Early antibody response and clinical outcome in experimental canine leishmaniasis

Infected dogs are the main reservoir of zoonotic visceral leishmaniasis, a widespread parasitic disease caused by Leishmania infantum. Therefore, the control of canine infections is required to reduce the incidence of human cases. Disease outcome in dogs depends on the fine balance between parasite virulence and efficacy of the immune system. Thus, knowledge of early response could yield relevant information for diagnosis and follow-up. In our study, 20 Beagle dogs were intravenously infected with 108 amastigotes of a fresh isolate of L. infantum and monitored along 16 weeks post inoculation. Specific antibody response and clinical evolution of infected animals were highly variable. Immunofluorescence antibody test (IFAT) and enzyme linked immunosorbent assay (ELISA) were useful to assess infection status, although only ELISA with promastigote-coated plates and, particularly, western blotting (WB) allowed an early diagnosis. Prominent antigens were identified by mass peptide fingerprinting. Chaperonin HSP60, 32 and 30 KDa antigens were recognized by all dogs on week 10 post infection. This suggests that these antigens may be valuable for early diagnosis. Advanced infection showed, in addition, reactivity to HSP83 and HSP70. Disease outcome did not show a clear relationship with ELISA or IFAT titers. Correlation between the clinical status and the combined reactivity to some antigens sustains their use for diagnosis and follow-up.

The main drawback of most cross-sectional studies relates to the potential cross-reactivity with other pathogens frequently coinfecting dogs (Babesia, Ehrlichia, Neospora, Toxoplasma) 33,34 and the lack of information on the actual time elapsed after inoculation. These limitations could be overcome by analyzing experimentally infected animals to determine early infection markers and, potentially, the value of the reactivity pattern of WB for clinical follow-up. Published longitudinal studies with experimentally infected dogs by L. infantum are hardly comparable due to the different infective doses and via of inoculation, age and breed of experimental dogs 35 . Most of them involved low numbers of animals [36][37][38][39][40] or the experiments did not include WB determinations 26,41-43 . In the course of an unrelated project, involving a considerable number of dogs experimentally infected with L. infantum, serial serum samples were obtained along infection. Humoral response of the animals was determined (IFAT, ELISA, WB) with the aim of identifying early infection markers, immune detection patterns, correlation between the diagnostic techniques and their relationship to the clinical status of the animals.

Results
Serum antibody response estimated by IFAT and ELISA. Female Beagle dogs (10-11 months old) were inoculated with 10 8 amastigotes of L. infantum freshly obtained from a naturally infected dog (n = 20) or kept as uninfected control animals (n = 4). Dogs were housed under controlled conditions precluding undesired arthropod-borne infections, daily observed and subjected to periodical clinical explorations and biochemical and immunological evaluations along 16 weeks post inoculation. Uninfected control dogs did not show any specific antibody response along the experiment. Inoculation of dogs with L. infantum elicited a time-dependent increase of IFAT titers along the infection and 5 weeks post infection (wpi) five animals were over the threshold titer (≥1/80) (Fig. 1); five weeks later (week 10 pi) the majority of the inoculated dogs (18 out of 20) were IFAT+ and 12 wpi all animals showed titers ≥1/160. Immune response was heterogeneous and on week 16 pi IFAT titers ranged from 1/320 to 1/2560. Specific response estimated by ELISA with soluble Leishmania antigen (ELISAsla) ( Fig. 2A) and ELISA using promastigotes as antigen (ELISAp) (Fig. 2B) showed a comparable pattern, all infected animals being positive by week 12. Despite individual variation, there was a strong correlation between both ELISA tests (r = 0.9376, P < 0.0001). IFAT values did correlate with ELISAp (r = 0.8632; P < 0.0001) and ELISAsla (r = 0.8487; P < 0.0001). ELISAp allowed an earlier diagnosis of L. infantum infection since 5 wpi the technique  Longitudinal study of western blotting pattern. A selection of dogs, representing the observed range of clinical presentations, was analyzed to determine the antigen recognition pattern at different post inoculation times (5, 7, 10 and 16 wpi) (Fig. 3). There was a notable individual variation, both in intensity of reactivity and immunodominant antigens recognized, despite the identical infective dose administered and the close genetic background of dogs. Sera from infected dogs showed extensive reactivity with antigens of MW ca. 93, 87, 85, 77, 72, 70, 66, 56, 50, 48, 46, 44, 41.5, 40, 38, 35, 32, 30, 28, 25.5, 23.5, 23, 21.5, 19.5, 17 and 15 KDa. Faint reactivities were found when testing the sera of the uninfected control animals, mainly on ~77 Da and >97 KDa (Fig. 3, Supporting Information Fig. 1).
