either secreted from myocytes or released upon myocyte death, could generate antibody responses or T cell help ( Fig. 1A).
The problem with this hypothesis is that myocytes lack important costimulatory molecules necessary for priming naive CD8 T cells for cytotoxic responses, and thus are unlikely to function as effective APCs (42). Pardoll and Beckerleg (42) have proposed two other mechanisms, both of which involve bone marrow-derived `professional' APCs such as dendritic cells or monocyte/macrophages. These cells express high levels of MHC class I and II molecules, as well as costimulatory molecules such as B7.1 and B7.2, and are highly efficient presenters of antigen to T lymphocytes. The first mechanism presumes that a small number of these professional APCs are directly transfected with the injected DNA ( Fig. 1B). These cells then traffic to regional lymphoid tissue, where they can activate CD4 T cells, B cells, and CD8 T cells (42). This mechanism seems conceivable in epidermal immunization, because the skin is known to contain a relatively high proportion of Langerhans cells, but appears less likely in intramuscular immunization, as dendritic cells and macrophages are thought to be scarce in muscle (41). The second hypothesis proposes that antigen produced by transfected myocytes (in the case of intramuscular inoculation) is transferred to bone marrow-derived APCs that have infiltrated the muscle as part of an inflammatory response to the immunization procedure ( Fig. 1C). With epidermal immunization, it is the transfected keratinocytes that would be producing the antigen and transferring it to dendritic cells located in the skin. The transferred protein would then cross over into the MHC class I-restricted processing pathway, allowing the APC to prime CTL responses. Though this would seem to contradict the dogma that only endogenously produced proteins can enter the MHC class I pathway, it would be consistent with the recent reports by Bevan and colleagues, Rock and colleagues, Srivastava and colleagues, and others (42–46) who have described how exogenously produced protein, particularly in particle form, can be taken up by APCs and then presented in the context of MHC class I molecules.
|
Corr and colleagues (47) performed a set of experiments designed to determine whether myocytes or bone marrow-derived APCs were responsible for priming CTL responses after intramuscular DNA immunization. They generated two sets of parent F1 bone marrow chimeras in which recipient H-2bxd F1 mice whose immune systems had been destroyed by irradiation received either H-2b or H-2d T cell-depleted bone marrow. Thus, these mice had myocytes whose MHC molecules expressed both parental haplotypes, T cells that could recognize both parental haplotypes (due to development within an F1 thymus), and bone marrow-derived APCs whose MHC molecules were only from one parental haplotype. These chimeras were then immunized twice with a plasmid encoding the influenza NP, which contains epitopes for both parental haplotypes; CTL assays were carried out 6 wk later. The authors found that NP-specific CTL responses were restricted only to the haplotype expressed by the bone marrow-derived APCs and not to the other haplotype expressed by the rest of the chimera's cells, including myocytes. The findings were the same for both sets of chimeras. Similar findings using bone marrow chimeric mice were reported for intramuscular immunization by Doe et al. (48) and for epidermal gene gun immunization by Iwasaki et al. (49). The latter group also showed that even if myocytes were cotransfected with a plasmid encoding a costimulatory molecule along with the influenza NP plasmid, they were still unable to prime CTL responses (49). These studies strongly suggest that bone marrow-derived cells are responsible for priming CTL responses after genetic immunization, and that nonhematopoietic cells such as myocytes or keratinocytes are not being converted into APCs by naked plasmid transfection.
What the above studies do not determine, however, is the mechanism by which bone marrow-derived APCs obtain antigen for presentation to CD8+ CTL, i.e., through direct transfection by plasmid DNA (
Fig. 1B) or through uptake of antigen released from transfected myocytes or keratinocytes ( Fig. 1C). Evidence supporting the former idea comes from the Bona laboratory (50), which recently reported that plasmid DNA could be isolated from lymph node-derived and skin-derived dendritic cells after intramuscular and intradermal immunization, respectively, demonstrating that direct transfection of APCs does occur. They also demonstrated that after intramuscular or intradermal immunization in mice, only dendritic cells derived from regional lymph nodes were capable of presenting antigenic epitopes to antigen-specific T cells (50). Chattergoon et al. (51) have demonstrated that direct transfection of macrophages and dentritic cells can occur after intramuscular DNA immunization, and have shown that these transfected cells are highly activated and can be found in the peripheral blood. Further support comes from a study by Torres et al. (52), who reported that complete surgical ablation of injected muscle as soon as 10 min after DNA immunization did not prevent the induction of antigen-specific antibody or CTL responses and that these responses were no different in magnitude or longevity than those in mice whose muscle was not ablated. This is somewhat surprising, as it suggests that the injected myocytes play almost no role in inducing the immune response, not even as an `antigen factory' for antibody induction. Nonetheless it supports the idea that direct transfection of APCs (in this case, distant from the injection site) may be the means whereby these APCs induce cellular immune responses.
