Assessment of fitness and vector competence of a New Caledonia wMel Aedes aegypti strain before field-release

Background Biological control programs involving Wolbachia-infected Aedes aegypti are currently deployed in different epidemiological settings. New Caledonia (NC) is an ideal location for the implementation and evaluation of such a strategy as the only proven vector for dengue virus (DENV) is Ae. aegypti and dengue outbreaks frequency and severity are increasing. We report the generation of a NC Wolbachia-infected Ae. aegypti strain and the results of experiments to assess the vector competence and fitness of this strain for future implementation as a disease control strategy in Noumea, NC. Methods/principal findings The NC Wolbachia strain (NC-wMel) was obtained by backcrossing Australian AUS-wMel females with New Caledonian Wild-Type (NC-WT) males. Blocking of DENV, chikungunya (CHIKV), and Zika (ZIKV) viruses were evaluated via mosquito oral feeding experiments and intrathoracic DENV challenge. Significant reduction in infection rates were observed for NC-wMel Ae. aegypti compared to WT Ae. aegypti. No transmission was observed for NC-wMel Ae. aegypti. Maternal transmission, cytoplasmic incompatibility, fertility, fecundity, wing length, and insecticide resistance were also assessed in laboratory experiments. Ae. aegypti NC-wMel showed complete cytoplasmic incompatibility and a strong maternal transmission. Ae. aegypti NC-wMel fitness seemed to be reduced compared to NC-WT Ae. aegypti and AUS-wMel Ae. aegypti regarding fertility and fecundity. However further experiments are required to assess it accurately. Conclusions/significance Our results demonstrated that the NC-wMel Ae. aegypti strain is a strong inhibitor of DENV, CHIKV, and ZIKV infection and prevents transmission of infectious viral particles in mosquito saliva. Furthermore, our NC-wMel Ae. aegypti strain induces reproductive cytoplasmic incompatibility with minimal apparent fitness costs and high maternal transmission, supporting field-releases in Noumea, NC.


Introduction
With an estimated 390 million infected people per year, dengue still represents a major public health problem throughout the tropics [1]. Dengue viruses (DENVs) are transmitted to humans by the bite of infected mosquitoes from the genus Aedes, with Aedes aegypti being the predominate vector. Dengue infection is a re-emerging disease caused by dengue virus (DENV) belonging to the genus Flavivirus. DENVs are divided in four serotypes (DENV-1 to -4), themselves subdivided in genotypes. Infection with one serotype is thought to provide lifelong protection from reinfection with the same serotype but does not prevent secondary infection by another serotype [1,2]. The spectrum of dengue clinical presentations is broad, ranging from asymptomatic to severe, sometimes fatal infections [2].
New Caledonia (NC), a French island territory located in the subtropical Pacific region with a population of approximately 280,000, has a history of recurrent dengue outbreaks. In the past decade, DENV circulation has increased in NC, causing recurrent outbreaks with cases detected every year [3] along with chikungunya virus (CHIKV) and Zika virus (ZIKV) circulation [3]. During the three last major DENV outbreaks in 2008-2009, 2012-2013, and 2016-2018, the NC Health Authorities reported 9,589, 11,240, and 7,266 DENV cases, respectively [3,4]. To date in NC, the only proven vector for DENV is Aedes aegypti. Until recently, the only means of controlling dengue were based on regular public prevention campaigns and vector control measures, consisting of elimination of larval habitats and pyrethroid insecticide application to kill adults. Although these campaigns have decreased the number of larval habitats and mosquitoes [3], this decrease has not been sufficient to prevent dengue circulation [3]. Furthermore, the low efficiency of outdoor space spraying and resistance to the pyrethroid deltamethrin has reduced the efficacy of control of Ae. aegypti adults in Noumea [5].
Given these concerns, population introgression strategy, based on the release of Wolbachiainfected Ae. aegypti mosquitoes in the environment, has been identified as a promising strategy to control dengue in NC. Wolbachia is a Gram-negative bacterium mostly present in arthropods with more than 40-65% of insect species harbouring Wolbachia [6,7]. Mainly transmitted vertically, this bacterium can manipulate the host reproduction in order to maximize its maternal transmission (MT) through the eggs. Females are favored by Wolbachia through parthenogenesis, feminization, male-killing, and cytoplasmic incompatibility (CI) [8]. CI is the most common alteration and occurs when Wolbachia-infected males mate with uninfected females, leading to death of embryos from uninfected eggs, which promotes the spread of Wolbachia and its maintenance in mosquito populations [9]. Wolbachia-infected females can rescue the lethality, providing them with a reproductive advantage over uninfected females [10]. Wolbachia can also alter responses to infections to reduce arbovirus transmission. Wolbachia transinfection into Ae. aegypti thus limits infection with DENV, CHIKV, and ZIKV [11][12][13][14][15]. Combining their ability to invade the host population by inducing CI and to interfere negatively with the transmission of viruses, Wolbachia has been deployed to prevent the transmission of mosquito-borne diseases. The goal is to establish Wolbachia in wild mosquito vector populations and to interrupt local virus transmission from mosquitoes to humans as Ae. aegypti carrying Wolbachia have a lower transmission potential for arboviruses [11,15].
This method is potentially applicable to NC for the following reasons: (i) the targeted mosquito species is Ae. aegypti, which is the only known DENV vector in NC to date [16], (ii) the method has already shown its efficiency elsewhere [17,18], (iii) the implementation of this method is considered to be safe for humans, animals, and the environment [19]. Finally, (iv) this method is self-sustaining through the CI and MT phenotypes which promote the maintenance of Wolbachia in mosquito populations [20].
As highlighted by [10], before being released, the Wolbachia-transinfected Ae. aegypti strain has to be tested to ensure its strong protection against virus replication, to demonstrate high levels of Wolbachia MT as well as CI. Even if the majority of the studies have shown no or low fitness impacts of wMel in Aedes aegypti [21], main fitness parameters have to be assessed to ensure that Ae. aegypti carrying Wolbachia will not be disadvantaged during the introgression in the environment. Here we report the generation of a Wolbachia-transinfected Noumea strain of Ae. aegypti and present results of laboratory experiments to assess its vector competence for DENVs, CHIKV and ZIKV. Wolbachia MT and CI, as well as fitness determinants (fertility, fecundity, and wing length) and insecticide resistance status were also evaluated.

