Outer Membrane Vesicle Vaccine Platforms

Many Gram-negative bacteria spontaneously release outer membrane vesicles (OMVs) during growth [1], very likely as a result of an imbalance between cell growth and outer membrane biosynthesis, with the excess membrane material released as OMVs [2]. A number of studies have reported the role of OMVs in enhancing the survival of bacteria in harsh environments and delivering virulence factors and DNA to host cells [3‐5]. OMVs appear involved in bacterial pathogenesis [6], help biofilm formation, and increase survival in hosts [7] or in soil [8]. Furthermore, OMVs support bacterial defense against antibiotics [9] and contribute to transfer of antibiotic resistance between bacteria [10].

OMVs mimic the outside of bacteria, resembling a pathogen without the ability to cause disease, and containing multiple surface-exposed antigens. In addition, OMVs hold pathogen-associated molecular patterns (PAMPs), such as lipopolysaccharides (LPS), peptidoglycans, lipoproteins, and flagella, that confer self-adjuvanticity [11]. Furthermore, the particle size of OMVs facilitates uptake by antigen-presenting cells (APCs), aiding presentation to cognate T cells, and by follicular dendritic cells (FDCs) that activate antigen-specific B cells, with the induction of adaptive immune response. For these reasons, over the last years, OMVs have been regarded as a versatile platform for vaccine development [12, 13].

OMVs naturally released from bacteria are designated as native OMVs (nOMVs), but for many bacteria this occurs at levels too low for vaccine manufacturing. To increase yields, OMVs have been chemically extracted from whole bacteria using detergents (e.g. deoxycholate), resulting in vesicle-like aggregates of insoluble outer membrane proteins called detergent-extracted OMVs (dOMVs). The supernatant of the fermentation containing spontaneous vesicles is usually discarded and extraction of the OMV is then performed. A sonication step can be introduced to obtain dOMVs of smaller and more homogeneous size, also to simplify sterile filtration [12]. The use of detergents largely reduces LPS and lipoprotein content, consequently reducing reactogenicity and improving tolerability of the OMV [2, 14, 15]. However, this approach also results in loss of important protective lipoprotein antigens and compromises vesicle integrity, with contamination of the resulting preparation with cytoplasmic proteins [16, 17].

Bacteria can be genetically manipulated to alter OMV properties (e.g., yield, endotoxicity, protein content) and classified as mutant-derived OMVs (mdOMVs) or generalized modules for membrane antigens (GMMAs). Mutations are often introduced to destabilize the linkage of the outer membrane with the inner membrane and the peptidoglycan layer in the periplasm. At this scope, different strategies have been pursued. A common one is deletion of the tolR gene of the Tol-Pal system present in most Gram-negative bacteria [18‐20]. Deletion of nlpI, with changes in peptidoglycan dynamics, has been used in Escherichia coli [21]; disruption of the VacJ/Yrb ABC (ATP-binding cassette) transport system, resulting in accumulation of phospholipids in the outer leaflet of the outer membrane, has been used in Haemophilus influenzae, Vibrio cholerae, and E. coli [22, 23]. Deletion of lpp [24], mltA (gna33) [16], rmpM [25], ompT [26], pagL [27], enterobacterial common antigen [22, 28], degP [1], and virk [29] also results in overblebbing of the bacterial outer membrane.

Because the LPS content is high in naturally released OMVs, additional mutations are introduced to reduce endotoxicity. This is usually achieved through modification of the lipid A structure, in particular reducing the number of acyl chains and phosphate groups of lipid A, impacting its ability to recognize/trigger toll-like receptor (TLR) 4 [30‐32] and decreasing the inflammatory response associated with lipid A [11]. Deletion of acyltransferases such as MsbB (lpxM), htrB (lpxL) and pagP (adding a myristoyl, lauroyl or palmitoyl chain, respectively, to the lipid A) has been used in E. coli, Shigella and Salmonella [33‐35]; deletion of lpxL1 and lpxL2 has been used in Neisseria meningitidis [36‐38].

Compared with dOMVs, the purification of nOMVs and GMMAs is simpler and requires less steps [12]. OMVs can be separated from whole bacteria by a first tangential flow filtration (TFF), and a second TFF retains OMVs while removing soluble proteins and other low molecular weight impurities [39, 40].

In this review, we provide an overview of OMVs and their genetically modified version, GMMAs, as a vaccine platform to present protein or glycan antigens, key achievements at clinical level, trends emerging in preclinical studies, and future perspectives of the field.

A few OMV-based vaccines are currently licensed and in use. These are vaccines against meningitidis, and target N. meningitidis serogroup B (MenB) and H. influenzae type b (Hib) infections (Table 1).

