Biofilm‐derived membrane vesicles exhibit potent immunomodulatory activity in Pseudomonas aeruginosa PAO1

Pathogenic bacteria form biofilms on epithelial cells, and most bacterial biofilms show increased production of membrane vesicles (MVs), also known as outer membrane vesicles in Gram‐negative bacteria. Numerous studies have investigated the MVs released under planktonic conditions; however, the impact of MVs released from biofilms on immune responses remains unclear. This study aimed to investigate the characteristics and immunomodulatory activity of MVs obtained from both planktonic and biofilm cultures of Pseudomonas aeruginosa PAO1. The innate immune responses of macrophages to planktonic‐derived MVs (p‐MVs) and biofilm‐derived MVs (b‐MVs) were investigated by measuring the mRNA expression of proinflammatory cytokines. Our results showed that b‐MVs induced a higher expression of inflammatory cytokines, including Il1b, Il6, and Il12p40, than p‐MVs. The mRNA expression levels of Toll‐like receptor 4 (Tlr4) differed between the two types of MVs, but not Tlr2. Polymyxin B significantly neutralized b‐MV‐mediated cytokine induction, suggesting that lipopolysaccharide of native b‐MVs is the origin of the immune response. In addition, heat‐treated or homogenized b‐MVs induced the mRNA expression of cytokines, including Tnfa, Il1b, Il6, and Il12p40. Heat treatment of MVs led to increased expression of Tlr2 but not Tlr4, suggesting that TLR2 ligands play a role in detecting the pathogen‐associated molecular patterns in lysed MVs. Taken together, our data indicate that potent immunomodulatory MVs are produced in P. aeruginosa biofilms and that this behavior could be a strategy for the bacteria to infect host cells. Furthermore, our findings would contribute to developing novel vaccines using MVs.


INTRODUCTION
Most bacteria release membrane vesicles (MVs), which are spherical nanoparticles with a diameter of 20-400 nm, into the extracellular environment. 1,2They consist mainly of membrane proteins and phospholipids filled with periplasmic and cytoplasmic components.MVs have multiple functions in bacterial-host interactions, serving as vehicles for toxins, enzymes, and immunomodulatory molecules, as well as various signals for cell-cell communication.MVs that are naturally shed from the outer membrane of Gram-negative bacteria are called outer membrane vesicles (OMVs) and carry pathogen-associated molecular patterns (PAMPs), including lipopolysaccharide (LPS) and outer membrane proteins such as lipoproteins. 3Toll-like receptors (TLRs), expressed by host cells, recognize PAMPs and proinflammatory cytokines and recruit immune cells to the infected site. 4][9][10][11] Therefore, OMVs are crucial carriers of virulence factors in host cells for innate immune responses.
Pseudomonas aeruginosa is a prominent human pathogen that establishes chronic biofilm infections, including cystic fibrosis.According to a report by the National Institutes of Health, 65% and 80% of all microbial and chronic infections, respectively, are associated with biofilm formation. 12Biofilms are formed in the lungs, corneas, and skin of individual animals, and contribute to the development and spread of infectious diseases. 13Therefore, a comprehensive understanding of the effect of biofilmderived MVs on immunomodulation is critical.
The physicochemical properties and components of P. aeruginosa MVs differ under different growth conditions. 14,158][19][20][21] Under DNA-damaged conditions, both OMVs and explosive cell lysis-mediated MVs are formed. 22][25][26][27] The internalization of virulence factors into MVs eliminates the need for direct interactions between the bacteria and the host cells.MVs released by P. aeruginosa contain virulence factors, such as alkaline phosphatase, β-lactamase, hemolytic phospholipase C, and CFTR-inhibitory factor Cif. 24,28,29 Thus, P. aeruginosa MVs are heterogeneous complexes of PAMPs such as LPS and virulence proteins.Macrophages, which initiate the inflammatory defense against colonizing bacteria, are more sensitive to the potent stimulus of P. aeruginosa MVs to pure LPS. 30,31Administration of P. aeruginosa MVs to mice was found to be an effective vaccine to prevent infection with P. aeruginosa. 32lthough the biofilm growth mode predominates in disease environments and MV formation is enhanced in biofilms, most studies on MV immunomodulation have been conducted using MVs derived from planktonic cells of P. aeruginosa.However, the components and formation mechanisms of planktonic-derived MVs (p-MVs) and biofilm-derived MVs (b-MVs) were significantly different, and it is speculated that their immunomodulatory activities were not identical.In this study, we evaluated the immunomodulatory activities of p-MVs and b-MVs.We found that the cytokine production was better induced by b-MVs, suggesting that b-MVs contain unidentified immunomodulatory substances.These results provide novel insights into the mechanism by which P. aeruginosa MVs induce inflammation in host cells and the development of MV vaccines against P. aeruginosa.

