Experimental risk assessment for chikungunya virus transmission based on vector competence, distribution and temperature suitability in Europe, 2018

Background Over the last decade, the abundant distribution of the Asian tiger mosquito Aedes albopictus in southern Europe and the import of chikungunya virus (CHIKV) by infected travellers has resulted in at least five local outbreaks of chikungunya fever in France and Italy. Considering the ongoing spread of Ae. albopictus to central Europe, we performed an analysis of the Europe-wide spatial risk of CHIKV transmission under different temperature conditions. Methods: Ae. albopictus specimens from Germany and Italy were orally infected with CHIKV from an outbreak in France and kept for two weeks at 18 °C, 21 °C or 24 °C. A salivation assay was conducted to detect infectious CHIKV. Results: Analyses of mosquito saliva for infectious virus particles demonstrated transmission rates (TRs) of > 35%. Highest TRs of 50% for the mosquito population from Germany were detected at 18 °C, while the Italian population had highest TRs of 63% at 18 °C and 21 °C, respectively. Temperature data indicated a potential risk of CHIKV transmission for extended durations, i.e. sufficiently long time periods allowing extrinsic incubation of the virus. This was shown for areas already colonised by Ae. albopictus, as well as for large parts of central Europe that are not colonised. Conclusion: The current risk of CHIKV transmission in Europe is not primarily restricted by temperature, which allows extrinsic incubation of the virus, but rather by the vector distribution. Accordingly, all European countries with established populations of Ae. albopictus should implement respective entomological surveillance and monitoring systems, as basis for suitable control measures.


