Leishmania infantum-specific IFN-γ production in stimulated blood from cats living in areas where canine leishmaniosis is endemic

Background Feline leishmaniosis caused by Leishmania infantum is considered a rare disease in endemic areas, whereas subclinical infections are common. Immune response plays a key role in driving the course of L. infantum infection in other host species; however, the feline cell-mediated immune response to L. infantum infection has not yet been investigated. The aim of this study was to determine the cell-mediated immune response specific to L. infantum by means of interferon (IFN)-γ release in whole blood assay from cats living in endemic areas (66 in Sicily and 113 in Catalonia) and to compare with antibody levels to L. infantum [enzyme-linked immunosorbent assay (ELISA) and immunofluorescence antibody test (IFAT)], blood parasite load and retroviral infections. Results Most cats (n = 140) were L. infantum antibody negative and only 22% (n = 39) were positive. Only 9 and 2% of tested cats had a feline immunodeficency virus (FIV) infection or a feline leukemia virus (FeLV) infection, respectively. Thirty-two cats out of 179 (18%) produced IFN-γ after stimulation with L. infantum soluble antigen (LSA) while the majority of cats (93%) produced IFN-γ after stimulation with concanavalin A (ConA). Six LSA-IFN-γ-producer cats were seropositive (three to ELISA and five to IFAT) but they were polymerase chain reaction (PCR) negative, while only one cat was antibody- and PCR-positive. Significant positive correlations were found between IFN-γ concentrations after stimulation with LSA and ConA, and between serology and PCR testing. No association was found between FIV status and LSA or ConA-IFN-γ production. Combining PCR, serology and specific IFN-γ concentration results, we found that 36% of cats studied were exposed to L. infantum. Conclusions As expected, cats from endemic areas produce IFN-γ after ex vivo blood stimulation with LSA and therefore are able to activate a cell-mediated adaptive immune response against the parasite that is variably associated with antibody or blood PCR positivity. The association of this assay to serological and molecular tests provides a better estimate of cat exposure to L. infantum.


Background
Leishmaniosis is a vector-borne disease of humans and animals caused in Europe by Leishmania infantum and transmitted by female sand flies of the genus Phlebotomus [1][2][3]. Dogs are considered the main reservoir of L. infantum but there is clear evidence that some wild and synanthropic mammals and domestic cats are able to infect sand flies and they play a variable role in a reservoir system according to local and ecological peculiarities [3][4][5].
Feline leishmaniosis (FeL) due to L. infantum infection was described, for the first time, in 1912 by Sergent et al. [6] and since then it has been globally reported in endemic areas [5,7]. While subclinical feline infections are common in areas where canine leishmaniosis (CanL) is endemic, clinical illness due to FeL is rare [7,8]. Clinical disease is frequently associated with possible impaired immunocompetence, as in case of retroviral co-infections, immunosuppressive therapy, or malignant neoplasia [7,8]. The immune response plays a crucial role in the control of Leishmania infection. Albeit some differences according to the host species, T cells modulate and orient, through cytokine production, macrophage reaction to the parasite [9][10][11]. Antigen processing cells (dendritic cells and macrophages) present L. infantum antigens to CD4 + T cells that modulate the type of immune response [12]. A T helper 1 (Th1) oriented immune response, associated with production of interferon-gamma (IFN-γ), interleukin (IL)-2 and tumor necrosis factor alpha (TNFα), stimulates phagocytosis by macrophages, their production of nitric oxide and reactive oxygen intermediate and consequent phagocyte-based parasite intracellular elimination [11,13]. Conversely, in dogs susceptibility to infection and disease progression is mediated predominantly by a non-protective T helper 2 (Th2) immune response and the production of cytokines such as IL-4, IL-10, IL-13 and transforming growth factor beta (TGFβ) which are associated with downregulation of the cellular immune response, a high level of antibodies, and L. infantum dissemination [9][10][11].
Different feline innate and adaptive immune responses might account for the observed lower prevalence of L. infantum infection as well as clinical leishmaniosis in cats as compared to dogs [7,14]. It is well recognized that cats appear to be less frequently affected by arthropodborne diseases when compared to dogs, although no important differences are known between the canine and feline immune systems [15]. Interestingly, to the best of our knowledge, no studies have so far evaluated L. infantum-specific cell-mediated immunity in cats. However, detection of IFN-γ in antigen stimulated whole blood or peripheral blood mononuclear cells (PBMCs) was used in cats to evaluate cellular immune response mechanisms to other pathogens (e.g. feline coronavirus or Toxoplasma gondii) [16][17][18].
The main aim of this study was to determine the L. infantum-specific cellular immune response in cats by means of the evaluation of IFN-γ production in stimulated blood from cats living in endemic areas of CanL (Catalonia and Sicily) and correlate it with L. infantum antibody levels, blood parasitemia, and retroviral status. and at Ambulatorio Veterinario Santa Lucia (Lipari, Messina). Inclusion criteria for enrollment included the exposure to at least one sand-fly season and no treatment with repellent ectoparasiticides (i.e. pyrethroid products). Sex, age class, breed, clinical status, lifestyle of pet cats (indoors, outdoors), and number of stray cats at each sampling site are summarized in Table 1. Age was classified as follows: young (6-18 months); adult (between 19 and 96 months); and old (> 96 months). The clinical status of cats was defined as "sick" or "apparently healthy" based on data available from history and physical examination. Sick cats were considered when clinical signs compatible with FeL (i.e. lymph node enlargement, skin, mucosal or eye lesions, stomatitis, weight loss, chronic kidney disease, anaemia) were found as previously described [7]. Apparently healthy cats were considered when no clinical signs were present with the exception of four cats admitted because of trauma that were also included in this group.

