DNA vaccines against mycobacterial diseases

For reprint orders, please contact reprints@expert-reviews.com DNA vaccines against mycobacterial diseases Expert Rev. Vaccines 8(9), 1237 1250 (2009) Marta Romano and Kris Huygen Author for correspondence Scientific Institute of Public Health-Site Ukkel, Program of Immunology & Vaccinology, 642 Engelandstraat, B1180 Brussels, Belgium Tel.: +32 2373 3370 Fax: +32 2373 3367 kris.huygen@iph.fgov.be Bacteria belonging to the genus Mycobacterium can cause several infectious diseases affecting humans and animals. Here, we reviewed the latest advances in the development of DNA vaccines against TB, Buruli ulcer and Johne s disease. Current understanding of the immunity to the respective causative pathogens indicates that the use of DNA vaccines encoding mycobacterial antigens could lead to efficient vaccination strategies. Moreover, characterization of protective mycobacterial antigens has been greatly facilitated by the ana lysis of immune responses induced after DNA vaccination. In addition, work aiming at optimizing DNA vaccines against mycobacterial diseases and research related to the controversial development of postexposure and therapeutic DNA vaccines are also discussed. Keywords: Buruli ulcer Johne s disease paratuberculosis postexposure DNA vaccine prophylactic DNA vaccine tuberculosis Infectious diseases caused by bacteria from the genus Mycobacterium Mycobacteria are weakly Gram-positive bacteria belonging to the genus Mycobacterium, the only genus in the family of the Mycobacteriaceae [1]. This genus comprises more than 70 pathogenic and nonpathogenic species, which can be divided into slowly growing (>7 days to form visible colonies on solid media) and rapidly growing species (<7 days to do the same). Most mycobacteria are saprophytes that are abundant in soil and water, but some are obligate pathogens and others are opportunistic pathogens that can be found in the environment, but can also occasionally cause infection. Slowly growing species are generally intracellular pathogens to humans or animals,while rapidly growing species are usually nonpathogenic to humans. Mycobacteria can cause leprosy, TB and Buruli ulcer (BU) disease in humans, and they can cause TB and Johne s disease in animals (mainly in ruminants). In this review, we will focus our discussion on the latest advances in the field of DNA vaccines for the development of vaccines for TB, BU and Johne s disease. First, we will briefly introduce these diseases. Tuberculosis Tuberculosis, HIV and malaria belong to the group of the big three infectious diseases claiming millions of lives every year [2]. In 2007, there were an estimated 13.7 million prevalent cases of TB worldwide (among these 9.27 million were new cases) and 1.3 million HIV-negative individuals died from it. In addition, 456,000 HIV-positive individuals died of TB (these are classified as HIV deaths) [3]. TB is a contagious infectious disease caused by infection with Mycobacterium tuberculosis and other mycobacteria belonging to the M. tuberculosis complex (Mycobacterium africanum, Mycobacterium bovis and Mycobacterium canettii). TB is spread by aerosols of M. tuberculosis or species of the M. tuberculosis complex, which are shed through coughs from open-cavitary pulmonary TB patients before the onset of multidrug therapy. Immunocompetent individuals inhaling the bacilli will be able to prevent the establishment of an infection in an estimated 70 90% of cases. The other 10 30% will become infected and develop acquired immunity to M. tuberculosis proteins (resulting in a positive purified-protein derivative skin test) [4]. Among these infected individuals, approximately 5 10% will develop active TB within 1 or 2 years after infection. The remaining 90 95% of infected individuals, due to the generation of an effective immune response, will become asymptomatically latently infected. These individuals have a residual population of viable, resting mycobacteria, which are mainly contained in well-organized pulmonary structures called granulomas. Latently infected individuals are a reservoir of quiescent TB bacilli, www.expert-reviews.com 10.1586/ERV.09.87 2009 Expert Reviews Ltd ISSN 1476-0584 1237

Romano & Huygen because they are potentially at risk to develop so-called postprimary TB at some stage of their life, following reactivation. It is estimated that a third of the worlds population is latently infected with M. tuberculosis. Comorbidity factors resulting in a state of immunodepression increase the risk of active TB. HIV coinfection is probably the most severe of these factors. Indeed, in individuals only infected with M. tuberculosis, the lifetime risk of reactivation is 10 20%, for HIV-infected individuals the risk is 10% annually [5] and TB accounts for 23% of HIV deaths [3]. The only vaccine currently available to prevent TB is the liveattenuated M. bovis bacille Calmette Guérin (BCG) vaccine. BCG is one of the most widely administered vaccines and in the year 2000 covered 86% of the world s population [6]. BCG vaccination protects children against TB meningitis and against disseminated, miliary disease, but confers a variable protection (ranging from 0 to 80%) against pulmonary TB in adults, and has been found to be of variable efficacy in a number of clinical trials [7 9]. When reviewing the current research for the development of a more effective vaccine for TB, it is important to distinguish between prophylactic vaccines (for individuals who are not yet infected with M. tuberculosis) and postexposure vaccines (for latently infected individuals). Moreover, we will also present results obtained when DNA vaccines encoding M. tuberculosis antigens are used in combination with antibiotics (so-called immunotherapeutic vaccines). Buruli ulcer disease Buruli ulcer disease is an infectious, necrotizing skin disease caused by infection with Mycobacterium ulcerans, occurring mostly in tropical and subtropical areas, which has been recognized by the WHO in 1998 as an emerging health problem [10,11]. Cases have been reported in several countries in west and central Africa, in central and south America, in southeast Asia, and in Australia. In west Africa, after TB and leprosy, BU is the third most common mycobacterial disease in humans. The natural history of M. ulcerans infection and the subsequent development of BU is not completely elucidated. Person-to-person transmission has not thus far been reported. In endemic areas, M. ulcerans bacteria have been found in stagnant water or slowly moving water sources, and in aquatic snails and carnivorous insects. Results from different