9.1 Who Could Contribute to an Arbovirus Outbreak

9.1.1 Taxonomy and Mosquito Surveillance in Europe

As the spread of mosquito-borne arboviruses is dependent on the presence of a suitable mosquito vector, the knowledge of the mosquito species distribution and vector competence of these mosquitoes belongs to the most crucial factors for estimations about the risk of mosquito-virus emergence to new areas or maintenance of (endemic) arboviruses within particular regions. The first critical issue for mosquito surveillance programs is the exact classification of species (Fig. 9.1).

Fig. 9.1
figure 1

Graphical representation of vector competence assay. The analysis of resident mosquito populations for virus presence and vector competence for the respective virus starts with the collection of mosquitoes (1). Subsequently, the mosquitoes are subjected to morphological taxonomic classification (2) and are pooled according to species and location. Mosquito pools are homogenized to isolate nucleic acids for PCR and proteins for MALD-TOF MS. These data are used for taxonomic confirmation (3) and abundance statistics (4) or virus screening. Virus-positive pools will be used for virus isolation (5) which can then be used for vector competence assays via oral infection (9). To obtain mosquito samples for vector competence assays, eggs of resident mosquito populations are collected (6) and reared in the laboratory (7). From each larval culture, some specimens will be used for taxonomic identification (8). Larvae from the same location and species are pooled, and emerging adult females will be used for vector competence assays. New virus isolates are mixed with blood and fed to 4–7-day-old female mosquitoes (9). After different times of infection, some mosquitoes are sacrificed, and body infection rates (IR), dissemination rates (DR), and transmission rates (TR) will be measured by virus titration (10)

To facilitate detection of different species, several methods have been proposed. Classical morphology is used as the first line of classification. Several keys for morphological discrimination have been published. The morphological characteristics described by Mohrig (1969) and Becker et al. (2010) have been most commonly used for species identification in surveillance programs in Germany. These programs include several university- and organization-driven approaches, some as a part of the European project VBORNET ( or the citizen science project “Mückenatlas” (Walther and Kampen 2017). All those projects have made large progress in redefining the mosquito fauna in Europe and Germany. Especially the “Mückenatlas” project has also proven a very sensitive tool to detect new and invasive species as, for example, several new populations of Aedes japonicus japonicus in North Rhine-Westphalia and Lower Saxony and Aedes albopictus populations in Baden-Wuerttemberg (Kampen et al. 2016a; Werner and Kampen 2013; Werner et al. 2012; Zielke et al. 2014).

Within all programs, the classical morphology has proven a useful tool. However, the accuracy of classical morphological classification is strongly dependent on expert knowledge and the availability of good-quality mosquito specimens. Furthermore, several cryptic species allow only for classification according to male mosquitoes, which are often not attracted by the traps used for surveillance programs. Especially females of the Culex pipiens complex (Fonseca et al. 2004) and the Anopheles maculipennis complex (Kronefeld et al. 2012, 2014; Proft et al. 1999) turned out to be difficult or impossible to distinguish in case of morphologically similar sibling species, such as Culex torrentium and the two Culex pipiens pipiens biotypes pipiens and molestus or mosquito species belonging to the Anopheles messeae/daciae complex. Both species complexes are of major importance for disease transmission: Culex pipiens a main vector for WNV, USUV, or RVFV and Anopheles maculipennis as a potential vector for Plasmodium parasites. Hence, classification methods besides morphology are needed to reach a satisfactory level of species discrimination (Bickford et al. 2007).

The use of morphometric analysis as a qualitative tool for species discrimination has expanded during the past years (reviewed by Lorenz et al. (2017)). In particular wing shape has been used for morphometric comparison in mosquito studies. Wilke et al. (2016) have established a protocol for geometric wing morphometries to identify a broad range of medically important mosquito species belonging to the Aedes, Culex, and Anopheles genera. To do so, 18 landmarks at wing vein intersections were collected from digitalized photographs of female wings. Mosquito genera were classified with 99% accuracy and species even with 100% accuracy, demonstrating the power of the approach (Wilke et al. 2016).

Several other groups also used this method to discriminate female samples of closely related cryptic species. Lorenz et al. (2012) analyzed the same 18 landmarks to distinguish between Anopheles cruzi, Anopheles homunculus, and Anopheles bellator mosquitoes and reached 78–88% accuracy. For the Culex complex, differences in wing venation were already described by Natvig (1948) and Mohrig (1969), who also proposed to use these differences for species discrimination. Especially the vein R2/3 was found informative for differentiation of Culex pipiens and Culex torrentium females. Borstler et al. (2014) used general wing morphology and the R2/3 indices for discrimination of Culex pipiens and Culex torrentium collected in Germany. Their study revealed more than 91% accuracy in the multivariant morphometric analysis using several wing landmarks and 90% correct species identification when only using the R2/3 vein indices. Thus, the morphometric discrimination method has been proven to be a stable and reliable method with success rates of 70–100% for correct reclassification (Lorenz et al. 2017). It is particularly tempting that this morphometric method has been shown to be most accurate in female mosquitoes, the main object of interest in the context of vector-borne diseases.

Although geometric morphometry is a quick and easy to use method, it should be noted that data capturing and identification of landmarks are still a critical issue. Furthermore, in large-scale surveillance programs, a certain degree of automatization of landmark detection and automatic species identification needs to be made, in order to ensure a timely species identification (Lorenz et al. 2017). Thus, molecular methods for large-scale species identification are still needed. In recent years, several advances in the use of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) have been explored to achieve species differentiation. MALDI-TOF MS has been extensively used in bacterial diagnostics (Dierig et al. 2015) and for species identification of Drosophila (Feltens et al. 2010) as well as of relevant vector species such as Culicoides biting midges (Kaufmann et al. 2012), Phlebotomus sand flies (Mathis et al. 2015), and Ixodes ticks (Yssouf et al. 2013a, 2015). Due to the extensive use in diagnostics, a lot of laboratories adjacent to clinics have already implemented MALDI-TOF MS facilities that can easily be used for mosquito surveillance programs. The adaptation of the MALDI-TOF MS for mosquito species identification has made great advances in the past years. Yssouf et al. (2013b) described this technique to analyze samples from tropical areas and were able to establish profiles from 20 mosquito species collected in La Réunion Island and Senegal. In this study, a reliable classification on subspecies level was achieved as demonstrated for the M and S forms of Anopheles gambiae. In total, 100% of the samples were identified correctly after generation of a spectra database. Therefore, this score was set a cutoff value for species identification. However, the method was not considered suitable for mosquito phylogeny yet. To refine the database and to add new species to the collection, Yssouf et al. (2014) conducted a subsequent study using mosquito samples of 11 different species collected at different sites in France and Sweden. After the generation of reference samples based on previous morphological characterization (Becker et al. 2010), 88.5% of the samples were identified correctly. These and other studies (Raharimalala et al. 2017; Schaffner et al. 2014) showed the feasibility and reliability of MALDI-TOF MS for mosquito species identification. Furthermore, the method is usually described as an inexpensive and easily implementable approach. However, a certain degree of instability was recently detected in mosquito samples collected in different countries, highlighting the importance to establish an international database to assure correct mosquito species identification.

Another very sensitive and reliable method for species identification and the differentiation of cryptic species or biotypes is their genetic characterization by conventional and real-time PCR using phylogenetic markers: either chromosomal markers such as the acetylcholinesterase 2 (ace2) gene, the second internal transcribed spacer (ITS2), and the microsatellite locus CQ11 or mitochondrial barcoding based on the cytochrome oxidase I (COI) gene. The use of the ace2 locus as a diagnostic criterion for the differentiation of Culex pipiens complex (Culex pipiens, Culex quinquefasciatus, Culex p. pallens, Culex australicus), Culex torrentium, and Culex pervigilans mosquitoes was developed by Smith and Fonseca (2004). Mosquitoes of these species were collected across the world and subjected to PCR analysis using diagnostic primers located between exons 2 and 3 of the ace2 genetic locus. The ace2 PCR assay was able to distinguish and to detect hybridization events between the mentioned Culex species, for example, hybridization of Culex quinquefasciatus and Culex pipiens, but the two bioforms Culex pipiens pipiens biotype pipiens and Culex pipiens pipiens biotype molestus could not be discriminated. The two biotypes show different feeding patterns and (breeding) habitat preferences with Culex pipiens pipiens biotype molestus being more anthropophilic and adapted to urban habitats, whereas the biotype pipiens is more ornithophilic and adapted to a wide range of natural habitats. These differences in lifestyle can have major impact on their ability to act as vectors for viruses such as WNV. Indeed, Culex pipiens pipiens biotype hybrids have been widely discussed as potential bridge vectors between birds and humans for bird-associated viruses such as WNV. Thus, the correct identification of these two biotypes can be crucial for correct risk assessment. To improve biotype differentiation, Bahnck and Fonseca sequenced microsatellite loci of the Culex pipiens complex and found that the CQ11 locus was suitable for the diagnosis and differentiation of two Culex pipiens biotypes (Bahnck and Fonseca 2006).

