Comparison of In Vivo and In Vitro Methods for Blood Feeding of Phlebotomus papatasi (Diptera: Psychodidae) in the Laboratory.

Phlebotomus papatasi Scopoli is a medically important insect that has been successfully colonized in the laboratory, and blood feeding is critical for colony propagation. There has been much interest in developing established protocols for in vitro blood-feeding systems. The objective of this study was to determine if a Parafilm membrane and a hog's gut membrane could be successfully used with in vitro feeding systems. We evaluated percentages of P. papatasi females that blood fed on different blood-feeding systems (a mouse, a Hemotek feeder, or a glass feeder) used with either a Parafilm or a hog's gut membrane, with cohorts of 250 and 500 P. papatasi females, and with or without external exhalations. For all feeding system combinations, female P. papatasi blood fed in higher percentages when in cohorts of 500 individuals and in the presence of exhalations. Higher percentages of P. papatasi fed on a mouse, but this study also demonstrates that P. papatasi will readily feed with in vitro feeding systems using a Parafilm membrane or a hog's gut membrane. This study suggests that female P. papatasi may use an invitation effect to blood feed and are attracted to blood sources via chemical olfaction cues, both of which have been characterized in other blood-feeding arthropods. Our study demonstrates that a Parafilm membrane or a hog's gut membrane, in conjunction with the Hemotek or glass feeder system, is potentially a viable alternative to live rodents to blood feed a colony of P. papatasi.

The establishment of laboratory colonies is critical for understanding the biology of arthropods that vector disease agents (Yaghoobi-Ershadi et al. 2007). Researchers using laboratory colonies of phlebotomine sand flies (Diptera: Psychodidae) have been able to study sand fly systematics, physiology, insecticide efficacy, disease transmission, and vaccine development (Rowton et al. 2008,Volf and Volfova 2011). Fewer than 60 sand fly species have been successfully reared in the laboratory, and even fewer have been reared in large numbers (Maroli et al. 1987,Harre et al. 2001,Chelbi and Zhioua 2007,Ivović et al. 2007,Mann and Kaufman 2010,Alarcón-Elbal et al. 2011,Castillo et al. 2015,Oliveira et al. 2015,Goulart et al. 2015).

Phlebotomus papatasi Scopoli, the principal vector ofLeishmania major, the agent of cutaneous leishmaniasis, is one sand fly species that has been successfully colonized in the laboratory (Chelbi and Zhioua 2007). LaboratoryP. papatasi females blood feed on anesthetized rodents (e.g., mice, hamsters, guinea pigs) to acquire a bloodmeal. The blood of these rodents yields sufficient sand fly fecundity, andP. papatasi females are able to readily adapt to feeding on these laboratory hosts (Modi and Rowton 1999,Harre et al. 2001,Volf and Volfova 2011).

To maintain colonies of sand flies, a large number of rodents are required to meet the sand fly feeding demands. The cost and maintenance of supporting rodent colonies have advocated for alternative blood-feeding methods to be investigated (Ward et al. 1978,Harre et al. 2001).Rowton et al. (2008) showed that membrane feeding was a viable alternative to anesthetized hamsters in terms of fecundity and the hatching success of eggs ofP. papatasi.

The use of Parafilm has garnered little attention as a potential membrane for in vitro membrane feeding.Ready (1978) found thatLutzomyia longipalpis fed more intensely through a chick skin membrane than a Parafilm membrane. In that same year,Ward et al. (1978) found thatLu. flaviscutellata did not successfully feed through Parafilm membrane. Overall, Parafilm has not been endorsed as a viable, alternative membrane (Volf and Volfova 2011). In addition, chicken membranes are often used with in vitro feeding systems forPhlebotomus andLutzomyia species (Harre et al. 2001,Noguera et al. 2006,Rowton et al. 2008), but a hog’s gut membrane has been used for feedingLu. shannoni (Mann and Kaufman 2010). In this study, we demonstrate thatP. papatasi females feed through a Parafilm membrane and a hog’s gut membrane using a Hemotek feeding system, as well as successfully demonstrate the use of a hog’s gut membrane with a glass feeder system.

Materials and Methods

Phlebotomus papatasi Colony

ThePhlebotomus papatasi sand flies used in this study were from a laboratory colony at Utah State University (USU, Logan, UT). This colony was derived from a long-establishedP. papatasi colony maintained at Walter Reed Army Institute of Research (Silver Spring, MD). All stages were reared in an environmental growth chamber at 25°C, 85% relative humidity, and a photoperiod of 16:8 (L:D) h according to methods ofLawyer et al. (1991) andModi and Rowton (1999). Larvae were fed a composted 1:1 mixture of rabbit feces and rabbit food (Young et al. 1981,Volf and Volfova 2011). Adults were provided 30% sucrose–water solution daily on saturated cotton balls.

