Induction of protection in mice against a respiratory challenge by a vaccine formulated with exosomes isolated from Chlamydia muridarum infected cells

The goal of this study was to determine if exosomes, isolated from Chlamydia muridarum infected HeLa cells (C. muridarum-exosomes), induce protective immune responses in mice following vaccination using CpG plus Montanide as adjuvants. Exosomes, collected from uninfected HeLa cells and PBS, formulated with the same adjuvants, were used as negative controls. Mass spectrometry analyses identified 113 C. muridarum proteins in the C. muridarum-exosome preparation including the major outer membrane protein and the polymorphic membrane proteins. Vaccination with C. muridarum-exosomes elicited robust humoral and cell-mediated immune responses to C. muridarum elementary bodies. Following vaccination, mice were challenged intranasally with C. muridarum. Compared to the negative controls, mice immunized with C. muridarum-exosomes were significantly protected as measured by changes in body weight, lungs’ weight, and number of inclusion forming units recovered from lungs. This is the first report, of a vaccine formulated with Chlamydia exosomes, shown to elicit protection against a challenge.


INTRODUCTION
Chlamydia trachomatis is the most common sexually transmitted bacterial pathogen in humans. According to the World Health Organization,~130 million people worldwide are infected each year with Chlamydia 1 . In the USA, depending on the population studied, 5-20% of men and women become C. trachomatis positive during their reproductive years 2 . In countries with poor hygiene, Chlamydia can infect the eye causing trachoma 3 . About 75% of the genital infections in females and 50% in males are asymptomatic [4][5][6] . Genital infections in females can produce cervicitis, acute urethral syndrome, ectopic pregnancy, pelvic inflammatory disease (PID) and infertility while in males it may result in urethritis, epididymitis, prostatitis and testicular infarctions 2,6,7 . If a Chlamydia infection is not treated during pregnancy, 70% of newborns are infected 8 . Moreover, genital C. trachomatis infections are associated with other serious diseases such as cervical hypertrophy, induction of squamous metaplasia, and HIV and HPV infections 9,10 . Antibiotic therapy is available, but due to the high proportion of asymptomatic patients antibiotic treatment have failed to eradicate these infections 11,12 . Countries that established screening programs have observed an increase in the prevalence of genital infections likely due to the failure to develop natural immunity as a result of antibiotic therapy 13 . Therefore, Chlamydia infections remain a key public health priority and a vaccine is required to control them [14][15][16][17][18][19][20] .
A number of chlamydial antigens have been tested for their ability to confer protection against genital, respiratory or ocular challenges [14][15][16][17][18][19][20] . Of these, the Chlamydia major outer membrane protein (MOMP) is the most promising vaccine candidate 16,[21][22][23][24][25] . A shortcoming of MOMP is that it induces serovar/serogroupspecific protection 23,26 . Thus, additional antigens may be required to induce broader protection 17,27 . Candidates, including a family of well-conserved polymorphic membrane proteins (Pmps) have been tested as vaccine antigens. Pmps have different relative expression levels in elementary bodies (EB) versus reticulate bodies (RB) 28 . Thus, it may be difficult to produce a vaccine that works against both forms. Furthermore; lower levels of protection were observed in Pmp-vaccinated versus MOMP-vaccinated mice 29 . Thus, there is a need to identify additional protective antigens.
