Abstract
A protocol has been developed for the indoor evaluation of candidate spatial repellents intended for use in push and pull systems. Single treatments (catnip oil, 1-methylpiperazine, and homopiperazine) and a mixture of catnip oil and homopiperazine were tested with yellow-fever mosquitoes (Aedes aegypti) in Y-tube olfactometers to determine 1) if these compounds inhibited mosquito host-seeking at short distances and 2) if results obtained in olfactometer tests can be correlated with a larger scale set-up, that is, a room test. All test materials significantly decreased the ability of mosquitoes to find host odors (from a human finger) by up to 96.7% (2.5% catnip and homopiperazine mix). Similar effects could be observed within a new room test set-up, which involved a repellent dispensing system and an attractive trap (BG-Sentinel). Mosquitoes captured by the BGS trap had to fly through a treatment-containing air curtain created by the dispensing system. Compared with the use of a control (ethanol solvent without candidate repellent), trap catch rates were significantly reduced when 5% catnip, 5% 1-methylpiperazine, and 5% homopiperazine were dispensed. Homopiperazine produced the greatest level of host-seeking inhibition with a 95% reduction in the trap catches. The experimental set-up was modified to test the viability of those technologies in a simple push & pull situation.. The combination of BGS trap and a 10% mix of catnip and homopiperazine helped to reduce human landing rates by up to 44.2% with a volunteer sitting behind the air curtain and the trap running in front of the curtain.
Mosquito-borne diseases are a major threat to human health. Half of the world's population is at risk of malaria, which caused an estimated 655,000 deaths in 2010 (World Health Organization [WHO] Fact Sheet N° 94, 2012) and around 2.5 billion people in >100 countries are at risk of dengue fever (DF). In contrast to malaria, no treatment and no vaccine are available against DF; therefore, control of this disease depends primarily on measures taken against the vectors (WHO 1997, Horstick et al. 2010), Aedes aegypti (L.) and Ae. albopictus (Skuse). Traditional control methods like adulticidal fogging are frequently inadequate because the adult mosquito rests in secluded sites (Matthews 1996). Furthermore, indiscriminate or inefficient application has led to increased development of insecticide resistance (Fonseca–Gonzales et al. 2010, Polson et al. 2011). An alternative could be push and pull, which has been reported as a strategy for integrated pest management (IPM) (Pyke et al. 1987). The approach of this seminal work used a combination of repelling and attracting stimuli to control the distribution of insecticide-resistant cotton moths (genus Heliothis). Through the use of both deterring and attracting stimuli, the abundance of insect pests can be changed in a given area by interfering with the ability of the target pest to find their resource (“push”) and luring them to an alternative source where they are trapped and killed (“pull”). Currently, most successful push and pull techniques are used in crop pest management, but similar strategies may improve the control of mosquitoes and other disease vectors (Cook et al. 2007). Our approach involves the BG-Sentinel trap (BGS) as the pull component, because it is a superior trap for Aedes species, such as Ae. albopictus, Ae. aegpyti, and Ae. polynesiensis (Marks), even without the use of CO2 (Azil et al. 2010, Farajollah et al. 2009, Meeraus et al. 2008, Schmaedick et al. 2008). The BGS trap attracts host-seeking females by mimicking convection currents produced by a human body, through visual cues and by emitting artificial host odors from a synthetic attractant dispenser, the BG-Mesh Lure (BG ML; Kroeckel et al. 2006). The synthetic lure is composed of lactic acid, caproic acid, and ammonia. These compounds are present on human skin and are known to play an important role in the host finding process of Ae. aegypti (Geier and Boeckh 1999). Numerous substances have been identified to act as mosquito repellents but in contrast to its Latin origins (“repellere” = to repulse, to drive away), some common mosquito repellents do not mediate a targeted movement away from their source (that would result in contact prevention) but rather work at a short distance or through direct contact and instead result in bite prevention (Bernier et al. 2007). Few substances with spatial properties have been discovered; however, catnip [Nepeta cataria (L.)] essential oil has been one of the most promising. Olfactometer bioassays detected its spatial effects against Ae. aegypti and Anopheles albimanus (Weidemann) (Bernier et al. 2005). In more recent research, 1-methylpiperazine and homopiperazine were reported as compounds that interfere with host-seeking ability and therefore act as attraction-inhibitors of kairomones (Bernier et al. 2012). In olfactometer bioassays, these compounds reduced the attraction of Ae. aegypti and An. albimanus toward a synthetic human odor blend from 92.7 to 12.8% and from 67.5 to 8.2%, respectively. To date, those compounds have only been evaluated in olfactometers and their performance under more realistic conditions, for example, in a room, or outdoors in a field setting is unknown. The use of a potent spatial repellent is crucial to the success of a push and pull control system. To address the issue of scaling up this technology for field use we have developed a room test protocol that involves the use of a simple repellent dispensing system in combination with an attractive BGS mosquito trap and human bait to investigate how trap catch rates and human landing rates are altered in the presence of test repellent compounds. Results from conventional olfactometer bioassays will be compared with room tests to determine how well results from laboratory olfactometers correlate with results from larger scale room tests.
