Virus-Like Particle-Based Countermeasures Against Rift Valley Fever Virus

Authors


R. Flick. NewLink Genetics Corporation, 2901 South Loop Dr., Bldg. 3, Ames, IA 50010, USA. Tel.: +1 515 598 5017; Fax: +1 515 296 3820; E-mail: rflick@bpsys.net

Summary

Rift Valley fever virus (RVFV) is an arbovirus that causes significant morbidity and mortality in both humans and livestock. With increased world travel and the threat of bioterrorism, there is a real risk of RVFV spreading to naïve geographical areas (Trans. R. Soc. Trop. Med. Hyg., 73, 1979, 618; MMWR Morb. Mortal. Wkly Rep., 49, 2000, 905). The introduction of RVFV would cause critical public health, agricultural and economic damage. Despite the clear need for an efficacious vaccine, there are no United States (US) Food and Drug Administration or US Department of Agriculture approved vaccines against RVFV. To address this need, a virus-like particle (VLP)-based vaccine candidate was developed. First, a non-replicating chimeric RVF VLP vaccine candidate was generated that protected mice and rats against a lethal RVFV challenge. This was followed by the development and optimization of conditions for production of RVF VLPs in insect and mammalian cells. Immunological studies demonstrated that VLP-based vaccine candidates elicit both humoral and cellular immune responses. Subsequent challenge studies using a lethal wild-type RVFV strain under high-containment conditions showed that RVF VLP vaccine candidates can completely protect mice and rats.

Impacts

  •  Virus-like particle (VLP)-based vaccines against Rift Valley fever virus (RVFV) can be easily and efficiently produced in insect and mammalian cells.
  •  RVF VLP vaccination results in immune responses against RVFV.
  •  A single vaccine dose of RVF VLPs can protect rats and mice against a lethal infection with RVFV.

Introduction

Because of the lack of prophylactic and therapeutic measures, potential for human-to-human transmission, and the serious threat to livestock, RVFV is considered a serious public health concern in endemic areas with history of previous RVFV outbreaks (Daubney et al., 1931, WHO, 2010). With increased world travel and the threat of bioterror, the risks associated with RVFV outbreaks extend to many non-endemic developed countries.

Rift Valley fever virus (RVFV) is an enveloped virus in the family Bunyaviridae characterized by a tripartite, single-stranded, negative-sense RNA genome (Elliott et al., 1992; Schmaljohn and Nichol, 2006) that consists of large (L), medium (M), and small (S) segments encoding a RNA-dependent RNA polymerase, the two glycoproteins (GN and GC) and a non-structural protein (NSm), and a nucleoprotein (N) and another non-structural protein (NSs), respectively (Elliott et al., 1992; Schmaljohn and Hooper, 2001). RVFV is one of several arthropod-borne viruses that can cause potentially lethal haemorrhagic fevers in humans and livestock (Gerdes, 2004; Flick and Bouloy, 2005). RVFV infections in humans are generally associated with benign fever but, in some cases, can lead to further complications such as retinal vasculitis, encephalitis, neurologic deficits, hepatic necrosis, or fatal haemorrhagic fever (Meegan, 1979; Geisbert and Jahrling, 2004; Gerdes, 2004; Flick and Bouloy, 2005). Historically, <2% of infected human cases are fatal, but recent case fatality rates of 20–30% were reported (WHO, 2007). RVFV infections in humans are generally preceded by outbreaks in livestock (Gerdes, 2004; Flick and Bouloy, 2005). While mortality rates in adult livestock vary from 10% to 70%, the mortality rates in young lambs can reach 100% (Gerdes, 2004). RVFV infections can also lead to 90–100% abortions in pregnant cattle and sheep. The introduction of RVFV would result in a significant negative economic impact because of embargoes on the export of animal products possibly infected with RVFV (Flick and Bouloy, 2005; Mandell and Flick, 2010).