To date, very little information is available on the role(s) of exosomes secreted from Chlamydia-infected epithelial cells. Exosomes from other organisms have been shown to either protect against infection or lead to pathogenesis 33,34 . Here, we studied in mice the protective ability of C. muridarum exosomes against a respiratory challenge. Exosomes from C. muridaruminfected HeLa cells elicited robust humoral and cell-mediated immune responses and conferred significant protection against a respiratory challenge. Thus, antigens found in these exosomes can now be explored as potential vaccine candidates.
Label-free peptide quantitation (data not shown) allowed us to approximate the absolute abundance of each protein based on the three most abundant tryptic peptide species from each. On this basis, we could estimate total protein in the chlamydia exosomes preparation attributable to the chlamydia proteome to be just 6.27 mole%, with the remainder (93.73 mole%) attributable the human host (HeLa 229 cells). Among Chlamydia proteins only, 10.04 mole% corresponded to MOMP, while 81.72 mole% was from Pmps.
Validation of chlamydial antigens in the exosomes To validate C. muridarum antigens present in the C. muridarumexosomes, these were probed by an indirect ELISA with mAbs and mouse polyclonal sera raised against Chlamydia antigens. As shown in Fig. 1, exosomes contained LPS, MOMP, Pmps, putative outer membrane protein (OMP 85) and the 60 kDa cysteine-rich protein. No antibody bound with the control-exosomes. C. muridarum-exosomes not only bound to mAbs that recognize linear epitopes in VD1 and VD4 of C. muridarum MOMP but also with a mAb that binds to a conformational epitope present in the native MOMP trimer.
Humoral immune responses following vaccination To determine humoral immune responses, serum samples were collected from exosomes vaccinated mice the day before the i.n. challenge. Pre-immunization sera were used as controls. Antibody titers were determined using EB and MOMP (Table 1). Animals vaccinated with C. muridarum-exosomes had an IgG geometric mean titer (GMT) to EB of 25,600 (range: 25,600-25,600) while negative controls immunized with control-exosomes or PBS had titers below the detection level (<100). Animals vaccinated with C. muridarum-exosomes had an anti-MOMP IgG GMT of 9,051 (range: 6,400-12,800). No anti-MOMP or EB antibody titers were detected in sera of mice immunized with control-exosomes or PBS.
In vitro neutralizing antibody levels were determined in sera the day before the challenge (Table 1). Animals immunized with C. muridarum-exosomes had a neutralizing GMT of 8611 while mice vaccinated with control-exosomes or PBS had titers below the detection level (<50).
The western blot using sera collected the day before the i.n. challenge is shown in Fig. 2. Mice vaccinated with C. muridarumexosomes had antibodies predominantly to bands of a high molecular weight protein (OMP 85), the 60 kDa cysteine rich protein, the 42 kDa MOMP, and LPS. Interestingly, no visible band corresponding to Pmps (≥100 kDa) was seen in the sera collected from C. muridarum-exosomes immunized mice. Control mice, immunized with control-exosomes or PBS, had no antibodies reactive with chlamydial components.
To determine MOMP epitopes-specific antibodies, serum samples were probed with 25 aa overlapping C. muridarum MOMP peptides (Fig. 3). Antibodies from animals immunized with C. muridarum-exosomes recognized peptides located almost exclusively in the four variable domains (VD) and a constant domain (CD) 5 while no reactivity was obtained with sera from mice inoculated with control-exosomes or PBS.

C. muridarum-EB C. muridarum-exosomes
Control-exosomes Fig. 1 Reactivity of mAbs and polyclonal sera to C. muridarum antigens against C. muridarum exosomes. mAbs and polyclonal sera to C. muridarum antigens were used in an ELISA to identify proteins present in C. muridarum exosomes. C. muridarum EB were used as a positive control antigen.

