Antibody response to Phlebotomus perniciosus saliva in cats naturally exposed to phlebotomine sand flies is positively associated with Leishmania infection

Background Zoonotic leishmaniosis, caused by the protozoan Leishmania infantum, is a public and animal health problem in Asia, Central and South America, the Middle East and the Mediterranean Basin. Several phlebotomine sand fly species from the subgenus Larroussius are vectors of L. infantum. Data from dogs living in endemic areas of leishmaniosis advocate the use of antibody response to phlebotomine sand fly saliva as an epidemiological biomarker for monitoring vector exposure. The aim of this study was to analyse the exposure of cats to phlebotomine sand flies using detection of IgG antibodies to Phlebotomus perniciosus saliva. The association between phlebotomine sand fly exposure and the presence of Leishmania infection was also investigated. Results IgG antibodies to P. perniciosus saliva were detected in 167 (47.7%) out of 350 cats; higher antibody levels were present in sera collected during the period of phlebotomine sand fly seasonal activity (OR = 19.44, 95% CI: 9.84–38.41). Cats of 12–35 months had higher antibody levels than younger ones (OR = 3.56, 95% CI: 1.39–9.16); this difference was also significant with older cats (for 36–95 months-old, OR = 9.43, 95% CI: 3.62–24.48; for older than 95 months, OR = 9.68, 95% CI: 3.92–23.91). Leishmania spp. DNA was detected in the blood of 24 (6.9%) cats, while antibodies to L. infantum were detected in three (0.9%). Only one cat was positive to Leishmania by both techniques. Cats presenting IgG antibodies to P. perniciosus had a significantly higher risk of being positive for Leishmania infection. Conclusions To our knowledge, this is the first study demonstrating anti-sand fly saliva antibodies in cats. The evaluation of the contact of this animal species with the vector is important to the development of prophylactic measures directed to cats, with the aim of reducing the prevalence of infection in an endemic area. Therefore, studies evaluating whether the use of imidacloprid/flumethrin collars reduces the frequency of P. perniciosus bites in cats are needed. It is also important to evaluate if there is a correlation between the number of phlebotomine sand fly bites and IgG antibody levels.


Background
Zoonotic leishmaniosis, caused by the protozoan Leishmania infantum, is a serious public and animal health problem in several countries of Asia, Central and South America, the Middle East and the Mediterranean Basin. Domestic dogs are the major hosts of the parasite and the main domestic reservoir hosts for human infection. Nevertheless, the number of feline leishmaniosis reports and subclinical L. infantum infections in cats living in endemic areas has increased in recent years [1]. In fact, there is an increasing trend to consider cats as a potential primary or secondary reservoir host of L. infantum, rather than being an accidental host [2]. This assumption is based on several premises, namely natural susceptibility to infection, suitability to serve as a blood source for phlebotomine sand flies, infectivity to the vector, and close contact with humans where the peridomestic and domestic transmission cycles of the parasite occur [3].
Leishmania parasites are transmitted by phlebotomine sand flies (Diptera: Psychodidae). During the blood meal, immunogenic components present in phlebotomine sand fly saliva are injected into the vertebrate host leading to the development of anti-saliva antibodies [4]. Data from dogs living in endemic areas of leishmaniosis caused by L. infantum suggest the use of antibody response to salivary antigens as an epidemiological biomarker for monitoring vector exposure [5][6][7][8][9][10][11]. The levels of specific IgG antibodies against phlebotomine sand fly saliva positively correlate with the number of blood-fed sand flies [6][7][8][9][10][11][12] and decays after the end of phlebotomine sand fly seasonal activity [7,10].
In the Old World, L. infantum is transmitted by several phlebotomine sand fly species belonging to the subgenus Larroussius, of which Phlebotomus perniciosus is the principal vector in the west part of Mediterranean, including Portugal [13]. Portugal is endemic for canine leishmaniosis [14] and hypoendemic for human visceral leishmaniosis [15]. Feline leishmaniosis [16] and L. infantum infection in cats have been documented in Portugal [17][18][19][20]. The phlebotomine sand fly season usually lasts from May until late October [13,21,22].
The aim of this work was to analyse the exposure of cats to phlebotomine sand flies through the detection of antibodies to P. perniciosus saliva, and to assess associated risk factors. The possible association between phlebotomine sand fly exposure and the presence of Leishmania infection was also investigated.

