Pathogenesis induced by (recombinant) baculoviruses in insects

<p>Infection of insect larvae by a baculovirus leads to cessation of feeding and finally to the death of the larva. Under optimal conditions this process may take as little as five days during which the virus multiplies approximately a billion times and transforms 30% of the larval weight into viral products. The key question addressed in this thesis is how virus infection spreads in the insect larvae and how the various tissues and organs respond to infection. The answer to these questions may explain the rapid insect pathogenesis induced by baculoviruses and may provide new leads for the specific engineering of the virus to convert it into an even more efficacious insecticide. In order to obtain such information, a unique explorer system, consisting of a recombinant <em>Autographa californica</em> nuclear polyhedrosis virus ( <em>Ac</em> NPV) <em></em> containing two reporter genes, was designed and used in combination with enzyme-histochemical techniques to record baculovirus infection <em>in situ</em> (Chapters 2, 4 and 5). A Lac-Z gene placed behind the constitutive <em>D. melanogaster</em> heat shock 70 promoter, is expressed prior to and independently of the viral replication and can thus report the successful entry of virus into cells and tissues. The GUS-gene, placed behind the late viral p10- promoter is only expressed in cells in which virus replication has occurred. This system provided the unique opportunity to distinguish early and late viral infection processes simultaneously in whole insect larvae. Using this baculovirus explorer, the pathway and pathogenesis of <em>Ac</em> NPV <em></em> infection in larvae of the beet army worm <em>(Spodoptera exigua) is</em> investigated and described in this thesis.<p>After ingestion polyhedra dissolve in the alkaline environment of the larval midgut whereby the rodshaped virus particles are released into the midgut lumen and pass through the peritrophic membrane. The virus particles bind to the microvillar membrane of columnar cells in a receptor-mediated manner (Hortan and Burand, 1993). The viral envelope fuses with this membrane and the nucleocapsids are released into a midgut epithelial cell. The midgut columnar cells and, at low frequency, the underlaying midgut regenerative cells are the primary targets of <em>Ac</em> NPV <em></em> infection (Chapter 2). Some of the parental nucleocapsids are transported directly through the columnar cells and infect nearby regenerative cells. Infected regenerative cells were only found underlaying infected columnar cells. Polyhedra morphogenesis took place in midgut columnar and regenerative cells. It was remarkable that the polyhedra formed in this epithelium were smaller than those formed in other tissues such as the fat body, and that they were often devoid of virus particles. The expression of very late viral genes in this primarily infected tissue may be of interest in the design of engineered virus that can express a gut-specific toxin or protease-inhibitor.<p>Invasion and infection of tissues other than the midgut epithelium was only recorded after the onset of virus replication in midgut epithelial cells. So, virus multiplication in the midgut cells is an essential prelude for secondary and further systemic infection. Our observations do not support the view that the direct passage of parental virus through the midgut basal lamina, as observed by Granados and Lawler (1981), is a biologically significant route for primary infection of underlying tissues. These authors reported that <em>Ac</em> NPV infection of hematocytes in the hemolymph of <em>Trichoplusia ni</em> larvae can be established by parental virus passing directly from the midgut lumen through the epithelium, basal lamina, and associated tissues into the hemocoel. Their conclusion was based on the detection of free (infectious) virions in extracts of the hemolymph at 0.5-2 h p.i., but not supported by evidence for an actual <em>in vivo</em> infection of other tissues. <em>In vivo</em> infection by viruses that directly passed through the midgut epithelium is not likely to occur as these virions lack the characteristics and surface projections of ECVs, including gp64, which are needed to establish secondary infections.