Analysis of WB with ImageJ software allowed the quantitation (expressed as DU or density units) of the total reactivity as well as the time-course recognition of individual antigens along the infection progress (Supporting   (Table 2). Thus, these antigens could be employed, if not shared with other dogs' pathogens, for diagnosis and follow-up. Despite the limited analysis, total WB reactivity (DU) of dogs did correlate both with ELISAp (r = 0.9132; P < 0.0001) and ELISAsla (r = 0.805; P < 0.0001) (r > 0.84; P < 0.0001) (Supporting Information Fig. 4). The 2D electrophoretic separation of L. infantum SLA (Fig. 4A) and the recognition by serum from a chronically infected animal (     Fig. 3,D). Analysis of reactivity to immunodominant antigens showed that 30 KDa and 32 KDa slightly correlated with CS (r = 0.763; P = 0.0389 and r = 0.80; P = 0.025, respectively). However, combined reactivity of some antigens displayed a higher correlation, particularly when DU of 85 + 66-56 + 32 + 30 KDa antigens on week 16 pi was considered (r = 0.9092; P = 0.005) (Supporting Information Fig. 8).

Discussion
All inoculated dogs were infected and developed clinical signs and lesions compatible with leishmaniasis as well as a strong anti-Leishmania specific antibody response. This supports the model and experimental design used, including L. infantum strain, infective dose, leishmanial stage and via of inoculation. Individual variability of the animals along the infection, both in the clinical course and the immune response, is the rule in experimental canine leishmaniasis 26,[40][41][42][43][44][45] despite the close genetic background of experimental Beagle dogs.
IFAT is considered the golden standard technique for diagnosis of canine leishmaniasis in clinical practice 9,46 and its sensitivity and specificity, compared to ELISA, for diagnosis and follow-up purposes has been tes ted 26,39,41,46,47 . As regards the diagnostic value of different formats of ELISA including recombinant antigens (rK39, rK28) 26,45,48 and synthetic peptides 30 , it has been determined in cross-sectional and longitudinal studies. Results have shown that, in general, ELISA has higher sensitivity than IFAT 47 . Both in-house IFAT and ELISA tests (ELISAp and ELISAsla) were, under our conditions, adequate diagnostic procedures after 12 weeks of infection as reported previously 41 . However, follow-up of inoculated dogs showed that early phases of the infection were only detected by ELISAp, and this method was on average >2.5 fold more sensitive than standard ELISA and IFAT (7 wpi). ELISAp allowed, in our case, an earlier diagnosis than those reported with standard ELISA and rK39 immunochromatographic test in experimentally infected dogs (90-120 days pi) 26,42,45 . There was a good agreement between IFAT titration and ELISA results (κ > 0.6) at 10 wpi, but not earlier, and ELISAsla and ELISAp produced comparable results at that time (κ = 1.0). Moreover, IFAT test is time consuming and requires skilled personnel and expensive equipment. Since earlier diagnosis is an advantage for the clinical management of dog leishmaniasis, ELISAp could be a convenient diagnostic choice compared to standard ELISA and IFAT to detect acute L. infantum infections in dogs.
