Research conducted for this manuscript was approved by the Institutional Biosafety Committee at the University of Minnesota, Twin Cities, under the protocol ID 2008-38359H and 2010-38519H. All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of University of Minnesota, Twin Cities, under the protocol ID 2006-38175A and 2005-38161A.

BHK21 Baby hamster kidney cells and Vero African green monkey kidney cells were grown in DMEM media (Fisher Scientific) with 10% fetal bovine serum (FBS) (Sigma) and 50 μg per ml penicillin and streptomycin (Invitrogen-LifeTechnologies). BSRT7-5 cells, obtained from K.K. Conzelmann (Ludwig-Maximilians-Universität, Germany), are BHK-21 cells stably expressing T7 RNA polymerase. BSRT7-5 cells were grown in minimal essential medium (MEM) (Invitrogen-LifeTechnologies) with 10% FBS, 1 μg/ml Geneticin (Invitrogen-LifeTechnologies), and 50 μg/ml penicillin-streptomycin. Recombinant PICV-vectored TB vaccines were amplified in BHK-21 cells. Infectious virus titer was determined by viral plaque assay in Vero cells as described previously (25). Mycobacterium tuberculosis Erdman is a fully virulent isolate and a BSL-3 agent. Mtb was routinely cultured at 37˚C with aeration in Middlebrook 7H9 liquid medium (Difco), which was supplemented with 10% albumin-dextrose-saline (ADS), 0.5% glycerol and 0.1% Tween-80.

Mtb proteins, peptide arrays, and antibodies were obtained from BEI Resources, NIAID, NIH. These include purified Ag85B (NR-14857) and ESAT-6/EsxA (NR-49424) proteins from Mtb strain H37Rv, peptide arrays of Mtb Ag85B (NR-34828) and ESAT-6 (NR-50711), rabbit polyclonal anti-ESAT6 (NR-13803) and anti-Ag85 Complex (NR-13800) antisera. Phycoerythrin (PE)-conjugated EsxH MHC-I tetramer (H-2K(b) IMYNYPAM), PE-conjugated MHC-II control and Ag85B tetramers were obtained from the NIH Tetramer Core Facility at Emory University. Antibodies used for T cell analysis of immunized mice were purchased from commercial sources, including allophycocyanin (APC)-labeled anti-CD3 (17A2) (Biolegend), fluorescein isothiocyanate (FITC)-labeled anti-CD4 (Biolegend), peridinin chlorophyll protein (PerCP)-Cy5.5-labeled anti-CD8 (53-6.7) (eBioscience), brilliant violent (BV) 510-labelled anti-CD44 (Biolegend), APC-Cy7 anti-CD69 (Biolegend), and BV421 anti-CD103 (BD Biosciences).

Recombinant tri-segmented PICV (rP18tri)-based TB vaccines were constructed as described previously (21, 22). Sequences of Mtb antigens Ag85B (NP_216402), EsxH/TB10.4 (NP_214802), ESAT6/EsxA (YP_178023), Rv1733c (NP_216249), and RpfA (NP_215382) were obtained from GenBank. Gene fragments encoding Ag85B single gene and dual antigens (EsxA-EsxH, Ag85B-EsxH, and RpfA-Rv1733c) linked by a P2A linker sequence (GSGATNFSLLKQAGDVEENPGP) (26) were codon optimized for human cell expression and chemically synthesized by Genewiz (Azenta Life Sciences). These gene fragments were cloned into the S1 or S2 vector of the rP18tri reverse genetics system between Kpn I and Xho I sites (21, 22). The resulting plasmids were verified by restriction enzyme digestion and sequencing. To rescue recombinant rP18tri vector-based TB vaccines, BSRT7-5 cells seeded at 1.5x10⁵ cells per ml in a 6-well plate were transfected with three plasmids expressing the L segment, the S1 and S2 segments encoding respective antigens as illustrated in Figure 1A , using Lipofectamine™ 3000 Transfection Reagent (Thermo Fisher Scientific, L3000015). From 2 to 4 days post-transfection, the supernatants were collected for plaque assay on Vero cells. Individual plaques were picked to infect BHK-21 cells on 10-cm plates for 48 h to prepare viral stocks. Viral RNA was extracted from viral stocks using the QIAamp Viral RNA kit (Qiagen, USA), subjected to RT-PCR and sequencing to confirm the identity of viral vectored TB vaccines. TBvac-1 encodes EsxA-EsxH on the S1 segment and Ag85B on the S2 segment, while TBvac-2 encodes Ag85B on the S1 segment and EsxA-EsxH on the S2 segment. TBvac-10 encodes Ag85B-EsxH on the S1 segment and RpfA-Rv1733c on the S2 segment.

