Seasonal Phenological Patterns and Flavivirus Vectorial Capacity of Medically Important Mosquito Species in a Wetland and an Urban Area of Attica, Greece

Seasonal patterns of mosquito population density and their vectorial capacity constitute major elements to understand the epidemiology of mosquito-borne diseases. Using adult mosquito traps, we compared the population dynamics of major mosquito species (Culex pipiens, Aedes albopictus, Anopheles spp.) in an urban and a wetland rural area of Attica Greece. Pools of the captured Cx. pipiens were analyzed to determine infection rates of the West Nile virus (WNV) and the Usutu virus (USUV). The data provided were collected under the frame of the surveillance program carried out in two regional units (RUs) of the Attica region (East Attica and South Sector of Attica), during the period 2017–2018. The entomological surveillance of adult mosquitoes was performed on a weekly basis using a network of BG-sentinel traps (BGs), baited with CO2 and BG-Lure, in selected, fixed sampling sites. A total of 46,726 adult mosquitoes were collected, with larger variety and number of species in East Attica (n = 37,810), followed by the South Sector of Attica (n = 8916). The collected mosquitoes were morphologically identified to species level and evaluated for their public health importance. Collected Cx. pipiens adults were pooled and tested for West Nile virus (WNV) and Usutu virus (USUV) presence by implementation of a targeted molecular methodology (real-time PCR). A total of 366 mosquito pools were analyzed for WNV and USUV, respectively, and 38 (10.4%) positive samples were recorded for WNV, while no positive pool was detected for USUV. The majority of positive samples for WNV were detected in the East Attica region, followed by the South Sector of Attica, respectively. The findings of the current study highlight the WNV circulation in the region of Attica and the concomitant risk for the country, rendering mosquito surveillance actions and integrated mosquito management programs as imperative public health interventions.


Introduction
Research on the distribution, abundance, and species composition of mosquitoes at a regional level is vital to estimate the risk of incidence of vector-borne diseases that are currently increasing in Europe, because of range expansion of native species and invasion events by alien species [1][2][3][4]. Factors, such as globalization of travel and trade, increasing land use and urbanization, high concentration of human populations, socioeconomics, and (i) Monitoring and recording of mosquitoes' species and population densities in the RUs under study; and (ii) detection and monitoring of the circulation of WNV and USUV in collected Cx. pipiens s.l. for possible co-circulation.
This manuscript aimed to highlight the importance of implementing surveillance programs for the prompt detection of viruses' circulation in mosquito populations and present, for the first time, surveillance data for USUV in mosquitoes.

Study Area
Two out of eight distinct RUs of Attica comprised the main study area, where traps were installed (Figure 1). The selection of the two RUs was based on the different environmental types, namely urban (UR) and wetland (WT) areas, as described in Table 1. Depending on the research coverage area in each RU, the corresponding number of traps was placed. Additionally, the epidemiological profile of the selected RUs of Attica played an important role concerning the risk for the residents, due to past WNV human infections in both RUs and previously recorded malaria cases in the East Attica Sector being reported to the NPHO since 2010 [38,63]. Data were collected from June 2017 to December 2018. The region of Attica, "Attiki" in the Greek language (38.0 • N 23.7 • E; total area: 3808.10 km 2 ; population: 3,828,434 inhabitants (2011 record data) [64], is the main metropolitan region of Greece, located on the eastern edge of the mainland, in Central Greece. The greater area of Attica region includes Athens (the capital of Greece) and Piraeus along with 62 other cities and settlements. It is bordered by the sea, to the east, including the south and southwest, while four mountains, Egaleo, Parnitha, Penteli, and Hymettus, delineate the hilly plain [65].  [64]) covers the eastern part of the urban agglomeration of Athens, and also the rural area to its east. It is the only Attica zone with significant agricultural activity, can be considered geographically isolated from the rest of the basin, and is inserted between the Penteli and Hymettus mountains. The selected RU of MS (38 • 9 N 23 • 57 E; total area: 222.75 km 2 ; population: 33,423 inhabitants [64]) is located outside the Athens Basin in the northeast part of the Attica region. The MS area lies 42 km away from the center of Athens. Marathonas is an area of intense agricultural activity, while Schinias is an area of marsh and coastal forest. The National Park of Schinias constitutes the most important coastal ecosystem in Attica [66,67].
(  [64]) is a coastal district, situated on the east coast of the Phalerum Bay of the Saronic Gulf, 6 km southwest of the Athens city center. The seaside area of PF is an important touristic attraction with a seaside promenade, several sports venues, and a marina. The Pikrodafni stream flows into the sea on the border of the RUs of Palaio Faliro and Alimos [68].