The two assays ace2 and CQ11 (Bahnck and Fonseca 2006; Smith and Fonseca 2004) were used to design a multiplex qPCR assay which allows the differentiation of Culex species, biotypes, and biotype hybrids within one reaction (Rudolf et al. 2013). Using a large collection of about 349 morphologically well-defined mosquito specimens (consisting of 227 Culex pipiens biotype pipiens, 3 Culex pipiens biotype molestus, and 119 Culex torrentium samples), the assay was evaluated and revealed 100% specificity for the respective Culex species or biotypes (Rudolf et al. 2013). The analysis of 16,566 Culex samples collected at different trapping sites in Germany with this multiplex qPCR revealed that Culex pipiens biotype hybrids are also present in Germany. Furthermore, the expansion of Culex torrentium in Central Europe was confirmed with more than 50% of the collected specimens containing Culex torrentium at some sample locations in Germany. The same multiplex qPCR method was also used for a surveillance study in the Emilia-Romagna in Italy, which revealed that all (100%) of the 24,165 tested mosquitoes were Culex pipiens and that Culex torrentium was absent at these sample locations (Calzolari et al. 2016). This is in agreement with other studies performed across Europe by Hesson et al. (2014) analyzing 2559 larval samples from 138 collection sites in 13 European countries. This study found Culex torrentium more prevalent than Culex pipiens in Central and Northern Europe but mostly absent in Southern Europe. The study by Hesson et al. (2014) used a different method based on the amplification of the 3′-end of the COI locus, subsequent restriction digest, and sequencing (Hesson et al. 2010) for genetic characterization of Culex pipiens and Culex torrentium. The mitochondrial COI gene is often used for species identification or confirmation of morphological classification. To do so, the 5′ part COI gene is amplified with generic primer sets, and the PCR products are usually sequenced and analyzed (Folmer et al. 1994). According to Hebert et al. (2003), this method is adequate to “barcode” most animal species with an intraspecies variation mostly below 2% and thus allows for reliable intraspecies identification. The COI barcoding has then been used in a large-scale approach such as the International Barcode of Life (iBOL) project creating a reference database BOLD ( In subsequent years, the method had become a standard technique to identify mosquito species from different countries around the world including China (Wang et al. 2012), Pakistan (Ashfaq et al. 2014), Chile, and Sweden (Engdahl et al. 2014). However, in the Swedish study (Engdahl et al. 2014), some inconsistencies between morphological discrimination and barcoding results were observed. Furthermore, the method may cause inconclusive results in closely related species such as species belonging to the Culex pipiens complex. In this case, additional methods such as the restriction analysis of the COI PCR fragment described by Hesson et al. (2010) or additional PCRs for the ace-2 and CQ11 loci (Bahnck and Fonseca 2006; Fonseca et al. 2004; Rudolf et al. 2013) can be advantageous.

The Anopheles maculipennis complex comprises 10–12 Palearctic species (Harbach 2004), and members of the complex have been associated with Plasmodium, Sindbis virus, and Batai virus transmission in Europe (Jost et al. 2010, 2011b; Kampen et al. 2016b). In light of the risk for reintroduction of Plasmodium species by enhanced global travel, identification of potential malaria vectors is of major interest. Therefore, in 1999, Proft et al. (1999) developed a diagnostic PCR method for identification of the members of this complex that are otherwise indistinguishable. This PCR assay was based on the ITS2 region, which had been previously used for differentiation of other complexes (Crabtree et al. 1995; Wesson et al. 1992). The PCR products were sequenced, results were compared with morphological classification, and a stable PCR assay for identification of Anopheles atroparvus, Anopheles melanoon, Anopheles sacharovi, Anopheles maculipennis s. s., Anopheles messeae, and Anopheles labranchiae was established. In the following years, surveillance studies revealed that particularly Anopheles messeae is widespread across Central Europe. However, in a study of Novikov and Shechenko in 2001, it became evident that Anopheles messeae was not a single species (Novikov and Shevchenko 2001) but represents two cryptic species, Anopheles messeae and Anopheles daciae, which was confirmed 3 years later (Nicolescu et al. 2004). To differentiate these cryptic species, the ITS2 assay was refined by the addition of a restriction fragment length polymorphism (RFLP) analysis after ITS2 amplification (Kronefeld et al. 2012, 2014). Also, Weitzel et al. (2012) refined the ITS2 analysis to facilitate Anopheles messeae and Anopheles daciae differentiation by adding a sequencing reaction after initial amplification. However, both protocols are somewhat laborious and prone to contamination. Thus, in 2016, Luhken et al. (2016) described a new multiplex qPCR method to discriminate between the most prominent members of the Anopheles maculipennis complex in Central Europe (i.e., Anopheles maculipennis, Anopheles messeae s.l., and Anopheles atroparvus) and a fluorescence resonance energy transfer (FRET)-based assay to distinguish between Anopheles messeae s.s and Anopheles daciae. As a result of the large-scale study following the establishment of this method, 1445 mosquitoes from Germany were screened, and the superior spread of Anopheles messeae in Central Europe was confirmed with approximately 70% of the samples belonging to this species.

9.1.2 Virus Surveillance in Europe

During the last decade, multiple previously exotic arboviruses that belong to different virus families which may have considerable implications on human and/or animal health have emerged in Europe. Notable examples are mosquito-borne viruses such as CHIKV (Togaviridae) and DENV and ZIKV (Flaviviridae) as well as Culicoides-borne viruses such as Bluetongue virus serotype 8 (BTV-8; Reoviridae) and Schmallenberg virus (SBV, an Orthobunyavirus within the Peribunyaviridae family and Bunyavirales order). Their unexpected emergence—facilitated by globalization and climate change—highlight the risk of future introductions and spread of additional pathogenic arboviruses to Europe such as (1) mosquito-borne Bunyamwera orthobunyavirus (BUNV, Peribunyaviridae), O’nyong-nyong virus (ONNV, Togaviridae), (2) mosquito- and Phlebotomus-borne RVFV (Phenuiviridae), (3) Culicoides-borne Oropouche virus (OROV, Peribunyaviridae), or (4) tick-borne Crimean-Congo Hemorrhagic fever virus (CCHFV, Nairoviridae) (Amraoui and Failloux 2016; Brustolin et al. 2017; Carpenter et al. 2013; Heitmann et al. 2017; Negredo et al. 2017; Rudolf 2015; Tappe et al. 2014). Importantly, the introduction of novel viruses to regions where related (endemic) viruses circulate can result in reassortment and a consequential change in pathogenicity and phenotype (Briese et al. 2013; Rudolf 2015). A large outbreak of hemorrhagic fever in humans was reported in Africa in the 1990s caused by a reassortant of African strains of BUNV and Batai virus (BATV, an infraspecies of BUNV), namely, Ngari virus (NRIV) (Gerrard et al. 2004). Repeated introductions of emerging zoonotic mosquito-borne viruses in addition to CHIKV and ZIKV have been reported in Europe, including DENV and Yellow fever virus (YFV, Flaviviridae) (Húbalek 2008). The detection of genome fragments of JEV, another flavivirus, in Culex pipiens mosquitoes caught in Italy (2010/2011) indicated a repeated introduction or enzootic circulation of JEV or of a related virus in Southern Europe (Cleton et al. 2014; de Wispelaere et al. 2017). A series of repeated disease outbreaks in humans were caused by various zoonotic mosquito-borne viruses endemic or emerging in Europe. In recent years, human virus infections or disease outbreaks were reported in Europe, including chikungunya in Italy (2007) and France (2010, 2014); dengue in Croatia (2010), France (2010, 2013, 2014), and Portugal (Madeira; 2012); usutu in Italy (2009) and Croatia (2013) (Kampen and Werner 2015); and West Nile fever in Austria (2009, 2010, 2014–2016), Croatia (2012–2013), France (2015), Greece (2010–2014), and Italy (2010–2015) (ages 2017; Gossner et al. 2017; Kampen and Werner 2015). WNV is considered endemic in Europe. However, there are several neglected zoonotic arboviruses circulating in Europe that may (at least occasionally) cause disease in humans and animals: BATV, Tahyna virus (TAHV, infraspecies of California encephalitis orthobunyavirus, Peribunyaviridae), SINV (Alphavirus, Togaviridae), and Inkoo virus (INKV, infraspecies of California encephalitis orthobunyavirus) (Eckerle et al. 2018; Húbalek 2008). In general, three groups of mosquito-borne viruses can be distinguished according to their clinical signs: (1) fever-arthralgia-rash (e.g., DENV, CHIKV, ONNV, ZIKV, WNV), (2) affection of the central nervous system (e.g., DENV, ZIKV, WNV), and (3) hemorrhagic fever (e.g., DENV, RVFV) (reviewed by Eckerle et al. (2018); Húbalek (2008); Kampen and Werner (2015)).

9.2 Where Would Competent Mosquito Species Meet Favorable Conditions for Transmission?

9.2.1 Factors for Arbovirus Transmission

Implications for public health emitting from arboviruses depend on key factors that influence vectorial capacity such as barriers to the infection and transmission of arboviruses to mosquitoes (Hardy et al. 1983; Mellor 2009) as well as biotic and abiotic factors that vice versa have an effect on the intrinsic infection barriers. Climate, in particular temperature and precipitation, and ecological factors (in particular land use, anthropogenic disturbance/urbanization) are main abiotic drivers determining the probability of transmission within a given region (Húbalek 2008; Junglen 2016; Kramer 2016). Biotic factors include (1) the susceptibility to infection (e.g., immunogenetics), host diversity, density, behavior, and seasonal abundance of vertebrate hosts (e.g., migration of birds) and their capability to efficiently amplify and transmit the virus to mosquitoes; (2) the vectorial capacity, density, and (opportunistic) feeding preferences concerning the blood source of invertebrate hosts; (3) the genotype and phenotype/pathogenicity of a virus; as well as (4) the interaction, variability, and adaptation between virus genotype x vector genotype and immune system x vertebrate genetics and immune system (Fros et al. 2015a; Kramer 2016; Lambrechts et al. 2009b).

