Background
Malaria remains a major public health problem throughout the tropics [
1] and is transmitted by
Anopheles mosquitoes. Vector control is an essential element of malaria control programmes. Throughout sub-Saharan Africa, indoor residual spraying (IRS) and insecticide treated bed-nets (ITN) have been successful vector control tactics because the primary vector species,
Anopheles gambiae feed almost exclusively on people sleeping in their houses at night and rest inside the house after blood-feeding [
2‐
4]. Other important malaria vector species in Africa (e.g.,
Anopheles arabiensis), Asia (e.g.,
Anopheles stephensi) and the neotropics (e.g.,
Anopheles albimanus) are exophagic (i.e., prefer to bite outdoors), exophilic (i.e., prefer to rest outdoors), and/or zoophagic (i.e., as likely to feed on non-humans as humans) [
5]. Effective control of
Anopheles species with these behavioural characteristics requires alternative tactics. The primary non-human blood sources for zoophagic malaria vectors are peridomestic livestock (cattle, goats). Livestock represent mosquito blood sources that humans can manage and control. It is logical that any comprehensive strategy against zoophagic malaria vectors should include some sort of livestock management component. One tactic to control zoophagic malaria vectors is to treat livestock with ivermectin.
Ivermectin is a lipophilic drug that belongs to the avermectin class of macrocyclic lactone compounds. Ivermectin acts as an endectocide (i.e., kills both endoparasites and ectoparasites). Ivermectin binds to and activates the glutamate-gated chloride channels of nerve and muscle cells of a wide variety of nematode and arthropod species, causing uncontrolled influx of chloride ions into the cells and leading to paralysis and death of the organism [
6]. Ivermectin has a high safety profile in humans and livestock [
7] and is one of the most commonly-used anti-helminthic drug in the livestock industry to control intestinal nematodes.
The large-scale administration of oral ivermectin to people has become a cornerstone in the global eradication campaign against human onchocerciasis and lymphatic filariasis [
8,
9]. The recommended dose of ivermectin used to treat humans for filarial infection [i.e., 150 mg/kg body weight (BW)] yields peak plasma concentrations of ca. 40–45 ng/ml [
10]. At these plasma concentrations, ivermectin can significantly reduce the survival of
Anopheles mosquitoes that ingest treated blood.
Anopheles species shown to be susceptible to ivermectin at these concentrations include major vectors from Africa (
An. gambiae [
11‐
13],
An. arabiensis [
14]), Southeast Asia (
Anopheles campestris,
Anopheles dirus,
Anopheles minimus,
Anopheles swandwongporni [
15]) and Latin America (
Anopheles aquasalis [
16],
Anopheles darlingi [
17]). Pilot field trials in western Africa indicate that the mass drug administration of ivermectin against onchocerciasis and lymphatic filariasis can simultaneously reduce survival of
An. gambiae and local malaria transmission [
18,
19].
Similarly, ivermectin treatment of livestock could theoretically reduce vector abundance and lower malaria transmission [
20,
21]. At the same time, treatment of herds may also reduce tick and intestinal worm burdens, leading to increases in livestock weight gain, milk production, and ultimately increases in the health and economic prosperity of livestock owners. However, it is currently unknown if all species of
Anopheles vectors are equally susceptible to ivermectin. This study compared the relative toxicity of ivermectin to two known malaria vectors:
Anopheles (Nyssorhynchus) albimanus, a major vector along the coastal regions of Mexico, Central America, the Caribbean, and northern South America [
22], and
Anopheles (Cellia) stephensi, an important malaria vector in southern and western Asia [
23]. Both species are highly zoophagic, exophilic and exophagic, making them appropriate species to target with a strategy involving ivermectin treatment of livestock.
Methods
Mosquitoes
Laboratory colonies of An. albimanus STECL strain and An. stephensi STE2 strain were obtained as eggs through BEI Resources (Manassas, VA USA). Mosquitoes were reared in the University of North Dakota insectary at a photoperiod of 12-h light:12-h dark and a temperature of 26 °C. Eggs were hatched in trays of dechlorinated water. Larvae were fed fish food (Tetra Pond Sticks, Tetra, Blacksburg, VA USA) ground to a fine powder in an electric coffee grinder. Pupae were placed into ca. 28 l wire mesh cages to emerge. Adults were given access to water and a sugar source. Female mosquitoes 3–7 days old for use in toxicity studies were aspirated from rearing cages into smaller (ca. 0.5 l), cylindrical cardboard containers with mesh tops at a density of 15–40 mosquitoes per cage.
