Alphavirus expression systems have been engineered to allow the utilization of recombinant viral particles, RNA replicons, and DNA replicons for the Semliki Forest virus (SFV) [30], Sindbis virus (SIN) [31], and Venezuelan equine encephalitis virus (VEE) (Figure 2) [32]. In the case of recombinant viral particles, both replication-deficient and replication-proficient systems have been developed. Briefly, the generation of replication-deficient particles takes place from co-electroporation or co-transfection of an alphavirus vector and a helper vector. The expression vector carries the nsP1–4 genes responsible for RC formation for massive viral RNA self-replication. The gene of interest (GoI) is introduced downstream of the highly efficient SFV subgenomic promoter (26S) located at the 3’ end of the nsP4 gene. As the alphavirus structural genes are deleted from the expression vector, they are provided in trans from the helper vector. RNA molecules from both expression and helper vectors are replicated in transfected host cells, but as the packaging signal is located in the nsP2 gene for SFV [33] or in the nsP1 gene for SIN and VEE [34], only RNA from the expression vector will be packaged into recombinant viral particles. These so-called suicide particles show the same broad host range as wild-type alphaviruses, but due to the absence of any viral structural genes, no further viral progeny will be produced. However, alphavirus suicide particles generate extreme levels of transgene expression. As no virus progeny is produced and the viral ssRNA is quickly degraded, active antigen production lasts for approximately 5–7 days in vivo, which is actually ideal for immunization with no risk of long-term extrachromosomal presence or chromosomal integration. In the case of replication-proficient alphavirus particles, transgene expression can be achieved by the introduction of the GoI either downstream of the non-structural genes or alternatively after the structural genes driven by a second 26S. Transfection of in vitro transcribed full-length alphavirus RNA and the GoI into mammalian host cells will generate fully replication-proficient particles. Importantly, as the generated alphavirus particles are both infectious and capable of generating viral progeny, the biosafety requirements are tighter compared to replication-deficient particles.

In the context of RNA-based delivery, in vitro transcribed RNA from the alphavirus expression vector containing the GoI can be directly administered to cell cultures or in vivo. However, the sensitivity of ssRNA to degradation [35] has triggered the encapsulation of RNA molecules in nanoparticles to improve their delivery, stability, and extended circulation [36]. Alternatively, DNA-based vectors can be used. In this context, the SP6 or T7 RNA polymerase promoter (T7) in alphavirus expression vectors has been replaced by a mammalian host cell compatible eukaryotic RNA polymerase type II promoter such as CMV [37]. The attractive features of DNA-based administration include their easy handling and rapid and inexpensive manufacturing. However, the application of these DNA replicons is dependent on efficient transfection methods, which are inferior to viral delivery. Another issue relates to the need for plasmid DNA to translocate to the nucleus for mRNA transcription followed by its transport to the cytoplasm for translation, which has substantially reduced transgene expression [38]. Different methods such as electroporation [39], gene gun [40], liposome [41], and polymer-based [42], nanoparticle formulation have been developed as well as the application of nuclear localization signals for optimized DNA [43].

In the case of flaviviruses, the Kunjin virus (KUN) has been subjected to the engineering of an expression system [44] (Figure 3). Generally, the GoI is introduced between the first 60 nucleotides of the C20 core protein and the last 22 codons of the E2 envelope protein [44]. As the GoI is expressed as part of a larger polyprotein, an FMDV2A protease sequence was engineered into the KUN vector for removing any foreign flanking regions [45]. Moreover, the generation of a packaging cell line has facilitated KUN particle production [46]. Other flaviviruses such as West Nile virus (WNV) [47], yellow fever virus (YFV) [48], Dengue virus (DENV) [49], tick-borne encephalitis virus (TBEV) [50], and Zika virus (ZIKV) [51] have been engineered. Additionally, expression vector systems have been constructed for the bovine viral diarrhea virus (BVDV) by insertion of the green fluorescent protein (GFP) reporter gene between the N(pro) and C genes in the non-cytopathic type-1 BVDV strain SD1 [52]. The classical swine fever virus (CSFV) has also been engineered as a bicistronic vector containing the GFP gene [53].

In contrast to alphaviruses and flaviviruses, the negative-stranded ssRNA viruses such as rhabdoviruses the RNA self-replication is based on the large protein (L) gene coding for the RdRp and its cofactor in the phosphoprotein (P) gene [28]. Expression vectors for vesicular stomatitis virus (VSV) have been engineered by insertion of the GoI between the glycoprotein (G) and the L genes for the generation of recombinant VSV particles or alternatively by replacing the VSV G by the GoI for pseudovirus production (Figure 4A) [54, 55]. Pseudovirus VSV particles can be produced from plasmid DNA in transfected mammalian cells followed by infection with the VSV G-complemented VSVΔG, which can be used for infection of mammalian host cells leading to transgene expression but no generation of viral progeny [54]. Fully infectious VSV particles can be produced by the infection of mammalian producer cells with VSV G-complemented recombinant virus replacing the VSV G [55]. Reverse genetics has been applied for VSV based on vaccinia virus vectors [56]. Moreover, a vaccinia virus-free system has been engineered to allow the production of infectious VSV particles from DNA [57]. In the context of large-scale production, an inducible VSV L cell line has been engineered for the packaging of recombinant VSV particles [58]. In the context of rabies virus (RABV), expression vectors have been engineered, where the GoI is introduced between the nucleoprotein (N) and P genes, or alternatively, between the G and L genes (Figure 4B) [59]. A reverse genetics system has also been constructed for RABV based on vaccinia virus-free vectors [60].