<p>In infected midgut columnar and regenerative cells the full complement of late viral structures (polyhedra and fibrillary structures) was observed depending on the larval stage and the age of the midgut epithelium (Chapter 3). Regeneration of the midgut epithelium of early instar larvae is part of the molting process. As part of this process virus infected cells may be rejected into the gut lumen. During molting, virus infection is eliminated (almost) completely from the epithelium. Elimination of infected midgut cells is not the only way to exclude infection from the midgut epithelium. In <em>Se</em> NPV infected <em>S. exigua</em> larvae the columnar midgut epithelial cells degenerate after the virus has been transmitted over the midgut epithelium (data not shown). This type of response provides an alternative explanation why late viral structures are not always found in the midgut of nuclear polyhedrosis virus infected lepidopteran larvae.<p>The requirement of <em>Ac</em> NPV to multiply in the midgut epithelium prior to further infection of other larval tissues sets limits to the speed of action that can be achieved by recombinant viruses designed for more effective insect control. Alternatively, the promoters of the very late viral genes, which are expressed in the midgut epithelium, may be used for expression of hormones and toxins in the early phase of larval infection, thus achieving an enhanced speed of kill if successfully targeted for the underlying tissues.<p>Viral replication in gut epithelial cells is also an important prerequisite to Systemic infection of other animal viruses, such as vesicular stomatitis virus, certain retroviruses, and some adenoviruses (Tyler and Fields, 1990). After infection with these viruses, primary infection and the first round of replication takes place in the gut epithelial cells, whereafter progeny virus buds from the cell at the basal site to infect the cells associated with the basal lamina (submucosa). Cells of the submucosa subsequently multiply the virus before it is transported to other tissues. Similarly, after primary infection of the midgut epithelium, secondary <em>Ac</em> NPV infection of <em>S.</em><em>exigua</em> larvae is also established in cells of the submucosa, i.e. muscle cells, tracheoblasts, and hematocytes (Chapter 2). These secondarily infected cells, associated with the basal lamina, multiply the virus and release vast amounts of particles into the hemocoel amplifying the virus titer in the hemolymph for further systemic infection (Chapter 4).<p>Systemic <em>Ac</em> NPV infection of other larval tissues is established by virus circulating in the hemolymph. Tracheoblasts of all tissues were the first targets for this circulating virus. From these prior infected cells infection radiated out into neighboring tissues. Larval tissues are physically separated from the hemolymph by a thick basal lamina, that forms an effective barrier to direct virus infection (Reddy and Locke, 1990). However, the basal lamina surrounding the tracheoblasts is very thin (or absent) and can apparently be easily penetrated by baculovirus ECVs. Therefore, by infection of tracheoblasts the virus circumvents the basal lamina and is able to establish further infections (Chapter 4).<p>A basal lamina is present around all tissues and plays an important role in preventing easy entry of the virus in tissues of <em>S.</em><em>exigua</em> larvae. The composition of the basal lamina and the mechanisms of transport through this lamina offer possibilities to design more effective recombinant viruses. For instance, degradation of this membrane induced by recombinant viruses may enhance the infection and thus the rate kill by the virus. In principle, the feasibility of this approach was shown by the observation that the underlying tissues could be directly infected after treatment of the basal lamina with dispase (Chapter 4).<p>In the case of <em>Ac</em> NPV infected <em>T. ni</em> larvae, an alternative route for a fast systemic infection was proposed by Engelhard <em>et</em><em>al.</em> (1994). This route involved transport of virus particles through the intercellular space of the tracheal epidermis. In <em>Ac</em> NPV infected <em>S.</em><em>exigua</em> larvae, we observed that the infection of distal tracheal elements progressed only in a cell-to-cell manner, starting in the tracheoblasts nested in the midgut and in other tissues (Chapter 4). The dimensional proportion of the intercellular space of tracheal elements does not favor long distance transport. Therefore effective systemic infection through the tracheal system is unlikely.<p>The mode of virus transport proposed by Keddie <em>et al.</em> (1989) whereby hematocytes infected at the midgut basal lamina near the primary site of infection carry the infection into other tissues, lacks experimental evidence from the <em>Ac</em> NPV- <em>S.