Background

Biting midges belong to the family Ceratopogonidae with the most common species being Culicoides [1]. Over 1400 species have been described in the genus Culicoides and, with the exception of Antarctica and New Zealand, they are prevalent worldwide [2, 3]. The distribution, abundance and seasonal occurrence of these holometabolous flies is determined by the availability of moisture-rich habitats that are essential for the development of immature stages [2]. They are associated with aquatic or semiaquatic habitats, e.g. mud or moist soil around streams, ponds and marshes [1]. They are also associated with the dung of various animals such as cattle, buffalo and equines, and breed in rotting fruit, cacti, banana stems and leaf detritus [1]. The presence of animals, especially livestock such as cattle, horses and sheep also play an important role in the abundance/distribution of Culicoides. Biting midges inflict painful bites and suck the blood of their hosts, which include humans, livestock and wild animals [3].

In areas where Culicoides are abundant they may constitute a biting nuisance to humans, domestic and wild animals [2, 4,5,6,7]. Their bites are sometimes associated with allergic skin reactions which in some individuals may result in urticaria. Culicoides are diverse in terms of both affected species and the pathogens they transmit. They are involved in the transmission of animal and human viruses as well as animal and human filariae [8].

The first definite association between flies of the family Ceratopogonidae and the transmission of filarial infections in Cameroon was made by Sharp in 1927 [9, 10] who reported on the development of Acanthocheilonema perstans (Manson, 1891) in Culicoides grahamii (Austen, 1909) and Culicoides austeni (Carter, Ingram & Macfie, 1920) in rainforest areas. In 1949, Henrard & Peel [11] and Chardome & Peel [12] suggested that Sharp was actually working with Dipetalonema streptocerca (Macfie & Corson, 1922) instead of A. perstans, now Mansonella perstans. Hopkins & Nicholas [13] also suggested that C. grahamii, C. austeni and C. inornatipennis (Carter, Ingram & Macfie, 1920) were potential vectors of M. perstans in South-West Cameroon and it remains unclear how abundant these species are in forest areas. Previously suitable ecological conditions for the Culicoides may have been modified due to human activity (land occupation for housing, farming, timber exploitation). Moreover, since these initial observations nearly half a century ago, few studies have updated either vector biology or the significance of the flies in disease transmission. Mansonella perstans has been reported to be endemic in Cameroon, with its distribution highly influenced by bioecology [14]. In the South-West region, where the prevalence of the parasite has not changed even after a decade of ivermectin mass drug administration (MDA), there is a need to investigate the vectors implicated in the transmission of the causative agent of the disease. Vector identification is an important step in the epidemiology of vector-borne diseases and information on the major vector species may provide a clearer indication of the disease distribution (or its potential distribution) and identify high-risk areas for human-vector contact. Of the 13 species of Culicoides previously identified in Cameroon, at least five feed on humans in South-West Cameroon [8, 15, 16]. Given the limited information on Culicoides species in South-West Cameroon and their possible role in the transmission of filarial parasites, this study was designed to determine species diversity of Culicoides, identify anthropophagous species, and elucidate their role in the transmission of M. perstans, an endemic and widespread filarial species in Africa.

Methods

Study design and study site

A cross-sectional survey was carried out in seven selected communities of five health districts in the South-West region of Cameroon (Fig. 1) during the rainy season (May to September). These communities were Bikoki, Bokwai, Ebam, Ediki, Mbule, Nlog and Ogurang (Table 1). The study sites were characterized by the absence of industrial livestock, but individual families reared chickens, pigs and cattle (max. three animals/family). Temperatures during this season are around 20 °C with a relative humidity of 75–80% and daily precipitation of 6–14 mm/day [17]. Due to the volcanic soil of this region, the inhabitants farm plantains around their dwelling/communities. After the harvest, decaying plantains become larval development sites of Culicoides species.

