Bacterial membrane vesicles, originating from both Gram-negative and Gram-positive bacteria, play various roles in bacterial survival and biological functions [1,2,3,4], including microbial virulence, cellular crosstalk, and host immune response modulation. Furthermore, they possess unique targeting and packaging abilities [5,6,7]. Bacterial membrane vesicles derived from Gram-positive bacteria are called membrane vesicles (MVs) as they are originating from the cytoplasmic membrane, while bacterial membrane vesicles derived from Gram-negative bacteria are called outer membrane vesicles (OMVs) as they originate from the outer membrane of the bacterial cell. They both range between 20 and 500 nm in size and contain parental bacterial cell materials [3,8,9,10,11,12], but since they originate from different parts, their contents vary accordingly. For instance, the surface of OMVs contains the same components of the outer membrane, while the surface of MVs contains the same components of the cytoplasmic membrane.

OMVs are lipid-membrane-bounded nanoparticles that are secreted via outer membrane vesiculation of Gram-negative bacteria to contribute to different biological processes [7,13]. Even if OMVs were observed in early reports, they did not receive any attention and their importance had been overlooked until they were found in the spinal fluid of patients infected with meningitis [14]. Since then, the understanding of OMVs’ biogenesis, function, production, and how they contribute to the interaction between the bacterial cell and the host has received increasing attention.

OMVs are composed of several components such as lipids, proteins (e.g., enzymes and structural proteins), carbohydrates, and genetic material (DNA and RNAs), which they originally inherited from the parental cells [3]. In the biomedical field, OMVs play essential roles in attenuating and treating diseases [15,16]. For instance, OMVs can play a significant anti-infection role by inducing and modulating immune responses or by inhibiting pathogen localization and proliferation. Thus, OMVs are recognized as promising candidates for various biomedical applications such as immune modulation, drug delivery, cancer therapy, vaccine development, and anti-bacterial treatments [15,17,18,19,20,21,22,23,24,25,26]; however, their full potential, advantages, future perspectives, and associated problems need to be further investigated.

Several models, briefly summarized below, have been reported to explain the vesiculation mechanism [3]. In all cases, vesiculation allowed the separation of the outer membrane from the below peptidoglycan layer and budding outward until a vesicle can form and separate from the bacterial cell surface. However, one exception was reported, which described the vesicle formation mechanism by the “explosive” cell lysis that is initiated via a prophage endolysin [27].

One of the currently reported models for OMV biogenesis is vesiculation via the VacJ/Yrb ATP-binding cassette ABC transporter, belonging to the family of phospholipid transporters. This mechanism is associated with the accumulation of phospholipids in the outer membrane’s outer leaflet due to the transcriptional silencing or inactivation of the VacJ/Yrb transporter, which is responsible for the maintenance of outer membrane lipid asymmetry [28,29]. This trafficking system is highly conserved in Gram-negative bacteria and was primarily reported to be responsible for phospholipid transportation from the outer membrane to the inner cytoplasmic membrane [30]. The accumulation of phospholipids inside the outer membrane leaflet due to the downregulation of the Vac/Yrb transporter induces the outward curvature which facilitates the formation of the outer membrane budding and OMV release.

The second model of OMV biogenesis involves the insertion of some molecules into the outer membrane outer leaflet (e.g., B-band of lipopolysaccharides (LPS) and Pseudomonas aeruginosa (P. aeruginosa) quinolone signal (PQS)), which can trigger the outward bulging of the outer membrane and promote OMV formation. For instance, the B-band of LPS was proposed to localize to a specific area of the outer membrane, and because of the close proximity of similar charges, it was hypothesized to force the outer membrane to bulge out [31]. Similarly, the insertion of PQS into the outer membrane’s outer leaflet was found to increase the OMV production in Gram-negative bacteria [32]. PQS interacts with lipid A of LPS and sequesters cations such as Ca²⁺ and Mg²⁺. The anionic repulsion that occurs between neighboring LPSs could lead to outer membrane blebbing and OMV formation. Moreover, the interaction of PQS with lipid A decreases LPS fluidity, facilitating the outer leaflet expansion and promoting curvature [33]. The PQS-based model is considered one of the best-investigated models. However, it is species-specific since PQS is only produced by P. aeruginosa.