9.2.2 Barriers to the Infection and Transmission of Arboviruses by Mosquito Vectors

Vector competence generally depends on intrinsic factors of the mosquito (Hardy et al. 1983) (Fig. 9.2). After the ingestion of an infectious blood meal, the pathogen has to overcome several barriers in the insect host before transmission with injected saliva during a blood meal on a vertebrate host can occur. The midgut infection barrier (MIB) and midgut escape barrier (MEB) may impede virus passing the midgut cells into the hemocoel (Hardy et al. 1983; Mellor 2009; Mellor et al. 2000). An interference of virus dissemination from the fat cells (dissemination barrier, DB) in the hemocoel is explained by the fact that the fat body plays a role in the insect immune response and prevents infection of other tissues (see section innate immunity below (Mellor 2009)). Further barriers include the salivary gland infection barrier (SGIB), salivary gland escape barrier (SGEB), and transovarial transmission barrier (TOTB) (summarized in Fig. 9.2). Mosquito females that survive the extrinsic incubation period (EIP, the interval between ingestion of a virus and the earliest time at which virus is released in saliva) potentially remain infectious throughout their life (Hardy et al. 1983; Mellor 2009; Mellor et al. 2000). Experimental vector competence studies regularly include analysis of the infection rate (IR), dissemination rate (DR), and transmission rate (TR). Due to the various possible barriers of an insect host, the TR (defined as the number of mosquitoes with virus-positive saliva per number of virus-positive mosquito bodies (Heitmann et al. 2017)) provides the most important information about the vector competence of a mosquito since only virus transmission by saliva during an insect bite is infectious for the vertebrate host. On the other hand, differences in the IR, DR, and TR may give useful information about possible barriers for a certain virus within an insect host. Furthermore, the EIP depends on the invertebrate host-virus interaction and on the ambient temperature (Mellor 2009) (Table 9.1). Mosquitoes that are not (typical) vectors for a given virus may get competent if reared at elevated temperatures, as reported for Culicoides nubeculosus biting midges (a potential vector of BTV). A possible reason is that an increased temperature during the immature stage of the mosquito may compromise the integrity of the gut wall enabling virus to bypass the gut barrier (“leaky gut” phenomenon) (Wittmann and Baylis 2000). In adult mosquitoes, crucial differences in vector competence of Aedes albopictus for ZIKV depending on the ambient temperature have been demonstrated by Heitmann et al. (2017). German and Italian populations of Aedes albopictus that were infected with ZIKV and kept at 18 °C were not found competent for ZIKV transmission (TR of 0%), while a TR of 18–20% was found in Aedes albopictus kept at 27 °C after an EIP of 14 days. In contrast, none of the Culex pipiens biotype pipiens, Culex pipiens biotype molestus, and Culex torrentium populations were found competent at 18 or 27 °C. Similar results were reported for Italian and French Aedes albopictus populations (TR of 4–29%) and for an Italian Culex pipiens (TR of 0%) population kept at 26 or 28 °C (Boccolini et al. 2016; Di Luca et al. 2016; Jupille et al. 2016). The midgut barriers (MIB and MEB) may be circumvented by using intrathoracic instead of oral infection. Intrathoracically infected mosquitoes show a considerable higher IR and TR (up to 100%) than orally infected mosquitoes as demonstrated for USUV (TR of 69%, Culex pipiens) and WNV (TR of 22–33%, Culex pipiens) (Fros et al. 2015a, b). This can lead to overestimation of IR and TR and consequently to misleading interpretation of vector competence (Fros et al. 2015a, b). Fu et al. (1999) suggested that, following intrathoracical inoculation, virus levels in the hemocoel exceed the virus amount that can be cleared by fat bodies. Another important factor for efficient infection of mosquitoes is the orally ingested virus dose. Only vertebrate species that produce viremia (sufficiently high for infection) can be regarded as amplifying hosts (Húbalek 2008) as it is the case for WNV in birds, but not for WNV in horses or humans (Angenvoort et al. 2013; Bunning et al. 2002; Hayes et al. 2005). When Culex quinquefasciatus mosquitoes are experimentally infected with a low (10^4 plaque-forming units per mL (PFU/mL)) or a higher (10^6 PFU/mL) dose of ZIKV, only mosquitoes that ingested the higher dose got infected (Guedes et al. 2017). In contrast, a considerably higher TR was found in Aedes vexans originating from a German colony infected with an avirulent RVFV strain (Clone 13) (TR of 25%) compared to a virulent RVFV strain (ZH548) (TR of 8.3%) (Moutailler et al. 2008).

Fig. 9.2
figure 2

Intrinsic barriers to infection and transmission of arboviruses in mosquitoes. The midgut infection barrier (MIB), midgut escape barrier (MEB), dissemination barrier (DB), salivary gland infection barrier (SGIB), salivary gland escape barrier (SGEB), and transovarial transmission barrier (TOTB) are potentially interfering with infection, dissemination, and transmission of viruses after the ingestion of an infectious blood meal by a mosquito. Only virus release in the saliva and transmission by bite of a vertebrate host confirm the completion of the extrinsic incubation period (EIP) and vector competence of a mosquito (Adapted from Mellor et al. 2000; Hardy et al. 1983)

Table 9.1 Summary of experimental studies of vector competence for ZIKV, DENV, CHIKV, USUV, RVFV, JEV, and WNV of mosquito species collected in Europe (data of positive control mosquito vectors used in the studies are not shown)

In European mosquito populations, transovarial transmission has only been investigated by Fortuna et al. (2015b) in four different Culex pipiens populations collected in Italy and experimentally infected with WNV. However, vertical transmission could not be confirmed in their offspring, although all four populations showed similar TR in their saliva (TR of 37–47%) and were therefore vector competent (Fortuna et al. 2015a) (Table 9.1). In contrast, transovarial transmission was found for WNV in Culex vishnui in India (Mishra and Mourya 2001) and for an insect-specific flavivirus (Culex flavivirus) by American Culex pipiens (Saiyasombat et al. 2011). Bagaza virus (Flaviviridae) was transovarially transmitted by Culex tritaeniorhynchus from India, but not by Aedes aegypti and Culex quinquefasciatus mosquitoes (Sudeep et al. 2013). Various studies in non-European countries confirmed the possibility of natural transovarial transmission by Aedes aegypti for different viruses such as DENV and ZIKV by analysis of immature mosquito stages (Gutiérrez-Bugallo et al. 2017; Li et al. 2017; Velandia-Romero et al. 2017). A high percentage of transovarial transmission of DENV (54.7% of immature stages in households) together with the possibility of transmission by the vector without a prior blood meal has been suggested a possible explanation for the persistence of DENV in (rural) areas (Velandia-Romero et al. 2017). However, the impact of transovarial transmission for DENV in other regions was found negligible, scrutinizing the elimination of larvae as intervention methods (Angel et al. 2016). On the other hand, elimination of larvae is not considered a powerful method for vector control (Pfeffer 2015) (see section vectorial capacity).

9.2.3 Vectorial Capacity

The vectorial capacity (VC) is defined as the efficiency of a mosquito species to serve as a vector for a given pathogen and can be estimated using calculations of the basic reproductive rate (R 0). VC is an entomological restatement of R 0 of a pathogen (Kramer 2016; Schaffner and Mathis 2014). R 0 is defined as the number of secondary infections expected to occur from the introduction of a single infection in a naïve population (Kramer 2016), and a key method to understand disease transmission. A major epizootic outbreak and spread of disease within a population are expected if R 0 > 1, while minor disease outbreaks that become extinct are expected if R 0 < 1. R 0 can be used to plan strategies for control of epizootics but also to estimate, quantify, and compare the outcome of control measures (Pfeffer 2015; Weesendorp et al. 2011). Out of different published equations, the following was proposed by Kramer (2016) and Pfeffer (2015):

$$ \overline{\mathrm{x}} $$

VC, vectorial capacity (R 0)

m, vector density in relation to the vertebrate host

a, probability that vector feeds on a host in 1 day (i.e., host preference index * feeding frequency)

p, probability that vector survives one day

t, duration of extrinsic incubation period (EIP) in days (latency period)

IR, infection rate (proportion of vectors infected after feeding on an viremic host)

TR, transmission rate (proportion of infected vectors that are able to transmit the virus to a host)

(IR * TR), vector competence (proportion of vectors ingesting an infective blood meal that are later able to transmit the infection to a host)

1/−ln (p), duration of the vector’s life in days after surviving the EIP (recovery rate)

Accordingly, viral factors are of major importance: a rapid dissemination of a virus from the midgut to the salivary glands would reduce the EIP and, hence, at the same time prolong the duration of the vector’s life after surviving the EIP (=1/−ln (p)). In contrast, host feeding (a), vector longevity (p), and EIP (t) would have a more powerful impact on VC (as square or component), while the vector-to-host density relation (m) and vector competence (IR * TR) of a mosquito population would have a linear and therefore weak effect on VC (Kramer 2016).

The control of malaria (caused by parasitic Plasmodium spp.) is a vivid example to demonstrate the power or weakness of different control strategies. Control of mosquito larvae affects the vector-host proportion, but a reduction of larvae (m) by 50% only results in a 50% reduction of the VC. However, a reduction of the daily survival time of mosquito vectors of Plasmodium (p) by 50% results in a 1000 times lower proportion of mosquitoes that transmit malaria since a reduction of p (survival time) has a direct effect on EIP (t) and the recovery rate (1/−ln (p)) (Pfeffer 2015 ).

9.2.4 Outcome of Experimental Vector Competence Studies by Virus Species CHIKV

Aedes albopictus , one of the most invasive mosquitoes now endemic across southern Europe, was the main vector for the initial CHIKV outbreak in Italy in 2007 (Bonilauri et al. 2008). Aedes aegypti, another primary vector of CHIKV, was introduced in Madeira (Portugal) in 2005 (CDC 2017; Sigfrid et al. 2017). Further autochthonous chikungunya outbreaks were reported in France in 2010 and 2014 (Delisle et al. 2015; Gould et al. 2010). The risk of CHIKV introduction and spread in Europe are highlighted by recent autochthonous outbreak of chikungunya in Italy and spread to France in 2017 (CDC 2017). Bioassays for vector competence studies have been conducted with four different Aedes albopictus field populations collected in Italy (n = 2) and Corsica, France (n = 2) (Moutailler et al. 2009; Talbalaghi et al. 2010). In both experiments, mosquitoes were infected with a CHIKV strain from the island La Réunion and kept at 28 °C. The TR was approximately between 10 and 80% up to 100% (Moutailler et al. 2009; Talbalaghi et al. 2010). These results are similar to the TR (61%) measured in a US Aedes aegypti strain infected with another CHIKV isolate (Blagrove et al. 2016). However, the latter experiment was conducted at a considerably lower temperature (21 °C). In comparison to the main vectors of CHIKV, the mosquito species Aedes detritus endemic to the UK, and a possible vector of JEV, RVFV, and WNV (Table 9.1), was CHIKV-infected and kept under the same experimental settings as Aedes aegypti (Blagrove et al. 2016). In contrast to Aedes aegypti, Aedes detritus was not susceptible to CHIKV infection, at least in this experimental setting (Blagrove et al. 2016). However, higher temperatures during the infection experiment or during the maturation of insects may affect their vector competence (Kramer 2016; Lourenço-de-Oliveira et al. 2013; Mellor 2009) for CHIKV. Hence, further vector competence studies are needed for abundant European mosquito species such as Culex pipiens and Aedes vexans to analyze their vector competence for CHIKV. A study of dissemination rates (DR) in Italian populations of Anopheles maculipennis (0%), Aedes vexans (7.7%), and Culex pipiens (0–33%) after CHIKV infection showed low susceptibilities suggesting a negligible role of these European mosquito species for CHIKV transmission (Talbalaghi et al. 2010 ). DENV