Materials and Methods
Chemicals, Equipment, and Laboratory Assays.
Test Materials. 1-methylpiperazine and homopiperazine were purchased from Sigma–Aldrich (Taufkirchen, Germany), pure catnip (Nepeta cataria), and thyme (Thymus vulgaris) essential oils were acquired from Aromaland (Röttingen, Germany). All single compounds were diluted vol:vol in ethanol (96%, p.a.) to final concentrations of 2.5 and 5%. In addition, 5% catnip oil and 5% homopiperazine were mixed at a 1:1 ratio to obtain a 2.5% ethanolic formulation of the two compounds. The 10% formulation was obtained by mixing 20% catnip oil and 20% homopiperazine at a 1:1 ratio. A proprietary repellent formulation (Autan Protection Plus, SC Johnson GmbH, Erkrath, Germany) containing 20% picaridin (hydroxyethyl isobutyl piperidine carboxylate) was acquired from a local drugstore.
Test Mosquitoes.Ae. aegypti females aged 10–21 d were used for all tests. Preliminary olfactometer tests revealed that our colony shows a comparable susceptibility for spatial repellents at days 6–20 after emergence while responses to a finger or spatial repellent show greater variations at a younger age (1–5 d). The colony was obtained originally from BAYER AG, Monheim, Germany, and has been maintained in our facilities over the past 15 yr. Mosquitoes were reared at 26 ± 1°C and 60 ± 5% relative humidity (RH) under a photoperiod of 12:12 (L:D) h. After hatching of the eggs, larvae were kept in a water basin (30 × 30 × 10 cm) filled with a 1:1 mixture of tap water and deionized water and fed with Tetramin fish food flakes (Tetra GmbH, Melle, Germany). Pupae were transferred into breeding cages (40 × 30 × 20 cm). Adult mosquitoes were provided with a 10% glucose solution on filter paper. Behavioral tests were performed with host-seeking females, which were lured out of their breeding cages at least 10 min before the start of the tests. The breeding cages contained a circular opening covered by fine mosquito netting in the left wall, while the right wall was fitted with a port and rotating door, where a transfer container could be attached. The transfer container consisted of a perspex cylinder with a rotating door on one end and a cover made from fine mosquito netting at the other end. A fan running at 7.5 V was connected to the opening in the left wall of the breeding cage to gently pull air into the cage, while a human hand was held to the transfer container on the opposite side of the cage and rotating doors were opened. Female mosquitoes seeking a bloodmeal flew upwind into the transfer container, attracted to the skin odors.