Formulated with three recombinant proteins in Bexsero (Bexsero is a trademark owned by or licensed to the GSK group of companies), the OMV detergent extracted from N. meningitidis (from strain NZ98/254) is indicated for prevention of meningitis and meningococcemia caused by N. meningitidis serogroup B. The vaccine works by stimulating the production of bactericidal antibodies that recognize the vaccine antigens (NHBA, NadA, fHbp) and the PorA, as a major component of the OMV. A review of the Meningococcal Antigen Typing System (MATS) by Muzzi et al. suggests that the vaccine covers between 81 and 84% of MenB isolates [41]. It was tested in several clinical trials showing a good immune response, as measured by human complement serum bactericidal activity (hSBA). In a study in infants [42], when the OMV component was increased while keeping the recombinant proteins the same, there was a corresponding augmentation of the immune response measured as proportions of infants with hSBA ≥ 5 and geometric mean titers (GMTs), which increased in a dose-dependent manner, lowest in infants receiving 6.25 μg and highest in those receiving the full dose of OMV in Bexsero (i.e. 25 μg). However, there was no similar trend for systemic reactogenicity, with a similar reactogenicity profile observed for recipients of one-quarter, half, and full doses of OMV. For local reactions, this was much higher among recipients of 4-component MenB vaccine containing OMVs compared with individuals who received a vaccine without OMV. Furthermore, the vaccine formulated with OMV, when compared with the recombinant proteins alone, gave a much higher frequency of fever. The vaccine was licensed in 2013 and is indicated for active immunization of individuals from 2 months of age against invasive meningococcal disease (IMD) caused by MenB [43]. The safety of the vaccine has been evaluated in 17 studies, including 10 randomized controlled clinical trials with infants, children, adolescents and adults, with a higher frequency of fever when coadministered with routine vaccines (69–79% of subjects), compared with those receiving routine vaccinations alone (44–59%), although most events were of mild to moderate severity and short-lived (lasting about 1 day) [43]. Although the vaccine was licensed using immunogenicity endpoints, within 3 years of introducing Bexsero in the UK there was a 75% reduction of MenB IMD cases in vaccine-eligible infants [44].

The same OMVs were formulated alone in the New Zealand MenB vaccine, used to fight an outbreak of meningitidis due to N. meningitidis serogroup B in that country. The OMV vaccine was tested in several clinical trials where it showed an acceptable safety profile and a strong immune response after three doses of the vaccine. The vaccination program was staggered throughout the different regions of New Zealand, delivered to different age groups at a time between 2004 and 2006, and used for infant routine vaccination until 2008 [45]. The vaccine was estimated to be 77% effective and prevented an estimated 210 cases of MenB IMD with its attendant sequelae between July 2004 and December 2008 [46]. Similarly, in Norway, a dOMV vaccine developed in 1983 against MenB was used to control outbreaks of MenB disease. The vaccine was administered to 171,800 adolescents from 1988 to 1991, with an estimated efficacy of 87% after 10 months and with a rapid decline thereafter, aligned with a reduction in serum bacterial activity [47]. The vaccine was also used between 2006 and 2009 to control an outbreak in Normandy, France, due to the same B14:P1.7,16 strain, with a decrease in the incidence of cases from 31.6 to 5.9 per 100,000 between 2006 and 2009 [48]. Another dOMV-based vaccine (VA-MENGOC-BC, Finlay Institute, Havana, Cuba) was licensed for use in Cuba against MenB in 1987 (VA-MENGOC-BC is a registered trademark of Finlay Institute of Vaccines, Cuba) [49], succeeding in reducing the incidence of MenB disease by 93–98% in the following 20 years, eventually resulting in MenB no longer being a public health problem in Cuba [50]. In 2014, the vaccine was acquired by Abivax for distribution outside Cuba in countries in Asia and Latin America.

Hib-OMPC was the first OMV-based vaccine to be licensed widely. It is a highly purified capsular polysaccharide of H. influenzae type b (polyribosylribitol phosphate [PRP]) covalently bound to outer membrane protein complex (OMPC), obtained by detergent extraction of OMVs from the B11 strain of N. meningitidis serogroup B. For licensure, PRP-OMPC was shown to be 93–100% efficacious in a pivotal trial that enrolled 3486 Navajo infants (a population at particularly high risk of disease) vaccinated at 2 and 4 months of age [51]. Licensed as PedvaxHib (PedvaxHIB is a licensed trademark of Merck Sharp and Dohme Corp.), the vaccine induced anti-PRP levels > 0.15 μg/mL in 88% and > 1.0 μg/mL in 52% of the infants, with a GMT of 0.95 μg/mL 1–3 months after the first dose; rising to 91% and 60%, respectively, after the second and third doses [51].

