The WHO announced the Global Arbovirus Initiative in 2022 in response to the growing concerns over expanding outbreaks of arboviral diseases (1), which are primarily caused by pathogenic viruses in the genus Alphavirus (family Togaviridae) and the genus Orthoflavivirus (family Flaviviridae) (2). Urbanization, globalization, human mobility, and climate change, with the ensuing expansion of mosquito vectors, are all anticipated to increase the global burden of arboviral diseases (1, 3). A key intervention has been the development of vaccines (2) and herein we describe an emerging set of technologies that use mosquito alphaviruses and orthoflaviviruses to generate chimeric arboviral vaccines for human and veterinary applications.

Insect-specific viruses (ISVs) are viruses that replicate only in insects, and are distinct from arboviruses, which can replicate both in arthropod vectors (including insects) and vertebrate animal hosts. A range of factors prevent ISVs from infecting vertebrate cells, with such restriction occurring at various stages in the replication cycles (4–11). ISVs arguably represent the dark matter of virology, with vast numbers of ISVs identified by metagenomics (12–17), but only a few isolated and their behavior studied in vivo and in vitro (18). Studies so far indicate that ISVs are generally transmitted vertically from infected females to their offspring via the eggs (9, 19). Some ISVs appear to exhibit a narrow host range, with, for instance, some mosquito ISVs reported to infect only a limited number of mosquito species (12, 20, 21). ISVs are being explored as potential biological control agents for insects that threaten agricultural crops (22). Of some interest has also been the infection of mosquitoes with certain ISVs in order to inhibit, via various mechanisms, replication of pathogenic arboviruses in those mosquitoes (2, 23–26). However, herein we focus on six ISVs from mosquitoes, specifically, two alphaviruses, Eilat virus (EILV) and Yada Yada virus (YYV) and four orthoflaviviruses, Binjari virus (BinJV), Aripo virus (ARPV), YN15-283-02 virus and Chaoyang virus (CYV). For these ISVs, their structural genes can be exchanged with the structural genes from a range of pathogenic alphaviruses and orthoflaviviruses, respectively. A resulting chimeric virus would thus encode the structural proteins of a pathogenic arbovirus and the non-structural polyproteins of one of the aforementioned ISVs. These chimeric viruses are able to replicate efficiently in mosquito cell lines, but are unable to replicate in vertebrate cells. This has allowed the development of a range of ISV-based chimeric vaccines for a number of arboviral diseases ( Figure 1 ).

A series of chimeric vaccines based on EILV, YYV, BinJV, ARPV, YN15-283-02 and CYV have now been described. They are manufactured in mosquito cell lines and resemble virus-like-particle (VLP) vaccines, as they essentially represent whole-virus, protein-based vaccines that cannot generate viral progeny in vertebrate vaccine recipients ( Figure 1 ). They differ from VLP vaccines (27) in that they contain a fully functioning viral genome that is replication competent in mosquito cells. ISV-chimeric vaccines also differ from licensed live-attenuated orthoflavivirus chimeric virus vaccines such as Imojev, for Japanese encephalitis virus (JEV) and Dengvaxia for dengue virus (DENV), which are replication competent in vaccine recipients. The latter encode the structural proteins of JEV and DENV, but are attenuated as they encode the non-structural proteins of the yellow fever virus (YFV) 17D vaccine strain (2). An experimental chimeric live-attenuated alphavirus vaccine has also been reported for Venezuelan Equine Encephalitis virus (VEEV), and encodes the structural proteins of VEEV (TC-83 vaccine strain) and the non-structural proteins of Sindbis virus (28). Thus, chimeric vaccines for alphaviruses and orthoflaviviruses are well described; however, in contrast to the aforementioned live-attenuated chimeric vaccines, ISV-based chimeric vaccines are naturally replication-defective in vaccine recipients as the ISV RNA replication complex, encoded by the non-structural proteins, is non-functional in vertebrate cells. Like many VLP vaccines [e.g. (29, 30)], ISV-based chimeric vaccines present authentically folded structural proteins of the pathogenic arboviruses to the immune system and thereby promote induction of effective neutralizing antibody responses (2).

Herein we describe the ISV-based chimeric vaccine technologies and focus on the immunological issues associated with these technologies including potential self-adjuvanting activity, formulation with adjuvants, provision of authentic tertiary and quaternary structures, safety issues and manufacturing.

