Currently, almost one billion people live in unstable or extremely low malaria risk areas, corresponding mainly to seasonal transmission, highlands (> 1,500 m), arid/semi-arid and urban areas . In the latter, malaria can be a serious public health problem, especially in several rapid growing African cities where migrations of human populations from the countryside is intensive [2–4]. While urban development was generally believed to reduce the risk of malaria vector breeding, and thus malaria transmission, many African countries have declining economies, and most of them are struggling to cope with the pace and the extent of urbanization in their cities [2, 5]. This may favour malaria vectors' breeding sites  and local malaria transmission . In addition, people living in urban regions could be, in spite of their low exposure level to Anopheles vector bites, at a high risk of malarial morbidity and mortality, because of their delayed acquisition or lack of protective immunity . The epidemiology of malaria in cities appeared therefore accurate, and urban malaria has been considered to be a major emerging problem of public health in Africa .
Dakar area is a typical example of sub-Saharan Africa's sprawling cities where malaria risk and transmission has been studied for several years. Malaria pathogens are transmitted by species of Anopheles gambiae complex (namely An. gambiae s. l.). Anopheles arabiensis is the most abundant Anopheles vector species and transmits mainly Plasmodium falciparum [9–12]. Anopheles melas, secondarily associated to P. falciparum transmission and An. gambiae s. s. M form have been described in low densities . Malaria risk in Dakar is very focal, due to a high diversity in the degree and type of urbanization, the variation of density of human populations, the quality of water and waste management, the differential nature and use of vector control strategies and other own household factors . A recent study has underlined possible changes in human exposure level to Anopheles bites in Dakar area. Indeed, due to an important increase of the building developments, estimated about 30% of the total area, the population at high risk for malaria fell from 32% to 20%, whereas the low risk population rose from 29 to 41% between 1996 and 2007 . These findings suggest a current need to evaluate accurately the malaria risk in Dakar settings.
The evaluation of malaria risk using classical entomological methods presents considerable limitations in urban settings. Heavy sampling efforts are required to assess exposure level to Anopheles bites and then to evaluate the risk of malaria by entomological tools (trappings, human-landing catches, residual sprayings, etc.), especially in low urban exposure [15–17]. In addition, such methods are mainly applicable at population level and do not allow the evaluation of the heterogeneity of exposure between individuals. Moreover, these methods are not adapted to consider differential attractiveness to mosquitoes between individuals or other environmental and socio-economic factors, which could induce important variations in individual exposure to vector bites . Such factors could be all the more considerable in urban areas. A simple, specific and highly sensitive indicator is therefore needed to evaluate the human exposure levels to Anopheles bites and potentially the risk of malaria in urban areas, at individual and population levels .
The measurement of antibody (Ab) response to vector saliva in human populations has been described to be a pertinent tool to assess the host exposure level to vector bites and the risk of vector-borne disease [19, 20]. Salivary proteins of haematophagous arthropods facilitate blood feeding by counteracting haemostatic and inflammatory reactions and by modulating the immune response of the human or animal host [21, 22]. Some of them can also induce a specific Ab response  which could represent a reliable indicator of vertebrate host exposure to vector-borne diseases in individuals bitten by arthropod vectors, such as ticks , sand-flies , Triatoma , Glossina , Aedes [28, 29], Culex , An. gambiae [31, 32], Anopheles dirus  and Anopheles darlingi . But, the use of whole saliva could not give a pertinent biomarker, because of i) the potential cross-reactivity with salivary epitopes of other haematophagous arthropods; ii) the lack of reproducibility between saliva batches, and iii) an inadequate production needed for large-scale studies [35, 36]. For these purposes, a specific, antigenic, easy synthesized and highly conserved peptide between Anopheles mosquitoes, the gSG6-P1 (An. gambiae Salivary Gland Protein-6 peptide 1), has been identified and validated as a more pertinent biomarker of Anopheles bites . Indeed, specific IgG responses to this gSG6-P1 peptide seemed to give an accurate evaluation of low and very low-level exposure to An. gambiae  as well as to Anopheles funestus bites, the second major malaria vector in Africa . Recently, it has been also shown that Ab response to gSG6-P1 peptide offers a useful biomarker for a reliable assessment of the efficacy of impregnated bed-net use against exposure to Anopheles bites . In addition, specific IgG response to gSG6-P1 does not seem to build up but wanes rapidly, when exposure failed. This property represents a major strength for its use for evaluating the human exposure to mosquito bites in low-exposure contexts .
The aim of the present study was to evaluate the exposure level to malaria vectors in individuals living in a supposed low endemic urban area (Dakar) by using specific human IgG responses to the Anopheles gSG6-P1 salivary peptide biomarker.