Mosquito Collection and Identification
The selection of the mosquito monitoring stations was performed following an on-site visit, and it was based mainly upon ecological and social characteristics, such as urban and rural sites, presence of vegetation and shading, occurrence of humans or livestock as potential hosts for adult mosquitoes, proximity to open sources of fresh or still water, nuisance complaints, and convenience of sampling.
The BG-sentinel trap (BGs) (Biogents AG, Regensburg Germany) baited with CO 2 and BG-Lure [69], which is considered an effective method for mosquito diversity and abundance characterization [2,18,32], was selected as the main monitoring tool.
The composition of mosquito fauna was investigated by the monitoring system of seven BGs traps that was established in seven selected monitoring stations. In particular, depending on the research coverage area in each RU, four traps were placed in MS area, and three in PF area, respectively ( Figure 1). A summary of all sampling data concerning the two studied areas is given in Table 1.
All collected mosquito samples were transferred weekly to the Laboratory of Medical Entomology, of Public Health Policy at the University of West Attica for further analysis. Closed and chilled containers containing dry ice were used for the transportation of samples, under the scope of morphological identification of mosquitoes and molecular detection of viruses (WNV and USUV), respectively.
Throughout the duration of the study, no male Culex torrentium adults were identified regarding the morphological identification of the members of Cx. pipiens s.l. complex [70,71]. Adult females were characterized morphologically as Cx. pipiens s.l., as the two species are indistinguishable morphologically [70].
Adult mosquitoes that were morphologically identified to belong to Anopheles maculipennis s.l. complex were further examined by molecular amplification methods, according to previously described protocols [74,75].

Flaviviruses Survey in Culex pipiens Pools
Screening of Culex pipiens s.l. pools was designed to evaluate the possible co-circulation of the two viruses (WNV and USUV) in the same mosquito pool. All collected mosquitoes were maintained under cold chain conditions to preserve the virus viability, pooled according to the collection site, date, species, and sex (up to a maximum of 200 individuals per pool), and stored at −80 • C.

WNV and USUV Detection
A total of 366 Cx. pipiens s.l. pools were analyzed for WNV (collected under the current surveillance program). Genetic material (RNA) from mosquito pools was extracted by using the Maxwell 16 Automated Nucleic Acid extraction system (Promega, Madison, WI, USA), according to the manufacturer's instructions for Maxwell16LEV Simple RNA Tissue kit [13]. A TaqMan Real-Time PCR protocol, specific for WNV lineages 1 and 2 detection, was implemented thereafter [76].
These 366 Cx. pipiens s.l. pools were also analyzed for the detection of USUV, with a reverse transcription real-time PCR protocol [77] and for selected samples also with a reverse transcription conventional PCR protocol [78].
Screening of the above-mentioned pools was designed to evaluate the possible cocirculation of the two viruses (WNV and USUV) in the same pool.

Infection Rates
The minimum infection rates (MIR) and maximum likelihood estimation (MLE) were calculated using the PooledInfRate program version 4.0 (available at https://www.cdc. gov/westnile/resourcepages/mosqSurvSoft.html, accessed on 20 August 2021) [13,79]. For each region included in the study areas, the respective MIR and MLE values were calculated per 1000 mosquitoes tested. The MLE has the advantage of considering variations in pool size, while the MIR in a study area assumes the presence of a single positive mosquito in a pooled sample.

