Infection of dogs by Leishmania infantum elicits a general response of IgG subclasses

Leishmania infantum is the etiological agent of zoonotic visceral leishmaniasis. In endemic areas, canine infections are considered the main source of infection for human populations. Therefore, any control of human leishmaniasis must include the control of canine infections. Chemotherapy of leishmaniasis is inadequate and canine immunoprophylaxis has important limitations. Reports on the response of infected dogs are abundant but no clear picture of immune events has emerged. To shed some light on these shortcomings the specific IgG subclass response was followed in 20 Beagle dogs experimentally infected with L. infantum using monoclonal antibodies (MAb) specific for canine IgG1, IgG2, IgG3 and IgG4, along with ELISA and flow cytometry. Results showed that parasitic infection elicits a general response of all IgG subclasses, with a predominant IgG1 response and without any evidence of IgG1/IgG2 dichotomy. These findings suggest that the inconsistent results reported previously could be related to the lack of specific reagents and not to the actual differences in the immune response of infected animals. Differential IgG subclass reactivity in ELISA and cytometry and the analysis of the reacting antigens could facilitate the diagnosis and prognosis of the disease and provide a useful tool for adequate therapeutics and vaccine development against leishmaniasis.


Scientific Reports
| (2020) 10:18826 | https://doi.org/10.1038/s41598-020-75569-6 www.nature.com/scientificreports/ Th2 polarization is far from clear 16 . In the case of L. infantum, dogs are also natural hosts for the trypanosomatid and the interest on the response elicited, both for diagnosis, prognosis or vaccination, has fueled research. Consequently, many reports on the cellular and humoral response and the cell types and cytokines involved have been published. Generally, resistance in canine leishmaniasis has been associated to upregulation of the Th1-cytokines (IFN-γ, TNF-α) with a significant cellular response; on the contrary, progression of the disease would be linked to a Th2-biased response (e.g. IL-4) with high production of IgG 17 . It is well established that leishmaniasis in dogs is characterized by the limited cellular response and the high levels of antibodies (Ab), particularly IgG [18][19][20][21][22][23] . However, reports on the IgG subclasses involved and their correlation to the clinical status of the infected animals are confusing. Thus, several studies have linked the active disease to a predominant IgG 1 response [24][25][26][27][28] whereas in other reports it has been associated to a significant IgG 2 29-36 or a mixed response 20,37-39 despite using comparable techniques and reagents.
Purification, identification and production of monoclonal antibodies (MAb) against the canine IgG subclasses (IgG 1 , IgG 2 , IgG 3 , IgG 4 ) [40][41][42] has allowed their comparison with marketed secondary antibodies (anti dog IgG 1 and IgG 2 ). Their use in the study on several dog diseases, including ZVL 43 , has highlighted the potential bias in the published reports on the IgG subclasses in ZVL 44 . Despite the elapsed time since those results were obtained (> 10 years), very few studies using MAb specific for canine IgG subclasses have been published [45][46][47][48] , whereas many contributions using commercial nonspecific immunological reagents have been published afterwards 28,33,35 . This is most critical since the inaccurate determination of the immune scenario (cytokines, cells and immunoglobulins, particularly IgG subclasses) in canine resistance, resilience or susceptibility is a strong shortcoming to develop vaccines or to assess the efficacy of chemotherapy or immunotherapy. In our case, the availability of sera from a large number of dogs subjected to an experimental infection with a recently isolated wild strain of L. infantum 49 allowed us to determine the dynamics of production of specific IgG subclasses and their relationship to the clinical course of the experimental animals.

