Novel vaccine delivery systems: solutions to HIV vaccine dilemmas?

Abstract

Introduction The history of successful vaccination against severe viral diseases such as smallpox, poliomyelitis or measles led to the initial hope that a vaccine against AIDS would be developed quickly. However, an effective vaccine against HIV needs to overcome substantial obstacles that emerged as research progressed. Due to its life cycle, HIV can effectively hide from the host immune response, integrating itself as proviral DNA in the host cell genome. A strategy pursued to deal with this difficulty is to include early viral regulatory proteins such as Tat, Rev, or Nef as vaccine antigens for induction of immune responses that can recognize and destroy HIV-infected cells as soon as the virus life cycle is activated [1,2]. The virus preferentially targets and destroys host immune cells such as T-helper lymphocytes, macrophages and dendritic cells that are probably essential to maintain an effective antiviral immune response. This would imply that vaccine-elicited immunity, unlikely to be able to prevent infection itself, must be able to quickly control virus replication to prevent harm to the immune system. The high antigenic variability of HIV can be considered as an extremely effective immune-evasion strategy. Because of the low fidelity of the viral RNA polymerase, virus progeny always represents a collection of RNA genomes (quasi-species) with random mutations. In vivo selection of immunodeficiency virus variants that can evade the recognition of neutralizing antibodies is common and strong virus-specific cytotoxic T-cell responses can select for escape variants already during resolution of primary viremia [3]. Most HIV infections are acquired sexually via the genital or rectal mucosae; however, at these entry sites it appears difficult to induce strong antiviral immunity by vaccination. Finally, HIV infection is a poverty-related disease that is particularly threatening health in societies of the developing world. Therefore, vaccine candidates must be safe and feasible in production and administration to be eligible for use where most needed. Because of these formidable obstacles, there have been years of scepticism about the possibility of success in AIDS vaccine development. By now, not least because research efforts have been intensified, the first encouraging data are emerging, especially from studies in the SIV/simian HIV (SHIV)-macaque model. In this article, we review the most recent progress in HIV-specific vaccinology, with the emphasis on delivery systems that promise to offer solutions to the obstacles to an effective AIDS vaccine. Live vector vaccines Progress in molecular biology and biotechnology has allowed production of recombinant viral or bacterial vectors capable of delivering vaccine antigens. This approach appears particularly attractive in the case of HIV because of the uneasiness with regard to live virus vaccines and the reasoning that induction of strong HIV-specific cell-mediated immunity may be important, if not essential, to achieve protection. Live vector vaccines are designed to mimic microbial infections allowing for de novo synthesis of vaccine antigens that appear particularly suitable for presentation via MHC-I molecules. Additionally, vaccination with these live vectors may elicit appropriate 'danger' signals to the immune system, resulting in preferential recognition and presentation of target antigens. Recombinant poxviruses have maintained a pole position among candidate viral vectors used in clinical trials for the delivery of HIV antigens. Particular features of poxvirus vectors include high stability and reasonable cost of manufacturing, which should help to make them suitable as future vaccines also for the developing world. Their capacity to accommodate large amounts of foreign DNA and to reliably induce immune responses to heterologous antigens is well established. From the evaluation in combined immunization protocols, recombinant poxviruses emerged as strikingly efficient vaccines for in vivo expansion of primed T-cell responses (for a review, see [4]). Most of the currently evaluated vaccines are based on vectors derived from highly attenuated vaccinia viruses such as modified vaccinia virus Ankara (MVA) [5] or NYVAC [6] and avipoxviruses [7,8]. These viruses all share the property of replication deficiency in mammalian cells that may have contributed to the already established clinical safety of the corresponding vectors. Importantly, even high-dose inoculation of MVA was shown to be safe in immune-suppressed macaques, suggesting the safety of recombinant MVA vaccines in potentially immunocompromised individuals [9]. Attenuated MVA vector vaccines induced significant levels of humoral and cellular immune responses to HIV-1 Env despite pre-existing vaccinia virus-specific immunity that considerably affected the immunogenicity of replication-competent vaccinia virus vectors [10]. When tested in the SIV-macaque challenge model, recombinant MVA vaccines delivering Env and Gag-Pol antigens primed for SIV-specific neutralizing antibodies and