MTBVAC-Based TB-HIV Vaccine Is Safe, Elicits HIV-T Cell Responses, and Protects against Mycobacterium tuberculosis in Mice

The tuberculosis (TB) vaccine MTBVAC is the only live-attenuated Mycobacterium tuberculosis (Mtb)-based vaccine in clinical development, and it confers superior protection in different animal models compared to the current vaccine, BCG (Mycobacterium bovis bacillus Calmette-Guérin). With the aim of using MTBVAC as a vector for a dual TB-HIV vaccine, we constructed the recombinant MTBVAC.HIVA2auxo strain. First, we generated a lysine auxotroph of MTBVAC (MTBVACΔlys) by deleting the lysA gene. Then the auxotrophic MTBVACΔlys was transformed with the E. coli-mycobacterial vector p2auxo.HIVA, harboring the lysA-complementing gene and the HIV-1 clade A immunogen HIVA. This TB-HIV vaccine conferred similar efficacy to the parental strain MTBVAC against Mtb challenge in mice. MTBVAC.HIVA2auxo was safer than BCG and MTBVAC in severe combined immunodeficiency (SCID) mice, and it was shown to be maintained up to 42 bacterial generations in vitro and up to 100 days after inoculation in vivo. The MTBVAC.HIVA2auxo vaccine, boosted with modified vaccinia virus Ankara (MVA).HIVA, induced HIV-1 and Mtb-specific interferon-γ-producing T cell responses and polyfunctional HIV-1-specific CD8+ T cells producing interferon-γ (IFN-γ), tumor necrosis factor alpha (TNF-α), and CD107a in BALB/c mice. Here we describe new tools to develop combined vaccines against TB and HIV with the potential of expansion for other infectious diseases.


INTRODUCTION
Today, tuberculosis (TB) has reached alarming proportions. An estimated 10 million people have developed TB in 2017 and 9% were people living with HIV (72% in Africa). There were an estimated 1.3 million TB deaths among HIV-negative people and an additional 300,000 deaths among HIV-positive people, as reported by the World Health Organization (WHO) 1 in 2018. TB is poverty related with a major burden in the poor and developing parts of the world, and it is aggravated by the HIV-AIDS pandemic, which greatly increases the risk of the infection evolving into active TB disease.
HIV-AIDS is a major global public health issue. Between 2010 and 2016, new HIV infections fell by 11% in adults and 47% in children, and AIDS-related deaths fell by 48% since the peak in 2005. This achievement was the result of great efforts by national HIV programs supported by civil society and a range of development partners. 2 However, sub-Saharan Africa accounted for 64% of new HIV infections in 2016, and, even though it is encouraging that 1.6 million people are currently receiving treatment in resource-poor settings, ensuring universal access to antiretroviral therapy still represents an enormous challenge. 3 Thus, the development of effective, safe, and affordable vaccines against both diseases could have a tremendous impact on public health.
The risk of active TB is estimated to be between 16 and 27 times greater in people living with HIV than among those without HIV infection. 4 Mycobacterium bovis bacillus Calmette-Guérin (BCG) has been the only licensed vaccine against TB for more than 90 years, 5 but the BCG-induced protective effects against pulmonary disease over all ages are variable. 6,7 Nevertheless, BCG 8 vaccination has several beneficial effects: (1) BCG vaccination reduces rates of Mycobacterium tuberculosis (Mtb) infection, aiding in the decrease of the pool of latent infections from which future cases of active disease may arise; 9 (2) BCG provides strong protection against disseminated forms of the disease in infants and young children; 10,11 (3) BCG revaccination of adolescents may provide additional benefits for the prevention of TB; 12 and (4) BCG vaccination reduces all-cause mortality through beneficial non-specific (heterologous) effects on the immune system. 13,14 These four effects strengthen the motivation for the inclusion of BCG in the global vaccination program. 15 We previously constructed and characterized MTBVAC, the first and only live-attenuated Mtb-based vaccine candidate in clinical development against TB disease in the pipeline. MTBVAC contains two independent deletions in the phoP and fadD26 genes without antibiotic resistance markers, and it fulfills the Geneva consensus requirements for progressing into clinical trials. 16 The vaccine candidate MTBVAC was safe, and it conferred superior protection in different animal models compared to the licensed BCG reference strain in use today. To date, phase I trials in adults and neonates (ClinicalTrials.gov: NCT02013245 and NCT02729571) have been successfully completed, and phase II trials for dose definition, at birth and in adults with and without latent TB, are in progress (ClinicalTrials.gov: NCT02933281 and NCT03536117). Clinical results showed that MTBVAC was immunogenic, in a dose-dependent manner, and it had a similar safety profile as that of BCG. 16,17 It is well known that there is strong evidence in favor of a role for HIV-1-specific T cell responses in the control of HIV-1 replication. 18,19 One promising approach for T cell induction is M. bovis BCG as a bacterial live recombinant vaccine vehicle. Specific humoral and cellular immune responses against HIV-1 have been detected after immunization of mice with recombinant BCG (rBCG) expressing HIV-1 antigens. [20][21][22][23][24] We previously developed several rBCG HIV-1 vaccine candidates with the aim of inducing protective cellmediated responses. Confirming the efficacy of an HIV-1 vaccine candidate in humans is as of yet not possible in animal studies alone. Achieving protection against HIV infection in humans following active vaccination, and subsequently identifying the correlates of protection, would allow the validation of protection in animal models. Our aim is to induce a strong CD8+ T cell response capable of aiding and complementing the protective efficacy of antibody-based vaccines, while providing viral control in the case of an infection occurring.
Our starting platform was based on a heterologous rBCG prime and recombinant modified vaccinia virus Ankara (MVA) boost regimen delivering a common immunogen called HIVA (MVA.HIVA), which is derived from consensus Gag protein of HIV-1 clade A, prevalent in central and eastern Africa, and a string of human CD8 + T cell epitopes. 25 Recently, we engineered a new BCG.HIVA 2auxo vaccine strain harboring an antibiotic-free plasmid selection system and maintenance. The BCG.HIVA 2auxo vaccine in combination with MVA.HIVA was safe, and it induced HIV-1 and Mtbspecific interferon-g-producing T cell responses in adult BALB/c mice. 26 In this study we have constructed a novel live-attenuated vaccine for HIV-1 and TB that is vectored by a lysine auxotroph of MTBVAC, MTBVAC.HIVA 2auxo . This is an innovative approach to develop bivalent TB and HIV vaccines that could be administered at birth and with the potential to confer protection against both diseases.