Cellular immune responses following vaccination
As a parameter of the C. muridarum-specific cellular immune responses, proliferation to Chlamydia antigens was determined using nylon wool purified spleen T-cells ( Table 2). The delta counts per minute (Δ cpm: cpm from antigen-stimulated wells minus cpm from no antigen-stimulated wells) of T-cells from mice vaccinated with C. muridarum-exosomes EB was 14,462 ± 1899, while from animals inoculated with control-exosomes or PBS were 4730 ± 1392, and 779 ± 125, respectively (P < 0.05). The Δ cpm of all groups of mice stimulated with a positive stimulant, Con A, ranged from 65,987 to 78,567 indicating viability of the T cells. All negative control wells, that received only medium, had cpm that ranged from 637 to 2754. Mean IFN-γ levels (pg/ml), as a measure of a Th1 response, were determined in supernatants from EB-stimulated T-cells (Table 2). Animals immunized with C. muridarum-exosomes secreted the highest levels of IFN-γ (4789 ± 1379 pg/ml) compared to mice immunized with control-exosomes or PBS (<15 pg/ml). No IL-4 (<4 pg/ml; a Th2 marker) was detected in the T cells supernatants from animals immunized with either C. muridarum-exosomes, control-exosomes, or PBS following stimulation with C. muridarum EB. T cells stimulated with Con A secreted low levels of IL-4 in all immunized mice (45-101 pg/ml).
Changes in body weight of mice following the i.n. challenge As a measurement of the systemic effect of the infection, the body weight was determined for 10 days following the i.n. challenge. All mice lost weight for the first 2-4 days post-challenge (d.p.c.) (Fig. 4). Subsequently, mice vaccinated with C. muridarumexosomes, slowly regained most of their initial body weight in contrast to mice immunized with control-exosomes or PBS. As determined by the repeated measures ANOVA, the cumulative body weight changes over the 10 days were significantly (P < 0.05) different between mice immunized with C. muridarum-exosomes and the two negative control groups. No significant differences were observed between the control-exosomes versus PBS group (P > 0.05).

Lungs weight
As a parameter of local inflammatory responses, lungs weights were measured at 10 d.p.c. (Fig. 5b) The mean lungs weight from mice vaccinated with C. muridarum-exosomes (0.20 ± 0.01 g) was significantly lower than negative control mice immunized with exosomes or PBS (0.29 ± 0.01 g and 0.30 ± 0.01 g), respectively (P < 0.05).
Local immune responses in the lungs at 10 d.p.c To evaluate local immune responses, the lungs supernatants were collected at 10 d.p.c. and levels of IFN-γ, IL-6, TNF-α, and C. muridarum-specific IgA were determined (Fig. 6a-d). The mean levels of IFN-γ (pg/ml) in mice vaccinated with C. muridarumexosomes was 175 ± 24 pg/ml indicating the control of C. muridarum infection in the lungs. Levels of IFN-γ were significantly higher in mice immunized with control exosomes (4434 ± 228), or PBS (3869 ± 162) (P < 0.05) suggesting the presence of active infection in the lungs (Fig. 6a). The levels of IL-6 follow a similar pattern to that C. muridarum EBs were run on a 10% SDS-PAGE and blotted onto nitrocellulose paper. EBs were probed with pooled sera collected from C. muridarum-exosomes, control-exosomes, or PBS immunized mice. Lane 1, molecular weight standards. Sera from mice: Lane 2, pool pre-immune; Lane 3, d60 pool sera from C. muridarumexosomes immunized mice; Lane 4, d60 pool sera from controlexosomes immunized mice; Lane 5, d60 pool sera from PBS immunized mice; Lane 6 mAb40 that recognizes the VD1 of C. muridarum MOMP; Lane 7, polyclonal sera from C. muridarum OMP85 immunized mice; Lane 8, pool sera from C. trachomatis serovar E Pmp (A-I) immunized mice; Lane 9, mAb02 to the C. muridarum 60 kDa cysteine-rich protein; Lane 10, mAb09 to C. muridarum LPS.  Fig. 3 Binding of serum antibodies from mice immunized with C. muridarum-exosomes to synthetic C. muridarum MOMP peptides. Serum samples from mice immunized with C. muridarumexosomes, control-exosomes, and PBS were collected the day before the intranasal challenge. Their reactivity to 25-aa overlapping peptides corresponding to the C. muridarum mature MOMP was analyzed by an ELISA.

