Vertical Transmission of Salmonella enterica ser. Gallinarum in Dermanyssus gallinae by the Mean of the Baudruche-Based Artificial Feeding Device

Abstract

The poultry red mite (PRM) Dermanyssus gallinae is well known for its vectorial role for pathogens, such as Salmonella enterica ser. Gallinarum, the causative agent of fowl typhoid. Here, we ascertained the vertical transmission of S. Gallinarum across the PRM life stages, combining the Baudruche-based in vitro feeding system and a PRM-fitting DNA extraction and detection method by qPCR. Small-sized pools (4–5 specimens) of adult mites, eggs, larvae, and protonymphs, as well as single eggs, were tested for S. Gallinarum. The pathogen was detected in 89% of adult mites, 5% of single eggs, 17% of pooled eggs, 9% of larvae, and 43% of protonymphs. Additionally, the feeding rate for infected and uninfected mites was similar, while differences in ovipositing and fecundity rate were observed. The method allowed to confirm the infection of mites through the bloodmeal and to strongly suggest the transmission of S. Gallinarum across the PRM life stages. Furthermore, it allows to avoid in vivo studies and it could be useful for further investigating the vectorial role of D. gallinae or other hematophagous arthropods for infectious agents.

1. Introduction

Dermanyssus gallinae (De Geer, 1778), the poultry red mite (PRM), is a serious concern for the poultry production system worldwide [1]. PRM infestations are associated with a wide range of repercussions, which can affect the infested animals, as well as the poultry workers [2]. Indeed, this hematophagous ectoparasite can severely impair the health, welfare, and productivity of birds, leading to restlessness, feather-pecking, cannibalism, anemia, and even death, as well as qualitative and quantitative decline in egg production [1,3]. Furthermore, in case of heavy infestations, humans also may be attacked by D. gallinae, showing conditions as itching, dermatitis, and erythema [4,5,6].

Besides the direct effects on animals and humans, the PRM is also well known for its role as a vector of infectious agents [7]. Several viruses and bacteria have been found to be associated with D. gallinae, such as avian pathogenic Escherichia coli (APEC) [8], Chlamydia psittaci [9], Pasteurella multocida [10], Newcastle disease virus [11], and Fowlpox virus [12]. However, mite-mediated transmission has been demonstrated only for a few pathogens, as in the case of Influenza A virus [13], Salmonella enterica subsp. enterica (S.) serovars Enteritidis [14] and Gallinarum [15]. The latter is the causative agent of fowl typhoid (FT), a septicaemic disease often associated with high morbidity and subacute mortality, which generally occurs in up to four days [16,17]. Although the disease has been eradicated in many countries, outbreaks occasionally occur in commercial poultry farms, severely impairing the health and productivity of the affected flocks [18].

Considering the heavy impact of FT, the relationship between the pathogen and D. gallinae has been widely investigated during the years. Zeman et al. [19] first reported that mites collected in a poultry farm with a history of FT were positive for the pathogen, carrying it for up to four months. The persistence of S. Gallinarum in the PRM has been better elucidated in a field study, in which the pathogen was detected in mites during a FT outbreak, in the following sanitary break, and even five months after the housing of new laying hens [20]. More recently, the mite-mediated transmission of S. Gallinarum from animals experimentally infected with the pathogen to healthy ones has been confirmed by an in vivo study in isolators [15].

Such a relationship may lead to a constant passage of bacteria from infected mites to birds and vice versa, making the eradication of the disease difficult to achieve as long as the PRM persists in the farm.

In the light of those considerations, the role of D. gallinae as a vector of S. Gallinarum becomes even more important in the perspective of a proper eradication of the disease in the affected poultry farms. Nevertheless, some crucial aspects of the PRM-pathogen relationship have not been yet assessed. Among these, the modality in which the mite acquires the pathogen is still not clear, as well as the possibility of the vertical transmission of S. Gallinarum across the different PRM life stages.

