We used BCG Tokyo type I, an attenuated strain of M. bovis, in this study. The frozen BCG stock was thawed and cultured in Middlebrook 7H9 broth (BD, Franklin Lakes, NJ, USA) supplemented with 0.2% (v/v) glycerol, 0.5% bovine serum albumin, 0.081% NaCl, 0.2% d-glucose, and 0.05% (v/v) Tween 80 (7H9ADN) (21). The morphology of BCG was assessed by scanning electron microscopy (SEM) (22).

To isolate static-cultured MVs, when the turbidity of the bacterial suspension (OD600) reached approximately 1.0, 10 mL of the BCG culture was added to 200 mL of Sauton’s medium (4.0 g/L L-asparagine monohydrate, 2.8 g/L Trisodium citrate dihydrate, 0.5 g/L K₂HPO₄, 0.5 g/L MgSO₄·7H₂O, 0.05 g/L ammonium iron citrate, and 6% glycerol), which is commonly used to produce the BCG vaccine. BCG was then statically cultured for 21 days. To isolate the planktonic-cultured MVs, 10 mL of the BCG culture was added to Mueller−Hinton II (MH-II) broth (BD) supplemented with 0.05% (v/v) Tween 80 or 7H9ADC broth, which is commonly used for mycobacterial experiments. Then, BCG was planktonically cultured for 10–14 days with gentle stirring by the magnetic string bar at 60 rpm.

For the mycobacterial challenge experiment, we used the pMV361-Km-containing BCG mutant strain (23, 24), which was kindly provided by Dr. Yoshitaka Tateishi, Niigata University. Kanamycin (Km) was purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan).

BCG-MVs were isolated via ultracentrifugation according to previous methods (11, 13, 25). Briefly, the culture supernatant was harvested via centrifugation (10,000×g for 30 min at 4°C) and passed through 0.45 μm and 0.22 μm filters. The supernatant was subsequently concentrated via ultrafiltration using an Amicon stirred cell and a 100 kDa Ultracel membrane (Millipore, Burlington, MA, USA). The MVs were sedimented via ultracentrifugation (100,000×g for 2 h at 4°C), washed once with phosphate-buffered saline (PBS) and resuspended in PBS. The BCG-MV protein concentration was determined via a BCA assay (FUJIFILM Wako) according to the manufacturer’s instructions. MVs from Escherichia coli Nissle 1917 ΔflhD (EcN-MVs), a probiotic strain lacking flagella, were purified from the supernatant of a 16-h bacterial culture of EcN after glycine induction performed as described previously (26). The EcN-MV protein concentration was determined via a Bradford assay (Bio-Rad Laboratories, Hercules, CA, USA). The morphology of the MVs was assessed by transmission electron microscopy (TEM) and field emission SEM (FE-SEM) (7). The size distributions of the MVs were assessed using a nanoparticle size analyzer NanoFCM (NanoFCM Inc., Xiamen, China) (7).

Human monocytic leukemia (THP-1) cells, which were kindly provided by Dr. Mayuko Osada-Oka, Kyoto Prefectural University, were cultured in RPMI-1640 medium (FUJIFILM Wako, Osaka, Japan) supplemented with 10% fetal bovine serum (FBS), and 100 units/mL penicillin plus 100 µg/mL streptomycin (P/S) in a humidified atmosphere of 5% CO₂ at 37°C. THP-1 cells were seeded onto 12- or 24-well plates at densities of 5×10⁵ or 2.5×10⁵ cells per well, respectively. The cells were differentiated into Mφs with 100 nM phorbol 12-myristate-13-acetate (Cayman Chemical, Ann Arbor, MI, USA) in Dulbecco’s modified Eagle’s medium (DMEM) (FUJIFILM Wako) supplemented with 10% FBS and P/S for 24 h (27, 28). Then, the cells were washed twice with PBS, and the culture medium was replaced with fresh DMEM containing 10% FBS and P/S for 24 h. The cells were subsequently stimulated with BCG-MVs for 3 h (for the RNA experiment) or 24 h (for the multiplex assay).

