The occurrence of Culicoides species, the vectors of arboviruses, at selected trap sites in Zimbabwe

Abstract A study of the distribution of Culicoides species was conducted by establishing 12 light trap sites over five rainy seasons between 1998 and 2003 covering all the geo-climatic natural regions of Zimbabwe. In total, 279 919 specimens of Culicoides were trapped over a total of 163 trapping nights. The highest median counts of Culicoides per trapping night were recorded in natural region III, which has climatic conditions conducive to the successful development of the larvae. Culicoides imicola, the major vector of bluetongue and African horse sickness viruses in Africa, was found to be the most abundant species (80.4%), followed by Culicoides enderleini (5.9%) and Culicoides milnei (5.2%). This study identified 10 species of Culicoides that had not been previously described in Zimbabwe, including Culicoides loxodontis and Culicoides miombo, which are members of the C. imicola complex. A total of 23 994 Culicoides midges were collected from five trap sites in Harare, Zimbabwe, with the dominant species, C. imicola, representing 91.6% of the total collection. Seventeen arboviruses were isolated from these midges, 15 of which were bluetongue virus. The predominant bluetongue virus serotype was serotype 11, followed by serotypes 1, 8, 12 and 15. Bluetongue virus serotypes 1, 2, 8, 10, 12, 15, 16 and 18, detected in this study, had not been previously reported in Zimbabwe.


Introduction
African horse sickness virus (AHSV), which infects all equine species, and bluetongue virus (BTV), which infects sheep and other ruminants, are endemic in Zimbabwe. High seroprevalences of AHSV in horses and donkeys and BTV in cattle and sheep were reported throughout Zimbabwe (Gordon 2010), with mortality rates because of African horse sickness and bluetongue reaching 95% (Coetzer & Guthrie 2004) and 33% (Musuka 1999) respectively. Based on their economic importance and capacity to spread rapidly between countries, the World Organisation for Animal Health (OIE) has designated both these diseases as notifiable animal diseases. Equine encephalosis virus (EEV), which infects all equine species, and epizootic haemorrhagic disease virus (EHDV), which infects domestic and wild ruminants, have also been detected in Zimbabwe (Gordon 2010;Musuka 1999;Paweska et al. 1999). Many of these viruses are also enzootic in neighbouring countries, namely Botswana (Mushi et al. 1998), Zambia (Mweene et al. 1996), Namibia and South Africa (Meiswinkel, Venter & Nevill 2004;Scacchia et al. 2009; World Organisation for Animal Health 2011). Oura et al. (2012) recently documented EEV in Ethiopia, Ghana, Gambia and Israel. Bluetongue today has a nearly worldwide distribution.
All of these diseases are caused by double-stranded ribonucleic acid (dsRNA) viruses of the family Reoviridae and genus Orbivirus, which are transmitted almost exclusively by Culicoides (Diptera: Ceratopogonidae) biting midges (Meiswinkel et al. 2004;Mellor, Boorman & Baylis 2000). Culicoides imicola and Culicoides bolitinos have been identified as vectors of AHSV and are widely distributed in sub-Saharan Africa (Meiswinkel & Paweska 2003;Mellor 1993). Culicoides bolitinos has also been implicated as an important potential vector of EEV and BTV in South Africa (Meiswinkel & Paweska 1998Venter et al. 2002).
Previous studies in Zimbabwe found C. imicola to be the most abundant livestock-associated Culicoides species, followed by Culicoides zuluensis (Blackburn, Searle & Phelps 1985;Braverman & Phelps 1981;Braverman et al. 1985;Gordon 2010;Musuka 1999;Musuka, Chihota & Kelly 1998;Phelps, Blackburn & Searle 1982). The vector competence of most Culicoides species in Zimbabwe has not been studied in detail and other species may, in future, be identified as vectors of these viruses.
A study of the distribution of Culicoides species was conducted by establishing 12 light trap sites over five rainy seasons between 1998 and 2003 covering all the geo-climatic natural regions of Zimbabwe. In total, 279 919 specimens of Culicoides were trapped over a total of 163 trapping nights. The highest median counts of Culicoides per trapping night were recorded in natural region III, which has climatic conditions conducive to the successful development of the larvae. Culicoides imicola, the major vector of bluetongue and African horse sickness viruses in Africa, was found to be the most abundant species (80.4%), followed by Culicoides enderleini (5.9%) and Culicoides milnei (5.2%). This study identified 10 species of Culicoides that had not been previously described in Zimbabwe, including Culicoides loxodontis and Culicoides miombo, which are members of the C. imicola complex. A total of 23 994 Culicoides midges were collected from five trap sites in Harare, Zimbabwe, with the dominant species, C. imicola, representing 91.6% of the total collection. Seventeen arboviruses were isolated from these midges, 15 of which were bluetongue virus. The predominant bluetongue virus serotype was serotype 11, followed by serotypes 1, 8, 12 and 15. Bluetongue virus serotypes 1, 2, 8, 10, 12, 15, 16 and 18, detected in this study, had not been previously reported in Zimbabwe.
Agro-ecological classification, known as natural region (NR) classification in Zimbabwe, divides the country into five regions based on mean annual rainfall (Vincent & Thomas 1960). The aim of this study was to identify the species of Culicoides distributed across these regions during the period 1998-2003. It is known that climate has a profound influence on the viability and fecundity of the Culicoides midges (Mellor 1993). To date no differences in disease abundance have been reported between these regions. The monthly abundance and distribution of Culicoides species in these five regions was, therefore, determined. In addition, this study aimed to isolate and identify arboviruses isolated from Culicoides midges collected in Harare over 7 days within the study period. Such studies will expand our knowledge of the biology of Zimbabwean Culicoides species and the potential for transmission of viral pathogens. This information will help assess the risks to both ruminant and equine livestock in Zimbabwe attributable to the presence of the Culicoides vector and hence the implementation of appropriate integrated control strategies (Rawlings et al. 1998).