A large increase in dengue fever cases has been experienced around the globe in the past decades. Between 2010 and 2014, repeated sporadic or large outbreaks have been reported in over 20 European countries (Kampen and Werner 2015; Sigfrid et al. 2017; WHO 2017). Aedes aegypti and Aedes albopictus are considered the two main vectors of DENV. Infection by one of the four DENV serotypes (DENV-1 to DENV-4) only mediates partial and temporary cross-immunity. Even more, additional infections with other serotypes can lead to severe dengue (Dejnirattisai et al. 2010). Despite to permanent risk of DENV introduction to Europe, only a few studies on the vector competence of European mosquito species have been conducted. One study uses British Aedes detritus and tropical Aedes aegypti mosquitoes for DENV infection and kept the mosquitoes at a low ambient temperature of 21 °C and 70% RH after infection to simulate low temperate temperatures of Great Britain (Blagrove et al. 2016). Similar to the results of CHIKV infection in these mosquito strains, Aedes detritus was not susceptible to DENV-2, while Aedes aegypti showed a high TR of 70%. Talbalaghi et al. (2010) and Moutailler et al. (2009) investigated dissemination rates (DR) of Italian and Corsican (France) Aedes albopictus populations after infection with DENV-2, but not TR. Italian Aedes albopictus (14–39%) and Corsican Aedes albopictus (13–69%) kept at 28 °C for 14 days showed similar DR. Because of intrinsic barriers in the mosquito potentially interfering with transmission, the TR as a proxy for vector competence is not necessarily similar to DR. Thus, vector competence for European mosquito populations of Aedes albopictus is not confirmed yet, but is likely considering the global role of Aedes albopictus as vector of DENV. Further studies are needed to investigate the vector competence for various potential European mosquito vectors and the four DENV serotypes. JEV

JEV is an exotic flavivirus to Europe. However, recent detection of fragmented JEV-RNA in Italian Culex pipiens mosquitoes and birds caught in 2010 indicated a sporadic introduction of JEV to Europe, although complementary studies to confirm the presence of JEV in Europe are required (Platonov et al. 2012; Ravanini et al. 2012; Zeller 2012). Several groups therefore aimed to investigate the vector competence of mosquito species endemic (Aedes detritus and Culex pipiens) or invasive (Aedes albopictus, Aedes japonicus japonicus) to Europe. While Aedes albopictus by now commonly occurs in large parts of Europe (in particular in Southern Europe), Aedes japonicus japonicus occurs considerably less frequent in Europe. However, this mosquito species is adapted to temperate regions, has been established in a few regions of Germany since 2008 (Kampen and Werner 2015), and was shown competent for JEV replication (Huber et al. 2014a). All four European mosquito species—Aedes detritus collected in the UK, Culex pipiens and Aedes japonicus japonicus collected in Germany, as well as Aedes albopictus collected in France that were orally infected with JEV strains of genotype 3 or 5 (Table 9.1)—were found competent for JEV transmission (de Wispelaere et al. 2017; Huber et al. 2014a; Mackenzie-Impoinvil et al. 2015). De Wispelaere et al. (2017) used two cDNA clones of field strains after their rescue in cell culture, while all other groups used field strains. Aedes albopictus, Aedes japonicus japonicus, and Culex pipiens species were kept at 25 or 26 °C and 80–85% RH, simulating intermediate to diurnal summer temperatures of Mediterranean Europe. TR ranged between 12 and 63% for Aedes albopictus and Culex pipiens. For Aedes japonicus japonicus, only the DR in the whole head (analyzed by PCR) was investigated, which was considerably higher (100%) (Huber et al. 2014a) compared to the TR of JEV found for the other mosquito species. Therefore, the high DR cannot necessarily be used to draw conclusions for the TR, which requires analyses of saliva or at least salivary glands (see section barriers). The study was included in this review since no other studies of vector competence for JEV in European Aedes japonicus japonicus mosquito populations have been conducted so far. The vector competence of local (temperate) British Aedes detritus mosquitoes was comparatively analyzed using 23 or 28 °C and a RH range of 70–90%. Interestingly, Aedes detritus mosquitoes were found competent at both temperatures, although the RT was markedly lower at 23 °C (TR of 3%) compared to 28 °C (TR of 17%). Interestingly, similar TRs were obtained for Culex quinquefasciatus, a tropical mosquito previously incriminated as vector for JEV (Mackenzie-Impoinvil et al. 2015). In summary, the results of the vector competence studies of JEV in three commonly occurring mosquito species in Europe suggest that JEV transmission is possible in various European countries especially during warm summer nights and in Mediterranean Europe. Complementary studies are necessary to determine the vector competence of different Aedes japonicus japonicus populations invasive in Europe for JEV. The results of the vector competence studies together with the recent detection of fragmented RNA of a JEV or a related virus highlight the need for comprehensive surveys of JEV in different mosquito species in Europe. RVFV

RVFV is an arbovirus mainly transmitted by a large number of different mosquito species to different mammals including humans in Africa. Multiple outbreaks of RVFV outside Africa, particularly in countries bordering the Mediterranean Sea, point to a high probability of RVFV outbreaks in Europe. Key drivers of seasonally high numbers of RVF disease outbreaks are heavy rainfalls following periods of drought that suddenly increase vector density (due to rain associated hatching of larvae to imago). The high vector density at water holes leads to a high probability of infection of susceptible vertebrate hosts that regularly visit water holes for drinking. The possibility of transovarial transmission of RVFV to the mosquito offspring as reported by Linthicum et al. (1985) contributes to efficient transmission of this virus (Brustolin et al. 2017; Moutailler et al. 2008).

Vector competence studies for RVFV in European mosquito species are scarce. Oral infection of Spanish Aedes albopictus, Culex pipiens biotype molestus, and hybrid Culex pipiens biotype pipiens x molestus with an South African RVFV strain resulted in the release of infectious virus transmission in saliva of a few individuals belonging to the species Aedes albopictus and the hybrid Culex pipiens biotype pipiens x molestus (exact proportion of the TR was not given) but not of the species Culex pipiens biotype molestus (Brustolin et al. 2017). The midgut barriers of infection (MIB) and escape (MEB) were comparatively analyzed in the species Culex pipiens biotype molestus and the hybrid species by virus isolation. Two different viral doses were used for oral infection (5.7 log10TCID50/mL or 5.7 log10TCID50/mL). Interestingly, while the lower and higher doses resulted in infection of the MIB in both species (IR of 7–20%), the MEB was only overcome in hybrid Culex pipiens biotype pipiens x molestus after infection with the higher virus dose (DR of 66.6%), but not in the species Culex pipiens biotype molestus (0%). A similar dependence of the viral dose on the infection and escape of midgut cells was previously reported for BTV in Culicoides (Mellor 2009). On the other hand, Culex pipiens biotype molestus is generally refractory to infection with various other viruses (WNV lineages 1 and 2, and ZIKV) (Brustolin et al. 2017; Heitmann et al. 2017) (Tables 9.1 and 9.2). Moutailler et al. (2008) studied various European mosquito species regarding their vector potential for RVFV by analyzing virus in head squashes by immunofluorescence assay, and hence the DR but not TR. At 14 days postinfection, Aedes vexans showed a considerably lower DR in virulent RVFV (ZH548, 8.3%) compared to an avirulent strain (Clone13, 25%) (Moutailler et al. 2008). In contrast, for the three European mosquito species, namely, Aedes detritus, Culex pipiens (France), and Culex pipiens (Cyprus) infected with both RVFV strains, DR were markedly higher after infection with the virulent ZH548 (13–30%) compared to the avirulent Clone 13 strain (0–14%) (Table 9.1). In the French colonies of Aedes caspius (7%) and Culex pipiens (9%) infected with the avirulent Clone 13 RVFV strain, DR were similarly low (7 and 9%, respectively) (Moutailler et al. 2008). In addition to vector competence of the European mosquitoes, results were compared with field strains of different Aedes and Culex species from different African and Asian countries. In general, similar dissemination of the virus is found in all tested species compared with the DR results of the European mosquito species, except for Aedes aegypti. Aedes aegypti showed a considerably higher DR of 20–90% for the virulent RVFV ZH458 and 24–73% for the avirulent Clone 13 strain suggesting that transmission of RVFV by Aedes aegypti is more efficient (Moutailler et al. 2008). On the other hand, mosquitoes belonging to the Culex pipiens complex are considered efficient vectors of RVFV in Africa, and virus isolation of RVFV from at least 40 mosquito species (Moutailler et al. 2008) indicates that the broad variety of competent vectors of RVFV primarily contributes to the efficient transmission of this virus in highly diverse habitats and climatic regions. The demonstration of vector competence of Spanish field populations of Culex pipiens and Aedes albopictus for RVFV and the potential vector competence of other European mosquitoes indicate that autochthonous outbreaks of RVFV are possible in Southern Europe.