Ethics statement
In NC, human blood for mosquito rearing and artificial blood feeding experiments were obtained from blood donor center (Service de Transfusion Sanguine, NC Hospital), upon consent of patients. Ethical

Mosquito rearing
All strains were reared and maintained in a controlled laboratory environment, at 28˚C ± 1˚C and 80% ± 10% relative humidity, with a 12:12 light: dark photoperiod in NC, and at 26˚C ± 1˚C with 65% ± 10% relative humidity and a 12 h:12 h light: dark photoperiod in Australia (mosquitoes used for intrathoracic injection experiments only). A 10% sucrose solution was provided to adults. Females were blood-fed with human blood twice a week (every 3 or 4 days) with artificial membrane feeding systems (Hemotek, United Kingdom). Eggs oviposited on cups lined with filter-paper were removed at each blood feeding and kept in a humid atmosphere for 48 h to allow embryos to fully develop before being dried. Hatching solution comprising one litter of pre-boiled water and 0.2 g of tetramin (Tetra, Melle, Germany) was prepared 24 h ahead of time. Synchronous hatching was induced by placing eggs in the hatching solution for 24 h, at room temperature. Then, larvae were allocated in trays to obtain a density of 100 larvae for 1.5 L of tap water. Larvae were fed ad libitum with a mix of 10% of yeast and 90% of Protinova (defatted dry powder made of Hermetia illucens larvae, supplied by Innova Feed).