Introduction
Chikungunya virus (CHIKV) is a mosquito-borne alphavirus (Togaviridae family) [1]. The virus was first isolated from humans, as well as from mosquitoes, during an outbreak at the border region between Mozambique and Tanzania in 1952. Human cases of chikungunya fever can cause severe, debilitating and often chronic arthralgia [2]. Several species of the mosquito genus Aedes were found competent to transmit CHIKV. However, Ae. aegypti and Ae. albopictus are considered the most important vectors. In the past, occurrences of CHIKV were restricted to the African and Asian continents, causing local, sporadic epidemics; during the last decade, however, the virus has expanded over a substantial geographical range, invading India, Indian Ocean islands and the Americas, where it has caused millions of human infections [2].
Three phylogenetic CHIKV lineages with distinct antigenic characteristics are known worldwide [2,3]. Two major lineages circulate in Africa (East, Central and Southern Africa region (ECSA) and West Africa) and a third is present in Asia. In Africa, CHIKV is maintained in an enzootic cycle. The transmission cycle includes forest-dwelling mosquito species of the genus Aedes and non-human primates [2]. Anthropophilic mosquito species, Ae. aegypti and Ae. albopictus, play a crucial role in the urbanisation of CHIKV. This is in contrast to Asia, where the virus has historically been maintained in an urban cycle between Ae. aegypti/albopictus and humans.
Ae. aegypti is the primary vector for CHIKV. However, Ae. albopictus is considered to have a high competence to transmit a specific variant of the ECSA lineage. The virus mutant has a single amino acid change from alanine to valine at the E1 envelope glycoprotein amino acid 226 (E1-A226V) [4]. This change leads to a better adaptation of the virus to the species Ae. albopictus, resulting in a 50-fold increase in vector competence in comparison to Ae. aegypti. The ECSA mutant is considered to be the most important factor for CHIKV outbreaks in regions where the primary vector, Ae. aegypti, is absent. This is potentially linked to the spread of the ECSA lineage to the Indian Ocean and the first autochthonous CHIKV transmission in Europe [5]. Thus far, all CHIKV strains isolated during outbreaks in Europe belonged to the ECSA lineage. However, some of the isolates contained the E1-A226V-mutation [6][7][8], whereas others did not [9,10]. Therefore, other yet undefined mutations in the envelope proteins may also affect the adaptation of CHIKV to Ae. albopictus [9][10][11].
Autochthonous transmission of CHIKV has been repeatedly observed in mainland Europe. The virus circulated in France in 2010 [9], 2014 [7] and 2017 [8], and two major outbreaks occurred in Italy in 2007 [6] and 2017 [10]. In total, at least 605 human CHIKV cases were reported, the majority in the two epidemics in Italy (n = 575) [8]. This was possible because of the establishment of Ae. albopictus in the region. One of the most invasive mosquito species in the world [12,13], its global spread is driven by transcontinental connectivity through shipping and flight routes. At present, Ae. albopictus has infested more than 25 European countries [12,14], with highest abundances reported from Italy. The species is repeatedly introduced to various locations in central Europe [15], even though regions north of the Alps were previously considered unsuitable for the establishment of Ae. albopictus. Still, overwintering was recently suggested in Germany [16][17][18]. This included local expansion of populations and detection of larvae already in spring. According to the European Centre for Disease Prevention and Control (ECDC), Solna, Sweden, Ae. albopictus is classified as 'established' along the Upper Rhine Valley in Germany and France [14].
There is currently no available vaccine or CHIKV-specific treatment [5]. The control of the disease primarily depends on reduction of the vector population. Further options are individual protection with repellents or behavioural avoidance. Therefore, the evaluation of the vector competence of local mosquito populations is crucial, as the assessment of CHIKV risk transmission allows for the adapting of surveillance and control systems. Vector competence is the ability of a mosquito to acquire a pathogen and subsequently transmit it to a new host [2]. This is commonly evaluated through experimental infection experiments. These studies aim to identify infectious virus particles secreted with the saliva to the vertebrate host. Transmission only becomes possible if the virus can overcome the midgut and salivary glands, which are the most important barriers for infection of the vector and final escape [19]. Vector competence depends on a complex interaction of vector population, virus strain and temperature [20]. Only a suitable combination allows the virus to replicate and disseminate. Invasion of the salivary glands may result in transmission through the next bite.
Vector competence studies using temperatures below 20 °C have become increasingly important since the observation of established populations of the Ae. albopictus north of the Alps [14]. We therefore conducted a vector competence study at three different temperatures, with a CHIKV outbreak strain from France, using Ae. albopictuspopulations from Germany and Italy, to comprehensively assess the risk of CHIKV transmission in Europe.  The detailed salivation assay to detect transmission was described before [21]. In brief, mosquitoes were anesthetised with CO 2 and demobilised. Infection rate (IR) is commonly defined as the number of CHIKV-positive mosquito bodies per number of fed females. Different definitions exist for the calculation of transmission rates (TRs). The study did not focus on the mechanistic processes of vector competence (e.g. infection barriers), but aimed to assess the risk for CHIKV transmission in Europe. Therefore, in order to simplify the sample processing, the calculation of TRs followed the definition by Fortuna et al. [22]: number of mosquitoes with CHIKV-positive saliva per number of mosquitoes with CHIKV-positive bodies.

Chikungunya virus transmission risk assessment
The risk map for CHIKV transmission in Europe was constructed by combining the current distribution of Ae. albopictus with temperature data from the infested regions. Current distribution data of Ae. albopictus at the regional administrative level (NUTS3), as at April 2018, were obtained from [14]. Time series of daily mean temperature data (European re-analysis and observations for monitoring, E-OBS v17.0) for the study area were downloaded from http://www.ecad. eu [23]. E-OBS data are available on a 0.25 ° regular latitude-longitude grid and were extracted for a period of 10 years between 2008-2017. For each grid cell, the number of days per year with preceding 14 days having a mean daily temperature ≥ 18 °C were calculated using the programme R [24]. The annual values were then averaged over the 10-year period.

