Next Article in Journal
Epstein–Barr Virus LMP1 Induces Soluble PD-L1 in Nasopharyngeal Carcinoma
Next Article in Special Issue
Comparative Study of Two Insulinlike Proteases in Cryptosporidium parvum
Previous Article in Journal
rt269I Type of Hepatitis B Virus (HBV) Polymerase versus rt269L Is More Prone to Mutations within HBV Genome in Chronic Patients Infected with Genotype C2: Evidence from Analysis of Full HBV Genotype C2 Genome
Previous Article in Special Issue
Role of Leishmania infantum in Meningoencephalitis of Unknown Origin in Dogs from a Canine Leishmaniosis Endemic Area
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Microorganisms
Volume 9
Issue 3
10.3390/microorganisms9030602
microorganisms-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Modelling Sand Fly Lutzomyia longipalpis Attraction to Host Odour: Synthetic Sex-Aggregation Pheromone Dominates the Response

by
Renata Retkute
1,2,*,
Erin Dilger
1,3,
James G. C. Hamilton
4,
Matt J. Keeling
1 and
Orin Courtenay
1,3,*
1
The Zeeman Institute for Systems Biology and Infectious Disease Epidemiology Research (SBIDER), The University of Warwick, Coventry CV4 7AL, UK
2
Epidemiology and Modelling Group, Department of Plant Sciences, University of Cambridge, Cambridge CB2 3EA, UK
3
School of Life Sciences, The University of Warwick, Coventry CV4 7AL, UK
4
Division of Biomedical and Life Sciences, Faculty of Health and Medicine, Lancaster University, Lancashire LA1 4YG, UK
*
Authors to whom correspondence should be addressed.
Microorganisms 2021, 9(3), 602; https://doi.org/10.3390/microorganisms9030602
Submission received: 27 February 2021 / Accepted: 12 March 2021 / Published: 15 March 2021
(This article belongs to the Special Issue Zoonotic Pathogens: A One Health Approach)
Browse Figures
Versions Notes

Abstract

:
Zoontic visceral leishmaniasis (ZVL) due to Leishmania infantum is a potentially fatal protozoan parasitic disease of humans and dogs. In the Americas, dogs are the reservoir and the sand fly, Lutzomyia longipalpis, the principal vector. A synthetic version of the male sand fly produced sex-aggregation pheromone attracts both female and male conspecifics to co-located insecticide, reducing both reservoir infection and vector abundance. However the effect of the synthetic pheromone on the vector’s “choice“ of host (human, animal reservoir, or dead-end host) for blood feeding in the presence of the pheromone is less well understood. In this study, we developed a modelling framework to allow us to predict the relative attractiveness of the synthetic pheromone and potential alterations in host choice. Our analysis indicates that the synthetic pheromone can attract 53% (95% CIs: 39%–86%) of host-seeking female Lu. longipalpis and thus it out-competes competing host odours. Importantly, the results suggest that the synthetic pheromone can lure vectors away from humans and dogs, such that when co-located with insecticide, it provides protection against transmission leading to human and canine ZVL.

1. Introduction

Understanding the blood-seeking behaviour of arthropod vectors has relevance to vector control against transmission of public and veterinary health diseases [1,2]. The emerging behavioural adaptations of important mosquito vectors to alter their blood-feeding host preferences, locations and feeding time periods, and thus their ability to evade fatal exposure to indoor insecticide interventions [2,3,4], highlight the need for sustained vector surveillance and development of complimentary vector control strategies.
Insect pheromones that mediate mating, aggregation, and invitation behaviours [5], could play a role in maintaining vector contact with insecticides, and consequently in reducing the human biting index. In the agricultural sector, insect pheromones and other semiochemicals have been widely exploited to monitor and reduce pest populations to protect crop yields [5,6,7]. In contrast, pheromones produced by vectors of public or veterinary health importance have not been widely identified or characterised [8] despite their potential to be used to help reduce infection or disease incidence.
One exception is the sand fly Lutzomyia longipalpis (Diptera: Psychodidae), which throughout the Americas is the predominant vector of Leishmania infantum (Kinetoplastida: Trypanosomatidae), a protozoan parasite that causes human and canine zoonotic visceral leishmaniasis (ZVL) which is usually fatal if not treated [9]. The males of this species produce a sex-aggregation pheromone that mediates attraction of female and male conspecifics to leks for mating [10]. Since leks are usually located on or near to animal hosts, lekking facilitates successful blood feeding by females and thus Le. infantum transmission.
Lu. longipalpis are catholic in their host choice, feeding on a wide range of non-reservoir (“dead-end”) hosts including household animals, livestock and wildlife species, as well as on domestic dogs which are the sole proven reservoir of Le. infantum [11]. Lu. longipalpis are not particularly endophagic or endophilic, and are usually trapped in greatest abundance in animal shelters, to which they show a degree of site loyalty as they have a short dispersal range (<100 m) [12,13,14,15]. Dead-end host populations in the near vicinity of human habitation help to maintain the sand fly populations through provision of blood [16], and consequently are likely to influence the epidemiology of ZVL through diversion of infectious vector bites away from humans and the reservoir (zooprophylaxis), or alternatively, by increasing vector densities as the numbers of attractive animals increase (zoopotentiation) [17]. Evidence for these possible outcomes in leishmaniasis epidemiology are currently contradictory or untested [18].
The recent development of a synthetic copy of the male produced Lu. longipalpis sex-aggregation pheromone placed in controlled-release dispensers, has provided a unique opportunity to test its attractiveness to conspecific males and blood-seeking females under experimental and field conditions [19,20,21,22,23]. The pheromone has been shown to be attractive at least 30 m from it’s source [23], and can attract many times more females and males to the source than controls [19,20,21,22,23]. Furthermore, co-location in close proximity with pyrethroid insecticides sprayed onto chicken roosting sites or onto household compound perimeter walls in control trials in Brazil, demonstrate that this “lure-and-kill” approach can reduce confirmed canine infection incidence and tissue Le. infantum parasite loads by 52–53%, and household vector abundance by 49%–70% [24,25]. A next step is to design and optimise community-wide deployment strategies of this vector control method. However, this requires a good understanding of the potential changes in host choice seeking behaviours in context of competing host odours and variations in demographic and epidemiological conditions.
Mathematical models aid in deciphering important drivers of insect behaviour and predicting changes in epidemiological dynamics when such factors vary. Feeding preferences of sand fly vectors have been modelled explicitely to identify the importance of host defensiveness [26], host irritability [27], host body surface area [28] and host species biomass [29]. However one of the key drivers of vector host-seeking behaviour of Lu. longipalpis and possibly other sand fly species is pheromone-mediated lek formation [12,15].
In this study, we report on the development of a mathematical framework to simulate and predict vector blood-seeking behaviour in the context of deploying a synthetic pheromone attractant, parameterised using data from relevant recent field and laboratory studies. To our knowledge this is the first time that the potential of a synthetic pheromone of an arthropod vector has been applied to control a microorganism of public health importance [20,24,25].