MATERIALS AND METHODS
Microbial strain and growth conditions P. aeruginosa PAO1 33 was routinely incubated aerobically at 37°C in Miller's Lysogeny Broth (LB) (1% w/v tryptone, 0.5% w/v yeast extract, and 1% w/v NaCl) or LB containing 1.5% agar.Culture conditions for the planktonic and biofilm states have been described by Kanno et al. 34 Briefly, preculture of PAO1 was inoculated to achieve an optical density of 0.02 at 600 nm and grown at 37°C for 12 h in 100 mL LB medium (500 mL flask) at 200 rpm for the planktonic state or in 10 mL LB medium in Petri dish without shaking for the biofilm state.

Extraction and purification of MVs
Extraction, purification, and quantification of MVs were performed as previously described. 34P. aeruginosa was grown overnight under shaking and static conditions.Under static biofilm conditions, the mucoid and viscous films attached to the bottom of the Petri dish were scraped using a scraper.The collected cultures were sonicated (38 kHz for 5 min) and vortexed to isolate vesicles from the bacterial cell surfaces.After centrifugation (6000g for 15 min at 4°C) of the bacterial cultures, the supernatant was filtered sequentially through 0.45 and 0.20 µm pore size cellulose acetate membranes (Advantec).The cell-free supernatants were ultracentrifuged (150,000g for 2 h at 4°C) using Himac CP80WX, and the pellets were resuspended in 50 mM HEPES (pH 6.8) and 0.85% NaCl (HEPES-NaCl buffer) and washed with the same buffer using a P45AT angle rotor (Eppendorf Himac Technologies).The MV samples were stored at −80°C prior to use.
The vesicles were purified by density gradient ultracentrifugation using different layered concentrations of iodixanol.The extracted vesicles were suspended in 1 mL of 45% (w/v) iodixanol (OptiPrep; Axis-Shield Diagnostics Ltd.) in HEPES-NaCl buffer and transferred to the bottom of ultracentrifuge tubes.Iodixanol gradients (2 mL of 40, 35, 30, and 25% and 1 mL of 20%) were layered over the crude vesicle samples and ultracentrifuged at 100,000g for 16 h at 4°C using a swing-out bucket rotor (P40ST; Eppendorf Himac Technologies).The MV fraction was further ultracentrifuged (200,000g for 1 h at 4°C) using a P50A3 rotor, and the pellet was reconstituted in HEPES-NaCl buffer or phosphate-buffered saline (PBS).

Treatment of purified vesicles
For polymyxin B treatment, purified MVs (10 μg protein/ mL) were preincubated with 100 μg/mL of polymyxin B for 30 min at room temperature.For heat treatment, the purified MVs were heated at 80°C or 100°C for 15 min.For ultrasonic disruption of MVs, purified MVs were treated using a homogenizer (NR-50M Microtec) equipped with an NR-50M-MT3 tip under the condition of pulsewidth modulation control at 50%-60% for 1 min on ice, kept for 1 min, and homogenized again for 1 min.

Quantification of proteins, phospholipids, LPS, and DNA
Protein concentration was quantified by the BCA protein assay using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific).Purified MVs were lysed with 1% SDS, and the protein concentration of homogenized vesicles was determined on a microtiter plate.Bovine serum albumin (BSA) was used as a standard, and the absorbance at 562 nm was measured using SpectraMax i3 (Molecular Devices).
The phospholipid concentration was determined using FM4-64 with linoleic acid as a standard, according to a previous report. 35The MV sample was mixed with 10 μg/mL FM4-64 in a black microtiter plate at 37°C for 30 min, and the fluorescence (excitation at 506 nm and emission at 750 nm) was measured.
3-deoxy-α-D-manno-octulosonic acid (KDO) of LPS associated with MVs was determined based on a previous report. 36Briefly, 50 μL of 200 μg protein/mL vesicles and 0.5 N sulfuric acid were mixed in 15 mL tubes and heated at 100°C for 15 min.After cooling to room temperature, samples were mixed with 200 μL of 100 mM arsenic acid/0.5 N HCl, followed by 800 μL of 4 mM thiobarbituric acid, and heated at 100°C for 10 min.After cooling to room temperature, the samples were mixed with 1.5 ml of butanol solution (n-butanol/sulfuric acid, 19:1), vortexed for 30 s, and left for 10 min. 1 ml of the butanol layer was transferred to a new microtube and centrifuged at 10,000g for 2 min to remove residual water.The absorbances of the butanol layer at 552 nm and 509 nm were measured, and KDO concentrations (mM/ μg protein) were calculated from the A 522 -A 509 value using purified KDO as a standard.
The concentration of DNA associated with MVs was determined using a Quant-iT PicoGreen assay (Thermo Fisher Scientific), according to a previous report, 37 with some modifications.Purified MVs (10 μg protein/mL) were disrupted using a homogenizer (NR-50M, Microtec) to measure both internal and external DNA.DNA concentration was measured according to the manufacturer's instructions.λDNA was used as a standard, and fluorescence (excitation at 480 nm and emission at 520 nm) was measured.

