An Improved Method for in Vitro Feeding of Dermanyssus Gallinae (Poultry Red Mite) Using Baudruche Membrane (Goldbeater’s Skin).

Background: Dermanyssus gallinae, or poultry red mite (PRM), is one of the most economically important ectoparasites in laying hens worldwide with infestations leading to both welfare and economic issues. Research into the testing and validation of novel vaccines and systemic acaricides to control PRM initially requires the use of in vitro methodologies before using in vivo assessments. In vitro methods to date have either involved PRM feeding on hen blood through biological membranes such as chick skin, which can be dicult to obtain and variable in quality, or through synthetic membranes that are unsatisfactory due to breakage and lack of natural feeding cues. The use of hen blood in these systems also requires multiple procedures (bleeds) to provide sucient material for experimental replicates and the potential for using other species, from which larger blood volumes can be withdrawn in a single procedure (e.g. geese), could improve this technique as a renement in the use of animals in research. Methods: Using adult female PRM, an initial experiment with an existing in vitro feeding device employing a Paralm™ M membrane [1] was used to investigate any differences in mite feeding, egg laying and mortality when fed goose or hen blood. Subsequently, any effects on these parameters when PRM were fed through two different membranes; Paralm™ M and Baudruche membrane, and a combination of these membranes with an overlaid polyester mesh to promote mite attachment, were then tested using goose blood as the food source. Results: PRM fed equally well on goose or hen blood through a Paralm™ M membrane and there were no statistically signicant differences in mortality of PRM fed with either blood type, although a signicant increase (p = 0.03) in eggs laid per fed mite when using goose blood was demonstrated, indicating that goose blood is a satisfactory replacement for hens blood. A 70% increase in PRM feeding was observed when mites were fed on goose blood through a Baudruche membrane when compared to the Paralm™ M membrane. Addition of an overlaid polyester mesh did not further improve feeding rates on either membrane. A signicant increase (p = 0.04) in PRM egg laying was observed in mites fed on goose blood through Baudruche membrane, compared to those fed through Paralm™ M. A mean of 1.22 eggs per fed mite was obtained using the Baudruche feeding device compared to only 0.87 eggs per fed mite using the Paralm™ M device when neither had a polyester mesh overlay. It was noted that PRM feeding through Baudruche membrane had fed to repletion when compared to those fed through Paralm™ M. Conclusion: The in vitro feeding of poultry red mite can be readily facilitated through the use of goose blood in feeding devices with Baudruche membrane.

statistically signi cant differences in mortality of PRM fed with either blood type, although a signi cant increase (p = 0.03) in eggs laid per fed mite when using goose blood was demonstrated, indicating that goose blood is a satisfactory replacement for hens blood. A 70% increase in PRM feeding was observed when mites were fed on goose blood through a Baudruche membrane when compared to the Para lm™ M membrane. Addition of an overlaid polyester mesh did not further improve feeding rates on either membrane. A signi cant increase (p = 0.04) in PRM egg laying was observed in mites fed on goose blood through Baudruche membrane, compared to those fed through Para lm™ M. A mean of 1.22 eggs per fed mite was obtained using the Baudruche feeding device compared to only 0.87 eggs per fed mite using the Para lm™ M device when neither had a polyester mesh overlay. It was noted that PRM feeding through Baudruche membrane had fed to repletion when compared to those fed through Para lm™ M. Conclusion: The in vitro feeding of poultry red mite can be readily facilitated through the use of goose blood in feeding devices with Baudruche membrane.

Background
Infestation of laying hen houses with the poultry red mite (PRM), Dermanyssus gallinae, is a major animal welfare and economic problem for the egg-producing industry worldwide, costing the poultry industry in excess of €231 million per year in the EU alone [2]. Current treatment options are limited and often ineffective. PRM infestations cause serious welfare and production issues in the birds as a result of their nocturnal blood-feeding activities and potential as competent vectors for bacterial and viral pathogens [2][3][4]. Novel methods of controlling PRM (novel biopesticides and plant derived products; semiochemicals and growth-regulators; vaccines etc.) are typically tested using small numbers of PRM in in vitro e cacy assays initially, followed by eld testing using large numbers of both parasites and hens [5][6][7]. The in vitro feeding devices employed in these assays contain the mites, allowing them to feed on hens' blood through a membrane such as day-old chick skin [8,9] or arti cial membranes such as Para lm™ M [1,10]. This methodology has technical limitations, not least the requirement for high numbers of replicates to overcome variable feeding rates when using Para lm™ M, and issues around supply, uniformity and quality of chick skin [8,9]. In addition, to supply su cient hen blood for these assays, multiple invasive procedures are required involving venepuncture on hens' wing veins. The aim of this study was to investigate the use of goose blood (which can be obtained in far higher volumes per procedure than hens' blood) as a higher-welfare alternative food source for the assays and goldbeater's skin (also known as Baudruche membrane) as an alternative feeding membrane that may enable improved feeding of adult female PRM.