An ISV-based chimeric vaccine would be viewed as a Genetically Modified Organism (GMO) in most regulatory environments. A series of potential ensuing risks can be perceived after release of these organisms into the environment (see below), with a range of varied restrictions applied in different countries (98, 99). To avoid such risks, ISV-based chimeric vaccines have been inactivated using traditional methods such as formalin fixation (35, 55), UV-irradiation (52) or X-ray irradiation (63). The former clearly reduced induction of neutralizing antibody responses to EILV/CHIKV chimeras, with formalin fixation preventing cell entry (35) and reducing immunogenicity (100) by inter alia irreversibly modifying lysine residues [discussed in (101)]. Nevertheless, formalin inactivated ISV-based chimeric vaccines, just like many formalin-inactivated whole virus vaccine products (2), can provide protective immune responses (35, 55). Manufacturing and regulatory processes for formalin inactivation of whole virus vaccines are also well established (102). UV-inactivation of RNA viruses is primarily achieved through uracil and cytosine dimer formation (103). However, irradiation technologies are still primarily in the research and development phase for whole virus vaccines (104, 105) and are not currently used for commercial whole arbovirus vaccine products (2). Delivering the correct irradiation dose safely, evenly and consistently during large scale manufacture (i.e. enough to inactivate, but not too much to damage immunogenicity) represent challenges for these processes. For instance, although a UV-inactivated BinJV/WNV_KUN vaccine still afforded protection in mice, neutralizing antibody levels induced by the UV-inactivated vaccine were significantly reduced (106).

Arguably, inactivation of ISV-based chimeric vaccines negates a key advantage of this technology as these vaccines are already intrinsically “inactive”, being unable to replicate in vertebrates. Avoiding the down-stream processes for inactivation and validation of inactivation protocols should represent a key advantage for these ISV-based technologies. Inactivation would remove the GMO classification, and thus the ensuing regulatory hurdles (99), as in most jurisdictions an organism would be defined (in this context) as a replication competent entity. Environmental risks of dissemination and recombination would clearly also be removed by inactivation. However, it should be noted that use of replication competent GMOs as vaccine products is already well established, with a number of licensed live-attenuated arboviral vaccines classed as GMOs e.g. (i) Ixchiq (Valneva), the recently approved CHIKV vaccine that has a 186 nucleotide (62 amino acid) deletion in nsP3, (ii) Imojev (Sanofi), a JEV vaccine, which has the non-structural proteins from YFV 17D, (iii) Dengvaxia (Sanofi), a DENV vaccine, also with a YFV backbone, and (iv) Qdenga (Takeda), where all four serotypes have the non-structural proteins of DENV2. Recombinant virally vectored vaccines are also classified as GMOs with a range of these now also licensed, e.g. (i) Covishield (AstraZeneca) a recombinant adenovirus vaccine for COVID-19, (ii) Raboral (Boehringer Ingelheim), a recombinant vaccinia veterinary vaccine for rabies, and (iii) Trovac-NDV (Boehringer Ingelheim), a recombinant fowlpox livestock vaccine for Newcastle disease (107).

The importance of presenting the immune system with authentically folded arboviral vaccine immunogens is widely recognized (92, 108), with protective antibodies often targeting quaternary epitopes (109–111). Currently licensed orthoflavivirus and alphavirus vaccines likely deliver antigens with authentically folded tertiary and quaternary protein structures, facilitated by the robust ability of orthoflavi- and alpha-viral structural polyproteins to self-assemble into virion particles in a range of settings (2).

ISV-based chimeric vaccines are fully functional viruses in insect cells and assemble into authentic virion particle structures (33, 48, 61, 112). For instance, low resolution analysis of BinjV chimeras by cryo EM revealed that the virus particles accurately mimic the virion structure of the wild-type pathogens (112). However, subtle differences may emerge to be important, for instance, ortho flavivirus envelope proteins are able to “breath”, adopting more “bumpy” versus “smooth” configurations at the virion surface. Ensuring that an ISV-based chimeric vaccine has the same structure or conformational dynamics as the contemporary pathogenic orthoflavivirus being targeted by the vaccine, may be important for optimizing vaccine efficacy (113).

Authentically folded immunogens are important for induction of antibodies capable of high-affinity and high-avidity binding to the target arbovirus virions and arbovirus infected cells. The vaccine-induced antibodies can be both neutralizing and non-neutralizing, with the latter also mediating a range of protective activities that are generally difficult to measure in clinical trial settings (2). Indirect evidence for a protective role for non-neutralizing, vaccine-induced, antibodies in humans comes from inter alia a phase III trial of Qdenga, where neutralizing antibody titers against DENV2, DENV3 and DENV4 did not correlate with protection (114).