Y-Tube Olfactometer Assays. Olfactometer tests were performed according to Geier and Boeckh (1999). In total, four Y-tubes, identical in construction, were used to measure the behavioral responses of host-seeking Ae. aegpyti females toward catnip oil (CN), 1-methylpiperazine (MP), homopiperazine (HP), thyme oil (THY), a mix of catnip and homopiperazine (CN-HP), and picaridin (PIC). Each olfactometer consisted of a transparent Plexiglas base leg, followed by a decision chamber and two branches which terminated into attached Teflon chambers, where the test stimuli were introduced (Fig. 1). A constant air stream from the institute's pressurized air system was purified with a filter of activated charcoal, heated up to 26 ± 1°C and humidified to a relative humidity of 70 ± 5% before it was transported through the tube system and into the base leg, with wind velocities of 0.4 m/s in the branches and 0.2 m/s in the base leg. Rotating doors in both branches, as well as at the downwind end of the base leg, allowed the release and entrapment of the test mosquitoes. Cohorts of 15–21 mosquitoes were attached to the apparatus at its downwind end.
Y-tube olfactometer according to Geier et al. (1999). Cage 1 (release chamber): cage that contains the test mosquitoes. Decision chamber: Box, in which the mosquitoes can decide between the two branches of the olfactometer. Cages 2 and 3: Test cage, where volatile stimuli are applied and control cage with pure air. Grid and rotating doors made from mesh gauze. Total length of the apparatus: 105 cm.
Test Procedure. Before the start of a test, 30 μl of an ethanolic test formulation were applied to a 1 × 3 cm filter paper strip (Schleicher & Schuell Microscience GmbH, Dassel, Germany), which was attached to a tempered metal wire and suspended into a Teflon chamber. The door of the base leg remained closed for 15 s to keep the test mosquitoes in the airstreams containing repellent. A forefinger was then inserted into the Teflon chamber behind the paper strip and the rotating door of the base leg was opened. Mosquitoes were allowed 15 s to fly upwind and decide between the test branch with volatile stimuli or the control branch with pure air. The rotating doors were closed and the number of mosquitoes that migrated from the release cage (=active), the number of mosquitoes inside the test cage (where the stimuli were applied) and the number of mosquitoes inside the control cage (with filtered air) were documented. At the conclusion of a test, the airflow in the apparatus was inverted and mosquitoes were lured back into the release cage by the palm of the hand and the next of four Y-tubes was used for testing. By rotating in this manner, mosquitoes had a minimum of 5 min to recover after each test. Treatments were tested in randomized order, and after each run, the control branch and test branch were changed to avoid position or adaptation effects. There were 10 replicates of each single compound or mix. All treatments were tested against a control of the forefinger and a paper strip treated with 96% ethanol. The spatial properties of 5% CN, 5% THY, 5% MP, and 5% HP were evaluated in experiment 1. Experiment 2 included the evaluation of 20% PIC in comparison to 2.5% THY and Y-tube experiment 3 compared the effects of CN, HP, and CN-HP at 2.5% and MP at 5%.
Data Analysis Olfactometer Assays. For each treatment, mean percentages of active test mosquitoes, mosquitoes inside test chamber, and control chamber were calculated, as well as corresponding standard errors. Means were subjected to an arcsine transformation and then compared using a one-way analysis of variance (ANOVA) and Tukey's honestly significant difference (HSD)-test as a post hoc test to verify significant differences between single treatments.