Nano-tracking analysis
The particle counts and size distributions were examined using nano-tracking analysis (NTA) with a ZetaView PMX 120 (Particle Metrix).Polystyrene nanospheres (100 nm) were used to assess the size and concentration measurement biases.Vesicles were diluted to allow for 50-200 particles in one frame and measured at 25°C and the value of 85 in the sensitivity using a 488 nm laser.

Scanning electron microscopy observation
For scanning electron microscopy (SEM) observations, purified vesicles were dropped onto glass coated with 0.5% poly-L-lysine hydrobromide (Sigma-Aldrich).Fixation, dehydration through an ascending acetone series, critical point drying, and osmium coating were performed as previously described. 38The samples were observed using an SEM Regulus 8220 microscope (Hitachi High-Tech).

Cell culture
J774.1 mouse macrophage-like cells (RIKEN BRC) were grown in RPMI-1640 medium (Gibco, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS) and penicillin-streptomycin at 37°C in 5% CO 2 .Confluent cells were harvested and 900 μL of 6.6 × 10 5 cells were seeded into wells of 24-well plates.Vesicles or PBS (100 μL) was added to each well, followed by incubation until the time of assay.

Quantification of cytokine gene expression by real-time PCR
Quantification of Tnfa, Il1b, Il6, Il12p40, Tlr2, Tlr4, and Tlr9 expression was performed using quantitative PCR (qPCR) as previously described. 39Briefly, J774.1 cells were incubated with MVs (1 ng-protein/mL or arbitrary concentrations) for 2 h in a 24-well plate.PBS was added instead of MVs into the J774.1 cell culture as an untreated sample.RNA was extracted using the RNeasy Mini Kit (QIAGEN), and cDNA was synthesized using the ReverTra Ace qPCR RT Master Mix with Genomic DNA Remover Kit (TOYOBO) according to the manufacturer's instructions.The TaqMan qPCR mixture was prepared by mixing Premix Ex Taq (probe qPCR) (Takara Bio), cDNA, primer pairs, and probes (Table 1).The sequences of specific primer pairs and probes were based on previous reports. 40,41qPCR was performed using an Applied Biosystems 7500 Fast Real-Time PCR System or a StepOne Real-Time PCR system (Thermo Fisher Scientific).Expression levels were normalized to ActB, which was used as the housekeeping gene.Each value was calculated relative to that of the untreated cell culture.

Toxicological evaluation of MVs on macrophages
The cytotoxicity of MVs on the macrophage-like cell J774.1 was evaluated by lactate dehydrogenase (LDH) 42 release, and cell viability was measured using the PrestoBlue assay.
J774.1 was inoculated into a 96-well microtiter plate at 4.0 × 10 5 cells and 50 μL of medium per well and grown to confluence.Arbitrary concentrations of MVs (50 μL) or PBS as a control were added to each well and incubated for 6 h.LDH assay was performed using the Cytotoxicity Detection Kit Plus (Roche) according to the manufacturer's instructions, and absorbance was measured at 492 nm using a SpectraMax i3 plate reader (Molecular Devices).Cells removed from the supernatants were washed with culture medium and resuspended with 100 μL medium in each well.The PrestoBlue assay was performed using a kit (Thermo Fisher Scientific) according to the manufacturer's instructions, and fluorescence (excitation at 560 nm and emission at 590 nm) was measured.