2. Materials and Methods

To address the question of the contribution of different attractive elements to the overall attractiveness of a household, we developed a mechanistic spatial model for sand fly host choice in response to a synthetic conspecific sex-aggregation pheromone, whereby the local distribution of sand flies was constructed in terms of attraction profiles. The attraction profile for a household is made up of multiple attractive (and repellant) elements, including hosts (humans, dogs and chickens) and synthetic pheromone traps (Figure 1a). The behaviour of a sand fly responding to a source of attraction depends on the distance required to travel and the strength of competing stimuli received.
We have defined an attraction profile as:
A S ( d , n , p ) = e x p ( − p 1 d ) p 2 S ( 1 − e x p a ( − p 3 S n ) )
where d is the distance, n is either amount of pheromone or a number of hosts; and p = { p 1 ,   p 2 S , p 3 S }   are parameters defining the shape of attraction profile for an attractor S = { P , H , D , C }   , where pheromone (P), human host (H), dog (D) or chicken (C) indicates a type of attractor. We made the following assumptions: (i) attraction decreases with a distance from a source with an exponential decay defined by a parameter p 1 ; (ii) there is a saturation effect with increasing amount of pheromone or hosts which is defined by parameter p 3 S ; (iii) the height of the attraction profile is equal to p 2 S .
We have parametrised attraction profiles using data on dispersal of Lu. longipalpis [12], host-biting preference [30], and sand fly capture success in synthetic pheromone-baited field traps generated by studies conducted in various locations in Brazil [22,23]. Details of the experiments are given below.
Dataset A. Capture experiments to assess attraction of Lu. longipalpis to the synthetic pheromone [22]. The attraction of individual sand flies to different amounts of pheromone was measured using a series of choice tests. The experiments aimed to test the dose response to the synthetic pheromone relative to chicken-only controls. The quantities of synthetic pheromone tested were 10, 50, 100 or 500 mg and at distances of 5, 10, 15, 20 and 30 m between test and control CDC light traps [22].
Dataset B. Capture–mark–recapture experiments to assess the attraction of Lu. longipalpis to the synthetic pheromone [23]. These experiments were conducted using wild-caught sand flies which were marked with fluorescent powders and released at a specific distance from test and control chicken boxes and collected with a modified (no light) CDC trap baited with the synthetic pheromone (10 mg). The synthetic pheromone was placed in the chicken boxes set at distances of 5, 10, 15, 20 and 30 m from the sand fly release point [23] and control traps (without pheromone) were also set at 5, 10, 15, 20 and 30 m from the sand fly release point and 5, 10, 15, 20 and 30 m from the pheromone-baited trap. All chicken boxes (test and control) contained 1 chicken.
Dataset C. Capture experiment to assess host preference [30]. Fieldwork was conducted to investigate the preference of Lu. longipalpis for dogs, humans and chickens. CDC light traps were set in three domestic locations: one trap in the bedroom of the house, one in the chicken shed and one trap above a wire mesh cage containing a dog, the relative positions rotated. The relative numbers of sand flies captured per night in each trap were counted [30].
Dataset D. Mark–release–recapture experiment to assess dispersal of Lu. longipalpis [12]. Wild-caught sand flies were marked with fluorescent powders and released for recapture at chicken sheds by CDC light trap. The relationship between the percentage of recaptured sand flies at distances (on a natural logarithm scale) was estimated to be proportional to −0.16√d, where d is the distance in meters [12].
A likelihood function for these aggregated data was formulated as a product of binomial distributions for the experimental datasets A and B, and multinomial distribution for experimental dataset C. The probability of success in a single trial was set equal to the probability that a sand fly would be attracted to a source. This probability was calculated by the integration of the convolution between the attraction profile and the dispersal kernel using the MCMC algorithm [31] to estimate the parameters from the aggregated dataset. More details on the likelihood function and fitting procedure are given in Appendix A.
To further understand how the presence of the synthetic pheromone might influence host choice in the heterogeneous landscape, an agent-based mathematical model was designed where a sandfly chooses a household, and a host or pheromone within the household, based on the parametrised attraction profiles. Model simulations were run 1,000 times. The fractions of sand flies attracted to humans, dogs and chickens in each household were calculated running simulations with and without the presence of 10 mg of the synthetic pheromone per household. Variations in the household characteristics were based on the observed geospatial and demographic characteristics of an example rural village (Caldeãro) in Marajó, Pará state, Brazil [30] (Figure 1b). More details on the simulations are given in Appendix B.

