Background
Anopheles darlingi is considered the most effective malaria vector in the Neotropical region [
1,
2] and is responsible for most malaria transmission where it is found, including areas of high deforestation [
3,
4]. This species efficiently transmits
Plasmodium falciparum,
Plasmodium vivax VK210, and
P. vivax VK247 across Latin America [
5‐
9], and its introduction to the northern Amazon has been linked to the considerable increase of
P. falciparum cases in Peru in the 1990s [
10,
11]. In Peru,
An. darlingi is the dominant malaria vector in the Loreto Department, which is the most affected malaria endemic region in the country.
Anopheles darlingi is present in both rural villages in the vicinity of Iquitos (15–20 km west-southwest) where it comprises 67–99 % of anophelines collected [
12], and riverside camps along the Mazan river (40–50 km northeast of Iquitos) where 99 % of anophelines collected were identified as
An. darlingi [
13]. Interestingly,
An. darlingi biting rates, and entomological inoculation rates (EIRs) differed considerably among these locations and are substantially higher at riverine sites [
13], possibly linked to the observed malaria transmission heterogeneity in the Peruvian Amazon. Therefore, knowledge of the biology, behaviour, and vectorial capacity of this vector is crucial for understanding malaria transmission dynamics in the Amazon region, as well as for developing and implementing effective malaria vector-targeted control strategies. Laboratory colonies of
An. darlingi reared up to the F9 generation have been reported before, producing less than 3,000 mosquitoes per generation [
14,
15] and without description of long-term continuity. The lack of a highly-productive and well-established
An. darlingi laboratory colony that can continuously provide large numbers of mosquitoes has limited these critical studies.
Previous reports regarding the establishment of a laboratory colony of
An. darlingi [
16‐
19] did not explain the main factors triggering sexual behavior and regulating colony productivity in this species. This historically limited replication of the methodology used and the establishment of new laboratory colonies. For example, low sexual activity resulting in limited oviposition in F
2 adult females made establishment of a colony in São Paulo unsuccessful [
19]. Recently, a method developed for
Anopheles pseudopunctipennis was used for colonization of
An. darlingi in the laboratory [
20], yet very low densities of laboratory-reared adults across six generations (1,618 adults per generation on average) were reported. Moreover, critical characterization of sexual behaviour or parameters demonstrating adaptation of F
1 to laboratory conditions and colony stability has not been reported to date [
16‐
20]. This is an area with limited to no evidence. There is only one report from an
An. darlingi mark-recapture study describing that sexual activity in field populations occurs in the evening shortly after the sunset and near human dwellings, yet no male swarms were observed despite the finding of inseminated recaptured females [
20].
Thus, the main factors stimulating
An. darlingi sexual behaviour and oviposition in the field and under laboratory conditions still remain unknown [
14,
21]. Limited success of copulation induction in artificial environments such as an insectary, where adults are generally confined to rearing cages [
19,
22,
23], remains the major blockage to colony establishment in
An. darlingi. Forced copulation technique was explored in
An. darlingi colony establishment, [
24,
25]. Not only is this very difficult and time-consuming [
14] but was also unsuccessful for
An. darlingi with a very low percentage of F
1 females (1 %) laying eggs (Escobedo and Huayanay, personal observation). The inability to induce copulation, in addition to the lack of an established method for mass larvae rearing, considerably impeded the establishment of a long-standing, highly productive
An. darlingi laboratory colony.
Therefore, a natural copulation induction technique that has effectively worked for colonization of
An. pseudopunctipennis and other anophelines including
An. darlingi in Mexico although only until the F
9 generation [
14,
22], was evaluated, adapted and optimized. Also, conditions favoring egg-laying in gravid females were standardized and larvae rearing was improved by providing a diet that meets nutritional needs, has appropriate particle size, and is easily digestible [
14,
22]. This combined approach has allowed to develop an effective method to establish and mass-rear
An. darlingi under insectary conditions. The methodology used is described in all its complexity and detail to contribute to the development of other laboratory colonies across the Americas and for potential replication in other
Anopheles species.