Until recently, those questions were still hard to address, due to the lack of proper experimental techniques for the peculiar feeding and reproductive behaviour of the PRM. Great support has come from the recent development of a novel in vitro feeding system, which allows mites to suck blood through a goldbeater’s skin (Baudruche membrane) [21]. These devices have been proven to be highly effective, ensuring a feeding rate of about 50% of the tested mites, and allowing putting the fed mites in the best conditions to perpetuate their biological cycle [21]. However, the small size of D. gallinae impairs downstream molecular analyses, requiring a large number of mites (50–100 specimens per aliquot) to be tested to obtain an adequate amount of DNA [7]. Those large-sized aliquots may impair, among others, the retrieval of accurate data about the pathogen load within mites [7]. Here, we ascertained the vertical transmission of S. Gallinarum in PRM by adopting the feeding system by Nunn et al. [21] and a detection method by real-time PCR (qPCR) from small pools of adults, larvae, and protonymphs of D. gallinae, as well as single eggs.

2. Materials and Methods

2.1. Mite and Blood Samples

Before each trial, mites were collected from a Salmonella-free commercial laying hen farm in Apulia, Italy. After the collection, mites were starved at room temperature for 7 days in plastic bags, to allow the proper digestion of blood. The day of the trial, chicken blood was collected into tubes with 20 units/mL heparin (APTACA, Asti, Italy) and then kept at 39 °C until needed.

2.2. Bacterial Strain and Experimental Infection of Mites

Feeding devices were constructed as previously described [21]. In brief, each device was composed of two chambers separated by a Baudruche membrane (Preservation Equipment Ltd., Diss, UK). First, the lower chamber was filled with 50 female adult mites per device, then, 1 mL heparinised blood was added to each upper chamber.

The S. Gallinarum strain used for the experimental infection of mites was a field strain isolated in 2009 during an outbreak of FT in a poultry farm, stored in 15% glycerol tryptic soy broth at −80 °C. Before use, it was revitalised on tryptic soy agar (Oxoid, Milan, Italy) and incubated overnight at 37 °C. The day before the trial, a single, well isolated colony was inoculated overnight at 37 °C in 25 mL tryptic soy broth (Oxoid). The suspension was titred by the plate count method.

The blood in each device was inoculated with 100 mL of the bacterial suspension at a final concentration of 10⁵ CFU/mL. For the negative controls, blood was added with an equal amount of 0.9% NaCl solution.

Once filled, the devices were incubated in the dark at 39 °C and 85% relative humidity (RH) for 3 h. At the end of the incubation period, the fed mites from each device were counted and transferred singularly into 48-wells tissue culture plates (APTACA), sealed with a breathable AeraSeal tape (Merck, Milan, Italy). The plates were incubated at 30 °C and 85% RH and they were opened at different moments, depending on the required PRM life stage. Specifically, half of the wells were opened after 72 h of incubation to count and collect adults, eggs, and larvae, and the others were opened at 120 h to also obtain protonymphs. Only adults, eggs, larvae, and protonymphs were used in the study, as protonymphs and deutonymphs would have needed additional bloodmeals to molt. Oviposition was calculated by summing the number of eggs, larvae, and protonymphs, each counted only one time.

Overall, seven experiment repetitions were performed using, in total, 350 and 1700 adult female mites placed in devices with uninfected (NC) and infected (SG) blood, respectively. The collected adults, eggs, larvae, and protonymphs were pooled as below described and used for the following investigations. Additionally, pooled adult mites fed in the NC devices, randomly selected from each repetition, were also tested to exclude the presence of S. Gallinarum.

2.3. Molecular Detection of Salmonella Gallinarum in Mites

DNA was extracted from D. gallinae specimens with different methods, depending on the life stage. Adults, larvae, and protonymphs were grouped in pools, each of 5 specimens of the same stage. The DNA was extracted with the ZymoBIOMICS DNA Miniprep Kit (Zymo Research, Irvine, CA, USA), according to the manufacturer’s instructions, in a final volume of 50 mL. The obtained DNA solution was divided into two aliquots, each of 25 mL, centrifuged at 16,000× g for 10 m at 4 °C, and then dried at 65 °C. The resulting pellet was resuspended in 12 or 6 mL TE buffer (10 mM Tris, 1 mM EDTA, pH 8.0).