Only femaleP. papatasi were used in this experiment. All females used were at least 2-d posteclosion and had never blood fed. The blood feed trials occurred on the same day and time (between 0900 and 1100 hours), and within the same growth chamber as the main laboratory sand fly colonies. Adult females used in the feeding trials were aspirated and counted from the main colony and released into 24 by 24 by 24” cages (BioQuip, Rancho Dominguez, CA).

Feeding Trials

Four replicates of each treatment combination (feeding system:membrane, 250 or 500 adult femaleP. papatasi, and with or without external exhalations) were completed. For trials with exhalations, the investigator exhaled in the direction of the feeder unit 10 times every five minutes. The same investigator exhaled for all the replicates. The exhalations were performed to simulate natural carbon dioxide emissions from an animal host. Female sand flies that had blood fed were visually confirmed and were counted as blood fed if they were fully engorged or if they had any blood that was visible in the gut.

In Vitro Membranes

Two membranes for in vitro blood feeding were used in this experiment: hog’s gut and Parafilm. Hog’s gut was cleaned with de-ionized water and stored at −20°C until used. On the day a piece was to be used, the membrane was brought to 25°C and blotted dry before being used with the feeder units. For Parafilm (Neenah, WI), on the day a piece was to be used, the piece was cut, stretched, and wrapped around an investigator’s arm for 10 min. Wrapping the Parafilm around the investigator’s arm was an attempt to allow sweat and odorants to adsorb onto the Parafilm to lure the females to the heated blood source. The Parafilm was then removed from the arm such that the surface in contact with the skin was the outer membrane and in direct contact with the probing sand flies. The Parafilm was further stretched, tightened, and sealed to an artificial feeding unit.

In Vitro Artificial Feeders

Two in vitro artificial feeding systems were used for this experiment: the Hemotek PS5 electrical feeder (Discovery Workshops, Accrington, United Kingdom) and glass feeders (Kontes Custom Glass, Vineland, NJ). For the Hemotek feeder, the Parafilm or hog’s gut membrane was secured to the feeder unit using an O-ring, and 1.5 ml of defibrinated bovine blood was added. The now-ready Hemotek unit was attached to the heating source and set to 38°C, placed inside the sand fly cage on a stand, and female sand flies were allowed to blood feed across the membrane for 1 h (Fig. 1A). For the glass feeders, the Parafilm or hog’s gut membrane was secured to the open end of the feeder using a rubber band, and 1.5 ml of defibrinated bovine blood was added. One feeder was placed inside the sand fly cage, horizontally secured to a stand, and the blood was heated to 38°C from circulating water using a peristaltic pump (Woessner 2007) and a Fisher Scientific Isotemp model 2340 water bath (Fisher Scientific, Marietta, OH). For 1 h, female sand flies were allowed to blood feed across the membrane.

Fig. 1.

Images of an aggregation of femaleP. papatasi sand flies feeding through a Parafilm membrane with the Hemotek PS5 in vitro blood feeder (A) and on anesthetized hairless mouse (B).

Mouse Blood Feed

One hairless mouse was anesthetized via intraperitoneal injection of a cocktail of ketamine/xylazine/acepromazine maleate. Once immobilized, the mouse was placed on its side inside the middle of the cage. Female sand flies were allowed to blood feed on the mouse for 1 h (Fig. 1B).

Statistical Analyses

Statistical analyses were performed using SAS/STAT 14.1 in the SAS System for Windows 9.4 TS1M3 using the GLIMMIX procedure (SAS Institute 2001). We conducted analyses of two data subsets to accommodate the fact that the mouse feeding system cannot be combined with “no exhalations,” as it is a living organism.

The effects of an in vitro feeder (Hemotek feeder or glass feeder), membrane (Parafilm or hog’s gut), exhalation (presence or absence), and number of females in the cage (250 or 500) on the percentage of femaleP. papatasi that blood fed were analyzed using a four-way factorial in a completely randomized design. The effects of feeding system:membrane (mouse, Hemotek feeder:Parafilm membrane, Hemotek feeder:hog’s gut membrane, glass feeder:Parafilm membrane, glass feeder:hog’s gut membrane) and number of females (250 or 500) on the percentage of femaleP. papatasi that blood fed, only with exhalations, were assessed using a two-way factorial in a completely randomized design.

Both analyses used a generalized linear model with a binomial distribution and a logit link, with observation-level variance estimated to address overdispersion. Pairwise comparisons among means were adjusted for inflated Type I error using the Tukey method. A threshold of α = 0.05 was used for all analyses.

