Diet, commensal microbiota‐derived extracellular vesicles, and host immunity

The gut microbiota has co‐evolved with its host, and commensal bacteria can influence both the host's immune development and function. Recently, a role has emerged for bacterial extracellular vesicles (BEVs) as potent immune modulators. BEVs are nanosized membrane vesicles produced by all bacteria, possessing the membrane characteristics of the originating bacterium and carrying an internal cargo that may include nucleic acid, proteins, lipids, and metabolites. Thus, BEVs possess multiple avenues for regulating immune processes, and have been implicated in allergic, autoimmune, and metabolic diseases. BEVs are biodistributed locally in the gut, and also systemically, and thus have the potential to affect both the local and systemic immune responses. The production of gut microbiota‐derived BEVs is regulated by host factors such as diet and antibiotic usage. Specifically, all aspects of nutrition, including macronutrients (protein, carbohydrates, and fat), micronutrients (vitamins and minerals), and food additives (the antimicrobial sodium benzoate), can regulate BEV production. This review summarizes current knowledge of the powerful links between nutrition, antibiotics, gut microbiota‐derived BEV, and their effects on immunity and disease development. It highlights the potential of targeting or utilizing gut microbiota‐derived BEV as a therapeutic intervention.


Introduction
The trillions of microorganisms inhabiting the gut, known as the gut microbiota, mostly comprise bacteria, which regulate many aspects of host physiology. Detrimental alteration of gut microbiota composition or function is associated with myriad diseases, in particular, noncommunicable diseases. Understanding how the gut microbiota regulates host processes is an active Correspondence: Prof. Laurence Macia e-mail: laurence.macia@sydney.edu.au area of research aimed at developing new strategies to prevent or treat disease. For instance, boosting production of, or directly administering, gut microbiota-derived short-chain fatty acids (SCFAs) shows immense promise in preclinical settings for alleviating or treating diseases, such as allergies and colitis [1][2][3], and in clinical settings for MS [4]. Indeed, human trials of supplementation of prebiotic fiber have shown evidence for alleviating airway inflammation, including asthma [5][6][7], however the direct link between SCFA levels and symptom alleviation was not demonstrated. Recently, bacterial extracellular vesicles (BEVs) have emerged as important mediators of microbe-host communication [8][9][10][11][12]. To date, the main focus has been on the effects of pathogen-derived BEVs on triggering and/or sustaining immune pathologies. However, the effects of commensal-derived BEVs are just starting to emerge in studies. Here, we provide an overview of the recent literature that supports commensal-derived BEVs as another key component of the gut microbiota instrumental in mutualistic host-microbial physiology.

General properties of BEVs
Extracellular vesicles are particles released by all living cells that play a role in intra-and interkingdom communication. BEVs are nanosized and formed from the parental cell membrane, and thus, comprise a lipid membrane layer with characteristics of the parental cell [13]. Biogenesis of BEVs from Gram-negative bacteria typically involves the blebbing of the outer membrane, forming outer-membrane vesicles [14]. Conversely, the biogenesis of BEV in Gram-positive bacteria requires disruption to the thick outer layer of peptidoglycan, allowing the blebbing of the inner cytoplasmic membrane, known as cytoplasmic membrane vesicles. During their formation, BEVs can encapsulate a variety of molecules from the cytosolic or periplasmic space including nucleic acids, proteins, lipids, and metabolites. BEVs are generated constitutively, however, their production is sensitive to stress conditions, and environmental stress generally promotes vesiculation. BEVs mediate a range of biological processes for bacteria within a microbiota, including nutrient and mineral exchange, waste and toxin removal, and dissemination of biological substances such as nucleic acids (horizontal gene transfer) and virulence factors [13]. BEVs are also involved in self-defense mechanisms, by delivering lysis factors to competitor species [15], or acting as decoys against bacterial and host defense factors [16,17].

BEV-host communication
Broadly, BEVs can mediate their effect on host cells via the physical interaction of membrane components (by ligand binding, membrane uptake, or membrane fusion with host cells) or by the delivery of a bioactive cargo to host cells, or a combination thereof.