3. Results

3.1. Parametrisation of Attraction Profiles

We fitted attraction profiles to the aggregated dataset containing 309 records from datasets A, B and C. The log-likelihood trace plot for the final 1,000 MCMC iterations and posterior distributions of parameters are shown in supplementary Figure S1.
The observed fraction of Lu. longipalpis captured on chickens, and fitted proportions α A   (for the experiment A) and α C   (for the experiment C) are shown in Figure 2.
The model fits indicate that a greater proportion of female Lu. longipalpis would be attracted to a trap baited with the synthetic pheromone relative to a control trap containing a single chicken, where the proportion depends on the amount of pheromone in the trap. Fitted attraction profiles indicate that the number of female Lu. longipalpis attracted to the synthetic pheromone can be increased through addition of more pheromone, but the effect saturates at approximately 100 mg (Figure 2a). In the absence of pheromone traps, a similar response can be seen when the number of chickens is increased: the fraction of Lu. longipalpis attracted to chicken sheds increases and then saturates at approximately 5 chickens per shed (Figure 2b).
Having a mechanistic model is advantageous to investigate hypothetical and real scenarios which otherwise would require expensive and time consuming field studies. We have simulated capture experiments to measure the effect of increasing the amount of pheromone vs increasing the number of chickens. We have assumed that the virtual experimental setup used one pair of chicken sheds set 30 m apart. The first shed contained a chicken and a different amount of pheromone and the second shed contained different numbers of chickens. The simulated fraction of sand flies attracted to the shed with pheromone is shown in Figure 2c. This fraction was between 0.4 and 0.85. It can be seen that chickens were more attractive only for the minimum amount of pheromone (10–20 mg per trap). In the case of 20 mg, the second shed had to have at least 25 chickens to attract more than 50% of the sand flies.

3.2. Effect of the Synthetic Pheromone on Lu. longipalpis Host Preference

The estimated effects on host preference in the heterogeneous spatial and demographic landscape is depicted in Figure 3 showing the predicted relative attraction of female Lu. longipalpis to the most commonly recorded blood-source hosts: chickens, dogs and humans, in the absence (3a) and presence (3b) of the synthetic pheromone. The variability in the fraction of Lu. longipalpis attracted to each source arises from the observed heterogeneity in host demography, i.e., the numbers of humans, dogs and chickens, recorded per household [29,30].
Across all households, in the absence of the synthetic pheromone, a median 6.7% (95% CIs: 3.8%–36.6%) of sand flies are preferentially attracted to humans, 15.1% (95% CIs: 7.8%–84.0%) to dogs, and 81.2% (95% CIs: 54.4%–87.6%) to chickens. In the presence of 10 mg of the synthetic pheromone, the expected median proportion of sand flies attracted to each of the three hosts decreased by approximately half: 3.1% (95% CIs: 2.2%–4.9%) (humans), 8.1% (95% CIs: 4.3%–22.2%) (dogs), and 42.1% (95% CIs: 14.8%–49.1%) (chickens), whereas the attraction to the synthetic pheromone accounted for 53.7% (95% CIs: 39.7%–86.4%) of total Lu. longipalpis (Figure 3b).