Statistical analysis
For all experiments, results were calculated from more than three replicates and are presented as mean ± standard error (SE) of the mean.Student's t tests were used for statistical analysis to compare the two groups.Comparisons among three or more samples were statistically analyzed using GraphPad PRISM V10 using one-way analysis of variance (ANOVA) with Tukey's multiple comparisons.

Biofilm-derived vesicles have high inflammatory activity
Previously, we found that MV formation in P. aeruginosa was enhanced in liquid static culture (referred to as the biofilm state) on a Petri dish compared with shaking culture in a flask (referred to as the planktonic state). 34The size distribution was not significantly different between p-MVs and b-MVs; however, the protein patterns were different, suggesting that the effects of these MVs on inflammation in cells are different.When the phospholipid and LPS concentrations per protein were evaluated in each purified MV, the phospholipid concentration of b-MVs was significantly higher than in p-MVs (Figure 1a).The LPS concentration was also higher in b-MVs (Figure 1b), suggesting that b-MVs are more inflammatory.Furthermore, MV-associated DNA was also abundant in b-MVs (Figure 1c).When NTA was performed at 10 ng protein/mL MV, particle numbers were not significantly different between p-MVs and b-MVs (5.1 × 10 9 and 6.0 × 10 9 particles/μg-protein, respectively).
To investigate the immunomodulatory activity of each MV, mouse macrophage-like cells J774.1 were exposed to MVs for 2 h, and cytokine mRNA expression was examined using real-time PCR.Cytokines TNF-α, IL-1β, IL-6, and IL-12p40 were the selected proinflammatory cytokines to be evaluated based on their previously described upregulation by P. aeruginosa infection. 30,43lthough Tnfa expression was not significantly different (Figure 2a), the expression of Il1b, Il6, and Il12p40 was significantly increased by the addition of b-MVs compared with p-MVs (Figure 2b-d).These mRNA expressions were also dependent on MV concentration (Supporting Information S1: Figure S1).These results suggest that b-MVs are highly capable of being recognized by macrophages.

Biofilm-derived vesicles alter the mRNA expression level of Tlr4
To investigate the factors responsible for the different immunostimulatory activities of these MVs, we focused on TLR2 and TLR4 because the microbial products that strongly stimulate innate immunity are associated with TLRs on the plasma membrane of macrophages.TLR2 recognizes lipoproteins, 8,44,45 whereas TLR4 recognizes LPS. 46Real-time PCR results of their gene expression showed that Tlr2 gene expression was enhanced by the addition of MVs, but the effects did not differ between p-MVs and b-MVs (Figure 3a).In contrast, Tlr4 gene expression was repressed by the addition of MVs (less than 0.4-fold), and the inhibitory effect was more pronounced for b-MVs (Figure 3b).Immune stimulation by LPS is very strong and tends to cause excessive immunity; a decrease in Tlr4 gene expression by LPS has been reported for self-protection. 47Therefore, the results of this study suggest that macrophages exhibit self-protection against LPS-containing MVs.

LPS present in MVs plays a critical role in causing inflammation
Since b-MVs contain more LPS and decrease Tlr4 gene expression compared to p-MVs, the prominent cytokine induction by b-MVs was thought to be due to LPS.Polymyxin B binds to LPS lipid A and counteracts the response of the TLR4 complex to LPS.To confirm the contribution of LPS to MV-mediated inflammatory response, the effect of polymyxin B-pretreated vesicles on cytokine production by macrophages was investigated.The effect of polymyxin B treatment on the gene expression of Tnfa was minimal (Figure 4a,b), but it significantly suppressed the expression of Il1b, Il6, and Il12p40 (Figure 4c-h).Notably, the induction of these cytokine productions by b-MV was completely inhibited by polymyxin B (Figure 4d,f,h).Although Tlr2 expression was downregulated by polymyxin B treatment (Figure 4i,j), Tlr4 expression was upregulated (Figure 4k,l).This result suggests that repression of Tlr4 by self-protection against LPS, induced by MV addition, was recovered by polymyxin B treatment.Thus, LPS is strongly F I G U R E 2 Biofilm-derived membrane vesicles (b-MVs) show higher inflammatory activity against macrophages than planktonic-derived MVs (p-MVs).Following incubation of J774.1 cells with 1 ng-protein/mL MVs for 2 h, the mRNA expression levels of Tnfa (a), Il1b (b), Il6 (c), and Il12p40 (d) were analyzed by real-time PCR.Shown are the data following normalization against the level of Actb.Each value was calculated relative to that of the untreated sample, in which PBS was added instead of MVs.PBS, phosphate-buffered saline.Data are means ± SE (n = 3).*p < 0.05; **p < 0.01; ns, not significant (Student's t test).
F I G U R E 3 Levels of Tlr2 and Tlr4 in macrophages against p-MVs and b-MVs.Following incubation of J774.1 cells with 1 ng-protein/mL MVs for 2 h, the mRNA expression levels of Tlr2 and Tlr4 were analyzed by realtime PCR.Shown are the data following normalization against the level of Actb.Each value was calculated relative to that of the untreated sample, in which PBS was added instead of MVs.b-MV, biofilm-derived membrane vesicle; p-MV, planktonic-derived membrane vesicle; PBS, phosphatebuffered saline.Data are means ± SE (n = 3).*p < 0.05; ns, not significant (Student's t test).involved in the inflammatory response induced by both types of MVs, and its contribution is particularly pronounced in b-MVs.