Construction and Characterization of a Lysine Auxotroph of MTBVAC, the MTBVACDlys Strain
For MTBVACDlys construction, the previously described recombineering-based technique was used. 27 Rv1293 (lysA) gene, which codes for the last enzyme involved in Lysine (Lys) synthesis, 28 was inactivated by homologous recombination with a PCR product containing a kanamycin (Km) resistance cassette disturbing the lysA gene (Figure 1A). To ensure proper selection of recombinants, the final MTBVAC transformed with the homologous PCR product lysA-Km was plated on 7H10 complete medium supplemented with Km and Lys. Correct recombination was confirmed by PCR using three different pairs of primers, which amplify the complete recombined region (Lys-fw/Lys-rv), the upstream (Lys-fw/km-OUT1-rv), or the downstream (km-OUT2-fw/Lys-rv) region ( Figure 1B). After MTBVACDlys construction, Lys auxotrophy was confirmed by plating on 7H10-ADC with and without Lys supplementation ( Figure 1C) and also by colony-forming unit (CFU) enumeration after removing Lys from medium (data not shown). Results showed the absence of MTBVACDlys growth in non-lysine-supplemented plates, whereas the MTBVAC strain grew at a similar level in both supplemented and non-supplemented plates. Accordingly, this auxotrophic MTBVACDlys strain was used in the subsequent experiments to generate recombinant vaccines expressing the HIVA immunogen.