4. Discussion

We modelled the localised influence of the synthetic pheromone in treated households based on empirical data showing that it can attract conspecifics from at least 30 m away [23], and that the attraction strength is non-linearly related to the pheromone quantity (dose dependent) [22]. When constructing the attraction profiles, we assumed that attraction decreases exponentially with distance from a source, and a saturation effect is reached as the amount of the synthetic pheromone or numbers of hosts (kairomone quantity) increases [15,22,23]. We introduce the term “attraction profile” to incorporate the synthetic pheromone as a potential “choice” for sand flies, in contrast to malaria models where the term “attraction rate” refers to the blood-meal choice dependence on the propensity of hosts to emit kairomone attractants, and on host accessibility [17]. Here we present an extension to the vector-host interaction model framework to inform interventions that include vector attractants, in this case a synthetic vector pheromone which can be co-located with insecticide as a “lure-and-kill” method against a public health disease. Our proposed mathematical framework explicitly models the quantitative interactions between the synthetic pheromone and key types of host (dead-end, reservoir and humans) within a realistic heterogeneous spatial and demographic dimension. Other influencing factors such as host accessibility [32,33] could easily be encorporated into the model framework.
The current simulations indicate that in the absence of the synthetic pheromone, chickens would be expected to be the preferred host (81.2%), with dogs being second choice (15.1%) and humans the least attractive (6.7%). This is expected as many studies demonstrate that Lu. longipalpis is trapped at greatest densities in animal shelters compared to inside houses [13,33]. In households fitted with the synthetic pheromone, model simulations predict that the synthetic pheromone attracts female Lu. longipapis away from the three potential alternative hosts, and in approximately similar proportions (46.3–53.6%), indicating substantial reductions in the absolute biting rates on humans and the canine reservoir. There is no evidence from community studies that the synthetic pheromone attracts larger numbers of Lu. longipalpis to households which could lead to zoopotentiation [24]. On the contrary, when co-located with insecticide, the pheromone can reduce vector numbers even in neighbouring houses which do not receive the lure-and-kill intervention [25]. Locating the synthetic pheromone and insecticide within household compounds contrasts to that of experimental human odour lures to attract mosquitoes, which are longer range attractants, and therefore best placed further away from human residencies and other mosquito aggregation sites in order to reduce the risk of zoopotentiation [34].
The model outcomes clearly demonstrate that the preference of Lu. longipalpis is skewed towards the synthetic pheromone within the context of competing host odours and the naturally-released male pheromone, but it appears not to alter the relative preference for the three types of hosts. A similar finding is observed from data independent from those used to parameterise the current model, collected elsewhere in Brazil where CDC traps were placed in houses, chicken roosting sites and above tethered dogs, in control and treated households [24]. In that study, the lure-and-kill approach reduced transmission in the canine reservoir by approximately 50% but did not increase household vector abundance [24].
The feasibility of a lure-and-kill strategy for community-wide deployment will depend on economic evaluation as an important consideration for public policy decision making. A key motivation of this study was the need to provide a mathematical framework to predict the likely outcomes on host choice underpinning field trial results, but it also serves as a framework to help optimise alternative interventions with spatial modes of action. The next steps will be to incorporate the predictive mathematical model of vector host-seeking behavior with spatially explicit VL transmission models to evaluate changes in human and canine spatial infection incidence under variable pheromone implementation scenarios and demographic conditions.
In conclusion, our analysis indicates that when the synthetic pheromone (10 mg) is co-located with insecticide, the lure-and-kill approach could dilute mean vector biting rates on humans and the canine reservoir by approximately half.

Supplementary Materials

The following are available online a https://www.mdpi.com/2076-2607/9/3/602/s1, Figure S1: MCMC chain and posterior distribution of parameters. Table S1: Household characteristics.

Author Contributions

Conceptualisation, R.R., O.C. and M.J.K.; methodology, R.R., O.C. and M.J.K.; software, R.R.; validation, R.R.; formal analysis, R.R.; investigation, R.R.; resources, O.C. and E.D.; data curation, O.C., E.D., and R.R.; writing—original draft preparation, R.R., O.C. and J.G.C.H.; writing—review and editing, R.R. and O.C.; visualisation, R.R.; funding acquisition, O.C. and J.G.C.H. All authors have read and agreed to the published version of the manuscript.

Funding

The work was conducted with the continued supported of the Wellcome Trust https://wellcome.ac.uk Strategic Translation Award (WT091689MF) to O.C. and J.G.C.H. The funding body played no role in the design of the study, the collection, analysis, interpretation of the data, or in writing the manuscript or decision to submit the paper for publication. For the purpose of open access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.

Data Availability Statement

The data supporting the conclusions of this article are included within the article. Data are freely available from DOI 10.5281/zenodo.4602849 on request to be used solely within the context of this study, following the ethical agreements and permissions granted from the relevant authorities for the conduct of this study.

Acknowledgments

The authors acknowledge the multiple collaborations that have led to the generation of the referenced data used in this study. We also thank members of the Zeeman Institute, University of Warwick, for helpful discussions during the preparation of this manuscript.