RVFV is classified as a Category A High Priority Pathogen by the National Institute for Allergy and Infectious Diseases (NIAID) (http://www.niaid.nih.gov/topics/BiodefenseRelated/Biodefense/Pages/CatA.aspx). Additionally, RVFV is on the Centers for Disease Control (CDC) Bioterrorism Agents list (http://www.bt.cdc.gov/agent/agentlist-category.asp#a) and is classified as a Department of Health and Human Services (HHS) and US Department of Agriculture overlap select agent.

Because RVFV has the potential for spreading to non-endemic regions with devastating consequences for both human and livestock health, there is an urgent need for a safe and efficacious vaccine against RVFV, especially for livestock, as such a vaccine would prevent or reduce the spread of RVFV infections from livestock to humans.

Vaccine Candidates

Several vaccine candidates against RVFV are in various phases of development (Bird et al., 2007b, 2009; Bouloy and Flick, 2009; Ikegami and Makino, 2009; Kortekaas et al., 2010c; LaBeaud et al., 2010; Pepin et al., 2010; Boshra et al., 2011a). The live-attenuated Smithburn vaccine against RVFV is effective in livestock, but is not recommended for pregnant animals (Barnard and Botha, 1977; Botros et al., 2006) and therefore not practical for a large-scale vaccination program. A formalin-inactivated RVFV vaccine, TSI-GSD 200, when tested in human volunteers, elicited neutralizing antibody titres against RVFV (Pittman et al., 1999; Rusnak et al., 2011). A similar formalin-inactivated RVFV vaccine was shown to be significantly efficacious in livestock (Barnard and Botha, 1977), but the vaccine is difficult to produce, expensive and requires multiple boosters to maintain immunity (Kark et al., 1982, 1985; Niklasson et al., 1985; Frank-Peterside, 2000). Because of these limitations, a live-attenuated vaccine, MP-12, was generated for human and animal use (Caplen et al., 1985). MP-12 is characterized by several mutations in the M and L RNA segments that substantially decrease virulence (Takehara et al., 1989; Vialat et al., 1997). MP-12 is immunogenic and protects young and pregnant livestock (Morrill et al., 1991; Baskerville et al., 1992). Mucosal or intramuscular administration of MP-12 elicits immune responses and protects Rhesus macaques against an aerosol RVFV challenge (Morrill and Peters, 2011a,b). The vaccine is currently in Phase II human clinical trials (http://clinicaltrials.gov/ct2/show/NCT00415051). Another live-attenuated vaccine candidate, Clone 13 (a plaque isolate of the RVFV 74HB59 strain), is efficacious in young calves and pregnant ewes (Swanepoel and Coetzer, 2004, Dungu et al., 2010; von Teichman et al., 2011). Attenuation is attributed to the deletion of 70% of the NSs open reading frame (Muller et al., 1995). Finally, R566 is a reassortant of MP12 that combines the attenuating mutations of MP-12 and the NSs gene deletion of Clone 13 (http://www-pub.iaea.org/mtcd/meetings/PDFplus/cn110bsyn.pdf).

As summarized recently by Bird et al. (2009) and Bouloy and Flick (2009), recombinant RVFV proteins have also been assessed as potential vaccine candidates. RVFV GN and GC proteins were expressed in Spodoptera frugiperda (Sf9) insect cells using recombinant baculovirus (rBV). Immunization with cell lysates induced protective immunity in mice (Schmaljohn et al., 1989; Takehara et al., 1990). Recombinant RVFV glycoproteins were also expressed through viral vectors such as the Venezuelan equine encephalitis virus (VEEV) and Sindbis virus (SINV) replicon for use as veterinary vaccines (Gorchakov et al., 2007; Heise et al., 2009). Newcastle disease virus-based vectors designed to produce RVFV glycoproteins elicited immune response in mice and livestock (Kortekaas et al., 2010a,b; Harmsen et al., 2011). A capripoxvirus (CPV) recombinant virus expressing RVFV GN and GC proteins protected both mice and sheep against a RVFV challenge (Soi et al., 2010). A non-replicating complex adenovirus (CAdVax) vector expressing RVFV glycoprotein genes elicit immune responses and protects mice against a lethal RVFV challenge (Holman, 2009). Lastly, DNA-based vaccines encoding for select RVFV proteins have been shown to elicit neutralizing antibody titres and protect against a lethal challenge with RVFV in mice (Spik et al., 2006; Bird et al., 2008; Lagerqvist et al., 2009; Bhardwaj et al., 2010; Boshra et al., 2011b) and sheep (Lorenzo et al., 2008).