Construction and Characterization of the MTBVAC.HIVA 2auxo Vaccine Strain
The plasmid p2auxo.HIVA ( Figure 2A) 26 was transformed into the MTBVACDlys host strain to generate the recombinant MTBVAC.HIVA 2auxo . The selection of positive recombinant MTBVAC.HIVA 2auxo colonies was performed by culturing the MTBVACDlys transformants on Middlebrook agar 7H10 medium without lysine supplementation. The MTBVAC.HIVA 2auxo strain harboring the lysine-complementing gene abolished the requirement for exogenous lysine, and colonies were observed in non-lysinesupplemented agar plates ( Figure 2B).
The expression of the full-size chimeric 19-kDa signal sequence-HIVA protein (total weight, 64 kDa) was confirmed by western blot analysis of the MTBVAC.HIVA 2auxo cell lysates ( Figure 2C) To assess the in vitro stability of the p2auxo.HIVA plasmid, subcultures of MTBVAC.HIVA 2auxo on selective media (no lysine supplementation) were carried out every 7 days. The maintenance of the p2auxo.HIVA plasmid DNA was evaluated by PCR analysis of HIVA and GlyA DNA-coding sequences. Bands corresponding to the HIVA DNA-coding sequence ( Figure 3A) and to the E. coli GlyA-coding sequence ( Figure S1A) were observed in all 6 MTBVAC.HIVA 2auxo subcultures (42 bacterial generations), indicating that there were no major genetic rearrangements in the HIVA and glyA genes of MTBVAC.HIVA 2auxo vaccine strain over the subsequent subculturing passages. In vivo stability of p2auxo.HIVA plasmid in MTBVAC.HIVA 2auxo was assessed in severe combined immunodeficiency (SCID) mice used in the safety trial. Homogenized spleens were plated on Lys-supplemented medium, and p2auxo.HIVA presence in the mycobacterial burden was analyzed by colony PCR using primers to detect the HIVA DNAcoding sequence ( Figure 3B) and the glyA gene ( Figure S1B). The analysis showed that 95.5% of the colonies retained the plasmid during in vivo infection. We previously demonstrated that heterologous BCG.HIVA prime boosted with MVA.HIVA elicited high-quality HIV-1-specific T cell responses. 26,29,30 In this study, we evaluated the specific HIV-1 T cell responses in adult BALB/c mice after intradermal immunization with MTBVAC.HIVA 2auxo or MTBVAC.Ø 2auxo prime and intramuscular MVA.HIVA boost ( Figure 4A). The intradermal route mimics the administration performed in human BCG vaccination, and it has been shown to elicit a higher HIV-1-specific CD8+ T cell response in adult BALB/c mice. 31 The immunogenicity readout was focused on the P18-I10 epitope, an immunodominant cytotoxic T-lymphocytes (CTL) epitope derived from HIV-1 Env and H-2D d murine restricted, which was fused to HIVA immunogen to evaluate the immunogenicity in mice.
On day 0, adult mice were either left unimmunized or primed with MTBVAC.HIVA 2auxo or MTBVAC.Ø 2auxo , and on week 6 the animals received an MVA.HIVA boost. Mice were sacrificed on week 8, and the functional quality of the elicited CD8+ T cells to produce interferon-g (IFN-g) and tumor necrosis factor alpha (TNF-a) and to degranulate (surface expression of CD107a) in response to P18-I10 peptide stimulation was measured by intracellular cytokine staining (ICS) ( Figure 4B). We observed in adult mice that MTBVAC.HIVA 2auxo prime and MVA.HIVA boost induced higher frequencies of P18-I10 epitope-specific CD8+ splenocytes producing IFN-g, TNF-a, and CD107 than mice primed with MTBVAC.Ø 2auxo , MVA.HIVA alone, or naive mice. We found that MTBVAC. HIVA 2auxo prime and MVA.HIVA boost induced higher frequencies of trifunctional specific CD8+ T cells compared with the MTBVAC.Ø 2auxo priming and MVA.HIVA boost and with MVA. HIVA alone ( Figure 4C). The capacity of splenocytes from vaccinated mice to secrete IFN-g was also assessed by the enzyme-linked immunosorbent spot (ELISPOT) assay. We observed the highest frequency of specific cells secreting IFN-g when stimulated with P18-I10 in mice primed with MTBVAC.HIVA 2auxo and boosted with MVA.HIVA, 1,280 spot-forming units (SFU)/10 6 splenocytes, compared to 1,043 SFU/10 6 splenocytes obtained when mice were primed with MTBVAC.Ø 2auxo and 1,095 SFU/10 6 splenocytes when mice were only boosted with MVA.HIVA ( Figure 4D). The capacity of splenocytes from vaccinated mice to secrete IFN-g after overnight stimulation with the Mtb-purified protein derivative (PPD) was also assessed by ELISPOT. The median SFUs per 10 6 splenocytes were similar in mice primed with MTBVAC.HIVA 2auxo and MTBVAC.Ø 2auxo (102 and 86 SFU/10 6 splenocytes, respectively; Figure 4E).