The eggs were placed singularly or in pools of four specimens in a sterile tube with 12 or 6 mL of sterile water and squashed with a sterile wooden toothpick. Then, the solution was heated at 90 °C for 10 m to cause the egg rupture releasing DNA.

For all the PRM stages, the final volume of DNA solution was calculated in order to be entirely used for the following investigation, reducing at the minimum the material loss.

The detection of S. Gallinarum in mites was performed by a multiplex qPCR with primers and probe previously described [20] and, as an internal control, primers and probe for D. gallinae designed using Primer3 software [22] on the basis of distinctive regions of D. gallinae (GenBank accession no. FN599090.1). Both primers and probe were checked for self-complementarity, hetero-complementarity, and melting temperature according to MIQE guidelines [23]. All primers and probes are listed in

Table 1. Primers and probes used for the amplification of Salmonella Gallinarum and Dermanyssus gallinae.

Oligomer Name	Nucleotide Sequence, 5′ to 3′	Length (bp)	Reference
RTSGf	CCGATATGAGGGATGTAC	144	[20]
RTSGr	AGGTCGTAATGAGTCAAA
RTSGp	FAM-ACATCGTAATTCATGCACTACCACCAT	___
RTELF1aF	GTTCGAAGCTGGTATCTC	183	This study
RTELF1aR	TAGCCGATCTTCTTGATG
RTELF1aP	HEX-ACAGCGACACCTCCTTCTTGAT	___

.

The multiplex qPCR was carried out on CFX Connect Real-Time PCR Detection System (Bio-Rad, Milan, Italy) in 96-well optical plates (Bio-Rad). Amplification was performed in a final volume of 15 mL, comprising SsoAdvanced Universal Probes Supermix (Bio-Rad), 20 nM of each primer-probe solution (Bio-Rad), and 3 mL of template. Non-template controls (NTC) were included in each run, adding an equal amount of sterile water instead of DNA solution. The amplification conditions were as follows: 95 °C for 3 m; 40 cycles of 95 °C for 10 s, 55 °C for 15 s; 60 °C for 15 s. Fluorescence from both FAM and HEX channels was acquired during each extension step. Cycle threshold (CT) and baseline were calculated automatically by the CFX Maestro Software v.5.2 (Bio-Rad). The samples were considered positive when the CT was lower than 35. Each DNA solution and the NTCs were amplified in quadruplicate or duplicate in the same plate.

2.4. Statistical Analysis

The sample size of the SG group was determined according to the method by Humpry et al., 2004 [24] for the estimated prevalence. Parameters were set up as follows: Assumed true prevalence = 0.03 (average infection rate [15]); assumed sensitivity = 0.4 (fed ratio [21]); assumed specificity = 0.9 (underestimated from previous studies, which reported a 100% specificity [15,25]). The sample size of the NC groups was calculated through the estimation of the apparent prevalence setting up the estimated true proportion at 0.03 [15]. In both cases, precision and confidence level were set up at 0.05 and 0.95, respectively. All calculations were carried out through the Epitools portal [26]. The estimated size of the samples was 1659 and 369 for SG and NC groups, respectively.

The normal distribution of all the sets of values was ascertained by the Shapiro–Wilk test. Considering the non-normal distribution of data, the median number of fed and egg-laying mites in NC and SG was calculated. Additionally, the central values and the 95% confidence intervals of the fecundity rate ratios (i.e., number of eggs by fed mites) for NC and SG groups were calculated with the Hodges–Lehmann estimator as previously described [15], and the comparison between the two groups was achieved by the Mann–Whitney U test. Fisher’s exact test was used to compare feeding, oviposition, and fecundity between the mites fed on NC and SG devices. When p was found to be less than 0.05, the difference was considered significant. All procedures were carried out in RStudio v. 2022.02.3 (Posit, Boston, MA, USA) with DescTools package.

3. Results and Discussion

Out of the 2050 mites used in this study, 153 (44%) and 658 (39%) were found engorged in NC and SG devices, respectively, with median values of 24 and 20 mites, but the difference was not found to be significantly different (p = 0.08), despite just above the significance level. The incubation of fed mites allowed the generation of offspring, as listed in

Table 2. Feeding and reproductive parameters observed in mites infected or not infected with Salmonella enterica ser. Gallinarum.