Study areas, cats and sampling
One milliliter of blood, aseptically put into a heparin tube, was used for whole blood assay within 24 h after blood sampling. EDTA blood and blood serum were also obtained and immediately aliquoted and stored at -20 °C until processed for DNA extraction and serological investigations, respectively.

Feline IFN-γ release whole blood assay
Whole blood assays were performed as previously described in dogs [10] with some minor modifications to adapt the technique to the smaller volume of blood available. Heparinized blood was diluted at a ratio of 1:10 with medium and after five days the supernatant was collected from unstimulated cultured cells. Two aliquots of diluted blood were respectively cultured with Concanavalin-A (ConA) or L. infantum soluble antigen (LSA) to subsequently recover supernatants from stimulated blood cells.

Leishmania infantum antibody detection IFAT
Anti-L. infantum IgG antibodies were detected using the L. infantum (strain MHOM/IT/80/IPT1) antigen produced by C.Re.Na.L. (Centro di Referenza Nazionale per la Leishmaniosi, Palermo, Italy). Fluoresceinated anti-cat immunoglobulin G (IgG) antibody (working anti-feline IgG (H+L)-FITC, Fuller Laboratories, Fullertone, CA, USA) was used. The manufacturer's protocol was followed, and the end point titer of positive samples was determined preparing serial 2-fold dilutions of serum starting from 1:20. The cut-off value for positivity was set at 1:80 [19,20]. Fluorescence microscope readings were made by a unique operator (MM).

In-house ELISA
ELISA was performed as previously described [20,21]. All plates included the serum from a sick cat from Cyprus with a confirmed infection with L. infantum as positive control, and the serum of a cat from an area where leishmaniosis was not endemic as a negative control. All samples were run in duplicate. The cut-off was established at 12.3 ELISA units (EU) (mean ± 3 standard deviations of sera from 81 cats from the UK, a non-endemic area).

Blood DNA extraction and Leishmania real-time PCR
Total DNA was extracted from EDTA blood using the DNA Gene extraction kit (Sigma Aldrich, Saint Louis, MO, USA) following the manufacturer's instructions with some modifications. Forty microliters of proteinase K solution were added to all samples. Four hundred microliters of whole blood were used for all the samples. Blood from a clinically healthy non-infected cat was used as a control for DNA contamination in every DNA extraction performed. The real-time polymerase chain reaction (RT-PCR) was carried out in a CFX96 Real-time System (Bio-Rad Laboratories s.r.l., Hercules, CA, USA) using TaqMan Master Mix (Applied Biosystems by Ther-moFisher, Waltham, MA, USA) and performed as previously described [22].

Statistical analysis
A minimum sample size of 173 cats was calculated for cat enrollment, based on the prevalence of feline L. infantum infection in areas under study (Catalonia and Sicily) [14,19,20,23], and on the assumptions of 99% confidence level and 5% precision [24]. Fisher's exact test was used to compare groups defined by categorical variables. Results of tests for L. infantum exposure of cats did not pass the D' Agostino-Pearson normality test. Accordingly, the Mann-Whitney U-test was used to compare unmatched continuous data and the Wilcoxon signed-rank test to compare paired continuous variables. Spearman's correlation coefficient was calculated to evaluate relationships between ELISA, IFAT and PCR results and between levels of IFN-γ, anti-Leishmania antibodies, and L. infantum DNA in blood of cats studied. P-values<0.05 were considered significant. The D' Agostino-Pearson normality test and Fisher's exact test were performed using Prism 7 for Mac IOS and all other tests were performed using SPSS software v.17.0 for Windows. Finally, Cohen's kappa coefficient was measured analyzing results of ELISA and IFAT with blood PCR (https ://idost atist ics.com/cohen -kappa -free-calcu lator ).