cytotoxic T lymphocytes (CTL) that were associated with reduced viremia on challenge [11-13]. Results from recent experiments in the SHIV-89.6P challenge model showed that vaccination with recombinant MVA can control replication of a highly pathogenic immunodeficiency virus and prevent disease progression [14]. While immunizations failed to elicit detectable virus neutralizing antibodies, the efficient control of viremia correlated well with vaccine-induced CTL responses. Vector vaccines on the basis of the genetically attenuated vaccinia virus NYVAC and delivering HIV-1, HIV-2, or SIV Env-Gag-Pol antigens reduced viral loads on challenge with HIV-2 and SHIV [15], and were used in combination with antiretroviral treatment for therapeutic immunization in SIV-infected macaques [16]. The safety and immunogenicity of recombinant canarypox viruses expressing HIV-1 gp 120, p55 and protease sequences was further established in a first phase II clinical trial involving higher-risk and lower-risk volunteers [17]. Alphavirus-based vector vaccines, including those derived from Semliki Forest Virus [18] and Venezuelan equine encephalitis virus [19], have shown promise when tested in the macaque-SIV challenge model [20,21]. Recombinant human adenovirus type 5 serves as a well-characterized candidate vector in HIV vaccine development [22,23]. Recent data suggest that strong endemic neutralizing immunity directed against the vector virus may be overcome by the use of non-human adenoviruses [24]. A number of vaccines based on the latter viral vectors are reportedly progressing towards clinical evaluation [25]. In addition to vector systems that already are or will soon be included in human phase I-III studies, several new delivery systems for a potential HIV-1 vaccine have recently been reported and may develop to attractive future vaccine candidates (Table 1). Recombinant, replication-competent rabies virus vaccine strain-based vectors expressing HIV-1 gp160 from both a laboratory-adapted (CXCR4-tropic) and a primary (dual-tropic) HIV-1 isolate were shown to induce strong cross-reactive CTL responses and high neutralization titers [26,27]. Similarly, live recombinant vesicular stomatitis viruses expressing HIV-1 antigens were shown to be highly effective in induction of neutralizing antibodies to a primary HIV-1 isolate [28], and recombinant, avirulent coxsackie virus expressing HIV-1 genes was proposed as novel vector for HIV immunogens [29]. Vaccine protection or sustained reduction in viral load following rectal challenge with pathogenic SIVmac239 was also achieved by recombinant strains of herpes simplex virus expressing SIV Env and Nef antigens [30]. Other herpesviral vectors like recombinant cytomegaloviruses are currently being developed (S. Dewhurst, M. Messerle, personal communication, 2001).Table 1: Comparison of viral vectors evaluated as candidate HIV vaccines.Besides viral delivery systems, bacterial vectors have been tested for vaccine approaches. A system with attractive characteristics is recombinant, attenuated Listeria monocytogenes since it may infect monocytes, the key antigen-presenting cells, and since natural infection with this bacterium originates at the mucosa. In mouse models, HIV-specific CD8 T-cell responses and protective capacity could be demonstrated after vaccination with Listeria strains engineered to express HIV-1 Gag or Nef [31-33]. Other bacterial vectors under investigation as experimental HIV vaccines are recombinant bacillus Calmette-Guerin [34,35] or attenuated strains of Salmonella [36]. HIV-vector-based and attenuated HIV vaccines Inoculation with live attenuated deletion mutants of SIV has afforded rhesus monkeys strong protection against subsequent infection by virulent wild-type SIV. However, the usefulness of live attenuated vaccines for controlling an HIV epidemic is being controversially discussed [37]. Most live attenuated HIV vaccine approaches have been abandoned since the occurrence of revertants is feared, which has been documented in vaccine studies in rhesus macaques inoculated with multiply deleted SIV [38,39]. Moreover, the SIV-macaque model yielded evidence for the possibility of in vivo recombination between the live attenuated vaccine and the wild-type virus given as challenge [40]. Nonetheless, in an interesting study using a mutant SIV vaccine strain with a deletion in the v1 and v2 variable regions of gp120, complete protection against a lethal strain of SIV was achieved [41]. The virus load after inoculation of the v1/v2 mutant SIV strain was extremely low, probably due to the induction of neutralizing antibodies able to contain the vaccine virus. The future for live attenuated HIV may be to develop vaccine constructs that avoid the risks associated with continuously replicating vaccine viruses. In such an approach, the Escherichia coli-derived Tet system for inducible gene expression was elegantly adopted to engineer HIV-1 genomes that replicate exclusively in the presence of the drug doxycycline [42]. DNA vaccines This attractive delivery system relies on the inoculation of plasmid vectors to produce the immunizing