Odour - mediated host - seeking behaviour of the Afro-tropical malaria vector Anopheles gambiae Giles

B.G.J. Knols

Research output: Thesisinternal PhD, WU


<br/>Malaria remains the single most important parasitic disease of man in tropical regions of the world. It is estimated that 40% of the world's population, in 102 <em></em> countries, is at risk from the disease. Some 100-200 million cases occur annually worldwide, of which 90 million in Africa, with 1-2 million deaths.<p>Efforts to control malaria by chemoprophylactic and/or curative drugs are seriously jeopardized due to widespread parasite resistance, and <em>Anopheles</em> mosquito vectors are rapidly developing physiological resistance against insecticides commonly used for indoor spraying or impregnation of bed nets. Trials with the malaria vaccine SPf66 have shown a 31% morbidity reduction in Tanzania but hardly had an effect on disease in The Gambia. The deteriorating malaria situation in the world thus urges the development of new control methods which need to be environmentally safe, socio-economically acceptable, and applicable for use within primary health care systems.<p>The African mosquito <em>Anopheles gambiae sensu stricto</em> Giles, is a prime vector of malaria due to its distinct preference for biting man (anthropophily). It generally feeds indoors (endophagy) between midnight and dawn and rests indoors after feeding for the first <em>1-2</em> days of its gonotrophic cycle (endophily). High survival rates and (seasonal) occurrence in high densities further increase its vectorial capacity. DDT resistance has been reported in several countries and other control measures such as environmental management are not possible due to its breeding in a wide variety of (sunlit) water bodies. Exactly how a female <em>An. gambiae s.s.</em> exploits human-related cues to orient itself while locating a host at night remains largely unknown. Field studies on closely related sibling species have shown that olfactory cues evoke behavioural responses from mosquitoes over considerable distances, but the identity of these kairomones has not been established. Furthering the knowledge of this process might ultimately result in the development of new control strategies analogous to those for tsetse flies. The host-seeking behaviour of tsetse flies has received considerable interest over the last two decades and has resulted in bait systems now widely applied for monitoring and control of trypanosomiasis in several African countries.<p>The goal of the work presented in this thesis was to gain more insight in the chemical ecology of <em>An. gambiae s.s.</em> and the resulting interactions with its human host. This encompassed identification of kairomones and the development of tools to study their effect on mosquito behaviour in the laboratory and field. The work was carried out in collaboration with research groups in the United Kingdom, Italy and Tanzania within a European Community funded project titled: 'Behavioural studies on malaria vectors'.<p><strong>Haematophagy, disease transmission, and factors affecting host-seeking by mosquitoes</strong><em>(Chapter 1)</em><p>The two possible evolutionary routes leading to haematophagy are described, these being the prolonged association with vertebrates and morphological pre-adaptation for piercing the skin surface. Three associations, depending on the time spent on the host, exist between vertebrates and haematophagous Diptera and pathogen transmission is suggested to benefit most in association with temporary ectoparasites. The host-seeking process, defined as any kind of behaviour that increases the chance to encounter a suitable host followed by blood-feeding, is categorised into two classes, namely 'appetitive search' behaviour prior to contacting host-related cues and activation/orientation upon perceiving these. Three classes of cues produced by hosts are defined namely olfactory, physical and visual cues. The relative importance of cues belonging to different classes depends on the behavioural ecology of the insect, and is affected by factors such as circadian rhythm and host range. Carbon dioxide, expired by all vertebrates, has been incriminated to affect the host- seeking process of all Dipteran blood-feeders. It is thought that this chemical dominated the olfactory-mediated part of host-seeking during the evolution of haematophagy. Specialism towards feeding on specific hosts required olfactory responses to odours other than carbon dioxide typifying the identity of those hosts. To date, only few of these compounds have been identified and for mosquitoes none of these are used on a large scale in the field.<p><em>An. gambiae s.s.</em> belongs to a complex of six sibling species with large variations in host-specificity and feeding habits. Host-specificity is genetically determined and differential responses to olfactory cues from different hosts have been observed in the field. Although carbon dioxide has been found to elicit responses from the various siblings, this chemical obviously can not be responsible for these differential responses. Odours other than carbon dioxide must influence this host-selection process and identification of kairomones for the highly anthropophilic <em>An. gambiae s.s. is</em> addressed in subsequent chapters of this thesis.