The third model is based on the presence of specific types of LPS, and/or phospholipids that are enriched in the outer membrane areas where vesiculation occurs [34]. These molecules have the ability to induce vesiculation due to their charges or atypical structure that promote the outward bulging of the outer membrane, and consequently the release of the OMVs into the external milieu. The elevated levels of the negatively charged LPS in certain areas of the outer membrane are an indication of the induction of OMV formation in these areas that could be triggered by specific conditions such as oxidative stress [35,36]. For instance, OMVs isolated from P. aeruginosa were found to primarily consist of negatively charged LPS.

The fourth model associates the accumulation of misfolded proteins, peptidoglycan fragments, and other molecules inside the periplasmic space with the increase in local pressure on the outer membrane responsible for OMV formation [37,38]. It was proposed that vesiculation occurs as a protective mechanism to remove toxic and/or unwanted cellular components. Indeed, vesiculation was increased in Escherichia coli (E. coli) carrying a deletion of DegP, which is a periplasmic chaperone/protease known to correlate with the stress response in the envelope. DegP activity prevents the accumulation of misfolded or damaged proteins inside the periplasm [38]. The role of the periplasmic chaperone/protease in avoiding the accumulation of toxic components was identified in several Gram-negative bacteria, which validates this model for OMV biogenesis  [38,39,40].

The fifth model proposes the disruption of crosslinks between the peptidoglycan layer and lipoproteins as a determining step for OMV formation. The outer membrane of Gram-negative bacteria is well known to be stabilized by the crosslinks between lipoproteins present in the outer membrane and the underlying peptidoglycan layer located in the periplasmic space. However, the lack of these crosslinks in some areas of the outer membrane allows the outer membrane in these regions to curve and form OMVs [41]. Thus, the biogenesis of OMVs is maintained by specific enzymes that are involved in the outer membrane–peptidoglycan layer interactions [42], such as enzymes involved in peptidoglycan synthesis and breakdown (e.g., peptidoglycan endopeptidases) [43].

One of the strengths of using OMVs in biomedicine is the possibility of using them as therapeutic vehicles. Various factors should be considered in developing extracellular vesicles for biomedical applications; e.g., they should be cost-effective, easy to synthesize, biocompatible, non-toxic, feasible to scale up, and with high therapeutic efficacy [44]. Other extracellular vesicles such as exosomes lack important aspects of their potentiality such as the unfeasibility of undertaking large-scale mammalian cultures for vesicle production. OMVs have been suggested as a viable alternative that can be manufactured and produced easily at a large scale and at lower cost [44]. Moreover, OMVs can combine several desirable effects such as delivering the targeted drug (e.g., chemotherapeutic agents) into the specific (e.g., tumor) microenvironment and at the same time recruit immune cells into it, and therefore, enhance their efficacy with no apparent toxicity [45,46,47]. In this regard, naturally acquired OMVs could be used directly or could be modified or genetically engineered to achieve this aim [20,24,25,48,49]. Naturally derived OMVs can serve as DNA, RNA, antigen, and antibody carriers, or as delivery vehicles for their natural cargos. Meanwhile, the modified OMVs can be used in various applications depending on the needed function: they can serve as carriers of specific nanoparticles, DNA and RNA molecules, antigens, antibodies, or drugs. The modifications of OMVs can be applied by loading the desired cargos inside the OMV lumen, by warping nanoparticles inside the OMVs, by concealing OMVs inside nanoparticles, or by embedding the desired components (e.g., antigen, antibody, ligand, etc.) within the outer membrane layer. Cargo loading into OMVs can be performed using different techniques, one of these being electroporation, which involves the use of high-voltage pulses to create pores in the membrane of OMVs, which leads to a temporary permeable state [50,51,52]. This temporary permeability allows the loading of drugs, proteins, nucleotides, small-sized nanoparticles (e.g., metallic gold nanoparticles, AuNPs), etc., which can be achieved using different electric pulses at different durations. After loading the desired molecules, the membrane of OMVs can recover its original structure and lose the temporary permeability without any damage. Similarly, the treatment of OMVs with saponin containing reagents increases their membrane permeability, which facilitates cargo loading without damaging the membrane structure [53,54]. The controlled and temporary disruption of OMV membranes allows cargo loading to also be achieved by applying multiple freeze–thaw cycles in a buffer that contains the material of interest [55,56,57].