DISCUSSION
This study aimed to determine the ability of exosomes, collected from C. muridarum infected HeLa cells, to protect against an intranasal challenge with C. muridarum. MSA of exosomes from C. muridarum-infected cells identified 113 C. muridarum proteins including some previously described protective antigens, such as MOMP and Pmps, and others not previously tested in vaccines 22,29,35 . NTA determined similar number and size exosomes from the C. muridarum-infected and the non-infected HeLa preparations indicating they are membrane vesicles. Vaccination with exosomes from C. muridarum infected HeLa cells adjuvanted with CpG and Montanide elicited preferentially Th1 type humoral and cell-mediated immune responses and produced high levels of C. muridarum-specific neutralizing antibodies in serum. Protection against a respiratory challenge with C. muridarum was determined by the changes in body weight during the 10 days post-infection, lungs weight and number of IFU recovered from the lungs at day 10 post-infection. To our knowledge, this is the first time that exosomes from C. muridarum infected cells have been shown to elicit protection against a chlamydial challenge. The novel discovery of Chlamydia exosomes as a source of protective antigen(s) should be further explored.
A large volume of literature had linked exosomes with the pathophysiology of many diseases 31,36 . In addition, the protective role of exosomes against bacterial or parasitic infection is emerging 33,37,38 . For example, vaccination with outer membrane vesicles (OMV) produced by ΔtolR deleted Salmonella strains was shown to lower Salmonella burden in mice 39 . Similarly, a vaccination study with exosomes collected from Plasmodium yoelli infected reticulocytes was shown to protect mice against a lethal P. yoelli challenge 33 . Mice, immunized with extracellular vesicles collected from the adult gastrointestinal nematode Heligmosomoides polygyrus, developed protective immunity against a larval challenge 38 . Furthermore, there is an effective OMV-based human vaccine against Neisseria meningitidis already in the market 40 . These studies, along with this one, support the protective role of exosomes against some infections.
MSA of C. muridarum exosomes, revealed several important findings. Of the proteins identified in the exosomes preparation, label free peptide quantitation, detected only 6.7 mole% of total protein attributed to Chlamydia and the remainder (93.3 mole%) to human host cells. Moreover, exosomes from C. muridaruminfected HeLa cells had 6-fold greater diversity of human proteins than control exosomes collected from uninfected-HeLa cells. This suggests that Chlamydia infection may lead to an increase in the number of host cells proteins in the exosomes. This may result in an increase in the quantity or quality of signals sent via exosomes to host cells. More work needs to be done to define the chlamydia-host cell interface at the exosomal level.
Among the 113 Chlamydia proteins identified by MSA,~10 mole % was attributable to MOMP while 82 mole% was assigned to Pmps. Importantly, these amounts were sufficient to elicit significant protective immunity against the pulmonary challenge suggesting high potencies of these antigens. Considering that exosomes elicited MOMP specific immune responses we have to assume that they played a role in protection. Also, T cells epitopes in Pmps 41,42 might have helped vaccinated mice to develop T cellmediated immunity against C. muridarum. By western blot, we could not detect Pmp reacting antibodies. Antibody responses to Pmps in humans are variable and even gender specific. Hence Pmps are considered to be mainly T cell antigens 42 .
Drawbacks of using exosomes as antigens, are the presence of host cell components and chlamydial heat shock proteins (hsp) (96, 70, 60, and 40 kDa). Antibodies to Chlamydia 60 kDa hsp have been associated with the development of long-term pathogenesis such as PID, infertility, and trachoma 45 . Thus, it is highly possible that chlamydial exosomes containing 60 kDa hsp may be involved with the development of long-term pathogenesis. In vivo studies are needed to delineate the role of exosomes in the pathogenesis of chlamydial infections.
One of the limitations of this study is that the role of mRNA or miRNA in eliciting protective immunity in mice was not evaluated. In addition to protein antigens, exosomes-associated mRNAs could potentially translate proteins in the target/antigen-presenting cells. Similarly, miRNA could modulate gene expression of immune cells 46,47 . Thus, studies are needed to define the role of mRNA and miRNA in exosomes. Also, the limited yield of exosomes from infected cells may discourage mass scale production for vaccine studies. This limitation of low yield may be overcome by producing recombinant OMV. Indeed, a recombinant OMV carrying C. muridarum serine protease HtrA has been shown to elicit C. muridarum-specific neutralizing antibodies in mice 48 . Further studies are needed to determine if recombinant OMV carrying MOMP, Pmps or some of the novel antigens found here, can produce protective immunity.
In conclusion, we have shown for the first time that exosomes from Chlamydia infected cells can induce protective immunity in mice against a respiratory challenge. These new Chlamydia antigens present in exosomes should be tested to determine their ability to elicit protective immune responses.