Room Tests With Repellent Evaporating System and BGS Trap. In experiment 4, treatments of 5% CN, MP, HP, THY, and a 2.5% CN-HP mix were investigated in a novel experimental set-up to simulate conditions of real world usage. Tests were performed in an air-conditioned 40.25 m3 windowless room (4.6 × 3.5 × 2.5 m) with artificial light from two fluorescent tubes (350 Lux). The temperature and relative humidity of the air in the room were 25 ± 1°C and 60 ± 5%, respectively. The ventilation from the air conditioner entered the room through an opening in the ceiling and exited the room through a second opening 4.5 m apart on the far side of the ceiling. A tent structure comprised of cotton fabric was built around the air entry with bottom edges held on the floor by wooden bars. The tent had three sides and contained a volume of 5.2 m3 (1.2 × 1.2 × 2.5 m) with the open side that was 2.2 m2 (1.2 × 1.8 m). A repellent dispensing system was placed at the top of the open side of the tent (Fig. 2). The dispensing system consisted of a polyethylene (PE) tube (length 1 m, diameter 0.5 mm; Festo AG & Co. KG, Esslingen, Germany) attached to a tripod, a 500 ml fritted gas wash bottle, flow meter and compressed air connection. The tube served as a dispensing device and contained fine holes (diameter 0.2 mm, distance between holes 2 cm) at the rear side to release the test volatiles. For each test, 500 μl of the ethanolic treatment formulations were pipetted onto round filter papers (Schleicher & Schuell Microscience GmbH) at the bottom of the fritted wash bottle. In control tests, 500 μl ethanol was used. Pressurized air was passed through the bottle at a flow rate of 15 liters/min, then loaded with treatment as it continued to flow into the PE tube attached to the tripod. In this way, a treatment enriched air curtain was released at the top of the tent window. To avoid a mixing of the treatments within the dispensing system, dedicated PE tubes and wash bottles were used for each treatment.
New experimental room test set-up containing a tent structure, repellent dispensing device and BGS trap.
Room Test. A BGS trap fitted with BG ML was placed inside the tent to attract host-seeking Ae. aegypti to fly through the curtain for potential capture. For each test, 10 mosquitoes were released into the room at the side furthest from the tent. After release, mosquitoes were allowed to respond for 15 min. At the end of the test time, the catch rate of the trap was documented and free flying mosquitoes were sucked away with a modified hand-held vacuum cleaner. Mosquitoes that did not approach the investigator or that were still sitting inside the transport cage were recorded as inactive. There were 10 replicates conducted for each single compound or mix and results were compared with the control (BGS trap and a repellent-free air curtain). Additional control tests without air curtain were performed to determine if air movement alone was a physical barrier that prevented mosquitoes from flying into the tent. Treatments were tested in a randomized order. To avoid an accumulation or mixing of the volatile stimuli, the room was aerated for at least 30 min before the next test was conducted.
Room Tests with Repellent Evaporating System and Human Bait. In experiment 5 a volunteer sat behind the air curtain in the middle of the tent and provided attractive odor cues in comparative tests of 10% CN-HP, 20% PIC, and repellent-free ethanol controls. The test procedure was consistent with that described above (experiment 4), apart from the following modifications: Mosquitoes that flew through the air curtain and landed on the volunteer were collected with a modified hand-held vacuum cleaner. The total numbers of mosquitoes entering the tent within 15 min as well as the times of landing were documented.
Room Test With Repellent Evaporating System, Human Bait, and BGS Trap. A simple push and pull situation was evaluated in experiment 6. A BGS trap was positioned on the left side of the tent opening and fitted with a BG ML dispenser. Two independent trials were conducted. The first used the regular tent opening from all former tests (experiments 4 and 5). In the second trial, the size of the opening was reduced to 80 × 80 cm that was centered in the tent. In both set-ups (regular and reduced tent opening), a 10% CN-HP mix was compared with 20% PIC and repellent-free ethanol controls. The test procedure was consistent with that described for experiment 4, apart from the following modifications: Mosquitoes which flew through the air curtain and landed on the volunteer were collected with a modified hand-held vacuum cleaner. The total number entering within 15 min and the times of landing were documented. At the end of the test, the catch rate of the BGS trap was recorded and free flying mosquitoes were removed.
Data Analysis Room Tests. For each treatment, mean percentages and standard errors of active test mosquitoes and mosquitoes caught by the BGS trap and/or volunteer were calculated. Means were subjected to an arcsine transformation and then compared using a one-way ANOVA and Tukey's HSD-test as a post hoc test to compare differences between treatments. In experiments 5 and 6, times of entry of the test mosquitoes were of a normal distribution, thus mean times of entry as well as corresponding standard errors were calculated and compared using one-way ANOVA and Tukey's HSD-test as a post hoc test.
Results
Olfactometer Assays.