Results
Experimental infection studies showed IRs of 100%, irrespective of whether Ae. albopictus populations from Germany or Italy were used and independent of temperature (18 °C, 21 °C or 24 °C) (

Chikungunya virus transmission risk assessment
The analysed data indicated that most parts of the vector's current distribution area allow transmission of CHIKV (Figure).

Discussion
The extrinsic incubation period (EIP) of CHIKV in Ae. albopictus can be very short, i.e. infectious particles can be present in the saliva within 2 to 3 days after ingestion of an infectious blood meal [25]. This has a direct impact on the epidemiology of CHIKV. It must be considered one of the most important factors allowing for transmission in areas without tropical climates. However, there is a lack of comprehensive knowledge regarding the EIP of CHIKV, especially at low temperatures [26]. Only a few studies have focused on temperate climatic conditions [2]. Such studies are especially important in light of the ongoing spread of Ae. albopictus from source populations around the Mediterranean Sea to central Europe [14][15][16][17][18].
The tested Italian mosquito population, as well as the more recently detected German mosquito population, showed IRs for CHIKV of 100%. TRs of more than 35% were observed for all three tested temperatures. It may appear surprising that the lowest CHIKV TRs (37.5%) for both populations were found at the highest temperature (24 °C) and TRs increased with decreasing temperatures; however, this is in line with previous infection studies with different arboviruses [27]. Some mosquito species have a reduced ability to modulate viral infections under low temperatures. The underlying mechanism might be a temperature-dependent deficiency of antiviral immunity, as RNA silencing is inhibited in mosquitoes subjected to low temperatures. Ae. aegypti specimens reared at cooler temperatures have an impairment of the antiviral immune RNA interference (RNAi) pathway. This pathway is critical to the mosquito's ability to control viral infections. The exact mechanism between temperature and CHIKV IRs needs to be further explored, even though it is well established that RNAi impairments occur downstream of the initial dicing step [27].
Human CHIKV infections are regularly imported to Europe [28].  (2017) were caused by virus strains without this specific mutation [8,10]. Further mutations might have a similar relevance for the probability of transmission in Europe [9,10]. Potential candidates are adaptive mutations in the E2 envelope glycoprotein (e.g. E2-L210Q) [11]. Such mutants were detected in CHIKV isolates from India and might play an important role in the spread and diversification of CHIKV lineages. In addition, one must keep in mind that vector competence is an important parameter of vector capacity, but it is not the only one. Vector capacity is determined by a complex interaction between different factors [30]. Important drivers include the population density and host-feeding patterns. The latter is strongly influenced by both host preference and host availability.
The risk assessment presented here is a conservative scenario. Both tested populations had substantial CHIKV TRs at 18 °C, i.e. 14 days post infection. In addition, temperature data for the risk maps were averaged over a 10-year period. For a more comprehensive and precise evaluation, further vector competence studies investigating lower temperatures and shorter EIPs are required. In addition, this analysis has only focused on Ae. albopictus, though other studies have demonstrated vector competence for CHIKV of further Aedes species [2]. Infection experiments have not identified any other European mosquito species except Ae. albopictus as potential vector for CHIKV, but only a few studies have been done for selected species (e.g. Ae. vexans and Culex pipiens).
In conclusion, Europe offers broad temperature suitability for CHIKV transmission and experiences regular travel-associated virus introduction. Therefore, all European countries with established Ae. albopictus populations should implement entomological surveillance programs as well as monitoring and notification systems for imported human cases to prevent further spread and autochthonous CHIKV transmission in Europe.

*Authors' correction
In the Discussion, the sentence "Further spread of the virus was observed during the two large outbreaks in Italy (e.g. to Germany; personal communication, Christina Frank, Robert Koch-Institute, Germany, November 2017).» was replaced by «The two large outbreaks in Italy also affected visitors (including a German traveller in 2017; personal communication, Christina Frank, Robert Koch-Institute, Germany, November 2017).» to avoid possible misinterpretation. The correction was made on 20 July 2018, as requested by the authors.