LNP-mRNA vaccines have emerged as the primary and highly effective approach for immunization against SARS-CoV-2 and possibly other infectious diseases as well. Nonetheless, despite their success, this technology does have intrinsic limitations that have posed challenges to its overall effectiveness. In this context, EV-protein vaccines offer a promising solution to overcome these limitations and enhance the field of vaccination.

One critical limitation of LNP-mRNA vaccines is the inherent toxicity of the LNPs used to deliver mRNA vaccine [29, 30]. In fact, the adverse side effects observed in LNP-mRNA vaccines, such as the Pfizer/BioNTech and Moderna vaccines, have been attributed to the inflammatory response induced by LNPs [31]. As such there are intense research studies to overcome this toxicity and improve safety and efficacy of LNPs for RNA-based therapeutics. For example, some of these efforts involve the synthesis of novel lipid chemical structures and the incorporation of different types of helper lipids and lipopolymers into nanoparticle formulations [32]. However, it remains to be seen if the toxicity of LNPs can be effectively improved.

In contrast, EVs offer inherent safety advantages compared to LNPs. Unlike LNPs, which have been documented in scientific literature to have toxic properties, EVs have not been reported to exhibit significant toxicity. This can be attributed to the fact that EVs are natural components present in all bodily fluids. A depiction of the general structure of EVs and LNPs is illustrated in Fig. 1. The safety of EV-rich tissues, such as blood and serum, has been extensively investigated in routine medical procedures like transfusion, establishing a solid foundation for the overall safety of EV treatments. Furthermore, clinical trials examining the use of exosomes or EVs for interventional studies further support their safety profile. Of the 21 clinical trials conducted as of 12th April 2023 (source: https://clinicaltrials.gov/), four have already been completed without any adverse events, emphasizing the favorable safety record of EVs as therapeutics or delivery vehicles.

The general structure of an EV and LNP. EV possesses a membrane structure made of a cell homologous lipid bilayer and carry a variety of biologically active substances such as proteins, nucleic acids, glycoproteins, metabolites. Some proteins represent EV markers (e.g., tetraspanins CD9, CD63, CD81), while other proteins are variable depending on the cell type origin, including adhesion molecules (ICAM and integrins), major histocompatibility (MHC) molecules, enzymes, and other factors. LNP is currently recognized as promising candidates for transferring vaccine mRNA, owing to their stability and biocompatibility. Typically, LNPs consist of four key components: cationic lipids, ionizable lipids, polyethylene glycols (PEGs), and cholesterol. The LNPs could be either mono- or bi-layer. These components collaborate synergistically to enhance the effective delivery of mRNA into the cytoplasm. EV extracellular vesicle, LNP lipid nanoparticle

In summary, the inherent toxicity associated with LNPs poses a critical limitation in their use as delivery vehicles, necessitating ongoing intensive research to enhance their safety and efficacy. In contrast, EVs exhibit remarkable safety advantages over LNPs, supported by the absence of documented toxicity and extensive research on the safety of EV-rich tissues. These characteristics make EVs highly promising for vaccine development, especially in scenarios requiring long-term or repeated use.

Scalable manufacture of mRNA vaccines presents challenges due to its non-continuous manufacturing process. Instead of a single streamlined process, it involves multiple discrete steps, including the production and purification of linearized DNA templates, transcription of the DNA to generate mRNA, capping of the mRNA, purification of the mRNA, encapsulation of the mRNA in LNPs, and finally fill and finish [33]. Moreover, the materials required for each of these steps are expensive and limited in supply, which further increases the cost and complexity of scaling up production [33].

On the other hand, the production of EV-protein vaccines can be achieved using two primary strategies. One approach involves obtaining EVs from cells genetically modified to express protein vaccines in EVs, while the other involves engineering isolated EVs to express protein vaccines. The former approach typically employs a large-scale, continuous manufacturing process, which begins with cell expansion, followed by the harvesting of the conditioned medium, EV enrichment, and fill and finish [34]. The latter method requires an additional step of engineering the native EVs produced by unmodified cells. Generally, manufacturing EV vaccines utilizing genetically modified cells offers clear advantages over the alternative strategy.

It is worth noting that many parallels can be drawn between the well-established manufacturing of cells for cell therapy or biologics, such as MSCs or antibodies, and the production of EVs from cells. For instance, the manufacturing process for MSC-EVs shares many similarities with that of MSCs themselves [35]. Since EVs are non-living entities, there is no need to implement the expensive monitoring and mitigation processes required for MSC products to ensure a viable living state. This factor can offset any additional manufacturing costs associated with MSC-EV production.

Independent of the processes used in the manufacturing of either EV- or LNP-based vaccines will have to adhere strictly to regulatory guidelines (http://www.fda.gov/downloads/BiologicsBloodVaccines/GuidanceComplianceRegulatoryInformation/Guidances/Vaccines/ucm092272.pdf). Specifically, the manufacturing processes, product characteristics, and product testing which are collectively known as CMC must be defined in order to to ensure that vaccines are safe, effective and consistent between batches [36]. The Table 2 provides an overview of the differences in key CMC of EV- vs. LNP-based vaccines.