<p><strong>Selection of biting sites by mosquitoes on man: a new approach to kairomone identification</strong><em>(Chapters 2 & 3)</em><p>A preliminary study on the selection of biting sites by haematophagous Diptera showed that only few species display a random biting pattern on their hosts. The vast majority preferentially bites specific parts of the body, and these preferences have been attributed to factors intrinsic to the behaviour of the insect (e.g. visual responses to extremities of hosts) or to the behaviour of the host (e.g. defensive behaviour). By studying the selection of biting sites by five mosquito species on the same volunteer under identical experimental circumstances it was observed that a) different species prefer different parts of the body, b) that for at least three of these species this process was influenced by odour emanating from the preferred biting region and c) that these preferences might not only reveal the origin of chemical cues involved in hostseeking, but that the biting pattern can provide information on the range of hosts fed upon. <em>An. gambiae s.s.</em> showed a strong preference for biting the legs and foot region (76% of all bites) and alteration of the odour from this region, by washing the feet with a bactericidal soap, significantly altered this preference (to 36%). Likewise, the significant preference of <em>An. atroparvus</em> and <em>An. albimanus</em> for biting the face (50 and 49% of all bites respectively) was influenced by breath since the pattern of biting changed dramatically after breath was removed from the host (to 19 and 20% respectively).<p>Studying the selection of biting sites by mosquitoes was proven to be a new method to reveal the origin of olfactory cues influencing their host-seeking behaviour.<p><strong>Odour-mediated host-seeking behaviour and kairomone identification: a laboratory approach</strong><em>(Chapter 4,5 & 6)</em><p>At the onset of the present study it was realised that a reliable laboratory bioassay was essential to facilitate rapid screening of behavioural effects of (human) odours on <em>An. gambiae s.</em> s.. The then existing windtunnel, which was constructed to study responses of individual females, was accordingly modified and resulted in a system closely resembling conventional dual-port olfactometers used for screening behavioural effects of odours on the yellow fever mosquito <em>Aedes aegypti</em> . Using this system, test odours were classified as 'attractive' whenever trap catches were significantly higher (using chi-squared) then those of control traps. Carbon dioxide at a human equivalent (4.5% in clean air) caught 4 times more <em>An. gambiae s.s.</em> than control traps baited with clean air only.<p>Results from chapters <em>2 & 3</em> showed that foot odour played a role in the selection of biting sites for <em>An. gambiae s.s.</em> and breath influenced this process for <em>An. atroparvus/ An. albimanus.</em> Since carbon dioxide at a human equivalent was clearly 'attractive' it was assessed whether whole human breath was similarly attractive. It was shown that human breath (of three volunteers) was not attractive when tested against the carbon dioxide concentration (3.30-3.65%) <em></em> of that breath only. When breath of the same volunteers (with carbon dioxide concentrations between 3.28-3.80%) <em></em> was tested against clean air this was also not attractive, nor was carbon dioxide at a concentration of 3.56%. <em></em> The difference in response to this latter concentration of carbon dioxide and the highly 'attractive' 4. 5% concentration was attributed to the dilution of the gas to levels which did no longer result in catches significantly different from those in the control traps.<p>The headspace of Limburger cheese, to the human nose reminiscent of foot odour, was significantly attractive to <em>An. gambiae s.s.</em> from East and West Africa. GC and GC-MS analyses of the acid fractions of Limburger cheese, human sweat and toe scrapings showed strong similarities in their aliphatic fatty acid content and composition. The fact that <em>An. gambiae s.s.</em> responded to odours of bacterial origin (produced by <em>Brevibacterium linens</em> on the cheese) strongly suggests that the human skin microflora is (at least in part) responsible for the production of kairomones for this species. Nineteen saturated and unsaturated fatty acids, ranging in carbon chain length from C <sub><font size="-2">2</font></sub> -C <sub><font size="-2">18</font></sub> were identified in the acid fraction of the cheese, whereas twelve of these were present in its headspace. In the windtunnel bioassay, the undiluted acid extract was repellent, but became significantly attractive at concentrations of 5*10 <sup><font size="-2">-4</font></SUP>to 10 <sup><font size="-2">-7</font></SUP>. A synthetic blend of the twelve acids identified in the headspace was significantly attractive at a concentration of 10 <sup><font size="-2">-8</font></SUP>. Electroantennographic studies showed significant and reproducible responses to (saturated) cheese headspace, the synthetic mixture, and all individual acids but hexadecanoic acid. These fatty acids form the first group of kairomones identified for <em>An. gambiae s.s.