protein and appears especially effective in priming long-lived cellular immune responses [43]. DNA vaccines encoding HIV or SIV antigens were used in a considerable number of vaccine studies. It was recently shown that DNA vaccines are able to induce antigen-specific T-cell proliferation, as well as antibody and cytokine responses in man [44]. However, as the immune response appears to be transient and low in non-human primates and man, most studies on DNA vaccines applied a variety of different strategies to increase their immunogenicity. An increasing number of DNA vaccines is now based on codon-optimized sequences, since it has previously been found that they are superior in the induction of both humoral and cellular immune responses [45]. Codon-optimized HIV-1 gag[46], SIV gag[47,48], and HIV-1 env genes of a primary isolate (BX08) [49] were shown to possess increased immunogenicity in comparison with wild-type sequences and to be Rev-independently expressed. Moreover, it was shown that codon-optimized sequences are able to elicit SIV-specific CTL responses in rhesus monkeys [47]. In several approaches, either multiple genes in combination or mutated proviral HIV or SIV plasmids producing non-replicating viral particles were used. A plasmid expressing Vif, Vpu and Nef as a fusion protein with proteolytic cleavage sites was able to induce significant levels of cellular immune responses [50]. Mucosal immunization with the SIV proviral clone induced high rectal IgA titers but a low CTL response, as well as an incomplete protection after rectal challenge [51]. Similarly, vaccination with the mutated HIV proviral clone led to a reduction of peak plasma viral loads but no protection [52]. Other strategies to increase the efficacy of DNA vaccines include DNA immunization as part of complex immunization protocols with several different components, so-called 'prime-boost' approaches (see 'Combined vaccine applications'). Adjuvants With varying success, the collection of classical adjuvants has been used in combination with HIV immunogens [53,54]. More recently, several newly developed adjuvants (some of them vector based) including co-stimulatory molecules or cytokine and chemokine adjuvants, immunostimulatory (CpG) sequences or dendritic cells (DC) were tested [55]. Intriguingly, cytokines or plasmids coding for cytokines were successfully applied in several new studies. An enhanced anti-gp 120 immune response was found in rhesus macaques by co-injection of interleukin (IL)-2-encoding and IL-4-encoding plasmids [56]. The protective efficacy of DNA vaccines expressing SIVmac239 Gag and HIV-1 89.6P Env was greatly augmented by the co-administration of a purified fusion protein IL-2/immunoglobulin, consisting of IL-2 and the Fc portion of IgG, or a plasmid encoding IL-2/immunoglobulin [57]. In chronically HIV-1-infected chimpanzees, co-administration of an IL-12-expressing plasmid along with plasmids encoding Env/Rev and Gag/Pol led to enhanced proliferative responses to HIV antigens and a transient decrease in viral load [58]. In a mouse animal model, IL-12 was found to enhance the specific anti-Env cell-mediated immune response in a DNA prime/VV boost vaccine regimen [59]. IL-18 was able to increase both humoral and cellular immune responses against HIV-1 Nef in a DNA prime/protein boost combination in mice [60]. In two interesting approaches, the IL-6 and IL-12 cytokine genes were directly integrated into the SHIV genome itself [61,62]. CpG immunostimulatory sequences were used as adjuvants in several vaccination studies against different pathogens [63]. The addition of immunostimulatory sequences to inactivated, gp120-depleted HIV-1 virus particles in incomplete Freund's adjuvant was the optimal combination for the production of HIV-1-specific immune responses as measured by interferon-γ and IgG antibody production [64]. Primed DC were found to be excellent stimulators of the immune system. Different techniques were used to deliver HIV proteins into these cells. The uptake and immunostimulatory effect of HIV-1 Gag, Pol and Env proteins into DC could be considerably enhanced by encapsulation into liposomes [65]. DC infected with non-replicating HIV-1 virus particles pseudotyped with vesicular stomatitis virus glycoprotein were able to induce strong proliferative responses of CD4 and CD8 cells [66]. Mature DC infected with recombinant canarypox virus producing HIV-1 antigens could be successfully used to ex vivo stimulate virus-specific CD8 and CD4 T-cell responses from chronically infected patients [67]. Combined vaccine applications The considerable number of experimental lentivirus specific vaccines prompted approaches based on the combination of vaccines for enhancement of lentivirus-specific immune responses. The rationale of these strategies includes the fact that vaccines probably differ in their suitability for optimal primary induction (priming) or efficient subsequent in vivo expansion (boosting) of antigen-specific responses. Among the first of such concepts were combinations of poxvirus vectors and recombinant subunit protein vaccines, which indeed prove to be immunogenic when clinically tested [17]. More recently, the combination of different genetically engineered vaccines has also been proposed [68]. Data obtained from vaccine studies in the SIV/SHIV-macaque model suggest recombinant DNA priming and viral vector boosting as an especially promising approach, which already heads for first clinical evaluation [69]. DNA prime/MVA boost regimens were found to elicit high levels of activated lentivirus-specific T cells in vaccinated non-human primates [70,71]. Moreover, Amara et al.