Molecular Methods for Identification of Anopheles Mosquitoes
Anopheles species have been incriminated as vectors for transmission of malaria worldwide [80]. Nevertheless, a common limitation in identifying morphologically related species (i.e., An. maculipennis s.l. complex) creates an urgent need for implementation of alternative laboratory approaches.
Unidentified by morphological characteristics, specimens of Anopheles spp., as well as subspecies belonging to the An. maculipennis complex (cryptic species), were subjected to molecular identification by PCR. The nucleotide sequence variation of the ITS2 ribosomal region (ITS2 rDNA) along with the nucleotide sequence of the mitochondrial gene region I (COI) of cytochrome oxidase, respectively, were used as the main targets of the implemented molecular protocol [81][82][83].
A total of 50 Anopheles adult mosquitoes, the majority of which were morphologically identified, were examined at the molecular level. Representative samples were isolated and sent for sequencing analysis [81,82].

Data Analysis
A Gaussian Generalized Estimating Equation (GEE) model was used to estimate the number of captures in the two areas, Faliro and Marathonas. GEE analysis was conducted using the package "geepack" [84,85] in R v4.0.0 (R Core Team 2013, R Foundation of Statistical Computing, Vienna, Austria). According to the results in Table 2 from the GEE analysis for the Culex and Aedes species collected in both study areas, we observed the following: Culex: there were significantly less Culex captures in Faliro than in Marathonas (p = 0.028).

Mosquito Fauna Identification
Aedes: the number of captures did not differ significantly between Faliro and Marathonas (p = 0.441).
A total of 37.810 (80.92%) individuals were captured in the MS area, corresponding to six genera and 15 species. Additionally, 8.916 (19.08%) individuals were captured in PF area, corresponding to three genera and four species, respectively. Three of the captured species, namely Anopheles sacharovi, Culex pipiens s.l., and Aedes albopictus, are of major medical importance. Furthermore, Cx. pipiens s.l. (88.25%) and Ae. albopictus (8.05%) were by far the most abundant from all the collected species (Table 3).
A total of 531 adult sampling collections (MS, n = 295; PF, n = 236) from established BGs traps were examined from both studied areas, as described in Table 1. Mosquito collection was carried out by 495 sampling collections (MS, n = 277; PF, n = 218), while 36 BGs (MS, n = 18; PF, n = 18) were problematic due to either technical failure that occurred while in operation or without catches, possibly due to the effectiveness of the local mosquito control programs conducted during the transmission period or due to unstable weather conditions, mainly in winter.
The observations regarding the adult sampling collections of the four BGs traps in the MS study area showed differences regarding species abundance, richness, and diversity. Comparing the findings from the four BGs traps in the MS area, we observed the largest numbers of mosquitoes were collected in the MS1 (33.6% of total MS catches) and MS4 (40.5%) traps, followed by MS2 (15.4%) and MS3 (10.5%), while the variety of species was enriched in the MS3 and MS4 traps, respectively (Table 3). Regarding the adult sampling collections of the three BGs traps in PF area, no differences were observed concerning the species diversity. A large number of mosquitoes collected in the PF1 (47.35% of total PF catches) trap were followed by the PF3 (38.72%) and PF2 (13.93%) traps, respectively (Table 3). Taking into consideration that the urban habitats contained more densely humanpopulated areas than the rural habitats, which had a higher density of livestock, we concluded that species richness and diversity recorded in the surveyed municipalities were within the expected range [32].
The results for Cx. pipiens s.l. and Ae. albopictus populations' fluctuations per week, from June 2017 to December 2018, concerning the surveyed RUs of MS and PF are presented in Figures 2 and 3, respectively. The results for Anopheles spp. population fluctuations per week, for the studied period concerning the surveyed RU of MS, are presented in Figure 4. presented in Figures 2 and 3, respectively. The results for Anopheles spp. population fluctuations per week, for the studied period concerning the surveyed RU of MS, are presented in Figure 4.   presented in Figures 2 and 3, respectively. The results for Anopheles spp. population fluctuations per week, for the studied period concerning the surveyed RU of MS, are presented in Figure 4.   High numbers of Cx. pipiens s.l. were observed in the MS RU from June to September 2017, while in 2018, the population reached a peak in June and then remained relatively low the following months ( Figure 2). In the RU of PF, the populations of Cx. pipiens s.l. were kept low during both years of entomological surveillance (Figure 2).
A gradual increase in Ae. albopictus population was recorded in the MS area since June, reaching a peak in August, followed by a gradual decline in 2017, while populations remained in low numbers in the year 2018 ( Figure 3). In the PF area, relatively low numbers of Ae. albopictus were recorded in 2017, while there was a peak in July 2018, followed by a gradual decline in the upcoming months (Figure 3).  High numbers of Cx. pipiens s.l. were observed in the MS RU from June to September 2017, while in 2018, the population reached a peak in June and then remained relatively low the following months ( Figure 2). In the RU of PF, the populations of Cx. pipiens s.l. were kept low during both years of entomological surveillance (Figure 2).
A gradual increase in Ae. albopictus population was recorded in the MS area since June, reaching a peak in August, followed by a gradual decline in 2017, while populations remained in low numbers in the year 2018 ( Figure 3). In the PF area, relatively low numbers of Ae. albopictus were recorded in 2017, while there was a peak in July 2018, followed by a gradual decline in the upcoming months ( Figure 3).
Of particular importance was the presence of An. sacharovi collected in the MS RU, showing an increase in the population in June 2017, then a gradual decrease of catches the following months, and a slight increase was observed in December 2017. In 2018, the catches of the above species were zero during the period of collection ( Figure 4). There were no collections of An. algeriensis in 2017; this species appeared in the area from March 2018, recording high numbers in April, May, and June, and reaching the peak in May; nevertheless, the population decline from July onwards (Figure 4). An. claviger was first collected in July 2017, reaching peak capture rates in August, and then gradually declined in the upcoming months. In 2018, the mosquito populations were kept in low numbers, and a few catches were recorded during the summer months ( Figure 4).