Total RNA was extracted with Isogen (Nippon Gene, Toyama, Japan) and a Direct-zol™ RNA MicroPrep kit (Zymo Research, Irvine, CA) as previously reported (29). After the quality of the RNA was checked with a NanoDrop 2000 (Thermo Scientific, Waltham, MA, USA), cDNA was synthesized using ReverTra Ace qPCR RT Master Mix with gDNA Remover (Toyobo, Osaka, Japan). Gene expression levels were quantified using Luna Universal qPCR Master Mix (New England BioLabs Inc., Ipswich, MA, USA). The sequences of the primer sets used in this study are listed in Table 1 . PCR was performed on a 7500 Fast instrument (Applied Biosystems, Carlsbad, CA, USA). To normalize the data, the relative expression levels of the target genes were calculated via the comparative CT (ΔΔCT) method and compared with those of 18S rRNA as an internal standard. The experiments were performed independently three times.

The culture supernatant was collected from THP-1 Mφs stimulated or not by BCG-MVs. The concentrations of cytokines and chemokines in the supernatants were assessed via the Bio-Plex Pro™ Human Cytokine 17-Plex assay (Bio-Rad Laboratories) according to the manufacturer’s instructions. The fluorescence intensity was measured and the concentrations of the target factors were calculated with a Bio-Plex 200 system on the basis of Luminex multiple analyte profiling technology.

All animal procedures were approved by the animal experiment committee of Osaka City University (approval no. 18077) and were performed in accordance with our institutional animal care guidelines. Female BALB/c mice aged 6–8 weeks at the time of immunization were purchased from Japan SLC (Hamamatsu, Japan). The mice were randomly divided into cages according to their group, provided ad libitum access to food and water, and maintained on a 12-h light/dark cycle at 22 ± 1°C. The mice were intranasally or subcutaneously immunized with 1 μg of MVs per mouse with or without adjuvants at a 3-week interval (30). The mice in the BCG immunization group were intradermally vaccinated with 5×10⁵ CFU of wild-type BCG suspended in 100 μL of sterile PBS. The adjuvants used in this study were 10 µg of poly(I:C) (Sigma−Aldrich, St. Louis, MO, USA), 100 µg of Imject Alum Adjuvant (Thermo Fisher Scientific) and 1 µg of EcN-MVs per mouse. Five weeks after the first immunization, sample collection was performed as previously reported (7, 30, 31). Briefly, saliva was collected after intraperitoneal administration of 0.08 mg/mouse pilocarpine and 0.04 mg/mouse isoproterenol. After saliva collection, the mice were anesthetized via intraperitoneal administration of medetomidine (0.3 mg/kg), midazolam (4 mg/kg) and butorphanol (5 mg/kg) (MMB) and sacrificed via exsanguination. To obtain serum, whole blood was centrifuged for 10 min at 3000×g and 4°C. Additionally, the lungs and nasal cavities were washed with PBS containing 1% bovine serum albumin.

The development of humoral immunity was confirmed by ELISAs as described previously, with some modifications (7, 31). Briefly, whole-cell lysates were prepared from BCG cells via bead beating 5 times (5000 rpm for 30 s each time) in carbonate buffer (pH 9.6), followed by centrifugation at 12,000 × g for 20 min at 4°C and passage through 0.45- and 0.22-µm filters. Then, the lysates were diluted to a concentration of 20 μg/mL and pipetted into a 96-well ELISA plate (SUMITOMO BAKELITE, Tokyo, Japan). The BCG lysates were immobilized overnight at 4°C. After the coated antigen was blocked by the addition of 5% skim milk in PBS with 0.05% Tween 20 (PBS-T), the samples collected from the immunized mice (serum, saliva, nasal wash and lung wash) were diluted in 5% skim milk in PBS-T and added to the wells. After overnight incubation at 4°C, diluted horseradish peroxidase-linked secondary antibodies (anti-mouse IgG and anti-mouse IgA antibodies (Cytiva, Wilmington, DE, USA)) were added to the wells at a dilution of 1:5000. The peroxidase substrate SureBlue (SeraCare Life Sciences, Milford, MA, USA) was subsequently added, and the reaction was stopped with 1 M HCl. BCG-specific antibodies were detected via colorimetry, and the absorbance was measured at 450 nm.