Table 9.2 Vector competence of different mosquito-vector species for endemic and emerging pathogens in Europe Usutu Virus

In a comprehensive field study of USUV infection in different mosquito species in Italy from 2009 to 2012, a substantial incidence of Aedes albopictus mosquitoes PCR-positive for USUV was found. However, USUV was not detected in any of the Aedes albopictus specimens collected in 2013 (Puggioli et al. 2017). Experimental infection of Aedes albopictus collected in the field in Italy with any of the three Italian virus strains (of 2011) and incubation at 28 °C and 80% RH showed RNA in a single individual after an EIP of 7 days, but no mosquitoes were found PCR-positive after an EIP of 14 days. Therefore, Puggioli et al. (2017) suggested that Aedes albopictus plays a negligible role in the epidemiology of USUV, but further studies are necessary using different experimental parameters. In contrast, Culex pipiens orally infected with USUV strain Bologna/09 showed a high vector competence (TR of 69%) at an EIP of 14 days at 28 °C and 60% RH, which is significantly higher compared to TR found for Culex pipiens infected with WNV lineage 2 strain Gr-2010 (TR of 33%) by the same group (Fros et al. 2015b). A considerable dependence on temperature was found comparing infection rates of Culex pipiens mosquitoes kept at 60% RH and the three different temperatures 18 °C (TR of 11%), 23 °C (TR of 53%), and 28 °C (TR of 90%). Since these three different temperatures represent the mean diurnal summer (July–August) temperature in North-Western Europe, an intermediate temperature, and the mean diurnal summer temperature for Mediterranean Europe, respectively, it can be assumed that particularly in Southern Europe, the transmission rate of USUV by Culex pipiens is considerably higher (Fros et al. 2015b). In a comprehensive field study of USUV occurrence in different mosquito species in Germany, USUV was detected or isolated from Culex pipiens (Jost et al. 2011a; Sieg et al. 2017). In field studies in Italy (Calzolari et al. 2012; Mancini et al. 2017) and other countries (reviewed in Nikolay (2015)), additional mosquito species were found PCR-positive for USUV, including Culex pipiens s.l., Aedes albopictus, Aedes caspius, Aedes detritus, Anopheles maculipennis, and Culiseta (Cs.) annulata. Similar to the results of the German studies (Jost et al. 2011a; Sieg et al. 2017), the cumulative results of the Italian field studies confirm that Culex pipiens likely is most involved in USUV circulation in Italy (Calzolari et al. 2012; Mancini et al. 2017) and in other European countries. West Nile Virus

In Europe, Culex pipiens is considered the main vector of WNV, but other species such as Aedes albopictus (Fortuna et al. 2015a), Aedes detritus, or Aedes japonicus japonicus (Wagner et al. 2018) may also act as competent vectors. Therefore, several research groups investigated the vector competence of these mosquito species in comparison to the main European vector Culex pipiens for WNV lineage 1 and 2 strains by using field and laboratory mosquito colonies collected in different European countries. Huber et al. (2014a) did not find replication of North American WNV lineage 1 strain NY-99 in a German Aedes japonicus japonicus population after artificial infection, while Wagner et al. (2018) found the Aedes japonicus japonicus populations collected in the neighboring country Switzerland susceptible for the same WNV strain and the Italian strain Italy/2009/FIN. Aedes detritus, a mosquito species endemic in the UK, were kept at 21 °C and 70% RH (according to climatic conditions in the UK during warmer seasons) during the experiment and were found competent for WNV strain NY-99 infection under these conditions (Blagrove et al. 2016). As expected, Culex pipiens endemic in Switzerland were found competent for the replication of WNV strain NY-99 (Wagner et al. 2018). A comparison of vector competence for European WNV lineages 1 and 2 strains was conducted by Brustolin et al. (2016) and Fros et al. (2015b). In contrast to other studies, Brustolin et al. (2016) used a fluctuating temperature regimen (mean of 21.3 °C at night and mean of 27.7 °C during the day, at 70% RH) to mimic natural conditions. For the comparative study of WNV line 1 and 2 strains, Aedes albopictus, Culex pipiens pipiens biotype molestus, and Culex pipiens pipiens hybrids of biotypes pipiens and molestus were collected in the field in Spain and orally infected with European WNV lineage 1 (France 2001) or 2 (Italy 178907/2013). The Culex pipiens hybrid was competent for lineage 2 but refractory to WNV lineage 1 (Brustolin et al. 2016). In contrast, Aedes albopictus was found competent for both strains (Brustolin et al. 2016). Similarly, a field colony of Aedes albopictus collected in Italy and orally infected with the European Sardinia 2011 lineage 1 strain Ma V3 kept at 27 °C and 70% RH showed a high vector competence (TR of 50%) (Fortuna et al. 2015b). A possible reason for a broader vector competence, more efficient transmission of arboviruses, and outbreak establishment might be that Aedes albopictus has a higher genetic variability due to independent and transcontinental introductions (Manni et al. 2017), which could therefore facilitate the adaption of this mosquito species to different regions and climates. Considerable genomic variations in Aedes japonicus japonicus due to similar reasons were also suggested by Kampen and Walther (Kampen and Werner 2014; Zielke et al. 2014, 2015, 2016). Fros et al. conducted vector competence studies with a laboratory colony of Culex pipiens collected in the Netherlands. After infection with the WNV lineage 1 strain NY-99 and the European lineage 2 strain Gr-2010 and maintenance at 23 °C (mean average temperature in Central Europe) and 28 °C (Mediterranean mean diurnal summer), a similar vector competence for both lineages and a slightly higher transmission rate at a higher temperature (TR of 33% compared to 24%) were found (Fros et al. 2015a, b). Interestingly, the vector competence and dissemination rate of these North-West European Culex pipiens was similarly high for both the NY-99 and Gr-2010 strains at 23 °C, while mosquitoes of North American origin infected with the same strains showed a significantly lower transmission rate for the WNV lineage 2 strain (Fros et al. 2015a). Unfortunately, the biotype of the Culex pipiens was not described to evaluate whether these mosquitoes were hybrids that may inherit a higher vector competence compared to Culex pipiens biotype molestus as described by Brustolin (Brustolin et al. 2016). Zika Virus

ZIKV has been circulating in Africa and South-East Asia for over 65 years. However, during the recent ZIKV endemic in the Americas, this Asian ZIKV genotype has been linked to different phenotypic characteristics (including congenital malformation and neurological disorders in humans, higher infection rates in Aedes aegypti) compared to the African ZIKV genotype (Willard et al. 2017). A risk analysis of Gardner et al. (2017) revealed that the vector status of Aedes species determines geographical risk of autochthonous ZIKV establishment. While the risk is geographically limited if Aedes aegypti is the only competent ZIKV vector, vector competence of Aedes albopictus would pose a risk of local establishment in all American regions including Canada and Chile, much of Western Europe, Australia, New Zealand, and South and East Asia, with a substantially increase in the risk of ZIKV outbreaks in Asia (Gardner et al. 2017). To estimate the risk of different mosquito species in different climatic regions, European Aedes albopictus were collected from the field in Italy, France, and Germany and experimentally infected with ZIKV belonging to the Asian genotype. Aedes albopictus were found competent at temperatures between 26 and 28 °C, but refractory to ZIKV at 18 °C (Di Luca et al. 2016; Heitmann et al. 2017; Jupille et al. 2016). In contrast, Culex pipiens collected in Italy and kept at 26 °C and 70% RH (Boccolini et al. 2016) as well as Culex pipiens biotype molestus and biotype pipiens and Culex torrentium collected in Germany incubated at 18 or 27 °C and 80% RH (Heitmann et al. 2017) were not found competent vectors of the Asian ZIKV genotype .

9.2.5 Lessons Learned by Experimental Vector Competence Studies

In summary, the varying results of the research groups regarding the proportion of mosquitoes of the same species that were found competent for WNV transmission (Table 9.1) may be due to considerable variations in specific mosquito genotype and virus genotype interactions (Lambrechts 2010; Lambrechts et al. 2009a). A considerable genetic variability in Aedes albopictus and Aedes japonicus japonicus due to independent and transcontinental introductions (Kampen and Werner 2015; Manni et al. 2017) can result in a broader vector competence, more efficient transmission of arboviruses, and regional outbreaks. A meta-analysis of laboratory experiments with DENV indicated that colonization of Aedes albopictus over a few generations might result in an increase of their susceptibility to DENV infection (Lambrechts 2010).

On the other hand, the effect of virus genotypes, serotypes, or lineages may be underestimated or overestimated regarding virulence and transmissibility for different mosquito populations of the same species. Vertebrate host factors such as differences in resistance to infection or low viremia may considerably impact virus transmission between hosts (Húbalek 2008; Reisen and Hahn 2007). Adaptation of new viruses to local hosts and vectors by initial positive (diversifying) selection with more virulent quasispecies, followed by negative (stabilizing) selection driven by strong evolutionary constraints, is reported for BTV (Boyle et al. 2012, 2014; Maclachlan et al. 2009; Schulz et al. 2016). For example, Culex pipiens populations occurring in North America showed a significantly lower transmission rate for a WNV lineage 2 strain compared to North-Western European Culex pipiens species, while a similar transmission rate was found for WNV lineage 1 (Fros et al. 2015a). However, even specific combinations of isofemale families and viral isolates may affect quantity of dissemination within mosquito vectors (Lambrechts et al. 2009a), challenging the validity and relevance of laboratory experiments with single virus-mosquito combinations (Lambrechts et al. 2009a). Furthermore, differences in mortality rates of virus-infected mosquitoes might be due to virus factors (see section virus adaptation to mosquitos). Aedes albopictus infected with CHIKV died a few days earlier than non-infected mosquitoes, while the primary vector Aedes aegypti survived the infection due to antiviral immune response (see section immune response against arboviruses). A higher frequency of cytopathological changes in salivary glands has been reported in WNV-infected mosquitoes (Girard et al. 2007). Furthermore, fast virus dissemination from the midgut impacting the duration of EIP, low mortality rate, and differences in feeding behavior influence the vectorial capacity of a vector (see section vectorial capacity). Interestingly, Aedes aegypti infected with DENV showed a significantly prolonged probing time and enhanced feeding frequency (Platt et al. 1997).

Therefore, the vector competence of various vector genotype and virus genotype combinations by studying different populations over time and space (from different regions/countries of interest) may result in an average and collective experience to allow an estimation of vectorial capacity of a mosquito species from different areas and over time (Fonseca 2016). In addition, a bias in results of vectorial capacity due to variations in methodologies used by different research groups may be mitigated by analyses of similar virus genotype and mosquito genotype combinations (Lambrechts 2010; Lambrechts et al. 2009a). On the other hand, harmonization of methods (e.g., temperature regimes) and analyses of experiments (representation of proportions of transmission rates by species) as well as the meticulous description of the origin and taxonomy of the used mosquito vectors and virus strains would be most valuable in terms of comparability and reproducibility. In a considerable number of studies, Culex pipiens was only superficially taxonomically classified. However, comparison of results of experimental infection of the Culex pipiens biotype pipiens, Culex pipiens biotype molestus, and hybrids of both forms revealed considerable differences in their susceptibility to different virus species and lineages (RVFV, WNV lineages 1 and 2, and ZIKV) under equal or similar experimental conditions (Brustolin et al. 2016, 2017; Heitmann et al. 2017) (Table 9.1) insofar that the parental forms molestus and pipiens of Culex pipiens seem to be refractory to the so far tested viruses (Table 9.2), while hybrids of Culex pipiens biotype pipiens and molestus were found competent for RVFV and WNV lineage 2 (Tables 9.1 and 9.2).