researchers indicate a transmission involving insect vectors. Indeed, the work by Marsollier et al. has shown that in Africa, some aquatic insects can harbor M. ulcerans in their salivary glands and transmit the disease to experimental animals by biting [12,13]. In addition, M. ulcerans DNA has been detected in salt marsh mosquitoes trapped in southeastern Australia where an outbreak of BU occurred, although transmission by this type of insect has not been proven [14]. Infection by M. ulcerans initially causes a painless nodular swelling that can eventually develop into an extensive necrotizing lesion. M. ulcerans has the ability to produce a family of toxin molecules, the so-called mycolactones. These are polyketides that can suppress the immune system and destroy skin, underlying fat tissue and bone, causing severe deformities [15 17]. BU results in considerable morbidity. Owing to the late detection of the disease, treatment is principally by excision of the lesion, sometimes necessitating skin grafting [18]. The WHO is currently recommending combined oral rifampicin and intramuscular streptomycin treatment of nodules for 8 weeks in the hope of reducing the need for surgery [19,20]. Unfortunately, there is currently no specific vaccine against BU [21]. The M. bovis BCG vaccine, used for the prevention of TB, has been reported to offer a short-lived protection against the development of skin ulcers [22 24] and to confer significant protection against disseminated cases of BU, such as osteomyelitis, both in children and adults [25,26]. The precise M. ulcerans antigens that induce a protective immune response are still poorly defined. Thus far, only two laboratories (Unit of Molecular Bacterial Genetics, Pasteur Institute, Paris, France, and our group) have analyzed the potential of DNA vaccines for the development of a specific vaccine against BU, and these works will be presented in this review. Johne s disease Johne s disease, also called paratuberculosis, is a chronic granulomatous enteritis in ruminants caused by Mycobacterium avium subsp. paratuberculosis (MAP), principally affecting cattle, sheep and goats. Principal clinical signs are cachexia and chronic diarrhea (less common in goats and sheep). This disease results in considerable economic losses in the livestock industry, particularly in the dairy sector, due to premature culling or death, reduced milk production and decreased fertility. The route of transmission is mostly via the fecal oral route, but hygienic measures and culling of shedding animals are insufficient to eradicate this disease [27]. Estimation of exact prevalence is hampered by the lack of specificity and sensitivity of diagnostic tests. However, based on serological assays, the herd prevalence (at least one animal testing positive in a farm) in dairy herds in the USA and Europe has been estimated to be between 7 and 66% [28]. A number of vaccines against Johne s disease, based on wholekilled or live-attenuated bacteria, are commercially available [27]. They are effective at delaying the onset of clinical symptoms but do not protect against infection. Moreover, their administration has been associated with a risk of accidental self-injection by the veterinarian. Finally, vaccinated animals develop antibodies that interfere with existing serodiagnostic tests for paratuberculosis and they become reactive in the tuberculin skin test, used for the control of bovine TB. Infection with MAP has been associated with Crohn s disease in humans. Different studies indicate that MAP may be at least one of the triggers in the development of this inflammatory bowel disease through a complex interplay between genetic, infectious and immunological factors [29 32]. Given the lack of efficient and safe vaccines against Johne s disease that do not interfere with paratuberculosis and bovine TB immunodiagnostic tests, and given the possible association between infection with MAP and Crohn s disease, it is important to develop novel effective vaccines against this pathogen. We will present current advances on the development of vaccines against paratuberculosis, in particular those involving the use of DNA vaccines. 1238 Expert Rev. Vaccines 8(8), (2009)

DNA vaccines against mycobacterial diseases Review Assets of DNA vaccines for the development of vaccines against mycobacterial diseases In a DNA vaccine, the gene encoding an antigen is inserted into a bacterial plasmid vector, the plasmid is amplified in transformed bacteria and the purified plasmid DNA (pdna) is administered to an immunocompetent host. Moreover, further manipulation of the bacterial plasmid vector can result in vectors encoding more than one antigen, the fusion of several antigens or antigens plus other proteins, leading to more efficient vaccines. The priming of the immune response after immunization with DNA vaccines involves professional antigen-presenting cells (APCs), such as dendritic cells and Langerhans cells, which endocytose DNA into acidic vesicles for subsequent transport to the nucleus, followed by transcription into mrna and protein translation in the cytoplasm. Bacterial pdna has intrinsic adjuvant properties. Indeed, it contains unmethylated CpG sequences that can act as polyclonal activators of B cells and as adjuvants. These CpG motifs stimulate the production of costimulatory molecules by APCs through interaction with a specific intracellular Tolllike receptor (TLR)9, which is present on the surface of the early endosome [33]. The major costimulatory cytokines induced upon TLR9 triggering are IL-12 (which stimulates natural killer cells to produce IFN-g and favors the development of a Th1-type T-helper [Th] subset), TNF-a and IL-6 (which favors antibody production but also plays a role in cytotoxic T lymphocyte [CTL] and Th17 differentiation). Following DNA vaccination, antigenic material is generated in the myocyte/keratinocyte and also within the APC [34], and exogenous and endogenous antigen processing can proceed in much the same way as following infection with intracellular pathogens. DNA vaccines stimulate both the exogenous (MHC class II restricted) and the endogenous (MHC class I restricted) antigen-presentation pathways [34]. Dendritic cells can also take up antigen-containing apoptotic bodies from transfected myocytes, and present the relevant peptides to CD4 + and CD8 + T cells (the so-called cross-priming phenomenon) [35]. By virtue of this induction of CD8 + T-cell responses, DNA vaccines strongly mimick infection with live pathogens, in contrast to vaccines based on protein antigens or killed pathogens that are preferentially processed