Stocks of C. muridarum
C. muridarum strain NiggII was grown in HeLa-229 cells using Eagle's minimal essential medium (MEM) supplemented with 5% fetal bovine serum (FBS). To produce antigens for ELISA and western blot C. muridarum was also grown in McCoy cells. EB were purified as described and stored in sugar phosphate glutamate buffer (SPG) at −80°C 49 . The number of C. muridarum inclusion forming units (IFU) was determined in HeLa-229 cells by an immunoperoxidase staining with C. muridarum-specific mAbs 21 .

Preparation of exosomes
Exosomes from C. muridarum infected HeLa-229 cells were prepared as described 50 . HeLa-229 cells, grown in T175 flasks, were infected with C. muridarum (MOI 1-3) and were cultured in serum free Dulbecco's modified Eagle medium containing cycloheximide (1 μg/ml). At 30 hours post-infection the media was collected and centrifuged at 16,500 × g for 1 h. The supernatants were passed through 0.22 μm membrane filters to remove EB and were centrifuged at 120,000 × g for 70 min. The pellets were collected, washed twice with PBS, resuspended in PBS, and stored at −80°C. Control-exosomes were produced using uninfected HeLa-229 cells. The protein content in exosomes was measured by a BCA assay (Pierce, Rockford, IL). Disease burden at day 10 following the i.n. challenge with C. muridarum. a Percentage change in body weight at 10 days following the i.n. challenge. The mean is shown as a horizontal line. Each symbol represents an animal. Symbol "*" represents P value < 0.05 by the one-way ANOVA with Holm-Sidak multiple comparisons test. b Lungs weight (g) at 10 days after the i.n. challenge. The mean is shown as a horizontal line. Each symbol represents an animal. Symbol "*" represents P value < 0.05 by the one-way ANOVA with Holm-Sidak multiple comparisons test. c Number of C. muridarum IFU recovered from the lungs at day 10 after the i.n. challenge. The median is shown as a horizontal line. Each symbol represents an animal. Symbol "*" represents P value < 0.05 by the Kruskal-Wallis test with Dunn's multiple comparisons.

Mass spectrometry analyses of exosomes
dried under vacuum and redissolved in 0.1% FA in water for mass spectrometry analyses (MSA; nanoLC-MS/MS).
Using an Easy-nLC 1000, a portion of each sample was injected to a 250 × 0.075 mm (ID) nanocapillary column packed in-house with C18 ReproSil-Pur (1.9 micron), eluting the column with a gradient of CH 3 CN in 0.1% FA (0-5% over 5 min extending to 25% over 205 min and to 35% CH 3 CN over a further 30 min) at a flow rate of 250 nano liter/min. The column eluate was sprayed into an LTQ Orbitrap Velos Pro mass spectrometer, collecting precursor spectra in the range 380-1600 m/z. Up to 15 of the most intense ions in each precursor spectrum with a charge of +2 to +4 and a minimum signal of 5000 were fragmented by HCD with normalized collision energy of 30%. Ions were dynamically excluded for 40 s after two fragmentations within 30 s, via a 500-entry list, with early expiration from the list after a detection within the exclusion period falling below S/N = 2.0.