For all the ISV-based chimeric vaccines the non-structural proteins of the pathogenic arboviruses are not present as vaccine antigens. These proteins are nsP1, nsP2, nsP3, nsP4 for alphaviruses, and NS1, NS2A, NS2B, NS3, NS4A, NS4B, NS5 for orthoflaviviruses ( Figure 1 ). The ISV versions of these non-structural proteins are synthesized in the insect cell lines used to manufacture the chimeric vaccines. ISV-based chimeric vaccines also contain ssRNA chimeric genomes, which encode the non-structural proteins of the ISV (but are not translated in vaccine recipients) ( Figure 1 ).

Ixchiq, the recently approved, live-attenuated CHIKV vaccine, would present CHIKV nsPs to vaccine recipients, whereas this would not occur for non-replicating VLP-based or ISV-based chimeric vaccines. Anti-nsP responses likely do not play a significant role in protection against pathogenic alphavirus infections, a contention supported inter alia by successful phase 3 trials recently reported for an aluminum hydroxide adjuvanted, CHIKV VLP vaccine (115). Anti-alphaviral CD8 T cell responses may play a minor role in protection, but they are generally viewed as secondary to antibody responses (116–120). CHIKV appears able to evade surveillance by antiviral CD8 T cells, in part, by nsP2-mediated disruption of MHC-I antigen presentation (121). Nevertheless, induction of CD8 T cells was shown for the EILV/CHIKV vaccine in mice (34) and NHPs (36), and for the ARPV/ZIKV vaccine in mice (59), suggesting that these vaccines can infect and endosome escape, thereby delivering structural protein antigens into the MHC-I processing pathway (122).

In contrast to alphaviruses, protective immune responses directed at orthoflaviviral non-structural proteins, in particular NS1, are well described, with a number of groups pursuing the development of NS1-based vaccines (123–127). Although cross-reactivity of anti-ZIKV NS1 responses with self has been described (128), a causal link to autoimmune disease (129) has yet to be established (130, 131). NS1 proteins of flaviviruses show a degree of sequence homology (132), so live attenuated vaccines such as Dengvaxia (Sanofi) and IMOJEV (Sanofi Pasteur) with a YFV 17D backbone, would likely induce anti-NS1 responses capable of cross-reacting with DENV and JEV NS1 proteins, respectively. Qdenga (Takeda), with a DENV2 backbone, would arguably induce better anti-DENV NS1 responses (2), as sequence homologies between NS1 from DENV2 versus DENV1, 3 and 4 are more substantial (132). However, even if anti-NS1 protein responses can mediate effective protection, a number of licensed orthoflavivirus vaccines achieve adequate protection without inducing such responses, specifically, the inactivated vaccines for (i) JEV, Ixiaro (Valneva), (ii) tick-borne encephalitis virus, Encepur (Bavarian Nordic), and (iii) West Nile virus, Innovator (a horse vaccine from Zoetis). An inactivated virus vaccine for ZIKV was also recently shown to be effective in pregnant marmosets (133).

An ideal characteristic for many alphavirus and orthoflavivirus vaccines would be provision of long-term protective immunity, an issue particularly pertinent for human vaccines in resource poor settings, although clearly less of an issue for short-lived livestock. Perhaps a standout is the YFV 17D vaccine, which provides life-long immunity in humans after a single vaccination (134). Non-replicating, protein-based vaccines are often viewed as providing poor long-term protection. However, this is not entirely accurate with, for instance, the whole-virion-based, formalin inactivated, aluminium hydroxide adjuvanted, hepatitis A vaccine providing protection for >20 years, and the VLP-based, aluminium hydroxide adjuvanted, human papilloma virus vaccine (Gardasil), providing protection for up to 14 years (135). The current COVID-19 mRNA vaccines would appear to perform poorly with respect to durability of protective responses, with neutralizing antibody responses waning within 6 months (136). However, this may be related to an inability to generate spike-specific, long-lived plasma cells (137). The durability of protection after vaccination with mRNA vaccines against arboviruses has yet to be evaluated (138). The same goes for ISV-based chimeric vaccines, and may depend on the choice of adjuvant. Perhaps encouraging, a single dose of a BinJV/ZIKV vaccine, with no adjuvant formulation, provided protective immunity in Ifnar^-/- mice for 15 months (139).