through the exogenous presentation pathway generating only MHC class II-restricted CD4 + responses. It is particularly this class I-restricted presentation, resulting in strong CD8 + -mediated immune responses, that is a hallmark of DNA vaccines and which makes them particularly attractive as vaccine formulations against viruses and intracellular bacteria, such as mycobacteria. Indeed, our current understanding of the immune responses against mycobacterial pathogens indicates that both CD8 + responses and Th1-biased CD4 + responses correlate with protective immune responses against M. tuberculosis, M. ulcerans and MAP [21,27,36]. Another asset of DNA vaccines for the development of vaccines against mycobacterial diseases is linked to financial considerations. Indeed, from a theoretical point of view, DNA vaccines are easy and cheap to design and produce. In addition, there is no need for a cold chain for their storage and distribution. Therefore, DNA vaccines are an attractive option when trying to develop vaccines against infectious diseases affecting socioeconomically weak populations, such as those mostly affected by TB and BU. DNA vaccines against mycobacterial diseases In 1996, the group of Jo Colston and Douglas Lowrie at the Medical Research Council (London, UK), and our group at the former Pasteur Institute (Brussels, Belgium; now integrated into the Scientific Institute of Public Health), were the first to report on the use of DNA vaccines against TB, using DNA encoding heat-shock protein (Hsp)65 of Mycobacterium leprae and Ag85A of M. tuberculosis, respectively [37,38]. A total of 13 years later, more than 30 M. tuberculosis antigens have been tested as encoded by DNA vaccines in preclinical TB models and it is impossible to discuss all the relative papers in this review. We will highlight some of the most recent results obtained in the field (summarized in Table 1) and refer to previous reviews for the less recent literature [39 41]. Overall, it is well established that the intramuscular immunization of mice with pdna encoding mycobacterial antigens is a potent inducer of strong Th1 immune responses, characterized by high levels of IL-2 and IFN-g, and little or no IL-4/IL-5 in antigen-stimulated spleen cell cultures. In addition, DNA vaccination has been a valuable tool for the characterization of mouse MHC class I-restricted epitopes for mycobacterial antigens. Indeed, pathogenic mycobacteria remain largely confined to the phagosome in experimentally infected mice. Therefore, identifying cognate specificities of CD8 + -mediated immune responses in infection models is very difficult. Some examples of mouse MHC class I-restricted epitopes for mycobacterial antigens characterized by means of DNA vaccination are those on the mycolyl-transferases of the Ag85 complex [42], the phosphate-binding proteins PstS-1 [43] and PstS-3 [44], the Mtb32 component of the 72F fusion protein [45], the latency-associated Rv2626c protein encoded by the dormancy DosR regulon [46] and the PPE44 antigen [47]. Moreover, the protective potential of a wide variety of subunit TB vaccines has been demonstrated by vaccinating mice with DNA vaccines encoding M. tuberculosis antigens. DNA vaccines encoding the mycolyl-transferases Ag85A and Ag85B are among the best- documented mycobacterial plasmids. Plasmids encoding these highly conserved mycobacteria-specific antigens have a definite vaccine potential, not only for human TB [38], but also for BU caused by M. ulcerans [48], for leprosy caused by M. leprae [49] and for MAP infections [50]. Very strong Th1 immune responses can be induced in mice with these plasmids [51], and Ag85-specific CD4 + Th cells induced by the vaccine can even function as novel adjuvants for the effective induction of HIV-1 specific CTL responses [52]. Recently, we have also reported on the immunogenicity of eight mycobacterial latency-associated antigens by pdna vaccination of BALB/c and C57BL/6 mice [46]. Expression of these proteins is upregulated in conditions mimicking a dormant M. tuberculosis infection and the same proteins are targeted by the immune system during latent infection in humans [53]. Finally, members of the Pro Glu (PE) Pro Pro Glu (PPE) family have also been reported to have a strong vaccine potential. The PPE protein www.expert-reviews.com 1239

Romano & Huygen Table 1. List of the principal and most recent mycobacterial antigens in preclinical animal models, and tested for their immunogenicity and protective efficacy against Mycobacterium tuberculosis, Mycobacterium ulcerans and Mycobacterium avium subsp. paratuberculosis. Antigen Readout and/or essential highlights Ref. Mycobacterium leprae Hsp65 Mtb Ag85A Immunogenic and protective in mice against experimental Mtb infection in terms of bacterial load reduction Immunogenic and weakly protective in mice against experimental Mu infection in terms of bacterial load in the infected tail Immunogenic and protective in mice against experimental Mtb infection in terms of bacterial load reduction [37] [60] [38] Mtb Ag85B Mtb Ag85A Mtb Ag85B Mtb Ag85C Mtb PstS-1 Mtb PstS-3 Mtb PPE44 Mtb dormancy regulonencoded antigens: Rv1733c Rv1738 Rv2029c (pfkb) Rv2031c/hspX/acr Rv2032c (acg) Rv2626c Rv2627c Rv2628 Mtb PE_PGRS: Rv3812 Rv3018c Mtb PE_PGRS: Rv0977 Rv1441c Rv1818c Mu Ag85A MAP Ag85A MAP Ag85B MAP MAP0586c MAP MAP0126 MAP antigens expression library immunization (mix of 26 MAP antigens) Protective in mice against infection with Mu in terms of bacterial load reduction in the foot pad [48] Immunogenic and protective against MAP in lambs [50] pdna Ag85B as an adjuvant for the induction of HIV-1-specific CTL responses by stimulation of Th1 immunity in mice Protective against leprosy in a mouse model as demonstrated by a reduction in bacterial number in footpad Characterization of the MHC-class I- and MHC-class II-restricted epitopes following pdna vaccination, Mtb infection and BCG vaccination in mice Protective in mice against Mtb infection as measured by reduced bacterial load Immunogenic and protective in mice against experimental Mtb infection in terms of bacterial load reduction. MHC-class I and MHC-class II restricted epitopes following pdna vaccination, Mtb infection and BCG vaccination characterized Analysis of the immunogenicity of DosR regulon-encoded antigens by pdna vaccination of mice, characterization of their MHC-class I- and MHC-class II-restricted epitopes Characterization of specific Th1 and CTL responses following pdna vaccination in mice [58] Immunogenic but not protective against an experimental Mtb infection in mice [59] Vaccination of mice with a pdna encoding species-specific Ag85A in a DNA prime protein boost protocol improves the protective efficacy in terms of delay in footpad swelling and reduction in bacterial load Immunogenicity and protective efficacy in mice of pdna encoding MAP antigens identified by postgenomic and proteomic ana lysis of MAP secretome Significantly protective against an experimental infection with MAP in mice (further characterization of the protective antigens needs to be performed) BCG: Bacille Calmette Guérin; CTL: Cytotoxic T lymphocyte; Hsp: Heat-shock protein; MAP: Mycobacterium avium subsp. paratuberculosis; Mtb: Mycobacterium tuberculosis; Mu: Mycobacterium ulcerans; pdna: Plasmid DNA. [52] [49] [42 44,51] [47] [46] [61] [69,75] [74] 1240 Expert Rev. Vaccines 8(9), (2009)