Five weeks after the first immunization, the mice were challenged as previously reported (24). In brief, 5×10⁵ CFU of BCG containing pMV361-Km was suspended in 40 μL of sterile PBS and intranasally administered to the mice under anesthesia with MMB. Two weeks after the challenge, the mice were sacrificed under anesthesia with MMB. The lungs were subsequently homogenized in 1 mL of sterile ultrapure water via bead beating with 28 mm diameter zirconia beads at 5000 rpm for 30 s, and lung cells were lyzed by 0.1% Triton X-100. The lysate was diluted with sterile ultrapure water and plated on Middlebrook 7H10 agar supplemented with oleic acid, bovine serum albumin, dextrose and 10 μg/mL kanamycin. The plates were incubated at 37°C for 20 days, after which the CFUs were counted.

Proteins from two biological replicates of both the s-MVs and p-MVs were identified via MS. Briefly, MV-associated proteins were concentrated by trichloroacetic acid precipitation and separated via sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS−PAGE) on a 12% polyacrylamide gel with 10 µg of protein in each lane. The resulting gel was stained with 10% Coomassie Brilliant Blue (G-250) solution, and 4 pieces were excised from each lane. The samples were reduced, alkylated and trypsinized according to an in-gel digestion method (32). The peptides extracted from the gel were subjected to nanoscale liquid chromatography (nanoLC)–MS/MS analysis with a system consisting of an Orbitrap Velos Pro (Thermo Fisher Scientific) coupled with an Advance UHPLC (Brucker, Billerica, MA) and an HTC-PAL autosampler (CTC Analytics, Zwingen, Switzerland). Each peptide sample was separated with both a trap column (5-µm C18 L-column) and an analytical column (3-µm C18 L-column) (Chemicals Evaluation and Research Institute, Tokyo, Japan) (32). The mobile phases were 0.1% formic acid in water (A) and 100% acetonitrile (B). Gradient elution of the peptides was performed from 5–40% B over 40 min and 40–95% B over 1 min, after which a concentration of 95% B was maintained for 9 min; the flow rate was 200 nL/min. Full MS spectra were obtained with an Orbitrap mass spectrometer in the mass/charge (m/z) range of 300–1800 with a resolution of 60,000 at 400 m/z. For automated gain control (AGC), the 12 most intense precursor ions were selected for MS/MS analysis. The raw data were processed using ProteoWizard msConvert (https://proteowizard.sourceforge.io/) and searched against the M. tuberculosis H37Rv protein sequence database (the SwissProt_2019_11 database) using the MASCOT algorithm (ver. 2.7.0, Matrix Science Inc., Boston, MA, USA).

Scaffold (version Scaffold_4.11.0, Proteome Software Inc., Portland, OR, USA) was used to validate peptides and proteins identified via MS/MS. We compared the proteins identified from the s-MVs and p-MVs and found those that were uniquely expressed in both types of MV. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID; https://david.ncifcrf.gov/home.jsp).

Proteins in the BCG-derived MVs were extracted in an equal volume of loading buffer containing SDS. Proteins separated by SDS−PAGE were transferred to Immobilon-P PVDF membranes (Millipore), and the membranes were probed with antibodies against lipoarabinomannan (LAM) (clone TB, kindly provided by Otsuka Pharmaceutical Corporation, Tokushima, Japan) and the 19-kDa lipoprotein LpqH (IT-12, BEI Resources).

THP-1 Mφs were trained as described previously, with some modifications (33, 34). Briefly, differentiated THP-1 Mφs in 24-well plates were incubated with PBS (negative control) or 1 µg/mL s-MVs for 24 h. The cells were washed twice with prewarmed PBS and then incubated for 2 days in culture medium; the medium was changed 6 h before restimulation. The cells were restimulated with PBS, 100 ng/mL lipopolysaccharide (LPS) (35–37), 100 ng/mL Pam3CSK4 (35, 38) or 1 µg/mL s-MVs for 3 h and then washed twice with ice-cold PBS. Finally, total RNA was purified using the Isogen and Direct-zol™ RNA MicroPrep kit.

The data obtained from in vivo or in vitro experiments are presented as the means and standard errors of the means or the means and standard deviations, respectively. Comparisons among more than three groups were performed via one-way analysis of variance followed by Tukey’s HSD test using GraphPad Prism (GraphPad Software, Boston, MA). Comparisons between two groups were performed via unpaired t -tests using GraphPad Prism. Differences were considered statistically significant at a value of P<0.05.