Vero cells in six-well plates were washed and infected with 0.5 ml of viruses in 10-fold serial dilutions. After incubation for 1 h at 37°C, the infection medium was removed. Fresh MEM supplemented with 0.5% agar and 10% FBS were added to the cells, which were cultured for 4 days at 37°C. Plaques were stained overnight with neutral red solution diluted (1:50) in 0.5% agar–MEM–10% FBS.

Vero cells grown on coverslips were mock infected or infected with the rP18tri vectors alone or respective TBvac viruses at MOI of 0.1. At 24 hours post infection (hpi), cells were fixed with 4% paraformaldehyde for 15 min at room temperature, and washed three times with phosphate buffer saline (PBS) (Invitrogen-LifeTechnologies). Cells were treated with 0.1% Triton X-100 for 12 min followed by incubation with rabbit polyclonal antibody against Ag85B or ESAT6 (BEI) for 1 h. After washing, cells were incubated with goat anti-rabbit Alexa Fluor-488 (Invitrogen) for 1 hr at room temperature. The cells were washed three times and mounted on a glass slide for examination by confocal microscopy.

For immunogenicity studies, female six- to eight-week-old C57BL/6 mice were obtained from Jackson Laboratories and housed for at least one week for acclimatization. Groups of three to five mice were immunized intranasally (IN) or intramuscularly (IM) depending on the study protocol with 1x10⁵ PFU of TBvac-2, TBvac-10, rP18tri vector, or phosphate buffered saline (PBS) as a mock control. To evaluate immune responses after primary immunization, blood was collected into lithium heparin tubes (Greiner Bio-One) via the facial vein. Peripheral blood mononuclear cells (PBMCs) were isolated by first lysing red blood cells (RBC) using 1X RBC lysis buffer (eBiosciences) and then washing in PBS with 2% fetal bovine serum (FBS) (Sigma Aldrich). Three weeks after primary immunization, mice were boosted IN or IM with the same vaccine candidate. Seven days after the boost, mice were euthanized by CO₂ inhalation and the spleen and lung collected into sterile PBS. To isolate splenocytes, spleens were gently crushed through a 40 µm cell strainer and washed with RPMI-1640 medium with L-glutamine (Cytiva HyClone). Red blood cells were lysed using 1X RBC lysis buffer and splenocytes resuspended in PBS with 2% FBS. To isolate lung lymphocytes, lungs were cut into small pieces and incubated for 1 hour at 37C in PBS+5% FBS with 1.5 mg/ml type I collagenase (Worthington Biochemical) and 2 μg/ml DNase I (New England BioLabs). Then, lungs were gently crushed through a 40 µm cell strainer, washed with RPMI-1640/1% FBS and 10mM HEPES (ThermoFisher). Cells were resuspended in 44% Percoll (GE Healthcare), underlain with 67% Percoll, centrifuged, and the lymphocyte layer collected. Lymphocytes were resuspended in PBS with 2% FBS.