Virus-Like Particles

The current vaccine candidates under development against RVFV have important limitations, including safety concerns (Bouloy and Flick, 2009; Ikegami and Makino, 2009; LaBeaud et al., 2010). To overcome some of these limitations, virus-like particle (VLPs) present a safer and often more immunogenic alternative vaccine platform. VLPs can be produced in insect cells (Liu et al., 2008; Monie et al., 2008; Harper, 2009; Metz and Pijlman, 2011; Vicente et al., 2011), mammalian cells (Roldao et al., 2010), yeast (Govan, 2008; Harper, 2009) and plants (Landry et al., 2010; Soria-Guerra et al., 2011).

Traditionally developed for the study of viral structure and assembly, genome packaging, virion budding, receptor analysis and virus entry (Bos et al., 1997; Johnson and Chiu, 2000; Schmitt et al., 2002; Li et al., 2003; Licata et al., 2004; Overby et al., 2006; Ye et al., 2006; Piper et al., 2011), VLPs can also serve as efficacious vaccines (Raghunandan, 2011; Schneider-Ohrum and Ross, 2011). VLPs are more immunogenic than simple recombinant proteins because they closely resemble the wild-type virus from which they are derived (Noad and Roy, 2003; Grgacic and Anderson, 2006). VLPs are also considered safer than live-attenuated vaccines because they do not contain any viral genetic material, thus eliminating the risk of reversion or reassortment to a virulent form of the virus (Noad and Roy, 2003; Grgacic and Anderson, 2006). Furthermore, there are precedents for the use of VLP-based vaccines in humans: Cervarix (GlaxoSmithKline, Rixensart, Belgium; Merck, Whitehouse station, New Jersey, USA.) and Gardasil (Merck) vaccines for human papilloma virus and Recombivax HB (Merck) for hepatitis B are Food and Drug Administration approved VLP-based vaccines (Ellis, 1993; Monie et al., 2008; Harper, 2009; Lu et al., 2011). In addition, promising VLP-based vaccine candidates for emerging viral pathogens such as Chikungunya virus (Akahata et al., 2010), Ebola and Marburg viruses (Warfield et al., 2005, 2007), H1N1 and H5N1 influenza viruses (Kang et al., 2009; Landry et al., 2010; Pushko et al., 2010; Ding et al., 2011a,b; Giles and Ross, 2011), Hantavirus (Li et al., 2010), arenaviruses (Casabona et al., 2009; Branco et al., 2010) and Nipah virus (Patch et al., 2007; Walpita et al., 2011) are in clinical trials or various stages of development (Roldao et al., 2010).

VLP-Based Vaccines for RVFV

VLPs are ideally suited as vaccine candidates for RVFV as they do not carry any viral genetic material and therefore adhere to the differentiating infected from vaccinated animals (DIVA) concept (Capua et al., 2004; Bird et al., 2008; Lee et al., 2011), which allows for the identification of vaccinated and infected animals in a given livestock population – a crucial requirement during an outbreak situation. For example, sensitive reverse transcriptase (RT) – PCR technology can be employed that can reliably detect the RNA genome of RVFV. In addition, VLPs do not contain all protein components of RVFV, allowing additional diagnostic assays (e.g. RVFV antibody-specific ELISAs) to differentiate between infected and vaccinated animals, because infected animals would contain antibodies to proteins not present in RVF VLPs.