Fig. 1
figure 1

Map of the South-West region of Cameroon showing the study sites with the Culicoides collection points

Full size image
Table 1 Grid-references of the study sites
Full size table

Collection of adult Culicoides species using CDC miniature UV-light traps

Due to the difficult access to some study sites during the rainy season, there were variations in the number of overnight collections per site. Collections of midges were carried out using CDC miniature black UV-light traps (Model 512, John W. Hock Company, Gainesville, USA). The number of collections ranged from 2 nights in Nlog to 6 nights in Ediki. To determine species diversity, we collected adults and potential larval developmental material from which larvae were isolated and reared to adults. Four UV-light traps were set at strategic positions around human dwellings at each of the seven sites. Hourly collections were made from 6 pm to 6 am each working day. For all study sites, the content of each trap was emptied hourly into labelled plastic bottles which were stored in a cooler box and transported to the laboratory for Culicoides identification.

Collection of adult Culicoides species using the human landing catch (HLC) technique

To determine which Culicoides species targeted humans, adult flies were collected using the HLC technique from members of four communities, Bokwai, Ediki, Mbule and Nlog (Fig. 1). HLC was carried out in the morning (6–9 am) and evening (4–7 pm) [15, 16]. This was done by four well trained collectors dressed in protective clothing against midges. The collectors were positioned in four different houses in each study site for 4 days of collection in each study month. Each collector was supplied with a torch to enable him to continue fly collection in darkness (6–7 pm). At collection sites the same collection effort was implemented: 3 h/morning and 3 h/evening for 4 days, making a total of 12 h collection for the morning periods and 12 h for the evening periods. Culicoides females seeking a blood meal were aspirated as soon as they landed on collectors. The aspirated midges were transferred immediately into well-labeled hourly-netted plastic cups and transported to the laboratory for morphological identification and preservation.

Collections of larval developmental material

Potential breeding material, banana/plantain stems in the 1st and 2nd stage of decay, were collected from four communities (Bokwai, Ediki, Mbule and Nlog). Breeding material was graded as described by Hopkins [16]. In the 1st stage of decay, the stems are either upright or prostrate and have concentric rings of fibrous tissue separated by the softer decaying intermediary cells. The 2nd stage of decay is characterized by a remaining basal disk which is hard outside but contains a mass of rotting cells inside which has fallen down. The material was transported to the laboratory where the larvae were isolated and reared to adult stages [15]. In the laboratory, five stems of 30 cm in length and 20 cm in diameter were selected per site. Culicoides larvae were isolated from the compost material as previously described by Hopkins and colleagues [13, 15]. Briefly, Culicoides larvae were isolated at room temperature by washing the compost material using sieves of different pore sizes followed by magnesium sulphate flotation. The larvae were reared in the laboratory using a 12 h day/night lighting regime at 22 °C and 80% relative humidity on compost which had been heat sterilized at 70–90 °C for 20 min. Prior to collection of adult midges (emergence), the compost was placed in cages designed for that purpose: square shaped 900 cm2 boxes covered with very fine netting of pore size 0.5 mm.

Collections of engorged Culicoides from a M. perstans microfilaremic volunteer using a drop trap

To elucidate the role of major anthropophagic species in the transmission of M. perstans, overnight HLC using a drop trap was conducted at Ediki. Collections were done from 6 pm to 6 am for four nights. An M. perstans microfilaraemic volunteer sat under a rectangular netting cage trap (2 × 2 × 2 m; Fig. 2). The M. perstans status of the volunteer had been ascertained using the thick blood film technique [18]. The cage was raised for 10–15 min to allow contact between the host and midges and then lowered to trap the attracted midges. After about 12 min, the expected time for trapped midges to be fully blood-fed, engorged flies were gently aspirated and blown into well-labeled 50 ml falcon tubes 3/4 filled with plaster of Paris (POP). The POP formed an absorbent layer at the bottom of the tubes to retain moisture.