Change in climate, land use, genetic diversity within mosquito species in combination with a rapid arboviral adaptation to alternative mosquito, and vertebrate hosts constitute a dynamic system that can substantially and rapidly change the epidemiological patterns of a viral disease as well as the disease expression in vertebrate hosts and therefore the impact on animal welfare and economy of affected countries (Kramer 2016; Lambrechts 2010; Schulz et al. 2016 ).

9.3 How Do Viruses and Vectors Interact to Facilitate Transmission?

Arboviruses can efficiently replicate in evolutionary distinct hosts, such as mosquitoes and humans; yet they seem to depend on specific mosquito vectors for transmission. The intrinsic factors that determine whether a specific mosquito can transmit a given virus (vector competence) remain poorly understood. Major factors defining vector competence of mosquito species are (1) the control of viral replication by the mosquito to an extent that the mosquito itself is not affected by the virus, (2) virus adaptation to the mosquito to increase viral replication, and (3) the microbiome in the insect vector (illustrated in Fig. 9.3). Within this part, a brief overview of these factors will be given, but since these factors are subject to intense research these days, not all details can be given in the frame of this chapter. For more detailed information please refer to recent reviews (Blair and Olson 2015; Donald et al. 2012; Johnson 2015; Sim et al. 2014).

Fig. 9.3
figure 3

Intrinsic factors that interfere with the vector competence of mosquitoes. The vector competence of a certain mosquito species is characterized by several factors: firstly, the ability of the virus to overcome the midgut barrier (midgut infection and escape barrier), secondly the ability of the virus to replicate in various tissues of the insect host, and most importantly the efficient dissemination of infectious viral particles to the saliva (salivary gland infection and escape barrier). The virus replication in the mosquito midgut is regulated by the gut microbiota represented by the Wolbachia endosymbiont (1) which can interfere with the virus replication by various ways including immune priming and completion for resources. After the dissemination of the virus to different mosquito organs such as the fat body and endothelial cells, the virus starts to replicate in these different tissues. The virus replication can trigger several antiviral pathways such as the RNAi pathways represented by Dicer cleavage and the inducible immune responses represented by the induction of Vago (2)

9.3.1 Immune Response in Insects Against Arboviruses RNAi Responses

Lacking an adaptive immune system, insects depend on different immune mechanisms for antiviral defense. Using the model insect Drosophila melanogaster, it has been demonstrated that RNA interference (RNAi) pathways are crucial to control various Drosophila viruses and also metazoonotic viruses such as SINV, WNV, v Vesicular stomatitis virus (VSV), and DENV (Chotkowski et al. 2008; Galiana-Arnoux et al. 2006; Mukherjee and Hanley 2010; van Rij et al. 2006; Wang et al. 2006; Zambon et al. 2006). The exogenous (antiviral) siRNA pathway (exoRNAi) is initiated by recognition and cleavage of long double-stranded (ds) RNA, deriving from viral replication intermediates or secondary RNA structures in viral genomes, by the RNaseIII enzyme Dicer-2 (Dcr-2). The resulting 21 nucleotide (nt)-long virus-derived small interfering RNAs (viRNAs) are then subjected to a multiprotein RNA-induced silencing complex (RISC). In this complex, the major component Argonaute-2 (Ago2) together with one strand of the viRNA initiates the sequence-specific degradation of viral genomes or transcription products (Liu et al. 2006; van Rij et al. 2006). Survival experiments in Drosophila lacking the key components of a functional exoRNAi response have demonstrated the exoRNAi-mediated control of arbovirus replication is crucial for the insects’ survival (Dietrich et al. 2017a; Kemp et al. 2013; Mueller et al. 2010; Mukherjee and Hanley 2010). The sequencing of full genomes of Aedes aegypti (Nene et al. 2007), Culex quinquefasciatus (Arensburger et al. 2010), and Anopheles gambiae (Holt et al. 2002) enabled the identification of orthologues of Dcr-2 and Ago2 in three important vector mosquito species (Campbell et al. 2008a) and subsequent description of further Dcr-2 and Ago2 orthologues in more vector species such as Aedes albopictus (Brackney et al. 2010). Furthermore, the production of viRNAs, a hallmark of exoRNAi pathway induction, has been shown in Aedes and Culex mosquitoes response to infection of mosquitoes with different arboviruses (Blair and Olson 2015; Brackney et al. 2010; Campbell et al. 2008b; Carissimo et al. 2015; Dietrich et al. 2017a, b; Leger et al. 2013). The full genome sequences further enabled to study the role of antiviral exoRNAi pathways for vector function of these mosquito species. For example, Keene et al. (2004) were able to show that knockdowns of Dicer and Argonaute genes in Anopheles gambiae lead to increased replication of ONNV. However, Carissimo et al. (2015) showed that the induction of the exoRNAi pathway is not essential to control the ONNV infection in the midgut and thus speculate that the role of exoRNAi may be more important during dissemination of the infection than at the initial site of infection. In contrast, Khoo et al. showed that the infection of Aedes aegypti with Togaviridae is controlled by exoRNAi pathways at the level of the midgut barrier (Khoo et al. 2010). The tissue-specific knockdown of Dcr-2 in the midgut leads to enhanced replication and increased viral escape from the midgut (Khoo et al. 2010). The importance of exoRNAi in the defense of Aedes aegypti against SINV was further demonstrated by Campbell et al. (2008b), Myles et al. (2008), and Cirimotich et al. (2009) of which the latter study demonstrated that suppression of the exoRNAi pathway leads to reduced survival of infected mosquitoes. The contradicting observations in two different vector species, Aedes and Anopheles, indicate that, although exoRNAi is accepted as the major antiviral response in insects (Blair and Olson 2015; Kemp et al. 2013), the importance of this response can be tissue- and vector species-specific. The major role of RNAi in Aedes aegypti mosquitoes was further underlined by the observations made by Sanchez-Vargas et al. (2009) showing that DENV is controlled by the exoRNAi pathway and that loss of this pathway leads to increased virus replication and a shortened EIP. Besides Aedes, Culex mosquitoes are major vectors for arboviruses. Despite their importance, less data on exoRNAi pathway induction and function are available for Culex mosquitoes. Brackney et al. (2009) demonstrated that WNV infection induces small RNA production in Culex quinquefasciatus mosquitoes indicating that the exoRNAi pathway plays a role in these mosquitoes. Also the production of viRNAs in RVFV-infected Culex quinquefasciatus mosquitoes (Dietrich et al. 2017a) and the demonstration of WNV- and USUV-derived small RNAs in Culex pipiens mosquitoes (Fros et al. 2015b) are suggestive for an antiviral role of the exoRNAi pathway in Culex spp. However, functional evidence as it is presented for Aedes and Anopheles mosquitoes is currently lacking for Culex mosquitoes.

Besides the exoRNAi pathway, the Piwi-interacting RNA (piRNA) pathway can be activated in mosquitoes after infection with arboviruses. This pathway was initially described in Drosophila melanogaster, where the expression of transposons in germline cells and ovarian follicle cells is controlled by piRNAs (Brennecke et al. 2007). The 24- to 29-nt-long piRNAs are generated in a Dicer-independent manner and show a characteristic molecular signature (Brennecke et al. 2007; Morazzani et al. 2012; Vodovar et al. 2012). The piRNA pathway is initiated by the long single-stranded precursor RNAs that transcribed from piRNA clusters in the genome (Brennecke et al. 2007). This signal is amplified by the so-called ping-pong amplification loop (Siomi et al. 2011) including the Argonaute-3 (Ago3), Aubergine (Aub), and Piwi proteins (Brennecke et al. 2007; Gunawardane et al. 2007; Saito et al. 2006). In contrast to Drosophila melanogaster, the piRNA pathway has undergone an expansion in aedine and culicine mosquitoes with seven Piwi proteins (Piwi1–7) in Aedes aegypti and six Piwi proteins in Culex quinquefasciatus (Campbell et al. 2008a; Schnettler et al. 2013). This expansion correlates well with the extended role of the piRNA pathway in mosquitoes. Up to date, virus-specific piRNAs have been found in Aedes mosquitoes infected with members of all major arbovirus families and orders Flaviviridae (DENV), Togaviridae (SINV, CHIKV), and Bunyavirales (Dietrich et al. 2017b; Hess et al. 2011; Morazzani et al. 2012; Vodovar et al. 2012). The mechanism by which virus-derived piRNAs are induced is still not completely understood, but a recent study has given some insight into the mechanism of virus-derived synthesis in mosquito cells showing its dependence on Piwi5 and Ago3 proteins (Miesen et al. 2015, 2016). In addition, the Piwi4 protein is shown to be essential to control Semliki Forest virus (SFV, Togaviridae), BUNV, and RVFV infection in Aedes aegypti mosquito cells (Dietrich et al. 2017a, b; Schnettler et al. 2013), and Ago3 is essential to control ONNV in Anopheles gambiae (Keene et al. 2004).

The role of the third RNAi pathway, the microRNA (miRNA) pathway in arbovirus infection, is less clear, but recent data point to an involvement of miRNAs in virus-vector interactions (extensively reviewed in Asgari (2014)). The microRNA pathway exists in most metazoans and was initially described as a posttranscriptional regulatory mechanism. The miRNAs are produced by a Dicer enzyme (in insects Dicer-1) and incorporated into RISC-containing Argonaute proteins. This miRNA aids the RISC to a target RNA sequence which is complementary to the 5′8 nucleotides (seed region) of the miRNA. In mammals the role of cellular as well as virus-derived miRNA in modulation of virus replication has been long known (Muller and Imler 2007); however, a lack of knowledge persists on the role of miRNA in arbovirus-vector interactions. After publication of whole genome sequences from Aedes aegypti, Culex quinquefasciatus, and Anopheles gambiae, also miRNAs have been identified (Aedes aegypti (Li et al. 2009), Culex quinquefasciatus (Skalsky et al. 2010), Anopheles gambiae (Winter et al. 2007)). A number of studies reported the differential expression of miRNA in these vector mosquitos after infection with arboviruses. For example, Culex quinquefasciatus miR-989 was downregulated, and miR-92 was upregulated during WNV infection, but the meaning of this regulation remains unclear since no target was yet identified for those miRNAs (Skalsky et al. 2010). In Aedes aegypti, the infection with DENV serotype 2 alters the abundance of 35 miRNAs of which some have target sequences in genes linked to signal transduction and the cytoskeleton, but to date, no experimental evidence links these potential miRNA-target interactions to virus-vector interactions (Campbell et al. 2014). In contrast, the downregulation of Aedes albopictus miR-252 leads to a 1.5-fold increase of DENV serotype 2 virus replication (Yan et al. 2014). Furthermore, Aedes albopictus miR-2940, which was found to be unregulated during WNV infection, positively affects WNV replication through the upregulation of metalloprotease m41 ftsh (MetP) (Slonchak et al. 2014). However, knockdown of Ago1, the key protein of the miRNA pathway in Anopheles and Aedes, does not alter replication of several viruses, whereas knockdown of Ago2 (exosiRNA) or Ago3 (piRNA) pathways has a major impact on virus replication. Thus, the role of cellular miRNAs is not entirely clear and needs further investigation . Inducible Antiviral Responses