Cats
Sex, breed, age class, lifestyle, origin (region and municipality) and clinical status of 179 cats included in the study are summarized in Table 1. The median of age of cats was 24 months [25-75 percentile (25-75) = 18-48 months; interquartile range (IQR) = 30 months]. There were no differences between the two regions related to the variables sex, breed, age class and lifestyle of cats. Conversely, cats from Sicily were more frequently found sick compared to Catalan cats (Fisher's exact test, P < 0.0001).

Leishmania infantum serological tests
Results of serological tests are shown in Table 2 and Fig. 1. A slight positive correlation (Spearman's correlation coefficient, r s = 0.342, P = 0.0001) was found between ELISA and IFAT. However, only 11 of the 39 cats positive to IFAT and/or ELISA were positive to both tests (28%) and an agreement of 78.1% was found when these tests were compared (Cohen's kappa coefficient = 0.18, i.e. slight agreement beyond chance). No differences were found regarding seropositivity and variables studied.

IFAT
IFAT results based on cat provenience and clinical status are summarized in

Leishmania real-time PCR
PCR results based on cat provenience and clinical status are summarized in Table 2. Only 9 cats (5%) were Leishmania-PCR positive (Fig. 1) and the median parasite load was 60 L. infantum amastigotes/ml [(25-75) = 10-128.5]. Leishmania-PCR positive cats were more frequently positive by serology (89% by IFAT and 78% by ELISA) than Leishmania-PCR negative cats (13% by IFAT and 7% by ELISA) (Fisher's exact test, P<0.0001). In addition, antibody level measured by IFAT (Mann-Whitney U-test, Z = 86.5, P < 0.0001) and ELISA (Mann-Whitney-U test, Z = 219, P < 0.0001) of Leishmania-PCR positive cats was significantly higher when compared to Leishmania-PCR negative cats. Moreover, a positive correlation was found between PCR and ELISA (Spearman's correlation coefficient, r s = 0.272, P = 0.0001) with an agreement of 86.1% (Cohen's kappa coefficient = 0.25, i.e. fair agreement) as well as a positive correlation between PCR and IFAT (Spearman's correlation coefficient, r s = 0.347, P = 0.0001) with an agreement of 80.1% (Cohen's kappa coefficient = 0.18, i.e. slight agreement). No other differences were found in positive PCR prevalence according to variables under study.

Anti-FIV antibodies and FeLV p27 antigen
Thirteen cats (9%) were antibody-positive to FIV and 4 (2%) were antigen-positive to FeLV (Table 2). FIV and FeLV co-infections were never detected. All FeLVpositive cats were from Catalonia while there was no  (2) difference in FIV prevalence between cats enrolled in Sicily (9%) and in Catalonia (8%) ( Table 2). Moreover, there were no statistical differences in prevalence of anti-L. infantum antibodies or Leishmania DNA positivity in cats positive to FIV or FeLV compared to the negatives. Similarly, no statistical difference was found for the levels of anti-L. infantum antibodies or L. infantum parasite load between FIV-or FeLV-positive cats and the negative ones. Interestingly, four FIV-and two FeLV-positive cats were also positive to one or more L. infantum diagnostic tests. Only one FeLV-and two FIV-positive cats produced IFN-γ after LSA stimulation. Moreover, two of them were positive to both serological tests and blood PCR. Conversely, 107 cats were L. infantum and retrovirus negative and 21 produced IFN-γ after LSA stimulation. There were no statistical differences in IFN-γ productions (after LSA or ConA stimulations) between FIV-or FeLV-positive and negative cats.

Overall L. infantum frequency of infection
An overall L. infantum prevalence of 36% calculated by PCR, serology and LSA IFN-γ assay was obtained in the population studied (Fig. 1) and no differences were found between Catalonia (35%) and Sicily (36%), the sites of each region studied, or between apparently healthy (33%) and sick cats (42%).