Animals and samples
From April to December 2017, a total of 350 cats with access to the outdoors from veterinary medical centres, animal shelters and from colonies (captured under the scope of trap-neuter-return programs) from Portugal, were studied. Cats were from three continental Portuguese NUTS II (Nomenclature of Units for Territorial Statistics): Centre (Coimbra and Guarda regions; n = 61), Lisbon (Lisbon and Setúbal regions; n = 266) and the Algarve region (n = 23).
Peripheral blood (1-2 ml) was obtained by cephalic or jugular venipuncture from each animal and collected into EDTA and serum-separating tubes. Serum and buffy coat were obtained by centrifugation and stored at -20 °C until use in serological analyses and DNA extraction, respectively.

Phlebotomus perniciosus salivary glands and detection of anti-P. perniciosus saliva antibodies
Salivary gland homogenate (SGH) was obtained by dissecting salivary glands from 4-6 days-old P. perniciosus females reared under standard conditions [23]. Groups of 20 salivary glands were pooled in 20 mM Tris buffer with 150 mM NaCl and then kept lyophilized at 4 °C until used.
Anti-P. perniciosus IgG was measured in all sera samples by indirect enzyme-linked immunosorbent assay (ELISA). The ELISA was performed in accordance with previous studies [7] with minor modifications. Briefly, flat-bottom microtiter plates (Nunc; VWR, Radnor, Pennsylvania, USA.) were coated with P. perniciosus SGH (0.2 salivary gland per well) in 20 mM carbonate-bicarbonate buffer (pH 9, 100 μl/well) and incubated overnight at 4 °C. The plates were washed with PBS + 0.05% Tween 20 (PBS-Tw) and blocked with 6% (w/v) low fat dry milk diluted in PBS-Tw at 37 °C for 60 min. Feline sera diluted 1/50 in 2% (w/v) low fat dry milk/PBS-Tw was added to the wells (100 μl/well) after washing twice with PBS-Tw. After 90 min incubation at 37 °C, the plates were washed with PBS-Tw and incubated at 37 °C for 45 min with secondary antibodies [AAI26P; Bio-Rad (AbD Serotec), Hercules, California, USA] (100 μl/well) diluted 1:5000 in PBS-Tw. Following another washing cycle, the ELISA was developed using orthophenylendiamine (P23938; Sigma-Aldrich, St. Louis, Missouri, USA) (0.5 mg/ml) in a phosphate citrate buffer (pH 5.5) with 0.001% hydrogen peroxide (30%; Merck, Darmstadt, Germany). The reaction was stopped after 5 min with 10% sulphuric acid and absorbance (OP value) was measured at 492 nm using a NanoQuant (Infinite M200 Pro; Tecan, Zürich, Switzerland). Each serum was tested in duplicate. Wells without serum (but coated with SGH) were used as blanks while sera from cats living in non-endemic countries, namely Germany and Switzerland, served as negative controls. The cut-off value was calculated by the addition of three standard deviations to the mean optical density of the control sera.

Detection of anti-Leishmania IgG
Anti-Leishmania antibodies were determined in sera by the immunofluorescence antibody test (IFAT) as previously described [18]. Briefly, a L. infantum MON-1 (MCAN/PT/05/IMT-373) suspension of 10 7 promastigotes was used as antigen and the anti-cat IgG (whole molecule)-FITC (F4262; Sigma-Aldrich) was used in a dilution of l:20. A serum sample from a seropositive cat (IFAT titre 1:1204) was used as positive control [16] while the serum sample of a cat from a non-endemic country of leishmaniosis was used as negative control. The IFAT cut-off value was established at a serum dilution of 1:64 (the same as used in the laboratory for dogs) [24].