Methods
Study design
The F1 offspring of wild-caught An. darlingi females were used to evaluate and optimize techniques for natural copulation stimulation, oviposition induction, and larvae rearing. Sexual behaviour was examined by direct observation of mating across 10 generations (F1-F10), and video recording of adult behaviour during mating in F1 and F14 individuals. Successful mating was assessed by estimating the rate of inseminated females via spermatheca dissections across 10 generations (F1-F10). Free-mating was confirmed by comparing the rate of insemination and egg-laying between stimulated and non-stimulated F9 and F10 adults. Copulation induction was deemed unnecessary from the F11 generation onwards.
Anopheles darlingi field collection and production of F1 adults
Adult females were collected in February 2013 (rainy season) in the community of Zungarococha (03º49′32.40″S, 73º21′00.08″W), 18 km southwest of Iquitos, Peru. Active transmission of
P. falciparum and
P. vivax historically occurs in this area [
26], where
An. darlingi has been incriminated as the main vector [
12]. Adult females were captured hourly from 18:00 to 22:00 using protected human landing collections [
25].
Anopheles darlingi females collected were placed in carton cups (250 ml) covered with nylon mesh, provided 10 % sugar solution and taken to the NAMRU-6 insectary in Iquitos. Here, cow, or chicken blood was offered through glass membrane feeders (3.8 cm outer diameter) to induce oviposition and obtain F
1 adults following previously described procedures [
21,
27]. Briefly, one wing of each blood-engorged female was cut and then females were placed individually in vials for egg laying. Eggs were inundated with water within 24 h and placed in plastic trays (26.5 × 16.5 cm). Larvae were transferred into plastic trays containing a mix of water from natural breeding sites and filtered water and fed a mixture of wheat flour and fish-meal with quantities increasing with larval development from 0.14 mg/larva (first instar) to 0.5 mg/larva (fourth instar). Pupae were transferred into plastic containers with water (200 pupae/container) and placed in screen cages for adult eclosion. F
1 adults were used in natural copulation induction assays and for optimization of oviposition and larvae rearing. Morphological species identification of wild-caught and F
1 adult
Anopheles darlingi was conducted using keys for Neotropical
Anopheles [
28,
29]. Molecular verification of F
10
An. darlingi adults was confirmed using mitochondrial
cytochrome c oxidase I (
COI) gene sequences, following the standard protocol of the Mosquito Barcoding Initiative [
30].
Standardization of the natural copulation induction technique
The natural copulation induction technique developed for
An. pseudopunctipennis [
22] was initially evaluated against laboratory reared
An. darlingi F
1 adults obtained in February 2013. As the appropriate/critical space for copulation, sugar source finding, and mosquito density were unknown, two sizes of screened cages, medium (46 × 46 × 46 cm) and large (61 × 61 × 61 cm) were evaluated simultaneously following modifications on documented environmental conditions known to trigger natural copulation in
An. pseudopunctipennis [
22]. Briefly, 2–5 day old
An. darlingi F
1 males and females were placed in the two sizes of screened cages: 1,500 adults in the medium cage (1:1 females to males) and 2,400 adults in the large cage (1:1 females to males), and fed
ad libitum with honey-water solution (10 %). Screened cages were kept in a room with controlled temperature that was maintained at 30 ± 1 ºC during the day (07:00 to 19:00) and at 25 ± 1 ºC during the night (19:00 to 07:00) and under a 12:12 light to dark photoperiod. Relative humidity was not controlled and ranged from 63 to 80 % following natural conditions in the city of Iquitos. F
1 adult mosquitoes were exposed to a beam of a Light-Emitting Diode (LED) projected with a flashlight of 1.5 W placed 50 cm away from the cage. Three artificial 15-min periods of light exposure interjected by 5-min dark intervals were conducted at dusk for five consecutive days in a bid to increase sexual arousal and trigger natural copulation. Light exposure was carried out in complete silence to avoid disruption of mosquito sexual behaviour. Three days after the first copulation induction, F
1 female mosquitoes were offered commercially purchased animal blood (either cow or chicken blood) via membrane feeders for 30–40 min. Blood meals were offered every two days for a period of 15 days from 1900 to 2000 h (recently induced females) and from 0900 to1030 h (ovipositing females); chicken blood was given twice during the first gonotrophic cycle (2–3 days) and either chicken or cow blood was provided at each subsequent feeding event. Fecundity (number of eggs laid per female) between females from cages that received only one type of blood versus those that received two types of blood was compared. In addition, wing length (mm) of a sample of females (10) per generation (F
1–F
10) was measured with a stereoscope (MEIJI EMZ-13, Saitama, Japan) at 30× magnification to examine changes in body size across generations, which could be linked to fecundity [
31,
32]. These standardized copulation induction conditions were used to rear
An. darlingi up to the F
10 generation but using large cages only based on results from the cage-size evaluation.