There is no global consensus on the WB recognition pattern by sera of L. infantum infected dogs. WB banding found in experimental infections [36][37][38][39]54 are simpler than those found in naturally infected dogs 39,51-53 although, in these cases, the possibility of coinfections with other antigen-sharing pathogens could not be ruled out. WB reactivities from our study, against immunodominant antigens of 85, 66, 56, 41.5, 32 and 30 KDa, were comparable to those obtained in experimentally infected dogs of the same breed 36 and in mixed breed animals 37 . Besides the diagnostic value of WB in established L. infantum infections, follow-up allowed the determination of time-related antigen recognition. Under our conditions, there was considerable variability among animals, and reactivity to the 56 KDa antigen and, less clearly, to the 32 and 30 KDa antigens was observed in all inoculated dogs only after 10 weeks of infection. It is possible that the apparent delay in the development of specific antibodies found, compared to previous reports 36,38 , would be related to the different experimental design, parasite strain and infective dose, individual immune response and methodology employed. Mass spectrometry allowed the identification of L. infantum antigens along the infection. Three of the immunodominant antigens were heat-shock proteins (HSP83, HSP70 and chaperonin HSP60), of poorly known biological functions 55,56 although significant immunogens in Leishmania infections 57 . HSP83 and HSP70 are recognized in rodent models and their simultaneous reactivity is considered a marker of visceral leishmaniasis 58 . These HSPs have been proposed as diagnostic antigens in dog leishmaniasis although HSP70 apparently cross reacts with Trypanosoma cruzi 59,60 this eventually leading to unspecific results in co-endemic areas. Diagnostic value of chaperonin HSP60 is less known although it has been reported to react with sera of dogs with subclinical natural infection 53 . Since, in our case, by week 10 pi L. infantum HSP60, 32 KDa and 30 KDa antigens were recognized by all inoculated dogs, their combination would allow an early diagnosis of canine infection. WB is not routinely employed in many diagnostic laboratories, but these antigens could be used in a dot-ELISA format, as suggested for human leishmaniasis 61 . Alternatively, epitope mapping could allow the construction of recombinant chimeric proteins. This approach has been followed with other proteins (PQ10, PQ20) in cross-sectional and a limited longitudinal study although the multiepitope-based ELISA required 4-5 months of infection to be positive 30,31 . Our results, and the present availability of recombinant Leishmania HSPs 56,62 and serial serum samples of experimentally infected dogs, could be used to confirm their diagnostic value under field conditions.
IFAT titration is frequently used in veterinary practice as a reliable method for monitoring the clinical evolution of L. infantum infected dogs, including their response after chemotherapy. Our results showed that, contrary to this assumption, IFAT did not show any significant correlation with the clinical status (CS) of the animals; therefore its value for disease follow-up and post treatment monitoring 44,63 should be reconsidered. However, it is worth indicating that results of IFAT are not lineal and serum titration was stopped at 1/2560 dilution whereas in ELISA actual OD values were considered. ELISA correlated better with the clinical status of animals than IFAT, although correlation was moderate. (r ca. 0.65). Several reports have associated WB patterns (IgG, IgG 1 , IgG 2 ) to the clinical status of naturally and experimentally infected dogs 36,38,40,51,52,64 and some antigens have been suggested as prognostic markers. In our study, combined reactivity to immunodominant antigens (HSP83, HSP70, HSP60, 32 KDa and 30 KDa) correlated with the clinical outcome. This suggests their potential value for both diagnosis and clinical follow-up and is consistent with the hyperglobulinemia found in dog leishmaniasis. Further research with accurate determination of immunoglobulin subclasses of dogs 65 along the L. infantum infection (2019) 9:18606 | https://doi.org/10.1038/s41598-019-55087-w www.nature.com/scientificreports www.nature.com/scientificreports/ course would clarify their role in the disease outcome. Whether these findings in experimental canine leishmaniasis, with a pure dog breed and intensively monitored animals, are also present in natural infections of different dog breeds, ages and management conditions needs further research under field conditions.

Material and Methods
Leishmania infantum strain. Inoculum was a fresh isolate of L. infantum obtained from the spleen of a naturally infected dog clinically and serologically diagnosed (Órgiva, Granada, Spain). After euthanasia, spleen was aseptically removed, and transported to our facilities under refrigeration. The organ was cut into small pieces (ca. 5 mm 3 ), and homogenized in a glass-in-glass tissue grinder (5 mL phosphate buffered saline, PBS). Suspensions were centrifuged twice (50 × g, 10 min; 1100 × g, 10 min, 4 °C). Cell pellets were treated for 30 seconds with cell lysis buffer (SDS 0.05%), resuspended in PBS and amastigotes counted in an improved Neubauer chamber. Isolation was performed ice-cooled under sterile conditions and amastigotes were kept at 4 °C and used to inoculate dogs after 24 h. The isolate was characterized using published kinetoplast primers 66  obtained from Envigo (France) when they were 4-5 months old and housed at the Faculty of Veterinary Medicine UCM (Madrid) (Animal facility Nr ES280790000091). Animal facilities were fitted with mosquito nets precluding the access of sand flies. Periodical complete physical exploration, biochemical, hematological and immunological evaluations showed physiological normality and negative IFAT test to L. infantum. When the animals reached 10-11 months age, 20 randomly selected animals were inoculated intravenously (cephalic vein) with 10 8 amastigotes of L. infantum/animal, administered in 1 mL. Four animals were kept as uninfected control dogs. After inoculation, dogs were daily observed, and every 2 weeks, weighed and subjected to complete clinical examination by a veterinarian blinded to the experimental design. Blood samples were obtained from the cephalic vein and routine immune response test (IFAT) was carried out by an external laboratory (Lab. Barba, Madrid). Dog sera were considered positive with IFAT titer ≥1/80. Infection status of inoculated animals was assessed on week 16 pi by popliteal lymph node sampling and microscope observation of amastigotes in stained smears (May Grünwald-Giemsa). Infected animals displayed a course-related range of clinical signs and lesions characteristic of leishmaniasis including lymph node enlargement, splenomegaly, skin lesions (e.g. erythema, alopecia), ocular lesions (e.g. conjunctivitis), paleness of mucosal membranes and muscular atrophy. Clinical status of the animals was quantified with a clinical score (CS) based on Manna et al. 68 and Foglia-Manzillo et al. 69 including clinical signs, lesions and hematological and biochemical abnormalities (maximum 35 points) (Supporting Information Table 2).