Support for the `transferred antigen' hypothesis exists as well, however. Ulmer and colleagues (53, 54) have reported that transplantation of myoblasts stably transfected with influenza NP DNA can induce both anti-NP antibodies and NP-specific, MHC class I-restricted CTL responses in recipient mice. These immune responses were similar in magnitude to those induced by intramuscular DNA immunization, and conferred protection from a lethal challenge with influenza A virus. As there was obviously no free plasmid available to transfect specialized APCs, these studies argue against the `direct transfection' hypothesis, and suggest that transfer of antigen from the transfected myoblasts to APCs for class I-mediated processing was the mechanism of CTL priming. These results are consistent with previous transplantation experiments using transfected tumor cells, in which presentation of class I-restricted tumor antigens was limited to host bone marrow-derived APCs, again implying in vivo transfer of antigen from transfected cells to APCs (43, 55). Corroboration of these transfected myoblast transfer experiments using different antigens, especially those that, unlike influenza NP, are not normally secreted as particles, would provide important confirmation of this hypothesis. It is possible that both mechanisms play a role in the generation of immunity through genetic immunization, and likely that as-yet-unimagined mechanisms may also be contributing.
IMMUNITY-MODIFYING FACTORS
CpG motifs and immunostimulatory sequences
One aspect of genetic immunization that has recently received significant attention is the immunostimulatory activity of DNA itself. It has been well demonstrated that DNA from bacteria, but not vertebrates, can induce a nonspecific immune response (reviewed in ref 56). It has been proposed that this is due to differences in the frequency of unmethylated cytosine-phosphate-guanine (CpG) dinucleotides found in the two genomes. While this dinucleotide appears in bacterial DNA at the expected frequency of 1 in 16, it is 10- to 20-fold less frequent in vertebrate DNA, a phenomenon termed `CpG suppression'. In vertebrates the majority of these dinucleotides contain a methylated cytosine, whereas in bacteria they are unmethylated (56). Krieg and colleagues (57) showed that oligonucleotides containing one or more CpG dinucleotides could trigger B cell proliferation and immunoglobulin secretion. Immune activation was optimized when the dinucleotide was flanked by two 5' purines and two 3' pyrimidines, forming a `CpG motif' or `immunostimulatory sequence'. This immune response was not seen with otherwise identical oligonucleotides that did not contain CpG motifs or whose CpG dinucleotides had been methylated by CpG methylase (57). Bacterial DNA can also stimulate natural killer (NK) cell activity and cellular immunity through the induction of inflammatory cytokines such as interferons (IFNs) and interleukin 12 (IL-12) (56). The relevance of these findings for genetic immunization became clear when Sato et al. (58) reported that a DNA vaccine whose plasmid backbone contained CpG ISS induced a more vigorous antibody and CTL response than an otherwise identical vaccine that did not contain the ISS, despite a higher level of gene expression produced by the latter plasmid. The immune response to the vaccine lacking the ISS could be restored to normal levels by coinjecting it with noncoding plasmid containing ISS; increasing the amount of coinjected noncoding plasmid increased the antibody response accordingly (58). This study demonstrated that the plasmid vector itself could have a significant adjuvant effect on DNA vaccine-induced immunity, and suggests that this effect may be mediated by CpG-containing ISS.
Subsequent studies have confirmed that CpG motifs can enhance immunity to genetic immunization (59, 60) and have shown that they can qualitatively modify the immune response by preferentially inducing a type 1 helper T lymphocyte (Th1) response. ISS-containing plasmids or oligonucleotides induce production of IL-2, IFN-
, and IL-12 (but not IL-4), increase the IgG2a:IgG1 ratio, and decrease immunoglobulin E (IgE) antibody production, all characteristic of a Th1 response (58–61). The Th1-promoting ability of plasmid DNA was further demonstrated by two recent studies in which mice were injected intradermally or intramuscularly with both plasmid DNA and a protein antigen. While protein vaccination typically induces a predominantly Th2 response, in both studies coimmunization with ISS-containing DNA shifted the immune response back toward a Th1 or Th0 response, with elevated IgG2a and IFN- production (60, 61). This Th1-inducing activity of bacterial DNA may be a reason why most DNA vaccines studied to date induce a predominantly Th1 response when injected intramuscularly or intradermally.
