Prevalence of Mosquito Populations in the Caribbean Region of Colombia with Important Public Health Implications

Mosquito studies are important for understanding their role in the transmission of pathogens including arboviruses, parasites, and protozoa. This study characterized the prevalence of Culicidae fauna in rural and peri-urban areas with human populations in the Colombian Caribbean region to establish the risk of transmission of mosquito-borne pathogens. From 2016 to 2017, adult mosquitos were collected in Turbaco (Bolívar), Sabanalarga (Atlántico) and Pueblo Bello (Cesar). The collections in rural areas were in the forest fragments using CDC, Shannon, and human bait traps. In peri-urban areas, Prokopack aspirator collections were used inside households. Entomological and ecological indicators were also calculated. A total of 11,566 mosquito specimens, from 13 genera and 63 species, were collected. The forests fragments of Sabanalarga and Turbaco had the highest species abundance and richness. Turbaco had the highest adult Aedes aegypti index. Arbovirus vectors were among the identified species, including Ae. aegypti, Culex quinquefasciatus, Haemagogus janthinomys, Sabethes chloropterus, Aedes angustivittatus, Mansonia titillans, Coquillettidia venezuelensis and the subgenera Culex Melanoconion. Overall, the diversity and abundance of mosquitoes present in these municipalities establish a potential disease transmission risk by these vectors.


Introduction
Mosquitoes (Diptera: Culicidae) are one of the most studied groups of arthropods worldwide due to their high transmission capacity for pathogens, including viruses, such as dengue, Zika, and yellow fever, and protozoa, such as Plasmodium spp., the causative agents for malaria [1]. Female mosquitoes are hematophagous insects and acquire these pathogens when they feed on infected blood, in turn transmitting them to other healthy hosts [1,2]. According to the World Health Organization (WHO), vector-borne diseases represent more than 17% of all infectious diseases and cause more than 700,000 deaths each year [1,3]. In the Americas, it is estimated that 50% of the population is predisposed to these pathologies [3]; specifically in Colombia, approximately 24% of the inhabitants are at risk of acquiring a disease transmitted by mosquitoes in endemic areas [4]. With the rapid expansion of mosquito distributions due to global warming and increased modes of transportation, epidemiological research of these vectors plays a key role in understanding the populations at risk for contracting diseases transmitted by mosquitos [5][6][7].
The tropical climate of Colombia presents ideal conditions for the migration, adaptation, and persistence of different species of culicids, with approximately 324 species from 28 genera classified throughout the territory [8]. In particular, the Caribbean region contains ecosystems that vary from tropical dry forests to humid mountainous forests with annual rainfall above 3000 mm [9]. This environmental heterogeneity supports the reproduction and colonization of a wide range of mosquitoes in urban, peri-urban, and rural areas. Previous studies in the region have reported the presence of vectors of the genera Aedes, Culex, Haemagogus, Psorophora, and Mansonia infected with arboviruses in areas deforested for expansion of rice fields and expansion of cattle ranching [10]. In addition, some mosquitoes from forest environments such as species of the genus Haemagogus (yellow fever vectors) and Anopheles (malaria vectors) have been reported in peri-urban environments in the Caribbean [11]. These studies represent only a fraction of the possible ecological niches in this region of Colombia and have not characterized a wide range of mosquitoes present, especially from the genus Culex that are known arbovirus vectors that pose a risk to humans. Hence, the current study aims to characterize the prevalence and biodiversity of Culex and other mosquito species over a wider geographic range than previously reported in the Caribbean Region of Colombia, to expand the understanding and geographic range of potential arbovirus vectors in support of arbovirus prevention and control programs in the Region.

Description of the Study Areas
The specimen collection areas, corresponding to three municipalities of the  (Figure 1). The municipalities of Turbaco and Sabanalarga are municipalities with a dry tropical climate and low or coastal ecotope characterized by large areas of tropical dry forests with altitudes between 90 and 200 m, average temperatures of 27 • C, relative humidity between 80 and 87%, and annual rainfall between 1189 and 1197 mm. In terms of territorial extension, Turbaco has an urban area of 8 km 2 inhabited by 95,000 people (~11,875 hab/km 2 ), while Sabanalarga has an urban area of about 7.5 km 2 and a population of 67,000 inhabitants (~8933 hab/km 2 ). In contrast, Pueblo Bello is a municipality in a high altitude (1200 m) ecotope that is predominantly tropical rainforest, with a temperate climate characterized by an average annual temperature of 21.8 • C, relative humidity of 91.6%, and annual rainfall of 26,572 mm. Pueblo Bello is the municipality with the smallest urban area (about 3.5 km 2 ) and population (8236 inhabitants;~2353 inhabitants/km 2 ) among the sampling sites. One of the main characteristics shared by these three municipalities is that they have extensive rural areas with agricultural and livestock exploitation, which are their main economic activities.