Assessment of An. darlingi sexual behavior and natural copulation under laboratory conditions
Direct observation of natural copulation in F1 mosquitoes was conducted in both medium and large cages during all three light stimulation periods for seven consecutive days. Mating pairs (copulations) were counted as mating pairs who encountered at flight and fell on the cage floor. Results are presented as total numbers of copulations/day/cage size as well as number of copulations/100 females, to account for the differences in total numbers of females placed in the medium and large cages. In subsequent generations (up to F10), the total number of copulations/female/cage was recorded but for only five days. Sexual behaviour of F1 and F14 males and females was recorded with a Nikon® D5100 video camera. The video was reviewed frame by frame in iPhoto’11 version 9.3 on a 1600 MHz iBook computer with an 11-in. video screen (Apple Computers, Cupertino, CA). Copulation time was recorded and sexual behaviour traits noted.
Insemination rates were estimated by dissecting the spermathecae of a representative sample of adult females collected from each cage across 9 generations (F
2–F
10), following the WHO standard methods [
25]. Fallen females were collected daily for a period of 5–17 days, starting one day after the first day of copulation induction.
Apparent copulation in the absence of stimulation was first observed in the F8 generation. Direct observation of copulation in adults not stimulated by light and thermoperiod was conducted between 18:30 and 19:30 for the F9 and F10 generations. Free-mating behaviour was evaluated by comparing insemination rates per cage and egg-laying rates per female in those copulation-induced females versus non-stimulated controls. Insemination rates were determined as described above; egg-laying rates were estimated by dividing the total number of eggs laid per cage by the total number of females per cage.
Optimization of An. darlingi oviposition and larvae rearing
Preliminary observations of oviposition preference of blood-fed F1 adult females were conducted in the NAMRU-6 Iquitos insectary, comparing light-coloured versus dark-coloured oviposition surfaces. Four white plastic trays (32 × 18 × 5 cm) and four additional identical trays covered with black plastic were filled with filtered water and placed into cages containing 2,000 F1 adults. Female attraction to each colored tray and oviposition rates were recorded. It was subsequently evaluated whether the addition of native aquatic plants (Ceratopteris pteridoides) commonly found in breeding sites would increase oviposition. Environmental conditions were identical for oviposition and natural copulation induction as eggs were collected in the same environment where cages with adults were kept. No honey solution was provided during the oviposition period (15 days). F1 eggs (c. 500) were transferred into larger plastic trays (37 × 25.5 × 5.5 cm) containing 1 - 1.5 L of filtered, dechlorinated water by directly pouring water from egg-collecting trays. These trays were kept at 28 ± 1 ºC, 66 ± 1 RH% and at a 12:12 light to dark photoperiod, conditions favourable for larval eclosion. In subsequent generations, eggs were collected on strips of filter paper (2 cm wide) lining tray edges to prevent egg damage.