The co-extrusion technique is a process of repeating mechanical extrusion using polycarbonate filter membranes that have various pore sizes, which allows the loading of the desired cargo into the OMVs [58]. In this method, OMVs are mixed with the cargo of interest (e.g., drugs, nanoparticles, etc.) and extruded together to force them to interact [59]. Similarly, sonication can be applied as a simpler method for OMV loading. Ultrasonic frequencies can be applied to a mixture of OMVs with the material of interest. This leads to their loading or could result in the attachment of the cargo to the surface of the OMVs due to the temporary disruption of their membrane [60,61]. On the other hand, a simpler alternative technique such as incubating OMVs with the material of interest can be applied. For example, OMVs from Klebsiella pneumoniae (K. pneumoniae) were loaded with doxorubicin hydrochloride (chemotherapeutic drug) by incubating the drug with the OMVs at 37 °C for 4 h [47]. Similarly, the loading process can also be applied by incubating the bacteria of interest with the cargo material during the bacterial growth phase. In this method, the bacteria engulf the material of interest that is present in the medium, pack it into the OMVs, and then release it into the extracellular medium. A study by Huang, et al. [62] successfully used this method to synthesize antibiotic-loaded OMVs from Acinetobacter baumannii (A. baumanii) that resulted in effectively killing certain bacteria in vitro and in vivo [62].

Genetic engineering could be applied to add certain molecules to the surface of OMVs [48,49]. Cargo loading can be applied by the transformation of bacteria using an engineered plasmid that expresses the desired cargo [63,64]. Using this method, the material of interest such as antibodies, antigens, enzymes, and proteins can be loaded into the OMVs [63,64,65,66,67]. Genetic engineering techniques allow the use of different methods to load various types of cargo into OMVs. For instance, recombinant DNA technology enables introducing specific modification into OMVs that can be beneficial for a specific desirable application (e.g., inserting antigens for immune modulation). In addition, genetic engineering can also be applied to knockout genes responsible for a specific undesirable function to be eliminated from the generated OMVs, such as knocking out genes responsible for toxic proteins [68,69].

In short, natural or modified/engineered OMVs could serve as nanopharmaceuticals based on their desired characteristics in a variety of biomedical applications, such as vaccines, adjuvants, cancer immunotherapy, drug delivery, and anti-bacterial adhesion agents  [15,70,71,72,73,74,75,76,77,78].

OMVs are considered excellent vaccine candidates against pathogenic bacteria, and can be used as antigens to induce cellular (cytokines and activated T cells) and humoral (antibody) immune responses after immunization of humans and animals [17,18]. The first vaccine trial of OMVs was in 1991 and was employed against Neisseria meningitidis (N. meningitidis) [79]. Meningitis type B (MenB) is an OMV-based vaccine that is currently approved to treat patients [80,81,82]. Subsequently, efforts continued for new vaccine development against various diseases caused by pathogenic Gram-negative bacteria [49,83,84]. However, there are no other OMV-based vaccines to treat pathogenic Gram-negative bacteria currently available on the market.

OMVs can also be employed as adjuvants to enhance the immune responses against an antigen. This can be achieved by mixing the OMVs with the antigen of interest in the vaccine preparations, linking the antigen to the OMV surfaces, loading the antigen inside the OMVs, or by genetically engineering the bacteria to express the antigen in their outer membrane and consequently into their released OMVs [15]. After immunization, the OMV-containing formulations could trigger robust cellular and humoral immune responses. Contrary to most of the other classic adjuvants that cause systemic and local hypersensitivity, OMVs were found to have low toxicity as well as high potency for inducing T cell responses [85].

The use as agents in cancer immunotherapy to annihilate tumor tissues is another intriguing OMV application. OMVs have been proposed as a good platform for anti-tumor vaccine development for several reasons, such as OMV strong immunogenicity, the ability of OMVs to carry the anti-tumor antigen (inside the vesicle or on its surface), to enhance the antigen presentation, and the lack of OMV proliferation [19,20]. OMV-based anti-tumor vaccines are primarily developed by genetic engineering to express a foreign protein inside the OMVs or linked to the OMV surfaces. This antigen should have the ability to induce the required immune response against cancer cells without causing undesired side effects. OMV-based anti-tumor vaccines can be used to kill cancer cells and/or to silence relevant genes [20]. Various bacterial components such as enzymes, peptides, and toxins have been investigated for cancer therapy [86]. OMVs provide a unique vehicle to combine several anti-tumor components that can initiate an immune response, which is considered a sought-after cancer immunotherapy agent. For instance, OMVs contain parental components (e.g., LPS) that can stimulate an immune response that enables immune cell maturation and tumor damage [45,46]. Moreover, OMVs can function as nanocarriers for loading chemotherapeutic agents. Indeed, the use of doxorubicin passive-loaded OMVs isolated from the attenuated K. pneumonia not only caused a cytotoxic effect and cell apoptosis resulting from the doxorubicin, but it was also observed in vivo that OMVs worked synergistically with their cargo to recruit macrophages into the tumor microenvironment, which enhanced the anti-tumor efficacy with no apparent toxicity [47].