Mosquito strains
The origin, status of Wolbachia infection and use of each strain are summarized in Table 1. The NC wild type strain (NC-WT) was established from larvae collected twice in the field at the same location, in Noumea, NC, in June and November 2018 (978 and 1526 larvae collected, respectively). Each time, this strain was maintained under laboratory conditions for a maximum of two generations to maintain genetic diversity and limit the impacts of inbreeding.
The Australian wMel strain previously described [11,22], referred to as AUS-wMel here, was used for backcrossing, and as a control for vector competence and fitness determinants analyses. The Australian tetracycline strain (AUS-Tet) is the AUS-wMel strain cured of Wolbachia by the use of tetracycline [10]. AUS-Tet was used in comparison with the AUS-wMel strain for this work for vector competence analyses. The NC Wolbachia strain (NC-wMel) was obtained by backcrossing AUS-wMel females with NC-WT males. Six backcrosses were made in order to obtain a strain infected with wMel that was genetically similar to NC-WT mosquitoes (i.e., 98% of the nuclear background similar between NC-wMel and NC-WT strains). Briefly, the first backcross was made between 250 NC-WT males and 250 AUS-wMel females. Then, the five following backcrosses were made between 250 NC-WT males and 250 female progeny from the previous cross. In order to ensure virginity, pupae were sexed and sorted according to size. Only virgin adults were used. Wolbachia frequency was checked at each generation of backcross on 160 3-5-day old females by quantitative qPCR (described below) to ensure the quality of the strain. When the six backcrosses were completed, the NC-wMel strain was maintained in our laboratory.
The first generation of NC-wMel strain after the completion of backcrossing (G0) was used for fitness assays. The NC-wMel from generation G0 and F2 NC-WT were used for the infected blood meal experiments. NC-wMel mosquitoes from generations G2-G5 and F2 NC-WT mosquitoes were used for intrathoracic injection experiments. Results were compared to those obtained for AUS-wMel and a tetracycline-treated version of this line (AUS-Tet) cured of Wolbachia as described previously [10].
Finally, the Bora strain, an Ae. aegypti laboratory strain, was used as an insecticide-sensitive control for deltamethrin resistance tests [23].

Wolbachia detection by qPCR
DNA was isolated from mosquitoes, as previously described [24]. Confirmation of Wolbachia infection status in mosquito tissue samples was performed using a duplex quantitative PCR targeting the Wolbachia-specific wsp gene and Ae. aegypti housekeeping RpS17 gene [25]. For each sample, qPCR was performed using a LightCycler 480 II Instrument (Roche) and the LightCycler 480 Probes Master kit (Roche).

Assessment of virus inhibition in the NC-wMel strain
In this work, artificial infectious blood meal and intrathoracic injection experiments were conducted in NC and Australia respectively.
Viruses and cells. All 4 DENVs, CHIKV, and ZIKV were used in this study (Table 2). Asian isolates of DENV-1 to 4 were obtained from the World Reference Center for Emerging Viruses and Arboviruses (WRCEVA). NC isolates of DENV-2 (this study), CHIKV [26], and ZIKV [27] were obtained from human sera [16] and are representative of recent arboviruses circulation in NC. Virus genotypes and origins are listed in Table 2.
Artificial infectious blood meal. For infectious blood meals, mosquitoes were reared under the same conditions as for the maintenance of the strains (see mosquito rearing section). Five-to-7 day-old nulliparous females of NC-WT and NC-wMel strains were allowed to feed for 20 minutes on a blood meal containing virus maintained at 37˚C using an a Hemotek system (Hemotek Limited, Great Harwood, UK) covered with pig intestine membrane [16]. Each blood meal contained either CHIKV, ZIKV, or DENV-2 diluted to the concentrations listed in Table 2 and was supplemented with a phagostimulant (5mM ATP). Fully engorged females were then transferred into cardboard containers covered with insect netting and maintained with 10% sucrose solution at 28˚C ± 1˚C, 80% relative humidity under a 12 h:12 h light:dark cycle (NC laboratory standard conditions). At 3 (only for CHIKV), 7, and 14 days post-exposure, a maximum of 30 mosquitoes of each Ae. aegypti population were randomly selected. Their saliva, head, and body were collected, ground (head and body), and treated as previously described [16]. Finally, 45 μL of DMEM (for CHIKV and ZIKV) or Leibovitz medium (for DENV) were added to the collected saliva. All the samples were stored at -80˚C. The detection of viral particles in each homogenate was performed by virus titration by plaque assay on VERO E6 cells for ZIKV and CHIKV and by immune-fluorescent focus assay on C6/36 cells for DENV [16,28]. The infection rate corresponds to the proportion of mosquitoes with infected bodies among all those tested. The proportion of mosquitoes with viral particles detected in saliva among all mosquitoes tested (i.e., engorged) represents the transmission efficiency.
Intrathoracic injection experiments. These studies were conducted in Australia and mosquitoes were reared as described in [29]. Seven-to-8 days old AUS-wMel, AUS-Tet, NC-wMel, and NC-WT Ae. aegypti were intrathoracically injected as previously described [24] with 69 nL of viruses diluted in RPMI to the concentrations listed in Table 2 using a