Approximately 500 F
1 first-instar larvae were reared in plastic trays (37 × 25.5 × 5.5 cm) until development of third-instar larvae. Third-instar larvae were divided into two trays so that a mean density of 250 larvae per tray was maintained until development of pupae. Larvae were fed commercially available rodent food (Laboratory Rodent Diet 5001, LabDiet®, St. Louis, MO) containing 23.0 % protein, 4.5 % fat, and 6.0 % crude fiber. Optimal feeding amount and frequency of dry rodent food fed were assessed and optimized for each larval instar as previously described [
14,
22].
When larval mortality was higher than average, microscopic examination (40×) of water samples was performed from trays to determine the presence of bacteria, fungi and protists. Selective, and differential media (MacConkey agar, xylose lysine deoxycholate agar, and thiosulfate citrate bile salts sucrose agar) were used to differentiate bacteria.
Standardized oviposition and larvae rearing procedures have been used to rear An. darlingi up to the F26 generation. The colony had to be relocated to another NAMRU-6 facility in Iquitos while rearing the F14 generation, adaptation to laboratory rearing facilitated this process.
Research ethics
Human landing collections were conducted following NAMRU-6′s security protocol, including explanation of risks, safety, healthcare, prophylaxis, and in case of disease, the appropriate care. Mosquito blood feeding using membrane feeders was conducted following the standard operating procedures of NAMRU-6.
Statistical analyses
Reductions in larval mortality across generations were evaluated with Spearman’s Rho non-parametric correlation tests. Changes in egg, larvae and pupae productivity, and reductions in mortality associated with natural copulation were tested using Wilcoxon rank-sum (Mann–Whitney) non-parametric tests. Independent two sample t-tests assuming unequal variances (separate-variances t-tests) were performed for mean comparison of insemination rates and number of eggs laid by stimulated versus not-stimulated females; number of copulations/100 females in large versus medium cages; and number of eggs laid on black trays versus white trays. A Least Significant Difference (LSD) test was performed for mean comparison of female wing length. All analyses were conducted with Stata v13.0 (StataCorp, College Station, TX, 2013) except the LSD test that was conducted with SAS 9.4 (SAS Institute Inc., Cary, NC, 2002–2013). P-values < 0.05 were considered statistically significant.
Discussion
Herein it is reported the successful establishment of the first free-mating, highly productive, and long-lasting
An. darlingi colony by: a) inducing natural copulation through a combination of optimal thermo-period and light stimulation; and b) optimizing egg oviposition and larval rearing parameters. Despite the relevance of
An. darlingi for malaria transmission in the Amazonian region [
10,
11], many aspects of its physiology, behavior, ecology, genetics, and interaction with
Plasmodium spp. are poorly understood due to the lack of a laboratory colony that can serve as a source of large numbers of individuals [
35,
36]. This large and highly productive colony is a unique, valuable resource for developing, and evaluating effective vector-based strategies against malaria transmission in the Amazon. Extensive methodological details are provided to the scientific community to encourage the reproduction of the results, establishment of additional colonies, and extrapolation of the described methods to critical
Anopheles vectors in the Amazon and in other regions of the world.
Historically the establishment of a laboratory colony of
An. darlingi proved difficult, and this is the first report of a free-mating, autonomous colony of this important vector species. A recent study showed successful colonization of
An. darlingi and
P. vivax infection of colonized mosquitoes from F
3 to F
6 generations [
15], yet no evidence of free-mating or a description of sexual behaviour was reported. Low sexual activity observed in adults confined to screened cages under insectary conditions has been the main cited obstacle in establishing an
An. darlingi colony [
23]. In this study it is shown that natural copulation occurs consistently and effectively in the laboratory with four-fold higher copulation rates in larger cages (61 × 61 × 61 cm). Additionally, external diurnal stimuli, was initially required to trigger sexual behaviour. This involved increased temperatures during the day and lower temperatures during twilight, in addition to exposure to a dim light beam at dusk, as to resemble field conditions. Light beams at dusk may resemble natural sunbeams decreasing at this time of the day, thus cueing mosquitoes that the sun is setting and potentially triggering their mating behaviour. These conditions sparked sexual activity in cages immediately after turning the lights off. In previous studies of
An. darlingi reproductive behaviour in the NAMRU-6 Insectary, a very small percentage of caged adult females were inseminated in the absence of any type of stimulus or when applying a light beam for 30-min during twilight [
14]. In this study, a specific thermoperiod regime coupled to exposure to a LED beam resulted in a considerably higher mean insemination rate (55 %). This is in line with field studies where virgin females were released in the field at sunset, and when recaptured two hours later, 60 % had been inseminated [
20]. Sexual activity increased across generations from 2.6 (F
1) to 40 (F
14) matings per minute. Light stimulation was not needed after the F
10 generation indicating the development and selection of a stenogamic colony.