Experiment 1 consisted of the evaluation of CN, THY, HP, and MP at 5%. The THY essential oil was selected as a negative control because it was reported to have weak spatial effects on Ae. aegypti in olfactometer tests (Drapeau et al. 2009). The flight activity was high when odors from the finger were presented in the airstreams in combination with ethanol (Fig. 3A, B). The proportion of mosquitoes that left the release cage and flew into the test cage averaged 65.9 to 72.6%. Compared with the ethanol controls, all test compounds had an inhibitory effect on the tests mosquitoes. With an average reduction of up to 45% in tests with MP, the overall flight activity was significantly reduced when CN, THY, HP, and MP volatiles were present in the olfactometer (F = 16.91; df = 4; P = 0.00). The proportion of mosquitoes reaching the test cage close to the stimulus source was also significantly reduced during compound tests (F = 68.93; df = 4; P = 0.00); however, CN, MP, and HP produced a significantly greater reduction than THY (P ≤ 0.038; Tukey's HSD-test).
Y-tube olfactometer experiment 1. Mean percentages and SE of active mosquitoes, mosquitoes inside test cage (=attracted) and inside control cage. Treatments: Finger (F) plus ethanol (EtOH) control in comparison to finger plus repellent. Each treatment was tested in 10 repetitions with Ae. aegypti.
Experiment 2 evaluated 20% PIC as a negative control. The selection of PIC was based on observations made during repellent efficacy tests of this substance in cages. Shortly after application, Ae. aegypti tends to land on the treated skin but then immediately takes off again, indicating that repellency requires direct contact at the normally applied topical concentration. The repellency of 20% PIC was compared with ethanol controls and 2.5% THY. Flight activity was high and reached an average of 81.4% in control tests. At the end of the tests an average of 69.3% of the test mosquitoes were found in the test cage. No inhibition of attraction was observed when 20% PIC was released and the proportion of activated mosquitoes (85.6%) and mosquitoes inside the test cage (72.5%) were not significantly different from ethanol controls (P = 0.97; Tukey's HSD-test). However, 2.5% THY produced a significant reduction in the proportion of activated mosquitoes (F = 10.3; df = 2; P = 0.00) and the proportion of mosquitoes that reached the stimulus source (F = 17.48; df = 2; P = 0.00) with average reductions of 28.9 and 56%, respectively. In experiment 3, the effects of 2.5% CN, HP, CN-HP, and 5% MP were compared with ethanol controls (Fig. 4). All single compounds and the mixture significantly reduced the proportion of activated mosquitoes (F = 13.01; df = 4; P = 0.00) and the proportion of mosquitoes attracted to the treatment stimuli (F = 28.52; df = 4; P = 0.00). The 2.5% CN-HP reduced the proportion of mosquitoes that flew into the test cage by 96.7%; this was the highest level of reduction recorded for these tests within the olfactometer.
Y-tube olfactometer experiment 3. Mean percentages and SE of active mosquitoes, mosquitoes inside test cage (=attracted) and inside control cage. Treatments: Finger (F) plus ethanol (EtOH) control in comparison to finger plus repellent. Each treatment was tested in 10 repetitions with Ae. aegypti.
Room Tests With Repellent Evaporating System and BGS Trap.
Experiment 4 was a novel test designed to evaluate the spatial effects of CN, MP, HP, and a CN-HP mix in a room by measuring BGS trap catch rates compared with THY and ethanol controls (Fig. 5). Two controls of the BGS trap were recorded with the evaporating system either switched on or off. Mean control catch rates showed no significant differences between these two test conditions (F = 0.80; df = 18; P = 0.42). All of the test compounds significantly reduced the catch rates of the trap (F = 37.46; df = 6; P = 0.00) but no differences in the general activity of the test mosquitoes were observed (F = 0.40; df = 6; P = 0.875). Compared with control tests with a repellent-free air curtain, 5% THY produced the weakest effect with a decreased average trap catch rate of 30%. Greater reductions in the trap catches were observed when 5% MP, 2.5% CN-HP, and 5% HP were dispensed into the room (P = 0.00; Tukey's HSD-test), the highest with a reduction of 95.3% in tests with 5% HP.
