Natural products for malaria vector control: flora, fish and fungi

Introduction
Despite international organisations providing much focus over the past 10 years, malaria is still killing vast numbers of Africans, especially children. It is agreed that malaria can only be successfully controlled by using different control tools simultaneously in the spirit of integrated vector management (IVM), and that African communities will need to become more directly involved in mosquito control (Chapter 2). Using mosquito control tools in a way that requires almost no technical equipment or knowledge will open them up to the rural communities that are best placed to deploy them. In addition, widespread insecticide resistance is reducing the ability of insecticide-based tools to control mosquitoes. For these reasons, biological control and other natural mosquito control methods are being researched by many institutions. Several potential natural control tools are readily available in sub-Saharan Africa. If these tools prove effective and become operational, then it is possible that they will be sustainable because communities can intentionally produce the biological agents themselves, bringing a source of money to rural communities. This would be especially important in areas where infrastructure is poorly developed, and repeat applications of chemical control tools are not easily made. This thesis was designed to test the feasibility and effectiveness of a variety of natural products against both larval and adult malaria vector mosquitoes using low-tech methods in laboratory and field trials.

Part I: Flora
Azadirachta indica A. Juss (Meliaceae) (the neem tree) was chosen due to the already proved mosquitocidal properties, and its ready availability in Africa. We wanted to use neem in a way that could easily be deployed in resource-poor rural areas. Laboratory studies were conducted to examine the larvicidal and pupicidal properties of a crude aqueous extract of neem wood against the principle African malaria vector, Anopheles gambiae Giles s.s. (Diptera: Culicidae) (Chapter 3) [1]. The results indicate that even a relatively low dose of 0.15 grams of dried neem wood in 1 litre of water was able to inhibit the emergence of 90% of mosquito adults when larvae were exposed during their first three larval instars. Even for the fourth (last) larval instar, just 0.6 g/l was required to prevent 90% emergence. Furthermore, neem-exposed larvae exhibited significantly increased development times when compared to the controls. Pupae were also killed by the aqueous neem extracts, and were subject to neem-induced emergence abnormalities, but the concentrations required to kill pupae were much higher than for larvae and not likely to be used operationally. High performance liquid chromatography (HPLC) analysis identified several polar constituents in the aqueous neem extracts including nimbin and salannin. However, azadirachtin was not present in significant amounts. The effect of this extract on the oviposition behaviour of adult female An. gambiae s.s. mosquitoes was then monitored (Chapter 4) [2]. The oviposition results show that when using 0.1 g/l of the crude aqueous neem extract, significantly more mosquitoes laid their eggs when compared to mosquitoes exposed to the control treatment. For the doses 10x and 100x higher, the same proportion of mosquitoes laid their eggs as in the control, indicating that even at much higher doses than required for successful larval control, female oviposition will not be detrimentally affected.

Part II: Fish
Larvivorous fish have previously been shown to effectively control mosquito numbers. Therefore, a census was carried out to examine the current status of fish farming in western Kenya (Chapter 5) [3]. Working with the Kenyan Fisheries Department we found that while fish farming is a favoured activity, 30% of the 261 ponds found did not contain fish. These “abandoned” ponds had significantly more An. gambiae s.l., Anopheles funestus Giles and culicine mosquitoes when compared to the ponds that still contained fish. Furthermore, An. gambiae s.l. was proportionally more abundant in the abandoned ponds when compared to the other mosquito types. Surprisingly, vegetation did not significantly affect mosquito distribution. Following our study, demand for fish to restock abandoned ponds increased by 67% when compared to the previous year. The overwhelming majority of fish being farmed in our census area were fish of the tilapiine subfamily. Given this finding, we set up a small-scale field experiment to study the larvivorous potential of the tilapiine fish Oreochromis niloticus L. (Perciformes: Cichlidae) (Chapter 6) [4]. Taking daily measurements of mosquito numbers, we found that immediately after fish introduction, the density of mosquitoes in the treated ponds dropped in comparison to the increase in the control pond. After 15 weeks, anopheline numbers had decreased by >94% in the ponds containing the fish, and we found that fish were able to sustainably control mosquitoes for at least 6 months, when our study finished. It is concluded that this type of fish could be an effective and sustainable way to control mosquito numbers in rural western Kenya. Furthermore, this fish provides a source of much needed income and protein to rural African communities.