It was also found that
An. darlingi copulation rates were higher in 2 to 8 day-old mosquitoes, suggesting that the sexual stimulation method should be applied to that specific age range. Older females may not respond as effectively. Observations in colonized
An. darlingi agree with those recorded in
An. gambiae, for which the optimal age for male swarming in the field and female copulation success in the laboratory was found to be 4–8 days [
37]. In the field,
An. darlingi males probably fly at low altitudes, close to the vegetation, and remaining near breeding sites waiting to copulate, and even some may travel to food sources following females [
20]. In this study, and in line with the hypothesized low flying altitude in the field, males were observed flying 5–40 cm above the cage floor, and forming a pseudo-swarm, and apparently attracting females. Adults copulated while flying and fell to the cage floor with most mating pairs separating in less than second. All of these behaviors are similar to that of Neotropical
Anopheles reared under laboratory conditions [
22,
38]. Also, some apparent male competition for females was observed in this study, with two males attempting copulation simultaneously with the same female and all three falling together on the cage floor, which could have interfere with copulation. Further examination is needed to assess if this behaviour had an effect on copulation success.
Previously, an
An. darlingi population from Lacandon Forest in Chiapas, Mexico was colonized by natural copulation induction and reared for nine generations. However, the colony collapsed after females stopped ovipositing in white containers placed inside cages [
14]. In this study, this problem was prevented by using highly-preferred dark oviposition trays that perhaps better resemble natural oviposition sites. Native aquatic plants in oviposition trays served as resting places for ovipositing adults and larvae, but were removed from the rearing protocol as they appeared to be a water contamination source.
In addition to optimizing oviposition, an optimal larval feeding regime was established using commercially available rodent food. This food appears to meet the nutritional needs of the immature
An. darlingi larvae in the laboratory, and has been tested in other
Anopheles spp. effectively [
22]. Other commercial animal foods (fish, dog, or monkey) appear to be not as suitable to rear
An. darlingi larvae possibly because their high fat content could lead to the formation of an oily/greasy layer on top of the water, which seems to kill larvae and promote the development of pathogenic microorganisms.
The establishment of autonomous neotropical anopheline colonies under insectary conditions has had variable success. Some anopheline species are considered relatively easy to colonize, such as
Anopheles albimanus that does not require copulation induction [
38]. In contrast,
An. pseudopunctipennis requires artificial induction of mating [
22]. In addition, the number of generations needed to develop an autonomous colony also varies by species and population origin. For example, the Tapachula, and Abasolo Mexican laboratory strains of
An. pseudopunctipennis required 5 and 12 generations to select a free-mating population, respectively [
22]. A stenogamous colony of
Anopheles albitarsis from Brazil was obtained in six generations [
39] while free-mating in
Anopheles aquasalis was recorded in the F
2 generation [
40]. Finally, the previously attempted Mexican
An. darlingi colony collapsed at the F
9 generation before becoming autonomous [
14], yet the Peruvian
An. darlingi colony became autonomous after nine generations. Therefore, it is suggested that free-mating in colonized
An. darlingi can only be demonstrated after rearing at least nine generations.
The fact that this study reports the first establishment of a stenogamic laboratory colony of
An. darlingi despite many previous attempts [
14,
16,
19] proves that colonizing and rearing this species is very challenging. The
An. darlingi colonization approach used involved three critical steps: 1) employment of a very rigorous yet replicable procedure to trigger sexual behavior leading to successful insemination, 2) optimization of proper conditions for consistent oviposition, and 3) standardization of an adequate immature feeding and rearing regime. This approach produced a practical method to establish
An. darlingi colonies and represents an alternative to the time-consuming and unpractical forced-mating method. The overall approach can be easily adapted and replicated for the establishment of other
An. darlingi colonies, and could be useful to establish colonies of other critical malaria vectors worldwide.