MSA search and quantitation
Raw file data were processed to peaklists by Mascot Distiller 2.7.1. Using Mascot 2.6.1, each resulting mass list was subjected to target-decoy searching against SwissProt (02/2019) plus a library of common contaminants (taxonomy: Human; Chlamydia muridarum), with Trypsin enzyme specificity, a maximum of 1 missed cleavage, parent and product mass tolerances of 20 ppm, and variable modifications of Deamidated (NQ) and Oxidation (M). Results were thresholded at p = 0.05 yielding a corresponding FDR of <5%. In-house software was used to generate a matrix of protein accessions (rows) vs. samples (columns). Using Mascot Distiller 2.7.1, data were re-searched as above, and the resulting peptides subjected to label-free quantitation with protein abundances scored from the summed intensities of the top 3 most intense peptides per protein and the resulting values ratioed for infected:uninfected samples. Only the highest scoring protein from each protein family was retained. Ratios were sorted in descending order and charted using Microsoft Excel.

Characterization of exosomes by nanoparticle tracking analyses (NTA)
Purified exosomes collected from C. muridarum infected-HeLa cells and from control non-infected HeLa cells were diluted to 1 μg/ml in 10 mM PBS, pH 7.2, for size characterization by NTA using a NanoSight 300 (Malvern Panalytical, UK). Each particle in a field of view was tracked by detecting scattered light at 488 nm and was analyzed by a proprietary software NTA (Malvern Panalytical, UK). Three videos were recorded (60 s each). All experiments were performed at room temperature.
Four weeks after the final immunization, anesthetized mice were challenged i.n. with 10 4 IFU of C. muridarum and mice were weighed daily for 10 days 22,52 . At 10 days post-challenge (d.p.c.) mice were euthanized, their lungs weighed and homogenized in SPG. Serial dilutions of the homogenates were used to infect Hela-229 cells grown in 48-well plates. Cells were then incubated for 30 h at 37°C in a 5% CO 2 incubator. Inclusions were visualized with Chlamydia-specific monoclonal antibodies (mAbs) and counted 53 . The limit of detection (BLD) was <50 C. muridarum IFU/lungs/ mouse. We had complied with all relevant ethical regulations for animal testing and research. Animal protocols were approved by the University of California Irvine, Animal Care and Use Committee.