</em><em>exigua</em> system. We were unable to find a relation between the infection and replication of the marked AcNPV recombinants in hematocytes and the infection of tissues in contact with these cells (Chapter 4). This observation suggests that infection was not transmitted by hematocytes in a cell- to-cell manner.<p>It has been shown for flavivirus, measlesvirus, and poliovirus that circulation of virus particles in the blood contributes to the spread of infection to different tissues of vertebrates (Tyler and Fields, 1990). Replication of viruses in hematocytes increases the virulence of these viruses (Tyler and Fields, 1990). In insects the hematocytes may boost the levels of ECVs at later times post infection.<p>Clem <em>et al.</em> (1994) described the infection of <em>p35</em> (apoptosis blocking gene) deletion mutants which also lacked the <em>p94</em> gene. Differences in infectivity were observed between orally infected larvae which were killed by both wt- <em>Ac</em> NPV and the dual deletion mutant, and virus injected larvae which were not killed. These authors suggested that tissue specific apoptosis may be responsible for this difference. This can indeed be the case as infection by either route needs different tissues to support the primary round of virus replication (Chapter 4). In orally infected larvae the virus replicates first in the midgut epithelium and then in cells associated with the midgut basal lamina, whereas primary infection of virus injected larvae is predominantly established in the hematocytes.<p>Although the majority of larval tissues, such as midgut columnar and regenerative cells, fat body, tracheal cells, hematocytes, and epidermis, supported full replication of the <em>Ac</em> NPV, this was not the case in some vital tissues such as midgut goblet cells, salivary glands, and Malpighian tubules (Chapter 5). In these tissues early viral gene expression was observed but this was not followed by late viral gene expression. Using immunogold labeling, to detect late viral structural proteins, revealed these proteins only rarely in midgut goblet cells and never in Malpighian tubules and salivary glands. The apparent tissue specificity of <em>Ac</em> NPV infection <em>in vivo</em> is <em></em> not regulated at the level of receptor binding, virus entry, or uncoating of the genome per se, as successful entry of the virus and initial transcription of the genome did take place in these cells as demonstrated by using <em>Ac</em> NPV/HSP-p10. The failure to establish full infection in these tissues must therefore either be regulated at the level of transcription and translation of viral genes or at the level of interaction of viral gene products with host factors. This type of tissue specificity is a rare phenomenon as tissue specificity of viruses is normally thought to be determined by host cell receptors (Tyler and Fields, 1990; Tyler, 1994).<p>Tissue specificity regulated by selective expression or interaction of viral and host proteins is reported for adenoviruses (Doerfler, 1994) and herpes simplex viruses (Aurelian, 1994). Adenoviruses enter cells, initiate early gene expression, and replicate normally in permissive cells. Ad12 infection of BHK21-cells is abortive. Early genes of Ad12 are expressed in BHK21-cells but viral replication and subsequent late gene expression do not occur. This defect can partially be complemented by expression of the Ad5-E1 region in these cells. The Ad12 will then replicate and transcribe its late genes, but synthesis of the late viral protein (and hence progeny virus) does not occur (Schiedner <em>et</em><em>al.,</em> 1994). So in this case there is a transcriptional and translational control of specificity. In the case of herpes viruses the latent or non-reproductive infection is determined by a viral latency factor and/or the expression of the immediate early gene IE1 10, the thymidine kinase gene, and the ribonucleotide reductase gene (Aurelian, 1994).<p>The mechanism determining the tissue specificity of <em>Ac</em> NPV in <em>S.</em><em>exigua</em> larvae may be of a similar nature. Expression of <u>one</u> of the <em>Ac</em> NPV immediate early genes alone may determine this tissue specificity. The <em>Ac</em> NPV- <em>PE38</em> and <em>-ME53</em> immediate early genes do not seem to be involved in this process (Chapter 5). Sequences that regulate the expression of these genes are found in most immediate early <em>Ac</em> NPV promoters and expression patterns homologous to that of <em>Ac</em> NPV pe38 and ME53 may be expected. Thus it is unlikely that <em>Ac</em> NPV specificity is determined through the expression of immediate early genes by host cell factors.