The mechanism by which CpG motifs trigger an immune response is not clear and merits further study. There is no evidence for a membrane CpG receptor, and there does not appear to be any difference in the cellular uptake of plasmids or oligonucleotides with and without CpG motifs (56). Therefore, it is likely that any specific recognition of the motif would be carried out by an intracellular protein, which could then lead to transcriptional regulation that effects the subsequent immune activation. Further investigation is also called for in assessing for any possible harmful effects of immunostimulatory DNA. At least one group has demonstrated a potential danger, reporting that intratracheal administration of bacterial DNA and CpG-rich oligonucleotides in mice resulted in marked lung inflammation (62). Although this finding raises some concern (63), it is still minimal at this point, as DNA vaccines administered by conventional routes (i.e., skin or muscle) have yet to induce any significant adverse effects. Finally, all of the data about CpG motifs and DNA vaccines reported thus far have been generated in rodents; it will be most interesting to see whether a similar immunostimulatory effect is seen in primates. Thus, the discovery that plasmid DNA itself, and CpG-rich immunostimulatory sequences in particular, can have potent immune-boosting and immune-modifying activity has provided another variable that can be used and manipulated in designing constructs for genetic immunization.
Method and location of immunization
One feature of genetic immunization that has become apparent over the past few years is that the way a DNA vaccine is delivered may have an effect on the type of immune response generated. Specifically, it appears likely that both the site of inoculation and the method by which the plasmid is delivered may independently affect the induced immunity in a qualitative, if not quantitative, manner. Successful DNA vaccination has been demonstrated via a number of different routes, including intravenous, intramuscular, intranasal, intraepidermal, intravaginal (3–7, 64, 65), and, more recently, intrasplenic (66) and intrahepatic (67). The majority of DNA vaccine studies so far, however, have used either skin or muscle as their immunization targets. Plasmid delivery at these sites is usually accomplished by one of two methods: needle injection of DNA suspended in saline or in a saline mixture containing a facilitator, such as bupivacaine, designed to enhance DNA uptake; or gene gun-mediated acceleration of DNA-coated microprojectiles directly into the cells of the target tissue. Both methods have been used in skin and muscle, although the gene gun has been more commonly used for epidermal rather than intramuscular administration. Multiple studies have reported that gene gun-mediated immunization is far more efficient than needle injection, eliciting similar levels of antibody and cellular responses with 100- to 5000-fold less DNA (5, 68, 69). Pertmer et al. (68) reported that as little as 16 ng of plasmid DNA delivered epidermally via gene gun could induce antibody and CTL responses in mice, whereas intradermal or intramuscular injection of the same plasmids required 10–100 µg of DNA to elicit comparable responses. Despite this difference in efficiency, however, there has been no compelling evidence that gene gun-based immunization leads to longer lived responses or greater protection from pathogenic challenge, even when higher doses of plasmids are used (70).
What has become evident is that the way a DNA vaccine is administered can affect the T helper cell profile that is ultimately generated. Upon activation, CD4+ T helper lymphocytes differentiate from precursor Th0 cells into two functionally distinct subsets ( Fig. 2).
Type 1 (Th1) cells activate macrophages and induce cell-mediated immunity, including CTL responses, whereas type 2 (Th2) cells primarily induce humoral immunity. Intramuscular needle injection of DNA produces a predominantly Th1-type response, with an elevated IgG2a:IgG1 ratio, IFN- production, and little IL-4 production (69, 71, 72). In contrast, epidermal gene gun inoculation generally induces a Th2 phenotype with successive immunizations, generating mostly IgG1 antibodies, less IFN-, and more IL-4 (69, 72, 73). Intradermal injections of DNA have been reported to induce both Th1 (69, 74) and Th2 (72) profiles, while intramuscular gene gun inoculation seems to generate a profile similar to epidermal gene gun inoculation ( Table 2)
(69). The primary immunization mode in some cases appears to irreversibly determine the profile produced, as subsequent DNA immunizations using the alternative method, subsequent immunizations using the protein encoded by the genetic vaccine, or even subsequent viral challenge were all unable to cause a shift from the originally induced Th profile (69, 72, 74). This profile can be shifted, however, through the coadministration of genes encoding various cytokines, a topic discussed in detail in the next section. Prayaga et al. (75) demonstrated that the Th profile induced after epidermal gene gun immunization with a plasmid encoding HIV-1 gp120 or influenza nucleoprotein, normally of the Th2 type, became more Th1-like when mice were coimmunized with plasmids encoding IL-2, IL-7, or IL-12. This suggests that the inherent bias of a particular immunization method can be overcome under certain conditions, allowing for greater manipulation of vaccines or immunotherapeutics toward the specific type of immune response desired.