ELISpot was performed using Mouse IFNγ ELISpot Development Module (R&D Systems, SEL485) following the manufacturer’s instruction. Briefly, 96-well PVDF-bottom Immunospot plates were coated with anti-mouse IFNγ capture antibody overnight at 4°C. The next day, plates were washed and blocked in 1% bovine serum albumin (BSA)/5% sucrose/PBS. 1x10⁵ splenocytes were added per well and incubated with medium only or with peptide pools of Ag85B or ESAT6/EsxA (2 µg/ml) (BEI) for 18 hours at 37°C and 5% CO2. After incubation, biotinylated anti-mouse IFNγ detection antibody was added and incubated overnight at 4°C. The following day, streptavidin-alkaline phosphatase and BCIP/NBT were added for color development. Spots were recorded using CTL ImmunoSpot Reader and quantified using immunospot Fluox suite software.

The PE-labeled MHC-I tetramer H-2K(b)/EsxH epitope IMYNYPAM was provided by the NIH Tetramer core facility at Emory University (Atlanta, GA). Isolated PBMCs, splenocytes, and lung lymphocytes were incubated with PE-labeled EsxH/H-2K(b) MHC-I tetramer, fixable viability stain 780 (BD Biosciences), APC-labeled anti-CD3 (Biolegend), PerCP-Cy5.5-labeled anti-CD8 (eBioscience), FITC-labeled anti-CD4 (Biolegend), and BV510-labeled anti-CD44 (Biolegend) for one hour at room temperature. Cells were washed three times with PBS/2% PBS. Sample acquisition was performed on FACSCelesta and data analyzed using FlowJo.

The PE-labeled MHC-II Ag85B tetramer I-A(b)/FQDAYNAAGGHNAVF and control tetramer I-A(b)/PVSKMRMATPLLMQA were provided by the NIH Tetramer core facility at Emory University (Atlanta, GA). Isolated PBMCs, splenocytes, and lung lymphocytes were incubated with fixable viability stain 780 (BD Biosciences), APC-labeled anti-CD3 (Biolegend), PerCP-Cy5.5-labeled anti-CD8 (eBioscience), FITC-labeled anti-CD4 (Biolegend), and BV510-labeled anti-CD44 (Biolegend), together with either control or Ag85B PE-labeled MHC-II tetramer. To detect tissue-resident CD4 T cells in the lung, lung lymphocytes were incubated with the same cocktails, together with APC-Cy7-labeled anti-CD69 (Biolegend). Cells were washed three times with PBS/2% PBS. Sample acquisition was performed on FACSCelesta and data analyzed using FlowJo.

Isolated cells were seeded in leukocyte medium (complete RPMI) at 1x10⁶ cells per well and incubated with medium alone or with purified Ag85B/ESAT-6 proteins (each 10 µg/ml) obtained from BEI for 2 hours at 37°C. Cells were further incubated for six hours at 37°C with anti-CD28 antibody and GolgiStop (BD Biosciences). Cells were collected, washed with PBS/2% FBS, and stained with fixable viability dye 780, PerCP-Cy5.5-labeled anti-CD3, FITC-labeled anti-CD8, APC-Cy7-labeled anti-CD4, and PE-Cy7-labeled anti-CD44 for 45 minutes on ice. Cells were then fixed and permeabilized using the Cytofix/Cytoperm kit (BD Biosciences) and stained with BV421-labeled anti-TNFα, PE-labeled anti-IL-2, and APC-labeled anti-IFNγ for 40 minutes on ice. Sample acquisition was performed on FACSCelesta and data analyzed using FlowJo.

Female six to eight-week-old C56BL/6 mice were immunized IM with 1x10⁵ PFU of rP18tri vector, TBvac-1, or TBvac-2 in a prime-boost strategy with a 21-day interval. At 28 days after immunization, vaccinated mice were challenged with ~400 CFU of virulent Mtb Erdman, using a Glas-Col Inhalation Exposure System, in a biosafety level 3 (BSL3) laboratory, as previously described (27). A group of unvaccinated mice (n=4) were euthanized at 24 hr post-infection to determine the Mtb dose. At 4- and 12-weeks post-infection, groups of mice (n=4-5) were sacrificed; lungs and spleen were collected for bacterial quantification. Tissues were homogenized (BioGen Pro200) in 3 ml of PBS containing 0.05% Tween-80 (PBS-T), serially diluted in PBS-T and plated on Middlebrook 7H10 agar (Difco) supplemented with 0.5% glycerol, 10% oleic acid-albumin-dextrose-catalase (OADC, Difco) and 100 μg/ml cycloheximide. Plates were incubated for 3-4 weeks at 37˚C before counting CFU.