Blood samples
Chicken and goose whole blood samples were collected by wing vein venepuncture into a 1.5ml capacity Eppendorf tube or syringe containing heparin, respectively, (Sigma-Aldrich Co Ltd, Dorset, UK) to give a nal heparin concentration of 20 units/ml blood. Hen blood from four hens was pooled for further use; individual goose bleeds were used for each in vitro experiment. Trials were carried out on the day of blood collection and blood was kept at room temperature (RT) until needed. Hens were housed in groups in enriched oor pens, geese were housed in predator-proofed, enriched paddocks.
Parasite material PRM were collected and stored as previously described [11]. Brie y, mixed stage and sex D. gallinae were collected at three weekly intervals from a commercial egg laying unit. Prior to feeding assays, mites were stored in vented 75cm 2 canted tissue culture asks (Corning, NY, USA) at RT for seven days to allow blood digestion to occur, after which they were stored at 4°C for up to 3 weeks.
In vitro feeding device Feeding devices (Fig.1) were constructed as described previously [1,4,12]. For initial experiments comparing hen and goose blood as food sources the feeding membrane used was Para lm™ M. For the experiments comparing feeding membranes, the membranes were either a 2cm x 2cm square of Para lm™ M or Baudruche membrane (Preservation Equipment Ltd, UK) held in place by wrapping a strip of Para lm TM M over the edges of the Baudruche membrane, around the outside of the tube, before placing the modi ed pipette bulb goose blood reservoir over the top (Fig. 1). In experiments to assess the impact of an overlaid polyester mesh on the feeding membranes, to improve mite attachment, a polyester mesh (Plastok Ltd, UK) of 105 µm aperture width and a 63 µm depth [10] was overlaid on the ventral aspect of the membranes used in the devices and heparinised goose blood was used as the food source.

Feeding Assays
On removing the mites from storage at 4 o C, they were maintained at RT for 20 mins to allow motile mites to migrate to the cap of the ask. Only motile, adult female mites were collected from the ask caps, transferred to the feeding devices (50 mites per device) and the devices sealed. Once all feeding devices contained mites, heparinised blood was added to the devices, which were then incubated in darkness in an incubator (Sanyo, MLR-351H) at 39 o C and 85% relative humidity (RH) for three hours. Following incubation, mites were recovered from the devices and each fed mite was transferred into a single-well of a 96-well tissue culture plate (Costar, Corning, NY, USA) which was then sealed using AeraSeal™ tape (Sigma-Aldrich Co Ltd, Dorset, UK). Plates were placed into an incubator at 25°C with 85% RH and the mites observed using a stereo microscope at 3, 24, 48, and 144 hours to record mite egg laying and mortality.
For the comparison of adult female PRM feeding rates, egg laying and mortality when fed goose blood or hens' blood through a Para lm™ M membrane, a single experiment was performed with 3 replicate assays per food source. For the comparison of adult female PRM feeding rates, egg laying and mortality when fed goose blood through either a Para lm™ M or a Baudruche membrane (or each of these membranes overlaid with the polyester mesh) two repetitions of the experiment were performed with 3 replicate assays per device in each.

Statistical analysis
Unpaired Student's t-tests were performed on mite feeding, mortality and fecundity data for the comparison of mites fed on goose or hens' blood. An Analysis of Variance (ANOVA) was used to determine variance between the two experiments held on different days comparing the different membranes. To compare any differences between the Para lm™ M membrane and Para lm™ M membrane + mesh, Baudruche membrane and Baudruche + mesh, an ANOVA using Dunnett's multiple comparisons was used. To examine any difference between Para lm™ M membrane and Para lm™ M membrane + mesh and Baudruche membrane and Baudruche + mesh, an ANOVA using Sidak's multiple comparisons test was used. All analyses were carried out using GraphPad Prism v8 (GraphPad Software, San Diego, California USA, www.graphpad.com).

Results
Comparison of feeding between goose and hen blood Goose and hen blood were equally successful in terms of the numbers of PRMs feeding on each blood source and survival of the mites following feeding (Table 1). There was a 26% increase in total egg production in mites fed on goose blood compared to hen blood and the number of eggs laid per fed mite was signi cantly higher in those mites fed on goose blood (p = 0.03). repetitions of the experiment in either feeding, egg laying, progeny per fed mite or mortality and therefore replicates from the two repetitions of the experiment were combined for further analysis. Comparison of feeding rates, fecundity and mortality between PRM fed with goose blood through the different membranes is shown in Table 2. The Para lm™ M membrane in one of the feeding devices developed a split during the incubation period, which is common with this type of membrane, whereas none of the Baudruche membranes failed. No feeding occurred in the failed device and so this replicate was not included in the analysis. A 70% increase in the mean number of PRM feeding was observed when the mites were fed on goose blood through a Baudruche membrane compared to a Para lm™ M membrane (Fig. 2, Table 2) though, because of the high levels of variability in feeding levels of mites feeding through a Para lm™ M membrane, this increase was not statistically signi cant (p = 0.1). Addition of an overlaid polyester mesh did not improve feeding rates on either membrane and a signi cant (