Flaviviruses Detection
Of the adult female Cx. pipiens s.l. captured, a total of 41.050 (MS, n = 34.358; PF, n = 6.692) were examined in pools for the presence of WNV and USUV. Cx. pipiens s.l. adults were treated as a single entity, without determining the relative composition of molestus and pipiens forms.
Out of the 366 mosquito pools tested, a total of 38 (10.4%) tested positive for WNV, including 30 positive samples in MS and 8 positive samples in PF, respectively ( Table 4).
The 366 mosquito pools tested for USUV were found to be negative (Table 4). One single pool of 200 Cx. pipiens s.l. collected in MS in June 2018 was suspected to be possibly USUV positive (low signal upon real-time PCR assay). A reverse transcription conventional PCR protocol was performed for further analysis. The sample produced a PCR product of low intensity (faint band), which did not contribute much to resolving this Of particular importance was the presence of An. sacharovi collected in the MS RU, showing an increase in the population in June 2017, then a gradual decrease of catches the following months, and a slight increase was observed in December 2017. In 2018, the catches of the above species were zero during the period of collection (Figure 4). There were no collections of An. algeriensis in 2017; this species appeared in the area from March 2018, recording high numbers in April, May, and June, and reaching the peak in May; nevertheless, the population decline from July onwards (Figure 4). An. claviger was first collected in July 2017, reaching peak capture rates in August, and then gradually declined in the upcoming months. In 2018, the mosquito populations were kept in low numbers, and a few catches were recorded during the summer months ( Figure 4).

Flaviviruses Detection
Of the adult female Cx. pipiens s.l. captured, a total of 41.050 (MS, n = 34.358; PF, n = 6.692) were examined in pools for the presence of WNV and USUV. Cx. pipiens s.l. adults were treated as a single entity, without determining the relative composition of molestus and pipiens forms.
Out of the 366 mosquito pools tested, a total of 38 (10.4%) tested positive for WNV, including 30 positive samples in MS and 8 positive samples in PF, respectively ( Table 4).
The 366 mosquito pools tested for USUV were found to be negative (Table 4). One single pool of 200 Cx. pipiens s.l. collected in MS in June 2018 was suspected to be possibly USUV positive (low signal upon real-time PCR assay). A reverse transcription conventional PCR protocol was performed for further analysis. The sample produced a PCR product of low intensity (faint band), which did not contribute much to resolving this issue. The PCR product was subsequently sent for sequencing analysis; however, due to the possible reduced concentration of DNA in the sample, the results were inconclusive and therefore the sample was not confirmed as a positive one. WNV screening in the same sample showed that it was positive for WNV.