Collection and Identification of Specimens
Mosquitoes were collected in four sampling events that occurred over five consecutive days in September 2016, December 2016, May 2017, and September 2017. The samplings covered both rural and peri-urban areas in each municipality.
In this study, rural areas were represented as those areas of forest adjacent to crop fields distant a minimum of 1.5 km from the urban center of each municipality. The collections in these areas occurred between 18:00 and 20:00 using three CDC light traps baited with CO 2 . Each trap was at a maximum distance of 10 m from another. Additionally, human bait and Shannon trap collections with mouth aspirators were also used. Three researchers with personal protection equipment to avoid mosquito bites and minimize risks carried out the human bait collections.
The collections in peri-urban areas were conducted within homes located in peripheral areas of each municipality during daylight hours. These areas were chosen due to the proximity of the houses to wooded areas that facilitate contact with mosquitoes. Sampling of the houses was by convenience and dependent on the acceptance and participation of the inhabitants. After verbal authorization from the homeowners, two investigators collected adult mosquitoes using Prokopack aspirators for 10 min [12]. All the places in the home that the residents allowed access to were inspected (living room, bedrooms, kitchen, bathroom, etc.), when access to the bedrooms was allowed, the aspirator was passed over walls, under furniture, and inside closets and other likely adult mosquito resting sites. Aspiration collections were similarly attempted outside the house from outside walls, under eaves, and outdoor stored materials [13].

Collection and Identification of Specimens
Mosquitoes were collected in four sampling events that occurred over five cons tive days in September 2016, December 2016, May 2017, and September 2017. The s plings covered both rural and peri-urban areas in each municipality.
In this study, rural areas were represented as those areas of forest adjacent to fields distant a minimum of 1.5 km from the urban center of each municipality. The lections in these areas occurred between 18:00 and 20:00 using three CDC light traps ba with CO2. Each trap was at a maximum distance of 10 m from another. Additionally man bait and Shannon trap collections with mouth aspirators were also used. Thre searchers with personal protection equipment to avoid mosquito bites and minimize r carried out the human bait collections.
The collections in peri-urban areas were conducted within homes located in per eral areas of each municipality during daylight hours. These areas were chosen due to proximity of the houses to wooded areas that facilitate contact with mosquitoes. Samp of the houses was by convenience and dependent on the acceptance and participatio the inhabitants. After verbal authorization from the homeowners, two investigators lected adult mosquitoes using Prokopack aspirators for 10 min [12]. All the places in home that the residents allowed access to were inspected (living room, bedrooms, kitc bathroom, etc.), when access to the bedrooms was allowed, the aspirator was passed walls, under furniture, and inside closets and other likely adult mosquito resting s Aspiration collections were similarly attempted outside the house from outside walls der eaves, and outdoor stored materials [13].
All specimens were identified with the available taxonomic keys [14][15][16][17][18][19][20][21] and w preserved in liquid nitrogen until stored at −80 °C at the Tropical Medicine laborator the UNIMOL research group of the University of Cartagena. No research permits w required for this study. All specimens were identified with the available taxonomic keys [14][15][16][17][18][19][20][21] and were preserved in liquid nitrogen until stored at −80 • C at the Tropical Medicine laboratory of the UNIMOL research group of the University of Cartagena. No research permits were required for this study.

Data Analysis
The data were analyzed in two sections. First, the relative abundance of species and ecological indices were calculated from the number of each mosquito species captured in rural areas using PAST v3.26b software. For each sampled fragment, the diversity was estimated using the Shannon-Wiener index, and the species richness was calculated using the Margalef index [22]. The species dominance was assessed using the Simpson index and the equity was estimated using the Pielou index [23,24]. The similarity between the municipalities was calculated using the Jaccard index, which makes comparisons from presence and absence data (qualitative data), and the Bray-Curtis index, which uses abundances (quantitative data) [23]. Second, the frequencies of the different species were calculated from the peri-urban collection sites. The infestation index of Ae. aegypti adults (percentage of houses positive for this species) and the density (total of individuals of Ae. aegypti/total of houses inspected) were estimated in each locality studied. The chi-square and Kruskal-Wallis tests were applied to compare these variables between municipalities with a statistical significance of 0.05 (p < 0.05).