Membrane feeding
For each membrane feeding trial, age-matched cohorts of An. albimanus and An. stephensi were blood fed simultaneously on identical ivermectin preparations. For each feeding trial, stock solutions were prepared fresh by diluting powdered technical grade ivermectin (Sigma-Aldrich, St. Louis MO, USA) into dimethyl sulfoxide (DMSO) at a concentration of 2 mg/ml. The stock solution was diluted to starting concentrations in phosphate-buffered saline (PBS). From that, serial dilutions were made in 1.5 ml polypropylene microfuge tubes containing whole bovine blood with sodium heparin (10 U/ml, Pel-Freez Biologicals, Rodgers AR, USA) for a total volume of 1 ml. A control group (i.e., 10 µl PBS added to 990 µl blood) was included with each trial. Prior to feeding, the tubes were inverted a minimum of three times to mix the solutions and kept in warm water. Natural sausage casing (Dewied International, San Antonio TX, USA) was rinsed thoroughly to remove salt preservative and cut to fit across the bottom of glass membrane feeders. Feeders were connected to one another with rubber tubing and heated water (ca. 37 °C) was circulated to warm the feeders. Membrane feeders were placed one per cage, and the blood-ivermectin mixtures were pipetted into the feeders. Mosquitoes were allowed to feed for ca. 30 min. Unfed mosquitoes were removed. Engorged mosquitoes were maintained in the insectary and provided with cotton soaked in a 10% sucrose solution. Cages were checked every day. Dead mosquitoes were counted and removed. For most trials, the number of surviving mosquitoes was counted at 4 days. In one trial, survival counts were extended to 5 days to determine if mortality had stabilized. By day 5, the rate of mortality stabilized and became asymptotic. Thus, 4 days was selected as an appropriate time on which to base LC50 determinations for oral toxicity. After preliminary trials to determine an appropriate range of concentrations, feedings were replicated five times using 4–5 ivermectin concentrations, plus a control, for each feeding trial.
Intrathoracic injection
Mosquitoes were immobilized by chilling (− 20 °C for ca. 20 s), then quickly transferred to filter paper onto a chill table (BioQuip, Rancho Dominguez, CA, USA). To perform injections, a semi-automated micro-injector (TriTech Research, Los Angeles CA, USA) was fitted with fine-tipped glass needles made with a micropipette puller (World Precision, Sarasota FL, USA). A stock solution was prepared as described above and from that, serial dilutions were made in 1.5 ml polypropylene microfuge tubes containing an insect salt solution [
24]. Mosquitoes were injected with 0.4 µl of solution into the thorax, near the base of the wing. Mosquitoes were then placed into appropriately labeled recovery containers. Injected mosquitoes were provided 10% sucrose solution and checked daily for 3 days. Dead mosquitoes were counted and removed over the course of 3 days, at which time the surviving mosquitoes were counted. In one trial, survival counts were extended to 4 days to determine if mortality had stabilized. By day 4, the rate of mortality stabilized and became asymptotic. Thus, 3 days was selected as an appropriate time on which to base LC
50 determinations for parenteral toxicity. After preliminary trials to determine an appropriate range of concentrations, injections were replicated three times for
An. stephensi and four times for
An.
albimanus, using 4–5 ivermectin concentrations plus a control at each injection trial.
Fecundity
Mosquitoes were fed with membrane feeders on blood containing various doses of ivermectin, as described previously.