DNA vaccines against mycobacterial diseases Review Table 1. List of the principal and most recent mycobacterial antigens in preclinical animal models, and tested for their immunogenicity and protective efficacy against Mycobacterium tuberculosis, Mycobacterium ulcerans and Mycobacterium avium subsp. paratuberculosis (cont.). Antigen Readout and/or essential highlights Ref. Cocktail immunization: MAP Ag85A MAP Ag85B MAP Ag85C MAP SOD MAP MAP2121c MAP Ag85A MAP Hsp65 Immunogenic and protective in mice against an experimental MAP infection in terms of bacterial load reduction and histopathology Immunogenic and protective against MAP in lambs [50] BCG: Bacille Calmette Guérin; CTL: Cytotoxic T lymphocyte; Hsp: Heat-shock protein; MAP: Mycobacterium avium subsp. paratuberculosis; Mtb: Mycobacterium tuberculosis; Mu: Mycobacterium ulcerans; pdna: Plasmid DNA. [77] family of M. tuberculosis includes 69 proteins rich in glycine and, together with the PE protein family, accounts for approximately 10% of the coding capacity of the M. tuberculosis genome. There is little functional information about PPE proteins, but their polymorphic nature suggests that they may represent antigens of immunological relevance [54]. Subcellular fractionation and immunoelectron microscopy studies have indicated that some PPE proteins are located at the periphery of the bacterial cell and could therefore be accessible to the host immune system. Moreover, they induce strong immune responses in animals and humans infected with M. tuberculosis [55]. We have evaluated the vaccine potential of PPE44 (Rv2770c), overexpressed in virulent M. tuberculosis H37Rv as compared with the attenuated H37Ra strain [56]. ppe44 gene expression shows high quantitative variations in clinical isolates selected to represent the major phylogenetic lineages of the M. tuberculosis complex and, more specifically, strains of the Beijing type demonstrate high ppe44 expression [57]. PPE44-specific immune responses could be detected in mice acutely, chronically and latently infected with M. tuberculosis. Vaccination of mice with a pdna vaccine coding for PPE44 or with recombinant PPE44 protein formulated in adjuvant generated strong cellular and humoral immune responses, and immunodominant T-cell epitopes were identified. Most importantly, vaccination of mice with both types of subunit vaccine followed by an intratracheal challenge with M. tuberculosis resulted in a protective efficacy comparable to the one elicited by BCG [47]. Chaitra et al. reported on DNA vaccines encoding two PE_PGRS proteins, encoded by Rv3812c and Rv3018c respectively. Strong immunogenicity was demonstrated in BALB/c mice, both against MHC class I- and MHC class II-restricted epitopes. Interestingly, an epitope-specific response was demonstrated by the lysis of peptide-pulsed APCs, release of perforin and IFN-g production [58]. Finally, Singh et al. reported recently on DNA vaccination encoding three M. tuberculosis proteins, namely PE_PGRS 16 (Rv0977), PE_PGRS 26 (Rv1441c) and PE_PGRS 33 (Rv1818c). All three PE_PGRS proteins were found to be cell-surface antigens, but immunization of mice with these PE_PGRS antigens as DNA vaccines showed no protection in a TB aerosol challenge model [59]. For BU vaccine development, subunit-based vaccines are an attractive alternative to live-attenuated vaccines. Indeed, they have the advantage of being well characterized and of posing no risk for application in HIV-positive populations, such as the ones mostly in need of an effective BU vaccine [21]. Thus far, only two subunit-based DNA vaccines have been tested in the mouse model. In our laboratory, Tanghe demonstrated that vaccination with DNA encoding the mycolyl-transferase Ag85A from BCG could significantly reduce the bacterial load in the foot pads of M. ulcerans-infected mice [48]. Furthermore, the vaccination of mice with a DNA vaccine encoding hsp65 of M. leprae was significantly protective against M. ulcerans infection, but nevertheless inferior to the one conferred by vaccination with BCG [60]. Recently, our group also demonstrated that a species-specific Ag85A DNA vaccine encoding M. ulcerans Ag85A used in a DNA prime recombinant protein boost immunization protocol resulted in a protective efficacy comparable to the one induced by the BCG vaccine [61]. In order to achieve major advances for the development of an effective subunit vaccine against BU, there is an urgent need to identify and characterize more protective antigens. For this purpose, the recent M. ulcerans genome sequencing [62], as well as the ensuing possibility to perform proteome studies to identify promising antigens candidates, will surely be beneficial [63]. Regarding the development of subunit-based vaccines against paratuberculosis, the identification of immunodominant protein antigens inducing strong Th1 immune responses during the first asymptomatic stage of the disease and the demonstration of their protective potential in experimental infection models (mouse and target species) will be crucial. A major benefit of the development of this type of vaccines is that they would allow overcoming the diagnostic interference issues linked to the available whole-cell based vaccines [27]. The publication of the entire genome sequence of the K-10 strain of MAP has provided a precious tool for the identification of MAP antigens useful for more specific immunodiagnosis and for more effective immunoprophylaxis [64]. Indeed, as a result of the MAP genome sequencing and through previous studies, a number of MAP immunodominant Th1 antigens have been identified. Interesting antigens are the superoxide dismutase www.expert-reviews.com 1241