PRP-OMPC stimulates a strong immune response after one dose in children with a high risk of Hib infection in early infancy [52]. However, the American Advisory Committee on Immunization Practices (ACIP) recommends that previously unvaccinated infants aged 2–6 months should receive two doses of PRP-OMPC administered at least 2 months apart, to ensure longevity of protection. They also recommend one or two doses in unvaccinated children depending on their age [53].

Eventually, PedvaxHib was combined with hepatitis B surface antigen (HbsAg) and licensed as Comvax (Procomvax in the EU). The combination resulted in anti-PRP antibody levels lower than the previously reported levels in the stand-alone vaccine (Procomvax–EU/Comvax–US) [54]. Procomvax and Comvax are registered trademarks of Merck and Co./MSD in Europe.

The immunogenicity of Comvax (7.5 μg H. influenzae type b PRP, 5 μg HBsAg) was assessed in 1602 infants and children 6 weeks–15 months of age in five clinical studies. In these studies, the immune response of Comvax was compared with that obtained using the monovalent vaccines, PedvaxHib (7.5 μg PRP) and RECOMBIVAX HB (5 μg HBsAg) administered at separate sites, either concurrently or 1 month apart. In infants not previously vaccinated with Hib or Hepatitis B vaccine, antibody responses to Comvax showed proportions achieving anti-PRP levels > 1.0 μg/mL after the second dose, similar to those of subjects receiving the monovalent vaccine (72.4%) [55]. The vaccine is indicated for vaccination against invasive Hib disease and HBV infection in infants born from HbsAg-negative women with three doses at 2, 4, and 12–15 months (MMR weekly, CDC). Procomvax was taken off the European market by the manufacturer in 2009, who opted against renewing the market authorization for commercial reasons. There was no safety concern linked to that decision.

The monovalent vaccine was also combined with diphtheria and tetanus toxoids, Bordetella pertussis antigens, HbsAg and inactivated polio virus as a hexavalent pediatric vaccine (VAXELIS). VAXELIS is a registered trademark of MSP Vaccine Company. This combination is indicated for primary vaccination and booster vaccination in infants and toddlers from 6 weeks of age, with a primary vaccination schedule of two to three doses, at an interval of at least 1 month between the doses. The safety profile of the vaccine is comparable with other multivalent vaccines for the same indication; however, a coadministration booster study with pneumococcal vaccine (Prevnar 13) revealed a frequency of fever of up to 52.5% among recipients of both vaccines, although the fevers were of mild to moderate intensity and short-lived (< 48 h) [56]. Overall, the safety of Hib PRP-OMPC vaccines is comparable with those of other Hib vaccines, with non-clinically significant differences between them [57].

Considering the advantages of OMVs in presenting multiple antigens in their natural configuration with self-adjuvanticity, there are a number of OMV-based vaccines in clinical development to fight pathogens from which they derive (Table 1).

GMMAs naturally display O-polysaccharides on their surface (Fig. 1) and have been exploited as vaccine candidates to prevent non-typhoidal Salmonella and Shigella, now in clinical trials [39, 58‐60].

In mice, O-polysaccharides displayed on Salmonella typhimurium and Salmonella enteritidis GMMAs have been shown to elicit a strong anti-O-polysaccharide immunoglobulin (Ig) G antibody response, at levels comparable with those induced by the corresponding CRM₁₉₇ conjugates formulated with alum [61]. Interestingly, GMMAs enhanced the IgG antibody isotype profile and resulted in greater serum bactericidal activity as compared with protein conjugates. The bivalent formulation of S. typhimurium and S. enteritidis GMMAs adsorbed on Alhydrogel is now being tested in a phase I trial in European adults (NCT05480800). This formulation has also been combined with a glycoconjugate vaccine against S. typhi (Typhibev), for which a phase I trial in healthy adults has also been started (NCT 05480800).

The first GMMA-based vaccine to be tested in clinical trials was the Shigella sonnei GMMA vaccine 1790 GAHB. GMMAs were purified from an S. sonnei strain genetically engineered to increase blebbing (ΔtolR), produce less reactogenic penta-acylated lipid A (ΔhtrB), and stably express the virulence plasmid encoding for the O-polysaccharide [39]. S. sonnei GMMAs, adsorbed on Alhydrogel, were tested in two phase I trials, one extension booster trial, one phase IIa clinical trial, and eventually in a phase IIb human challenge infection model (CHIM) trial [60]. O-Polysaccharide-based S. sonnei GMMAs have been demonstrated to be well tolerated and immunogenic in healthy adults in European and disease-endemic countries, eliciting bactericidal anti-O-polysaccharide IgG response after one dose of vaccination [62‐65]. Noteworthy, revaccination 3 years apart from the primary immunization showed the ability of the vaccine to induce a robust anamnestic memory response [66].