Antigen preparation. Promastigotes obtained by back transformation of amastigotes from the original isolate used for infection, were cultured in 175 cm 2 culture flasks at 27 °C in RPMI 1640 modified medium (BioWhittaker) supplemented with 10% heat-inactivated (30 min at 56 °C) fetal bovine serum (Gibco), 100 U/ mL penicillin plus 100 μg/mL streptomycin (BioWhittaker), 1% L-glutamine (BioWhittaker) and 1% human urine. To obtain SLA for ELISA and WB mid-log phase promastigotes were frozen at −80 °C, subjected to 5 freezing-and-thawing cycles (liquid nitrogen-water bath at 37 °C) and centrifuged at 18000 × g for 20 min at room temperature (RT). Supernatants were collected and protein concentration was determined with RC-DC Protein Assay (BioRad). For ELISAp, promastigotes (10 9 cells/mL) were fixed with 0.025% formaldehyde (Panreac) in PBS for 2 h at RT, counted in improved Neubauer chamber and used to coat microtiter plates.
2D Electrophoresis, 2D WB and peptides identification. First dimension (50 μg SLA) was run in parallel on two 3-11 NL pH gradient 8 cm strips (GE Healthcare) in Ettan IPGphor 3 IEF System (GE Healthcare) until 5 kVh. Second dimension was run on hand cast 10% acrylamide gels (BioRad gel caster and MiniProtean II chamber) at 100 V in Laemmli buffer. One gel was stained with colloidal Coomassie blue g250, and the other transferred (2019) 9:18606 | https://doi.org/10.1038/s41598-019-55087-w www.nature.com/scientificreports www.nature.com/scientificreports/ to a PVDF membrane (BioRad, Mini Trans-Blot cell) (150 V, 2 h at 4 °C) in Towbin buffer. WB was carried out with 1/50 diluted serum from a dog chronically infected with L. infantum. Peptides were selected in the stained gel by comparison with 2D WB reactivity, in-gel reduced and digested with trypsin 71 . Analysis of peptides was performed using a 4800 Plus MALDI TOF/TOF mass spectrometer analyzer (Applied Biosystems, MDS Sciex), at the Proteomics Unit of UCM, Madrid. Peptide mass fingerprint and peptide fragmentation spectra were combined and searched in MASCOT v2.3 (http://www.matrixscience.com) through Global protein Server software (Applied Biosystems) against NCBI database. Search was performed without taxonomy restriction and the following parameters: carbamidomethyl cysteine as fixed modification and oxidized methionine as variable modification; peptide mass tolerance, 80 ppm; one missed trypsin cleavage site allowed, and MS/MS fragments tolerance, 0.3 Da. In all protein identifications, the probability scores were greater than the score fixed by Mascot as significant (P < 0.05).
Statistical analysis. For statistical analysis, IFAT values were transformed (≤1/40 = 1; 1/80 = 2; 1/160 = 3; 1/320 = 4; 1/640 = 5; 1/1280 = 6; ≥2560 = 7). ELISA values were expressed as percentage (%) of the OD value found for 20 pooled sera of dogs naturally infected with L. infantum, serologically and parasitologically confirmed, obtained from the Clinical Services of the Faculty of Veterinary Medicine UCM (OD of samples/Average OD of positive control population x 100). Agreement of diagnostic techniques was determined with Cohen's Kappa index. Relationship between the different diagnostic techniques, as well as between CS and diagnostic techniques, were evaluated using the non-parametric Spearman correlation 72 . Differences between DU of individual antigens recognized in WB by infected and control dogs were determined with Mann-Whitney U non-parametric test. In all statistical analyses level of significance was set at P < 0.