Discussion
Here we have demonstrated that goose blood represents an excellent food source for laboratory maintained PRMs and that Baudruche membrane is a superior feeding membrane to the traditionally used Para lm TM M for in vitro feeding assays. Geese are well suited as blood donor animals as their size means they can donate more blood per procedure than hens (approx. 20 fold). Generally, blood collected from individual hens is pooled for use in feeding assays and the amount collected can be limiting in terms of replicates and/or size of experiment. A statistically signi cant increase in eggs laid per fed mite was demonstrated using goose blood compared to hen blood in the initial experiment using the Para lm TM M only device.
To our knowledge, this is the rst published report on the use of Baudruche membrane in the feeding of D. gallinae, although it has been used for tick feeding in conjunction with silicone as recently as 2019 [13] and for mosquitoes as early as 1964 [14]. Although chick skin has been used successfully to allow in vitro feeding of PRM [5,8,9] availability and ease of use of arti cial membranes has led to their more frequent use. Membranes such as Para lm™ M [1] or Nesco lm coated with a hen skin extract [10] have been employed and can work well. However, Para lm™ M membranes often fail due to the need to stretch the membrane to make it thin enough for mites to feed successfully and Nesco lm is no longer commercially available. Membrane failure makes recovery of fed mites from the device di cult and can lead to misleading mortality data [6] and the rate of device failure and highly variable feeding rates requires additional replicates with the associated increased invasive sampling of hens. Therefore, a more reliable in vitro feeding device increases hen welfare in terms of re nement in number of procedures required for successful data collection.
Krull et al., 2017 [15] suggested that thin membranes for the feeding of tick larvae and nymphs could be obtained using Baudruche membrane coated in silicone, or by using Baudruche membrane alone. As previous attempts at using silicone with lens tissue as described in [16] had not resulted in PRM feeding in our laboratory (data not shown) it was hoped that the reported high tensile strength of the Baudruche membrane alone would su ce and that it would be thin enough to allow the mites to feed, without damaging the membrane. The addition of a mesh support was successfully used in previous tick studies [14,16] which resulted in enhanced attachment times of ticks to membranes. It was decided to test a mesh that had already demonstrated utility in the feeding of PRM [11] to determine if attachment was enhanced. No difference was observed in feeding rates or egg laying with the Para lm TM M only membrane or the Para lm TM M combined with mesh. A signi cant decrease in feeding was demonstrated between the Baudruche membrane combined with mesh in comparison to Baudruche membrane alone. This might be attributed to a lack of tension in the Baudruche membrane in this system and the subsequent di culty for the mites trying to attach through both mesh and membrane compared to membrane alone. Greater variability was observed in both feeding rates and in the number of eggs laid per fed mite when using only Para lm TM M when compared to Para lm TM M used in conjunction with mesh, whereas neither feeding rates or eggs laid per mite were affected by the addition of the mesh. In addition, no membrane failures were observed when using the Para lm TM M with the overlaid mesh.
It was observed in this study that mites that had fed on goose blood through the Baudruche membrane were fully engorged when compared to those that had fed through the Para lm™ M membrane. This repletion level may explain the differences in the trend of increased egg production in the Baudruche fed mites. Therefore, the textured surface of the Baudruche membrane may facilitate easier mite attachment, enabling them to feed to repletion, when compared to the smooth Para lm™ M membrane. The robust performance, availability and good feeding rates makes Baudruche membrane a useful alternative to chick skin and arti cial membranes for use in PRM feeding.

Conclusion
When applied to novel interventions being delivered to PRMs via hen blood (e.g. novel vaccines and systemic acaricides) in vitro studies have previously suffered from highly variable mite feeding rates and high background mortality of mites when using the in vitro feeding system [5,6]. A more reliable in vitro feeding device reduces the necessity of higher replicates and therefore the volume of blood required, leading to a re nement in animal procedures.
A more reliable in vitro feeding method is ideal for initial screening of new control compounds and for those studies that don't require on-hen testing e.g. RNAi studies. Here, we have demonstrated the potential of using Baudruche membrane for in vitro feeding of adult poultry red mites with further studies planned to examine its use for the other hematophagous life stages of the parasite.

Declarations
Ethical Approval

Availability of data
The datasets supporting the conclusions of this article are provided within the article and can be acquired from the corresponding author on request.

Competing interests
The authors state they have no competing interests.  Sequence of construction of In vitro feeding device for poultry red mites. The closed end of a 5ml cryotube (A) is cut off, the tube is inverted and a piece of lter paper inserted (B). Mites are added to the tube and a square of mesh, Baudruche membrane or Para lm™ M is placed over the cut off end and held in place by a strip of Para lm™ M stretched around the outside of the tube (C). A cut down pastette bulb is then placed over the membrane and a watertight seal is obtained by using a castration ring (D). Blood is placed inside the pastette bulb and the device is placed in an incubator in the dark to allow mite feeding to commence.