Selection of Culicoides trap sites in Zimbabwe
Culicoides midges were collected from a convenience sample of 12 trap sites in Zimbabwe. Collections were made over five consecutive rainy seasons (November-April) from 1998 to 2003. Collection site selection was based on several criteria, including: a minimum number of either horses or ruminants (> 8); historical presence of clinical disease or deaths in animals; presence of suitable Culicoides vector breeding conditions (e.g. low-lying areas, such as river banks or marshy fields, with clay soils); security of the equipment; and the availability of an alternating 220V current power supply. Trap sites were distributed to include each of the five NRs of Zimbabwe ( Figure 1). Two trapping sites were established in NR I, six in NR II, one in NR III, two in NR IV and one in NR V. Twenty-four individual collections (trapping nights) were made from NR I, 109 from NR II, 18 from NR III and 9 from NR IV across the five rainy seasons. Trapping was to be conducted for one night every 2 weeks during each rainy season, although this protocol was not followed consistently at each trapping site. Only three individual collections were made in NR V during the entire period under study.
Rainfall is highest in NR I (> 1050 mm per annum) whilst NR II and III contain the most arable land and receive 500 mm -1050 mm of rain per annum. NR I is a specialised and diversified farming region with plantation forestry, fruit and intensive livestock production, whilst NRs II and III are suitable for intensive farming based on crops and livestock production. Severe mid-season dry spells make NR III marginal for maize, tobacco and cotton, or for enterprises based on crop production alone. The farming systems, therefore, are based on both livestock (assisted by the production of fodder crops) and cash crops. NR IV and V have a lower annual rainfall (< 600 mm per annum) and experience periodic seasonal droughts and/or severe dry spells during the rainy season. The farming, therefore, is based on extensive livestock ranching and drought-resistant fodder crops (Vincent & Thomas 1960).