Overall, although the manufacturing of EV-protein vaccines utilizing living cells might appear intimidating, it is not inherently more challenging than the production of LNP-mRNA vaccines. The scalability and continuous manufacturing precedents established in the field of cell therapy and biologics can be easily applied to EV and EV vaccine production. This has the potential to make EV production as, if not more cost competitive as the discontinuous manufacturing of mRNA vaccines.

Unlike most traditional vaccines that can be stored at 2–8 °C (https://ldh.la.gov/assets/oph/Center-PHCH/Center-PH/immunizations/vaccine-storage-handling.pdf), the current LNP-mRNA COVID-19 vaccines developed by Moderna and BioNTech/Pfizer require ultra-cold temperatures for storage. Moderna’s vaccine needs to be kept between − 15 and − 25 °C(https://www.cdc.gov/vaccines/covid-19/info-by-product/moderna/downloads/storage-summary.pdf), while BioNTech/Pfizer’s vaccine requires storage temperatures between − 60 and − 90 °C (https://www.cdc.gov/vaccines/covid-19/info-by-product/pfizer/downloads/storage-summary.pdf). This presents significant challenges for distribution due to limited infrastructure for ultra-low temperature storage and transportation.

The reasons for the differing storage requirements of Moderna and BioNTech/Pfizer COVID-19 vaccines are not fully understood. Although significant improvements have been achieved in enhancing the stability and efficacy of mRNA-LNP vaccines, a notable gap still exists in our understanding of their long-term storage stability. To address this issue, it is imperative to adopt a systematic approach that enables the identification and understanding of the physicochemical degradation mechanisms affecting LNP-mRNA stability [37, 38]. This will help in the design of stable LNP-mRNA formulations that can be stored, transported and administered at refrigerated or ambient temperatures.

Recent studies have shown promising results regarding the stability of mRNA-LNP formulations. Some formulations have demonstrated the ability to be lyophilized and stored at room temperature or even up to 37 °C for several weeks [39, 40]. However, further validation is required to ensure reproducibility and reliability of these findings.

In the case of EVs, which are also generally stored at − 20 °C, lyophilization has become a common practice. Lyophilized EVs have been reported to remain stable at higher temperatures, such as 25 °C for four weeks [41] and 40 °C for three weeks [42]. Lyophilized EVs have also been utilized for bio-functional testing in pre-clinical animal models [43, 44].

Although both LNP-mRNAs and EVs are typically unstable and require ultra-low temperature storage, there is increasing evidence that they can be lyophilized and stored at higher temperatures in their lyophilized states. This advancement has the potential to make LNP-mRNA or EV-protein vaccines more accessible by eliminating the need for ultra-cold storage and transportation.

Both EVs and LNPs face a critical challenge as delivery vehicles for vaccines, which is their limited ability to escape the endosome-lysosome pathway after being internalized by cells. Recent research suggests that only a small fraction of LNPs (approximately 1–2%) can successfully escape from the endosomes and avoid degradation in the lysosomes [45, 46]. The principal challenge in RNA drug development lies in achieving efficient cytosolic delivery, hindered by the inefficiency of traversing both plasma and endosomal membranes to reach the cytosol [47]. In the context of LNP-mRNAs, once the LNPs successfully escape, they liberate their mRNA cargo within the cytoplasm, facilitating the translation of RNA into proteins. These proteins can then undergo proteolysis via the ubiquitin proteasome pathway to facilitate antigen presentation on MHC-class I, which is necessary for generating an immune response (Fig. 2A). Alternatively, the proteins may either be displayed on the cell surface membrane or released into the extracellular space [48, 49], where they can be endocytosed by antigen-presenting cells (APCs) for presentation on MHC-class II [48, 49] (Fig. 2B).

Although EVs also exhibit inefficient endosomal escape when internalized by cells [50, 51], this is less detrimental in terms of immune system recognition of EV-protein vaccines. EV-protein vaccines, unlike LNP-mRNA vaccines, do not require translation once they have escaped from the endosomes. Instead, they can undergo immediate proteolysis by the ubiquitin-proteasome system and subsequent loading onto MHC-class I for antigen presentation [48, 49]. (Fig. 2C). They also do not require translation to be presented on the cell membrane or secreted into the extracellular space for recognition by APCs. Furthermore, the poor endosomal escape of endocytosed EV-protein vaccines in APCs would likely promote lysosomal processing of the protein vaccines, facilitating antigen presentation on MHC-class II [48, 49] (Fig. 2D). Therefore, EV-based protein vaccines possess a significant advantage over LNP-mRNA vaccines in terms of endosomal escape and subsequent antigen presentation.

Endocytosis of LNP-mRNAs / EV-proteins for immune interaction. A Most of the internalized LNP-mRNAs will be shuttled to the lysosome and a small fraction will escape from the endosomes to release mRNA for protein translation. Some of the newly translated proteins could undergo degradation by the proteasome for MHC class I antigen presentation to CD8+ T cells Others could be displayed on the cell surface or secreted into the extracellular space. B The membrane-bound or secreted proteins when endocytosed by APCs will be processed in the lysosomes for MHC class II antigen presentation to CD4+ T cells. C Most of the internalized EV-proteins will be shuttled to the lysosome and a small fraction will escape from the endosomes to release protein for degradation by the proteasome and MHC class I antigen presentation to CD8+ T cells. D When EV-proteins are endocytosed by APCs, the poor endosomal escape enhances lysosomal processing of the proteins and facilitates MHC class II antigen presentation to CD4+ T cells. ER endoplasmic reticulum, TCR T cell receptor, APC antigen-presenting cell

Another advantage of EVs as protein vaccine delivery systems over LNP-mRNA vaccines is the potential for wider dissemination of the protein from the site of administration to immune inductive sites [28]. In contrast, the protein antigen produced by the translation of the mRNA vaccine is retained at the cellular site of mRNA translation thereby reducing the accessibility of the translated proteins to immune cells.

Overall, EV-based protein vaccines are promising alternatives to LNP-mRNA vaccines, particularly in terms of endosomal escape and antigen presentation. In LNP-mRNA vaccines, the inability to escape from the endosome can result in the degradation of the vaccine in the lysosomes, thereby limiting its efficacy. In contrast, in EV-protein vaccines, even if they fail to escape the endosome in APCs, the proteins can still be processed in the lysosomes and presented by APCs to activate the immune system. This makes EV-based protein vaccines a more reliable and efficient option for inducing an immune response compared to LNP-mRNA vaccines.

The COVID-19 pandemic has significantly accelerated the progress of vaccine development, evaluation, manufacturing, and deployment across various platforms. Notably, mRNA-based vaccines, viral-vectored vaccines, inactivated virus-based vaccines, and subunit recombinant proteins have emerged as prominent approaches. According to a comprehensive review by Mouro and Fischer [52], over 10.5 billion doses of COVID-19 vaccines have been administered worldwide in just over a year, highlighting the scale of the vaccination efforts. By March 2022, ten vaccines had received emergency or full use approval from WHO-recognized regulatory authorities, with additional authorizations granted in specific countries. Furthermore, there are currently 346 COVID-19 vaccine candidates in development, with 151 undergoing clinical trials.

The review underscores the efficacy of most vaccines in preventing symptomatic infection, reducing the severity of illness, and decreasing mortality rates in both clinical trials and real-world settings. However, it also acknowledges that these vaccines have not proven effective in suppressing community transmission of the virus. Consequently, non-pharmaceutical protective measures like mask wearing, social distancing, and border closures have remained necessary. Unfortunately, these measures have imposed a significant socioeconomic cost, disproportionately impacting vulnerable populations.

Retrospective analyses have indicated that the current generation of LNP-mRNA COVID-19 vaccines has primarily been unsuccessful in suppressing community transmission due to their limited ability to induce mucosal immunity, despite their success in stimulating systemic immunity [52,53,54,55]. Since viruses such as SARS-CoV-2, avian influenza, SARS, MERS, and Nipah primarily target the upper respiratory tract and establish colonization on mucosal surfaces, effective prevention of such viral infections requires measures that impede initial mucosal colonization. In this regard, bolstering mucosal frontline immunity becomes crucial as it plays a critical role in preventing mucosal colonization by these viruses. By effectively impeding mucosal colonization, the subsequent dissemination of the virus into systemic tissues can be curtailed. This notion finds support in a recent report demonstrating that mice, even after receiving two intramuscular applications of an adenovirus-based mRNA vaccine, still require additional intranasal boosts with the same vaccine to stimulate high levels of mucosal IgA and lung-resident memory T cells and achieved complete protection against SARS-CoV-2 infection [56]. However, intranasal immunization with these adenoviral vectors did not elicit the expected pre-clinical response in a human clinical trial [57].

The current LNP-mRNA COVID-19 vaccines are administered intramuscularly, and they induce robust systemic immunity but weak mucosal immunity. To date, there have been no reports on intranasal administrations of LNP-mRNA vaccines to induce a mucosal immune response. In contrast to LNP-mRNA vaccines, intramuscular injection of EV membrane bound GFP enhances both IgA and IgG production, indicating that intramuscular injection of EV-protein vaccines can stimulate both mucosal and systemic immune responses [28]. Moreover, EV membrane decorated with a recombinant SARS-CoV-2 receptor-binding domain can be administered through inhalation and provide protection against live SARS-CoV-2 challenge in animals [44].

These findings underscore the significance of mucosal immunity in preventing mucosal infection and reducing the risk of severe systemic diseases. Vaccines based on EV-protein, which are more amenable to mucosal vaccination, such as intranasal administration, may hold promise for inducing effective mucosal immunity.

One of the advantages of LNP-mRNA vaccines is the rapid response time in developing a vaccine. In principle, the mRNA can be rapidly designed in silico and then chemically or biochemically synthesized to produce the LNP-mRNA as soon as the pathogen genome is known as well as the viral antigen candidate is identified. By the same token, the DNA sequence of viral antigen candidates can be readily inserted into EV engineering platforms as such those described above [22, 23] to generate EVs carrying the desired protein antigens.