Interaction with membrane components
As BEVs are derived from the bacterial membrane, they carry a variety of surface membrane components that can directly interact with eukaryotic host cells. In general, the major membrane components, LPS (Gram-negative) and lipoteichoic acid (Grampositive), activate TLR 4 and TLR2, respectively. Other surface components can also activate TLRs, such as capsular polysaccharide A from Bacteroides fragilis BEVs, which activates TLR2 [18]. Furthermore, LPS derived from Helicobacter pylori strongly activates the TLR2-TLR1 complex but weakly activates TLR4 [19], or antagonizes its signaling [20]. These differences may relate to structural variation in LPS via differential lipid A acylation and O-antigen expression [21][22][23]. In general, highly acylated lipid A (hexa-acylated) transduces a stronger signal via TLR and is more immunogenic [22]. Membrane composition also influences uptake by eukaryotic cells: Gram-negative BEVs that lack O-antigen are endocytosed via clathrin-mediated endocytosis, while the presence of O-antigen allows more rapid lipid raftmediated uptake [24]. Other than TLRs, BEVs can interact with host nucleotide-binding oligomerization domain proteins, which recognize peptidoglycan components [25].

Impact of commensal BEVs on host immunity
The majority of studies on gut microbiota BEVs have focused on pathogenic bacteria, as extensively reviewed elsewhere [13,32]. Mounting evidence indicates that gut BEVs pass through mucosal barriers under healthy conditions [33], thereby having the capacity to modulate innate immune cells, including APCs (such asDCs, monocytes and macrophages), as well as adaptive immune cells including T cells and B cells [8].
Intestinal epithelial cells lining the gut provide the first cellular barrier between the microbiota and host, maintaining appropriate compartmentalization of the microbiota in the lumen. These cells express various PRRs and are one of the first cells to interact with BEVs. BEVs demonstrate modulatory activity on cellular junctions, PRR activation, and cytokine expression [34,35]. Both changes in the type or amount of microbiota-derived BEVs can affect epithelial cell signaling [34,36]. We have shown that fecal-derived BEVs stimulated NF-κB signaling in epithelial cells in a dose-dependent manner, which promoted B-cell recruitment to the small intestine lamina propria and their differentiation to IgA-producing plasma cells [34]. This suggests that BEV levels may function as a "readout" of the gut microbiota status for the host, with increased BEVs triggering secretory IgA production to ensure the efficient exclusion of bacteria from traversing the epithelium.
Gut-derived BEVs also activate other cell subsets located within the gut as well as in distal organs. BEVs from different commensals appear to have unique effects on PRR activation and cytokine secretion. In general, however, commensal BEVs appear to promote "anti-inflammatory" phenotypes such as the induction of M2-polarized macrophages [37] and the activation of tolerogenic DCs that promote Treg [18]. A summary of the effects of commensal BEVs on host epithelial and immune cells is presented in Table 1.

Environmental factors regulating BEV production in vivo
While many studies have examined factors regulating BEV production in vitro, variables affecting their production in the intestine in vivo are poorly understood. Dietary macronutrients (protein, fats, and carbohydrate), as well as micronutrients and minerals (such as iron, zinc, and vitamins), regulate gut bacterial composition and growth, but may also regulate BEV production. Exposure to bactericidal agents, such as antibiotics and food additives, triggers BEV release in vitro and potentially in vivo. These dietary factors are summarized below in Fig. 1.

Proteins
Dietary protein regulates BEV production by modulating gut microbiota metabolism toward succinate production. Elevated succinate levels induce bacterial stress, thus, triggering increased total intestinal BEV release and increasing host secretory IgA production [34]. High succinate is associated with intestinal inflammation [38], psoriasis [39], and obesity [40], which are characterized by increased gut permeability. However, we found that neither increased succinate nor elevated intestinal BEVs affected gut permeability [34]. Regardless, this suggests that there is a link between increased succinate, BEV levels, and increased inflammation that is worthy of more investigation.
The effects of isolated amino acids on BEV production have also been identified. Glycine weakens peptidoglycan structure, alters membrane crosslinking, and inhibits bacterial growth at high concentrations [41]. The addition of glycine in media promoted vesiculation for E. coli Nissle 1917, and BEVs had altered size distribution, protein profile, and lowered endotoxin activity [42]. Cysteine deprivation induces oxidative stress and BEV production by Neisseria meningitidis [43] and Francisella tularen-sis [44]. Interestingly, D-form isomers of AAs play a role in regulating peptidoglycan structure [45], inhibiting biofilm formation [46], and synergize with antimicrobials [47]. Most D-form AAs in mammals are derived from the gut microbiota or through the intake of fermented foods [48]. Whether fermented foods stimulate gut BEV production is unknown but could be a factor underlying their health benefits. N-acetyl cysteine, a cysteine precursor found in protein-rich foods, is also marketed as an antioxidant supplement. N-acetyl cysteine altered BEV production of respiratory pathogens and blunted their immunogenicity in vitro [49].
Altogether, evidence suggests that dietary AAs and proteins modulate bacterial vesiculation (Fig. 1), offering opportunities to manipulate BEV production through dietary intervention.