To investigate whether BCG-MVs could be used as novel TB vaccines, we first aimed to determine the optimal culture conditions for the MV donor bacilli by comparing MVs obtained from statically cultured BCG in nutrient-restricted Sauton’s medium (s-MVs) and those obtained from planktonically cultured BCG in nutrient-rich medium via ultracentrifugation (p-MVs). We first examined whether the morphological and biochemical properties of MVs were affected by the culture conditions of the donor bacilli. SEM analysis revealed that BCG produced MVs by blebbing ( Figure 1A ). We then confirmed the vesicular structure of the purified BCG-MVs by FE-SEM and TEM ( Figure 1B ). EM analysis revealed that smaller MVs (φ 20–40 nm) were enriched in p-MVs than in s-MVs. On the other hand, according to the result of the nanoparticle size analyzer, the mean diameter of the s-MVs (65.25 ± 7.33 nm) was smaller than that of p-MVs (72.43 ± 13.45 nm) ( Figure 1C ). These results showed that differences in the donor cell culture conditions may affect MV morphology.

Because the difference in size distribution between the s-MVs and p-MVs may be caused by the presence of different proteins and lipids in the MVs, we compared the proteins expressed in both types of BCG-MVs ( Figure 2A ). Global LC−MS/MS analysis revealed that 58 proteins were unique to p-MVs and 110 proteins were unique to s-MVs, while 235 proteins were expressed in both types of MVs ( Figure 2B ). Overall, the number of identified proteins was greater in s-MVs than in p-MVs. Moreover, functional annotation clustering via GO and KEGG enrichment analyses revealed that various metabolic pathways (such as valine synthesis, gluconeogenesis, TCA cycle, pyrimidine metabolism and biosynthesis of amino acids) and proteasomal catabolic pathways, which consist of the prokaryotic ubiquitin-like protein (pup)-dependent protein degradation system (39), were enriched in the proteins expressed in the s-MVs ( Supplementary Figures S1A, B ). Next, we compared the amounts of LAM and LpqH, which are components of the mycobacterial cell wall and cell membrane, in the MVs. These molecules are well-known virulence factors in Mtb infection and have immunostimulatory properties that are mediated by Toll-like receptor (TLR) 2 activation (40, 41). The amounts of LAM and LpqH were significantly greater in s-MVs than in p-MVs ( Figure 2C ). These results showed that differences in the donor cell culture conditions may affect the protein and lipid composition in the isolated MVs.

Next, we examined whether alterations in the molecular composition of MVs caused by the different donor cell culture conditions may affect the immunostimulatory properties of the MVs. Differentiated THP-1 Mφs were stimulated with s-MVs or p-MVs for 3 h ( Figure 3A ), after which the transcription levels of inflammation-related genes were analyzed. Compared with PBS treatment, s-MV treatment significantly increased the expression levels of interleukin (IL)-6. With the exception of monocyte chemotactic protein 1 (MCP1), the levels of IL-1β, IL-10 and tumor necrosis factor alpha (TNFα) tended to increase by s-MV treatment ( Figure 3B ). However, treatment with p-MVs did not affect the expression levels of these genes.

Next, the levels of secreted cytokines were analyzed via multiplex analysis. Treatment with s-MVs increased the production of various cytokines, such as IL-1β, IL-2, IL-4, IL-5, IL-6, IL-10, IL-17 and TNFα, but not interferon gamma or MCP1 ( Figure 3C ). However, p-MVs did not affect the production of these cytokines. The immunostimulatory property of s-MVs depends on the TLR2-mediated signaling pathway because the increased production of cytokines, except MCP1, induced by treatment with s-MVs was abolished in TLR2-knockout THP-1 cells ( Supplementary Figure S2 ). These results showed that the culture conditions of the MV donor bacteria affect not only the components of MVs but also their immunostimulatory activities. Moreover, s-MVs strongly stimulate the host innate immune response via TLR2.