The development of various VLP-based vaccine candidates against RVFV has been reported by several laboratories. Liu et al. (2008) used recombinant baculoviruses (rBV) expressing RVFV GN, GC and N proteins to successfully generate RVF VLPs in Spodoptera frugiperda insect cells. RVF VLPs were purified by ultracentrifugation and transmission electron microscopic (TEM) analysis revealed that the VLPs structurally resembled the wild-type RVFV virions. Transcriptionally active RVF VLPs produced in mammalian cells by Naslund et al. (2009) were used to protect mice from lethal challenge with RVFV (Habjan et al., 2009; Pichlmair et al., 2010). In addition, RVF VLPs lacking the N protein produced in Drosophila insect cells were immunogenic and protected mice against a RVFV lethal challenge (de Boer et al., 2010).

Chimeric VLPs

Mandell et al. (2010a) produced chimeric VLPs (RVF chimVLPs) in HEK-293 cells constitutively expressing a Moloney murine leukaemia virus (MoMLV) gag protein (293-gag). The presence of gag in VLPs has been shown to increase yield and stability (Gheysen et al., 1989; Haffar et al., 1990; Rovinski et al., 1992; Hammonds et al., 2003; Haynes et al., 2009). RVFV chimVLPs were produced by transient transfection of 293-gag cells with expression plasmids encoding the RVFV glycoproteins and nucleoprotein. Chimeric RVF VLPs were concentrated from supernatants by tangential flow filtration (TFF) and ultracentrifugation through 20% sucrose cushions. The presence of RVFV G, N and gag proteins was confirmed by Western blot analysis with antibodies to the RVFV proteins and MoMLV gag. TEM analysis confirmed that the size, uniformity and overall structure of the VLPs were similar to RVF virions (Mandell et al., 2010a).

Process development was performed to optimize RVF chimVLP production. First, the use of a codon-optimized RVFV G gene marginally increased the yield of RVF chimVLPs. Second, for the generation of RVF chimVLPs, a molar ratio RVFV N and G expression plasmids of 1:4.3 resulted in the maximum yield. Third, time-course experiments demonstrated that RVF chimVLPs harvested every 12 h from the cell culture supernatant resulted in the highest yield of RVF chimVLPs. Finally, contrary to published results (Overby et al., 2006; Habjan et al., 2009), the inclusion of a M segment-based minigenome (under the control of an RNA polymerase I promoter) affected neither the yield nor the amount of N incorporated into the RVF chimVLPs (Mandell et al., 2010a).

Suspension Cell-Derived VLPs

VLPs or recombinant proteins are generally produced at a small scale in adherent (monolayer) mammalian cells or at a larger scale in insect cells (Schmaljohn et al., 1989; Rovinski et al., 1992; Wurm and Bernard, 1999; Maranga et al., 2002, 2003; Hunt, 2005; Warnock and Al-Rubeai, 2006; Liu et al., 2008; Habjan et al., 2009; Pillay et al., 2009; Mandell et al., 2010a). Small-scale production is usually sufficient for structural and qualitative analysis. Unlike the adherent mammalian cells, cells grown in suspension cultures are more amenable for large-scale production of recombinant proteins (Nallet et al., 2009) or virus (Paillet et al., 2009). Cost-effective large-scale production is critical for any VLP-based vaccine candidate. To address this issue, Mandell et al. (2010b) compared the production of RVF VLPs in insect and mammalian suspension cell lines. RVFV VLP production in these two systems were optimized by the identification of suitable cell lines, transfection parameters, culture conditions, molar ratios of transfected plasmids, infection conditions and harvesting schedules. As a result, a novel, scalable system for production of RVF VLPs in insect and mammalian suspension cells was established (Mandell et al., 2010b).