DNA extraction and PCR amplification
DNA was extracted from buffy coat using the High Pure PCR Template Preparation Kit (Roche Diagnostics GmbH, Mannheim, Germany) according to manufacturer's instructions. Detection of Leishmania DNA was done using a nested PCR protocol with primers targeting the small subunit ribosomal RNA (SSU rRNA) gene [25]. A positive control containing L. infantum MON-1 (MHOM/PT/88/IMT318) DNA and a negative control without DNA template were included in each amplification. The DNA amplicons were resolved by conventional electrophoresis on 1.5% agarose gels stained with Green Safe Premium (Nzytech, Lisbon, Portugal), using a 100 bp DNA ladder as a molecular weight marker, then visualized under UV illumination.

Statistical analysis
An exploratory and descriptive data analysis was conducted for the main variables of the dataset. Cats were considered infected with Leishmania if they tested positive for at least one of the techniques (i.e. PCR or IFAT). For the quantitative variable "age in months", the normality and the homogeneity of variance were evaluated using Kolmogorov-Smirnov/Shapiro-Wilk tests and the Levenne test, respectively. When these prerequisites were not valid, the non-parametric Mann-Whitney test was used. To explore the associations between qualitative variables and to compare proportions a Chi-square test, the alternative Fisher's exact test or the Freeman-Halton test was performed. Confidence intervals (95% CI) for proportions were obtained by the Wilson method. This initial approach was followed by multivariate analysis that was developed to evaluate, in an integrated way, possible factors associated with the presence of antibodies against P. perniciosus saliva and with the presence of Leishmania DNA and/or antibodies to the parasite (outcome variables). First, crude odds ratios (OR crude) and 95% CIs were obtained by a simple logistic regression model to screen the effect of each explanatory variable on the outcome variables. In a second step, explanatory variables with a P-value ≤ 0.2 in the univariate analysis were selected and included in the multiple logistic regression model. Finally, a backward stepwise elimination procedure was implemented, using a P-value ≤ 0.05 as the criterion for variables to remain in the model. The Hosmer & Lemeshow goodness-of-fit test, residual analysis and determination of the area under the receiver operating characteristic curve (ROC) were performed. All statistical analyses were conducted using IBM ® SPSS ® Statistics version 25.0 and OpenEpi version 3.01 software.
Only one cat was positive to Leishmania by both techniques. No significant differences were detected in positivity to L. infantum among all the variables/categories studied (Table 1).
Leishmania DNA or specific antibodies to the parasite were detected in 18 cats seropositive to phlebotomine sand fly saliva. Of these 18 cats, all but one had a blood sample taken during phlebotomine sand fly activity. Cats presenting IgG antibodies to P. perniciosus had significantly higher risk ( χ 2 Wald = 4.893, df = 1, P = 0.027; OR = 2.64, 95% CI: 1.12-6.25) of being infected with Leishmania (Table 3).