Infection Rates
In the present study, MLE values both for the wetland and the urban area were almost zero (Table 4), suggesting very low circulation of the virus in the study areas, which is in accordance with the human WNF cases recorded (no cases for 2017 and three cases in the MS area in 2018) [38]. MIR was calculated by extrapolation from the real-time PCR results (the total number of positive pools in the area/total number of mosquitoes sampled in this area × 1000) and is presented in Figure 5. MIR rates indicate that the peak in WNV-infected mosquitoes coincides with high numbers of Cx. pipiens populations. issue. The PCR product was subsequently sent for sequencing analysis; however, due to the possible reduced concentration of DNA in the sample, the results were inconclusive and therefore the sample was not confirmed as a positive one. WNV screening in the same sample showed that it was positive for WNV.

Infection Rates
In the present study, MLE values both for the wetland and the urban area were almost zero (Table 4), suggesting very low circulation of the virus in the study areas, which is in accordance with the human WNF cases recorded (no cases for 2017 and three cases in the MS area in 2018) [38]. MIR was calculated by extrapolation from the real-time PCR results (the total number of positive pools in the area/total number of mosquitoes sampled in this area × 1000) and is presented in Figure 5. MIR rates indicate that the peak in WNV-infected mosquitoes coincides with high numbers of Cx. pipiens populations.