For the production of RVF VLPs in insect cells, rBV containing RVFV G or N coding genes were generated. Although Sf9 and Sf21 cells were initially cultured in spinner flasks, square bottles and Erlenmeyer flasks for the production of rBV carrying RVFV G or N genes, the highest infectious rBV titers were obtained using Sf9 cells cultured in square bottles. Similarly, RVF VLP production in square bottles with insect suspension cells infected with rBV carrying RVFV G or N genes resulted in the highest yield of RVF VLPs. The use of rBV RVFV G or N at a multiplicity of infection (MOI) of 1 and 5, respectively, resulted in the highest yield of RVF VLPs. Additionally, Sf21 cells produced substantially higher amounts of RVF VLPs than Sf9 cells. Finally, analysis of harvest frequencies demonstrated that 24 h harvest schedule resulted in the best RVF VLP yields (Mandell et al., 2010b).

RVF VLPs were also produced in mammalian cells by transient transfection of HEK-293 cells with RVFV G and N expression plasmids using the cost-effective and scalable polyethylenimine (PEI) technology. RVF VLPs were concentrated by TFF and purified by ultracentrifugation through a 20% sucrose cushion. Direct comparison of RVF VLP production in adherent and suspension mammalian cells demonstrated that, despite a lower transfection efficacy, suspension cells produced approximately 10–60× more RVF VLPs. In addition, for the highest VLP yield, the optimal molar ratio of expression plasmids carrying the genes for RVFV G and N was found to be 1:3. Finally, harvesting RVF VLPs every 12 h up to 144 h post-transfection resulted in the maximum production of RVF VLPs (Mandell et al., 2010b).

Immunogenicity of RVF VLPs

The chimVLPs produced by Mandell et al. (2010a) were analyzed for their ability to elicit an immune response in mice against RVFV. Vaccine efficacy often correlates to the generation of virus-neutralizing antibodies (Khanam et al., 2006; Ye et al., 2006), especially in the case of RVFV (Peters et al., 1986, 1988). To study the ability of RVF VLPs to elicit long-lasting neutralizing antibodies, mice were subcutaneously (s.c.) immunized 3× with RVF chimVLPs and neutralizing antibody titers in the immunized mice were measured 6 months post-first vaccination (Mandell et al., 2010a). While Plaque Reduction Neutralization titers (PRNT80) of 1:40 were considered to be protective in rodents (Peters et al., 1986; Anderson et al., 1987, 1991) and primates (Peters et al., 1988) against RVFV, Mandell et al. (2010a) demonstrated that immunization with RVF chimVLPs results in >1:640 neutralizing antibody titers even 6 months post-first vaccination. Interestingly, immunization with RVF VLPs without N protein elicits similar high titer, long-lasting neutralizing antibodies. These results suggest that immunization with RVF VLPs can provide long-lasting immunity to RVFV antigens.

The amount of antibodies in RVF VLP-vaccinated mice against RVFV G or N was quantified by a recently developed ELISA using bacterially expressed RVFV GN, GC or N proteins as standards. Vaccination of mice with either insect or mammalian cell-derived RVF VLPs elicits antibodies against both RVFV G and N proteins (Koukuntla R, Mandell RB and Flick R, unpublished data). In addition, a durability study over a period of ∼15 months post-vaccination demonstrated that RVF VLP-vaccinated mice have long-lasting anti-N protein antibodies. In similar studies conducted in rats, it was observed that even a single vaccination with RVF VLPs elicits humoral response to RVFV G proteins (unpublished data).

Vaccine efficacy is often correlated with both humoral and cellular immune responses (Wack et al., 2008). To study the cellular immune response to RVF VLPs, antigen-specific secretion of select cytokines by splenocytes isolated from RVF VLP-vaccinated mice was examined. Splenocytes harvested from mice vaccinated with RVF chimVLPs were cultured in the presence of RVFV antigen [heat-inactivated MP12 or a non-specific control antigen (heat-inactivated influenza A virus strain, A/HK/1/68)]. Analysis showed that splenocytes from RVF chimVLP-vaccinated mice secrete cytokines in response to MP12 but not to the non-specific antigen. The secreted cytokine profile was consistent with both humoral (TH2) and cellular (TH1) responses (Mandell et al., 2010a).