Room test experiment 4. Mean percentages and SE of active mosquitoes and proportion of active mosquitoes caught by the BGS trap within 15 min. Each treatment was tested in 10 repetitions with Ae. aegypti. BG-S air+: control tests with air curtain, BG-S air−: control tests without air curtain.
Room Tests With Repellent Evaporating System and Human Bait.
In experiment 5, a volunteer sat inside the tent and documented the number of mosquitoes entering and the times of landing when treatments of either ethanol, 20% PIC or a 10% CN-HP were dispensed at the tent opening. None of the treatments prevented test mosquitoes from entering into the tent. All of the mosquitoes that were released had reached the volunteer in control tests and tests with 20% PIC and an average of 97.2% were caught when 10% CN-HP was dispensed. The mean landing/catch rates were not significantly different (F = 2.184; df = 2; P = 0.132); however, the mean times when mosquitoes entered and landed on the volunteer were significantly delayed when 10% CN-HP was released (F = 15.26; df = 2; P = 0.00) (Fig. 6).
Room test experiment 5. Mean times and SE of the first, fifth, and ninth test mosquito landing on the volunteer. Means were generated from 10 replicates per treatment.
Room Tests With Repellent Evaporating System, Human Bait, and BGS trap.
We next investigated whether the presence of a BGS trap adds to a reduction of mosquitoes entering the tent and landing on the volunteer. In experiment 6, 10% CN-HP and 20% PIC were used in combination with the BGS trap and evaluated in two settings: 10 replicates with the regular tent opening and 10 replicates using a window opening on the front side. During tests with the regular tent opening, 91 and 82.1% of the test mosquitoes reached the volunteer during ethanol control tests and tests of 20% PIC, respectively. Trap catches averaged 9% for controls and 17.9% for PIC. The number of entering & landing mosquitoes was significantly reduced when 10% CN-HP was dispensed (F = 16.95; df = 2; P = 0.00) while the number of mosquitoes caught by the BGS trap increased (F = 13.37; df = 2; P = 0.00). Compared with control tests, the proportion of mosquitoes landing on the volunteer was reduced by 42.4% when 10% CN-HP was dispensed while trap catch rates increased almost fivefold. The mean landing times of the test mosquitoes were also significantly delayed when 10% CN-HP was used in combination with the BGS trap (F = 6.42; df = 2; P = 0.005). During control tests, the first mosquito was caught after an average of 31.7 ± 4.6 s compared with an average of 82.2 ± 12.6 s when 10% CN-HP was present. The second room test used a 80 × 80 cm window opening in the front side of the tent. Compared with trials with the regular opening, the mean catch rates of the BGS trap increased in tests with the control and 20% PIC; however, the majority of the released mosquitoes still landed on the volunteer (Fig. 7). In contrast, the proportion of mosquitoes that landed on the volunteer was significantly decreased when 10% CN-HP was dispensed (F = 4.53; df = 2; P = 0.02) while the proportion caught by the BGS trap increased. At the end of the test time, an average of 37.8 ± 6.5% had reached the volunteer while an average of 46.6 ± 5.7% were caught by the BGS trap. This equates to a 44.3% reduction in landings rates and a 150% increase in trap catch rates. The mean landing times of mosquitoes were again significantly delayed during tests of 10% CN-HP (F = 13.21; df = 2; P = 0.00). The first mosquito was caught after an average of 37.3 ± 5.6 s in control tests; compared with an average of 150.8 ± 21.7 s for 10% CN-HP.
Room test experiment 6 with centered window opening. Mean percentages and SE of active test mosquitoes and mosquitoes caught by the BGS trap and volunteer within a testing time of 15 min. Each treatment was tested in 10 repetitions with Ae. aegypti.