MTBVAC.HIVA 2auxo Prime and MVA.HIVA Boost Were Well Tolerated in Mice
As shown in Figure 5, the body mass was monitored over time and recorded to depict any adverse events and body mass loss due to A B C vaccination. To detect vaccine-derived adverse events, a 12-week period between MTBVA-C.HIVA 2auxo and MVA.HIVA boost was established for this trial. Importantly, no statistically significant difference (by ANOVA) was observed between the vaccinated mouse groups and the control mouse group at the final time point. Furthermore, between weeks 1 and 14, the body mass monitored in all vaccinated mouse groups was found to lie between the mean ± 2 SD body mass curves in control mice. It is also important to mention that no mice died during the trial, and no local adverse events or associated systemic reactions were observed.

MTBVAC.HIVA 2auxo Protective Efficacy against M. tuberculosis in Mice Was Similar to MTBVAC
We evaluated the efficacy of the bivalent vaccine strain MTBVAC. HIVA 2auxo with respect to the parental MTBVAC vaccine. Groups of 6 C57BL/6 mice were left unimmunized or vaccinated with MTBVAC, MTBVAC.HIVA 2auxo , or MTBVACDlys by subcutaneous injection, a route previously used for efficacy studies in mouse models. 32 At 7 weeks post-vaccination, the mice were challenged with the pathogenic H37Rv strain by the intranasal route. Bacterial load in lungs and spleens was examined 4 weeks post-challenge by plating homogenized organs on complete 7H10 medium ( Figure 6). In all vaccinated groups, the bacterial reduction was significant with respect to the unvaccinated group, both in lungs and spleens. The auxotrophic strain MTBVACDlys, which was expected not to be able to survive without lysine, also displayed significant protection against Mtb H37Rv when compared to the naive group. No differences were found between the different MTBVAC strains tested, which validates the protective behavior of MTBVAC.HIVA 2auxo vaccine against Mtb despite the genetic manipulations introduced.

MTBVAC.HIVA 2auxo Was Highly Attenuated in the SCID Mouse Model
As well as affecting vaccine efficacy, genetic manipulation may also affect attenuation of live vaccines. 32 With the aim of corroborating the attenuation status of MTBVAC.HIVA 2auxo and MTBVACDlys strains, SCID mice were inoculated with 10 6 CFU by the intraperitoneal route, and the survival of animals was monitored (Figure 7). SCID mice are the reference model for safety assessments of live vaccines in preclinical TB studies, as recommended by regulatory bodies. 33 Intraperitoneal, as well as intravenous, administration is a systemic inoculation route that allows rapid dissemination of the bacteria and, thereby, virulence assessments. 32 The auxotrophic MTBVACDlys strain showed a hyper-attenuated profile; all mice inoculated with this strain survived until the endpoint of the experiment at week 31. Bacterial burden per spleen of these SCID mice vaccinated with MTBVACDlys at week 31 was approximately 5 Â 10 3 CFU ( Figure S2), which demonstrated that this strain survived in vivo.
When we analyzed survival time of MTBVAC.HIVA 2auxo -vaccinated mice, data revealed a marked attenuation profile (they survived approximately 160 days) when compared to the BCG-vaccinated mice (deceased by day 90) and the MTBVAC-vaccinated mice (deceased by day 120).

DISCUSSION
Despite the progress made in the development of a safe, effective, and affordable vaccine against HIV-1 and TB shortly after birth, the prevention of mother-to-child HIV-1 transmission via breast milk and childhood TB still remain great challenges.
The use of mycobacteria as a vaccine vector is an attractive option; on top of the previously mentioned advantages (cheap mass production, good safety profile, suitable for neonates, etc.), it induces a potent Th1 type immune response (the central defense mechanism against intracellular pathogens) in humans and mice. 34,35 Three experimental systems must be orchestrated to develop a recombinant Mycobacterium-based HIV vaccine: (1) a live vaccine vehicle based on mycobacteria, (2) an E. coli-mycobacterial expression vector without antibiotic resistance markers, and (3) an HIV immunogen design. In this study, we have engineered a novel live-attenuated vaccine for HIV-1 and Mtb infection that is vectored by a lysine auxotroph of MTBVAC, 16 which expresses the HIV-1 clade A-derived immunogen HIVA. 25 The use of mycobacterial vectors 20,36,37 for antigen expression in M. bovis BCG, 20,38,39 M. smegmatis, 40 or M. vaccae 41,42 has been documented. Through a Barcelona-Oxford collaboration, we previously engineered a mycobacterial vaccine platform for HIV-TB by using lysine auxotrophic strains of BCG as vectors: an E. colimycobacterial expression vector containing an antibiotic-free selection system and different HIV immunogen designs to improve the specific HIV-1 immunogenicity. 26,29,31,43,44  The auxotrophy-complementation strategy has also been used in the context of other intracellular pathogens as Listeria monocytogenes to express simian immunodeficiency virus (SIV) or HIV antigens. 45 Despite being highly effective and safe in mice, including neonatal animals, the success observed in the murine model was not translatable upon immunogenicity and efficacy assessment of the Lmdd-BdopSIVgag in non-human primates. 46 This should be taken into consideration during further stages of MTBVAC.HIVA 2auxo development.
In 2014, the promising live-attenuated Mtb vaccine, MTBVAC, was developed. 16 It conferred improved immunogenicity and protection and had a higher safety profile compared to BCG in preclinical studies. Furthermore, MTBVAC has advanced through phase I (ClinicalTrials.gov: NCT02013245 and NCT02729571) 17 and into phase II (ClinicalTrials.gov: NCT02933281 and NCT03536117) clinical trials.
It is important to recall that, of the 1,603 experimentally validated T cell epitopes present in M. tuberculosis, 433 of these epitopes were lost during the in vitro attenuation of BCG. Conversely, MTBVAC maintains the whole antigenic repertoire of the human pathogen Mtb. 47 Further, MTBVAC displayed increased secretion and immunogenicity against some key antigens (such as those of the Ag85 complex) as a consequence of the phoP mutation. [48][49][50] Thus, MTBVAC may well be a better platform than BCG when it comes to the expression of heterologous antigens, based on its higher immunogenicity.
The MTBVAC.HIVA 2auxo and MTBVAC.B 2auxo strains constructed in this study grew properly without lysine supplementation in the medium, which demonstrated the usefulness of the auxotrophycomplementation system for the selection and maintenance of plasmids in Mtb-based vaccines. 26,51 Use of the lysine auxotrophlysine complementation system was previously shown to increase plasmid stability and to prevent extensive genetic rearrangements in recombinant mycobacteria. 26,43,51 Plasmid stability in recombinant mycobacteria has always been a critical and controversial issue. Méderlé et al. 52 evaluated the genetic stability of recombinant BCG strains harboring episomal or integrative vectors expressing nef and gag genes from SIV. They observed a higher genetic stability in vivo and in vitro as well an increased duration of heterologous gene expression in vivo using the integrative plasmid. 52 Recently, we published that the in vitro stability of the integrative plasmid p2auxo.HIVA int was increased 4-fold compared with the BCG strain harboring the episomal plasmid. 30 These results were in concordance with Méderlé et al. However, as there are more copies of episomal vector per cell, higher levels of recombinant protein can be expressed than from integrative vectors. Although our previous results and a number of others have demonstrated that integrative vectors are more stable than episomal vectors, 52,53 it has also been demonstrated that an initial, high level of antigen expression is necessary to prime an immune response. 52 In this study, we demonstrated that the episomal p2auxo.HIVA plasmid transformed into the MTBVACDlys strain was retained in vivo over 220 days in SCID mice and after 42 bacterial generations in vitro. Sequencing of the recovered plasmids in vitro and in vivo would be desirable to accurately confirm the genetic stability of p2auxo.HIVA plasmid at the nucleotide level.
We assessed the HIV-1-and Mtb-specific T cell-mediated immune responses elicited after BALB/c prime-boost immunization with MTBVAC expressing HIVA immunogen (MTBVAC.HIVA 2auxo ). Mice were primed with MTBVAC.HIVA 2auxo or MTBVAC.B 2auxo (plasmid with no HIVA DNA-coding sequence) and boosted with MVA.HIVA. This way, we were able to detect the non-specific immune responses due to intrinsic immunogenic properties of mycobacteria. 54 The frequencies of CD8+ T cells producing IFN-g, TNF-a, and CD107a were higher in mice vaccinated with MTBVAC.HIVA 2auxo in comparison with mice primed with MTBVAC.B 2auxo . A similar profile was observed in spleen cells producing IFN-g after P18-I10 peptide stimulation measured by ELISPOT assay. These results are in concordance with our previously published data using BCG.HIVA 2auxo and BCG wild-type as priming agents. 26 The extent of T cell polyfunctionality was correlated to protection against leishmaniasis in mice, HIV-1 in humans, and SIV in non-human primates. 55 It has also been demonstrated that the magnitude and polyfunctionality of virus-specific CD8+ T cell responses were associated with the control of viral replication after SHIV-89.6P challenge in rhesus monkeys. 56 The construction of Mtb auxotrophs (DpanCD, leuCD, and secA2) expressing HIV or SIV antigens with the aim of developing a dual pediatric vaccine against TB and HIV has been described as oral vaccination of macaques at birth. 57 However, unexpectedly, vaccinated infants required fewer SIV exposures to become infected compared to naive controls. Considering that the current TB vaccine, BCG, can induce potent innate immune responses and confer pathogen-unspecific trained immunity, they hypothesized that an imbalance between enhanced myeloid cell function and immune activation might have influenced the outcome of oral SIV challenge in AMtb-SIVvaccinated infants. Ideally, an appropriately targeted specific response directed toward beneficial epitopes of HIV would overcome the dangers of trained immunity while maintaining its positive aspects, which have been linked to decreased mortality in children receiving BCG, as well as lead to the control or prevention of HIV infection. 58 When developing a dual vaccine against TB and HIV, it is important to confirm that the inclusion of HIV immunogens does not increase the metabolic burden in the TB vaccine that could affect its protective efficacy. We have already described this issue in previous publications, demonstrating that recombinant BCG-based HIV vaccine conferred the same level of protection as the wild-type. 29 In the present study, we confirm that MTBVAC and MTBVAC.HIVA 2auxo confer equivalent levels of protection against Mtb challenge and, consequently, vaccine efficacy is not compromised after genetic manipulations for HIVA expression.
The recombinant MTBVAC.HIVA 2auxo strain showed an increased safety profile in comparison with the parental MTBVAC strain. This finding might support the potential use of this vaccine in individuals at risk of immunosuppression. On the other hand, the auxotrophic MTBVACDlys strain had a highly attenuated profile, with 100% survival in this group after 220 days of inoculation. This hyper-attenuated profile of MTBVACDlys was observed with other auxotrophic BCG and Mtb strains, 59-61 due to lower viability and growth limitation without amino acid supplementation in vivo. Despite the growth limitation, the strain could be isolated in the spleens of SCID mice at the end of the trial, demonstrating that the auxotrophic strain was not fully cleared on day 220 after inoculation. It may suggest that the increased attenuation profile of the recombinant MTBVAC.HIVA 2auxo could be due to episomal plasmid loss. However, in our study, the persistence of plasmid was 95% in the isolated colonies from spleens of SCID mice inoculated with MTBVAC.HIVA 2auxo .
We also demonstrated that MTBVAC.HIVA 2auxo prime and MVA. HIVA boost were well tolerated in adult BALB/c mice by monitoring and recording body mass over time. The findings were along the lines of previous observations with recombinant BCG expressing HIVA. 26 In conclusion, the development of a recombinant MTBVAC-based HIV-TB vaccine may provide a new and improved tool for mycobacterial-based vaccine design. In the future, MTBVAC could be further developed as a vector to express optimized HIV immunogens and utilized in prime-boost vaccination protocols along with new boosting agents. It could, furthermore, be used as a novel mycobacterial vaccine platform for infectious diseases, such as malaria, whooping cough, and other tropical diseases, to prime protective immune responses shortly after birth.

Bacterial Strains and Culture Conditions
Mycobacterial cultures were grown in Middlebrook 7H9 broth medium or on Middlebrook agar 7H10 (Becton Dickinson) medium supplemented with albumin-dextrose-catalase (ADC; Difco Laboratories) and containing 0.05% Tween 80 and 20 mg/mL kanamycin. The L-lysine monohydrochloride (Sigma) was dissolved in distilled water and used at a concentration of 40 mg/mL. Mycobacterial suspensions for animal inoculation were prepared in PBS from frozen glycerol stocks previously titrated by plating serial dilutions. The strain MTBVACDlys was constructed following the bacterial artificial chromosomerecombineering (BAC-rec) strategy described by Aguilo et al. 27 Briefly, the target gene rv1293/lysA was identified in the E. coli BAC library (kindly donated by Roland Brosch from Institut Pasteur, Paris). In this clone, the target gene lysA was interrupted with a kanamycin resistance cassette by heterologous recombination mediated by PCR product, which was amplified by using specific primers LysA-P1-pKD4-fw and LysA-P2-pKD4-rv (Table S1). The disrupted lysA-Km gene was amplified by PCR using ArgS1-fw and ThrB1-rv primers (Table S1). This PCR product containing the lysA-Km fragment was introduced into the MTBVAC genome by homologous recombination. MTBVACDlys recombinant colonies were selected by plating on 7H10-ADC supplemented with Km and Lysine, and proper recombination was checked by PCR using the pairs of primers Lys-fw/km-OUT1-rv and Lys-rv/km-OUT2-fw (Table S1).
For the construction of MTBVAC.HIVA 2auxo and MTBVAC.B 2auxo strains, MTBVACDlys culture was grown until log phase and conditioned for electroporation by washing with 10% glycerol, according to the method described by Wards and Collins. 62 Electrocompetent MTBVACDlys was transformed with 0.5-1 mg p2auxo.HIVA or p2auxo.B, 26 using a Bio-Rad gene pulser electroporator at 2.5 kV, 25 mF, and 1,000 U, and plated onto Middlebrook agar 7H10-ADC medium without lysine supplementation. Recombinant MTBVA-C.HIVA 2auxo and MTBVAC.B 2auxo colonies containing the corresponding plasmids p2auxo.HIVA and p2auxo.B were selected by PCR using the pair of primers 19kDss-fw/HIVA-rv and 19kDss-fw/ Pglya-rv, respectively (Table S1). The MTBVAC.HIVA 2auxo colonies were assessed for heterologous HIVA protein expression, and clone 2 was selected and preserved using the seed-lot system. A master seed stock and derivative working stock, which we used also as a vaccine stock, were prepared and stored at À80 C with 20% glycerol as a preservative.
In Vitro Stability of the MTBVAC.HIVA 2auxo Strain Six subcultures ($42 bacterial generations) of MTBVAC.HIVA 2auxo (working vaccine stock), harboring the episomal p2auxo.HIVA plasmid DNA that contains the lysine-complementing gene, were grown in 7H9 broth on selective media (no lysine supplementation). Subcultures were performed every 7 days by transferring 100 mL stationary-phase culture to 5 mL fresh medium. PCR analysis of HIVA DNA-coding sequence and GlyA gene (Table S1) was performed to detect plasmid genetic rearrangements.

Mouse Trials
For immunogenicity studies, adult (7-week-old) female BALB/c mice were left either unimmunized or immunized with MTBVAC. HIVA 2auxo or MTBVAC.Ø 2auxo , and they were boosted with MVA.HIVA at doses, routes, and schedules outlined in the Figure 4 legend. On the day of sacrifice, individual spleens were collected, and splenocytes were isolated by homogenizing spleens using a cell strainer (Falcon; Becton Dickinson) and a 5-mL syringe rubber plunger. Following the removal of red blood cells with ACK lysing buffer (Lonza, Barcelona, Spain), the splenocytes were washed and resuspended in complete medium (R10 [RPMI 1640 supplemented with 10% fetal calf serum and penicillin-streptomycin], 20 mmol/L HEPES, and 15 mmol/L 2-mercaptoethanol).
Body mass was monitored over time and recorded to depict any adverse events and body mass loss due to vaccination. To detect vaccine-related adverse events, a 12-week period between MTBVAC.HIVA 2auxo prime and MVA.HIVA boost was established for this trial.
For safety studies, groups of 6 female CB-17/Icr-Prkdc SCID mice (Janvier Biolabs) received a single intraperitoneal inoculation of 10 6 CFU BCG Pasteur, MTBVAC, MTBVACDlys, or MTBVAC.HI-VA 2auxo . Mice were monitored for any sign of disease and body mass measurements were performed weekly. Experimental endpoint was set as a 20% body weight reduction. In the case of the MTBVACDlys group, the endpoint was established at 220 days post-inoculation, upon which surviving animals were humanely euthanized and bacterial load in spleen was quantified. Samples were also obtained from MTBVAC.HIVA 2auxo animals euthanized during the protocol to check for plasmid stability by colony PCR analysis using 19kDssfw/HIVA-rv and HIVA-fw/Pglya-rv primers (Table S1).

Peptides
For assessing the immunogenicity of HIVA in the BALB/c mice, the following peptides were used: H-2D d -restricted epitope P18-I10 (RGPGRAFVTI). The PPD (AJVaccines, Copenhagen, Denmark) was used to assess the immunogenicity induced by MTBVAC.

Ex Vivo IFN-g ELISPOT Assay
The ELISPOT assay was performed using a commercial IFN-g ELISPOT kit (Mabtech, Nacka Strand, Sweden), following the manufacturer's instructions. The ELISPOT plates (MSISP4510, 96-well plates with polyvinylidene difluoride membranes; Millipore, Billerica, MA) were coated with purified anti-mouse IFN-g capture monoclonal antibody diluted in PBS to a final concentration of 5 mg/mL at 4 C overnight. A total of 5 Â 10 5 fresh splenocytes was added to each well and stimulated with 2 mg/mL P18-I10 peptide or 5 mg/mL PPD for 16 h at 37 C. Wells were washed four times with PBS 0.05% Tween 20 and twice with PBS before incubating with 100 mL 5-bromo-4-chloro-3-indoyl-phosphate/nitro blue tetrazolium substrate solution (Sigma). After 5-10 min, the plates were washed with tap water and dried, and the resulting spots were counted using an ELISPOT reader (Autoimmun Diagnostika, Strassberg, Germany).

Intracellular Cytokine Staining
One million splenocytes were added to each well of a 96-well roundbottomed plate (Costar, Corning, NY), pulsed with 2 mg/mL P18-I10 peptide and kept at 37 C and 5% CO 2 for 60 min, followed by the addition of GolgiStop (Becton Dickinson) containing monensin. After 5 h of incubation, the reaction was terminated by transferring the plate to 4 C. The cells were washed with wash buffer (PBS, 2% fetal calf serum, and 0.01% azide) and blocked with anti-CD16/32 (BD Biosciences) at 4 C for 30 min. All subsequent antibody stains were performed using the same conditions. Cells were then washed and stained with anti-CD8-PerCP (BD Biosciences) and anti-CD107a-fluorescein isothiocyanate (FITC), washed again, and permeabilized using the Cytofix/Cytoperm kit (BD Biosciences). Perm-wash buffer (BD Biosciences) was used to wash cells before staining with anti-IFN-g-APC and anti-TNF-a-PE (BD Biosciences). Cells were fixed with CellFIX (Becton Dickinson) and stored at 4 C until analysis. All chromogen-labeled cells were analyzed in a Becton Dickinson FACScalibur, using the CellQuest software (Becton Dickinson) for acquisition and the FlowJo software (Tree Star, Ashland, OR) for analysis.

Statistical Analysis
Immunogenicity data are shown as group means or group medians as well as individual responses. Statistical significance was determined by ANOVA and Bonferroni post-test. In safety experiments H37Rv bacterial burdens are shown as mean ± SEM. Groups were compared using one-way ANOVA and Bonferroni post-test. Survival data were analyzed applying Mantel-Cox test. In all cases, confidence intervals were as follows: *p < 0.05, **p < 0.01, and ***p < 0.01. GraphPad Prism 5.0 software was used for representation and statistical analysis of the data.

CONFLICTS OF INTEREST
C.M. and J.G.-A. are co-inventors in patent applications entitled "TB vaccine," filed by the University of Zaragoza (PCT/ES 2007/070051), and "Compositions for use as a prophylactic agent to those at risk of infection of TB, or as secondary agents for treating infected TB patients" (218382097.6-1112), University of Zaragoza/Biofabri. There are no other conflicts of interest. J.J.-M. and N.S. are co-inventors in a patent application entitled "Mycobacterium comprising expression vector with two auxotrophic selection markers and its use as vaccine" (EP 12382336.1), Laboratorios Esteve, S.A. and Fundación Privada Institut de Recerca de la SIDA. There are no other conflicts of interest. The rest of the authors certify that they do not have a commercial or other association that might pose a conflict of interest in the subject matter or materials discussed in this manuscript.