Discussion
To our knowledge, this study demonstrated for the first time that cats naturally exposed to L. infantum infection produced IFN-γ after ex vivo whole blood stimulation with L. infantum antigens, as occurs in dogs, humans and laboratory animals [10,25,26].
Parasite-specific IFN-γ production was found in 18% of enrolled cats and it was associated with antibody production in only seven of the 32 IFNγ-p cats of this study. Both types of adaptive immune responses were therefore variably combined in single cats and a wide immunological spectrum may exist also in cats as already reported in dogs and humans [10,27].
The present study did not aim to evaluate the parasite T cell mediated immunity of clinical cases of FeL but based on data available from physical examination, the frequency of IFNγ-p individuals was not different in apparently healthy and sick cats. Similarly, there was no difference in the prevalence of sick cats between IFNγ-p and IFNγ-np cats positive to serological or blood PCR tests. However, sick IFNγ-p cats produced a significantly higher level of IFN-γ. Moreover, the unique cat with clinical FeL confirmed by cytology was IFNγ-p and, despite possible immunosuppression due to FeLV infection and to neoplasia, this cat reached an IFN-γ level around the median value. In the present study, retroviral infection per se did not affect adaptive immune response of cats to L. infantum, but a limitation was the very low number of FIV (n = 13) or FeLV (n = 4) infected enrolled cats.
Dogs with mild or moderate CanL produce IFN-γ and the lack of production is restricted to severe disease [10,28]. Prospective clinical studies with cats with confirmed clinical leishmaniosis are needed to further evaluate if IFN-γ is as possible marker for staging the severity of disease and evaluating efficacy of therapy as documented in dogs [10,28].
IFN-γ production after stimulation with ConA was obtained in almost all cats and it was not associated with any variable, including clinical status, retroviral infections or exposure to L. infantum. These findings are similar to data from studies of CanL [10,28]. As expected, concentrations of ConA-IFN-γ were significantly higher than LSA-IFN-γ and they were positively correlated. Production of ConA-IFN-γ suggests non-specific T cell activation. Surprisingly, when considering cats exposed to L. infantum (i.e. positive to at least one the specific tests performed), sick individuals produced a significantly higher level of ConA compared to apparently healthy individuals. This finding is unusual, and it is difficult to when a significantly lower concentration is observed [10]. The prevalence of anti-L. infantum antibodies obtained in this study by combination of both IFAT and ELISA results was similar in the two regions with an approximate value of 20%. It is not easy, therefore, to explain the higher frequency and level of positivity of ELISA in Sicilian cats but these cats were more frequently found sick compared to Catalan cats ( Table 2). Previous studies conducted in Sicily were based on IFAT and, when the same cut-off was used, anti-L. infantum antibody prevalence ranged between 6.6-29% [19,[29][30][31][32]. Conversely, only two studies evaluated anti-L. infantum antibody prevalence in Catalonia; they both used ELISA and antibody prevalence was 1.7 and 5.3%, respectively [14,33]. However, as already reported, we found a positive correlation between ELISA and IFAT [21].
The parasite load of cats was measured by blood PCR, thus some positive cats were possibly missed because blood is not the most sensitive tissue for the detection of L. infantum DNA in dogs and likewise with cats [7,34]. However, all PCR-positive cats were also antibody-positive and the level of L. infantum DNA in blood was positively correlated with the antibody level of cats. This is not a common finding in cats from L. infantum endemic areas but, compared to other studies, we used two serological techniques and therefore the sensitivity of serology was increased [19,20,35].
In this study, most IFNγ-p cats were negative to L. infantum antibody and DNA detection in blood (25/32, 78%). This finding means that the combination of serological and molecular tests with L. infantum-specific IFN-γ evaluation offered a more accurate overall estimation of the exposure to L. infantum of cats under study. In fact, the prevalence of L. infantum was 5% for blood PCR and 22% for antibody detection but the combination of results obtained by serological and molecular testing did not increase the percentage of positivity because all PCR positive cats were also positive for L. infantum antibodies (Fig. 1). However, when also considering positivity for L. infantum-specific IFN-γ production, the overall L. infantum prevalence raised to 36%. These data confirm that a considerable percentage of the studied cats had contact with L. infantum and that detection of cell-mediated immune response by measuring specific IFN-γ production provides a better estimation of exposure of cats to L. infantum in endemic areas as seen in dogs with different techniques [34,36].

Conclusions
As expected, cats from endemic areas produce IFN-γ after ex vivo blood stimulation with LSA and therefore are able to activate a cell-mediated adaptive immune response against the parasite that is variably associated with antibody or blood PCR positivity. The association of this assay to serological and molecular tests provides a better estimate of cat exposure to L. infantum.