Trapping, sorting and processing of Culicoides samples for identification
Dusk (18:00) till dawn (06:00) trapping was conducted at each trap site using Onderstepoort 220 V down draught suction light traps each fitted with an 8 W black fluorescent (ultraviolet) light tube. The traps were hung on the branches of trees or on the eaves of buildings approximately 1.4 m above ground level and within close proximity to the animals. Culicoides midges were collected into phosphate-buffered saline (PBS) and 0.5% chlorhexidine (Savlon ® , Novartis, U.K.). After collection specimens were re-suspended and preserved in 80% ethyl alcohol and stored in the dark at room temperature until sorted, identified and counted.
Species were identified and counted under a binocular dissecting microscope using a wing pattern atlas of southern African Culicoides species (R. Meiswinkel, unpublished atlas 1996). For large collections in excess of 1000 midges, the total numbers were estimated by measurement of a minimum of five 1 mL subsamples according to the protocol of Van Ark and Meiswinkel (1992).

Trapping and sorting of Culicoides midges in Harare for virus isolation
Culicoides midges were collected for virus isolation at five sites around Harare (17°51′50′S, 31°1′47′E, located in NR II) using the method described above. These five Harare sites formed part of the 12 trappings mentioned previously. These Culicoides species were age graded according to the method described by Dyce (1969). Harare is 1508 m above sea level and situated in the northeast highveld of Zimbabwe. It is characterised by moderate subtropical conditions with an average monthly rainfall of 120 mm and air temperature between 15 ºC and 27 ºC. All trap sites were on smallholdings with vertisol soil types and mixed forest vegetation where urban-intensive farming was practised. The main species of livestock were cattle, goats, sheep and horses.
The traps were hung from branches or eaves of buildings within close proximity to animal premises. All sites' traps were operated from dusk (18:00) to dawn (06:00) for seven consecutive days in March 2003. Culicoides were collected into a phosphate-buffered saline (PBS) solution containing 100 mg/mL of gentamicin and 0.1% chlorhexidine (Savlon ® , Novartis, U.K.) and stored at 4 ºC until processed.
Individual species of Culicoides were stored at 4 ºC in 2 mL Eppendorf tubes (a maximum of 50 flies per Eppendorf tube) with 1 mL of PBS containing 100 mg/mL of gentamicin and 0.1% chlorhexidine (Savlon ® , Novartis, U.K.). For each species or species group, the Culicoides specimens were sorted into the following categories: males, blood engorged females, non-engorged nulliparous females and non-engorged parous females, according to Dyce (1969).

Preparation of Culicoides specimens for virus isolation and identification
Midges were grouped into individual pools of up to 100 nonengorged females of each species. Each pool was homogenised in a mortar using a pestle with 1 mL of Eagles maintenance medium containing 1% foetal calf serum and antibiotics, using a modification of the technique proposed by El Hussein et al. (1989). Each homogenate was subsequently titrated for infectious virus in BHK-21 cells, brains from suckling mice and 11-day-old embryonated hens' eggs. Virus isolates were characterised using serogroup-specific, indirect, sandwich enzyme-linked immunosorbent assay (ELISA) for BTV (Thevasagayam et al. 1996), AHSV (Hamblin, Mellor & Boned 1991), EEV (Crafford et al. 2003) and EHDV (Thevasagayam et al. 1996) antigens. The BTV isolates were subsequently serotyped using virus neutralisation index assays (Hazrati & Ozawa 1965).

Collection of Culicoides across Zimbabwe for species identification
A total of 279 919 specimens of Culicoides, representing 51 different species, were trapped across Zimbabwe over five rainy seasons.   (Table 3), bringing the total number of Culicoides species recognised in Zimbabwe to 60.