Fig. 2
figure 2

Rectangular drop trap set in the night in the human dwelling

Full size image

The midges were maintained in the tubes in the field using the protocol of Hopkins [15]. The collected midges were fed with 15% sugar solution and transported to the laboratory in a cooler box for maintenance. The Culicoides were maintained at 23 ± 2 °C and relative humidity (RH) of 75 ± 5%. A hole (0.5 cm in diameter) was made in the lid of each rearing tube to permit ventilation and feeding of the flies in captivity. This hole was covered with fine netting (pore size 0.5 mm) on which the sugar solution-moistened cotton was placed for fly-feeding. Flies were fed daily with 15% sugar solution soaked in cotton gauze. Additionally, 3–4 drops of distilled water were added daily using a 10 ml syringe to keep the tubes moist. Mortality was recorded daily, and the insects were identified and dissected 12 days post-feeding to retrieve infective larvae of M. perstans. During dissection, the head, thorax and abdomen were separated and gently teased open in individual dissection wells containing incomplete culture medium (RPMI-1640 medium; Sigma-Aldrich, Munich, Germany) supplemented with a 2% antibiotic cocktail (penicillin-streptomycin-neomycin; Thermo Fisher Scientific, Schwerte, Germany). The species that yielded infective larvae (L3) as well as any other developmental stages (L1 or L2) were noted and the numbers of larvae recorded.

Morphological identification of adult Culicoides species

Morphological identification of Culicoides species was done by examination of the wing pigmentation pattern under a dissecting microscope. In cases difficulties were identifying species based on wing pattern proved inconclusive, other morphological characteristics such as maxillary palps inter-ocular space and male genitalia were used. A combination of identification keys were used for species identification [8, 19, 20].

Data analysis

Data from fly collections, site of collections, identification and dissection were entered in a template designed in Epi info v.3.5.3 (CDC, Atlanta, USA) and exported to SPSS v.20 (IBM, Armonk, NY, USA) for statistical analysis. Culicoides diversity and abundance was expressed as the number of different species of Culicoides collected per site per force of work made at that site. The number of flies captured by a trap in a day (fly/trap/day) was expressed as the number of flies collected divided by the number of traps used divided by the number of collection days. The number of flies per person per day (fly/person/day) was expressed as the number of flies attempting to bite a collector divided by the product of the number of collection days and the number of collectors. The Chi-square test was used to compare the abundance of the different species across the communities as well as the proportions of adult species collected by the different trapping methods. P-values less than 0.05 were considered statistically significant.

Results

Species of Culicoides collected with CDC miniature UV-light traps

Eight Culicoides species were collected over a 4 month period at different time points at 7 sites: C. bedfordi (Ingram & Macfie, 1923), C. inornatipennis, C. fulvithorax (Austen, 1909), C. grahamii, C. imicola (Kieffer, 1913), C. milnei (Austen, 1909), C. neavei (Austen, 1912) and C. kumbaensis (Callot, Kremer, Mouchet & Bach, 1965) (Table 2). Generally, the most abundant species as determined with UV-light trapping was C. grahamii (n = 2865), with an abundance of 31.1 flies/trap/night. The species with the lowest fly/trap/night ratio was C. imicola with 0.01 fly/trap/night. Culicoides grahamii represented 41.6% of the 6889 specimens collected by this method and was closely followed by C. milnei (30.7%). The Chi-square test demonstrated a significant difference in the total number of flies collected at the different sites (χ2 = 7504.2, df = 42, P < 0.0001).

Table 2 Culicoides species collected using UV-light traps during May-September 2016 at 7 study sites in the South-West region of Cameroon
Full size table

Species of Culicoides from larval developmental medium

Fisrt and second stage decaying banana/plantain stems (five stems per site) constituted the larval developmental medium. Upon isolation of larvae and rearing, three species, C. milnei, C. grahamii and C. inornatipennis, were common to all the sampled study sites (Table 3). The highest percentage emergence was achieved from larval developmental medium collected from Mbule and the lowest from Bokwai. In general, C. inornatipennis was the first (6 days) to emerge and was the most abundant (46.5%; n = 1307), while C. milnei was the last (10 days) to emerge and the least abundant (11.5%; n = 303). Emerging Culicoides species were monitored over a 14-day period.