Heat treatment enhances the inflammatory activity of biofilm-derived MVs
The effect of heat-treated MVs on cytokine production by macrophages was examined to determine whether substances other than LPS influence the inflammatory response to MVs.LPS is heat-stable, whereas the structure of proteins changes with heat.Contrary to our expectations, heattreated b-MVs significantly enhanced cytokine production, including Tnfa, Il1b, Il6, and Il12p40, and these effects were observed by heat treatment at 100°C but not at 80°C (Figure 5a-d).In contrast, the heat treatment of p-MVs did not increase the inflammatory response.
SEM observations of the state of the MVs after heat treatment showed no significant change in p-MVs.In contrast, the number of b-MVs decreased after heat treatment at 100°C (Figure 6a).This decrease in the number of b-MVs was also confirmed by NTA (Figure 6b).These results indicate that the structure of b-MVs is affected by heat treatment, and endogenous b-MV substances enhance cytokine production.The small number of MVs, approximately 100 nm in size, were probably residual MVs from heat treatment or naturally reconstituted MVs.
To further investigate the effect of heat-treated b-MVs on macrophage inflammation, the expressions of Tlr2, Tlr4, and Tlr9 were examined.Although heat-treated b-MVs showed enhanced Tlr2 expression (Figure 7a), there was no F I G U R E 4 LPS is strongly associated with the inflammatory activity of vesicles to macrophages.J774.1 cells were exposed to 1 ng-protein/mL p-MVs (a, c, e, g, i, k) or b-MVs (b, d, f, h, j, l) treated without (control) or with 100 μg/mL polymyxin B (PMB) for 2 h and the mRNA expression levels of Tnfa (a, b), Il1b (c, d), Il6 (e, f), Il12p40 (g, h), Tlr2 (i, j), and Tlr4 (k, l) were analyzed by real-time PCR.Shown are the data following normalization against the level of Actb.Each value was calculated relative to that of the untreated sample, in which PBS was added instead of MVs.b-MV, biofilm-derived membrane vesicle; LPS, lipopolysaccharide; p-MV, planktonic-derived membrane vesicle; PBS, phosphate-buffered saline.Data are means ± SE (n = 3).*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant (Student's t test).change in Tlr4 or Tlr9 expression induced by b-MVs with or without heat treatment (Figure 7b,c).These results suggest that substances recognized by TLR2 influence the inflammatory response enhanced by heat-treated b-MVs.

Heat treatment does not alter the cytotoxicity of biofilm-derived MVs
MVs released from P. aeruginosa biofilm contain substances that increase the gene expression of Tlr2.It has been reported that TLR2 induces NF-κB activation or apoptosis via MyD88, 10 and P. aeruginosa MVs induce lung inflammation in vivo via TLR2 and TLR4. 31In addition, some pattern recognition receptors (PRRs) also recognize danger-associated molecular patterns (DAMPs) released by injured cells and tissues that activate inflammatory responses. 48o test whether the differences in cytokine induction were due to the toxicity of the DAMPs contained in MVs to macrophages, cytotoxicity, and cell viability assays were performed.Cytotoxicity analysis using LDH showed that 5 ng/mL MVs, a concentration sufficient to induce cytokine production, was not cytotoxic to macrophage cells, whereas excessive amounts of MVs (5 μg/mL) were cytotoxic (Figure 8a).Notably, heat treatment did not affect b-MV cytotoxicity.The results of the cell viability assay using PrestoBlue showed that viability was inhibited by 2 and 5 μg/mL MVs, whereas 5 ng/mL MVs, on the contrary, promoted cell proliferation (Figure 8b).These results suggest that the difference in induced cytokine production between p-MVs and b-MVs and the enhanced cytokine production by heat treatment of b-MVs are not due to differences in cytotoxicity.