</em><p><strong>Host attractiveness and odour-mediated host-seeking behaviour of mosquitoes: a field approach</strong><em>(Chapter 7,8,9 & 10)</em><p>The study of the effects of kairomones on mosquito behaviour in Tanzania was initially hampered by the absence of sampling devices. Standard CDC light traps, commonly used for sampling malaria vectors indoors, were not applicable for outdoor use. Tents ('mobile huts) were modified for sampling mosquitoes by fitting exit traps and constructing artificial eaves for mosquito entry. These tents could be positioned at various sites, in randomized block designs, and could be washed which make them ideal for studying the attractiveness of different odours. Initially it was assessed whether three isolated individuals varied in their attractiveness to <em>An. gambiae s.l., An. funestus, Culex quinquefasciatus</em> and <em>Mansonia</em> spp. <em>,</em> which for all but the latter species was the case. These differences were attributed to the odour of the volunteers since this was the only variation factor not eliminated in the experimental setup. Also, this experiment showed for the first time that <em>isolated</em> individuals vary in their attractiveness to mosquitoes.<p>In addition to these results it was found that different mosquito species show preferences for different individuals within a group. Catches of <em>An. funestus</em> and <em>An. squamosus</em> were inversely correlated for two individuals of the group. Furthermore, a negative correlation between group size and individual catch was observed, which suggests that mosquito sampling is more efficient if individuals are not grouped.<p>The tents were also used to assess the relative importance of the physical presence of man versus his odour only in the number of mosquitoes attracted (and caught). This was done by positioning a man inside a bed net or underground in a pit from which odours were pumped into the bed net. Both treatments caught equal numbers of <em>An. gambiae s.l.</em> and <em>An. funestus,</em> which indicates that human odour is the main factor in attracting these species to a bait inside a house. When carbon dioxide (at a human equivalent) was released inside a bed net the catches were 9 and 27% of those by man for <em>An. gambiae</em> s.l. and <em>An. funestus</em> respectively. A five-fold increase in carbon dioxide dose increased the catch of the latter species to 69% but those of <em>An. gambiae s.l.</em> remained the same. The relative contribution of carbon dioxide to the overall attractiveness of man to <em>An. gambiae s.l.</em> was thus limited and underlined the importance of other human odours in host-seeking by this species.<p>Electric nets (a grid of wires electrocuting insects when they contact them) have played a paramount role in the study of tsetse fly behaviour, and it was studied whether similar sampling systems were applicable for outdoor use, and more specifically to study responses to odours and analyse flight behaviour. Various designs were tested, and when baited with carbon dioxide, sampled all mosquito species present in the study area. Furthermore it was shown that primarily zoophilic mosquito species were attracted to human breath or carbon dioxide, but that the anthropophilic species <em>(An. gambiae s.1.</em> and <em>An. funestus)</em> hardly responded to these baits. Whole human odour baited nets caught higher numbers of these latter species, thus confirming the results from the tent experiments. A large cylindrical net was used to assess the flight direction of mosquitoes upon contacting carbon dioxide and showed that they engage in positive anemotaxis. The overall efficiency (number caught/number approached) of a rectangular electric net (28 x 40 cm, wires at 4 mm), when baited with carbon dioxide, was 42%. This work showed that electric nets can be used for outdoor sampling of mosquitoes, for analysing responses towards laboratory identified kairomones, and for the analysis of mosquito flight behaviour. <strong></strong><p><strong>Odour-mediated host-seeking behaviour of <em>An. gambiae s.s.</em> in relation to its human host: a case for adding the skin microflora and <em>Plasodium</em> parasites as interactants</strong><em>(Chapter 11)</em><p>The general discussion reviews the results of the research in a broader context. The preceding chapters suggested that olfactory cues influencing the host- seeking behaviour of <em>An. gambiae s.s.</em> originate <em></em> from the human skin. Their production is (at least partially) the result of the metabolic activity of the skin microflora. It is argued that the influence of the skin microflora on the interaction between man and mosquito is of such importance that this justifies the recognition of the skin microflora as a separate entity. In addition to this it is hypothesized that the influence of the <em>Plasmodium</em> malaria parasite may affect the production of infochemicals mediating the interaction between man and mosquito, or, alternatively, is responsible for the production of these allomones. It is concluded that the study of odour-mediated host-seeking behaviour of <em>An.</em><em>gambiae s.s.</em> will <em></em> benefit from analysing the interactions between man and mosquito within a multipartite, rather than the conventional bitrophic (man- mosquito) context.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • van Lenteren, Joop, Promotor
  • Takken, W., Promotor
Award date23 Feb 1996
Place of PublicationS.l.
Print ISBNs9789054854876
Publication statusPublished - 1996


  • Culicidae
  • Anopheles
  • malaria
  • animal behaviour
  • host-seeking behaviour
  • nematocera


Dive into the research topics of 'Odour - mediated host - seeking behaviour of the Afro-tropical malaria vector Anopheles gambiae Giles'. Together they form a unique fingerprint.

Cite this