|
|
The reasons why different T helper subsets arise after different immunization methods are not clear. One possibility is that it may be a dosage phenomenon. As mentioned above, needle injection requires a much greater amount of DNA than gene gun inoculation. This large amount of bacterially derived DNA, with its inherently Th1-inducing immunostimulatory sequences, may be one reason why injected DNA vaccines induce a Th1 response whereas gene gun-delivered vaccines, which use much less DNA, do not. Evidence supporting this claim comes from the Johnston lab (70), which has reported that when only a few micrograms of DNA are used, both gene gun and i.m. needle injection induce predominantly IgG1 antibodies, a Th2 response, but when 50 µm of DNA are used, both methods induce more IgG2a antibodies, a Th1 response (54). Conflicting evidence, however, comes from the Robinson lab (69), which reported that even with as little as 1–2 µm of DNA, i.m. injection still elicited predominantly IgG2a antibodies, while gene gun inoculation of the same amount elicited IgG1 antibodies. Thus it remains possible that other factors, such as the host haplotype or the nature of the processed antigen, may be involved. Another explanation may derive from the different way DNA is delivered by these two methods. Needle injection delivers DNA to the extracellular space, where it is taken up into cells by some poorly understood mechanism, perhaps inducing a set of intracellular signals that ultimately leads to the increased production of Th1-type cytokines. The gene gun, on the other hand, directly transfects its target cells with plasmid, bypassing any uptake mechanism and perhaps generating a different set of intracellular signals. This remains an area of active study, and obviously only further research will clarify this issue.
Another area of active exploration is DNA vaccination at mucosal sites. The earliest study of mucosal genetic immunization used a plasmid encoding the influenza hemagglutinin protein. Two intranasal administrations of this plasmid to mice primed serum IgG antibody responses and protected 76% (13/17) of immunized mice from lethal influenza challenge (5). Subsequent studies of DNA vaccines given intranasally have used plasmids encoding HIV-1 or herpes simplex virus antigens, and have been successful in inducing immunity both systemically and at distant mucosal sites, with significantly higher titers of vaginal and fecal secretory IgA than those elicited by intramuscular immunization with the same plasmid. The immune responses generated appear to be primarily of the Th2 type, as determined by IgG subclass ratios and cytokine production, and can be boosted by coadministration of cholera toxin as an adjuvant (76). Intranasal immunization with a HSV-1 DNA vaccine did not, however, protect mice from subsequent vaginal challenge with live HSV-1 (76). Antiviral immunity can be elicited by intravaginal DNA immunization as well: Wang et al. (64) have reported the induction of anti-HIV envelope antibodies (both IgG and IgA) in vaginal secretions after intravaginal immunization of mice and demonstrated that these vaginal secretions have neutralizing activity against homologous cell-free HIV-1. Bagarazzi et al. (65) have extended these studies into a primate model, and have shown that intravaginal DNA vaccination of infant chimpanzees can elicit serum antibodies to HIV-1 envelope protein. Whether mucosal immunization will confer protection from mucosal viral challenge remains to be seen, but it is likely that at a minimum it will serve as a valuable means to supplement and strengthen the immune response generated by systemic immunization.