Conflicts of Interest

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Appendix A

The dispersal was evaluated through the mark–release–recapture method [12]. The relationship between the percentage of recaptured sand flies (Iog scale) has been estimated to be proportional to − 0.16 d   , where d is the distance in meters. Consequently, we have set the dispersal kernel function as:
K ( d ) = K 0 e x p ( − 0.16 d )
where
K 0 = ( ∫ 0 ∞ exp ( − 0.16 x ) d x ) − 1   .
A proportion of sand flies attracted to a household h with n S ( t ) number of attractors at time t can be calculated by summing the integrals:
f h ( t )   =   ∑ S ∈ { P , H , D , C } ∫ 0 ∞ K ( x ) A S ( | x h − x | , n S ( t ) , p S ) d x F ( t )   ,
where S is source, {P, H, D, C} stands for pheromone lure, human, dog and chicken; K(x) is the dispersal kernel, attraction profile is given by Equation (1), and xh is coordinate of the household h.
Normalisation is obtained by summing over all households, i.e.,
F ( t )   =   ∑ h ∑ S ∈ { P , H , D , C } ∫ 0 ∞ K ( x ) A S ( | x h − x | , n S ( t ) , p S ) d x
We can also calculate a fraction of sand flies attracted to a particular host in a household h, for example for dogs:
f h D ( t )   =   ∫ 0 ∞ K ( x ) A D ( | x h − x | , n D ( t ) , p S ) F ( t )   .
There are nine parameters defining attraction profiles: { p 1 , p 2 P , p 3 P , p 2 H , p 3 H , p 2 D , p 3 D , p 2 C , p 3 C }   , where the upper indices are for pheromone (P), human hosts (H), dogs (D) or chickens (C), and the lower indices correspond to decays with distance (1), relative height of the profile (2) and saturation with amount (3). We set p 2 P = 1   , i.e., the height of the attraction profile for a maximum amount of pheromone to be equal to one.
Below we describe how the likelihood function for each of the experiment has been formulated.
In the experiment A, there were two chicken sheds positioned at x t e s t and x c o n t r o l sites. In each shed, there was a synthetic pheromone trap equipped with φ i (i = {contr, test}) amount of the synthetic pheromone, where φ i   =   0 corresponds to the control with no synthetic pheromone present. We set up a virtual experiment within a rectangle around a control and test trap and assumed that sand fly can enter at a random point along the perimeter. The proportion of sand flies attracted to the test trap is calculated as:
α j A   =   β ( d t e s t , φ t e s t , x t e s t ) β ( d t e s t , φ t e s t , x t e s t )   +   β ( d c o n t r o l , φ c o n t r o l , x c o n t r o l )   ,
β ( d , φ , y )   =   ∫ 0 d K ( x ) A ( | y   −   x | , φ , n C , p ) d x / d .
where d t e s t is the distance between the test trap and entry point; d c o n t r o l is the distance between a control trap and entry point. As a chicken was placed in each shed to provide a source of host odour, we assume that:
A ( | y   −   x | , φ , n C , p )   =   A P ( | y   −   x | , φ , p )   +   A C ( | y   −   x | , 1 , p ) ,  
i.e., the attraction profile is a sum of attraction profiles for the synthetic pheromone and chickens with n C   =   1 . We run r simulations with randomly positioned entry point along perimeter of the rectangle and set α A   =   1 r ∑ j = 1 r α j A   .
In experiment B, sand fly was captured, colour coded, released and recaptured. We assumed that released sand flies which were not recaptured in the test trap have dispersed elsewhere. Therefore, the proportion of sand flies attracted to the test trap was calculated as
α B   =   ∫ R 2 K ( y ) A ( | x   −   y | , φ , n C , p ) d y ∫ R 2 K ( y ) A ( | x   −   y | , φ , n C , p ) d y   +   ∫ R 2 K ( y ) d y .
In experiment C, the proportions of sand flies attracted to humans, dogs and chickens are calculated as:
α s , i C   =   ( 1 / d s ) ∫ 0 d s K ( x ) A ( | x s   −   x | , n s , p ) d x ∑ s ∈ S ∫ 0 d s K ( x ) A ( | x s   −   x | , n s , p ) d x ,
where S = {H;D;C}, dS is the distance between a trap S and entry point, nS is a number of individuals of type S. We run r simulations with randomly positioned entry point and
α C   =   1 r ∑ j = 1 r α s , i C   .
Finally, we assume that the number of sand flies caught in a trap follows a binomial distribution with success probability equal to α A   for the experiment A, α B   for the experiment B, and multinomial distribution with probabilities α C   for the experiment C. Then the likelihood function is given by:
L ( p )   =   ∏ i = 1 n A B i n o m ( n i t e s t , n i t e s t   +   n i c o n t r o l , α A )   ×   ∏ i = 1 n B B i n o m ( n i r e c a p t u r e , n i r e l e a s e , α A ) ×   ∏ i = 1 n C M u l t i n o m ( n i C , n i D , n i H , n i C   +   n i D   +   n i H , α C ) .  
We used the MCMC algorithm to estimate the parameters from 309 available datasets. We set a number of entry points to r = 10^5. We chose uniform priors U[0; 1] for all parameters. Parameters were updated using an adaptive random walk Metropolis algorithm with proposal distribution given at iteration k [31]:
Q k ( x , . ) = { N ( x , 0.1 2 I m / m )   i f     k   ≤   2 m , ( 1   −   ξ ) N ( x ,   2.38 2 Σ m / m )   +   ξ N ( x , 0.1 2 I m / m ) ,         o t h e r w i s e .  
where m = 8 is a number of parameters to estimate, Σ m is an empirical estimate of the covariance matrix at iteration k, and ξ   =   0.05   .