Homogenized biofilm-derived MVs possess inflammatory activity
Since structurally disrupted b-MVs enhance the inflammatory response, we confirmed this phenomenon using the other common disruption method.Purified b-MVs were homogenized by ultrasonication, and the effect of the lysate on cytokine production was examined by qPCR.The results showed that lysed b-MVs significantly enhanced cytokine production (Tnfa, Il1b, Il6, and Il12p40) compared to intact b-MVs (Figure 9), suggesting that immunoreactive PAMPs are present in b-MVs.These PAMPs are thought to be displayed on the surface of reconstituted MVs, as ultrasonically disrupted extracellular vesicles can be reconstituted. 49I G U R E 5 The inflammatory activity of biofilm-derived vesicles is enhanced by heat treatment.MVs were incubated at 80°C or 100°C and J774.1 cells were exposed to each MV at a concentration of 1 ng-protein/mL in the culture for 2 h.The mRNA expression levels of Tnfa (a), Il1b (b), Il6 (c), and Il12p40 (d) were analyzed by real-time PCR.N, 80, and 100 mean no heat treatment, 80°C, and 100°C, respectively.Shown are the data following normalization against the level of Actb.Each value was calculated relative to that of the untreated sample, in which PBS was added instead of MVs.b-MV, biofilm-derived membrane vesicle; p-MV, planktonic-derived membrane vesicle; PBS, phosphate-buffered saline.Data are means ± SE (n = 3).*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant (one-way ANOVA with Tukey's multiple comparisons).

DISCUSSION
It is well known from previous studies that MVs derived from pathogenic bacteria have an immunomodulatory effect.Several studies have shown that MVs produced by P. aeruginosa induce inflammation in macrophages via the activation of TLR2 and TLR4. 30,31Among P. aeruginosa, MVs derived from a multidrug-resistant strain showed a particularly high proinflammatory effect compared to those from an antibiotic-susceptible strain and a standard laboratory strain. 50However, the characteristics of MVs from the laboratory strain P. aeruginosa PAO1 differ under various culture conditions and growth phases. 14,16,51This study is the first to evaluate differences in cytokine secretion