Fats
Few studies have examined the effects of lipids on BEVs. Choi et al. identified that feeding mice a high-fat diet increased the size and protein composition of fecal BEVs, compared to normal chow [50]. High-fat diet-derived fecal BEVs also had increased LPS and decreased lipoteichoic acid content. High-fat diet induced changes to microbiota composition, however, BEV composition differed from these changes and was more dramatically altered between diets [50]. Fatty acid composition, particularly the ratio of saturated to unsaturated fatty acids, influences the rigidity of the bacterial phospholipid membrane [51]. Modulating fatty acid incorporation can be part of the bacterial response to environmental stress [52]. In vitro, saturated fatty acid supplementation induced BEV production by B. fragilis in a dose-dependent manner but did not affect Bacteroides thetaiotaomicron. Supplementation with low doses of unsaturated fatty acids increased and decreased BEV release for B. thetaiotaomicron and B. fragilis, respectively [53], suggesting there are both fatty acid and strain-specific effects on BEV production. Thus, the influence of fatty acids on vesiculation in vivo may be highly complex, and further work is needed to determine the effects of saturated and unsaturated fatty acids on vesiculation.
Lipids can also indirectly affect gut BEV production via bile acids. Lipid intake triggers bile secretion by the host, and both conjugated and unconjugated bile salts increased B. fragilis BEV production (as well as biofilm formation), likely by inducing bacterial stress [54]. Bile acids also reportedly induce membrane damage, increasing Bifidobacteria BEV generation [55]. This aligns with the known antimicrobial properties of bile acids, which contribute to changes in gut microbiota composition under high-fat feeding conditions [56].

Carbohydrates
Dietary fiber (complex carbohydrates) is the primary source of energy for the gut microbiota, and by-products of  Polysaccharides from dietary fiber serve as a primary energy source for many commensals, and polysaccharide digestion is facilitated by the expression of carbohydrate-active enzymes (CAZymes), which can be carried on BEVs [58]. High protein intake can cause elevated bacterial succinate production, which triggers a bacterial stress response that enhances BEV release [34]. Various properties of AAs may cause structural changes to the bacterial cell wall and/or membrane that modulate vesiculation [45][46][47]. Fatty acid ratio (saturated:unsaturated) and their availability regulates the fluidity of the bacterial phospholipid membrane [51] and may modulate BEVs in vivo. Fat intake triggers bile secretion into the intestinal lumen by host cells, which can induce bacterial stress and, thus, promote BEVs [54]. BEVs also play a role in bacterial acquisition of essential micronutrients such as iron [67] and vitamin B12 [78]. Finally, antimicrobial intake, via oral antibiotic therapy or food additives, can have bactericidal impacts on gut bacteria [90,94], and hence, likely regulates BEV levels.
fiber metabolism include fatty acid derivatives such as SCFAs. Fiber supplementation supports the expansion of polysaccharide-digesting bacteria in vivo, particularly those expressing carbohydrate-active enzymes [57]. Carbohydrateactive enzymes may be carried by BEVs, as glycosidase-enriched BEVs released by Bacteroides facilitate fiber digestion [58]. Thus, BEVs can facilitate "cross-feeding" for nonpolysaccharide-utilizing species, which may be important for supporting a healthy gut microbiota [59]. Similarly, cellulose-degrading enzymes have been identified in ruminant microbiota BEVs [60]. Carbohydrates can also directly modulate BEV production in vitro. Beta-mannan, a wood-derived saccharide with a similar structure to dietary fiber, induced BEVs in vitro from porcine gut microbiota [61], particu-larly Clostridiales, Bacilli, and Enterobacteriales phyla. Here, betamannan exposure altered BEV protein content but not expression of polysaccharide-digesting enzymes. Thus, dietary fiber supplementation in vivo may increase BEV generation by carbohydratedigesting bacteria, by increasing the amount of polysaccharide substrate. However, the impact of dietary fiber on BEV generation needs to be further characterized. Food allergy is linked to a microbiota with a diminished capacity to digest dietary fiber and, thus, release the SCFA, butyrate [62,63]. Administration of BEV enriched with carbohydrateactive enzymes could be a novel strategy to rescue SCFA generation by a dysbiotic microbiota, and thus, prevent or treat allergies.