Appendix B

The spatial and demographic data for a rural setting came from the fieldwork and survey data from Calderao village, Marajó [30]. The data consisted of GPS coordinates of households, and the number of humans, dogs and chickens resident in the household. Households with missing data for human numbers were assigned a median of five closest neighbours, and number of chickens were sampled from the Marajó dataset. The latitude and longitude coordinates of households were converted to Universal Transverse Mercator zone 22 projection and translated so that the left bottom corner of a local coordinate system is (0,0). There were 235 households in the village. Locations of the households are shown in Figure 1b. Table S1 gives statistics on demographic data.
To account for a local impact of synthetic pheromone traps, we apply following steps:
1. We calculate a number of sand flies in each household without pheromone traps V h .
2. For each household with a pheromone lure, we redistributed sand flies to pheromone lure and hosts according to a ratio between attraction profiles:
f s   =   ∫ 0 ∞ K ( x ) A s ( | x s   −   x | , n s , p s ) d x ∑ S ∈ { P , H , D , C } ∫ 0 ∞ K ( x ) A S ( | x S   −   x | , n S , p S ) d x .  
3. Number of sand flies attracted to a host after introduction of the synthetic pheromone is equal:
V h S   =   f S V h   .
We set total number of female sand flies to V h   =   1000   . We calculate sand flies attracted to hosts and pheromones in each household and aggregate these proportions in order to draw a distribution. The boxplots for estimated sand fly preferences for hosts and pheromones in the absence and presence of pheromone traps are shown in Figure 3.