Discussion
To our knowledge, this study describes for the first time feline antibody response against P. perniciosus saliva in cats naturally exposed to phlebotomine sand flies. The detection of antibodies to P. perniciosus in 47.7% of sera shows that cats are frequently bitten by this sand fly, which is the most abundant Phlebotomus species in the three studied regions [13,26]. The presence of IgGs in 73.6% of sera tested during phlebotomine activity period corroborates the results obtained in dogs from the Lisbon metropolitan area, where antibodies to P. perniciosus SGH were detected in 181 (75.1%) out of 241 animals at the beginning of phlebotomine sand fly activity (i.e. May) and in 209 (86.7%) out of 241 at the end of phlebotomine sand fly season (i.e. October) [27].
Previous studies have demonstrated that canine antibodies to phlebotomine sand fly-saliva correlate with biting intensity, fluctuate within phlebotomine sand fly season and decline significantly after the end of the biting season [6,7,12], emphasizing their usefulness as biomarkers for evaluating the exposure to phlebotomine sand flies and efficacy of vector control campaigns [4,28].
In the present study, two non-casual associations were observed in the univariate analysis, namely the presence of a higher percentage of P. perniciosus antibodies in domestic and treated cats with ectoparasiticides than in stray and untreated cats, respectively. The reasons for these non-casual associations can be explained with the fact that most (73.8%; 107/145) of the blood samples of the domestic cats with access to the outdoors were taken during the exposure period to phlebotomine sand fly bites, while only 35.8% (72/201) of stray cats were sampled during phlebotomine sand fly season activity. On the other hand, the fact that cats treated with ectoparasiticides did not show a lower prevalence of positivity to P. perniciosus saliva than untreated cats was not entirely surprising because the only repellents effective against phlebotomine sand flies, the pyrethroids, are toxic to cats, with the exception of flumethrin. However, the application of imidacloprid/flumethrin collars in cats is still quite low in Portugal [29]. Nevertheless, and despite the lack of repellent effect of the most common ectoparasiticides applied to cats, they can potentially prevent parasite transmission from treated animals to other vertebrate hosts.
However, based on multivariate analysis, the presence of P. perniciosus antibodies in the peripheral blood of cats was neither associated with lifestyle nor with the use of ectoparasiticides, suggesting that these two variables are confounders. These results reinforce the importance of multivariate analysis in addressing confounding in epidemiological studies [30]. Based on this analysis, an association between IgG positivity and phlebotomine sand fly seasonal activity was observed in the present study, being significantly higher between May and October, than during winter months, when phlebotomine sand flies are inactive, suggesting that feline antibodies to saliva are relatively short-living. Unfortunately, no data are available  for cats regarding the kinetics of specific antibodies to phlebotomine sand fly saliva or their correlation with the number of phlebotomine sand fly bites; therefore, it is not possible to precisely correlate feline antibodies against sand fly SGH and the seasonal abundance of P. perniciosus.
In cats, the antibody levels to P. perniciosus saliva were significantly increased with age group, suggesting accumulative exposure of older animals to sand fly bites. A similar positive correlation was repeatedly demonstrated in dogs [8] which is probably related to the reexposure of vertebrate hosts to phlebotomine sand flies following antigenic priming in the previous seasons. Interestingly, cats presenting antibodies to saliva were significantly more at risk of being positive to Leishmania infection. Whether saliva antigens could be used as biomarkers for Leishmania infection remains controversial, since both positive [7,10,11] and negative [6] associations between anti-P. perniciosus SGH antibodies and active L. infantum infection have been observed in dogs from endemic areas of leishmaniosis [4].
Regarding Leishmania infection, antibodies to the parasite or its DNA were detected in 26 cats (7.7%). The positivity of detection of Leishmania DNA (6.9%) was higher than the 0.3% previously obtained in the north and centre of Portugal [17], but lower than the one (9.9%) obtained in the south of the country [19], reinforcing that the rate of Leishmania infection is dynamic over time, depending on the density of proven vector population and on the number of infected vertebrate hosts.
Antibodies to Leishmania were detected by IFAT in 3 cats (0.9%), which is also in agreement with previous studies performed in domestic and stray cats from the metropolitan Lisbon area [18,20] but lower than the 3.8% of seropositivity obtained in cats from the Algarve region [31]. This strengthens the assumption that IFAT might not be sensitive enough to detect Leishmania infection in cats, or at least in those subclinically infected [2,32].

Conclusions
To our knowledge, this is the first study demonstrating the development of anti-sand fly saliva antibodies in cats. Due to the potential role of this animal species in sustaining and spreading L. infantum infection, the evaluation of the contact of cats with the vector is important in the development of prophylactic measures directed to cats with the aim of reducing the prevalence of infection in an endemic area. Further studies are needed to evaluate if there is a correlation between the number of phlebotomine sand fly bites and the dynamics of antibody production and if the use of imidacloprid/flumethrin collars reduces the frequency of P. perniciosus bites and L. infantum positivity in cats.