A couple of inducible mechanisms have been described in Drosophila and mosquitoes during the past years. The Toll and immune deficiency (IMD) pathways , initially characterized for their role in the control of bacterial and fungal infections in Drosophila (reviewed in Mussabekova et al. (2017)), are now widely recognized immune pathways in mosquitoes (reviewed by Sim et al. (2014)). In mosquitoes, Toll and IMD pathways are induced after pathogen recognition through peptidoglycan recognition proteins (PGRPs). Subsequent intracellular signaling is induced by Spätzle-MyD88 interaction (Toll) or IMD protein (IMD) which leads to the activation of nuclear factor “kappa-light-chain-enhancer” (NF-kB)-like transcription factors, namely, Rel1A (Toll) and Rel2 (IMD). Both pathways trigger the expression of antimicrobial effectors such as cecropins or defensins. The antiviral role of the Toll and IMD pathway was first shown in Drosophila after infection with several viruses (Toll, Drosophila X virus (Zambon et al. 2005); IMD, SINV and Cricket paralysis virus (Avadhanula et al. 2009; Costa et al. 2009)). In mosquitoes first evidence of a potential involvement of the Toll pathway in antiviral defense came from DENV-infected Aedes aegypti mosquitoes where 240 genes including key components of the Toll pathway, e.g., Spätzle, Toll, and Rel1A, were differentially regulated (Xi et al. 2008). A functional role of the Toll pathway was further confirmed in DENV-infected Aedes aegypti mosquitoes showing that transient Rel1 activation significantly reduces DENV titers, whereas silencing of MyD88 increased virus replication (Xi et al. 2008). Along this line, the induction of the Toll pathway in Wolbachia-infected Aedes aegypti is believed to be one way how the bacterium interferes with virus replication (Pan et al. (2012); see also section Wolbachia below). The impact of Toll pathway activation in other arbovirus infections and other mosquito species is less well studied. SINV and WNV induce the Toll pathway in Aedes aegypti (Colpitts et al. 2011; Sanders et al. 2005), while the latter fails to induce the Toll pathway in Culex quinquefasciatus (Bartholomay et al. 2010). Thus, the induction of the Toll pathway due to virus infection might be mosquito species-specific, or orthologues of the Toll pathway have not been completely characterized in other mosquito species, which could explain the lack of detection (e.g., the WNV-induced transcript CQ G12A2 in Culex quinquefasciatus shares 33% homology with the Toll-like receptor of Aedes aegypti; Smartt et al. (2009)). The IMD pathway plays a major role in mosquito antibacterial and antiparasite defense (Dong et al. 2009; Garver et al. 2012; Meister et al. 2005). The antiviral role has only been studied recently and in less detail than the Toll pathway. The upregulation of IMD pathway components was shown for Aedes aegypti mosquitoes infected with DENV and SINV (Barletta et al. 2017; Luplertlop et al. 2011; Sanders et al. 2005). First indirect evidence for a functional role for the IMD pathway in virus infection was presented by Sim et al. (2013) who showed that silencing of the pathway leads to enhanced viral replication in DENV-refractory strains of Aedes aegypti. However, transient activation of the pathway does not influence DENV infection (Xi et al. 2008). Recent findings by Barletta et al. (2017) point to an indirect role of the IMD pathway by controlling the gut microbiota, which then controls SINV replication. Further studies are necessary to clarify the role of the IMD pathway in antiviral defense. Specifically, attention needs to be payed to the clear distinction between the impact of the IMD pathway and the Janus kinase transducer and activator of transcription (JAK-STAT) pathway which both can be activated in mosquitoes by similar stimuli. The insects’ JAK-STAT pathway was initially described as a response to stress in Drosophila but has been linked with antiviral response in the fly through a microarray study (Dostert et al. 2005). Further evidence of a functional involvement of JAK-STAT pathway in antiviral defense arose from infection experiments of flies with mutations in the Janus kinase gene hopscotch (hop) with a panel of viruses. These experiments showed that the JAK-STAT pathway is essential to control Dicistroviridae (e.g., Drosophila C virus) infection in Drosophila but is dispensable for antiviral immunity against other viruses tested (Kemp et al. 2013). Bioinformatic analysis of mosquito genome data showed that orthologues of JAK-STAT pathway components, namely, the domeless (dome) receptor, the hop kinase, and STAT transcription factor, are also found in Anopheles gambiae and Aedes aegypti mosquitoes (Souza-Neto et al. 2009; Waterhouse et al. 2007). Infection of Aedes aegypti mosquitoes with DENV significantly induces the JAK-STAT pathway, and silencing of dome or hop leads to increased virus replication (Souza-Neto et al. 2009; Xi et al. 2008). Furthermore, a recent study by Jupatanakul et al. (2017) demonstrated that genetically engineered mosquitoes overexpressing dome and hop in the fat body have significantly reduced DENV replication in their bodies and most importantly largely reduced DENV infection rates in the salivary glands. However, the infection rates of ZIKV and CHIKV were not affected in the same mosquitoes. In contrast, Angleró-Rodríguez et al. (2017) demonstrated that ZIKV modulates the expression of Toll-, IMD-, and JAK-STAT-associated genes in Aedes aegypti and that the activation of Toll and JAK-STAT pathway significantly reduces ZIKV replication. Thus, it is not clear whether the JAK-STAT pathway is a pan-flavivirus-specific antiviral pathway similar to what was observed for Dicistroviridae in Drosophila or whether the antiviral function of this pathway is strictly virus species-specific. Furthermore, it is not clear if JAK-STAT pathway induction has a similar antiviral effect in other mosquito species. Data from WNV infection in Culex quinquefasciatus mosquitoes indicate that activation of the JAK-STAT pathway controls virus replication in these mosquitoes. Interestingly, the pathway is activated through secreted Vago, which is induced in a Dicer-2-dependent manner, thereby providing first evidence for a JAK-STAT-RNAi pathway cross talk (Paradkar et al. 2012). In contrast to Aedes and Culex mosquitoes, Anopheles gambiae mosquitoes do not show any transcriptional activation of JAK-STAT or Toll and IMD pathways after experimental infection with ONNV nor did a knockdown of components of this pathway impact ONNV replication (Waldock et al. 2012).

9.3.2 Virus Adaptation to the Mosquito: Immune Evasion and Immune Suppression by Arboviruses

Viruses are constantly exposed to the immune system of their hosts/vectors, which seeks to eliminate viral infection. In consequence, viral pathogens have evolved mechanisms to evade the immune system and infect new vectors.

Genetic reassortment is an important source of antigenic variability for segmented RNA viruses. It allows the fast antigenic shift instead of the slower antigenetic drift and, therefore, is one important factor for the evolution and emergence of viruses with an altered phenotype, disease potential, or host range (Gerrard et al. 2004; Kilian et al. 2013). Extinct or “new” viruses with greater pathogenicity might be created by natural or laboratory reassortment (Briese et al. 2013). An introduction of BUNV (Orthobunyavirus, Peribunyaviridae) or La Crosse virus (LACV) exotic to Europe and the possibility of reassortment with BATV (infraspecies belonging to the Bunyamwera orthobunyavirus species and serogroup), respectively, and TAHV (infraspecies belonging to the California encephalitis orthobunyavirus species and serogroup) endemic in Europe that may lead to reassortants with greater pathogenicity for humans or other vertebrates have to be considered (Briese et al. 2013; Eiden et al. 2014; Rudolf 2015). Bunyaviruses inherit a tripartite genome consisting of a small (S), medium (M), and large (L) segment. In Africa, NRIV and BUNV have similar geographic distributions across a broad region of sub-Saharan Africa, and both viruses have been isolated from the same species of Aedes mosquitoes (Gerrard et al. 2004). Importantly, a large outbreak of hemorrhagic fever in humans in East Africa in late 1997 and early 1998 was related to NRIV, which was found a reassortant of BUNV (S and L segment) and BATV (M segment) (Gerrard et al. 2004). Vector competence studies with Culex quinquefasciatus, Anopheles gambiae, and Aedes aegypti revealed considerable differences in their susceptibility to oral BUNV and NRIV infection. Culex quinquefasciatus was refractory and Anopheles gambiae moderately susceptible to both viruses. Interestingly, Aedes aegypti was moderately susceptible to BUNV but refractory to NRIV infection (Odhiambo et al. 2014). Therefore, considerable differences in the host range of viruses within the same Orthobunyavirus serogroup may occur.

Reassortment is a major driver of rapid evolution in viruses, such as genetic reassortment of avian and human influenza A viruses, bunyaviruses, or bluetongue viruses. Bunyaviruses are considered to originate from strictly inter-mosquito-transmitted viruses, and evolution led to adaptation to vertebrate hosts (Junglen 2016). A marked number of orthobunyaviruses lack the open reading frame encoding nonstructural NSs protein. For example, a novel clade of mosquito-associated bunyaviruses (herbeviruses) and viruses belongs to the Anopheles A, Anopheles B, and Tete serogroups. The orthobunyaviruses that lack the NSs protein fail to prevent induction of interferon-beta mRNA in mammalian cells (Hollidge et al. 2011; Marklewitz et al. 2013; Mohamed et al. 2009; Shchetinin et al. 2015; Weber et al. 2002). Reports about whether NSs may have an effect on virus growth and RNAi in insect cells are controversial (Hart et al. 2009; Hollidge et al. 2011; Rudolf 2015). The phenotype of closely related virus strains of the same species may even differ in their route of transmission. An originally strictly inter-arthropod-transmitted virus circulating within arthropod populations may convert to an arthropod-borne virus (may be due to evolutionary advantages, faster temporal and spatial distribution, and therefore higher fitness) that is more efficiently transmitted using an intermediate amplifying host, e.g., vertebrate hosts. Furthermore, a change in phenotype from a strictly inter-arthropod-transmitted virus toward a virus that may be directly transmitted between mammals cannot be precluded, as previously described for the two atypical bluetongue virus (BTV) serotypes 26 and 27 (Batten et al. 2014; Bréard et al. 2018).