Among the negative strand ssRNA paramyxoviruses, measles virus (MV) vectors have been engineered for the insertion of the GoI between the P and the M genes or alternatively between the hemagglutinin (HA) gene and the L genes (Figure 5) [27]. Packaging systems and helper cell lines have been engineered for MV, which will allow the rescue of replicating MV from plasmid DNA vectors [61, 62].

Self-replicating RNA viruses have been used for a large number of studies in preclinical animal models and a limited number of clinical trials. This is also reflected in the very low number of approved vaccines, so far. In this section, examples of applications of self-replicating RNA viruses from preclinical studies to clinical trials are described and summarized in Table 1 and Table 2.

The common factors for all vaccine development for both infectious diseases and cancers are the application of self-replicating RNA viral vectors and the overexpression of recombinant proteins used for antigen production. Moreover, the route of administration is of importance, where intramuscular injection is the most common approach. Below are summarized the findings from preclinical studies and clinical trials.

Preclinical studies have provided proof-of-concept for the vaccine establishment to demonstrate on a small scale as cost-effectively as possible that robust immune responses can be induced and in the case of infectious diseases protection against challenges with lethal doses of infectious agents can be achieved. In the case of cancer vaccines, demonstration of tumor growth inhibition, tumor eradication, and complete cure in animal models have been indications that expensive clinical trials can be initiated. Moreover, cancer vaccines are expected to provide protection against challenges with tumor cells in animal models. As described above and summarized in Table 1 and Table 2, these criteria have been reached both for infectious diseases and various cancers, which has allowed moving forward to clinical evaluation.

In the case of clinical assessment, relatively few studies have been conducted, so far. In quite a few cases of infectious diseases, relatively modest immune responses have been obtained in phase I and II trials in healthy volunteers, which is a clear indication of the less straightforward transition from animal studies to humans. For example, clinical evaluation of both MV- and VSV-based vaccine candidates against SARS-CoV-2 was discontinued in phase I despite promising results in animal models [78, 81]. In contrast, VSV-based vaccine development against EDV was highly successful, showing excellent vaccine efficacy in phase III studies [69, 70], leading to vaccine approval by both the FDA and the EMA [71].

The trend for cancer vaccines is similar. Several studies have demonstrated tumor inhibition [89], tumor eradication [93], and even long-term tumor-free survival [91] in different animal models. However, the transition to humans has resulted in less efficient tumor eradication, although PR, SD, and CR have been obtained in individual cases [90, 97].

Although plenty of engineering has been dedicated to the development of more efficient self-replicating RNA virus vectors, there is still room for improvement (Table 3). Related to vector development, several efforts have been made to specifically generate less cytotoxic alphavirus vectors by introducing point mutations in the non-structural protein genes [101, 102]. In the case of cancer vaccines and especially targeting existing tumors, this approach is not necessarily of advantage. Instead, the application of naturally occurring or engineered oncolytic viruses, which specifically replicate in tumor cells leading to their killing [103] might be a better approach. Recently, trans-amplifying SFV vectors, a split-vector derivative of self-replicating RNA vectors, have been engineered for the amplification of transreplicon RNA coding for antigens, which requires minimal amounts of RNA for vaccine development [104].

In the context of antigen targets, the choice of regions or epitopes generating the best possible immune responses is important. In the case of infectious viruses, surface proteins easily recognized by the host cell immune system and/or structures involved in viral replication are desirable. For example, the replacement of the VSV-G protein with target surface proteins such as the SARS-CoV-2 S protein [82]. As epitopes have proven good as antigens, the combination of various epitopes can provide a larger spectrum against infectious viruses or their variants providing enhanced vaccine efficacy. Related to cancer vaccines, desirable antigen targets are those that are commonly expressed or uniquely overexpressed in cancer cells. More recently, personalized antigen approaches have become feasible, where neoepitopes derived from autologous tumors can potentially elicit more potent and personalized immune responses [100].

The delivery of viral vectors has been an area of optimization too. The introduction of targeting sequences by generating pseudoviruses or using strains or mutants capable of selectively targeting antigen-presenting cells (APCs) can improve and target delivery. Related to delivery, the route of administration and dose optimization are key factors for the improvement and optimization of immune responses and vaccine efficacy.

In the case of using RNA replicons instead of recombinant viral particles, prevention of ssRNA degradation has been addressed by stabilization of RNA by chemical modifications [35] and encapsulation of RNA replicons in nanoparticles to protect against degradation, enhance delivery, and prolong circulation of RNA in vivo [36].

Combination therapy of self-replicating RNA viruses has also been attractive. In this context, prime-boost vaccinations with other viral-based vaccines [100], other antiviral and anti-cancer drugs [105], and ICB drugs [88] have been tested.