The genome of wild-type PICV consists of two RNA segments, each encoding two genes in opposite orientations (25). The L segment encodes a matrix protein Z and a large RNA polymerase L, while the S segment encodes the envelope glycoprotein GPC and the nucleoprotein NP (28). By splitting the S segment into the S1 and S2 segments, we were able to insert an ORF of up to 2 kilobases in each segment ( Figure 1A ) (21, 22).

We focused on two immunodominant Mtb antigens EsxH/TB10.4 and Ag85B as they have well-established major histocompatibility complex (MHC) class I (MHC-I) and MHC-II tetramer assays to conveniently evaluate the antigen-specific CD8 and CD4 T cells, respectively, in mice (29–33). To evaluate whether multiple antigens expressed from the same vector would interfere with the immune development, we included additional antigens ESAT-6/EsxA, Rv1733c, and RpfA that are associated with different stages of Mtb infection (acute, latency, and resuscitation, respectively) and that have been included in other candidate TB vaccines (12). To maximize the coding capacity of an ORF, we used a P2A linker between two antigens. The P2A linker sequence induces ribosomal skipping during translation, allowing the production of multiple proteins from one ORF (26, 34). We aim to test whether effective immune responses are induced to each antigen linked by the P2A sequence. This may represent a valid strategy to generate rP18tri-based multivalent vaccines, which confer a greater protection against disease than those containing a single antigen (33, 35, 36).

Codon-optimized genes were cloned into either the S1 or S2 vector as illustrated in Figure 1A . Recombinant rP18tri-based TB vaccines, TBvac-1, TBvac-2, and TBvac-10, were generated using reverse genetics as described previously (21, 22, 37). TBvac-1 and TBvac-2 encode the same Mtb antigens, Ag85B and EsxA-EsxH, except that the antigen location is swapped between the S1 and S2 segments ( Figure 1A ). TBvac-10 encodes Ag85B-EsxH within the S1 ORF, and RpfA-Rv1733c within the S2 ORF ( Figure 1A ). In this study, we used TBvac-2 and TBvac-10 to characterize T cell immune responses and evaluated TBvac-1 and TBvac-2 for protective efficacy against virulent Mtb challenge.

To detect antigen expression from viral vectored vaccines, Vero cells were mock infected or infected with rP18tri vector or viral vectored TB vaccines at MOI of 0.1 for 24 h. Ag85B expression was detected in cells infected with vaccines but not with vector or mock infection by both Western blotting ( Figure 1B ) and immunofluorescence assay (IFA) ( Figure 1C ). We also detected EsxA expression in cells infected with TBvac-1 and TBvac-2 but not with mock, vector, or TBvac-10 by IFA ( Figure 1C ). We were unable to assess EsxH, RpfA or Rv1733c expression due to lack of antibody reagents.

We first evaluated vaccine-induced T cell responses by IFNγ-based enzyme-linked immunosorbent spot (ELISpot) assay, which detects production of IFNγ from lymphocytes after antigen stimulation. We have previously shown that the rP18tri vector-based influenza vaccine induces even stronger antibody and T cell responses upon boost than after prime (21), so we evaluated T cell responses only after the boost dose. C57BL/6J (BL6) mice (n=3) were immunized with either the rP18tri vector (V) or TBvac-2 by an IM-IN prime-boost strategy ( Figure 2A ). At 7 days post-boost (dpb), splenocytes were analyzed by mouse IFNγ-based ELISpot assay after stimulation with medium alone or respective peptide pools ( Figure 2B ). After normalization to the medium control, the number of IFNγ-secreting cells after Ag85B or ESAT-6/EsxA peptide stimulation was significantly increased in TBvac-2 immunized compared to vector immunized mice ( Figure 2C ). These findings show that immunization with TBvac-2 elicits IFNγ-producing, antigen-specific T cell responses.