Discussion
This is the first published study to describe a laboratory test set-up that allows the evaluation of mosquito repellents and inhibitors under standardized conditions, while providing insight into the correlation between results obtained in olfactometer assays and in larger scaled tests. Our results also provide evidence, that the combination of a suitable spatial repellent or attraction-inhibitor with an attractive trapping system such as the BGS trap can yield a significant reduction of human landing rates (Fig. 8).
Summary of room tests experiments with 10% CN-HP. Mean mosquito landing rates and SE on the volunteer during different treatments: (−) R/(−) BGS: no repellent, no trap; (−) R/(+) BGS: no repellent, trap installed; (+) R/(−) BGS: repellent dispensed in regular tent opening, no trap; (+) R/(+) BGS: repellent dispensed in regular tent opening, trap installed; (+) R/(+) BGS/WO: repellent dispensed in centered window opening, trap installed. Each treatment was tested in 10 repetitions with Ae. aegypti.
Y-Tube Olfactometer Assays.
Olfactometer tests are a quick and efficient way to evaluate the behavioral responses of mosquitoes toward volatile stimuli; however, they sometimes overestimate efficacy because of the restrictions related to a confined volume and short distances from the point source release of the odors. Thus, results obtained from olfactometer assays may not correlate well with field results obtained with the same test chemicals. To address this issue, we screened potentially nonspatial or contact repellents to find out if the chosen experimental set-ups can provide reliable results for (contact) repellents, as well as spatial repellents that function by inhibiting the ability of mosquitoes to detect and find the source of attractive odors. Results from olfactometer assays with 20% PIC supported its action as a contact repellent whereas THY exhibited inhibition of test mosquitoes to find the attractive odor source, even though they were not as potent as effects produced by the release of CN, MP, HP, and CN-HP. Thus, 20% PIC is better suited than THY to serve as a negative control in these spatial efficacy evaluations. The potential of CN as a spatial repellent (better described now as an attraction-inhibitor) has been reported previously (Bernier et al. 2005) and results presented here support those findings. The attraction-inhibitors MP and HP showed a significant reduction of the test mosquitoes' response toward human odors. These inhibitors were reported as superior to CN in recent studies (Bernier at al. 2012). However, the use of MP as a push component is less favorable because it has an unpleasant and obtrusive odor. We decided to focus on HP that also has a distinct but less perceivable smell. The experiments with the formulated CN-HP mixture exhibited the strongest inhibition of Ae. aegypti host-seeking, with an average reduction in the attraction of 96.7%, indicating the importance of including a piperazine-based inhibitor.
Room Tests.
A novel test was used to investigate the efficacy of the candidate materials on a larger scale (in a room) by measuring and comparing BGS trap catch rates in the presence and absence of the experimental treatments. In contrast to the olfactometer assays, CN had a weaker effect on mosquitoes in room tests. The best results were obtained from use of HP, which was as efficient in reducing the BGS trap catches as was the CN-HP combination. Room test experiment 4 was designed to investigate if mosquito host-finding can be decreased via the use of an attraction-inhibitor. Only three treatments were evaluated to reduce bites received by the volunteer. Even though CN-HP held great promise based on results from olfactometer assays, it did not provide adequate protection within room tests with landing rates on the volunteer of nearly 100%. In contrast to olfactometer tests, additional cues like CO2, body heat, and vision contribute to the test mosquito attraction toward the volunteer and these may override the ability of the attraction-inhibitors to block the detection of other host-produced kairomones. The discrepancy between our results from Y-tube and room tests also demonstrates that although olfactometer tests are suited to discriminate between spatial and contact repellent properties (and thereby represent a quick and efficient way to screen a large number of interesting candidate compounds in a short period of time), they do not provide a reliable indication on the magnitude and quality of effects in a larger area. Room test experiment 5 involved a simple push and pull situation to determine if the combination of an attractive trap and an inhibitor leads to a reduction in human-mosquito contact. The push and pull system led to human landing rates that were reduced by 42.4% and trap catches increased fivefold when 10% CN-HP was dispensed and tested in combination with the BGS trap.