Table 3 Percentage emergence of Culicoides species larvae isolated from plantain stems in four study sites of the South-West region of Cameroon
Full size table

Anthropophilic Culicoides species as determined with the HLC technique

Of the eight species collected in this survey, only four (C. inornatipennis, C. fulvithorax, C. grahamii and C. milnei) were collected from humans (Table 4). Culicoides grahamii was the most anthropophylic species with a morning fly/person/hour ratio of 48.5 and an evening fly/person/hour ratio of 90.5. This was seconded by C. milnei. Culicoides inornatipennis and C. fulvithorax had the lowest fly person ratios in the evening.

Table 4 Culicoides species collected during morning and evening periods in 4 sites using the HLC technique
Full size table

In total, 6919 Culicoides were collected by the HLC technique with a significantly higher number collected in the evening period (4–7 pm) compared to the morning period (6–9 am) (χ2 = 69.9, df = 4, P < 0.0001) at all of the study sites except in Mbule. The number of Culicoides per person per hour was higher in the evening (92.7) when compared to the morning (51.5) periods (Table 4).

Susceptibility of C. milnei to M. perstans

Engorged Culicoides (n = 2100) were collected with a drop trap and maintained in the laboratory for 12 days. Culicoides milnei was the most abundant species collected, followed by C. grahamii, C. inornatipennis, C. fulvithorax and C. neavei (Table 5). During the laboratory maintenance period, survival of the midges decreased gradually from the 3rd day post-infection (Fig. 3, Additional file 1: Table S1). After 12 days of maintenance, 53.4% (n = 1091) of engorged flies had survived and were dissected to monitor for the presence of M. perstans infective larvae. Upon dissection, C. milnei was the main species that allowed the development of ingested microfilariae of M. perstans to third-stage infective larvae. Out of 807 C. milnei dissected, 333 were infected and 584 infective larvae were recovered giving an output of 1.75 L3s/infected fly. Only 7 infective larvae were obtained from 5 C. grahamii females out of 194 collected from the same donor. In contrast, the C. inornatipennis, C. fulvithorax and C. neavei engorged on the same donor were all negative for infective larvae of M. perstans at dissection (Table 5).

Table 5 Developmental stages of Mansonella perstans recovered from Culicoides fed on a microfilaraemic donor and reared in the laboratory for 12 days following infection
Full size table
Fig. 3
figure 3

Engorged Culicoides survival curve following rearing in the laboratory (note a steady reduction of survival rate over time)

Full size image

Hourly Culicoides collection using UV-light traps

Overall, C. grahamii was the most abundant species (average of 200 midges/h) compared to the other 7 captured species (Fig. 4, Additional file 1: Table S2). A peak in the total numbers of Culicoides collected was observed between 10 and 11 pm. This peak followed a lower peak observed between 2 and 3 am. Culicoides fulvithorax seemed to become active during the second half of the night, peaking between 1 and 3 am. Culicoides milnei was characterized by two biting peaks, one in the late evening hours, 10–11 pm, and in the early morning hours between 2 and 4 am (Fig. 4, Additional file 1: Table S2).

Fig. 4
figure 4

Average number of the most abundant Culicoides visiting cycles in all collection sites after 18 nights of collection using four traps

Full size image

Comparing the light trap visiting cycles of the different species of Culicoides at the different collection sites (Fig. 5, Additional file 1: Tables S3–S7), C. grahamii displayed 2 major biting peaks: one in the early hours of the morning (5–6 am) and the other in the early hours of the evening (7–8 pm) in Mbule (Additional file 1: Table S5). In Ediki, a steady increase in biting intensity was observed from the late hours of the evening towards the early hours of the morning (Additional file 1: Table S6). With the exception of Ebam, C milnei was the second most abundant species collected at all sites (Fig. 5, Additional file 1: Tables S3–S7). Two very conspicuous biting peaks were observed in Ebam at 8–10 pm and 4–6 am and relatively very few other Culicoides species were collected (Fig. 5, Additional file 1: Table S4). Using the Chi-square test, the hourly trap visiting cycle differed significantly for the different species collected (χ2 = 3342.3, df = 84, P < 0.001).