References

  1. Wilson, A.L.; Courtenay, O.; Kelly-Hope, L.A.; Scott, T.W.; Takken, W.; Torr, S.J.; Lindsay, S.W. The importance of vector control for the control and elimination of vector-borne diseases. PLoS Neglected Trop. Dis. 2020, 14, e0007831. [Google Scholar] [CrossRef] [Green Version]
  2. Carrasco, D.; Lefèvre, T.; Moiroux, N.; Pennetier, C.; Chandre, F.; Cohuet, A. Behavioural adaptations of mosquito vectors to insecticide control. Curr. Opin. Insect Sci. 2019, 34, 48–54. [Google Scholar] [CrossRef]
  3. Killeen, G.F.; Chitnis, N. Potential causes and consequences of behavioural resilience and resistance in malaria vector populations: A mathematical modelling analysis. Malar. J. 2014, 13, 97. [Google Scholar] [CrossRef] [Green Version]
  4. Kreppel, K.S.; Viana, M.; Main, B.J.; Johnson, P.C.D.; Govella, N.J.; Lee, Y.; Maliti, D.; Meza, F.C.; Lanzaro, G.C.; Ferguson, H.M. Emergence of behavioural avoidance strategies of malaria vectors in areas of high LLIN coverage in Tanzania. Sci. Rep. 2020, 10, 14527. [Google Scholar] [CrossRef]
  5. Shani, A. Chemical communication agents (pheromones) in integrated pest management. Drug Dev. Res. 2000, 50, 400–405. [Google Scholar] [CrossRef]
  6. Cook, S.M.; Khan, Z.R.; Pickett, J.A. The Use of Push-Pull Strategies in Integrated Pest Management. Annu. Rev. Entomol. 2007, 52, 375–400. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  7. Mansour, R.; Belzunces, L.P.; Suma, P.; Zappalà, L.; Mazzeo, G.; Grissa-Lebdi, K.; Russo, A.; Biondi, A. Vine and citrus mealybug pest control based on synthetic chemicals. A review. Agron. Sustain. Dev. 2018, 38, 37. [Google Scholar] [CrossRef] [Green Version]
  8. Logan, J.G.; Birkett, M.A. Semiochemicals for biting fly control: Their identification and exploitation. Pest Manag. Sci. 2007, 63, 647–657. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  9. Alvar, J.; Vélez, I.D.; Bern, C.; Herrero, M.; Desjeux, P.; Cano, J.; Jannin, J.; den Boer, M.; WHO Leishmaniasis Control Team. Leishmaniasis Worldwide and Global Estimates of Its Incidence. PLoS ONE 2012, 7, e35671. [Google Scholar] [CrossRef]
  10. Bray, D.P.; Hamilton, J.G.C. Host Odor Synergizes Attraction of Virgin Female Lutzomyia longipalpis (Diptera: Psychodidae). J. Med. Entomol. 2007, 44, 779–787. [Google Scholar] [CrossRef] [Green Version]
  11. Quinnell, R.J.; Courtenay, O. Transmission, reservoir hosts and control of zoonotic visceral leishmaniasis. Parasitology 2009, 136, 1915–1934. [Google Scholar] [CrossRef] [PubMed]
  12. Dye, C.; Davies, C.R.; Lainson, R. Communication among phlebotomine sandflies: A field study of domesticated Lutzomyia longipalpis populations in Amazonian Brazil. Anim. Behav. 1991, 42, 183–192. [Google Scholar] [CrossRef]
  13. Morrison, A.C.; Ferro, C.; Morales, A.; Tesh, R.B.; Wilson, M.L. Dispersal of the Sand Fly Lutzomyia longipalpis (Diptera: Psychodidae) at an Endemic Focus of Visceral Leishmaniasis in Colombia. J. Med. Entomol. 1993, 30, 427–435. [Google Scholar] [CrossRef] [PubMed]
  14. Galvis-Ovallos, F.; Casanova, C.; Bergamaschi, D.P.; Galati, E.A.B. A field study of the survival and dispersal pattern of Lutzomyia longipalpis in an endemic area of visceral leishmaniasis in Brazil. PLoS Negl. Trop. Dis. 2018, 12, e0006333. [Google Scholar] [CrossRef] [Green Version]
  15. Kelly, D.W.; Dye, C. Pheromones, kairomones and the aggregation dynamics of the sandflyLutzomyia longipalpis. Anim. Behav. 1997, 53, 721–731. [Google Scholar] [CrossRef] [Green Version]
  16. Alexander, B.; De Carvalho, R.L.; McCallum, H.; Pereira, M.H. Role of the Domestic Chicken (Gallus gallus)in the Epidemiology of Urban Visceral Leishmaniasis in Brazil. Emerg. Infect. Dis. 2002, 8, 1480–1485. [Google Scholar] [CrossRef]
  17. Saul, A. Zooprophylaxis or zoopotentiation: The outcome of introducing animals on vector transmission is highly dependent on the mosquito mortality while searching. Malar. J. 2003, 2, 32. [Google Scholar] [CrossRef] [Green Version]
  18. Bern, C.; Courtenay, O.; Alvar, J. Of Cattle, Sand Flies and Men: A Systematic Review of Risk Factor Analyses for South Asian Visceral Leishmaniasis and Implications for Elimination. PLoS Negl. Trop. Dis. 2010, 4, e599. [Google Scholar] [CrossRef]
  19. Bray, D.P.; Bandi, K.K.; Brazil, R.P.; Oliveira, A.G.; Hamilton, J.G.C. Synthetic Sex Pheromone Attracts the Leishmaniasis Vector Lutzomyia longipalpis (Diptera: Psychodidae) to Traps in the Field. J. Med. Entomol. 2009, 46, 428–434. [Google Scholar] [CrossRef] [Green Version]
  20. Bray, D.P.; Alves, G.B.; Dorval, M.E.; Brazil, R.P.; Hamilton, J.G. Synthetic sex pheromone attracts the leishmaniasis vector Lutzomyia longipalpis to experimental chicken sheds treated with insecticide. Parasit. Vectors 2010, 3, 16. [Google Scholar] [CrossRef] [Green Version]
  21. Bray, D.P.; Carter, V.; Alves, G.B.; Brazil, R.P.; Bandi, K.K.; Hamilton, J.G.C. Synthetic Sex Pheromone in a Long-Lasting Lure Attracts the Visceral Leishmaniasis Vector, Lutzomyia longipalpis, for up to 12 Weeks in Brazil. PLoS Negl. Trop. Dis. 2014, 8, e2723. [Google Scholar] [CrossRef]
  22. Bell, M.J.; Sedda, L.; Gonzalez, M.A.; De Souza, C.F.; Dilger, E.; Brazil, R.P.; Courtenay, O.; Hamilton, J.G.C. Attraction of Lutzomyia longipalpis to synthetic sex-aggregation pheromone: Effect of release rate and proximity of adjacent pheromone sources. PLoS Negl. Trop. Dis. 2018, 12, e0007007. [Google Scholar] [CrossRef] [Green Version]
  23. González, M.A.; Bell, M.; Souza, C.F.; Maciel-De-Freitas, R.; Brazil, R.P.; Courtenay, O.; Hamilton, J.G.C. Synthetic sex-aggregation pheromone of Lutzomyia longipalpis, the South American sand fly vector of Leishmania infantum, attracts males and females over long-distance. PLoS Negl. Trop. Dis. 2020, 14, e0008798. [Google Scholar] [CrossRef]