1790GAHB failed to show clinical efficacy in the CHIM trial, likely due to inadequate immune response linked to the low O-polysaccharide dose tested. However, based on those results, a new generation S. sonnei construct has been developed, allowing to increase 10 times the O-polysaccharide dose. Such GMMAs have been combined in a 4-component formulation with Shigella flexneri 1b, 2a, and 3a GMMAs that is currently being tested in a phase I/II trial (NCT05073003) [60].

There is also epidemiologic evidence of a moderate level of effectiveness of N. meningitidis B OMV vaccines against Neisseria gonorrhoeae [67‐69], leading to suggestions that an OMV-based vaccine could be suitable for prevention of gonococcal infection, both through a vaccine against meningococcal B (NCT04297436 and NCT04415424) and with GMMAs specific for N. gonorrhoeae, with clinical trials underway (NCT05630859).

Vaccines with heterologous targets, in which the OMVs are conjugated to other antigens, are also in early clinical development, such as the intranasally delivered vaccine against coronavirus disease 2019 (COVID-19; Avacc 10) developed by Intravacc BV (NCT05604690).

Traditionally, OMVs have been proposed to fight diseases caused by pathogens from which they derive [12], with some of them licensed or in clinical development and many others at the preclinical stage [2, 12]. The immune response can be directed to the LPS rather than protein components, as shown in early studies with naturally released V. cholerae OMVs, which elicited an antibody response by mucosal immunization protecting mice offspring from oral V. cholerae challenge [70] and ensured cross protection against the two variants Inaba or Ogawa of the most epidemiologically relevant serotype O1. OMVs failed to cross-react with O139 [71].

Recent studies have shown how genetic manipulations can be easily performed with the aim to further improve OMV safety, immunogenicity, and protective efficiency. For example, it has been shown that disruption of small noncoding RNA improves the protective efficacy of OMVs against Helicobacter pylori infection in a mouse model [72]. Both intranasal and intraperitoneal immunization with flagellin-deficient S. typhimurium OMVs resulted in efficient protection against heterologous S. choleraesuis and S. enteritidis challenge [73] and immunization with OMVs from major outer membrane protein-deficient S. typhimurium mutants enhanced cross-protection [74].

Recently, OMVs have also been proposed as delivery systems for heterologous antigens. Indeed, OMVs can be modified to display either proteins or polysaccharides (Fig. 1) derived from different (even phylogenetically distant) pathogens (viral, bacterial, parasitic).

OMVs have a long and rich history of use, particularly in vaccines against N. meningitidis and H. influenzae type b. A large body of preclinical data has been generated to fight many other pathogens, supporting the use of OMVs and their genetically modified version (GMMAs) as a promising affordable platform to deliver both homologous and heterologous protein and carbohydrate antigens.

The nanosize of OMVs is thought to enhance uptake by APCs, whose activation occurs through recognition of multiple PAMPs on OMVs by TLRs. B-cell activation is facilitated by the repetition of epitopes on the OMV surface and by the size of OMVs, which allows them direct access to the lymphatic system. T-cell/B-cell cooperation is essential for the generation of high affinity antibody-producing plasma cells and memory B cells. More studies on the OMV mode of action will allow to better understand how immunity to this class of vaccines is induced, and will help elucidating whether the mechanisms identified are platform-related or pathogen-specific.

The COVID-19 pandemic, as well as the emergence of antimicrobial resistance, are catalyzing the progression of novel vesicle-based candidate vaccines into clinical development (Table 1). Homologous meningococcal OMVs are included in licensed vaccines, and GMMAs to fight homologous pathogens from which they are derived are being tested in clinical studies. Data from current trials, both in terms of safety and immunogenicity, will be key to further support the use of this platform for novel vaccine development.

Current advancements in the field focus on OMVs and GMMAs as carriers for heterologous protein and carbohydrate antigens (Fig. 1). Vaccines against COVID-19 developed using this technology (Avacc 10) are undergoing clinical testing. Generation of this type of multicomponent vesicles has been shown to be feasible by different new technologies, including chemical conjugation, molecular engineering and Spy-tag/Spy-capture. These new approaches will favor simplification of vaccines targeting a variety of different pathogenic mechanisms or even different pathogens. We expect that in the future, this class of vaccines will advance significantly, and more vesicle-based vaccines will become available to help fight emerging diseases and unmet medical needs.