Results

Mean percentages of femaleP. papatasi that blood fed in each trial combination are shown inTable 1. The mouse system had the highest observed mean percentage of females that blood fed of any of the treatment combinations (38.3% with 500 females). The glass feeder with a hog’s gut membrane, 500 females, and exhalations had the highest observed mean percentage of females that blood fed of any in vitro combination (26.5%). The Hemotek system with a Parafilm membrane, 250 females, and no exhalations had the lowest observed mean percentage of females that blood fed of any in vitro combination (0.8%).

Table 1.

Mean percentage (± SD) (n = 4 per treatment combination) of femaleP. papatasi that blood fed

	Exhalations
	No		Yes
	No. of flies in replicate		No. of flies in replicate
	250	500	250	500
Feeder:Membrane
Mouse			26.4% (± 9.2%)	38.3% (± 3.4%)
Glass feeder:Parafilm	1.0% (± 1.0%)	3.6% (± 0.6%)	7.0% (± 1.7%)	21.0% (± 3.1%)
Glass feeder:Hog’s gut	15.6% (± 5.9%)	18.2% (± 3.9%)	14.9% (± 6.7%)	26.5% (± 10.9%)
Hemotek:Parafilm	0.8% (± 0.7%)	1.3 % (± 0.7%)	8.4% (± 4.6%)	22.3% (± 2.7%)
Hemotek:Hog’s gut	6.1% (± 3.1%)	5.5% (± 2.8%)	23.7% (± 12.6%)	23.6% (± 7.5%)

In Vitro Feeding Outcomes

A higher percentage of femaleP. papatasi blood fed in larger cohorts of 500 than in cohorts of 250 (P = 0.011). Presence of exhalations increased the percentage of femaleP. papatasi that blood fed (P < 0.001), as well as the increase was more pronounced with the Hemotek feeder (P = 0.028). The percentage of femaleP. papatasi that blood fed was higher with hog’s gut membranes than Parafilm (P < 0.001), particularly in the absence of exhalations (P < 0.001).

In Vivo and In Vitro Feeding Outcomes in the Presence of Exhalations

The effect of cohort size on percentage of femaleP. papatasi that blood fed was not the same for all five system:membrane combinations (P = 0.041). With a cohort of 500 femaleP. papatasi, the percentage of females that blood fed with the mouse system was higher than the percentage with any in vitro feeding system (Hemotek feeder:Parafilm membrane system,P = 0.004; Hemotek feeder:hog’s gut membrane system,P = 0.010; glass feeder:Parafilm membrane system,P = 0.002; and glass feeder:hog’s gut membrane system,P = 0.055). With a cohort of 250 females, the percentage of females that blood fed with the mouse system was larger than any in vitro feeding system using Parafilm (Hemotek feeder:Parafilm membrane,P = 0.008; glass feeder:Parafilm membrane,P = 0.004). The mouse feeding system was not distinguishable from either hog’s gut membrane feeding system (Hemotek feeder:hog’s gut membrane,P = 0.987; glass feeder:hog’s gut membrane,P = 0.185). With the mouse system, the percentage of females that blood fed was higher with cohorts of 500 than with 250 (P = 0.028).

Discussion

The findings from this study demonstrate that an in vitro feeding system using Parafilm and/or hog’s gut membrane can be successfully used to feed femaleP. papatasi a bloodmeal, which counters previous reports that suggest that in vitro feeding systems with a Parafilm membrane will not adequately provide the required bloodmeal to sand flies (Ready 1978,Ward et al. 1978). This option reduces the cost burden, potential safety hazards, and the need for additional protocols associated with maintaining and handling live animals and controlled substances (Costa-da-Silva et al. 2014).

Our initial efforts with the Hemotek feeder and a Parafilm membrane involved numerous iterations to find an effective blood-feeding method. Preliminary attempts included the use of baited lures with octenol to attract femaleP. papatasi to the blood source, but these were not very effective at increasing the feeding rate. Using arm-wrapped Parafilm with intermittent exhaling into the cage near the Hemotek feeder was a successful combination to lureP. papatasi females to the blood source. Many hematophagous Diptera are attracted to some combination of chemicals including carbon dioxide, water vapor, and host odors (Gibson and Torr 1999).Pinto et al. (2001) found that the closer carbon dioxide traps and human-baited traps were positioned to one another, the fewerLutzomyia sand flies were trapped in carbon dioxide traps compared to human-bait traps.Bernier et al. (2008) found that traps baited with carbon dioxide and human hair captured more sand flies, although not significantly more, than traps with only carbon dioxide or with carbon dioxide plus octenol.Kline et al. (2011) discovered that black traps with body heat, moisture, and carbon dioxide captured roughly 40 times moreP. papatasi than equivalent traps without carbon dioxide. The feasibility of humans to provide carbon dioxide in the form of human exhalants may not be deemed practical for long-term, large-scale mass rearing ofP. papatasi in the laboratory. Other sources of carbon dioxide, such as compressed carbon dioxide, or less frequent intervals of human exhalations, should be considered when using in vitro systems of blood feeding. The research findings from this study though, suggest and support a combination of body odorants with carbon dioxide as a potent lure for sand flies.