Romano & Huygen (SOD) MAP2121c [65], and the three members of the highly conserved mycobacterial proteins of the Ag85 complex Ag85A (MAP1609c), Ag85B (MAP0126) and Ag85C (MAP3531c), which induce strong (Ag85A and Ag85B) to regular (Ag85C) Th1 T-cell responses [66], and which have also been reported as immunodominant in experimentally infected cattle and mice [67], in mice vaccinated with recombinant Ag85B protein [68] and in mice vaccinated with DNA encoding Ag85A and Ag85B [69]. Other immunogenic MAP proteins are Hsp65 (GroEL) and Hsp70 (DnaK), which can also induce specific immune responses in MAP-infected and MAP-vaccinated cattle [70], and the P22 (22-kDa) exported MAP protein belonging to the LppX/LprAFG family of putative mycobacterial lipoproteins, which leads to specific immune responses (antibodies and IFN-g production) in sheep vaccinated with the live-attenuated Neoparasec vaccine, and in clinically affected and subclinically infected cows [71]. In addition, two MAP proteins belonging to the PPE family, MAP1518 and MAP3184, elicit significant IFN-g levels in peripheral blood mononuclear cells from experimentally infected Holstein calves [72]. These may have vaccine potential, as PE/PPE proteins are implicated as virulence factors in M. tuberculosis and, as mentioned before, a number of PE/PPE proteins of M. tuberculosis are promising TB vaccine candidates [45,47]. Concerning the development of DNA vaccines against paratuberculosis, it is important to first highlight the fact that DNA vaccines are already available for some veterinary viral diseases [73] and this might theoretically facilitate the development of a DNA vaccine against paratuberculosis. Some studies have already proved that the use of DNA vaccines encoding MAP antigens can protect against paratuberculosis. For example, using expression library immunization, Huntley et al. reported on the protective potential of a plasmid mix (encoding 26 MAP antigens) that conferred significant protection to BALB/c mice against intraperitoneal challenge with 10 8 colony-forming units (CFUs) of MAP [74]. Genes in the protective mix coded for transport/binding, membrane and virulence proteins, and mycobactin/polyketide synthases, but to our knowledge further ana lysis of the respective antigens in this mix has not yet been performed. In addition, our group has recently evaluated the DNA vaccine potential of MAP0586c and MAP4308c [75]. These two MAP proteins were previously identified by postgenomic and immunoproteomic ana lysis of the MAP secretome as novel serodiagnostic antigens [76]. Immunization of mice with pdna encoding MAP0586c and MAP4308c induced strong Th1 immune responses, whereas only DNA encoding MAP4308c stimulated antibody responses. MAP-infected BALB/c mice also generated strong MAP0586c-specific T-cell responses and could be partially protected against infection following DNA vaccination, indicating that this putative transglycosylase warrants further investigation [75]. In addition, Park et al. recently reported on a DNA vaccine cocktail immunization of C57BL/6 mice, using a mix of five plasmids encoding Ag85A, Ag85B, Ag85C, SOD and a 35-kDa protein (MAP2121c) [77]. Mice were vaccinated three times at 3-week intervals and challenged by an intraperitoneal injection of 10 9 CFU of MAP 3 weeks after the second booster. This resulted in a significant reduction in the number of bacteria in the spleen and liver of vaccinated mice. The potency of DNA vaccines in larger animals and humans has usually been found to be considerably lower than in small rodents, but a recent study in sheep provided sufficiently encouraging results to continue optimizing this form of vaccination for paratuberculosis. DNA vaccines encoding Ag85A from M. bovis BCG and from M. avium subsp. avium and Hsp65 from MAP were evaluated in groups of five lambs each [50]. Lambs were vaccinated intramuscularly three times (0, 20 and 40 days) at 5 months of age and challenged with 2 10 9 MAP, 3 months after the last vaccination. Histopathology of postmortem tissue sections after 1 year revealed an absence of lesions in all three DNA-vaccinated groups, whereas lesions were readily observed in the control group. Optimization of DNA vaccines against mycobacterial diseases Attempts to increase delivery to APCs A major problem with DNA vaccines is their transfection efficacy and the amount of actual protein synthesized. In vivo electroporation dramatically increases the number of DNA-transfected myocytes, and both in mice and farmed ruminants, this electroporation technique was shown to increase the immunogenicity of TB DNA vaccines [78,79]. Li et al. reported that an in vivo DNA electroporation prime and protein-boost strategy can also enhance humoral immunity of TB DNA vaccines encoding Ag85A and early secretory antigenic target (ESAT)-6 in nonhuman primates [80]. Cross-priming, in which antigen-containing apoptotic bodies from dead, plasmid-transfected muscle cells are engulfed by immature dendritic cells, is thought to be essential in the priming of the immune response upon intramuscular DNA vaccination [35]. In an attempt to increase cross-presentation through apoptosis, we inserted the DNA encoding caspase-2 prodomain followed by wild-type or catalytically inactive mutated caspase-3 into a plasmid encoding Ag85A. Transient transfection showed that the mutated caspase induced slow apoptosis, normal protein expression and NF-kB activation, while wild-type caspase induced rapid apoptosis, lower protein expression and no NF-kB activation. Ag85A-specific antibody production was increased by coexpressing the mutated and decreased by coexpressing the wildtype caspase. Vaccination with pro-apoptotic plasmids triggered more Ag85A-specific IFN-g-producing spleen cells, and more efficient IL-2 and IFN-g-producing memory cells in spleen and lungs after M. tuberculosis challenge. Compared with DNA encoding secreted Ag85A, vaccination with DNA coexpressing wild-type caspase increased protection after infection with M. tuberculosis, while vaccination with plasmid coexpressing mutated caspase did not, possibly due to exaggerated production of proinflammatory cytokines IL-6, IL-10 and IL-17A in this vaccination group [81]. Complexation of DNA to adjuvants such as the cationic lipid Vaxfectin [82] can increase the immunogenicity and protective efficacy of DNA, but as this adjuvant is particularly effective for increasing antibody responses through the stimulation of the Th2 cytokine IL-6, its use needs to be studied carefully, particularly with respect to the possible induction of deleterious Th17 cells [83]. 1242 Expert Rev. Vaccines 8(9), (2009)

DNA vaccines against mycobacterial diseases Review For the mucosal delivery of DNA vaccines, it is essential to protect them against host endonucleases by formulation in a carrier system. DNA vaccination has been successfully used for the definition of human MHC class I-restricted epitopes, using HLA-A*0201 transgenic mice [84], and using these HLA-A*0201 transgenic mice, Bivas-Benita et al. demonstrated that pulmonary delivery of chitosan DNA nanoparticles could induce (be it weak) spleen cell IFN-g responses against four out of seven of the HLA-A2 predicted peptides encoded by a polyepitope DNA [85]. Manganelli and colleagues have developed a nonpathogenic invasive commensal Escherichia coli BM2710 strain that can be used for the intranasal delivery of pdna encoding Ag85A and HtpX [86]. Antigen-specific T-cell responses and protection against M. tuberculosis challenge were induced with DNA doses at least 100-fold lower that those needed for intramuscular immunization. This mucosal approach using nonpathogenic E. coli for needle-free plasmid delivery certainly merits further studies. Attempts to increase vector immunogenicity A promising approach to maximize the expression of microbial genes in the plasmid is the optimization of codon usage. A synthetic humanized Ag85B gene codon optimized for expression in human cells was reported to display an approximately sixfold increased in vitro expression in Cos-7 cells. Antibody responses were not affected by the optimization, but Th1-like and CTL responses (assessed in BALB/c mice) induced with the humanized gene were higher than with the plasmid encoding wild-type Ag85B. Finally, reduced CFU counts in the spleen and lungs from animals challenged with M. tuberculosis 4 weeks after the third DNA immunization were also indicative of increased potency of the humanized pag85b DNA [87]. Multisubunit vaccination by coimmunization with different DNA vectors that are not very effective, as single vaccines may result in a greater degree of protection as indicated by reduced CFU counts [88] and, more convincingly, by up to sevenfold prolonged survival times following high-dose aerosol challenge compared with mice vaccinated with vector DNA only [89]. For industrial purposes, however, the use of hybrid genes or of multipromoter plasmids is more interesting than a combination of plasmids. In this line of thinking, Steven Reed and his colleagues at Corixa Corporation (WA, USA), have shown very convincing results for a Mtb72F DNA encoding a fusion of the Mtb39 PPE protein and a 32-kDa serine protease [45]. Similarly, we have analyzed the use of a pbudce4.1 vector encoding the genes for the mycolyl-transferase Ag85A and the 40-kDa phosphatebinding protein PstS-3 under control of the promoters for IE1 of cytomelagovirus and of human elongation factor EF1-a, respectively. Although both antigens were expressed by the pbudce4.1 vector, humoral and cellular immune responses clearly indicated an antigenic competition at the level of CD4 + (but not of CD8 + ) T cells, between the Ag85A and PstS-3 protein, with the mycolyltransferase being the dominant partner over the phosphate transport receptor [90,91]. We hypothesize that prior activation of Ag85A-specific CD4 + T cells directed against this common mycobacterial antigen may have led to cross-competition for MHC class II-restricted peptide complexes between the Ag85A and the Pst-3 antigens. This might have implications for future combination vaccines using components of the Ag85 complex. Fadda et al. reported on a multigene DNA combination, encoding the combination of Ag85B coupled to ESAT-6 [92], a fusion protein that has been entered into clinical Phase I and Phase II trials as hybrid1 protein-subunit vaccine [93]. They showed that this Ag85B ESAT-6 vaccine could not be enhanced by broadening the antigen repertoire by adding other highly immunogenic secreted proteins [92], which is largely in agreement with our findings on the combination of Ag85A and PstS-3. Another way to increase the immunogenicity of DNA vaccines is the coadministration of plasmids encoding costimulatory molecules (CD80/CD86) or adjuvant cytokines. Coimmunization with plasmids expressing granulocyte macrophage colonystimulating factor or IL-12 can enhance the T-cell immunity of DNA vaccines encoding Ag85B or MPT64 by approximatley twofold. However, this was not sufficient to improve their protective efficacy at the peak of infection after an aerosol challenge with M. tuberculosis [94,95]. DNA encoding IL-23, but not IL-27, was also reported to increase the efficacy of DNA encoding Ag85B from M. tuberculosis [96]. Dou et al. addressed the immune adjuvant effects of IL-21 on a DNA vaccine encoding Ag85A. Their results showed that the DNA vaccine construct prsc IL21 Ag85A elicited stronger immune responses in BALB/c mice than prsc Ag85A [97]. Finally, DNA encoding the cytokine IL-15, with a pivotal role in the maintenance of memory T cells, holds particular promise [98], although to our knowledge pil-15 has not been tested in combination with mycobacterial antigens. Prime boost strategies By virtue of the strong Th1-biased and MHC class I-restricted immune responses that DNA vaccines can induce, they are particularly attractive as priming agents in prime boost regimens. For HIV and malaria, these prime boost regimens have progressed successfully to clinical trials [99]. Systemic boosting of DNA-primed animals with mycobacterial proteins [100 103], with recombinant modified vaccinia Ankara type (MVA) [104] or recombinant replication-deficient adenoviruses [105] have been reported to increase immunogenicity and/or protective efficacy of TB DNA vaccines. Similarly, boosting DNA-primed mice with a live Salmonella typhimurium vaccine carrier expressing ESAT-6, has also been attempted [106]. Xing and colleagues have recently demonstrated that airway delivery of soluble Ag85A protein can restore protective mucosal immunity induced by single intramuscular Ag85A DNA vaccination. Recruitment of systemically activated antigenspecific T cells into the airway lumen of DNA-vaccinated mice appears to be critically dependent not only on proinflammatory signals induced by the intranasal instillation but equally on parenteral immunization and robust systemic antigen-specific T-cell priming, which we think is an important conceptual finding [107]. www.expert-reviews.com 1243

Romano & Huygen Combinations of pdna with the existing M. bovis BCG vaccine Feng et al. were the first to report that sequential immunization with mycobacterial antigen 85B-expressing DNA followed by M. bovis BCG Tokyo was more effective than BCG immuni zation alone in protecting B6 mice against an aerosol M. tuberculosis infection [108]. It is now well established that immune responses induced by the existing BCG vaccine can be augmented by combinations with pdna in mice, guinea pigs and cattle, either by priming with DNA [108 110], or by boosting with DNA [111 113]. In most of these studies, increased efficacy of BCG has been measured by CFU counting or pathological scoring in infected organs, but effects on long-term survival were rarely reported. By performing long-term survival studies in BALB/c mice infected intravenously with M. tuberculosis, we have shown that priming with DNA prior to BCG, but not boosting after BCG with DNA (or recombinant protein or recombinant MVA for that matter) encoding Ag85A, could increase the potency of the BCG vaccine, resulting in 7 9 weeks longer mean survival times [114]. A BCG prime/dna boost protocol, using plasmid encoding a 72F fusion protein, did increase the long-term survival of guinea pigs infected with a lowdose aerosol [113]. The reason for the discrepancy with our report is not clear, but might be linked to the different animal species used. It is not very likely that boosting with BCG after priming with DNA will be applicable as a vaccine regimen for developing countries, where neonatal BCG vaccination is part of the expanded program of immunization by the WHO. On the other hand, this regimen may have a better chance for testing in populations at risk (healthcare workers, prisoners and exposed household contacts) in developed Western countries where BCG is not routinely administered. Finally, the actual BCG vaccine is not very effective in inducing immune responses against HspX (Rv2031c) [115] and other latency-associated antigens [116], and this could partially explain the low efficacy of BCG against pulmonary, reactivation TB [117]. A tempting experimental approach would be to administer BCG and DNA vaccines encoding latency-associated antigens simultaneously. However, a proof-of-concept of this combined vaccination approach remains to be given. Postexposure & therapeutic DNA vaccines against mycobacterial diseases A number of groups are exploring the potential use of DNA vaccines encoding M. tuberculosis antigens to be administered as postexposure vaccines or as immunotherapeutic vaccines. The aim of postexposure vaccines is to prevent reactivation of TB in individuals latently infected with M. tuberculosis. Indeed, as already mentioned, it is estimated that a third of the world s population is latently infected with M. tuberculosis and is potentially at risk to develop active TB at some stage of their life. Immunotherapeutic vaccines are rather considered as an adjunct to the antibiotic cure of active TB, with the aim of shortening the duration of the drug treatment. Indeed, active TB is treated by the administration of several antibiotics for 6 24 months. Choice of the antibiotics and the duration of treatment is dependent on the susceptibility of the infecting strain to the available antibiotics, and the incidence of drug-resistant strains has, alarmingly, increased in the last decade [118,119]. A first study published by Lowrie et al. in 1999 indicated that, in mice, a M. leprae Hsp65-encoding DNA vaccine, initially designed to prevent infection, could also have a pronounced therapeutic action by switching the immune response from one that is relatively inefficient and gives bacterial stasis to one that kills bacteria [120]. Recently, Zàrate-Bladés et al. performed a comprehensive gene profiling in the lungs of mice infected with M. tuberculosis following DNA hsp65 immunotherapy. The ana lysis suggests that DNA hsp65 therapy can not only boost the Th1 immune response, but also inhibit Th2 cytokines (IL-4) and regulate the intensity of inflammation through fine-tuning of gene expression of various genes, particularly IL-6 [121]. However it is important to note that other studies have shown that postinfection administration of this same DNA vaccine and of a M. tuberculosis Ag85A-encoding DNA vaccine result in the development of classical Koch reactions. Indeed, histological ana lysis of the lungs of the post infectionimmunized mice showed the presence of multifocal discrete regions of cellular necrosis throughout the lung granulomas [122]. Furthermore, in a comprehensive study Derrick et al. reported on the safety of postexposure vaccination of C57BL/6 mice with 12 different M. tuberculosis preparations, four of which were DNA vaccines. In a low-bacterial-burden model (drug treatment for 3 months started 1 month after infection), postexposure vaccination did not induce significant reactivation and only injection of BCG evoked increase in lung inflammatory responses at 1 month after immunization. In addition, although significant increases in lung inflammation were seen for animals injected with the hsp65 DNA vaccine or a M. tuberculosis culture filtrate, no differences in the survival periods were detected between vaccinated and nonvaccinated mice at 10 months postvaccination. [123]. In an active disease model of immunotherapeutic vaccination, significantly more lung inflammation was observed 1 month after administration of hsp65 DNA but none of the vaccine formulations tested at this time increased (or decreased) the lung bacterial burden at this early time point. Furthermore, vaccination of diseased mice with BCG or TB DNA vaccines did not significantly affect mortality rates compared with unvaccinated controls [123]. These results are thus contradictory to the initial results reported by Lowrie et al., who showed that M. tuberculosis-infected BALB/c mice given four doses of pdna encoding Hsp65 of M. leprae demonstrated a rapid and spectacular decline in live bacteria in the spleen and lungs up to 5 months later [120]. On the other hand, they indirectly confirm the findings that the Ag85A DNA vaccine, which is known to induce protective immunity and prevent long-term necrosis in guinea pigs, failed to protect mice when given in an immunotherapeutic model in mice earlier infected by aerosol with M. tuberculosis [124]. When reviewing these types of studies, it is important to highlight several differences in the experimental protocols used, which can lead to different experimental outcomes. For example, it is possible that the genetic background of the mice used in these studies (C57BL/6 vs BALB/c) and the origin of the hsp65 gene (M. tuberculosis vs M. leprae) are responsible for these 1244 Expert Rev. Vaccines 8(9), (2009)