Results
In total, 11,566 culicids were captured from Turbaco (n = 3935), Sabanalarga (n = 4601), and Pueblo Bello (n = 3030). A total of two subfamilies, six tribes, 13 genera, and 63 species were identified (Supplementary Table S1, Figure 2). It is important to note that, despite the poorly defined taxonomic status, it was possible to identify at least ten species of the genus Culex. Even so, many specimens of the genus Culex were only identified up to subgenera (Culex and Melanoconion); therefore, in this study, they are reported as Culex (Culex) spp. and Culex (Melanoconion) spp., respectively. Likewise, all specimens that suffered damage or loss of important structures for identification during collection were classified as genus or subgenus.
presence and absence data (qualitative data), and the Bray-Curtis index, which uses abundances (quantitative data) [23]. Second, the frequencies of the different species were calculated from the peri-urban collection sites. The infestation index of Ae. aegypti adults (percentage of houses positive for this species) and the density (total of individuals of Ae. aegypti/total of houses inspected) were estimated in each locality studied. The chi-square and Kruskal-Wallis tests were applied to compare these variables between municipalities with a statistical significance of 0.05 (p < 0.05).

Results
In total, 11,566 culicids were captured from Turbaco (n = 3935), Sabanalarga (n = 4601), and Pueblo Bello (n = 3030). A total of two subfamilies, six tribes, 13 genera, and 63 species were identified (Supplementary Table S1, Figure 2). It is important to note that, despite the poorly defined taxonomic status, it was possible to identify at least ten species of the genus Culex. Even so, many specimens of the genus Culex were only identified up to subgenera (Culex and Melanoconion); therefore, in this study, they are reported as Culex (Culex) spp. and Culex (Melanoconion) spp., respectively. Likewise, all specimens that suffered damage or loss of important structures for identification during collection were classified as genus or subgenus.
The most efficient capture method was human bait, collecting 52.45% (n = 5509) of the total captured among the municipalities, followed by the CDC trap with 45.07% (n = 4734) and the Shannon trap with 2.47% (n = 260). The human bait method found 54 species, followed by the CDC trap with 47, and the Shannon trap with 29. The species abundance and richness for each capture method in each municipality are presented in Supplementary  Table S2.
The dendrograms demonstrate that Sabanalarga and Turbaco have a similar composition and abundance of species, with similarity percentages > 40% as determined by the Jaccard and Bray-Curtis indices ( Figure 3). These same forest fragments also showed the highest values of diversity and species richness evaluated by the Shannon and Margalef indices. Pueblo Bello had the lowest diversity and richness as calculated by the indices ( Table 1)

Indoor Collections (Peri-urban Area)
In the peri-urban areas, a total of 125 homes were visited in the three municipalities where 1063 total mosquitos from 18 species were collected (Supplementary Table S1). Forty percent of the homes inspected (n = 50) among the three municipalities were found to be positive for adult Ae. aegypti. There was no statistical difference in the adult infesta-

Indoor Collections (Peri-urban Area)
In the peri-urban areas, a total of 125 homes were visited in the three municipalities where 1063 total mosquitos from 18 species were collected (Supplementary Table S1). Forty percent of the homes inspected (n = 50) among the three municipalities were found to be positive for adult Ae. aegypti. There was no statistical difference in the adult infestation index between the three municipalities, although Turbaco was slightly higher than Pueblo Bello and Sabanalarga (chi-square, p > 0.05). Likewise, there were no significant differences in the density of adult mosquitoes (Kruskal-Wallis, p = 0.567) ( Table 2). In the municipalities of Turbaco and Sabanalarga, the Ae. aegypti was the most frequent, totaling 48,24 and 59.3% of the samples corresponding to each municipality. In Pueblo Bello, Cx. quinquefasciatus was the most abundant species (59.79%). The indoor samplings also showed the presence of different species of the genera Culex, Aedes, and Mansonia within the houses, as well as the species Psorophora ciliata, Limatus durhamii, and An. neomaculipalpus (see Supplementary Table S1).