Form of the encoded antigen
Another parameter that may have an effect on the magnitude and orientation of the immune response is the localization of the encoded protein. Genetic vaccination has successfully produced immune responses to proteins that are cytoplasmically sequestered (e.g., ß-galactosidase), membrane bound (e.g., rabies G-protein), and secreted (e.g., hepatitis B surface antigen). It is possible that one form of a protein may be better at inducing an immune response than another: for example, a secreted antigen may be more effective at generating antibody and CD4+ T helper lymphocyte responses than one that is not secreted. This hypothesis, however, has so far not been borne out in experimental studies. Secreted forms of rabies G-protein and hepatitis B surface antigen (HBsAg) were no better at inducing humoral and cellular immunity than their membrane-bound counterparts in murine studies (71, 77), and the plasmid encoding the secreted rabies protein conferred less protection against lethal viral challenge than did the one encoding a membrane-bound antigen (71). Though studies with other antigens are needed, it has not been conclusively shown that the form of the plasmid-encoded protein plays a critical role in DNA vaccine-induced immune responses.
Immunization regimen
The optimal regimen for administering a DNA vaccine (e.g., the dose, number, and/or frequency of immunizations) is far from determined, and is the subject of considerable debate. Although most researchers would agree that multiple immunizations will likely be necessary to maximize immune responses, the questions of `how many?' and `how often?' have not yet been answered. These are important questions, however, as these factors have been shown to affect the nature of the resulting immune response. In a study using gene gun delivery, immunization of mice with an HIV-1 env DNA vaccine induced strong CTL and weak antibody responses when one, two, or three immunizations were given, but a fourth immunization caused a drop in CTL activity and a marked rise in antibody titers. This shift in the nature of the immune response was accompanied by a shift in cytokine production by antigen-stimulated splenocytes, with declining IFN- and increasing IL-4 production (73). This decline in IFN- production with successive immunizations has also been reported with an influenza nucleoprotein-encoding plasmid (72). The timing of the immunizations may also affect the subsequent immune response: mice vaccinated twice with 3 months between injections generated much higher antibody titers and cytokine production than mice vaccinated two or three times at monthly intervals, suggesting that a longer rest period between immunizations may be immune enhancing (75). Whether these findings will hold true for different disease models or different species remains to be seen.
Another way to modulate the immune response may be to devise a regimen that combines genetic immunization with other, more traditional, forms of immunization. For example, combining a DNA vaccine with a recombinant protein vaccine in a prime boost regimen may elicit both a Th1- and Th2-type profile, thereby maximizing both cellular and humoral immunity to induce more complete protection. We have examined such a regimen in mice with an HIV-2 envelope DNA construct, and found that priming with DNA and boosting with protein significantly enhanced antibody and T cell proliferative responses and increased antibody neutralization activity (7). A prime boost regimen using multiple doses of an HIV-1 env plasmid, followed by a boost of HIV-1 env plasmid plus HIV-1 Env protein, has been evaluated in rhesus monkeys. This protocol induced strong CTL and neutralizing antibody activity in vaccinated animals, and completely protected the animals from i.v. challenge with a chimeric SHIV virus expressing an HIV-1 envelope on a SIVmac backbone. Animals receiving protein vaccination alone were not protected (30). Even though this study used a large number of vaccinations (10) and a small number of animals, and therefore requires repeating, it nonetheless provides a glimpse of the potential power of DNA–protein vaccine combinations. Another prime boost regimen that may be particularly interesting would be to combine poxvirus or other recombinant live vaccines with a DNA-based approach. Such an approach might be expected to broaden cellular immunity. Also interesting to consider would be a multicomponent vaccine that combines protein, live viral vectors, and DNA in an effort to achieve the broadest possible cellular and humoral immunity.
ENGINEERING A BETTER IMMUNE RESPONSE
Cytokines
Of the different ways to modulate the immune response to DNA immunization, the most promising may be through the coadministration of `biological' adjuvants such as cytokines. Cytokines are molecules secreted mainly by bone marrow-derived cells that act in an autocrine or paracrine manner to induce a specific response in cells expressing a particular cytokine receptor. Some cytokines and their actions are listed in Table 3.
Since Raz et al. (78) reported in 1993 that injection of cytokine genes into muscle resulted in the characteristic biological actions of these cytokines in vivo , and could enhance the immune response to a protein antigen, many laboratories have reported that coinjection of plasmids encoding cytokines can have a substantial effect on the immune response to a plasmid-encoded antigen (described below and summarized in Table 4).