Discussion
Trap sites chosen were believed to be representative of the climatic and geo-physical conditions found within each NR. However, the significance of the distribution of Culicoides across the five NRs of Zimbabwe must be interpreted with caution. It is important to realise that these five agroecological regions were characterised more than 50 years ago (Vincent & Thomas 1960). Mugandani et al. (2012) have argued that this is misleading as it implies that the climatic conditions have remained stable over this time. Research, however, points to the contrary; Makarau (1999) and Low (2005) noted increased variability of rainfall, rain days and temperature in Zimbabwe. Most meteorological stations in Zimbabwe have shown a decline in rainfall over the past 100 years (Mugandani et al. 2012). Furthermore, Corbett and Carter (1996) noted that continued use of these agroecological regions was at odds with the political, social and agrarian reform that has taken place in Zimbabwe since Vincent and Thomas (1960) established these regions.
Caution should also be observed when interpreting absolute counts of Culicoides as there may be some variation dependent on the trap sites within each NR. The number of host animals and the distance of these animals from the trap site may also influence the numbers of Culicoides collected. Trap sites in NR II recorded the greatest diversity of Culicoides species, with 43 species identified. The highest median count of Culicoides was recorded in trap sites located within NR III. Natural region III includes the central highveld of Zimbabwe, averaging a total annual rainfall of 500 mm -700 mm, with predominantly greyish brown clay soils and sandy loams derived from granitic rocks (Vincent & Thomas 1960).
nr. angolensis (species closely associated with C. angolensis) High median counts of Culicoides midges were recorded in March and April, at the end of the rainy season. Numbers of adult flies steadily increase after the start of the rainy season as the availability of breeding sites increases (Musuka 1999). Similar findings have been reported previously in Zimbabwe and South Africa, with large populations of Culicoides vectors found in areas with high average annual rainfall and summer temperatures (Musuka 1999;Musuka et al. 2001;Venter, Nevill & Van der Linde 1997). Temperature influences the abundance of Culicoides vectors through its effect on the availability of suitable breeding sites, larval development and adult mortality (Mellor & Hamblin 2004). High environmental temperatures lead to a shorter time span between blood-feeding events and an increase in the rate of virogenesis within the adult vector, leading to earlier virus transmission, although at high temperatures the rate of adult mortality also increases (Mellor & Hamblin 2004;. Previous studies have demonstrated that soil type appears to be very important in determining the distribution and abundance of C. imicola (Meiswinkel 1998;Musuka et al. 2001). In concurrence with these previous findings, the largest numbers of C. imicola were found in areas with a high, moisture-retentive clay soil, whilst the lowest numbers were encountered in rapidly draining sandy soils. The biggest collections of Culicoides were obtained from a large equine stud on the outskirts of Harare. This farm was situated in a low-lying area on the banks of a major river and had clayey soils and open grassland vegetation.
This study showed that large numbers of C. imicola were collected in light traps located at sites where horses were stabled indoors at night even though some stables were fitted with protective fly screens. This suggests that C. imicola were still feeding successfully as they depend upon blood meals for the development of their eggs. Successful feeding may have been facilitated by the common practice of leaving the upper half of the stable door open for ventilation (Meiswinkel 1998). Horses kept out in the open at night appear to be more susceptible to midge bites than stabled horses and these outside horses may act as decoys deflecting C. imicola away from the stabled animals. If these decoys are removed or stabled, then the midges may attempt to feed on stabled animals (Meiswinkel 1998).
This study has identified 10 species of Culicoides that previously have not been recorded in Zimbabwe (Blackburn et al. 1985;Musuka 1999;Musuka & Kelly 2000;Musuka et al. 1998Musuka et al. , 2001Phelps et al. 1982), bringing the total number of recognised species of Culicoides in Zimbabwe to 60. Of particular note was the identification of two additional C. imicola complex species, Culicoides loxodontis and C. miombo, which were distributed across four of the NRs of Zimbabwe.