In nature, reassortment may occur in the vertebrate host or the mosquito vector. However, studies of genome reassortment of orthobunyaviruses in vertebrates were unsuccessful (Beaty et al. 1985). In contrast, reassortment of heterologous as well as homologous orthobunyaviruses (Peribunyaviridae, Bunyavirales) was demonstrated in the mosquito vector (Borucki et al. 1999). In Aedes triseriatus mosquitoes, 20% of the offspring transovarially infected with LACV became superinfected when challenged with a second LACV strain or the serologically closely related Snowshoe hare virus (SSHV) (Borucki et al. 1999). Furthermore, a greater viral intra-host diversity was detected in ticks infected with Crimean-Congo hemorrhagic fever orthonairovirus (CCHFV; another member of the Bunyavirales order) compared to the vertebrate host (Xia et al. 2016) suggesting the arthropod vectors are the primary source of antigenic shift and drift. Accordingly, there is an urgent need to continually monitor emergent arboviral genotypes circulating within particular regions as well as vectors mediating these transmissions to preempt and prevent their adverse effects, genetic mechanism for species specificity, and vector competence owing to reassortment that needs further investigation (Odhiambo et al. 2014).

Besides such drastic measures as exchange of genome segments, viruses have developed other mechanisms to avoid or circumvent the vector and host immune systems. A study by Brackney et al. (2009) showed that siRNAs generated from WNV genomes in Culex mosquitoes mostly matched specific “hot spot” regions in the virus genome. These specific regions were more prone to mutations than other so-called cold spots, indicating that an enhanced mutation rate is one mechanism to escape siRNAs targeting. Also, abundantly expressed sub-genomic RNAs derived from the highly structured RNA encoded in the 3′ untranslated region of all flaviviruses (sub-genomic flavivirus (sf) RNAs) have been described (Pijlman et al. 2008). The sfRNAs derived from DENV and WNV genomes during infection have been shown to suppress Dcr-2-dependent dsRNA cleavage most likely through a direct inhibition of Dcr-2 (Moon et al. 2015; Schnettler et al. 2012). Furthermore, this inhibition in mosquito immunity by sfRNAs is curtail for successful virus transmission as shown by decreased transmission of sfRNA-deficient WNV in Culex mosquitoes and enhanced transmission of sfRNA overexpressing DENV in Aedes aegypti (Moon et al. 2015; Pompon et al. 2017).

Looking at classical viral suppressors of RNA silencing proteins (VSRs), diverse examples with multiple modes of action can be found. However, some common themes were established during evolution. The NSs protein of the plant-infecting Tomato spotted wilt orthotospovirus (TSWV; Tospoviridae, Bunyavirales), the B2 protein of the insect Flock house virus (FHV; Nodaviridae), and the VP3 protein of the mosquito-specific Culex Y virus (CYV; Birnaviridae) all sequester dsRNA molecules of different lengths and by this inhibit the recognition of these RNA molecules by Dcr-2 and incorporation of small dsRNA species into RISC, respectively. For arboviruses, not a lot of those classical VSRs have been identified; some authors even speculate that arboviruses do not express VSRs as matter of adaptation to avoid undue replication of the virus in the mosquito vector. Nevertheless, the DENV NS4B protein was shown to have VSR activity. While it could not bind to dsRNAs, it interferes with dicing (Kakumani et al. 2013 ).

9.3.3 Impact of Wolbachia on Vector Competence and Virus Replication

Several studies during recent years have shown that Wolbachia, a family of endosymbiotic Alphaproteobacteria, are associated with resistance to viral infection in several insect species including Drosophila (Hedges et al. 2008) and mosquito species (Glaser and Meola 2010; Moreira et al. 2009). Wolbachia pipientis was first described in the mosquito Culex pipiens and is inherited maternally via egg cytoplasm. The effects of Wolbachia infection in mosquitoes are widespread and somewhat contradictory. Several studies describe inhibitory effects on the infection with pathogens, especially in mosquitoes that have been artificially infected. However, others have reported no effect on virus infection or even enhanced virus infection rates (reviewed in Johnson (2015)).

Concerning the Culex complex, up to now, reported observations are contradictory. Glaser and Meola described an enhanced resistance of Wolbachia-infected Culex quinquefasciatus toward WNV (Glaser and Meola 2010), whereas Dodson et al. (2014) reported an enhancement of WNV replication in artificially Wolbachia-infected Culex tarsalis mosquitoes. Furthermore, the resistance phenotype in Culex quinquefasciatus as well as Culex pipiens toward WNV is not only dependent on the presence or absence but also on the Wolbachia density (Micieli and Glaser 2014). In Aedes mosquitoes, repression of virus infection but again lack of effect on virus replication due to Wolbachia infection was observed. The most impressive phenotypes were observed in Aedes aegypti, which is one of the few mosquito species not naturally infected with Wolbachia. Artificial transfection of the Drosophila wMel and wMelPop strains leads to severe reduction of virus replication, dissemination, and transmission of DENV, YFV, CHIKV, and WNV (Hussain et al. 2013; Moreira et al. 2009; van den Hurk et al. 2012; Walker et al. 2011). The transinfection with the wAlbB strain from Aedes albopictus also reduced DENV infection (Bian et al. 2010). However, naturally Wolbachia (wAlbA and wAlbB)-infected Aedes albopictus did not show virus repression phenotypes when transinfected with the Drosophila-specific Wolbachia strain wMel. Only on rare occasions, i.e., when Blagrove et al. (2012) replaced the natural Wolbachia strain by the wMel strain, they were able to observe reduced transmission efficiency in those mosquitoes for DENV.

The mechanisms underlying the resistance phenotype are poorly understood. Some studies link the effect of Wolbachia infection in mosquitoes to immune priming. In Drosophila several studies argue against an involvement of Toll an IMD pathway priming, since Rancès et al. (2013) showed that priming of these pathways is not necessary for the Wolbachia-mediated blocking of DENV and Chrostek et al. (2014) found a high antiviral protection without immune upregulation after interspecies transfer of Wolbachia. Most studies analyzing the resistance phenotype in mosquitoes used Aedes aegypti mosquitoes and DENV infection. This combination has shown the most pronounced resistance phenotype, which might be due to the fact that Aedes aegypti mosquitoes are not naturally infected with Wolbachia. The activation of the Toll, IMD, JAK-STAT, and melanization pathways in Aedes aegypti was investigated among others by Kambris et al. (2009). Other studies have demonstrated the crucial role of the Toll pathway to control DENV infection in these mosquitoes (Xi et al. 2008). Hence, it is rational to suspect a causative link between the blocking phenotype and immune activation. Indeed, such a link could be established by Pan et al. (2012). Recent data collected from wAlbB-infected Aedes aegypti mosquitoes even show that the permanent activation of the Toll and IMD pathways by Wolbachia is needed by the bacteria to establish a stable infection in the mosquito (Pan et al. 2018). However, the same is not the case in the natural DENV vector Aedes albopictus. An additional transinfection with the heterologous wMel strain does not lead to significant upregulation of the innate immune pathways (Blagrove et al. 2012). A recent study with Wolbachia-infected Aedes aegypti cell lines demonstrated that the induction of RNAi, Toll, and IMD pathways must not necessarily be the cause of the protective effect since only the knockdown of the RNAi pathway leads to a small but significant reduction of the protective phenotype (Terradas et al. 2017). It has been demonstrated that a Vago homolog in Aedes aegypti (AedesVago1) inhibits the replication of DENV (Asad et al. 2017). This is of special interest, since Vago has been demonstrated to be induced in a Dcr-2-dependent manner in Culex quinquefasciatus and facilitated crosstalk between the exoRNAi pathway and the JAK-STAT pathways. Taken together, all data on immune priming and the inhibition phenotype induced by Wolbachia show that immune priming might explain this effect in parts. However, other mechanisms need to be considered to fully explain the inhibition phenotype. A couple of studies create possible links between insect cell physiology and infection resistance. The infections of Aedes aegypti cells with the Wolbachia strain wMelpop leads to a downregulation of MCT1 expression (Osei-Amo et al. 2012). Since alteration of MCT expression has been shown to induce apoptosis in insects (Jang et al. 2008), there might be a link between enhanced apoptosis and reduced virus replication. Also the energy metabolism of cells is discussed as possible cause for virus inhibition. DENV is known to manipulate the cellular fatty acid biosynthetic pathway to create a favorable environment for viral replication complexes at intracellular membranes (Perera et al. 2012). It has been also shown that Wolbachia requires unsaturated fatty acids from host cells, since it cannot synthesize those. Thus, both bacteria and viruses need unsaturated fatty acid from the host; by limiting this resource, bacterial growth could suppress virus replication. Cholesterol was also shown to be crucial for DENV and alphavirus replication (Hafer et al. 2009; Lu et al. 1999; Rothwell et al. 2009) as well as for Wolbachia replication. A high growth rate of Wolbachia could deplete the insect cells from cholesterol and by this block the virus from essential resources for its replication (Moreira et al. 2009). This hypothesis was also confirmed in Drosophila when a cholesterol-enriched diet reduced the protective effect of Wolbachia against virus infection (Caragata et al. 2013). In Aedes aegypti cells, treatment with 2-hydroxypropyl-β-cyclodextrin to restore cholesterol homeostasis rescued DENV replication (Geoghegan et al. 2017). To confirm this mechanism in different mosquito and virus species, future studies will be necessary .

9.4 Conclusions

It can be summarized that virus replication in vector mosquitoes and thus the emergence of arbovirus infection are controlled by a myriad of different factors, including the mosquito abundance, temperature profiles, habitats, abundance of susceptible hosts, and other external factors. Also, the importance of intrinsic factors including the immune system of the insect and the presence of microbiota and endosymbionts such as Wolbachia cannot be stressed enough. The available data draw a complicated picture of the importance of all different factors which all warrant further research for clarification, altogether making vector competence studies a challenging but no less fascinating future topic.