Anopheles stephensi were given a single experimental blood meal and monitored for egg production. However during colony rearing of
An. albimanus, it was noted that this species required more than one blood feeding in order to produce sufficient eggs to maintain the colony. To accurately assess the effect of ivermectin on
An. albimanus fecundity, it was necessary that control-fed females undergo full gonotrophic development. Therefore,
An. albimanus were first given a normal blood meal (i.e., cow blood with no drugs). Unfed mosquitoes were removed and engorged
An. albimanus were held for 2 days without oviposition sites. The mosquitoes were then offered a second, experimental blood meal. Unfed mosquitoes were removed. One day after experimental blood meals, fully engorged mosquitoes were immobilized by chilling and each mosquito was placed into an individual 30 ml glass vial with screened top and single strip of filter paper on which to rest. Raisins were placed on top of each vial as a sucrose source. After mosquitoes fully recovered and could fly, 2 ml of aged tap water was introduced into the vial for mosquitoes to lay their eggs. Vials were maintained at 26 °C for 5 days, after which vials were filled with ethanol to kill the parent mosquito and the eggs and hatchlings. Eggs and hatchling larvae were counted the same day. Midguts and ovaries from the parent mosquitoes were dissected to quantify blood digestion and ovarian development, based on the Sella scale [
25]. For simplicity, Sella stages were combined into two groups: early stages (Sella stages II, III, IV) indicating that minimal vitellogenesis had occurred, and late stages (Sella stages V, VI, VII) indicating that complete or nearly complete vitellogenesis had occurred.
Feeding on treated mice
Ivermectin stock solution (2000 µg/ml DMSO) was diluted in saline to a working concentration of 80 µg/ml. Three outbred, white mice were weighed and subcutaneously injected with an appropriate volume of ivermectin solution to achieve a dose of 600 µg/kg BW. Three control mice were injected subcutaneously with equivalent volumes of saline. The next day (= 24 h after treatment), each mouse was anesthetized (pentobarbital 60 µg/gm BW, ip) and individually placed onto a screened-top container containing 20–30 An. albimanus mosquitoes. Mosquitoes were allowed to feed for ca. 20 min, after which anesthetized mice were immediately transferred to screen-top containers containing 20–30 An. stephensi mosquitoes. Mosquitoes were allowed ca. 30 min to feed. Afterwards, mice were returned to the vivarium. Unfed mosquitoes were removed from the cages. Engorged mosquitoes were provided with moistened cotton and a sucrose source (chopped prune), and maintained at 26 °C. Four days after blood feeding, surviving mosquitoes from each mouse were subdivided into aliquots of 3–6 mosquitoes each and gently introduced into new containers containing a cup of water for oviposition. Three days after mice had received injections, a fresh batch of An. albimanus and An. stephensi mosquitoes were fed on the mice and the process described above was repeated. Dead mosquitoes were removed and counted daily for 6 days, after which the number of mosquitoes left alive and total number of eggs laid were counted.
Data analysis
Mosquito mortalities observed within experimental groups were adjusted for any mortality that occurred within corresponding control groups using Abbott’s formula [
26]. Only experimental trials having control mortalities less than 20% were used for further data analyses. Log-probit analyses were conducted on the corrected percent moralities to estimate LC
50 values (Minitab Inc., State College PA, USA). Mosquito survivorship was analysed with a Kaplan–Meier survival analysis and Log-rank Mantel-Cox test (GraphPad Software, La Jolla, CA, USA). Mosquito fecundities (i.e., average number of eggs laid per female) were compared among treatments using one-way analysis of variance (ANOVA) on log
10 transformed egg counts. If an ANOVA showed a significant effect due to treatment, the Tukey Honest Significant Differences (Tukey HSD) post hoc test was used to separate statistical differences between groups. Rates of mosquito oviposition and egg hatch, and blood digestion (based on early Sella scores) in treatment groups were compared with their corresponding controls with Chi square analyses (Statistix, Tallahassee FL, USA). A 0.05 level of significance was used throughout.
Discussion
Ivermectin ingested in a blood meal has been shown to be extremely potent against at least eight different anopheline species that have been tested to-date [
11‐
17]. This study found ingested ivermectin to be similarly potent against
An. stephensi (Table
1). However,
An. albimanus is the most ivermectin-tolerant
Anopheles species evaluated to-date, with an oral LC
50 over 25-fold greater than other species tested. The relative insensitivity of
An. albimanus to ingested ivermectin was evident regardless of whether ivermectin was administered in cow blood via membrane feeders or via mosquito feedings on treated mice (Table
3).