Fitness determinants
For fitness assays, larvae and adults were reared under the same conditions as for strain maintenance (see mosquito rearing section). Between five and eight trays for each strain were created (300 larvae for 3 L of water). After larval development, the pupae were sexed by their size and placed in cups for emergence (25 male or female pupae per cup). After adult emergence, and to guarantee that adults were virgin before crossing, solely the cups containing only males or only females were used. Different crosses have been made to evaluate MT, CI, and fertility of the different strains. Those crosses consisting of a group of 50 virgin males and 50 virgin females were performed with strains varying according to each test. Each cross was replicated three times. When adults were between 5 and 7 days old, females were blood-fed with human blood collected from donors treated with therapeutic phlebotomy (blood donor center: Service de Transfusion Sanguine, NC Hospital). The number of living females and blood-fed females were counted immediately after the blood meal. Then, 3 days after the blood meal, one egg cup with wet filter paper was placed each cage for 3 days to allow oviposition. Eggs were kept in the humid atmosphere of the insectarium for 48 h to allow embryos to fully develop before being dried. One week after egg production, filter papers were split to obtain five batches of around

Wolbachia fitness determinants
Maternal transmission. To quantify the success of Wolbachia MT, MT crosses were set up between WT males and Wolbachia-infected females (crosses of NC-WT males x NC-wMel females, compared to crosses of NC-WT males x AUS-wMel females). One week after egg production, eggs were submerged in hatching solution separately for the three replicates of each cross. Larvae were reared to 4-6 day old adults, then 160 females of each replicate were sampled and screened by qPCR to detect Wolbachia.
Cytoplasmic incompatibility. To investigate the level of Wolbachia-induced CI, CI crosses were set up between Wolbachia-infected males and WT females (crosses of NC-wMel males x NC-WT females, compared to crosses of AUS-wMel males x NC-WT females). One week after the eggs were produced, five egg batches from each cage were submerged individually in hatching solution and first instar larvae counted the following day. CI was estimated by dividing the total number of hatched larvae by the total number of counted eggs for each replicate.

Mosquito fitness determinants
Fertility. Fertility crosses were set up between males and females of the same strains (NC-WT males x NC-WT females; NC-wMel males x NC-wMel females, and AUS-wMel males x AUS-wMel females). The same protocol as for CI experiments was applied to the five batches of eggs from each cage. Fertility was determined by dividing the total number of 1 st instar larvae by the total number of eggs counted for each replicate.
Fecundity. Fecundity was assessed on fertility crosses. For each female strain, fecundity was estimated by the mean number of eggs laid per blood-fed female per cage.
Wing length. Wing length, as a proxy measurement of body size, was measured for 30 to 40 specimens of each sex and strain. The wing was removed and placed between slide and cover slip on a white surface. A picture was taken with a camera (Leica DMC2900) plugged into a stereomicroscope (Leica M205C). The measurement was made thought the LAS X software of Leica (V3.0.4). Wing length was calculated as the distance from the wing base to the wing tip.
Insecticide resistance. The insecticide susceptibility tests were conducted with standard WHO test tubes [31] on NC-wMel, NC-WT, and Bora Ae. aegypti strains. This device allows exposing sets of 25 adult females (2-5 days old) to a filter paper impregnated with insecticide. For each dose, the insecticide was diluted in a mixture of acetone and silicone oil and 2 mL of solution was applied to each paper. Different doses of deltamethrin were tested for the resistant strains: 0% (control), 0.02%, 0.05%, 0.1%, 0.4%, and 0.9%. For the sensitive strain, doses of 0%, 0.0004%, 0.001%, 0.003%, 0.005%, and 0.01% were used. For each strain and each dose tested, four exposure tubes containing around 25 females (2-5 days old) were used. Females were exposed for 1 h. After exposure, 10% sugar solution was provided to females and mortality was recorded at 24 h.