Manufacture of human and animal vaccines need to adhere to a series of production and quality control standards, with requirements for human vaccines considerably more onerous than for animal vaccines (140, 141). The number of cell lines available for manufacture of human products is limited, but include CHO (142), HEK293 (143) and Vero cells (144), as well as baculovirus systems that use insect cells from Spodoptera frugiperda (armyworm moth) (29, 145). However, to date, no large-scale vaccine production systems have been developed for mosquito cell lines. A desirable property of C6/36 cells is that they do not contain adventitious viruses (146), with such agents having caused problems for human and veterinary vaccine products in the past (147, 148). Another desirable property of ISV-chimeric virus vaccines is that they often grow to higher titers in insect cell lines than their parental viruses (34), with, for instance, the BinJV/ZIKV chimera produced yields of up to ~10^9.5 CCID₅₀/mL or ~7 mg/liter (48). Production and purification of EILV/CHIKV vaccine in C7-10 cells has been described, with culture in serum free medium for 18 h before harvest, followed by Cellufine sulfate column and then sucrose gradient purification (34). This purified vaccine, delivered once i.m. at up to 1.3 x 10⁸ PFU, showed no overt adverse effects in NHPs (36). Manufacturing and purification processes for BinJV chimeric vaccines have also been developed and involve growth in C6/36 cells and purification by density gradient centrifugation (54, 68). A serum free, suspension culture system was recently described for production of BinJV chimeras in C6/36 cells, representing an important step in the development of a scalable manufacturing system (149).

A number of critical steps remain for development of human ISV-based vaccines. These include (i) regulatory clearance for a master cell bank of a suitable mosquito cell line, (ii) characterization of impurities and development of GLP/GMP vaccine production and purification processes, (iii) identification of suitable excipients and storage protocols that preserve vaccine immunogenicity, and (iv) development of quality control processes to monitor vaccine production. Thereafter, phase I human trials, which primarily determine safety, might be initiated. For veterinary vaccines, depending on the animal species, a priority is often cost-effectiveness, so manufacturing costs will generally need to be low (150).

Needle-free vaccine delivery overcomes needle-phobia, eliminates needle-stick injuries and the necessity for sharps disposal, and may emerge to be attractive and cost-effective in mass vaccination campaigns, especially in resource poor settings (179). Two devices have been licensed for delivery of vaccines, Stratis (PharmaJet) for delivery of inactivated influenza vaccine (Afluria, Seqirus) in the USA, and Tropis (PharmaJet) for delivery of the ZyCoV-D DNA COVID-19 vaccine (Zydus Lifesciences) in India (180) for Restricted Use in Emergency Situations. These devices are applied to the skin and deliver vaccines via a narrow precise fluid stream that penetrates the skin. Similar needle free systems have also been developed for the livestock industry and have the advantage of reducing pain, saving time, avoiding lesions in muscles (meat products), as well as avoiding the aforementioned safety issues associated with needle use (181).

An alternative needle-free vaccine delivery technology involves microarray patches or microprojection arrays (182), which involves dry coating of vaccine onto microprojections arrayed on small patches that are applied to the skin and that deliver antigen to skin antigen presenting cells (183, 184). A number of microarray patches have been evaluated in phase I human trials for inter alia influenza, measles, rubella, and SARS-CoV-2 vaccines. They were generally regarded as safe and well tolerated, inducing similar or increased immune responses when compared with conventional needle-based vaccination (185–188).

The technology developed by Vaxxas (188–190) has been used to deliver a BinJV-based chimeric vaccine for DENV that provided protection in a mouse model (68, 191).

A number of mosquito ISV-based chimeric vaccines for human arboviral pathogens have been developed and been shown to generate protective immunity in preclinical studies in mice and NHPs. However, human vaccine applications will require a certified mosquito cell line and GMP manufacturing and purification processes to be established, with evaluation of different adjuvant formulations also likely required. Ultimately, ISV-based chimeric vaccines may have to compete with mRNA vaccine technologies (192, 193) on inter alia cost of goods, durability of responses, and/or safety. Manufacture of ISV-based vaccines is at an early stage and data is currently limited. Conceivably, the manufacturing performance of mosquito cell line systems might emerge to be comparable with baculovirus systems (145); however, cell line-derived, whole-virus vaccines that need inactivation and adjuvanting generally have a relatively high cost of goods (194). Future developments in mRNA vaccine technology may also reduce the costs of this new technology (195). Specific adjuvants may be required to promote responses and durability of protection of ISV-based chimeric vaccines in humans (2, 135). New developments may also lead to improvements in the durability of responses after vaccination with mRNA vaccines (196). The safety profile of mRNA vaccines is now well established with billions of doses delivered globally (197). The safety profile for ISV-based vaccines has yet to be formally established and must await phase I human trials.

Livestock applications have been illustrated for BinJV/WNV_KUN for crocodiles and BinJV/JEV for pigs. Although unable to replicate in vaccine recipients, inactivation (e.g. formalin) of ISV-based chimeric vaccines removes the regulatory hurdles associated with release of GMOs and avoids any risk of transmission to mosquitoes. Whether such vaccines would emerge to be cost-effective enough for livestock markets has yet to be determined.