<p>Having obtained information on the infection pathway and the tissues involved, effects of the deletion of viral genes can be studied <em>in vivo.</em> In Chapter 6 the effect of deletion of the <em>egt</em> gene from the <em>Ac</em> NPV genome was studied. In larvae infected with this recombinant, Malpighian tubules were found to degenerate at an early stage in the infection. This early degeneration most probably causes the increased speed of kill of <em>S. exigua</em> larvae by the <em>Ac</em> NPV <em>egt</em> deletion mutant. As this degeneration was not observed in uninfected or wild type <em>Ac</em> NPV infected larvae, it can be concluded that, during viral infection, ecdysteroid hormones play a role in the degeneration of vital tissues such as the Malpighian tubules. This early degeneration caused by the infection with <em>egt</em> -deletion mutants is dictated by the level of ecdysteroid hormones in combination with other hormones, by the viral infection, or by the activation of hematocytes by ecdysteroid hormones (Vinson, 1994). This results in a hypersensitive response.<p>Design of a baculovirus containing a dual reporter gene system that allows identification of early and late stages of viral infection allowed us to follow baculovirus infection in the whole insect. Such a marked recombinant also allowed the investigation of tissue specificity and the study of the effect of gene deletions with only very little distortion of the insect. Using this recombinant virus, the most likely pathway of infection of <em>Ac</em> NPV in <em>S. exigua</em> was determined. The virus enters the insect via the midgut columnar cells. At a low frequency parental nucleocapsids can be transported to midgut regenerative cells. The necessity of infection of columnar cells, through which the infecting nucleocapsids are transported, that overlay these regenerative cells is not known. Parental virus does not establish further infection in the insect. Secondary (i.e. systemic) infection occurs only after the virus has replicated in the midgut epithelium. Secondary infection is first established in cells closely associated with the midgut basal lamina close to the primarily infected loci. The secondarily infected cells are muscle cells, tracheoblasts, and hematocytes. How the virus passes the basal lamina is unclear, but it is possible that the infected epithelial cells are not able to maintain this layer. The virus produced in the midgut epithelium starts to circulate in the hemolymph after which infection of other tissues is established. As the virus replicates in, and is released from, the cells associated with the basal lamina, the virus titer of the hemolymph will be enhanced. The circulating virus directly infects the tracheoblasts nested in all tissues. The basal lamina surrounding these cells is thin or absent and does not form a physical barrier for virus passage. In other areas of the tissue this membrane is too thick to be penetrated by virus particles. Transmission of infection by infected hematocytes in a cell-to-cell mechanism does not occur. These infected hematocytes do not pass the basal lamina nor transmit the virus through this membrane. From infected tracheoblasts the infection quickly radiates out into the tissues by cell-to-celI translocation following replication. Infection is not restricted to tissues that support viral replication and subsequent polyhedra morphogenesis, such as fat body, tracheolar cells, epidermis, and hematocytes. ECVs will also enter and initiate early viral gene expression in apparently non-permissive, vital tissues, such as salivary glands and Malpighian tubules. These vital tissues further remain unaffected by the virus. This allows the infected insect to function, resulting in a high production of progeny virus and polyhedra in the insect.<p>The information obtained on the pathogenesis induced by <em>Ac</em> NPV and the developed reporter virus enables us to evaluate the risks of the use of genetically modified baculoviruses. For example, it allows the study of infection in semi-susceptible and apparently non-susceptible insects and may reveal changes of host range by the virus or potential risks for non-target insects. The study of the effect of <em>egt</em> deletion illustrates the further potential the use of these genetically marked baculoviruses to understand the pathogenesis of baculovirus infections in insects