Statistical analysis
Statistical analysis and graphics were performed using R software (R Core Team (2017). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria). Comparisons of proportions were made using Fisher's exact tests. For continuous data, the normality was assessed by group using a Shapiro-Wilk normality test. ANOVA was used to compare means across multiple groups. If any differences were found, groups were compared two by two using a Student's t-Test. Non-parametric Kruskal-Wallis test was carried out to compare multiple groups when number of replicates was low. Comparison between two groups were made using Wilcoxon test. If multiple tests were performed, the p-values were adjusted using the Holm method. The statistical significance threshold for these tests was set at 0.05.
The analyses of dose-mortality responses were performed using the R script BioRssay 6.2 [32,33] as previously described [34]. Briefly, this script computes the Lethal Doses of insecticide killing 50% of the tested strains (LD 50 ) and the associated confidence intervals. The comparison of strains was made by calculating the Resistance Ratios 50, or RR 50 (= LD 50 of tested strain/LD 50 of the sensible strain) and their 95% confidence intervals (CI 95 ). A RR 50 in which the confidence interval does not include 1 was considered statistically significant.

Vector competence
Oral challenge with infectious blood meals. NC-WT and NC-wMel Ae. aegypti were orally challenged with DENV-2, CHIKV or ZIKV. Pairwise comparisons of infection rates showed that NC-wMel mosquitoes were significantly less susceptible to DENV-2, CHIKV, and ZIKV compared to NC-WT Ae. aegypti irrespective of the day post infection (Fisher's exact test, p-values < 0.001; Fig 1A-1C). Infection rates of NC-WT mosquitoes reached more than 90% irrespective of the day of post infection, except for ZIKV for which infection rates ranged from 72% to 85%. For NC-wMel mosquitoes, infection rates ranged from 13% to 39%, 0% to 13%, and 0% to 10% for DENV-2, CHIKV, and ZIKV respectively. No infection was detected for CHIKV-exposed NC-wMel Ae. aegypti at 14 days post challenge and for ZIKV-exposed NC-wMel Ae. aegypti at 7 days post-challenge.
Infectious viral particles were detected in saliva of NC-WT mosquitoes for the three viruses tested irrespective of the day post infection, except for ZIKV-infected mosquitoes at 7 days post-infection (Fig 1D-1F). Overall, transmission efficiencies did not exceed 20% for NC-WT mosquitoes. No transmission was observed for NC-wMel Ae. aegypti regardless of the viruses tested and the incubation time. Significant differences were observed in pairwise comparisons of transmission efficiencies between NC-WT and NC-wMel mosquitoes for CHIKV at 3 days and for ZIKV and 14 days (Fisher's exact test, p-values = 0.01 for CHIKV and 0.02 for ZIKV).
DENV viral titers in NC-wMel and AUS-wMel infected mosquitoes were also significantly lower compared with those obtained for NC-WT and AUS-Tet respectively regardless of the serotype (Wilcoxon test, p-values < 0.001; Fig 2E-2H). The medians of viral titers for NC-WT and AUS-Tet ranged from 1.1 x 10 5 to 1.3 x 10 6 DENV copies/mosquito, whereas the medians of DENV viral titers for NC-wMel and AUS-wMel ranged from 6.2 x 10 3 to 1.6 x 10 5 DENV copies/mosquito.

Wolbachia fitness determinants
Maternal transmission. MT is one of the key factors for the introgression of Wolbachia in field mosquito populations. The MT rate measured for the NC-wMel strain was high with 96% of offspring infected on average, compare with 98% for the AUS-wMel strain (Table 3). No significant difference was observed between these two strains (Wilcoxon test, pvalue = 0.7). Cytoplasmic incompatibility. CI is the second factor favouring the introgression of Wolbachia in field populations. For eggs obtained from the cross of NC-wMel males and NC-WT females, no larvae were observed (total number of eggs = 4174). The same result was observed for AUS-wMel males crossed with NC-WT females showing that NC-wMel was as effective as AUS-wMel males at inducing CI (total number of eggs = 2048) ( Table 4).