The epidemiological significance of these species as vectors of arboviruses in Zimbabwe needs further investigation, although the findings of this study significantly add to the available knowledge on the distribution of these species in southern Africa. Musuka (1999) has shown that C. bolitinos has a widespread distribution in southern Africa. Furthermore, C. bolitinos, which was found at all five trap sites in Harare in the present survey, has been implicated in AHSV, EEV and BTV transmission in South Africa (Meiswinkel & Paweska 1998Venter et al. 2002) and, therefore, could play a role in the spread of the disease in Zimbabwe. In our study, the highest prevalence (5.7%) of C. bolitinos was recorded in NR V (Table 2). This region is characterised by low rainfall and high environmental temperatures, although in South Africa C. bolitinos has been reported to be more abundant in the cooler mountainous regions (Meiswinkel & Paweska 2003;Venter et al. 2002). This species breeds in the dung of cattle and buffaloes and the traps in the NR V trap site were located near a cattle pen containing 30 cattle. The trap sites also included a wildlife sanctuary in north Zimbabwe that hosts large populations of wild buffaloes (Syncerus caffer).
The distribution of Culicoides species observed during this study closely matched the sero-incidence and prevalence of AHSV, EEV, BTV and EHDV reported by Gordon (2010). In order to fully understand the role of the various Culicoides species in the transmission of arboviruses in Zimbabwe, further studies are needed to determine the vector competence of the sibling species of the C. imicola group and other Culicoides species that are potential vectors in southern Africa (Musuka et al. 1998). It has recently been demonstrated, for example, that C. magnus could become infected with both BTV and AHSV under laboratory conditions (Paweska, Prinsloo & Venter 2003;Paweska, Venter & Mellor 2002;Venter et al. 2004Venter et al. , 2009. In this study, virus isolations of EEV and BTV were made only from C. imicola and C. magnus. According to Meiswinkel et al. (2004), C. miombo is suspected to be a vector of BTV. In the Harare study, however, no viral isolates were obtained from this species (although only nine individual specimens were collected and assayed). No AHSV or EHDV were isolated from any of the trapped Culicoides species despite the traps being located at horse studs and cattle farms. Low AHSV isolations from Culicoides midges were also reported by Rawlings et al. (1998) in a study conducted in Gambia, which was attributed to an extremely low infection prevalence of AHSV in the vector species of Culicoides. These observations were further authenticated by Mellor, Osborne and Jennings (1984), in Sudan, who could only make two AHSV isolations out of 7000 Culicoides specimens. Many viral, genetic and environmental factors can influence the susceptibility of Culicoides to infection and some species may be refractory to infection (Mellor et al. 1984). It is recommended, however, that large numbers of flies from many different collection sites should be collected and assayed before defining a species as refractory to infection.
The predominant serotype of BTV isolated from the Culicoides vectors was serotype 11, followed by serotypes 1, 8, 12 and 15. Previously only BTV serotype 11 has been isolated from the Culicoides vector in Zimbabwe (Blackburn et al. 1985), so this is the first time that BTV serotypes 1, 2, 8, 10, 12, 15, 16 and 18 have been recorded in the country. The possibility that some of these additional BTV serotypes isolated may be vaccine viruses should be considered, however. The vaccine currently used in Zimbabwe is an Onderstepoort Biological Product ® attenuated cell culture adapted vaccine. Serotypes 15, 16 and 18 are not included in this vaccine and thus vaccinated cattle and sheep may not be adequately protected against all BTV serotypes present in Zimbabwe. There is, therefore, a need to consider the possibility of developing a local polyvalent vaccine that protects against all the BTV serotypes circulating in Zimbabwe.

Conclusion
The greatest diversity and highest median counts of Culicoides were found in the central highveld region of Zimbabwe, which enjoys climatic and geophysical conditions favouring optimal vector breeding and larval development. The discovery of new Culicoides species in Zimbabwe adds to the existing knowledge on the distribution of Culicoides species in southern Africa. These findings, however, highlight the need for further studies to measure the vector competence of existing Culicoides species and to fully understand the role of these species in the transmission and epidemiology of arboviruses in Zimbabwe.
Furthermore, the discovery of novel BTV serotypes in this study highlights the need to review bluetongue vaccines currently in use and vaccination and control strategies being practised in Zimbabwe.