Anopheles Specimens' Molecular Identification
The species Anopheles maculipennis s.l., Anopheles sacharovi, and Anopheles algeriensis were identified by PCR and RFLP analysis. Further analysis of the sequencing chromatograms in the ITS2 gene identified An. maculipennis s.l. and An. sacharovi species in eight of the nine tested samples, with the associated traps located in the marsh and the rural environment, respectively. Species identification based on PCR amplification of the COI gene following sequencing of the 522 bp fragment revealed the presence of An. algeriensis and An. sacharovi, with the relevant traps located in marsh, rural, and semi-arid areas, respectively.
Vector competence plays an important role in vectorial capacity helping in identifying species that might be important contributors to flaviviruses transmission, implementing control measures to reduce the potential of WNV/USUV transmission [28,42], and indicating the possible role of supporting the spread of WNV/USUV during winter [20,53,88,[92][93][94][95].
In this study, mosquito screening for WNV showed that the majority of positive samples for WNV were detected in the East Attica, followed by the South Sector of Attica (Table 4). In the RU of PF, WNV-positive pools in mosquitoes were detected in both years, while in the study of Bisia et al. (2020) [18], no positive pool was detected in 2018. The region of MS has a warm temperate climate with hot dry summers and mild winters and displays characteristics to sustain WNV transmission cycles [96]. Due to its ecological and geographical features, this region is considered a risk area for flavivirus transmission [67].
The higher diversity and abundance of mosquito fauna were observed in Marathonas (intense agricultural activity) and Schinias (swamp and coastal forest), confirming similar remarks from previous studies [67,97]. The spatiotemporal monitoring of land cover changes studied by Gaitanis et al. (2015) [66] in the RU of MS showed a reduction of the areas covered by semi-natural and agricultural and cover types (forests, wetlands, shrublands, and cultivated fields) and the increase of urban and mixed areas during the last 60 years. According to the results of the study and records from the resident population, the mosquito nuisance is serious from early spring onwards [97].
Regarding data on the WNV circulation in equids and birds, according to the Ministry of Rural Development and Food, no cases were detected in 2017 in the Attica region. For 2018, confirmed WNV cases in equids were recorded in West Attica and East Attica RUs, and canary birds that were positive for the virus were also detected in the Athens West Sector RU [98].
For 2018, Greece reported 317 WNV infections and 51 deaths, representing 20% of total EU cases and displaying a 6,6 higher rate than in 2017, with 48 human cases being recorded. Regarding the areas of the present study, 11 and 14 human cases were recorded in the South Sector of Athens and the East Attica RUs, respectively, with a total number of 160 cases being recorded in the Attica region for 2018 [99].
Minimal infection rates of Cx. pipiens adults for WNV in both study areas were in accordance with the mosquito population density, and the low infection rates detected are consistent with the low human case rates observed. No human cases were recorded for 2017 and three cases were detected in the wetland area in 2018 [38,99].
The molecular identification of Anopheles spp. proved to be a useful tool for supporting morphological identification. This molecular approach using two genetic markers increased taxonomic resolution helped to identify damaged specimens and to distinguish species within a complex. A deeper study on the molecular identification of the Anopheline mosquito complex is required [3], as many of the Anopheles species in the MS area are malaria vectors [96], and indigenous cases of malaria have been recorded in East Attica the years 2009, 2010, 2011, 2012, and 2015 [63]. An. maculipennis s.l. is a potential vector of malaria, and it has been considered as an important vector in the past for the spread of this disease in various regions of Greece. An. sacharovi is considered to be the principal vector of malaria, from all subspecies of the An. maculipennis s.l. complex for Mediterranean countries and of course for Greece [67,75,83]. An. claviger is a potential vector of malaria, although its medical significance is not considered to be great for our country. Relatively in high numbers, An. algeriensis is a common and very abundant species in the area, captured in all four traps, and although it can easily be infected with malaria plasmodium, it is considered a secondary vector due to its exophily [70].
The findings of this study revealed different assemblages of mosquito species in each targeted RU. Regarding the selected RUs, there were significant differences in many of their ecological characteristics and that was the main factor for their selection. All of the species recorded in this study were collected in the MS RU, while in the PF RU, the main collected species were Cx. pipiens s.l., Ae. albopictus, and Cs. longiareolata, which is in accordance with the study of Bisia et al. (2020) [18]. In both surveyed RUs, Cx. pipiens s.l. and Ae. albopictus were by far the most abundant species.
This study also provides baseline information and acts as a starting point for further investigation of USUV circulation. With the continuing spread of USUV since 2001 in neighboring countries of Greece [42,47,58], it is important to monitor both viruses before the possible occurrence of an epidemic. According to the epidemiological model, Greece belongs to the areas where the USUV can be transmitted, causing a possible epidemic. It mainly indicates areas in the north of the country, for two possible reasons. First of all, in these areas, especially in the river deltas where the number of mosquitoes has increased, migratory birds appear to have stopped moving from Europe to Africa [101], and secondly, in Northern Greece, the only case so far with antibodies to the virus has been recorded [62].
In the present study, samples from Attica for the years 2017-2018 were found negative for USUV. However, despite the limitations that emerged for confirming one possibly positive sample, USUV and WNV co-circulation cannot be excluded in the future. Up to date, no cases of USUV have been reported in humans or equids in Greece, suggesting that there is no circulation of this virus or, at least, its prevalence is very low. Given the knowledge we have from other relatives of flaviviruses, such as the WNV, the risk of causing even more outbreaks in areas endemic to the USUV or in new ones that have not yet spread are quite high. It must be noted that USUV might be misdiagnosed as WNV when the diagnosis is based only on antibody detection, without testing by PCR or neutralization assays, due to cross-reactivity in serology [42,60,102].
Further investigations of both viruses could provide answers to our suspicions about both the USUV circulation in our country and the interaction with its related flavivirus. Furthermore, entomological surveillance activities should be extended to Attica, especially urban ones, as USUV appears to be equally transmitted in urban and rural areas, in contrast to WNV, where higher transmission rates are recorded in rural areas [26]. Virus surveillance within the native mosquito populations offers an opportunity to detect a virus before the emergence of disease in the susceptible host population [103].
A general comment to be made refers to the fact that Anopheles species identification exclusively by morphological features often presents difficulties, highlighting the necessity for implementation of a targeted molecular protocol to the species level [104]. In our case, Anopheles mosquitoes collected in the Marathonas-Schinias area created a need for the development of a special molecular protocol. In this manuscript, we aimed to highlight that the implemented combined research methodology proved to be a useful tool for supporting morphological identification. Furthermore, the applied molecular methodology was found to be specific and sensitive, regarding the possibility of finding positive mosquito pools for WNV.
In conclusion, the findings of this study emphasize the need for regular monitoring of the mosquito fauna in all regions of Greece, which will contribute to increasing the current knowledge about the diversity, distribution abundance, and ecology of species that are present in the regions. Related studies on mosquito fauna should be performed in all RUs of the Attica region as data on mosquito population and species distribution would be valuable, in particular on those species that are of zoonotic relevance. The implementation of integrated arbovirus surveillance programs represents a relevant and necessary assessment of the risk of pathogen transmission in a given region, allowing for the establishment of the appropriate preventive measures.