Taken together, these results show that immunization with RVF VLPs elicits an immune response with both humoral and cellular components.

RVF VLP Vaccine Efficacy Studies

Although the RVF chimVLPs were shown to be immunogenic and elicit both humoral and cellular immune responses, vaccine immunogenicity does not always correlate with protection against a lethal challenge. Therefore, vaccine efficacy was measured in a mouse lethal challenge model (Mandell et al., 2010a). Mice were vaccinated 3× with either RVF VLPs or RVF chimVLPs and challenged with 103 pfu RVF ZH501 virus under biosafety level 4 (BSL-4) conditions. While none of the unvaccinated mice survived the lethal challenge, 56% of the RVF VLP-vaccinated mice and 68% of the chimVLP-vaccinated mice survived the challenge with a lethal dose of ZH501 virus. The presence of gag protein in the RVF chimVLPs did not significantly increase vaccine efficacy (Mandell et al., 2010a).

Mandell et al. (2010a) also tested the vaccine efficacy in rats as they offer an important alternate animal model (Anderson and Peters, 1988; Anderson et al., 1991; Bird et al., 2007a). Rats were vaccinated 3× at 2-week intervals with RVF chimVLPs and then challenged with 105 pfu of RVF ZH501 virus ∼9 weeks post-third vaccination. While all of the unvaccinated control rats succumbed to disease by day 4 post-challenge, 100% of the rats vaccinated with RVF chimVLPs survived the lethal challenge with ZH501 virus. In addition, unlike the control rats that showed rapid and substantial weight loss before succumbing to challenge, RVF chimVLP-vaccinated rats maintained their weight throughout the course of the study (Mandell et al., 2010a). Mandell et al. (2010a) also conducted vaccine efficacy studies with insect cell-derived RVF VLPs. Rats were vaccinated either 3×, 2× or 1× and challenged 8 weeks post-first vaccination with 105 pfu of RVF ZH501 virus. While five of six control animals succumbed to RVFV challenge within the first 5 days, 100% of the rats vaccinated with RVF VLPs, even with a single dose, were protected against a lethal challenge. Control rats rapidly lost weight while vaccinated rats maintained their bodyweight throughout the course of the study. Recent studies have shown that RVF VLPs produced in mammalian suspension cells can protect 100% of vaccinated mice, even at low doses or with a single vaccination (unpublished data).

Conclusion

We have successfully generated RVF VLPs in insect (suspension) and mammalian (adherent or suspension) cells (Mandell et al., 2010a,b). RVF VLP production conditions were optimized in different cell lines to increase the yield of RVF VLPs, thus substantially reducing the cost of production. The use of serum free media provides an added advantage for quality control and eliminates the risk associated with use of animal products in vaccine production (Mandell et al., 2010b). VLPs are safe to use and conform to the DIVA concept.

RVF VLPs elicit antibodies against RVFV G and N proteins in both mice and rats (Mandell et al., 2010a and unpublished data). Further immunological studies demonstrated that RVF VLPs are effective at stimulating a cellular immune response in vaccinated mice. Vaccine efficacy studies show that RVF VLP-based vaccine candidates can protect vaccinated mice and rats against a lethal challenge with wild-type virus. In addition, the VLP-based vaccine candidate is efficacious even at a single dose. Thus, RVF VLPs are safe to use, DIVA conforming, immunogenic and efficacious against a lethal challenge. The reviewed data strongly presents a case for the use of RVF VLPs as a viable vaccine candidate against RVFV. Further studies in primates and livestock are required to conclusively establish the protective potential of the RVF VLPs for both human and veterinary applications.

Conflicts of interest

The authors work at Newlink Genetics Corporation/BioProtection Systems developing the described vaccine candidate (RVF VLP-based) and own stocks/stock options.

Ancillary