To investigate the mechanism(s) of this insensitivity, serial dilutions of ivermectin were injected directly into the haemocoels of
An. albimanus and
An. stephensi and the resulting parenteral LC
50 values were compared with the oral LC
50 values derived from membrane feeding trials (Table
1). The rationale for injecting ivermectin is that the target molecules for ivermectin (i.e., glutamate-gated chloride channels) reside within nervous and muscle tissue outside of the midgut [
27]. Injection bypasses the midgut. When injected directly into the haemocoel of
An. albimanus, ivermectin not only worked more rapidly (Fig.
2B) but was also tenfold more toxic than when it was ingested in a blood meal (Table
1). This indicates that ingested ivermectin was poorly absorbed across the gut of
An. albimanus. Even so, injected ivermectin was still significantly less toxic to
An. albimanus (LC
50 = 188 ng/ml) than to
An. stephensi (LC
50 = 49 ng/ml), suggesting that the binding affinity of ivermectin and/or its stimulatory effect on the glutamate-gated chloride channels of
An. albimanus was less than that of the glutamate-gated chloride channels of
An. stephensi. The relative insensitivity of
An. albimanus to ingested ivermectin may be attributable to (1) poor absorption of ivermectin across the gut of
An. albimanus, (2) potential structural differences between the glutamate-gated chloride channels among anopheline species, or (3) more efficient detoxification processes occurring in
An. albimanus compared to other species of
Anopheles tested.
Curiously, ivermectin was significantly more toxic to An. stephensi when ingested (LC50 = 7 ng/ml) than when injected (LC50 = 49 ng/ml). This suggests that ivermectin metabolites resulting from mosquito digestion of the blood meal may have a contributory role in the overall oral toxicity of ivermectin to An. stephensi.
Ingestion of ivermectin also reduced the reproductive potential of
An. albimanus and
An. stephensi by inhibiting blood meal digestion, vitellogenesis, oviposition rate, fecundity, and egg hatch (Fig.
3, Tables
2 and
3). However, statistically significant reductions in these reproductive parameters occurred only when mosquitoes ingested ivermectin at or above the respective LC
50 concentrations for each species. This is similar to that reported for ivermectin in
An. arabiensis [
14] but stands in contrast to results reported for
An. aquasalis [
16], where significant reductions in egg production and hatching were observed after mosquitoes had ingested ivermectin at a substantially lower potency, i.e., a concentration equivalent to the LC
5 for that species (= 18 ng/ml). Defining the effects of ingested ivermectin on mosquito reproduction is important because significant net reductions in mosquito reproduction could still act to reduce mosquito populations, even at sub-lethal concentrations (see Table
2).
According to the recommendations of most commercial manufacturers of endectocide livestock products, a subcutaneous dose of ivermectin at 200 mg/kg BW is standard for control of intestinal nematodes. At that dose and mode of delivery, the peak plasma concentrations of ivermectin in cattle ranges between 30 and 46 ng/ml [
28‐
30]. Field studies have shown that this amount of ivermectin (200 mg/kg BW) in the blood of treated cattle is sufficient to cause significant mortality and reproductive losses in several species of
Anopheles mosquitoes, including
An. arabiensis [
31],
Anopheles culicifacies and
An. stephensi [
32]. This study suggests that plasma levels of ivermectin in this range would not produce any mortality or loss of reproduction in
An. albimanus that feed on ivermectin-treated cattle.
To successfully implement endectocides for vector control, it is essential to fill the gaps in knowledge about how different mosquito species react to ivermectin. This study found wide differences in oral and parenteral toxicities of ivermectin to the Asian vector, An. stephensi, and the Latin American vector An. albimanus. These studies utilized long established laboratory strains of mosquitoes. An important next step will be to define comparative toxicities of ivermectin and other endectocides against field strains of Anopheles mosquitoes.
Authors’ contributions
SMD and JAV conceived and designed the experiments; SMD, KJM and JAV performed the experiments; SMD and JAV analysed the data; SMD and JAV wrote or revised the manuscript. All authors read and approved the final manuscript.