Vitamins and minerals
The gut microbiota competes with the host for the uptake of minerals and vitamins. The most well-characterized example of this is iron, an essential element for bacterial growth, and free iron is generally low in the mammalian host. During bacterial infection, PRR activation triggers epithelial and immune cell secretion of iron-binding proteins, such as lipocalin-2, which sequester iron to limit pathogen growth [64,65]. BEVs play a role in bacterial iron acquisition, as they may contain iron-binding moieties (siderophores) that scavenge free iron and facilitate delivery to the parental cell [66]. Growth in iron-limiting conditions triggers BEV production for Mycobacterium tuberculosis [67], H. pylori [68], pathogenic and commensal E. coli [69,70], and similarly, KO of the iron-sensing locus in Citrobacter rodentium increased vesiculation [71]. Mechanisms of iron acquisition by commensal species are less studied, although Bacteroides do not produce siderophores, but may utilize siderophores of other bacteria and scavenge host-bound iron via surface-localized proteins [72][73][74][75]. Other metals can also modulate BEV production, for example, in the detoxification of bismuth for H. pylori [76] and copper for marine bacteria [77]. This may apply to the gut microbiota under conditions of metal toxicity, which could increase production of BEVs and potentially expose the host to increased metal concentrations.
Recently, Bacteroides BEVs were shown to contain proteins with high affinity for vitamin B12, an essential mammalian cofactor exclusively derived from the gut microbiota. The addition of these BEVs, inhibited the growth of a vitamin B12 auxotrophic Salmonella strain in vitro, suggesting the BEVs sequestered vitamin B12. This could be part of the strategy that gut commensals have evolved to outcompete pathogenic species. These BEVs were also shown to deliver vitamin B12 to colonic epithelial cells (CaCo 2 ), indicating that BEV-mediated transfer of metabolites and micronutrients may play an important role in the gut microbiota-host interaction [78].
Dietary components can actively modulate BEV generation, with differential effects depending on the bacteria. While a deeper understanding of the impact of gut-derived BEVs on host physiology is still needed, diet composition could be tailored to manipulate BEV production, which may support host health by optimizing fiber digestion, outcompeting pathobionts, and supplying vitamins.

Impact of bactericidal agents
Exposure to sublethal antibiotics triggers BEV release from a variety of bacterial species, which is well documented in vitro [79][80][81], and extensively reviewed previously [82]. Briefly, Liu et al. [82] classified three pathways through which antibiotics induce BEVs: (i) envelope stress; (ii) induction of the SOS response; and (iii) weakening of cell walls (the mechanism of the β-lactam class of antibiotics). Primarily, most studies have examined the effect of β-lactams on promoting vesiculation. β-lactams (which include penicillins, cephalosporins, and carbapenems) are the most commonly prescribed antibiotic class [83] and target bacterial cell wall production by inhibiting peptidoglycan synthesis, hindering growth, and cell division. The destabilization of peptidoglycan and outer membrane crosslinking decreases membrane integrity, thus, promoting vesiculation. While to date, no studies have demonstrated direct effects of antibiotic consumption on intestinal BEV production in vivo, given the well-documented disruptive effects of antibiotics on gut flora [84], it is likely that oral antibiotic therapy could significantly influence BEV production.
Intestinal bacteria may also be exposed to antimicrobial compounds through the form of food additives. Antimicrobial preservatives are used to prolong shelf life and inhibit food-borne pathogen growth, however, these compounds can have bactericidal activity against commensals [85]. The most commonly used antimicrobial agents in food (sodium benzoate, potassium sorbate, and sodium nitrite) are believed to act by a variety of mechanisms including disruption to the cell membrane and metabolism, accumulation of toxic anions [86], and oxidative stress [87]. These preservatives have been tested as an alternative to antibiotic supplementation in pig feed [88], and reduced pig intestinal bacterial load [89], showcasing their potentially dramatic impact on gut flora.
In mice, antimicrobial preservatives are associated with changes to microbiota composition and lowered diversity [85,[90][91][92] as well as changes to their functional profile (as inferred from 16S sequencing data with the PICRUSt algorithm) [93]. Commonly used antimicrobial additives show bactericidal action in vitro against commensals, such as Lactobacillus [94,95], Bacteroides, and Clostridia [94], even at concentrations considered safe in foods. Intake of these additives in processed foods is, thus, likely to regulate BEVs in vivo, which may affect humans health.
Recognizing the impact of bactericidal agents on BEV release in vivo is important because of the critical role BEVs play in antibiotic resistance. BEVs mediate the dissemination of virulence and antibiotic-resistance genes across species [96], and in particular, the transfer of β-lactam resistance elements has been identified [97]. BEVs may also facilitate antibiotic resistance by serving as "decoy" targets for antibiotics, sequestering the drug, and enabling the parental cell to persist [98]. Enzymes carried by BEVs, such as proteases and peptidases, have been shown to facilitate the degradation of antimicrobial peptides [98]. BEVs from antibiotic-resistant bacteria may protect susceptible bacteria, as Bacteroides BEVs carrying β-lactamase protected β-lactamsusceptible pathogenic and commensal bacteria in vitro [79,99]. In vivo BEVs carrying β-lactamase may have a protective effect on pathogens during respiratory co-infections [100,101]. Overuse of antimicrobials that may select for antibiotic resistance is considered an increasing threat worldwide, and thus, understanding how BEVs may contribute to this phenomenon in vivo is essential.
Many other dietary components, such as salt, emulsifiers, and artificial sweeteners, modulate host immunity via the gut microbiota [102,103]. These components also alter bacterial membrane properties and vesiculation in vitro [104,105]. While their effects on gut BEV production in vivo have not been studied, BEVs may potentially contribute to some of the effects of these food products on host immunity.