Fig. 5
figure 5

Average number of Culicoides species using four light traps during three nights at Bikoki, Ebam, Mbule, Ogurang and six nights at Ediki

Full size image

Discussion

In a defined area, the abundance and diversity of Culicoides species strongly depends on the availability and type of breeding sites [8]. Overnight UV-light trap collections indicated that eight Culicoides species were common in the surveyed rain forest villages in the South-West region of Cameroon. All eight species had previously been reported in Cameroon [8, 20]. The high number of decaying plantain stems around homes in the seven surveyed communities is no doubt a contributing factor for the abundance of the Culicoides species that breed in rotting vegetation. Culicoides grahamii was the most abundant species caught overnight and was closely followed by C. milnei.

Some authors have suggested that C. milnei is identical to C. austeni [15, 21]. Ecologically, C. austeni is a species native to mangrove and brackish water areas while C. milnei is a forest species, as observed in our studies performed here. In 1961, Murphy [22] confirmed that C. milnei is a separate species from C. austeni. Our observations of the palps, inter-ocular space, the wing patterns and the male genitalia of this species correspond to the description given by Murphy for C. milnei [22].

It is well documented that most Culicoides species have a crepuscular periodicity, peaking at dusk and dawn [7, 23]. Moreover, crepuscular periodicity can be influenced by nocturnal temperatures and it is suggested that seasonal shifts in the periodicity can occur [24,25,26]. In this study, C. grahamii was dominant in morning, evening and overnight collections whilst C. milnei was mainly collected during overnight collections (Additional file 1: Table S8). These findings confirm those of Hopkins [15] that indicated C. milnei is essentially a nocturnal species. However, our results differ with regard to C. grahamii since they found it to be only a diurnal species but here, this species surpassed all others in all the collection techniques used both during the day and night, except with the drop trap where C. milnei-dominated. Our studies concur with the observations of Vattier-Bernard and colleagues [16] that C. grahamii has a high biting rate in the early morning and evening hours although in that study there were no overnight collections.

To determine the breeding sites of the different species we investigated first and second stage decaying plantain and banana stems. Three Culicoides species, C. milnei, C. grahamii and C. inornatipennis, were common to almost all the sites with C. inornatipennis being the most abundant species, while the least was C. milnei. All these species were collected with the UV-light traps. Hopkins [15] also reported C. milnei and C. grahamii in decaying banana and plantain stems and also observed C. inornatipennis in third stage vegetation. In this study, the majority of breeding material was collected in the first and second stages of decay. Although we did not separate the material according to the stage of decay, the second stage was relatively more productive compared to the first. We had more of C. grahamii than C. milnei from this pool of material. Previous studies have shown that C. milnei prefers to breed in first rather than second stage decaying material and moreover, it is rare to find this species in third stage material [13]. Likewise, C. grahamii also prefers to breed in first stage material [15]. Culicoides inornatipennis on the other hand breeds in rot holes in addition to third stage decaying banana and plantain stems which is in contrast to most non-anthropophilic species which prefer moist soil microhabitats of various kinds, including rubbish heaps and dung [15, 27, 28]. An advantage at our study sites lies in the year-round planting and harvesting of bananas and plantains which provides continual breeding sites for Culicoides and, in turn, has an implication on yearly fly abundance. Only four anthropophilic species, C. inornatipennis, C. fulvithorax, C. grahamii and C. milnei, were collected in the early morning and evening hours. Species diversity was greater using the miniature UV-light trap compared to HLC (Additional file 1: Table S8). When determining the importance of a particular vector species and related dynamics of disease transmission, HLC appears to be a more suitable approach. Most vector species are polyphagous and some midges attracted to light may not normally bite humans and therefore will have no role in disease transmission. Agbolade et al. [29] used HLC methods to distinguish C. fulvithorax as the vector of M. perstans in the northern area of Western Nigeria. White [30] also captured C. fulvithorax, C. grahamii and C. milnei as anthropophilic species. The high diversity of Culicoides species captured with UV-light traps indicates that these serve as a great attractant to midges. It will, however, be difficult to determine whether these species are disease vectors without further investigation [31].

