Experimental study of the susceptibility of a European Aedes albopictus strain to dengue virus under a simulated Mediterranean temperature regime

Dengue virus (DENV) has re‐emerged in Europe driven by the geographic expansion of the mosquito species Aedes albopictus and Aedes aegypti (Diptera: Culicidae) and the introduction of the virus by viraemic travellers. In the present study, the vector competence (VC) of Ae. albopictus collected in Catalonia (northeast Spain) was evaluated for two different DENV strains, DENV‐1 and DENV‐2, the serotypes responsible for all outbreaks of dengue that have occurred in Europe. Mosquitoes were reared under environmental conditions mimicking the mean temperature and humidity recorded in July on the Mediterranean coast of Catalonia. Mosquitoes were fed on an artificial infectious bloodmeal and, at 14 days post‐exposure, infection, disseminated infection and transmission rates (IR, DIR, TR) and transmission efficiency (TE) were determined by testing the virus in the body, legs and saliva. The tested Ae. albopictus strain was found to be susceptible to both DENV‐1 and DENV‐2 and to be able to transmit DENV‐1. This is the first time that the VC of Ae. albopictus for DENV has been tested in Europe in this specific context (i.e. mimicking the Mediterranean temperature and humidity recorded in Catalonia in July). This study confirms the potential of Ae. albopictus to start autochthonous DENV transmission cycles in the Mediterranean basin.


Introduction
Dengue virus (DENV) is a vector-borne Flavivirus of the family Flaviviridae and is mainly transmitted by Aedes aegypti (Linnaeus) and Aedes albopictus (Skuse). Dengue virus is the most widespread of the arboviruses that affect humans, with more than 390 million cases of infection estimated per year (Bhatt et al., 2013). Infection with any of the DENV serotypes (DENV-1, DENV-2, DENV-3 and DENV-4) can either be asymptomatic (in 75% of cases) or result in one of Further autochthonous outbreaks were reported in France in 2013, 2014 and 2015 (Marchand et al., 2013;Giron et al., 2015;Succo et al., 2016). Aedes albopictus was indicated as the vector involved in all local transmissions in continental Europe, and the DENV strains belonged to either serotype 1 or serotype 2. Aedes albopictus was first introduced into Europe in 1979 (Adhami & Reiter, 1998). Since then, it has spread rapidly to several European countries, especially in the Mediterranean basin, and was first identified in Spain in 2004 (Aranda et al., 2006).
To assess the risk for local transmission events, it is essential to evaluate the vector competence (VC) of local mosquito populations. This is particularly important for Ae. albopictus as this species' VC for DENV has been shown to vary substantially among populations of different geographic origins (Gubler & Rosen, 1976;Boromisa et al., 1987). Previous studies in European Ae. albopictus populations have been performed at a constant temperature (28 ∘ C) to simulate the mean temperature in tropical countries where DENV is endemic. Only a few studies have addressed the effects of realistic temperatures on VC for DENV, and all of them were carried out in Ae. aegypti (Lambrechts et al., 2011;Carrington et al., 2013aCarrington et al., , 2013bGloria-Soria et al., 2017). These studies support the notion that local temperatures may alter VC. Therefore, measuring VC while simulating environmental conditions at the mosquito collection site is the best approach to estimating the risk for an outbreak in the case of DENV introduction. Consequently, this study was performed using the mean temperature and humidity recorded in July on the Mediterranean coast of the Iberian peninsula to estimate the risk for autochthonous transmission of DENV.