OMVs are also considered an efficient delivery system to transport their cargo to other cells and/or microenvironments. OMV contents can be transported to any targeted cell through two possible mechanisms. The first system proposes the spontaneous lysing of OMVs, which allows their contents to diffuse. The second mechanism is based on the OMV fusion with the targeted cell, on their proximal lysis or internalization [21]. OMV ability for long-distance transportation is one of the main strengths of the OMVs as a vehicle for drug delivery. OMVs can enhance the pharmacodynamics and pharmacokinetics of the loaded drugs by extending the blood circulation time as well as by protecting the loaded molecules from degradation [66]. OMVs have an outstanding targeting capability for bacteria, cells, or inflammatory sites through surface functional protein modifications [24]. Genetic engineering can be applied to express specific targeting ligands onto OMV surfaces [48]. Moreover, pathogen-associated molecular patterns (PAMPs) on the OMVs facilitate their recognition and ingestion by immune cells, which hold great potential for targeted drug delivery against immune cells [87]. OMVs inherit the same surface antigens as the parental cell. Thus, they have the ability to be ingested and recognized by the immune cell, and this can be beneficial in targeted therapy.

The identification of pathogenic bacteria in humans is sometimes difficult for several reasons such as localized infections or aggressive antibiotics concurrent treatments as well as slow-growing bacteria and poor sensitivity of the diagnostic methods [22]. For instance, the analysis of more than 2.5 million sepsis cases using the Premier Healthcare Database in the United States showed that the specific causal organism responsible for sepsis could not be identified in over 70% of the cases [88]. In another study by Stranieri et al. [89], the causal organism of neonatal sepsis was identified only in 41% of the blood cultures from patients [89]. Indeed, bacterial cultures can take up to 24 h to grow and thus they are not compatible with a quick diagnosis and proper specific antimicrobial treatment that is suggested within 1 to 3 h from the recognition of the infection types (e.g., sepsis) [90]. While bacterial cultures fail to provide an accurate and fast method to identify the bacteria that caused the infection, OMVs can persist and permit a more definitive diagnostic approach [91,92]. Due to OMV size, they can widely circulate in the body and freely cross tissue barriers, which allows efficient diagnosis from easily obtained biofluids (e.g., urine or blood) [22]. As stated above, OMVs contain various components of the parental cells. The OMV cargo conserved among bacteria can act as ideal biomarkers for their presence; meanwhile, the OMV cargo specific for a given bacterium can be extremely useful as a rapid differentiation and identification tool for bacterial species identification in OMVs isolated from biofluids [22]. For instance, the widely expressed LPS can serve as a biomarker for Gram-negative bacteria, whereas a species-specific component (e.g., 16S r RNA, urease A (UreA) and heat shock protein (Hsp60) for H. pylori) that is conserved within the targeted bacterial species can be used as a biomarker for the bacteria of interest following its characterization [93,94,95,96]. The presence among pathogenic bacteria of species-specific repeats, both at genomic and protein levels, can be identified using computational methods [97,98,99]. This allows the recognition of highly conserved species-specific hallmarks. Subsequently, they can be used as reliable biomarkers for identifying a specific species present in biological fluids that contained the targeted bacteria or its OMVs.

OMVs also play a vital role in cell–cell communication since signaling molecules can be protected inside the OMV lumen until they reach the target location. For instance, it was reported that 86% of total PQS were packed inside the OMV-derived P. aeruginosa [23]. When these OMVs were removed from the bacterial population, the PQS-controlled group behavior and cell–cell communication were inhibited. Similarly, the hydrophobic quorum-sensing molecule CAI-1 and the hydrophobic signal N-hexadecanoyl-L-homoserine lactone from Vibrio harveyi and Paracoccus sp., respectively, which are responsible for coordinating bacterial group behavior and involved in long-distance communication, were found to be present inside the OMVs [100,101]. OMVs spread far from their parental cell, and therefore, they can be considered as an intra-kingdom communication mechanism that enables the transportation of signaling molecules [102,103]. In addition, OMVs can facilitate trans-kingdom exchange and the delivery of biomolecules between bacteria and their hosts [104].