|
|
The first studies using cytokine genes as adjuvants to genetic immunization used plasmids encoding IL-2, a potent stimulator of cellular immunity that induces proliferation and differentiation of T cells as well as B cell and NK cell growth (79, 80). Watanabe et al. (81) showed a fivefold increase in antibody responses to a kappa L chain V gene when IL-2 plasmid was coinjected with plasmid encoding the V gene, while Abai et al. (82) reported no increase and a possible decrease in antibody responses when IL-2 plasmid was coinjected with a plasmid encoding influenza NP. However, in the latter experiments, coinjection of a noncoding plasmid also decreased antibody responses, suggesting that the decrease may not have been specifically due to the biologic activity of IL-2. More recently, Chow et al. (77) demonstrated that injection of a vector that encoded HBsAg and IL-2 on the same plasmid induced marked increases of antibody responses and T cell proliferation compared to a plasmid encoding HBsAg alone, and enhanced T cell production of IL-2 and IFN-. Similar results were reported for hepatitis C virus (HCV) core protein by Geissler et al. (12), who also showed augmented CTL responses with IL-2 gene coinjection. Taken together, these results suggest that IL-2 gene coinjection can increase both humoral and cellular immunity to a plasmid-encoded antigen and enhance Th1 cytokine production.
IL-4 induces differentiation of T helper cells into the Th2 subtype, enhances B cell growth, and mediates Ig class switching (80, 83). Injection of plasmid encoding IL-4 3 days before immunization with a protein antigen increased antigen-specific antibody levels compared to protein immunization alone, but had no effect on the delayed-type hypersensitivity (DTH) response in antigen-challenged animals (78). Coinoculation of mice with plasmids encoding IL-4 and HCV core protein resulted in augmented antibody and T cell proliferation responses, but decreased specific CTL responses compared to HCV plasmid alone (12). Likewise, intranasal administration in liposomes of plasmids encoding HIV-1 env/rev and IL-4 resulted in higher antibody responses but diminished DTH and CTL responses in mice (84). In a tumor challenge model, mice with established pulmonary metastases who received recombinant IL-4 protein after gene gun administration of a DNA vaccine encoding the appropriate tumor-associated antigen failed to demonstrate any reduction in number of metastases, whereas mice who received recombinant IL-2, IL-6, IL-7, or IL-12 all had significant improvement in their tumor burden (85). Thus, while IL-4 may enhance the humoral response after genetic immunization, its relative inhibition of Th1-mediated responses may limit its usefulness as an adjuvant in viral or tumor vaccines or immunotherapy.
Other cytokines that have been studied in concert with genetic immunization include granulocyte-monocyte colony-stimulating factor (GM-CSF) and IL-12. GM-CSF increases the production of granulocytes and macrophages, and induces the maturation and activation of APCs such as dendritic cells (79, 86, 87). In theory, coexpression of GM-CSF and a plasmid-encoded antigen could augment the host's immune response against the antigen by expanding the pool of activated APCs at the injection site. Xiang and Ertl (88) tested this theory in vivo by coinoculating mice with plasmids encoding GM-CSF and rabies glycoprotein. Coexpression of GM-CSF increased antibody responses to rabies glycoprotein in a dose-dependent manner and enhanced T helper cell responses compared to injection with rabies glycoprotein vector alone, whereas coinjection of IFN- gene had no enhancing effect. GM-CSF also improved the efficacy of the rabies glycoprotein plasmid in protecting from lethal virus challenge. Subsequent studies of GM-CSF plasmid coinoculation with DNA vaccines against HIV-1 (89), influenza (90), encephalomyocarditis virus (91), and HCV (12) have confirmed the boosting effect this cytokine has on both humoral and cellular responses to plasmid-encoded antigens. This effect may be dependent, however, on the route of immunization, as reported by Conry et al. (92). In contrast to the above-mentioned studies, which all used intramuscular immunizations, they found that codelivery of plasmids encoding GM-CSF and human CEA via epidermal particle bombardment decreased antibody and T cell proliferative responses compared to CEA plasmid alone, whereas administering GM-CSF plasmid via this route 3 days prior to administering CEA plasmid enhanced antibody and T cell responses. Further study of this cytokine delivered by different routes is needed to determine the reasons for this apparent difference in effect.