F I G U R E 7
The response of Toll-like receptors in macrophage to heat-treated MVs.MVs were incubated at 80°C or 100°C and J774.1 cells were exposed to each MV at a concentration of 1 ng-protein/mL in the culture for 2 h.The mRNA expression levels of Tlr2 (a), Tlr4 (b), and Tlr9 (c) were analyzed by real-time PCR.N, 80, and 100 mean no heat treatment, 80°C, and 100°C, respectively.Shown are the data following normalization against the level of Actb.Each value was calculated relative to that of the untreated sample, in which PBS was added instead of MVs.b-MV, biofilm-derived membrane vesicle; p-MV, planktonic-derived membrane vesicle; PBS, phosphate-buffered saline.Data are means ± SE (n = 3).*p < 0.05; ns, not significant (one-way ANOVA with Tukey's multiple comparisons).The data presented here indicate that b-MVs stimulated the expression of Il1b, Il6, and Il12p40 by macrophages better than p-MVs, but the expression of Tnfa did not differ between the two MV treatments, under these experimental conditions.The increased expression of these cytokines was due to the higher LPS content in b-MVs, as the treatment of b-MVs with polymyxin B significantly inhibited the expression of Il1b, Il6, and Il12p40.This result is consistent with a previous report that stated naturally produced MVs from P. aeruginosa induce a strong innate immune response in the presence of LPS and flagellin. 30SEM observations confirmed that MVs purified by density gradient ultracentrifugation in this study did not contain liberated flagellin debris; however, the flagellin proteins found in MV may play a role in immunogenesis.The reason for the high level of LPS in b-MV is not clear, but the presence of biofilm-specific genes involved in the induction of MVs in P. aeruginosa suggests that the mechanism of MV formation is different in planktonic and biofilm conditions.In the future, a more detailed elucidation of the mechanisms will clarify what causes the high levels of LPS in MVs released in biofilms.
Increased cytokine production by bacterial LPS is due to the recognition of LPS by TLR4.We showed that b-MVs decreased the gene expression of Tlr4 compared to p-MVs, and treatment of b-MVs with polymyxin B promoted its expression.3][54] This has been suggested to be a self-protection mechanism against an excessive immune response 47 and is likely due to the fact that LPS rapidly induces the degradation of the critical adaptor Mal. 55urprisingly, cytokine stimulation by b-MVs was enhanced by heat treatment or homogenization.Notably, this enhancement was observed in b-MVs but not in p-MVs, suggesting that biofilm-specific compounds in MVs are associated with the induction of cytokine production.Furthermore, an enhanced effect was observed not only for Il1b, Il6, and Il12p40 but also for Tnfa, whose production was not altered when exposed to either p-MVs or b-MVs.This unique observation was likely due to the release of internal PAMPs in the b-MVs, and the released PAMPs or reconstituted MVs covered with PAMPs enhanced the cytokine production.No change in Tlr4 gene expression was observed between the untreated and heat-treated b-MVs, suggesting that PAMPs other than LPS are involved in excessive cytokine production.
Several studies have investigated the differences in DNA, proteins, and phospholipids in P. aeruginosa MVs derived from the planktonic and biofilm states.DNA is an important component of P. aeruginosa MVs, 56 and the release of extracellular DNA (eDNA) is enhanced in biofilm [57][58][59] ; eDNA is highly associated with MVs in biofilms, 51 and b-MVs have high efficiency for bacterial horizontal gene transfer. 60DNA was also more abundant in b-MVs than in p-MVs in this study; however, the expression of Tlr9, which binds to bacterial CpG DNA, did not change after exposure to p-MVs and b-MVs or between heat treatments, suggesting that MV components other than DNA are likely to be immunoreactive.In contrast, P. aeruginosa MVs have been shown to serve as DNA cargo for transfer into host cells 61 and to alter DNA methylation patterns in certain host cells. 62Therefore, it is possible that the DNA contained in MVs derived from P. aeruginosa biofilms may have a substantial influence on macrophage characteristics other than cytokine production, which were not analyzed in this study.
Although the gene expression of Tlr4 and Tlr9 did not change between heat-treated and non-treated b-MVs, that of Tlr2 was significantly increased by the heat treatment of While TLR2 recognizes lipoproteins in Gram-negative bacteria, 5,55 the mRNA expression of Tlr2 is upregulated by LPS, 63,64 peptidoglycan, 9,65,66 and lipoproteins 67,68 in Gram-negative bacteria.In contrast, another report indicated that both peptidoglycan and high concentrations of Escherichia coli LPS activate Tlr2 expression but suggested that the latter is probably due to contaminated lipoproteins. 65Our results indicated that treating b-MVs with polymyxin B represses Tlr2 expression, suggesting that P. aeruginosa LPS affects TLR2.Given the lack of transcriptional changes in Tlr4, lipoproteins or peptidoglycans are likely to be the Tlr2 gene transcriptional inducers in b-MVs.Different MV protein profiles between planktonic and biofilm states have been reported in several studies, [17][18][19][20][21] and the speculation that Tlr2-inducing lipoproteins are highly concentrated in b-MVs is consistent with a previous report showing that the abundance of outer membrane proteins is higher in b-MVs than in p-MVs in P. aeruginosa. 17The effects of heattreated conformationally altered lipoproteins on Tlr2 expression need further investigation.
Although several reports have shown the upregulation of Tlr2 expression by peptidoglycan, 9,65,66 whether TLR2 recognizes peptidoglycan is controversial; some reports support this, 7,9 whereas others do not. 11,69Peptidoglycan turnover is crucial for MV formation and toxicity in P. aeruginosa, 70 and MVs deliver peptidoglycans to host cells. 71Additionally, the OmpA family lipoproteins contain a predicted peptidoglycan-binding region at the C-terminus, and P. aeruginosa PAO1 contains four OmpA homologs (OprF, OprL, PA0833, and PA1048). 72,73Therefore, peptidoglycans and lipoproteins may also be involved in increased Tlr2 expression induced by heat-treated b-MVs.
Importantly, infection models often use heat-killed bacteria that can alter the predominant roles of participating PRRs.A previous study indicated that TLR2 is activated by heat-killed Treponema denticola but not live bacteria and suggested that heat-labile inhibitor signaling as well as ligands activating TLR2 are present in T. denticola. 74In contrast, heat-killed P. aeruginosa PAO1 showed the same effect of increasing Tlr2 expression.Our results were slightly different from these reports because Tlr2 expression was activated not only by heat-labile b-MVs but also by homogenized b-MVs, and differences in the effectiveness of TLR2 inhibitors localized to b-MVs should also be considered.
It should be noted that MVs possess self-adjuvant properties that induce both cellular and humoral immune responses, making them promising vaccine candidates.The utility of MV as a vaccine against P. aeruginosa for the treatment of pulmonary infections and keratitis has been shown. 32,75Approaches to developing more effective P. aeruginosa MV vaccines have included the removal of virulence factors, the expression of bivalent antigens, and the conjugation of the diphtheria toxoid to MVs. [76][77][78] Our finding that the effect of MV on immunity varies T A B L E 1 Primers and probes used in this study.Primer namePrimer sequence (5′-3′) Probes were labeled with FAM (5′) and TAMRA (6-carboxytetramethylrhodamine) (3′).F I G U R E 1 Characteristics of planktonic-derived membrane vesicles (p-MVs) and biofilm-derived MVs (b-MVs).(a) The ratio of phospholipid/protein in the vesicles.(b) The ratio of KDO/protein in the vesicles.(c) The ratio of DNA/protein contained in the vesicles.Data are means ± SE (n = 3).**p < 0.01; ****p < 0.0001 (Student's t test).