OMVs are a secretory system, as they have the ability to disseminate bacterial products to their environment, but with unique features. Besides the cargo protection described above, unlike other systems, OMV-mediated secretion can allow the simultaneous secretion of various soluble and insoluble compounds such as membrane proteins, lipids, and insoluble molecules [31,105,106]. Furthermore, the OMV-mediated secreted materials can be delivered and transported at high concentrations, which is often needed for proper efficacy [31,107], and a specifically targeted delivery of molecules can be obtained via selective binding between surface bacterial adhesins and the receiver’s receptors and ligands [31,108]. The selective transportation of OMV cargo to other bacterial cells was observed both in the same or different species. The targeting abilities of OMVs to specific cells hold great potential for the targeted delivery of molecules.

OMVs are mimics of their parental bacteria, and thus they have the ability to inhibit the adhesion of their parental pathogenic bacteria onto the host cell by competitively binding to the same target site [15,24]. Bacterial infections are initiated by bacterial cell adhesion to the targeted cell, and therefore, the anti-adhesion treatment can offer a promising therapy in comparison to other conventional therapies that might induce antibiotic resistance [109]. In addition, the combination of antibiotic and anti-adhesion therapies shows collaborative antibacterial efficacy [110]. OMV-derived H. pylori were reported to block H. pylori adhesion to gastric epithelial cells by developing OMV-coated nanoparticles that preserved the H. pylori surface antigens [24]. Moreover, these developed OMV nanoparticles reduced H. pylori attachment in mouse stomach tissues, which indicates OMV potential in the inhibition of bacterial adhesion to the host tissues. OMV anti-adhesion efficacy can be further improved through genetic engineering to regulate adhesion expression as well as by choosing the ideal nanoparticle cores with desired properties [111,112]. In addition, the use of OMVs as anti-bacterial adhesion tools can promote bacterial clearance, uptake, and recognition by immune cells [15,113]. Overall, OMVs can be considered an efficient anti-adhesion intervention to fight bacterial infections.

In recent years, several reports have investigated the use of OMVs as antibiotic delivery carriers or as active antibacterial agents [15,25,26], showing the potential role of OMVs in antibacterial therapy. Traditional antibiotic therapy has faced many challenges due to the emerging antibiotic resistance in bacteria that has resulted in treatment failure and is becoming a serious threat to human health. The earliest observation of the bactericidal effect of OMVs was reported in 1996, when OMVs from P. aeruginosa were found to contain peptidoglycan hydrolases (autolysins) [21]. Autolysins are intracellular bacteriolytic peptidoglycan hydrolases that are commonly found in bacteria and play major roles in various essential functions such as protein transport, cell division, and peptidoglycan recycling [114]. OMVs have the ability to transport autolysins into other competitor bacteria (Gram-positive and Gram-negative) and negatively impact them by causing disintegration through hydrolyzing their peptidoglycan. Moreover, OMVs from Myxococcus xanthus (M. xanthus) were reported to have many types of enzymes such as phosphatases, hydrolases, and proteases with bactericidal activity against E. coli [115]. The antibacterial activity of these OMVs against E. coli was improved when the fusogenic enzyme (glyceraldehyde-3-phosphate dehydrogenase) was present, which facilitated the fusion-based interaction between M. xanthus OMVs with the targeted cells. Similarly, OMVs from Lysobacter capsici (L. capsici) were found to have a bactericidal effect due to their bacteriolytic enzymes [116] and OMVs from P. aeruginosa were found to have significant antimicrobial activity against Staphylococcus epidermidis (S. epidermidis) due to the presence of quinolines within the OMVs [23]. In this regard, the interbacterial antagonism between at least two bacteria can be explored in order to use their OMVs as antibacterial agents.

In addition to their natural antibacterial activity, OMVs can be also used as antibiotic delivery carriers due to their efficient targeting capacity, drug loading, cargo protection, prolonged circulation time, etc. [107,117]. It is reported that the targeting capacity, pharmacokinetics properties, and chemical stability of antibiotics can be enhanced when loaded inside the OMVs [25]. Gentamicin-loaded OMVs were found to have a strong bactericidal effect against the gentamicin-resistant P. aeruginosa [21]. Similarly, the gentamicin-loaded OMVs from Buttiauxella agrestis (B. agrestis) exhibited a strong bactericidal effect against their parental cells as well as against E. coli and P. aeruginosa [118]. Interestingly, the bactericidal effect of gentamicin-loaded OMVs was stronger against B. agrestis (parental cell) and other bacterial species of Buttiauxella spp. than those of E. coli and P. aeruginosa, suggesting a bacterial species specificity. Despite the multiple advantages that OMVs can offer as antibiotic delivery vesicles, only a few reports have been published so far, indicating the need for further investigation.