Mosquito fitness determinants
Fertility. In contrast to incompatible crosses, fertility crosses (males crossed with females of the same strain) produced viable eggs, with mean hatch rates of 53%, 62%, and 77% for NC-wMel, AUS-wMel, and NC-WT respectively (Table 4), with significant differences between crosses (Kruskal-Wallis Test, p-value = 0.04).
Fecundity. Regarding fecundity, the mean number of eggs laid per blood-fed female per cage was significantly different for all strains (Kruskal-Wallis Test, p-value = 0.04). The AUS-wMel strain had the highest fecundity with on average 61 eggs laid per female, while NC-WT and NC-wMel females laid approximately 43 and 22 eggs per females respectively (Table 5).
Wing length. Mean wing lengths were significantly impacted by strain and sex of the mosquitoes (ANOVA; p-values < 0.001). For females, all wing length means were significantly different with wing lengths of NC-WT specimens shorter (mean = 3.15 mm) when compared with Wolbachia-infected strains (means = 3.40 and 3.32 mm for NC-wMel and AUS-wMel respectively) (Student's t-Tests; p-values < 0.009) (Fig 3). The same trend was observed for males, with lower average wing length for the NC-WT strain (2.37 mm) compared to both Wolbachia-infected strains (2.52 and 2.54 mm for NC-wMel and AUS-wMel respectively; Student's t-Tests; p-values < 0.001). No significant difference was found between the mean wing lengths of males of the two strains carrying Wolbachia (Student t-Test; p-value = 0.53).

Discussion
Biological control programs involving Wolbachia-infected Ae. aegypti must be sufficiently robust to limit arbovirus transmission in different epidemiological settings. NC is an ideal