Results
Specificity of MAb anti-IgG subclasses. Since no purified canine IgG subclasses were available, specificity of MAb (B6, E5, A3G4, A5) was assessed by capture ELISA. MAb A5 (anti-IgG 4 ) reacted only with the serum captured by itself as coating Ab, thus confirming its specificity for dog IgG 4 (Table 1). Preliminary trials performed with biotinylated MAb E5 (anti-IgG 2 ) yielded inconsistent results. Therefore, its determination required the use of peroxidase-labeled anti-mouse IgG 2b . With this approach E5 reacted with the serum captured by itself besides showing nonspecific recognition of anti-IgG 3 . The cross-reactivity observed was expected since both the revealing labeled Ab (E5, anti-IgG2) and the coating anti-dog IgG 3 MAb (A3G4) were of mouse IgG 2b isotype. Interestingly, A3G4 showed a strong self-recognition (anti-IgG 3 ) and lower reactivity with the serum captured by B6 (anti-IgG 1 ). MAb B6 (anti-IgG 1 ), employed as secondary (detection) Ab, slightly reacted when the coating MAb was itself (anti-IgG 1 ) whereas it strongly reacted with dog serum captured by A3G4 (anti-IgG 3 ). These results with capture ELISA confirmed that MAb recognized the four canine IgG subclasses. Therefore, they could be used to determine the dynamics of IgG subclasses. However, MAb displayed variable affinities and their value had to be confirmed in the indirect ELISA determinations in the infected dogs along the infection with L. infantum.
Experimental L. infantum infection in dog elicits the elevation of all IgG subclasses determined by ELISA. ELISA results indicated that inoculated Beagles showed an increase of all anti-L. infantum IgG subclasses related to the infection while the uninfected control dogs remained under the cut-off value during the experimental period. However, increase over pre-infection levels varied among IgG subclasses (Fig. 1A-D). Thus, while on week 7 post infection (pi), 30% (6/20) of the animals were already positive with total IgG, IgG 1 , the most abundant subclass was only elevated in 3/20 dogs and IgG 2 and IgG 3 levels were over the cut-off value in 5% and 10% of the animals, respectively. Specific IgG 4 response was delayed and it was detected only from week 10 pi (7/20, 35%) onwards. At this sampling time (week 10 pi) 95% of the dogs (19/20) were IgG and IgG 1 positive. Considering all the serum samples only 14 out of 20 infected dogs showed detectable levels of anti-Leishmania IgG 4, and less than half of the animals displayed positive IgG 3 reactions (40%, 8/20) (Fig. 2A). All dogs, except #24, were anti-Leishmania IgG 2 positive at some sampling along the infection. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Average values of total IgG and IgG subclasses increased in a time-dependent manner despite the individual variations (Fig. 2B). Specific anti-Leishmania IgG levels strongly correlated with IgG 1 (r = 0.9438), IgG 2 (r = 0.8558) and IgG 4 (r = 0.7152) (P < 0.0001) but not with IgG 3 (r = 0.2961; P = 0.0010). Accordingly, whereas IgG 1 , IgG 2 and IgG 4 were highly correlated (r = 0.7155-0.8441; P < 0.0001) no clear relationship emerged with IgG 3 (r = 0.250-0.269; P < 0.01). Considering the cumulative Ab response, estimated by the trapezoidal method (AUC), IgG levels up to week 16 pi strongly correlated with IgG 1 and IgG 2 but not with the other subclasses. IgG 2 correlated with IgG 1 and IgG 4 but not with IgG 3 ( Table 2).
Given the wide variation found in the immune response, levels of Ab were compared to the clinical score (CS) of the experimental dogs (Supporting Information Table 1). No correlation was found between Ab response and the clinical status in the early phases of the infection (5 weeks pi). However, IgG 1 and IgG 2 were positively  Table 2. Correlation (Spearman ρ value) between the cumulative Ab response of IgG subclasses along the experimental infection of dogs (16 weeks) with L. infantum. **P < 0.01; ***P < 0.001; ****P < 0.0001; ns: not significant. www.nature.com/scientificreports/ correlated with the value of CS of experimental Beagle dogs on week 10 pi, and IgG 1 also on week 16th. IgG 2 :IgG 1 and IgG 3 :IgG 1 ratios did not significantly vary with the clinical status of the animals but the IgG 4 :IgG 1 ratio showed a negative correlation with CS from week 10 pi onwards (Table 3).