Mosquito strain
The strain of Ae. albopictus used in this study was collected in Sant Cugat del Vallès in 2005 and reared in the laboratory to obtain a stable colony. Mosquitoes were reared under conditions that mimicked the environmental field conditions in their natural habitats in Catalonia during the month of July. Climatological data were provided by the Meteorological Service of Catalonia (www.meteo.cat). Mean day and night temperatures were calculated with respect to the summer photoperiod in Catalonia [26 ∘ C for 14 h (light) and 22 ∘ C for 10 h (dark)]. An average relative humidity of 86% was calculated.
Virus stocks of DENV-1 were produced following two passages on C6/36 cells, whereas DENV-2 was obtained after one passage on C6/36 cells. Supernatants were collected and stored at − 80 ∘ C prior to their use for mosquito oral feeding.
Vector competence assay design To investigate VC, four different rates were considered: infection rate (IR); disseminated infection rate (DIR); transmission rate (TR), and transmission efficiency (TE). The IR was defined as the proportion of mosquitoes with virus-positive bodies (abdomen, thorax and head) among the tested mosquitoes. The DIR was defined as the proportion of mosquitoes with infected legs among those with infected bodies. The TR was defined as the proportion of mosquitoes with DENV RNA in saliva among the number of mosquitoes with disseminated infection. Transmission efficiency was defined according to the proportion of mosquitoes with DENV RNA in saliva among the total number of mosquitoes tested.
Female mosquitoes aged 7-10 days and not previously blood-fed were fed using the Hemotek feeding system (Discovery Workshops, Accrington, U.K.) with a pathogen-free chicken skin as a membrane. The bloodmeal was prepared by mixing DENV-infected bovine blood with heparin and ATP (5 × 10 −3 M) as phagostimulant (Sigma-Aldrich Corp., St Louis, MO, U.S.A.). The viral load was 5.5 log 10 TCID 50 /mL. The viral titre was chosen based on the reported viraemia level reached in human patients infected with DENV (Tang et al., 2010). After blood feeding, 10% of fully engorged females were randomly collected and killed to confirm virus exposure. Then, fully engorged females were selected under carbon dioxide (CO 2 ) anaesthesia and individually transferred to cardboard cages (Watkins & Doncaster Ltd, Leominster, U.K.). Females were kept for an extrinsic incubation period (EIP) of 14 days inside the climatic cabinet. Sucrose solution (10%) was administered on cotton pledgets placed on the mesh screen. At 14 days post-exposure (d.p.e.), all mosquitoes were anaesthetized using CO 2 and dissected. The legs and wings were detached from the body, and both parts were separately homogenized in 0.5 mL of Dulbecco's modified Eagle's medium (DMEM) (Lonza Group AG, Basel, Switzerland). The samples were homogenized at 30 Hz for 1 min using TissueLyser II (Qiagen GmbH, Hilden, Germany) and stored at − 80 ∘ C until tested for DENV. The same protocol was used in all assays aimed at estimating the mortality rate, as well as in those designed to obtain IR and DIR data.
To estimate transmission, two separate assays were designed. In the first assay, transmission was tested using FTA™ cards at two different time-points: 9 d.p.e. and 14 d.p.e. On those days, cotton pledgets were replaced by FTA™ cards (GE Healthcare, Little Chalfont, U.K.). The FTA™ cards were soaked with Manuka honey (Manuka Health New Zealand, Te Awamutu, New Zealand) mixed with a blue alimentary colorant. After collection, FTA™ cards were resuspended in 0.3 mL of phosphate-buffered saline (PBS) and stored at − 80 ∘ C until tested. In the second assay, transmission was tested at 14 d.p.e. with two different methods: (a) using FTA™ cards, as described above, and (b) performing a direct salivary extraction using a capillary technique, as previously described (Dubrulle et al., 2009). Briefly, after the dissection of the legs and wings, the proboscis was inserted into a P20 pipette tip filled with 7 μL of a 1 : 1 solution of fetal bovine serum (FBS) and 50% sucrose solution. To stimulate salivation, 1 μL of 1% pilocarpine (Sigma-Aldrich Corp.) prepared in PBS at 0.1% Tween 80, was applied to the thorax of each mosquito. After a period of 60 min, the solution containing the saliva was expelled into 1.5-mL tubes containing 193 μL of DMEM; 150 μL were used for viral RNA extraction and the remaining 50 μL were used for DENV isolation.
Amplification was performed using the AgPath One-Step ID RTPCR Kit (Ambion, Applied Biosystems, Inc., Foster City, CA, U.S.A.), and a 7500 Fast Real-Time PCR System (Applied Biosystems, Inc.) programmed as follows: 45 ∘ C for 10 min; 95 ∘ C for 10 min, and 40 cycles of 97 ∘ C for 15 s and 55 ∘ C for 30 s.