Parameter	Negative Control		Salmonella Gallinarum		p
	n	CI95	n	CI95
Tested mites	350	-	1700	-	-
Fed mites	153	-	658	-	0.08
Egg-laying mites	65	-	146	-	<0.01
Laid eggs *	135	-	313	-	<0.01
Fed mites per device †	24	-	20	-
Egg-laying mites per device †	12	-	3	-
Fecundity rate #	0.9 a	(0.5, 1.2)	0.4 a	(0.3, 0.6)	<0.01

CI95: 95% confidence interval. * The number includes eggs, larvae, and protonymphs. See the text for details. † Each device contained 50 adult female mites. ^# Eggs laid/fed mites ratio. ^a Central value.

and Table S1.

While the feeding ability of mites was not influenced by the presence of S. Gallinarum in the blood, the latter affected mite reproduction. In fact, ovipositing fed females were found in 43% of NC and 22% of SG mites, respectively, with a significant difference (p = <0.01). Similarly, the number of laid eggs by SG-fed mites was found to be significantly reduced (p < 0.01) with respect to the NC-fed mites. Likewise, the fecundity rate was significantly lower in the SG mites (p = <0.001), with 0.9 eggs laid per female in the NC group and 0.4 in the SG group.

Those results suggest that, while D. gallinae has almost the same chance to take a bloodmeal on infected or non-infected animals, S. Gallinarum seems to impair the reproduction of the infected mites, in terms of both oviposition and fecundity. Similar findings about the negative impact of S. enterica were previously reported for mites contaminated by S. Enteritidis [26], relating such evidence to the localization of the pathogen in the reproductive apparatus of female mites.

Actually, S. Gallinarum was detected in adult PRMs fed on experimentally infected blood, as well as in eggs, larvae, and protonymphs from adults fed in SG devices. Out of them, 89% of pooled adult mites, 5% of single eggs, 17% of pooled eggs, 9% of pooled larvae, and 43% of pooled protonymphs were positive for S. Gallinarum (

Table 3. Detection of Salmonella enterica ser. Gallinarum in the tested Dermanyssus gallinae at different life stages.

Life Stage	Specimens per Sample (n)	Samples Tested (n)	qPCR-Positive Samples (n)	qPCR-Positive Samples (%)
Adult	5	19	17	89
Egg	1	41	2	5
Egg	4	23	4	17
Larva	5	11	1	9
Protonymph	5	7	3	43

). All pools of adults, eggs, larvae, and protonymphs from NC devices resulted negative.

Those data evidenced the vertical transmission of S. Gallinarum in D. gallinae from adults to eggs, larvae, and protonymphs. Considering the impact and the sanitary consequences of FT in affected farms, such evidence is particularly worrying. Indeed, the vertical transmission of S. Gallinarum across the PRM life stages may play an important role in enhancing the circulation of the pathogen within and even among the chicken flocks.

Reduced fertility does not appear to be a limiting factor for such a spread, as previously demonstrated [20]. This fact may be due to the huge number of mites that may infest the poultry farms, reaching up to 500,000 per hen [27]. In those conditions, a percentual reduction in fertility does not produce significant effects on the whole mite population.

Apart from demonstrating the vertical transmission, the adopted system allowed to confirm that S. Gallinarum can be orally introduced by PRM. Previous in vivo studies could not ascertain such a contamination route, since the experimental design allowed the mites to be infected following the bloodmeals or by contact with feces or other excretes [15]. In fact, both the transcuticular absorption of surface pathogens and their entry through the stigmata, followed by an internal infection, have been suggested [28].

In vitro devices offer the possibility to confine the mites in a separate environment, focusing on blood as the only infection source. Such an approach confirmed the oral entry of S. Gallinarum, despite it not being able to exclude the other route.