We next evaluated the vaccine-induced antigen-specific CD8 T cell responses after either systemic (IM) or mucosal (IN) inoculation using an established MHC-I EsxH tetramer assay. BL6 mice were immunized with PBS/mock, rP18tri vector alone, or TBvac-2 through either IM or IN route ( Figure 3A ). At 7 d post-prime (dpp), lymphocytes from peripheral blood (PBMC), spleen, and lung were analyzed by EsxH-specific MHC-I tetramer analysis ( Figure 3B ). A single dose immunization of TBvac-2 through either IM or IN route generated a significantly higher percentage of MHC-I EsxH tetramer-positive CD8 T cells in the peripheral blood ( Figure 3C ) and spleen ( Figure 3D ) than mock or vector immunization. No significant difference in the PBMC CD8 T cell response was detected between the IM and IN routes ( Figure 3C ), while IN appears to elicit higher CD8 T cell responses than IM in the spleen ( Figure 3D ). We also quantified the tetramer-positive CD8 T cells in the lung at 7 d post-boost after prime-boost immunization by the IM-IM or IN-IN routes ( Figure 3A ). Consistent with the results observed in PBMCs and spleen, TBvac-2 immunization via both routes generated a significantly higher percentage of tetramer-positive CD8 T cells in the lung than mock or vector alone, with a trend toward higher responses in mice immunized by the IN route compared to IM, though this did not achieve statistical significance ( Figure 3E ).

As both routes of inoculation effectively stimulated immunity, we elected to move forward with a heterologous prime-boost strategy (IM-IN), which has been shown to result in greater protection against Mtb in guinea pigs (38). We evaluated CD8 T cell responses induced in mice immunized IM-IN with TBvac-10 by the MHC-I EsxH tetramer assay ( Figure 3F ). Mice immunized with TBvac-10 had a significantly higher percentage of tetramer-positive CD8 T cells compared to vector immunized animals in the peripheral blood, spleen, and lung ( Figure 3G ). The percentage of antigen-specific CD8 T cells was significantly higher in the lung than in the spleen.

Taken together, these results show that two different rP18tri-based TB vaccines (TBvac-2 and TBvac-10) administered systemically, mucosally, or a combination of the two, elicit strong antigen-specific CD8 T cells in the spleen, peripheral blood, and lungs. As EsxH is expressed after the P2A linker in both TBvac-2 and TBvac-10, these results also demonstrate that P2A-linked antigens in the rP18tri viral vector can stimulate robust immune responses.

We next evaluated the vaccine-induced CD4 T cell responses by MHC-II tetramer analysis. BL6 mice were prime-boosted with vector (V), TBvac-2, or TBvac-10 by either the IM-IN or IN-IN routes ( Figure 4A ). Spleen and lung lymphocytes were analyzed for Ag85B MHC-II tetramer-positive CD4 T cells ( Figure 4B ). A gating strategy for analyzing the flow cytometry data is shown in Figure 4B . Both spleen and lung lymphocytes were gated for CD3⁺CD4⁺CD44^h population and lung lymphocytes were further gated by CD69⁺, which is a tissue-resident cell surface marker (39). Tetramer-positive cells in the target cell populations were specifically detected in TBvac-2 and TBvac-10-immunized mice after incubation with Ag85B MHC-II tetramer ( Figure 4C ). After normalization with the control MHC-II tetramer, the percentage of Ag85B MHC-II tetramer-positive cells among CD4 T cells in the spleen ( Figure 4D ) and lung ( Figure 4E ) were determined for each mouse. In both tissues, significantly higher percentages of tetramer-positive cells were detected in mice immunized with TBvac-2 and TBvac-10 than with the vector, demonstrating the induction of antigen-specific CD4 T cells by both vaccines. TBvac-10 induced a higher percentage of tetramer-positive CD4 T cells in both the spleen ( Figure 4D ) and lung ( Figure 4E ) compared to TBvac-2, suggesting a stronger Ag85B-specific CD4 T cell immunity. No significant differences in lung tetramer-positive CD4 T cells were detected between IN-IN and IM-IN routes of inoculation with TBvac-2 ( Figure 4E ).