Discussion
Understanding the presence of culicids in the different regions of Colombia is of great interest to public health. Therefore, this study bridges this gap by enhancing our knowledge of the abundance and diversity of mosquito species present in rural and peri-urban areas in the municipalities of Turbaco (Bolívar), Sabanalarga (Atlántico), and Pueblo Bello (Cesar).
Studies on the mosquito fauna in the Caribbean Region of Colombia are minimal, and thus far, the investigations carried out report a relatively low diversity of species (4-25 identified species) [10,11,25,26]. By contrast, this study collected 63 species of mosquitoes from 13 genera, in this region. The increase in the diversity of species captured may be due to the characteristics of the collection sites, the type of study, the collection techniques, and the taxonomic identification efforts. Unfortunately, the taxonomy of several mosquito species is poorly understood, making it difficult to describe their exact distribution and abundance. This was particularly problematic for the species of the subgenera Cx (Cx.) spp. and Cx (Mel.) spp. found in this study. Despite extensive taxonomic efforts, it was not possible to identify a high proportion of these mosquitoes. However, the number of Culex species that could be identified here (at least ten species) was considerably higher than those reported by the studies mentioned above (zero to two species identified).
In general, the highest abundance and richness of culicids obtained in this study were collected using the human bait technique, followed by the CDC trap and the Shannon trap. Comparatively, these results are similar to those obtained by Parra et al. [27] in a study conducted with mosquitoes from the Urabá, Antioquia. In the former study 59.7% specimens were captured by human bait compared to 52.45% in the current study. However, compared to Parra et al. [27], we captured 18.5% more specimens in CDC traps and 11.2% less in Shannon traps. Additionally, similar results have been obtained in studies conducted on anopheline mosquitoes in Venezuela and Brazil [28,29].
In this study, the forest fragment of the municipality of Turbaco presented the highest species richness, while Sabanalarga had the highest abundance, diversity, and equity (Table 1). This may be due to the presence of more stable and diverse natural breeding sites than those found in other municipalities. Likewise, when determining the similarity between the structure of the mosquito community, Sabanalarga and Turbaco had considerably highest diversity, richness, and equity and a lower dominance index compared to the municipality of Pueblo Bello. These differences are likely because coastal areas with lower altitudes and higher temperatures regularly have a higher abundance and richness of culicids in relation to the more temperate and high zones [30].
It is important to note that the density of Cx. quinquefasciatus, a mosquito capable of transmitting multiple diseases, including Venezuelan equine encephalitis and West Nile virus [31,32], was found in the rural environments of all three municipalities. This is of particular concern for public health since this species is usually found in urban environments. These findings are consistent with previous reports from forest areas in Antioquia and Córdoba, indicating Cx. quinquefasciatus as a species that effectively adapts to multiple environments [11,27]. Another member of the same genus, Culex nigripalpus, was found in these regions. This mosquito, with an ornithophilic tendency, is generally found in dense and humid vegetation, but has also been found to proliferate in urban areas [33]. In the present study, Cx. nigripalpus showed a preference for forest spaces in each municipality. However, its proximity to the homes in Turbaco and Pueblo Bello increases the epidemiological importance of this species, which is also involved in the transmission of Venezuelan equine encephalitis and St. Louis encephalitis in Florida [34,35].
The subgenus Ochlerotatus of the genus Aedes includes some species involved in the transmission of pathogens in humans. This study found the species Ae. angustivittatus, a mosquito found in both forest and peri-urban environments. Here, it was found in the rural areas of the three municipalities and in the homes of Turbaco. This species is of importance to public health since it has been reported to be naturally infected by the Venezuelan equine encephalitis virus in Colombia [36]. Additionally, the species Ae. scapularis is a culicid found at low elevations and has the ability to reproduce in a wide variety of temporary or semi-temporary fresh waters [18]. Our collections of Ae. scapularis were consistent with descriptions as we found this species in the municipalities of Turbaco and Sabanalarga, located at an altitude between 90 and 200 m. This species is also of high importance as a potential vector of various arboviruses, such as Venezuelan equine encephalitis and yellow fever, as well as a secondary vector of bancroftian filariasis [18]. On the other hand, Ae. taeniorhynchus is a competent vector for Venezuelan equine encephalitis and the parasite Dirofilaria immitis [37,38]. This culicide reproduces in brackish water, so it is considered a species mainly in coastal environments; however, its adaptation to freshwater has led to the colonization of areas far from the sea [39]. This explains the presence of this species in the municipality of Pueblo Bello, which is located approximately 100 km from the coast with an altitude of 1200 m. In addition, its presence in homes in the municipality of Turbaco exemplifies its wide ecological plasticity being found in both urban and rural areas.