Smoke experiments with ammonium chloride mist revealed that the air curtain had greatest density at the top of the tent opening where it was released and that the density was decreased in the lower half. The air curtain also thinned out toward the sides of the opening and toward the floor. Based on these findings, a modification of the set-up was introduced that used a window opening on the front side of the tent to create a defined area with high repellent density. Even though the window opening restricted the ability of mosquitoes to fly into the tent, the majority of the test mosquitoes still reached the volunteer in control experiments. However, use of the 10% CN-HP mixture led to an inversion of the catch ratio between volunteer and BGS trap. For the first time in our experiments, more mosquitoes were found in the trap while <40% reached the volunteer (Fig. 7). An optimization of the volatile release device (see above) might further increase the efficacy of the system. The establishment of push and pull strategies in vector control is a subject of great interest. Recently, new promising research on behavior modifying chemicals has been published on a mosquito odorant co-receptor-agonist that could disrupt olfactory-mediated behaviors, such as host-seeking (Jones et al. 2011). However, results of this work are based on electrophysiological assays and further testing under realistic conditions is required to verify and confirm the promising effects. Turner et al. (2011) identified odors that inhibit CO2-sensitive neurons or evoke a CO2-like activity and their use as more powerful tools in repelling and trapping mosquitoes was suggested. Semifield tests in experimental green houses involving CO2-emitting counter-flow traps revealed that trap catch rates could be decreased by ≈25% when CO2-response-modifying odors were dispensed. These findings, however, were not confirmed in tests with human odors as some of the described agonists and antagonists may have undesirable safety profiles at higher concentrations that could disqualify them for human use. In 2010, semifield tests evaluated the bite-reducing efficacy of the Mosquito Magnet trap in a confined area when used in combination with conventional and commercially available repellents (Kitau et al. 2010). The use of the trap and a skin repellent led to a significant decrease in the human biting rates but further modes of application of the repellent component need to be investigated to reduce the personal effort within the presented approach. Recently, an initial assessment study on the acceptability of a push and pull control strategy which involved common household insecticides in combination with the BGS trap as an outdoor trapping tool was published (Paz–Soldan et al., 2011). Results indicated that the chosen concept could be well accepted by the communities, but it implies the use of insecticides which produce increased resistance in a target mosquito population when used inefficiently or indiscriminantly (Fonseca–Gonzales et al. 2010, Polson et al. 2011) and does not comply with the general idea of push and pull as a nontoxic means of pest control (Cook et al. 2007). In addition, depending on the concentration used, insecticides can intoxicate the target mosquito and cause it to rest or seek shelter instead of being attracted to a trap and get caught, a behavior that was also observed during the semifield tests by Kitau et al. (2010). Our work therefore focused on evaluating non- to low-insecticidal compounds which mediate distance effects without killing or intoxicating the target mosquitoes. In a similar study, repellent plants like wild sage, neem, lemongrass, and West Indian lantana in outdoor plantation were studied for their effect on mosquito house entry in rural tropical areas (Mngongo et al. 2011). When Lantana camara was planted outdoors, up to 83% fewer A. funestus were collected indoors compared with control houses. The project aimed to identify affordable means of mosquito control for developing countries and did not include research on pull components. All these publications indicate that push and pull could be a viable concept for mosquito control; however, there is still a great need for 1) standardized methods to assess and evaluate spatial effects of both repelling and attracting compounds and 2) elaborate techniques to apply single components in the most efficient yet easy way in a natural setting. Even though our experimental design represents a simple and basic approach, our findings from laboratory tests indicate that a push and pull system based on an attractive trapping tool and volatile inhibitors or repellents is capable of reducing human-vector contact in a confined area. Field tests involving experimental huts or tents need to be the next step and will help to investigate if distance effects persist under realistic conditions. Future studies should also include modifications of the repellent dispensing system, to create long-lasting effects as well as the use of multiple traps and various ways of trap arrangement to find out if human landing rates can be further decreased.