IL-12 is a prototypic Th1 cytokine that is known to be a potent inducer of cellular immunity, leading to production of IFN- and enhancement of NK and cytotoxic T cell activity (93, 94). Kim et al. (89) examined the effects of IL-12 gene coadministration on the immune response to plasmids encoding HIV-1 env, gag/pol, and two accessory proteins, nef and vif. As expected, IL-12 induced a Th1-type response, with decreased antibody production and increased T cell proliferation, as well as markedly enhanced CTL responses to all four antigens. In fact, IL-12 codelivery could induce direct CTL activity in the absence of any in vitro stimulation. IL-12 codelivery was compared to GM-CSF codelivery in the same study. GM-CSF had a beneficial effect on serology and T cell proliferation, but little or no effect on CTL induction was observed. Iwasaki et al. (90) demonstrated that coimmunization with an IL-12-encoding plasmid can convert a weak plasmid DNA immunogen into one that induces a strong CTL response. Other groups (84, 95–98) have confirmed this activity of IL-12, using multiple different antigens, as a powerful inducer of cell-mediated immunity to DNA vaccines given intramuscularly or intranasally, although when given intranasally the suppression of humoral immunity was not seen (84).
Kim and others (99) have recently reported on the effects of several other cytokine plasmids when coadministered with different plasmids encoding HIV-1 proteins. The proinflammatory cytokines IL-1, tumor necrosis factor 
(TNF-), and TNF-ß are important mediators of the host response to tissue injury or infection. When these cytokine genes were coinjected into mice with HIV-1 env or gag/pol plasmids, all three significantly boosted specific antibody responses compared to immunization with the HIV-1 plasmids alone. In addition, both TNF- and TNF-ß enhanced T cell proliferation. With regard to CTL responses, TNF- had the greatest positive effect. Another proinflammatory cytokine, TCA3, a ß-chemokine chemotactic for monocytes/macrophages and neutrophils, has also been reported to increase Th and CTL activity when coadministered with a DNA vaccine, though it did not enhance antibody activity (100). In contrast, Kim et al. (100a) demonstrated that CD8-elaborated chemokines could drive CTL, Th, and antibody responses under different conditions. Kim et al. (89, 95) also compared genes encoding the putative Th1 cytokines IL-2, IL-15, and IL-18 with IL-12. Coinjection with IL-15 resulted in modest enhancement of antibody and proliferative responses and marked enhancement of CTL responses, on a par with that induced by IL-12 coadministration and yet without the depression of humoral responses previously reported with IL-12. By comparison, IL-2, which shares receptors with IL-15, stimulated strong antibody and proliferative responses but had little effect on CTL responses. Both IL-12 and IL-18 are believed to be IFN--dependent cytokines. However, while IL-12 polarized the resulting immune response toward strong cellular immunity, IL-18 induced dramatic increases in antibody production and T helper responses, with only modest increases in CTL activity. Finally, the Th2 cytokines IL-4, IL-5, and IL-10 were studied. As prototypic Th2 cytokines, these cytokines markedly augmented humoral and T cell proliferative responses without affecting CTL responses. These cytokines induced the highest antibody titers of all the coinjection groups (99).
The studies discussed above indicate the ability of cytokine gene coadministration to modulate the immune response to many different DNA immunogens. To date, however, these effects have only been examined in rodents. The next step should be to examine the effect of cytokines on DNA immunization in animals of different species, especially nonhuman primates. The long-term effects of cytokine gene expression also need to be examined, from both an efficacy and safety perspective. For example, Wolf and colleagues (101, 102) demonstrated that coadministering IL-12 with a protein antigen induced a primary Th1 response, a result consistent with the known biological actions of IL-12 and similar to that seen with DNA vaccines. However, the recall response appeared to be characterized by both a Th1 and Th2 response, despite subsequent immunizations with IL-12. While these results may be explained by the fact that these experiments used a protein immunogen, which tends to bias the immune response more toward a Th2 phenotype than a DNA immunogen, they nonetheless underscore the importance of looking beyond an initial response to assess the true effect of cytokines on vaccine-induced immunity. Recall studies have yet to be reported for IL-12 as an adjuvant for DNA immunization, and such studies will add significant data to protein–DNA comparisons. It will also be important to assess the safety of cytokine constructs in nonhuman primates, given concerns about persistent cytokine production leading to unwanted local or systemic inflammatory responses (103).