PLOS NEGLECTED TROPICAL DISEASES
New Caledonia Aedes aegypti wMel strain epidemiological context for the implementation and evaluation of such a strategy as the only proven vector for dengue is Ae. aegypti, DENV outbreaks are frequent, and efficient epidemiological and entomological networks are deployed [3]. We report the development of the NC-wMel Ae. aegypti strain and present results of experiments assessing its vector competence, MT, CI, fitness determinants and insecticide resistance status, as the first steps in implementing Wolbachia to limit arbovirus transmission in NC.
As expected, NC-wMel Ae. aegypti orally challenged with DENV, ZIKV or CHIKV were significantly less susceptible to infection than WT mosquitoes. More importantly, in our study, no NC-wMel Ae. aegypti were able to transmit the arboviruses tested whereas WT mosquitoes did as reported previously [16,26,28]. These data were in accordance with previous studies on pathogen blocking in wMel-infected mosquitoes, which indicate that strong pathogen blocking occurs against a range of viruses [12,13,35,36]. In an intrathoracic virus challenge model, the midgut barrier is bypassed; high infection rates, and even more importantly, high dissemination rates can be achieved more quickly. NC-wMel strains demonstrated significantly reduced infection rates and viral titers with all DENV tested. To our knowledge, this study is the first to assess Wolbachia-blocking in a recently backcrossed wMel strain using both oral feeding with epidemiological relevant viruses and IT with reference viruses. By standardizing virus dose and bypassing midgut barriers to infection and dissemination, IT gives a standardized evaluation of virus blocking by Wolbachia infection. Conversely, oral feeding has the advantage of being more representative of the life-cycle of the virus in the vector as it recapitulates intrinsic barriers to infection. With significant results with both techniques, the findings presented here indicate the blocking observed in this new NC wMel-infected Ae. aegypti strain is robust.
Even if strong virus blocking is exhibited in the NC-wMel strain, its effectiveness depends on its ability to obtain high levels of introgression of Wolbachia in Ae. aegypti field populations via MT and CI. MT of Wolbachia by the new NC-wMel strain of Ae. aegypti was comparable to other strains used in field deployment in other countries [10,11,37]. The Ae. aegypti NC-wMel also showed complete CI. These results suggest that the NC-wMel should introgress and be maintained in the field at high frequency, as shown in other trials [38].
Transinfection of Ae. aegypti by Wolbachia may be costly to mosquito fitness. These costs vary depending on the Wolbachia strain [9,10,39] and, to some extent, can negatively impact the introgression and long-term stability as observed with the wMelPop strain whose high fitness costs do not allow its maintenance in the field [40]. Among the various fitness parameters that we studied using the NC-wMel strain, several seemed to have been impacted by Wolbachia infection, in particular the fecundity and fertility of females. Females of the NC-wMel strain laid fewer eggs than NC-WT females and the egg hatch rate was reduced compared to NC-WT eggs. In the future, it would be interesting to increase the time for embryogenesis (i.e., before drying the eggs) for NC-wMel strain in order to improve the hatching rate, as previously described [41]. The relatively low fertility and fecundity of the NC-wMel strain may be due to the cumulative fitness costs of Wolbachia and insecticide resistance. In fact, this phenomenon has already been observed in Brazil where a strain highly resistant to insecticides exhibited fecundity and hatch rates close to those we observed in our study [42]. These levels of Wolbachia cost on the fertility and fecundity of Ae. aegypti have already been observed in Australian and Brazilian strains, without impacting the establishment of wMel in field populations [38,42]. The number of replicates on the fecundity and fertility experiments was limited. They should be repeated on isolated female mosquitoes to ensure that this cost does not limit the installation of Wolbachia in the field. Other fitness parameters were evaluated. No reduction in wing size was observed in mosquitoes infected with wMel. In contrast, a slight, but significant, increase in the size of the wings of males and females of wMel-infected strains (NC and AUS) was observed. Although this difference in wing size between the NC-WT and AUS-wMel strains could be due to a genetic background difference, it is less likely that wing-size difference between the NC-WT and NC-wMel strains relies on genetic background differences. Indeed, the NC-wMel strain was obtained through six generations of backcrossing with the NC-WT strain; these two strains therefore share a large part of their nuclear genome. This observed increase in wing size is consistent with previous work by [39] in which a similar increase in the size of Wolbachia-infected adults was observed, which was potentially due to a longer duration of larval development and a positive impact of Wolbachia.
The last fitness component that can negatively impact the establishment of Wolbachia in the field is a difference in insecticide resistance levels. If insecticide treatments are applied, a susceptible mosquito strain will be disadvantaged compared with resistant WT mosquitoes. This phenomenon was observed in Brazil, where the first Wolbachia-infected Ae. aegypti strain released exhibited much lower resistance levels than the field populations. As a result, Wolbachia did not establish in this first trial, necessitating the generation of a new insecticide resistant strain [42]. To avoid this problem, the NC-WT strain was generated through backcrossing with a field strain resistant to deltamethrin, the only insecticide currently used in Noumea. The level of resistance to deltamethrin was similar between NC-wMel strain and WT field mosquitoes, which should allow Wolbachia establishment in Noumea, even in the presence of insecticide treatments.
Such an operational program needs to be carefully monitored to assess the evolution of wMel Ae. aegypti strain in the environment. Abiotic and biotic factors could impact the strategy, particularly viral adaptative evolution and specific environmental conditions that may impact Wolbachia efficiency in the long term [43]. In the long-term, DENV genetic adaptation is eventually expected. This could reduce complete blocking by Wolbachia but partial DENV blocking should persist indefinitely [44], still reducing the risk of dengue outbreaks. Concerning environmental conditions, heat stress caused by elevated ambient temperatures (above 30˚C during the day in summer in Noumea) could have a negative impact on Wolbachia density [45]. However, it should be transient [46] given the temperatures recorded in Noumea (25˚C in average in dry season) [47]. The strategy could also be affected by the introduction of new species of mosquitoes, especially those transmitting arboviruses as different species have been recorded in neighboring islands [48,49]. In New Caledonia, Aedes scutellaris has been  [50,51]. Since then, no new detection occurred despite a regular monitoring specifically dedicated to this species. To reduce the risk of introduction, a regular monitoring of the main international entry points is managed by the Department of Health and Social Affairs of New Caledonia in the framework of the International Sanitary Regulations.
Despite the different factors which could influence the strategy, results now available from the field show that the method is stable after several months [25] to several years [52]. Furthermore, the Vector Control Advisory Group (WHO) has recently concluded that "wMel introgression into populations of Ae. aegypti demonstrates public health value against dengue" [53]. As Ae. aegypti is the only proven vector for dengue in NC, we expect a high efficiency of the method on the reduction of dengue and other arboviruses cases in Noumea as recently shown in Yogyakarta, Indonesia [18].