A further, not negligible advantage is the possibility to avoid the use of animal models for investigating the transmission of pathogens mediated by hematophagous mites, such as D. gallinae. The topic raises great interest, insomuch that the vectorial role has been widely investigated during the decades through the adoption of in vitro and in vivo approaches [7]. In vivo studies are usually considered more reliable, but they need to deal with live animals, with all the related ethical concerns. On the other side, most of the in vitro alternatives showed several drawbacks. Among the first methods to be described, one-day-old chick skin has proven to be effective, allowing the assessment of the vectorial role of the PRM for S. Enteritidis [14]. In any case, this option does not solve the problem of using live animals, as chicks must be sacrificed to obtain their skin, and the technique was time-consuming. On the other hand, the synthetic membranes tested (e.g., Nescofilm, Parafilm, or rayon) did not ensure the same results of chick skin in terms of feeding rates [28]. Goldbeater’s skin (Baudruche membrane) has been proposed as a valid alternative to chick skin, offering a feeding rate of 50% with goose blood as food source [21]. Although made by bovine intestine, this membrane is commercially available as preservation equipment.

In this case, the in vitro feeding system allowed us to successfully confirm that D. gallinae can acquire S. Gallinarum from infected blood, and to strongly suggest the transovarial and trans-stadial transmission of the pathogen, as previously demonstrated for S. Enteritidis [29]. In this case, the sensitivity of pathogen detection methods might represent an experimental weakness.

Thus, the detection technique proposed herein ensured to successfully test PRM aliquots composed of few specimens (4–5) and even single eggs. The first studies about the vectorial role of D. gallinae were mostly performed detecting pathogens through cultural [19,28] or immunological methods [29]. The application of molecular techniques (i.e., PRC and qPCR) allowed to reach more refined results about this topic, especially in terms of quantification of pathogens within the mites [8,15,20]. However, these studies are usually made from pools each of 50–100 individuals, hence gathering only estimated information about the pathogen load per mite. The main issue relies on obtaining good quality DNA suspensions, due to the small size of the PRM and the presence of amplification inhibitors [30]. Several extraction methods have been proposed, such as cethyl trimethyl ammonium bromide buffer, a Chelex resin, a filter-based technology, and commercial kits, with the latter as the most effective and widely used [29,30].

Here, we obtained good quality results adopting two different DNA extraction methods, depending on the PRM life stage. DNA from pools of adults, larvae, and protonymphs, each of five specimens, were extracted using a commercial kit with bead beating, ensuring uniform mechanical lysis. Due to their small size, DNA solutions from single or pooled eggs (n = 4) were obtained by squashing them entirely in sterile nuclease-free water, and then incubating them at 90 °C for 10 m. Most of the PCR and qPCR protocols involve the use of a small amount of the template DNA solution. In this study, we dried the solutions and resuspended the pelleted DNA in a final volume small enough to be entirely tested by qPCR. Although this was not the aim of this research, such a strategy could be useful for the quantification of pathogens within the mites. In fact, it offers the possibility to successfully amplify the pathogen and to achieve more reliable data about its load, by testing the entire DNA from small numbers of specimens.

Overall, the combination of the in vitro feeding system and the detection technique proposed herein paves the way for assessing additional aspects of the potential pathogen-PRM interactions. Indeed, further studies may be performed to evaluate in vitro the re-contamination of blood by mites experimentally infected with S. Gallinarum. Although the mite-mediated transmission of S. Gallinarum to healthy animals has been demonstrated in vivo [15], the possibility that hens can acquire the pathogen by ingesting the infected mites cannot be excluded. This method could be applied to determine the transmission of S. Gallinarum through the bloodmeal. Furthermore, it could be useful to assess by qPCR the multiplication of the pathogen within the infected mites, as well as across the different PRM life stages.

4. Conclusions

The method described herein allowed us to successfully detect S. Gallinarum in all the tested life stages of D. gallinae, strongly suggesting its vertical transmission and confirming that mites can acquire the pathogen through the infected bloodmeal. This strategy shows promising applications for further studies about the vectorial role of D. gallinae, without the need to deal with live animals, and it could be applied for evaluating additional aspects about this topic. For instance, the quantification of the pathogen load in small-sized pools of adult mites and their offspring could be useful to evaluate its multiplication within the mites and its fluctuations across the life stages. In addition, the life cycle of the infected mites may be completed by feeding protonymphs to allow their molt to deutonymphs, and deutonymphs to new adults, which may be tested for the presence of the pathogen. Furthermore, the technique might be extended to other pathogens or different haematophagous arthropods.