  24. Courtenay, O.; Dilger, E.; Calvo-Bado, L.A.; Kravar-Garde, L.; Carter, V.; Bell, M.J.; Alves, G.B.; Goncalves, R.; Makhdoomi, M.M.; González, M.A.; et al. Sand fly synthetic sex-aggregation pheromone co-located with insecticide reduces the incidence of infection in the canine reservoir of visceral leishmaniasis: A stratified cluster randomised trial. PLoS Negl. Trop. Dis. 2019, 13, e0007767. [Google Scholar] [CrossRef]
  25. Gonçalves, R.; de Souza, C.F.; Rontani, R.B.; Pereira, A.; Farnes, K.B.; Gorsich, E.E.; Silva, R.A.; Brazil, R.P.; Hamilton, J.G.C.; Courtenay, O. Community deployment of a synthetic pheromone of the sand fly Lutzomyia longipalpis co-located with insecticide reduces vector abundance in treated and neighbouring untreated houses: Implications for control of Leishmania infantum. PLoS Negl. Trop. Dis. 2021, 15, e0009080. [Google Scholar] [CrossRef] [PubMed]
  26. Kelly, D.W. Why are some people bitten more than others? Trends Parasitol. 2001, 17, 578–581. [Google Scholar] [CrossRef]
  27. Zahid, M.H.; Kribs, C.M. Impact of dogs with deltamethrin-impregnated collars on prevalence of visceral leishmaniasis. Infect. Dis. Model. 2020, 5, 235–247. [Google Scholar] [CrossRef]
  28. Smith, T.; Killeen, G.F.; Maire, N.; Dietz, K.; Molineaux, L.; Vounatsou, P.; Tanner, M. Relationship between the entomologic inoculation rate and the force of infection for plasmodium falciparum malaria. Am. J. Trop. Med. Hyg. 2006, 75, 11–18. [Google Scholar] [CrossRef] [Green Version]
  29. Buckingham-Jeffery, E.; Hill, E.M.; Datta, S.; Dilger, E.; Courtenay, O. Spatio-temporal modelling of Leishmania infantum infection among domestic dogs: A simulation study and sensitivity analysis applied to rural Brazil. Parasit. Vectors 2019, 12, 215. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  30. Dilger, E. The Effects of Host-Vector Relationships and Density Dependence on the Epidemiology of Visceral Leishmaniasis. Ph.D Thesis, University of Warwick, Coventry, UK, 2013. Available online: http://webcat.warwick.ac.uk/record=b2704268~S1 (accessed on 27 February 2021).
  31. Roberts, G.O.; Rosenthal, J.S. Examples of Adaptive MCMC. J. Comput. Graph. Stat. 2009, 18, 349–367. [Google Scholar] [CrossRef]
  32. Agrela, I.; Sanchez, E.; Gomez, B.; Feliciangeli, M.D. Feeding Behavior of Lutzomyia pseudolongipalpis (Diptera: Psychodidae), a Putative Vector of Visceral Leishmaniasis in Venezuela. J. Med. Entomol. 2002, 39, 440–445. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  33. Quinnell, R.J.; Dye, C. An experimental study of the peridomestic distribution of Lutzomyia longipalpis (Diptera: Psychodidae). Bull. Entomol. Res. 1994, 84, 379–382. [Google Scholar] [CrossRef]
  34. Okumu, F.O.; Govella, N.J.; Moore, S.J.; Chitnis, N.; Killeen, G.F. Potential Benefits, Limitations and Target Product-Profiles of Odor-Baited Mosquito Traps for Malaria Control in Africa. PLoS ONE 2010, 5, e11573. [Google Scholar] [CrossRef] [PubMed] [Green Version]
Microorganisms 09 00602 g001 550
Figure 1. (a) Conceptual representation of host choice model: the behaviour of a sand fly depends on the distance and the strength of stimulus received from various sources: hosts (humans, dogs and chickens) and synthetic pheromone traps. (b) Location of households [30].
Figure 1. (a) Conceptual representation of host choice model: the behaviour of a sand fly depends on the distance and the strength of stimulus received from various sources: hosts (humans, dogs and chickens) and synthetic pheromone traps. (b) Location of households [30].
Microorganisms 09 00602 g001
Microorganisms 09 00602 g002 550
Figure 2. Proportions of Lu. longipalpis as a function of: (a) amount of pheromone (for the experiment A); (b) number of chickens (for the experiment C); (c) amount of pheromone and number of chickens (simulation results). Measured mean and standard deviation are shown in blue and fitted curves in black in (a) and (b).
Figure 2. Proportions of Lu. longipalpis as a function of: (a) amount of pheromone (for the experiment A); (b) number of chickens (for the experiment C); (c) amount of pheromone and number of chickens (simulation results). Measured mean and standard deviation are shown in blue and fitted curves in black in (a) and (b).
Microorganisms 09 00602 g002
Microorganisms 09 00602 g003 550
Figure 3. Lu. longipalpis estimated preference for hosts and pheromones: (a) in the absence of the synthetic pheromone; (b) in the presence of the synthetic pheromone.
Figure 3. Lu. longipalpis estimated preference for hosts and pheromones: (a) in the absence of the synthetic pheromone; (b) in the presence of the synthetic pheromone.
Microorganisms 09 00602 g003
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Retkute, R.; Dilger, E.; Hamilton, J.G.C.; Keeling, M.J.; Courtenay, O. Modelling Sand Fly Lutzomyia longipalpis Attraction to Host Odour: Synthetic Sex-Aggregation Pheromone Dominates the Response. Microorganisms 2021, 9, 602. https://doi.org/10.3390/microorganisms9030602

AMA Style

Retkute R, Dilger E, Hamilton JGC, Keeling MJ, Courtenay O. Modelling Sand Fly Lutzomyia longipalpis Attraction to Host Odour: Synthetic Sex-Aggregation Pheromone Dominates the Response. Microorganisms. 2021; 9(3):602. https://doi.org/10.3390/microorganisms9030602

Chicago/Turabian Style

Retkute, Renata, Erin Dilger, James G. C. Hamilton, Matt J. Keeling, and Orin Courtenay. 2021. "Modelling Sand Fly Lutzomyia longipalpis Attraction to Host Odour: Synthetic Sex-Aggregation Pheromone Dominates the Response" Microorganisms 9, no. 3: 602. https://doi.org/10.3390/microorganisms9030602

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop