Lysine Auxotrophy Combined with Deletion of the SecA2 Gene Results in a Safe and Highly Immunogenic Candidate Live Attenuated Vaccine for Tuberculosis

Tuberculosis (TB) caused by Mycobacterium tuberculosis remains a major global health problem, despite the widespread use of the M. bovis Bacille Calmette-Guerin (BCG) vaccine and the availability of drug therapies. In recent years, the high incidence of coinfection of M. tuberculosis and HIV, as well as escalating problems associated with drug resistance, has raised ominous concerns with regard to TB control. Vaccination with BCG has not proven highly effective in controlling TB, and also has been associated with increasing concerns about the potential for the vaccine to cause disseminated mycobacterial infection in HIV infected hosts. Thus, the development of an efficacious and safe TB vaccine is generally viewed as a critical to achieving control of the ongoing global TB pandemic. In the current study, we have analyzed the vaccine efficacy of an attenuated M. tuberculosis strain that combines a mutation that enhances T cell priming (ΔsecA2) with a strongly attenuating lysine auxotrophy mutation (ΔlysA). The ΔsecA2 mutant was previously shown to be defective in the inhibition of apoptosis and markedly increased priming of antigen-specific CD8+ T cells in vivo. Similarly, the ΔsecA2ΔlysA strain retained enhanced apoptosis and augmented CD8+ T cell stimulatory effects, but with a noticeably improved safety profile in immunosuppressed mice. Thus, the M. tuberculosis ΔsecA2ΔlysA mutant represents a live attenuated TB vaccine strain with the potential to deliver increased protection and safety compared to standard BCG vaccination.


Introduction
Tuberculosis is a devastating disease, and despite the availability of therapeutic drugs and widespread vaccination with BCG, it remains a serious global health issue, infecting nearly a third of the world population and causing 2-3 million deaths annually [1]. Since initiating BCG vaccination in the 1920s, more than three billion people have been vaccinated with BCG [2]. However, the benefits and drawbacks of BCG have been debated since its early use. Ongoing concerns and controversies include the potential for disseminated infection by BCG in immunosuppressed hosts, loss of sensitivity to tuberculin as a diagnostic reagent, and the failure of BCG to show efficacy in the prevention of pulmonary TB in a number of trials in developing countries [3,4,5]. There have been numerous hypotheses to explain the inadequate protective effect of BCG against pulmonary TB. Factors related directly to the vaccine, such as inappropriate treatment and storage of the vaccine and strain variability have been suggested. Additionally, several deficits of a more fundamental immunological character have been proposed, such as a lack of important antigens, interaction with environmental mycobacteria, and dwindling efficacy of BCG vaccination over time [6,7]. Because various T cell subsets have been shown to participate in the immune responses against mycobacteria, the lack of an effective stimulation of the optimal blend of T cell populations, and particularly the failure to induce CD8 + T cells, may also explain the insufficient levels of immunity promoted by the BCG vaccine strains [7,8]. In fact, increased vaccine efficacy against tuberculosis has been observed in association with enhanced priming of CD8 + T cells in vivo [9,10]. As there is an urgent need for innovative protective vaccine strategies, we must identify and exploit the mechanisms that can facilitate or inhibit immunity to M. tuberculosis, while simultaneously maintaining an adequate safety profile.
In previous work, we described a novel immune-evasion strategy of M. tuberculosis by which secretion of superoxide dismutase A (SodA) inhibits the apoptotic program of infected macrophages, effectively blunting the specific CD8 + T-cell responses to the bacterium. In those studies, inactivation of the M. tuberculosis secA2 gene, which encodes a virulence-associated protein secretion system, enhanced the apoptosis of infected macrophages and increased priming of antigen-specific CD8 + T cells in vivo. Vaccination of mice and guinea pigs with the DsecA2 mutant significantly increased resistance to virulent M. tuberculosis challenge [9]. Thus, inactivation of the secA2 gene represents a novel way to improve the immunogenicity of mycobacterial vaccines. However, in subsequent studies we observed that inactivation of secA2 in BCG did not lead to enhanced immunogenicity (unpublished data) and the DsecA2 mutant retained a moderate level of virulence, making it unsuitable as a vaccine. In order to improve vaccine efficacy and safety of the DsecA2 mutant, we generated multiple double and triple mutants of the DsecA2 mutant strain via incorporation of other attenuating mutations of M. tuberculosis. In the current study, we have evaluated one of the double mutants, DsecA2DlysA, and show here that this strain displays an excellent safety profile in immunocompromised SCID mice while retaining the enhanced immunogenicity of the parental DsecA2 mutant strain.

Results
Improved attenuation of M. tuberculosis DsecA2DlysA in SCID mice The principal advantage of incorporating auxotrophic mutations into vaccine candidates is the associated reduction in virulence. Previously, the DlysA strain of M. tuberculosis was shown to be unable to replicate in vitro in the absence of lysine supplementation of the growth medium, and to be markedly attenuated for growth in vivo in SCID mice, consistent with the inability of the bacteria to scavenge sufficient lysine from the intracellular environment [11,12]. Strain DsecA2DlysA is strictly auxotrophic for lysine, and no growth of the mutant was observed in the absence of lysine supplementation. To investigate the virulence of the double mutant, SCID mice were injected intravenously with 10 5 CFU of either wild-type M. tuberculosis (strain H37Rv), a DsecA2 mutant, or a double mutant strain DsecA2DlysA. As expected, the DsecA2 single mutant of M. tuberculosis showed a high level of virulence in this experiment, as all animals died in less than 50 days. In contrast, a significant reduction of virulence of the DsecA2DlysA strain was observed, as the group infected with this strain had a median survival of .300 days. In comparison, BCG infected SCID mice showed a median survival of ,150 days, and all BCG infected animals succumbed by day 165 (Figure 1). These data clearly indicated that DsecA2DlysA was strongly attenuated and even less virulent than the standard BCG vaccine strain in the context of severe immunodeficiency.

Induction of macrophage apoptosis by M. tuberculosis DsecA2DlysA
The secA2 gene encodes an accessory secretion factor required for the export of superoxide dismutase protein (SodA), an enzyme that catalyzes conversion of superoxide anions to hydrogen peroxide. In previous work we have shown that deletion of secA2 gene results in enhanced apoptosis of infected macrophage, an effect that was reversed by the restoration of secretion of SodA through a secA2 independent pathway [9]. To test whether the DsecA2DlysA has retained this pro-apoptotic property, transformed human monocyte/macrophage cells (THP-1 cells) were infected with at an MOI of 10 bacilli per macrophage with H37Rv, DsecA2, DlysA, or DsecA2DlysA. At 72 hours following infection, quantitative assessment of apoptosis by staining for DNA strand breakage (TUNEL) showed that both DsecA2 and DsecA2DlysA demonstrated high levels of apoptosis, whereas WT M. tuberculosis and the single mutant auxotroph DlysA did not ( Figure 2). These results indicated that the defect in apoptosis inhibition resulting from the loss of DsecA2 was retained in the DsecA2DlysA double mutant.
Augmented CD8 + T cell priming following immunization with DsecA2DlysA As illustrated in earlier work, an important consequence of blocking apoptosis is an effective reduction of the priming of CD8 + T cell response to M. tuberculosis. M. tuberculosis antigens are produced within a phagosomal compartment which is sequestered from the classical cytoplasmic pathway of antigen processing. As with other vacuolar pathogens, such as Salmonella, display of these antigens on MHC class I molecules likely occurs by the uptake and presentation of apoptotic material by bystander dendritic cells (DC) [13]. Consistent with this idea, we have previously shown that mice vaccinated with the DsecA2 mutant strain exhibited a marked enhancement of CD8 + T cell priming. To test whether the DsecA2DlysA mutant strain retained the enhanced immunogenicity observed in the parental DsecA2 strain, we used a CD8 + T cell adoptive transfer strategy similar to that described previously [9]. Recipient mice injected with low numbers (5610 5 ) of OT-I T cells were vaccinated subcutaneously with recombinant M. tuberculosis strains expressing the 19-kDa lipoprotein-SIINFEKL fusion protein. This experimental design enabled accurate quantification of antigenspecific CD8 + T cells by tetramer staining in mice vaccinated with H37Rv-OVA, DsecA2-OVA, DlysA-OVA, and DsecA2DlysA-OVA. As anticipated, mice vaccinated with H37Rv-OVA and DlysA-OVA showed little detectable expansion of transferred T cells, while the pro-apoptotic strains, DsecA2-OVA, and DsecA2DlysA-OVA induced significant expansion of antigen specific CD8 + T cells ( Figure 3A). We also measured expression of CD62 ligand (CD62L) and CD44 on tetramer + cells to distinguish between naive (CD44 low CD62L high ), central memory (CD44 high CD62L high ), and effector memory cells (CD44 high CD62L low ). Our results showed enhanced expansion of both subsets in mutant strains deficient in secA2 ( Figure 3B, C). In addition, we were able to demonstrate a measurable induction of endogenous CD8 + T cells in mice vaccinated intravenously with DsecA2-OVA and DsecA2DlysA-OVA, with little detectible expansion of tetramer positive cells in mice vaccinated with H37Rv-OVA and DlysA-OVA (Figure 4). Protective immunity against virulent M. tuberculosis challenge following vaccination with DsecA2DlysA As discussed above, although DlysA has a compelling safety profile in immunocompromised animals, and performs much better than BCG in our studies conducted in SCID mice, its poor growth in vivo comes at the cost of weaker immunogenicity. In order to determine whether the enhanced T cell priming observed with DsecA2DlysA correlated with improved performance as a live vaccine strain, we conducted immunization and challenge studies designed to evaluate potential vaccine efficacy of the double mutant. C57BL/6 mice that were either naive or immunized 8 weeks earlier by subcutaneous immunization with live DsecA2DlysA, DsecA2 or BCG (Pasteur strain) were challenged by low-dose aerosol infection with virulent M. tuberculosis (Erdman). In naive mice, substantial growth in the lungs and dissemination to spleens were detected within 1 month of challenge. In contrast, vaccination with DsecA2DlysA, DsecA2 and BCG considerably reduced M. tuberculosis bacillary loads in both lungs and spleens of aerosol-challenged mice as compared with naive controls ( Figure 5A).
Immunization with DsecA2 and DsecA2DlysA also showed a more prolonged effect on control of M. tuberculosis infection, which was still apparent 3 months after challenge. More importantly, the protection evoked by DsecA2DlysA immunization was significantly greater than that obtained with BCG vaccination, as measured by bacterial burden in both the lung and spleen at 3 months post challenge (0.66-log reduction in lungs of DsecA2DlysA versus BCG vaccinated mice; P,0.001, 0.68-log reduction in spleens DsecA2-DlysA versus BCG vaccinated mice; P,0.01). Interestingly, DsecA2DlysA vaccinated mice also showed significantly lower lung bacillary loads than mice vaccinated with DsecA2 (0.35-log reduction in lungs of DsecA2DlysA versus DsecA2 vaccinated mice; P,0.05) ( Figure 5A), an effect that may relate to the enhanced stability of memory cells in response to non-persistent live vaccines [14,15,16]. The one month time point represents the peak bacillary burden and the 3 month time point represents the set point of the bacillary burden in the chronic phase of infection. Histopathological examination of the lungs from mice immunized with BCG, DsecA2, or DsecA2DlysA showed relatively mild inflammation with less severe spreading lung lesions, less extensive granulomatous pneumonia, and reduced consolidation compared with naive mice ( Figure 5B).

Discussion
In recent years a number of attenuated strains of M. tuberculosis, many constructed through the engineered deleting of genes important for growth in vivo, have been shown to confer protection against challenge with virulent M. tuberculosis [17,18]. The appeal of auxotrophic mutant strains, from the standpoint of vaccine development, relates to their marked growth attenuation in vivo and exceptional safety profile in immunocompromised animals. However, a significant drawback of many of these strains is poor immunogenicity in immunocompetent animals, which appears to be a function of reduced growth and decreased antigen load. Consistent with this view, the DlysA mutant strain, while shown to be profoundly attenuated in SCID mice, required multiple rounds of immunization to achieve levels of protection comparable to BCG in immunocompetent mice [19]. In contrast, we have shown in the current study that a single immunization with the DsecA2DlysA strain conferred markedly enhanced protection from challenge with virulent M. tuberculosis when compared to BCG. A similar result was previously observed for the parental DsecA2 strain, which exhibited enhanced CD8 + T cell priming when compared with wild-type M. tuberculosis, while conferring enhanced protection to vaccinated mice and guinea pigs compared to BCG. Enhanced immunogenicity of DsecA2 correlated with the loss of apoptosis inhibition by the mutant strain. This proapoptotic phenotype was preserved in the DsecA2DlysA strain, along with the enhanced CD8 + T cell stimulating effect. In the present study, we confirmed our previous finding that DsecA2 conferred enhanced protection against virulent challenge when compared with BCG, and demonstrated that vaccination with the DsecA2DlysA double mutant strain also gave protection that was significantly better than that achieved with BCG. In fact, DsecA2DlysA vaccination led to a significantly greater The explanation for this augmented protective efficacy of DsecA2DlysA vaccination may relate to the inability this strain to persist in vivo, leading to more rapid antigen clearance after T cell priming and more stable memory T cell populations [20]. This is a substantial difference from the parental DsecA2 single mutant, which is known to be only moderately attenuated for growth in mice and most likely persists at least at low levels for many weeks and perhaps permanently [21]. Research on T cell responses to viral pathogens including LCMV, Hepatitis C, and SIV, suggests that the persistence of antigen stimulation associated with chronic infection can lead to the functional impairment or ablation of effector and memory CD8 + T cell responses. In this model, too little antigen leads to suboptimal responses, whereas an overabundance or prolonged expression of antigen leads to a suboptimal reservoir of memory T cells [22,23,24,25]. Through the pairing of the augmented CD8+ T cell priming properties of the DsecA2 mutation with the restrained growth imparted by DlysA, we may have achieved a more favorable formulation to generate robust, yet stable, memory responses in vaccinated animals.
A salient advantage of the auxotrophic double mutant, illustrated here, is its marked attenuation in immunocompromised mice. Due to the high incidence of HIV-infection in regions where TB is endemic, safety of immunocompromised animal models has become an important consideration for experimental vaccines. Even BCG, widely regarded as a safe vaccine, has been shown to progress to disease in a significant subpopulation of HIV infected individuals. In a study of BCG vaccinated HIV-positive children conducted in 2003, 10% of mycobacterial disease was attributed to BCG [26]. HIV infection leads to a disproportionate loss of CD4 + T cells, which in experimental models have been shown to be essential for controlling infection with BCG [27]. Although CD4 + T cells are critical for protective anti-TB protective immunity [28], CD8 + T cells have been shown to play an important role in the control of TB, particularly in CD4-deficient mice [5,6,29]. Thus, the DsecA2DlysA strain, with an improved safety profile and augmented CD8 + T cell immunogenicity, may serve as an effective substitute for BCG in the HIV infected population.
Consistent with results from our previous work characterizing the immunogenicity of DsecA2, the mechanisms underlying the augmented vaccine properties of DsecA2DlysA appear to stem, at least in part, from a loss of apoptosis inhibition, enhanced CD8 + T cell priming and memory T cell generation. In addition, by coupling the moderately attenuating DsecA2 mutation with DlysA we were not only able to achieve a superior safety profile, but also significantly improved control of infection relative to the parental strain. The rational design of novel, live attenuated vaccines for  tuberculosis depends on integrated disruption of multiple pathogen virulence strategies. This study clearly demonstrates a proof of this principle by combining DsecA2 with a second mutation that interferes with virulence of M. tuberculosis, thus generating a novel attenuated candidate mycobacterial vaccine strain with significantly improved safety and efficacy.

Construction of recombinant mycobacterial strains
The double mutant strain DsecA2DlysA was constructed using specialized transduction to disrupt the chromosomal copy of the lysA gene of an unmarked secA2 clone, as described [30]. To generate OVA expressing strains of wild type M. tuberculosis (H37Rv-OVA) or of mutants (DsecA2-OVA, DlysA-OVA, and DsecA2DlysA-OVA), the wild type and mutant strains were transformed with the plasmid p19OVA which encodes the M. tuberculosis 19kD lipoprotein fused at the C-terminus with the sequence for ovalbumin that includes the MHC class I presented epitope SIINFEKL [9]. Transformations of M. tuberculosis were performed according to a published protocol [31].
Apoptosis studies in THP-1 cells THP-1 cells (ATCC) were grown in RPMI-1640 (Gibco BRL; Invitrogen) supplemented with 10% FBS, 1% HEPES, 1% nonessential amino acids and essential amino acids. THP-1 cells were treated with 8 ng/ml PMA (Sigma-Aldrich) in order to induce differentiation into macrophage-like cells. Differentiated THP-1 cells were plated in 6-well plates at 2610 6 cells per well and allowed to adhere overnight. Dispersed bacilli were incubated with cells for 3 hours using a multiplicity of infection (MOI) of 10 bacteria to 1 THP-1 cell. The cells were then washed twice with PBS and incubated at 37uC in a 5% CO 2 incubator for 72 hours. The presence of apoptotic cells in the cell cultures was analyzed with In Situ Cell Death Detection Kit, Fluorescence (Roche Diagnostics), following the manufacturer's instructions.

Vaccination and challenge studies
All animal studies were approved by the institutional animal care and use committees of the Albert Einstein College of Medicine (A3312-01), the Center for Biologics Evaluation and Research, and the Duke University (A071-10-03). C57BL/6 mice were vaccinated subcutaneously with 1610 6 CFU of M. tuberculosis DsecA2, DsecA2DlysA, or M. bovis BCG-Pasteur as described previously [32]. Aerogenic challenge was performed 2 months later using a Glas-Col inhalation chamber to deliver 50-100 CFU per animal of M. tuberculosis Erdman strain. Mice were sacrificed at 1 and 3 months after challenge. Lungs and spleens of individual mice were aseptically removed and homogenized separately in 5 ml normal saline plus 0.05% Tween-80 using a Seward Stomacher 80 blender (Tekmar). The homogenates were diluted serially and plated on Middlebrook 7H10 agar. Lung tissues were processed for histopathology using standard paraffin fixation, sectioning, and H&E staining. In the survival study, animals infected with M. tuberculosis Erdman were observed daily until they died or became moribund and were euthanized.

Statistics
Two-way ANOVA with Bonferroni's post test and one-way ANOVA with the Tukey post test were used for analyses using GraphPad Prism version 5 for Windows (GraphPad; [www. graphpad.com]). P values of less than 0.05 were considered significant.