BEVs in health and disease
Evidence suggests there are gut microbiota-derived BEVs circulating in the blood of healthy individuals [106] and increased BEV levels in patients with intestinal barrier dysfunction [107]. Both dysbiosis and increased gut permeability have been identified as standard features of noncommunicable diseases (such as colitis, obesity, diabetes, and MS) and increased "leaking" of intestinal bacterial components is believed to contribute to the pathogenesis of chronic inflammation [108]. However, whether gut BEVs that translocate to distal organs contribute to the inflammatory process and disease pathogenesis is unknown. As such, there is considerable interest in interventions that target microbiota composition and intestinal barrier function to prevent or treat diseases. Since BEVs are nonreplicative and can recapitulate the effects of their parental cell, they offer an exciting alternative to current clinical therapeutics, such as live probiotics and fecal transplants, which have mixed efficacy and can be high risk.
Fecal transplants carry the hazard of introducing pathogens and antibiotic-resistant strains, and this risk can be mitigated by use of BEVs. As BEVs are nonreplicative, they also have reduced risk of introducing pathobionts. The role of BEVs in disease pathogenesis as well as their utilization in preclinical models of disease are summarized in Table 2.

Concluding remarks
Since BEVs were first discovered over 50 years ago, there has been growing emphasis on the role of gut microbiota-derived BEVs on host physiology and immunity. Working with BEVs presents technical limitations (including isolation, tracking, and characterization), which underlie the dominant focus on probiotics and microbiota-derived metabolites in the field. For example, there is currently no standardized protocol for BEV isolation, and methods vary across the literature [109,110]. Reporting of BEV concentrations used also varies, with some quantifying concentration as particle number, and others as total protein or lipid amount [111]. Thus, it can be difficult to make comparisons across studies. BEV biodistribution and uptake is most commonly tracked with the use of nonspecific membrane labeling dyes, which can have significant artifactual considerations [112]. A major limitation of the field is the understanding of how BEVs are generated and what factors regulate this process. While microbiota BEV production can be regulated by a range of environmental factors, particularly diet and antibiotics, the precise mechanisms are not well characterized. Despite this, substantial evidence highlights the broad impact of BEVs. BEVs first interact with the gut-mucosal epithelial barrier, but also show systemic effects on organs, including the nervous and metabolic systems, and on a wide range of innate and adaptive immune responses.
These broad impacts highlight the potential role of BEVs as a new therapeutic, not only in supporting gut homeostasis, but also for treating a wide range of pathologies such as dementia, obesity, or autoimmune disease. While BEVs appear to offer an alternative to probiotics, which have variable efficacy, greater understanding of their role in the host-microbiota axis is needed. Taken together, while this field is still in its infancy, BEVs offer promising therapeutic potential for a broad range of diseases.