Virus isolation and titration
Virus isolation was performed in a monolayer of C6/36 cells. Cells were incubated for 6-7 days (28 ∘ C, 5% CO 2 ). As a cytopathic effect was not observed, DENV replication was detected in the supernatant using the real-time RT-PCR.
Both DENV stocks were titrated in a monolayer of C6/36 cells. Routinely, eight wells were infected for each 10-fold dilution and 20-μL quantities of inoculum were spiked into each well. After that, 150 μL MEM (Gibco, Thermo Fisher Scientific, Inc., Waltham, MA, U.S.A.) supplemented with 2% FBS (Life Science Co., London, U.K.), 2 mM l-glutamine, non-essential amino acids, 1000 U/mL of penicillin, 10 mg/mL of streptomycin and 500 U/mL of nystatin (all from Sigma Corp.) were added per well and the plates were incubated at 28 ∘ C and 5% CO 2 for 7 days. Calculation of the viral titre was performed by virus detection in each well using the real-time RT-PCR described above. Cycle threshold (Ct) values ranged from 18.38 to 22.45 in those wells in which the virus replicated. The TCID 50 /mL was calculated using the method of Reed and Muench (Villegas, 1998).

Estimations of IR and DIR
Assays to estimate IR and DIR were conducted using mosquitoes from the same colony and an identical protocol (i.e. they were two replicates of the same experiment). Thus, the results from the first and second assays could be combined and the differences between DENV-1 and DENV-2 evaluated using a chi-squared test.
In addition, the ability of each strain of DENV to replicate in different mosquito tissues (body and legs) was also evaluated using retrotranscriptase quantitative PCR (qPCR). Mean Ct values of the bodies and legs infected with DENV-1 and DENV-2, respectively, were compared. Briefly, the normality of the data was assessed using the Shapiro-Wilk test, and then data were compared using the t-test (in the case of normally distributed data) or the Wilcoxon rank sum test (in the case of non-normally distributed data). All calculations were carried out using R statistical software (http://cran.r-project.org/).

Mosquito infection
The results of the assays for the estimations of mortality, IR and DIR are shown in Table 1.
The finding of positive bodies indicated that the virus was able to cross the midgut infection barrier successfully at 14 d.p.e. Positive legs showed mosquitoes in which the virus was also able to cross the midgut escape barrier. Assessment of infection showed that the cumulative IRs for DENV-1 and DENV-2 were 53% (49/93) and 33% (26/78), respectively. The statistical analysis revealed the IR was significantly higher for DENV-1 than for DENV-2 (P = 0.041). The cumulative DIR values were also higher for DENV-1 than for DENV-2 at 53% (26/49) and 35% (9/26), respectively (Table 1). However, the difference was not statistically significant (P = 0.12).
The result of the Wilcoxon rank sum test indicated that the difference in mean Ct values between DENV-1 and DENV-2 was statistically significant (P = 0.00016). This result suggested that DENV-1 was able to replicate more efficiently than DENV-2 in the body of Ae. albopictus. By contrast, the mean ± SD Ct values of the legs infected with DENV-1 and DENV-2 were 32.1 ± 2.0 and 34.0 ± 3.4, respectively (Fig. 1). The result of the t-test indicated that the difference in mean Ct values between DENV-1 and DENV-2 was not statistically significant (P = 0.151). Therefore, although DENV-1 was able to replicate more efficiently than DENV-2 in mosquito bodies, females with disseminated infection had similar amounts of virus in the legs regardless of viral strain.