F I G U R E 6
Effect of heat treatment on MVs.(a) Scanning electron microscopy (SEM) observation of MVs of p-MVs (left) and b-MVs (right) without and with heat treatment at 80°C or 100°C.Bar = 100 nm.(b) The nano-tracking analysis (NTA) result of p-MVs (left) and b-MVs (right) with and without heat treatment at 100°C.b-MV, biofilm-derived membrane vesicle; p-MV, planktonic-derived membrane vesicle.
P. AERUGINOSA BIOFILM-DERIVED VESICLES ARE IMMUNOMODULATORY | 231 by MVs derived from planktonic and biofilm states.Our results indicate that b-MVs have greater immunomodulatory effects against macrophages and that lysed b-MVs contain more PAMPs.

F I G U R E 8
Low concentrations of vesicles with inflammatory responses do not cause cytotoxicity.MVs were incubated at 80°C or 100°C and J774.1 cells were exposed to p-MVs or b-MVs at a final concentration of 5 ng/mL, 2 μg/mL, or 5 μg/mL MVs.(a) The cytotoxicity of MVs to macrophage cells evaluated using an LDH detection reagent.(b) Cell viability of macrophage cells evaluated using PrestoBlue reagent.The percentage was calculated by live cells treated with PBS (live cell control) and those treated with 1% Triton X (dead cell control).N, 80, and 100 mean no heat treatment, 80°C, and 100°C, respectively.b-MV, biofilm-derived membrane vesicle; LDH, lactate dehydrogenase; p-MV, planktonic-derived membrane vesicle; PBS, phosphate-buffered saline.Data are means ± SE (n = 3).BD, below the detection limit.F I G U R E 9 Lysed biofilm-derived vesicles show high inflammatory activities.The mRNA expression levels of macrophages exposed to non-treated or homogenized b-MVs (1 ng-protein/mL) are shown.The relative gene expressions of Tnfa (a), Il1b (b), Il6 (c), and Il12p40 (d) were analyzed by real-time PCR.Shown are the data following normalization against the level of Actb.Each value was calculated relative to that of the untreated sample, in which PBS was added instead of MVs.b-MV, biofilm-derived membrane vesicle; p-MV, planktonic-derived membrane vesicle; PBS, phosphate-buffered saline.Data are means ± SE (n = 3).*p < 0.05; **p < 0.01; ns, not significant (Student's t test).
P. AERUGINOSA BIOFILM-DERIVED VESICLES ARE IMMUNOMODULATORY | 233 significantly with culture conditions and heat treatment be an important insight for vaccine development.Taken together, this study provides new insights into the previously uncharacterized role of P. aeruginosa biofilmderived MVs on host immunomodulation.While LPS is a potent stimulator of proinflammatory responses in b-MVs, the disruption or reconstruction of b-MVs has a stronger effect on macrophages.PAMPs other than LPS encapsulated in b-MV increase Tlr2 expression, suggesting that they are potent immunomodulatory factors.Future studies should focus on determining the factors that are critical to the observed responses.The findings of this study could potentially influence the advancement of diagnostic tools, vaccines, and treatments for chronic pulmonary diseases, including cystic fibrosis.