Part III: Fungi
For the control of mosquito adults using natural products, entomopathogenic fungi hold the most promise. In this thesis the entomopathogenic fungi Beauveria bassiana and Metarhizium anisopliae were separately suspended in mineral oil and applied to polyester netting. A laboratory experiment was then conducted to investigate the fungal susceptibility of insecticide-susceptible and insecticide-resistant strains of An. gambiae s.s.. In addition, fungal conidial viability was tested at various time points after application onto polyester netting (Chapter 7) [5]. Whilst both mosquito strains were susceptible to both species of fungal infection, the pyrethroid-resistant An. gambiae s.s. VKPER strain was significantly more susceptible than the insecticide-susceptible SKK strain, dying more quickly. Conidial viability was significantly lower for both species after application onto the polyester netting when compared to the viability in suspension. However, the ability of the treated netting to infect and kill mosquitoes was not significantly diminished over the one week trial period. Given the finding that fungal-treated polyester netting could infect and kill mosquitoes, an experimental hut field trial was conducted in Benin, West Africa, to investigate the effect of fungal treatment on blood feeding behaviour and survival of wild insecticide-resistant mosquitoes. Benin was chosen due to the presence of multi-insecticide-resistant mosquito populations that are threatening the effectiveness of current vector control. We used fungal-treated netting to infect mosquitoes entering the hut windows, and either an untreated or insecticide-treated bednet was placed into each hut to examine how the entomopathogenic fungi would work with current control tools (Chapter 8) [6]. Only enough Culex quinquefasciatus Say (Diptera: Culicidae) mosquitoes were collected from the huts for accurate analysis. Our study was the first to monitor the effect of entomopathogenic fungi on blood feeding of wild mosquitoes. We found that the B. bassiana treatments caused significant and instantaneous reductions in blood feeding. No significant effect of the fungi on mosquito mortality was found. Conidial viability of B. bassiana and M. anisopliae was found to decrease rapidly under field conditions [7].

Conclusions
This thesis used several different experimental techniques to examine the potential of three natural products to control mosquitoes. For the flora, it was found that even a small amount of neem wood in water would control mosquitoes (Chapter 3), and at this and higher doses, the oviposition behaviour was not adversely affected (Chapter 4). Neem trees are readily available in many areas of Africa, and promising field trials indicate that the use of this tree species should be incorporated into malaria control trials.
This thesis reports that edible native African fish can be effective at controlling mosquitoes (Chapter 6), but if fish farming is abandoned and the ponds not filled in, then they can allow large numbers of the most effective malaria vectors to breed (Chapter 5). Fish have been successfully used for malaria vector control in many countries and this could be rolled out across appropriate areas of Africa, as long as it is accompanied with adequate education about the dangers of abandoned ponds.
We found that insecticide-resistant mosquitoes were more susceptible to fungal infection than the insecticide-susceptible strain. Under field conditions fungi were able to prevent blood feeding but did not cause significant mortality in the wild-caught mosquitoes. Although entomopathogenic fungi produce high levels of mortality in laboratory settings, (Chapter 7), their use under field conditions still has a long way to go and is not yet at the operational stage. Although the results found in this thesis are encouraging for the use of fungi in African situations (Chapter 8), further work should be carried out to maximise fungal persistence under field conditions.
The current emphasis is on IVM for malaria control (Chapter 2), and focus is turning to biological control tools that can help manage insecticide-resistant populations. With this in mind, the natural products investigated in this thesis have produced encouraging results that show they have the potential to be integrated into malaria control strategies. Furthermore, flora and fish are readily available in the areas where they are most required, and could be used almost immediately to help reduce mosquito numbers and correspondingly, malaria disease transmission.