Historically, use of whole, infectious but attenuated virus has been more effective in producing a strong, lasting, and effective immune response, and this may be especially true for Alphaviruses and Flaviviruses. The membrane glycoproteins of Alphaviruses [and probably Flaviviruses] are folded during assembly into compact high energy metastable structures [Carleton et al 1997, Mulvey and Brown 1994, 1995, 1996]. The energy stored in the proteins is likely employed to drive the events that lead to infection of a host cell by the virion [Paredes et al 1993]. The unstable high-energy configuration of the protein renders it exquisitely sensitive to treatments used to remove it from the virion. Protocols to release and purify the protein result in its collapse into non-native, low-energy configurations [Mulvey and Brown, 1994]. The folding process resulting in the production of the high energy, native configuration is complex and involves interactions with other virus proteins and molecular chaperones [Carleton et al 1997; Mulvey and Brown, 1995]. This latter point suggests that vaccine strategies that employ expression of domains of these proteins as subunits or as components of other virus proteins may not produce the conformation of the protein that exists in the mature infectious virus, and thus will not allow production of antibodies that would neutralize the proteins of the wild-type virus [i.e. immunity to the native virus would not be conferred]. Neutralizing antibodies, which will bind and inactivate a viral invader, are a necessary component of a successful vaccine.

However, other types of antibodies that bind, but do not inactive virus (non-neutralizing or sub-neutralizing antibodies) can also be generated by vaccines that do not use the intact virus, such as chimeras or sub-unit vaccines. In the case of Dengue virus, virus exposure that leads to high levels of non-neutralizing antibodies is linked to occurrences of the more severe and often fatal Dengue hemorrhagic fever upon infection with subsequent Dengue virus serotypes [16]. In order to make a safe and effective vaccine against Dengue, the vaccine must protect from all 4 Dengue virus serotypes and initiate the desired immune response (neutralizing antibodies) while not generating the potentially deleterious non-neutralizing antibodies.

We have overcome the problems presented above by exploiting the fact that Alphaviruses and Flaviviruses have evolved to effectively replicate in the unique biochemical and genetic environments of both vertebrate and invertebrate hosts (Brown and Condreay, 1986). This implies that they have adapted to engage various components of these diverse hosts. It is reasoned that certain elements of virus structure may be essential for assembly in the mammalian host cell that are not essential for growth in the insect host cell. The Alpha- and Flaviviruses are hybrid structures; their proteins are the product of virus genes but their membranes are the product of the host cells biochemistry. When a virus is grown in insect cells it acquires insect membranes and when it is grown in mammalian cells it acquires mammalian membranes. There are dramatic differences in the physical and chemical properties of mammalian and insect membranes. Our approach to vaccine development takes advantage of these differences.

In cells initially infected by the virus through vaccination, the virus can partially replicate to form capsids, but cannot form infectious virions as it has no envelope – the capsids of the virus cannot infect another cell. This gives the host immune system a chance to mount a defense. The protein coat [capsid] of the virus is unchanged, allowing full host recognition. This should stimulate an immune response with high titers of neutralizing antibody to provide immunity to the wild-type virus, without production of disease. The timing and duration of the immune response and titers of non-competent virus produced eventually will be tested in primates to determine dosage required to confer sufficient resistance. Because of the method by which the genetic changes are engineered, there is almost no potential for reversion to wild type virus. The changes have been shown to be stable in vivo. In principle, this technology can be used to produce a live virus vaccine against any of the arthropod-vectored diseases.