Unlimited access to
An. darlingi colony material will allow us to examine and understand unknown biological aspects of this important malaria vector including longevity, gonotrophic cycle length, oviposition behaviour, fecundity and so forth, as well as behavioural, and genetic traits that can be exploited for the development of novel and effective vector control approaches. It is also important to better understand male traits that determine copulation success, such as body size, nutritional status, and any associated genetic components, to facilitate future release of genetically altered males refractory to
Plasmodium spp. Additionally, the interaction of
An. darlingi with
Plasmodium spp. parasites under controlled laboratory conditions and the performance of novel drug therapies targeting parasite development in the mosquito can now be better examined. Preliminary
P. vivax infection studies via membrane feeding assays (4 successful experimental infections) conducted in colonized
An. darlingi adult females from the F
19–F
21 generations have shown that these colonized mosquitoes are susceptible to
P. vivax infection, with number of oocysts per mosquito ranging from 2 to 197 (mean ± SD = 38.6 ± 41.5), and number of sporozoites per mosquito ranging from 152 to 61,867 (mean ± SD = 8,141.4 ± 9,489.7). These values are comparable with those reported for colonized
An. darlingi (F
1–F
6) experimentally infected with
P. vivax, with oocyst number per mosquito ranging from 1 to 57 and sporozoite number per mosquito ranging from 150 to 7,380 [
15]. These preliminary results will be confirmed by additional
P. vivax experimental infections of colonized
An. darlingi from the F
22–F
26 generations. Overall, information gathered from physiological, behavioural, and experimental infection studies conducted in colonized
An. darlingi will be useful in developing strategies for comprehensive, integrated malaria vector control.
Acknowledgements
We would like to thank the Ministerio de Agricultura y Riego de Perú, Dirección General Forestal y de Fauna Silvestre, and the Dirección de Salud, Gobierno Regional de Loreto for permission to conduct these studies: Anopheles darlingi collections in Loreto were conducted under the auspices of Resolución Directoral No. 0406-2013-MINAGRI-DGFFS/DGEFFS; An. darlingi molecular identification was conducted under the auspices of Contrato de Acceso Marco a Recursos Geneticos No. 0017-2014-MINAGRI-DGFFS/DGEFFS. We would like to thank Fanny Castro, Geidin Chavez, Hugo Jaba, Luz Romero, Miguel Vásquez, and Alex Vásquez for technical support. Additionally, we would like to thank Dr. Craig Stoops for providing helpful comments on the manuscript. Financial support was provided by the Global Emerging Infections Surveillance and Response System (AFHSC/GEIS) of the U.S. Department of Defense sustainment funding award (FMS, GMV), the Consejo Nacional de Ciencia y Tecnología (CONACYT) México, grant CB2008-105806-M (CVT), and by the training grant 2D43 TW007393-06 awarded to NAMRU-6 by the Fogarty International Center of the U.S. National Institutes of Health. YML was supported by the National Research Council (NRC) Research Associate Program.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
CVT designed, supervised, and performed the experiments, interpreted data, and drafted the manuscript. GMV designed and supervised the experiments, analyzed and interpreted data, and drafted the manuscript. VLS organized field operations to collect wild mosquito populations, identified anophelines collected during the study, supervised colony rearing, and participated in manuscript revision. KEV and AHR conducted mosquito rearing, performed natural copulation induction experiments, entered data, and participated in manuscript revision. YML confirmed the molecular identification of the An. darlingi colony and participated in manuscript revision. CFM participated in study design, data interpretation, and manuscript revision. AGL participated in data analysis, data interpretation, and manuscript revision. FMS participated in study design and manuscript revision. All authors read and approved the final manuscript.