Recent studies suggest that tissue-resident memory T cells (Trm) play an important role in the primary defense against infection (40) and that Mtb-specific CD4 and CD8 T cells with tissue-resident memory phenotypes are associated with protection (12). To determine whether the vaccine-induced lung CD4 T cells exhibit tissue-resident phenotypes, we gated the CD69⁺ (a tissue-resident cell marker) population of the CD4 T cells (CD3⁺CD4⁺CD44^hCD69⁺) for quantification of tetramer-positive cells. Both TBvac-2 and TBvac-10 stimulated high levels of tetramer-positive CD69⁺ CD4 T cells in the lung ( Figure 4F ). Consistent with the results for CD4 T cells ( Figures 4D, E ), TBvac-10 immunization induced a higher level of CD69⁺CD44^hCD4 T cells than TBvac-2 ( Figure 4F ).

As multifunctional T cells are associated with control of multiple intracellular pathogens (41, 42) including Mtb (43), we also evaluated whether the viral vaccines can induce antigen-specific functional CD4 T cells by intracellular cytokine staining after ex vivo stimulation with antigen peptide pools ( Figure 5A ). CD4 T cells that were positive for dual cytokine expression (IFNγ/IL-2, IFNγ/TNFα, and IL-2/TNFα) were specifically detected after peptide stimulation in TBvac-2 and TBvac-10 immunized mice but not in vector-immunized mice ( Figure 5B ). The average percentages of dual- and triple-positive (IFNγ/IL-2/TNFα) CD4 T cells were significantly higher in mice receiving either vaccine than in the vector group ( Figure 5C ). There was a trend toward greater percentage of multifunctional cells for TBvac-10 than TBvac-2, but this was not statistically significant.

Taken together, these results suggest that both TBvac-2 and TBvac-10 can elicit functional antigen-specific CD4 T cells systemically and locally (lung) and that TBvac-10 may induce a stronger CD4 T cell immunity than TBvac-2.

We conducted an initial study to evaluate the protective efficacy of TBvac-1 and TBvac-2 in a standard Mtb low-dose aerosol challenge mouse model (44). BL6 mice prime-boost immunized through the IM route with 1 x10⁵ PFU of vector alone, TBvac-1, or TBvac-2 were infected by aerosol with virulent Mtb Erdman and euthanized at 4 or 12 weeks post-infection for bacterial load quantification ( Figure 6A ). At the 4-week timepoint, TBvac-1 and TBvac-2-immunized mice showed a similarly decreased bacterial load in both lungs and spleens ( Figure 6B ). The average Mtb burden in the lungs of TBvac-1 and TBvac-2-immunized mice decreased by 0.8 to 1 log compared to vector-immunized mice ( Figure 6B ). Although not statistically significant, this decrease is similar to what is seen in BL6 mice immunized with BCG alone (45, 46). TBvac-1 and TBvac-2 immunization also significantly lowered the bacterial load in the spleens at the 4-week time point compared to vector alone ( Figure 6B ). At the 12-week time point, we included the vector and TBvac-1 groups, but not TBvac-2 group, due to insufficient number of animals. TBvac-1 immunized animals had significantly decreased Mtb loads in the lung but not in the spleen ( Figure 6B ). These results establish that rP18tri-based vaccines expressing the Mtb antigens Ag85B, EsxA, and EsxH reduce the bacterial burden in mice challenged with virulent Mtb, demonstrating their potential for further development.