Vector competence refers to the ability of arthropods to acquire, maintain, and transmit pathogenic agents [32]. For example, besides the transmission of filarial parasites, studies in Tunisia, South Africa, Senegal and Nigeria reported that Culicoides spp. are able to transmit viruses [33,34,35,36]. However, to our knowledge, until now there have been no reports of transmission of viruses from Culicoides spp. in Cameroon. In regard to filariae, upon dissection of the engorged specimens, C. milnei was the most competent at permitting the development of the infective larval stages from the ingested microfilariae. From 807 fed C. milnei, 333 (41.3%) flies were found to harbor M. perstans infective larvae, compared to 5 (2.6%) out of 194 fed C. grahamii. In 1952, Hopkins [15] did not find any developing larvae of M. perstans in 1500 engorged C. grahamii. Hopkins & Nicholas [13] also engorged 418 C. grahamii on an M. perstans donor but found only two flies with infective stage larvae upon dissection. The current data supports those earlier findings since we found only a few infective larvae in C. grahamii that had fed on the M. perstans donors. Therefore, although the larvae are yet to be confirmed molecularly, C. grahamii seems to be an inefficient vector of M. perstans infections. Our study therefore confirms the works of Sharp [9, 10] and Hopkins & Nicholas [13] who first made an association between A. perstans and the possible vectors. The results of the present study confirm that C. grahamii is a poorly competent vector of M. perstans while C. milnei is a highly competent vector of M. perstans [37]. This confirmation on the potential role of C. milnei in the development of M. perstans, six decades after the first observation by Hopkins & Nicholas [13], has several implications, with the first being the possibility of carrying out surveys on the transmission of M. perstans targeting C. milnei as the major vector. However, before such programmes can begin, it is important to confirm the observations made here in South-West Cameroon with other bioecological zones of Cameroon. One impediment for the study of transmission of M. perstans remains the tedious microscopic dissection of Culicoides. Future studies including PCR identification/confirmation of the M. perstans vector would constitute a major step for the transmission survey of mansonellosis. Another important finding of this study is that the major potential vector of M. perstans, C. milnei, is a nocturnal species. It is known from other filariae that parasitize humans that the abundance of blood-dwelling microfilariae correlates with whether the vector bites during the day (Loa loa, [38]) or night (Wuchereria bancrofti, [39]). Mansonella perstans is considered a diurnal species meaning that microfilariae are potentially circulating during day and night [40, 41]. The biting behavior of C. milnei, as observed in this study, contradicts the possible diurnal periodicity of M. perstans microfilariae since they appeared more abundantly in the peripheral blood during the day. Indeed, in a recent survey for lymphatic filariasis carried out in 31 health districts in the forested area of Cameroon, where we collected blood from the same individuals during the day and night, we observed that the microfilarial load of M. perstans was systematically higher during night than during day collection (unpublished data). This apparently contradictory information calls for further studies to clarify the periodicity of microfilariae of M. perstans in the peripheral blood in humans and correlate it with the biting behaviors of the Culicoides vectors. Such clarification would be necessary to know which is the most convenient period (day or night) for new epidemiological surveys of mansonellosis.

Conclusions

This study has confirmed observations made 60 years ago on the biodiversity, biting behavior and breeding habitats of Culicoides species in South-West Cameroon. It has clarified the important role which C. milnei play in the transmission of M. perstans, giving the opportunity to target this species for studies on the transmission of mansonellosis in Cameroon and elsewhere.