Costimulatory molecules
Another strategy for strengthening the effectiveness of genetic immunization is through the codelivery of genes for costimulatory molecules such as B7.1 (CD80), B7.2 (CD86), and CD40, with the goal of improving the antigen-presenting capabilities of transfected cells. These molecules play an important role in antigen presentation by providing a `second signal' necessary for efficient MHC-restricted T cell activation ( Fig. 3)
(104). Studies in mice have shown that intramuscular injection of a B7.2 gene expression cassette along with cassettes for influenza nucleoprotein (90) or HIV-1 proteins (105) results in a significant enhancement of cytotoxic T cell responses to the encoded antigen. This effect was seen whether the B7.2 gene was delivered in the same plasmid as the antigen gene (90) or in a separate plasmid (105). The effect of B7.1 is less clear-cut. Although codelivery of B7.1 plasmids with viral protein plasmids have not demonstrated any enhancement of immune responses (90, 105), two groups have reported that codelivery of B7.1 and a tumor-associated antigen led to increased protection from subsequent tumor challenge (92, 106). The reason for this difference in activity is not clear, but it has been suggested that these two molecules may differentially induce T cells down the Th1 or Th2 pathway (107). More recently, a plasmid encoding CD40 ligand (CD154), which is transiently expressed on T cells and serves to stimulate CD40-bearing APCs, inducing them to proliferate and differentiate, has been reported to augment antibody and CTL activity when coadministered with a DNA immunogen (108). Taken together, it appears that mimicking costimulatory molecules or molecules that facilitate interaction between T cells and APCs can specifically benefit vaccine-induced responsiveness ( Table 5).
|
|
DEVELOPING VACCINES AND IMMUNOTHERAPIES
Genetic immunization is a powerful new tool for both prophylactic vaccination and immunotherapy for human disease. One of the most attractive features of this approach is its malleability: altering the conditions under which the vaccine is administered can modulate both the magnitude and orientation of the subsequent immune response. Some of these conditions and the way they bias the immune response are depicted in Fig. 4.
|
By using this technology, researchers could rationally tailor the immune response to DNA-based vaccines and immunotherapies based on the particular correlates of immunity of each individual disease. For example, if antibodies are thought to be the most important correlate of protection for a particular disease, then designing and testing a vaccine that includes a Th2-type cytokine and a costimulatory molecule that augments the humoral response, and administering it via the epidermal gene gun method (which has been shown to induce a Th2-mediated response), may be most effective. If Th1-biased, cell-mediated immunity is the most important element and antibody response is less critical, then the optimal DNA vaccine would contain multiple CpG motifs, be coadministered with a Th1-inducing cytokine gene, and be given intramuscularly. For diseases in which the correlates of immunity are still unknown, such as HIV, a combination of strategies that induce both Th1 and Th2 responses could be sought.
These applications are not limited to infectious diseases. By redirecting the orientation of the immune response, genetic immunization may have a potential use as immunotherapy for autoimmune disease (109) and allergies (74). Immunizing with the genes for tumor-specific antigens may generate tumor-specific immune responses that could play a role in cancer treatment (21, 22, 85, 110). In this case, codelivery of immunomodulatory genes may be particularly effective, as successful tumor immunotherapy would require breaking tolerance, a significant immune feat. Of course, a great deal of further study is required. More testing of DNA vaccines given under various conditions in primates is needed to determine whether the immune-modulating effects (observed mostly in mice so far) hold true in a more relevant animal model. At this time, however, the ability of DNA vaccine-induced immune responses to be manipulated, especially by coadministered cytokines, appears to be a real phenomenon. This phenomenon is likely to have a profound effect on the future of genetic immunization, and may ultimately usher in a new age of rationally designed vaccines and immunotherapeutics.
FOOTNOTES
1 Correspondence: 505 Stellar-Chance Bldg., Department of Pathology, University of Pennsylvania School of Medicine, 422 Curie Blvd., Philadelphia, PA 19104, USA. E-mail: dbweiner{at}mail.med.upenn.edu 
2 Abbreviations: APC, antigen-presenting cell; CEA, carcinoembryonic antigen; CpG, cytosine-phosphate-guanine; CTL, cytotoxic T lymphocyte; DTH, delayed-type hypersensitivity; GM-CSF, granulocyte-monocyte colony-stimulating factor; HBsAg, hepatitis B surface antigen; HCV, hepatitis C virus; HIV, human immunodeficiency virus; HSV, herpes simplex virus; IFN, interferon; Ig, immunoglobulin; IL, interleukin; ISS, immunostimulatory sequence; MHC, major histocompatibility complex; NK, natural killer; NP, nucleoprotein; SIV, simian immunodeficiency virus; Th, helper T lymphocyte; Th1, type 1 helper T lymphocyte; TNF, tumor necrosis factor.
REFERENCES