DNA vaccines against mycobacterial diseases Review discrepancies. Other differences relate to the M. tuberculosis strains used for infection (H37Rv vs Erdman), to the routes of infection (intravenous, intraperitoneal or aerosol), to the antibiotics used for the treatment (combination of different antibiotics or single antibiotic) and durations of the treatment. Overall, we feel that a major problem when trying to develop postexposure TB vaccines is the absence of a small animal model that consistently mimics M. tuberculosis latency as developed by the human host. Besides, a lack of standardized experimental protocols for mouse studies makes it difficult to compare results obtained by different groups. The use of DNA vaccines as postexposure vaccines against reactivation TB therefore remains a controversial matter. However, it is important to note here that the observed inflammatory responses cannot be attributed to the fact that the mycobacterial antigens are administered as encoded by DNA vaccines. Indeed, typical Koch reactions have also been observed when protein preparations of mycobacterial antigens or whole-cell preparations were administered to infected mice [125]. Finally, combinations of DNA vaccines administered simultaneously with chemotherapy have been described. In a report by Ha et al., combinations of chemotherapy with a double-gene DNA vaccine composed of Ag85A and PstS-3 completely blocked TB reactivation and significantly prevented a secondary infection when chemotherapy was combined simultaneously [126]. Likewise, a report by Nuermberger et al. has suggested a beneficial effect of the combination of rifapentine, moxifloxacin and DNA encoding Hsp65 in a mouse model of latent M. tuberculosis infection [127]. Recently, Yu et al. have presented data showing that boosting the efficiency of the immune system with a combination of three DNA vaccines encoding the M. tuberculosis antigens (Ag85B, MPT-64 and MPT-83) may be a valuable adjunct for shortening the duration of antibacterial chemotherapy [128]. To our knowledge, no reports are currently available on the potential use of postexposure or immunotherapeutic BU vaccines. Subunit-based DNA vaccines encoding M. ulcerans antigens could have a role in the antibiotic therapy of BU, by shortening the duration of antibiotic treatment. In addition, they could also prevent recurrences and severe forms of this disease [21]. It would therefore be very interesting to assess if such an approach would be applicable in the context of M. ulcerans infection. Similarly, also in the case of of M. avium subspecies paratuberculosis infection, no reports on postexposure or immunotherapeutic vaccines exist. However, promising results have been obtained by treating M. bovis-infected cattle with combinations of DNA vaccines. In this study, Hu et al. have treated 16-week-old M. bovis-infected calves at 2-week intervals with four injections of 1.5 mg of a mix of three pdna vaccines encoding mycobacterial antigens. This resulted in significantly reduced pulmonary bacterial loads compared with the untreated group, or with the BCG-treated group [129]. Expert commentary & five-year view The development of vaccines for the prevention of TB and BU would be highly effective for decreasing the morbidity and mortality caused by these poverty-related human diseases. A number of promising TB candidates have reached the level of clinical trials [130], but thus far no DNA vaccines are among these. Currently, it is not possible to predict whether one or more of them will prove to be efficient. Moreover, recently characterized latency-associated proteins could possibly prove in ongoing studies to be essential protective antigens, and thus will be included in more effective TB vaccines. Continuing research for the identification of protective antigens to be included in new TB vaccines is therefore important, and the DNA vaccination approach is a very powerful tool for this purpose. For BU vaccine development, it is clear that a major problem resides in the minimal amount of research devoted to this disease, attributable to weak financing possibilities for this neglected tropical disease, resulting in a small number of lab oratories active in this field of research. However, increased awareness of the importance of finding a solution to prevent this mutilating disease will most probably boost future BU vaccine development research. A hopeful sign was recently given by the EU in its willingness to fund the FP7-health collaborative project BuruliVac, which aims to identify and develop vaccine candidates for BU disease. Regarding paratuberculosis, the development of a subunit veterinary vaccine would mostly benefit the livestock industry and could additionally have implications related to Crohn s disease control (if the association between MAP infection and this pathology is confirmed by ongoing studies). From a regulatory point of view, advances towards a DNA-based subunit vaccine for paratuberculosis will probably benefit from the fact that a few DNA vaccines are already used for the prevention of some veterinary viral diseases. Research on the possible use of DNA-based subunit vaccines encoding mycobacterial antigens is helpful, not only for research into vaccine development against mycobacterial pathogens, but also for other diseases. Indeed, results of the studies related to the optimization of DNA vaccines against mycobacterial diseases can be applied to the development of DNA vaccines against other infectious and noninfectious diseases, particularly those in which cell-mediated immune responses are needed for protection. On the other hand, a number of protective M. tuberculosis antigens currently evaluated in clinical trials such as the mycolyl-transferase 85A expressed in the recombinant MVA vector [131] have been identified thanks to studies involving immunizations with DNA vaccines. For the development of subunit vaccines against paratuberculosis, antigens specific to MAP have also been defined by the use of DNA vaccines. This could lead to a paratuberculosis subunit vaccine, which would not interfere with the diagnostic tests for paratuberculosis and bovine TB. Unfortunately, until now only a few studies have exploited the potential of DNA vaccines for the prophylaxis of BU. Moreover, as M. ulcerans protective antigens are still poorly defined, more research should be performed and DNA vaccines will surely be a highly valuable tool for that purpose. The best candidates could be included in a BU subunit vaccine, be it protein or plasmid based. In addition, the use of DNA vaccines encoding M. ulcerans antigens should also be tested for immunotherapeutic purposes. This could improve the efficacy and shorten the duration of the antibiotic treatment currently used, which is long-lasting and requires injections. Regarding the development of postexposure TB vaccines, the lack of a reliable latent www.expert-reviews.com 1245