Next, we examined whether the culture conditions of the MV donor bacteria affect MV immunogenicity in vivo. We intranasally immunized mice twice with s-MVs or p-MVs at a 3-week interval and assessed BCG-specific antibody production ( Figure 4A ). These findings showed that the MVs could not induce the production of BCG-specific antibodies in the absence of an adjuvant. In contrast, significant induction of BCG-specific IgA was observed in the lung wash, nasal wash and saliva samples, but BCG-specific IgA in serum tended to be induced only when the mice were immunized with s-MVs and poly(I:C) but not alum ( Figure 4B ). Moreover, significant induction of BCG-specific IgG was detected in nasal washes and saliva upon coadministration of s-MVs and poly(I:C) ( Figure 4C ). However, p-MVs could not induce antibody production even when adjuvants were coadministered. These results showed that s-MVs stimulated adaptive immune responses more strongly than p-MVs did. Furthermore, s-MVs induced antimycobacterial immunity only when combined with poly(I:C). Poly(I:C), a double-stranded RNA, is an agonist of TLR3 and stimulates the T-helper 1 (Th1)-polarizing immune response by activating NF-κB-regulated gene expression (42). On the other hand, alum usually stimulates the Th2-polarizing immune response (43). Thus, Th1-type adjuvants may be more favorable than Th2-type adjuvants when combined with s-MVs.

A previous study revealed that EcN-MVs, another Th1-type adjuvant (44), have greater adjuvanticity than poly(I:C) (26). We next investigated whether poly(I:C) or EcN-MVs are more suitable adjuvants for s-MVs ( Figure 5A ). First, BCG-specific antibody production in intranasally immunized mice was assessed. Compared with the control, both the poly(I:C) and EcN-MV adjuvants significantly induced the production of BCG-specific IgA, as detected in the lung washes, nasal washes, saliva and serum ( Figure 5B ). However, with respect to adjuvanticity for IgG production, poly(I:C) tended to be superior to EcN-MVs. TB-specific secretory IgA production is important for TB prevention, whereas TB-specific IgG enhances TB infection (45). However, IgG is not necessarily a harmful factor for TB (46, 47). We also compared the adjuvanticities of poly(I:C) and EcN-MVs in subcutaneously immunized mice ( Figure 5A ). Unlike intranasal immunization, the adjuvanticity of EcN-MVs tended to be superior to that of poly(I:C) upon subcutaneous immunization ( Figure 5C ). MVs have both high adjuvanticity and high immunogenicity; however, s-MVs alone did not induce significant production of mycobacterial antibodies upon either intranasal or subcutaneous immunization ( Figures 4B, C ). These data suggest that EcN-MVs are needed to enhance the mycobacterial humoral immune response induced by subcutaneous immunization with s-MVs.

Next, we investigated whether immunization with s-MVs confers efficient protection against mycobacterial infection in a BCG challenge study. Mice that were intranasally or subcutaneously immunized with s-MVs with or without poly(I:C) or EcN-MV as an adjuvant twice at a 3-week interval were intranasally challenged with live Km-resistant BCG (5 × 10⁵ CFU/mouse) at 5 weeks after the first immunization. Moreover, mice intradermally immunized with live BCG were also challenged with live Km-resistant BCG at 5 weeks after immunization. At 2 weeks after challenge, intranasal immunization with s-MVs did not significantly reduce the bacterial burden in the mouse lungs ( Figure 6 ) despite inducing BCG-specific antibody production ( Figure 5B ). Among the groups receiving subcutaneous immunization, the combination of s-MVs and EcN-MVs significantly reduced the bacterial burden in the lungs to a level comparable to that in the mice intradermally immunized with BCG ( Figure 6 ).

The BCG vaccine confers not only TB immunity but also trained immunity (17, 18, 33). A recent study revealed that MVs from E. coli induce trained immunity in mice (16). We therefore investigated whether s-MVs could induce trained immunity in vitro. We trained THP-1 Mφs with BCG-MVs, and the trained cells were restimulated with the TLR4 agonist LPS, the TLR2 agonist Pam3CSK4, or s-MVs followed by 48 h of rest ( Figure 7A ). In the absence of restimulation (PBS treatment), the basal expression levels of IL-6, IL-1β and TNFα in THP-1 Mφs tended to increase with training ( Figure 7B ). The expression level of IL-6 in trained cells also tended to increase when the cells were restimulated with Pam3CSK4 and s-MVs ( Figures 7C, D , respectively). Moreover, the expression level of IL-6 in trained cells significantly increased when the cells were restimulated with LPS ( Figure 7E ). These data may indicate that vaccination with s-MVs confers not only specific protection against TB infection but also broad protection against gram-negative pathogen infection in a TLR4-dependent manner.