[72] demonstrated that immunization with multiple SIV and HIV-1 antigens, giving two DNA inoculations followed by a single booster with recombinant MVA, effectively controlled a mucosal challenge with highly virulent immunodeficiency virus SHIV 89.6P administered 7 months after the last vaccination. Particular encouraging take-home messages were the possibility of rapid containment of infection by vaccine-induced memory immune responses and the relative simplicity of vaccine design, dosage and administration. Another DNA/MVA virus prime-boost experiment, including mucosal immunizations in addition to intramuscular delivery, resulted in reduction of virus loads or protection from infection after intravenous challenge with chimeric immunodeficiency virus SHIV-4 [73]. Similar to other studies, T-cell immunity, here determined by antigen-specific interferon-γ production, appeared to correlate best with protection. Further encouraging results have recently been obtained after combined DNA and recombinant adenovirus type 5 vaccination yielding solid protection from immune damage in the SHIV 89.6P challenge model [74]. Recent data, again from the SIV-macaque model, suggest that extension of the prime-boost strategy to the additional use of different viral vectors may be promising to further improve HIV-specific immunity induced by vaccination [68]. Combinations of Semliki Forest virus vectors and recombinant MVA co-delivering SIV Env, Gag-Pol, Tat, Rev, and Nef antigens significantly enhanced immunodeficiency virus-specific immune responses in vaccinated cynomolgus macaques but failed to protect against intrarectal challenge with a heterologous SIV strain [75]. However, protection of cynomolgus macaques against intravenous infection with homologous SIV could be demonstrated using the same vector vaccine combination in a dose-escalation protocol and focusing on viral regulatory proteins Tat and Rev as antigens [72]. Interestingly, in a follow-up experiment repeating this scheme of immunization and including one group of monkeys receiving Gag-Pol-specific vectors, Tat/Rev vaccinees were able to more efficiently control replication of the challenge virus as compared with Gag-Pol-vaccinated animals (A. D. Osterhaus, K. J. Stittelaar, personal communication, 2001). The recent promise of prime-boost protocols is mainly based on the robust vaccine-induced immune responses that can be achieved. Moreover, looking at the number of already established vaccine vectors and the many more systems in the pipeline, it appears fortunate that, if necessary, there is a wealth of combined vaccination strategies still on substitute. Conclusion In spite of the complexities associated with the development of HIV vaccines, there is good reason for confidence. Significant progress has been made in enhancing the efficacy of prophylactic vaccination strategies (Table 2). This conclusion can be made from recent data collected in the up-to-date best animal model system, the infection of immunized macaques with virulent SIV or SHIV isolates. One main achievement is to learn that vaccine-induced immune responses may be able to control the replication of even agressive challenge viruses that are inoculated at doses far higher than an HIV infectious dose in humans. Furthermore, this immune control can be accomplished with rather feasible vaccine formulations and can provide protection from disease at least over a significant time period. The learned use of novel or improved vaccine delivery systems has significantly contributed to this success. Here, important strategies include the simultaneous delivery of multiple antigens, the increasing evaluation of new prime-boost immunizations, and the intended development of particular T-cell-inducing vaccines. However, formidable tasks remain to be tackled. The prevention of infection itself certainly is a supreme challenge. More realistic yet ambitious goals may be: (i) to realize even greater reduction of initial peak virus growth after infection with highly replication-competent viruses; (ii) to demonstrate protective capacity against challenge with heterologous viruses; (iii) to improve the potential of therapeutic immunizations based on the knowledge of how vaccine-induced responses can best control virus replication; and (iv) to transfer vaccine effectiveness from primate models to human clinical trials.Table 2: Advantages and disadvantages of different approaches in AIDS vaccine development.Acknowledgement The authors wish to acknowledge support from the European Commission (QLK2-2000-1040) and the Deutsche Forschungsgemeinschaft (SPP 1089).