Higher percentages of femaleP. papatasi blood fed when in cohorts of 500 compared to 250. This effect may be explained by an aggregation behavior on hosts or blood-feeding sites (Tripet et al. 2009). We observed that an aggregation would initiate when a single female probed the Parafilm membrane, hog’s gut membrane, or mouse until the sand fly found a suitable location to blood feed (Fig. 1A, B).Schlein et al. (1984) was able to characterize the invitational effect forP. papatasi via a pheromone released from the palps of females. The invitational effect has been characterized in another sand fly species,Lutzomyia longipalpis, as well as ceratopagonids, simuliids, andAmbylomma ticks (Norval et al. 1989,Blackwell et al. 1994,McCall and Lemoh 1997,Tripet et al. 2009). Aggregations ofLu. longipalpis during blood feeds have been suggested to benefit individual females by needing to produce less saliva, truncating the time needed complete blood feeding, and having higher fecundity (Tripet et al. 2009).

The use of biological membranes with in vitro feeding systems has been demonstrated to be effective in blood-feeding sand flies (Harre et al. 2001,Noguera et al. 2006,Rowton et al. 2008,Mann and Kaufman 2010). Even with the seemingly lowP. papatasi blood-feeding rates demonstrated in this study, a Hemotek feeder with a Parafilm membrane has been used successfully at Utah State University to establish new colonies, as well as to maintain longstanding colonies. For example, we used a Hemotek with Parafilm membrane system to obtain sufficiently large quantities of flies for the analysis of insecticide resistance (Denlinger et al. 2015,2016). Even with an in vitro system feeding rate ranging from 8% to 22% (Table 1), researchers working to establish a newly formed colony are capable of successfully feeding and capturing substantial numbers of female sand flies needed to oviposit on a weekly basis through multiple blood feeds, thereby establishing a colony within a few generations. We hypothesize that a larger colony (e.g., 750, 1,000, or 2,000 females) would increase the percentage of femaleP. papatasi that blood feed. A limitation of this study was that fecundity rates were not evaluated for all feeding system combinations. It is important to note though, that during the initial months of establishing aP. papatasi colony at Utah State University when Parafilm was discovered to be an effective membrane, the colony consistently yielded sufficient numbers of viable eggs from generation to generation.

TheP. papatasi colony used in this study originally derived from a 30-yr established colony maintained at the Walter Reed Army Institute of Research. That colony has a history of blood feeding using hamsters and was not preadapted for feeding across a membrane used with an in vitro feeding system. This history suggests that host-seeking traits in laboratoryP. papatasi can be quickly selected for and that feeding on a different host or membrane (i.e., mouse, Parafilm membrane, or hog’s gut membrane) does not have significant detrimental effects on fecundity.

Further studies could be developed to understand the success, utility, and impacts of an in vitro feeding system, with a Parafilm or hog’s gut membrane, with respects to fecundity and hatching rates as a viable alternative where live animals are not feasible as a blood source. For example, studies could include analyzing impact on fecundity and survival of recently field-collected sand flies, its utility in mass-rearing other laboratory-colonized sand fly species capable of vectoringLeishmania, and its ability for initiating and maintaining sand fly species that are not yet successfully colonized in the laboratory. Our study demonstrates that in vitro feeding system combinations were effective for a single colony ofP. papatasi. Feeding success may vary for different geographicP. papatasi collections from around the world. In addition, in vitro blood-feeding systems, especially with a Parafilm membrane, may be viable for other laboratory uses, like vector competence analysis. For example, in studies examining sand fly vector competence, chick skin membranes have been used with in vitro blood-feeding systems for many types sand fly species withLeishmania-infected blood (Hlavacova et al. 2013;Pruzinova and Volf 2013;Sadlova et al. 2013). The effect of Parafilm in lieu of a biological membrane needs to be investigated as a potential membrane in vector competence studies.

The findings from this study suggest that a Parafilm or hog’s gut membrane used with either the Hemotek or glass feeder system is well-suited for maintaining largeP. papatasi colonies. These combinations can be considered as alternative feeding systems in lieu of rodents if the costs and maintenance of keeping rodents is prohibitive. This option could also potentially be used to conduct additional studies to further the understanding of vector competence and the sand fly’s contribution to disease transmission.