Occasional collections of the genus Haemagogus were obtained with a mouth aspirator between 18:00 and 18:30 h. Additionally, three specimens were captured in CDC light traps (supplementary Table S2). Species of the genus Haemagogus are characterized by diurnal habits and are active at noon hours [19,40]. Although activity and biting have also been reported at approximately 18:00 h [41], this allows us to suspect that the specimens captured by the CDC traps were possibly attracted by CO 2 in the first minutes after the activation of the traps. This study was directed toward nocturnal and twilight mosquitoes; therefore, no samples were taken during the hours of the day of the greatest activity of the species of this genus, explaining the low density of individuals collected. However, we do report the twilight activity of Haemagogus mosquitoes when their habitat is invaded. Among the species found, Haemagogus janthinomys and Hg. equinus are considered the main vectors of yellow fever in Colombia [25,40]. Here, Hg. janthinomys was found in the three municipalities sampled, being more frequent in Turbaco and Sabanalarga. In previous studies, the species Hg. janthinomys had been recorded in all the departments that make up the Caribbean region with the exception of the department of Bolívar [10]. However, with these findings, the distribution of this species now extends the entire territory of the region.
Other species with diurnal activity were found, including Sabethes belisarioi and Sa. chloropterus. The species Sa. chloropterus in particular, has been reported as a vector of yellow fever in the Americas [42][43][44]. To complement the studies of diurnal mosquitoes such as those of the genera Haemagogus and Sabethes, it is necessary to perform vertical sampling between 0 and 30 m, given that the preference of the species of these genera is to the highest strata of the forest [45].
The species Ae. aegypti was found abundantly in indoor collections due to its known adaptation and preference for urbanized environments [17]. However, it is plausible to consider that the presence of some individuals of Ae. aegypti in the forest fragments of the municipalities of Turbaco and Pueblo Bello is due to human activity in the crop fields surrounding the sampled area. Aedes aegypti is responsible for the transmission of the urban cycle of arboviruses such as Dengue, Zika, Chikungunya, Yellow fever, and several types of encephalitis [7,[51][52][53][54][55][56]. The entomological index established during the home inspections classified Turbaco as the municipality with the highest presence and abundance of Ae. aegypti within households. These results can be explained by two observations: first, Turbaco was more urbanized than the other municipalities, which expands the settlement space of the mosquito Ae. aegypti in the sampled area. The second reason is attributed to the greater presence of containers with clean stagnant water in the inspected homes of this locality compared to Sabanalarga and Pueblo Bello. Although no immature forms of Ae. aegypti were collected in this study, the presence of larvae was observed in these artificial hatcheries.
The municipality of Pueblo Bello was the most rural and was surrounded by forest fragments that presented continuity with the vegetation of the slope. This characteristic generally led to the fact that the dwellings of the inspected neighborhoods were a few meters away from each other and these in turn were surrounded by shrub areas with semipermanent natural breeding sites such as swamps and puddles. These conditions possibly facilitated the presence of forest species such as An. neomaculipalpus, Li durhamii and Ps. ciliata (Supplementary Table S1). It is important to note that the species An. neomaculipalpus is reported in the literature as a potential vector of malaria [57].
The inspected homes were in areas bordering the urban center of each municipality. In addition, some bordered semi-permanent natural breeding sites such as swamps and puddles, suitable for the proliferation of culicid, represent a scenario of easy interaction between these mosquitoes and the surrounding populations. These conditions possibly favored the presence of mosquito species from both peri-urban and rural environments. On the other hand, the 64 species of mosquitoes reported in this study were found in rural areas, some of which are recognized vectors of arboviruses such as dengue, malaria, yellow fever and different types of encephalitis. The rural areas sampled in the three municipalities were generally near areas deforested for the expansion of agriculture and livestock activities, which favors human contact with mosquitos in the area [27]. Therefore, these sampling sites in both peri-urban and rural areas of the Caribbean Region of Colombia undoubtedly pose a potential risk of transmission of diseases that could affect human populations living in or deployed to these regions.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/tropicalmed8010011/s1, Table S1: Distribution and frequency of mosquito species in rural and urban areas from three municipalities; Table S2: Distribution of species according to collection method in rural areas.

Data Availability Statement:
All data used to support our findings are included in the manuscript. The corresponding author may provide any additional request information.