Transmission of DENV
In the first assay, the FTA™ cards were collected at 9 d.p.e. and at 14 d.p.e. The FTA™ cards collected at 9 d.p.e from both groups (DENV-1 and DENV-2) showed negative findings (Table 2). However, two FTA™ cards collected at 14 d.p.e. from two mosquitoes exposed to DENV-1 were positive. These Rates were obtained at 14 days post-exposure for dengue virus (DENV) serotypes 1 (DENV-1) and 2 (DENV-2) in the first and second assays. specimens also showed positive legs, indicating the presence of a disseminated infection. Therefore, the TR for DENV-1 was 20% (2/10). Conversely, at the same time-point (14 d.p.e.), all the FTA™ cards from mosquitoes exposed to blood infected with DENV-2 were negative. In the second assay, all the FTA™ cards and saliva samples were collected at 14 d.p.e. The FTA™ cards and saliva samples from mosquitoes exposed to both viral strains tested negative by RT-qPCR. Moreover, direct isolation in Vero cells from saliva samples gave negative results.
Given the results obtained using molecular techniques with FTA™ cards, the estimated TE for DENV-1 was 4.7% (2/43).

Discussion
For the first time in Europe, the VC of Ae. albopictus for DENV was estimated under a protocol that mimicked the temperature conditions present in a Mediterranean area during the summer month of July. Recent data suggest that the accurate assessment of the risk for DENV transmission in a mosquito population requires the incorporation of local temperature in the VC experiments (Gloria-Soria et al., 2017). The results of the present study provide evidence that the Ae. albopictus strain tested is susceptible to oral infection with both DENV-1 and DENV-2 at a viral load (5.5 log 10 TCID 50 /mL) within the range of viraemia (3-8 log 10 TCID 50 /mL) observed in humans (Salazar et al., 2007). The IR was statistically higher for DENV-1 than for DENV-2, which indicates that susceptibility is dependent on the serotype of DENV, as observed by Gubler & Rosen (1976).
The Ae. albopictus strain tested in the present study was more susceptible to DENV-1 infection in terms of disseminated infection (DIR: 53%) than the French population of Ae. albopictus (DIR: 28-45%) exposed to 10 5.3 focus-forming units (FFU)/mL of bloodmeal and maintained at a constant temperature (28 ∘ C) (Vega-Rua et al., 2013). Such variation may be explained by the different environmental conditions assayed, but geographical differences in the vector population and the virus strain cannot be ruled out.
Previous studies examining VC in Mediterranean populations of Ae. albopictus (from France, Italy and Lebanon) Talbalaghi et al., 2010;Haddad et al., 2012) for the same Asian DENV-2 genotype and conducted under identical environmental conditions (constant temperature of 28 ∘ C) showed a wide range of DIRs of 12-69%, 14-38% and 32-47%, respectively. The fluctuating conditions of the present study (daytime temperature of 26 ∘ C and night-time temperature of 22 ∘ C) produced a DIR of 35% for the DENV-2 strain, which is within the range of DIRs reported in the earlier works Talbalaghi et al., 2010;Haddad et al., 2012).
In addition to differences in IRs between DENV strains, the present study found a significant difference in viral infectivity. The comparison of mean Ct values in the body for both DENV strains indicated a statistically significantly higher viral replication for DENV-1 than for DENV-2, which may reflect differences in the immune response to viral infection (Sánchez-Vargas et al., 2009). By contrast, there were no statistically significant differences between the strains tested in the mean Ct values for the legs. This suggests that the efficiency of viral replication in the midgut did not determine the amount of virus in other tissues after dissemination, as reported in previous studies (Bosio et al., 1998).