Immunological assays
Blood from the periorbital plexus was collected before immunization and the day before the challenge 27 . To avoid the cross-reactivity between the HeLa cells proteins present in the EB preparation and the antibodies that  Fig. 6 Local immunological responses in vaccinated mice at 10 days after challenge. a IFN-γ levels in the lungs at 10 days after challenge. The mean is shown as a horizontal line. Each symbol represents an animal. Symbol "*" represents P value less than 0.05 by the one-way ANOVA with Holm-Sidak multiple comparisons test. b IL-6 levels in the lungs at 10 days after challenge. The mean is shown as a horizontal line. Each symbol represents an animal. Symbol "*" represents P value < 0.05 by the one-way ANOVA with Holm-Sidak multiple comparisons test. c TNF-α levels in the lungs at 10 days after challenge. The mean is shown as a horizontal line. Each symbol represents an animal. Symbol "*" represents P value < 0.05 by the one-way ANOVA with Holm-Sidak multiple comparisons test. d C. muridarum-specific IgA levels in the lungs at 10 days after challenge. The mean is shown as a horizontal line. Each symbol represents an animal. Symbol "*" represents P value < 0.05 by the one-way ANOVA with Holm-Sidak multiple comparisons test. might have developed in mice vaccinated with exosomes collected from HeLa cells, EB grown in McCoy cells were used as antigen. C. muridarum EB produced in McCoy cells (10 μg/ml) or recombinant C. muridarum MOMP (1 μg/ml) were used as antigens. Goat anti-mouse IgG, IgG1, and IgG2a (BD Bioscience, San Diego, CA) diluted 1:5000 for IgG and 1:1,000 for the two isotypes, were used to determine the sub-class or isotype-specificity. ABTS [2, 2′-azino-bis-(3-ethylbenzthiazoline-6-sulfonate)] (Sigma-Aldrich, St. Louis, MO) was used as the substrate and the, plates were scanned in an ELISA reader at 405 nm (Labsystem Multiscan; Helsinki, Finland).
In vitro neutralization assays were performed according to Peterson et al. with modifications 54 . In brief, duplicate sets of two-fold serial dilutions of serum were made in Ca +2 /Mg +2 free PBS containing 5% guinea pig serum as a source of complement. The serum samples were incubated with 10 4 C. muridarum IFU for 45 min at 37°C. The mixture was used to inoculate HeLa-229 monolayers grown in flat bottom 96-well plates which were centrifuged for 1 h at 1000 × g. The cells were incubated for 30 h. in culture medium with cycloheximide (1 µg/ml). Monolayers were then fixed and chlamydial IFU were stained with a C. muridarum-specific mAb 40 recognizing the C. muridarum MOMP 54 . The number of IFU was counted and neutralization was defined as greater than, or equal to, a 50% decrease in the number of IFU with respect to the controls incubated with pre-immunization sera 21 .
A T-cell lymphoproliferative assay (LPA) was performed using splenocytes collected from the immunized mice on the day immediately prior to challenge 53 . Splenic T-cells, purified using nylon wool (>95% purity), were stimulated with C. muridarum EB in the presence of antigen-presenting cells (APC). APC were pre-prepared by irradiation (3300 rads, 137 Cs) of syngeneic splenocytes and 1.25 × 10 5 cells were incubated in round bottom 96-well plates (Costar, Corning Inc.) at 37°C for 2 h with EB at a 1:1 ratio. T cells were added to APC at a ratio of 1:1. Concanavalin A (5 μg/ml) served as a positive stimulant and cell culture medium (RPMI with 10% FBS) used as a negative antigen control. Following 96 h of incubation, cell proliferation was measured by adding 1 µCi of [methyl 3 H] thymidine per well. The mean counts per minute (CPM) were determined from triplicate culture wells using a scintillation counter.
Western blot analysis of sera from exosomes-immunized mice was conducted as described before 52 . Briefly,~200 μg of C. muridarum EB were solubilized in SDS then separated on a 7.5 cm wide Tricine-sodium dodecyl sulfate-polyacrylamide slab gel. Serum samples were diluted 1:100 and added to the membrane strips in BLOTTO followed by overnight incubation at 4°C. Detection of antibody binding employed horseradish peroxidaseconjugated goat anti-mouse antibody, and the signal was developed with 0.01% H 2 O 2 and 4 chloro-1-naphthol. mAbs recognizing MOMP, the 60 kDa cysteine rich protein, and LPS and polyclonal sera from Pmp or OMP85 immunized mice were used as positive controls. All blots were derived from the same experiment and processed in parallel.
To detect antibodies, elicited by exosomes vaccination, to linear epitopes of C. muridarum MOMP, overlapping 25-mers corresponding to the amino acid sequence of mature MOMP, were chemically synthesized (SynBioSci Corp.; Livermore, CA) 24 . Peptide 25 (p25) overlapped the N-and C-termini of MOMP. The peptides were adsorbed onto high binding affinity ELISA plates (1 µg/well of a 96-well plate) and antibody binding was determined in triplicates as described above using anti-mouse IgG 55 .

Statistical analyses
The Kruskal-Wallis test with Dunn's multiple comparisons was used to compare the numbers of C. muridarum IFU. The one-way ANOVA with Holm-Sidak multiple comparisons test was used to compare T-cell proliferative responses, levels of cytokines, lungs, and body weight changes of mice. Repeated measures ANOVA was employed to compare changes in mean body weight over the 10 days of observation.

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.