The use of FTA™ cards was originally developed for field studies (Hall-Mendelin et al., 2010. However, FTA™ cards may also be used in VC assays and allow the collection of saliva from a single mosquito at different time-points without requiring the specimen to be killed. The technique does not require trained personnel, whereas the capillary technique does, and reduces the operative time required to collect the sample. The principal disadvantage of the technique is that the saliva cannot be used for direct virus isolation in cell culture. Thus, although the use of FTA™ cards cannot replace virus isolation, it represents a good strategy for evaluating transmission capacity Table 2. Transmission rates of dengue virus (DENV) serotypes 1 (DENV-1) and 2 (DENV-2) in the first and second assays. at intermediate time-points without the need to create subgroups of specimens and sacrifice them periodically. The transmission capacity of each mosquito was tested by using FTA™ cards at two different time-points (9 d.p.e. and 14 d.p.e.). At 9 d.p.e., no mosquitoes were able to transmit either strain. However, at 14 d.p.e., the mosquito strain tested was able to successfully transmit DENV-1. By contrast, DENV-2 was not detected in mosquito saliva at 14 d.p.e. at the conditions used in the present work. These negative results do not imply that DENV-2 transmission will not occur under different conditions. Thus, the extended period needed to complete the EIP of DENV (from 9 d.p.e. to 14 d.p.e.) may have an impact on the VC of this Spanish mosquito strain, potentially reducing the risk for outbreak and providing a larger time window in which to implement surveillance and vector control measures. In previous VC studies using Mediterranean mosquito populations (Haddad et al., 2012;Vega-Rua et al., 2013), EIP varied although the mosquitoes were held at a constant temperature of 28 ∘ C. The DENV-1 strain presented a shorter EIP and could be transmitted by a French population of Ae. albopictus at 9 d.p.e. (67%) (Vega-Rua et al., 2013). By contrast, DENV-2 seemed to require a longer EIP to achieve transmission (Vega-Rua et al., 2013). The same strain was not detected in saliva of a Lebanese Ae. albopictus population at 10 d.p.e., but achieved a 38% TR at 21 days (Haddad et al., 2012). It is important to note that, by contrast with the previous studies, the present experimental infections were performed under a protocol that mimicked fluctuating temperature conditions, which strengthens the findings of this work. The average temperature within the current profile (22-26 ∘ C) is 24 ∘ C, which is 4 ∘ C lower than temperatures used in the comparable studies mentioned above, and which may potentially result in lower infection and transmission rates, and also lengthen the EIP (Poole-Smith et al., 2015). The results of previous studies in Ae. aegypti VC for DENV under realistic temperature regimes indicate that fluctuations around low temperatures (20 ∘ C) enhanced the mosquito's VC (Carrington et al., 2013a). This was not reported with fluctuations around high temperatures (26 ∘ C) (Lambrechts et al., 2011;Carrington et al., 2013b). In Europe, DENV epidemics have occurred in the Mediterranean area during the summer. These environmental conditions allowed Ae. albopictus mosquitoes to acquire and subsequently transmit DENV from viraemic travellers to local populations (Semenza et al. 2014). The results of the present study confirm that the Ae. albopictus strain from Catalonia is competent for DENV under the temperature conditions that prevail on the Mediterranean coast of Catalonia during the summer. Although the present results point to a low TE (4.6%) for DENV-1, this parameter is only one of the multiple components used to determine VC. Additional ecological (mosquito population densities, feeding behaviour, daily mosquito survival) and anthropogenic (frequency of arrival of viraemic hosts, population density) factors must be considered to determine the real risk for DENV transmission in the Mediterranean area.

Conclusions
The present results indicate that the tested Ae. albopictus strain was susceptible to both DENV-1 and DENV-2. The mosquito strain tested was also able to transmit DENV-1, demonstrating that a local transmission event is possible in the Mediterranean region. This study contributes to knowledge of the VC of Ae. albopictus for DENV, which may be useful in the development of DENV risk models and surveillance programmes.