Flow cytometry determination of specific IgG subclasses against L. infantum in experimentally infected dogs.
MAb allowed the detection of the four IgG subclasses produced against L. infantum using flow cytometry. All inoculated animals were IgG and IgG 1 positive (over cut-off value) at some moment of the infection whereas six dogs did not show any detectable specific IgG 2 and in one dog no specific IgG 3 and IgG 4 were found ( Fig. 1E-H). FACS-estimated response against L. infantum showed a steady time-dependent increase of specific IgG subclasses, besides total IgG. An early response of IgG 1 was found and by week 5 pi 55% (11/20) of the infected Beagles were positive, whereas response of other subclasses was low (IgG 3 and IgG 4 , 5% of the animals) or absent (IgG 2 ) in this sampling time (Fig. 2C). This late response of IgG 2 was also evident by week 7 pi, when 80% of the inoculated dogs were positive to IgG 1 and only 2 out of 20 animals had IgG 2 levels over the cut-off value. Contrary to findings in ELISA, relative increase of IgG 4 was higher than that from IgG 3 and IgG 2 (Fig. 2D). Moreover, in a similar way to that found with ELISA, the widest variation among inoculated dogs was observed in specific IgG 4 response; values in the last sampling of some dogs experienced > 200% increase. Concentration of IgG subclasses determined by flow cytometry were correlated among them and with total IgG (r = 0.8245-0.9405; P < 0.0001) but no relationship between their levels and the CS of the dogs was observed. FACS values were correlated with ELISA values, particularly for total IgG (r = 0.8427; P < 0.0001) and IgG 1 (r = 0.8399; P < 0.0001). Other subclasses also showed significant correlations (P < 0.0001) but with lower "r" values (IgG 2 : 0.7234; IgG 4 : 0.7001; IgG 3 : 0.6161).

Discussion
Inoculation of age-and sex-matched Beagle dogs with amastigotes of a fresh canine isolate of L. infantum elicited, under our conditions, an infection in all of them, developing clinical signs and lesions compatible with leishmaniasis. Progression of the disease was accompanied by a steady increase of all specific IgG subclasses determined (IgG 1 , IgG 2 , IgG 3 , IgG 4 ). These results do not support the polarized framework described in natural and experimental canine leishmaniasis with either a predominant IgG 1 24-28 or IgG 2 29-35,50 response in clinically affected dogs.
Several causes can account for the inconsistent findings on anti-Leishmania specific IgG 1 and IgG 2 response in ZVL, including the general absence of quantitative determinations of IgG and IgG subclasses and the variable experimental designs (e.g. infective dose, number of animals, epidemiological surveys), methodologies (e.g. ELISA, Western blot, antigen and antigen preparation). For sure these factors must play important roles but, in our view, at least two main reasons have hampered the determination of the relative value of IgG 1 :IgG 2 and its diagnostic and prognostic potential in ZVL. The first relates to the scarce available information on the real role of IgG subclasses in dogs and their relationship to that found in mice 14,44,51 or hamsters [52][53][54] . The second, is linked to the lack of specificity of immunological determinations which, in turn, are dependent on the specificity of the reagents. Mazza and coworkers characterized and developed MAb against canine IgG subclasses [40][41][42] . In our case purified canine IgG subclasses were not available, and the specificity of the MAb obtained from the hybridoma lines (B6: anti-IgG 1 ; E5: anti-IgG 2 ; A3G4: anti-IgG 3 ; A5: anti-IgG 4 ) could not be tested directly but rather by capture ELISA. A5 and E5 showed a good recognition of IgG 4 and IgG 2 , respectively. The cross reactivity found in capture ELISA between IgG 2 and IgG 3 was expected as anti-mouse IgG 2b was employed for the amplification to detect E5 (anti-IgG 2 , IgG 2b ). Cross reactivity between IgG 1 and IgG 3 with MAb B6, already reported, could be due to a prozone phenomenon or competition for epitopes. Serum IgG 1 concentration in healthy dogs is over 20 fold that of IgG 3 (8.17 ± 0.95 mg/mL vs. 0.36 ± 0.43 mg/mL), and quantitative assays suggested that there was no simultaneous reactivity 41 . We have not performed quantitative ELISA but no correlation (ρ) between the IgG 1 and IgG 3 response against L. infantum was observed in our study (see below). Thus, despite some shortcomings, the available MAb could be used to evaluate canine IgG subclasses along the experimental infection.
Our results show that the concentrations of all IgG subclasses were increased along the experimental infection with L. infantum in a time-dependent manner. This elevation of the four subclasses confirms and extends the results obtained with the same MAb by other authors 43,[45][46][47][48] . Response was variable among experimental dogs. IgG 1 and IgG 2 , the most abundant subclasses, were also the dominant ones reacting in ELISA with leishmanial soluble antigen (SLA). Up regulation of specific IgG 1 , IgG 4 and IgG 3 , compared to IgG 2 , has been reported in naturally infected dogs 43,46 . In our case, all inoculated animals developed clinical signs, lesions and biopathological Table 3. Correlation (Spearman ρ value) between serum concentration of IgG subclasses and ratios, and the clinical score (CS) of dogs infected with L. infantum, at different pi times. wpi: weeks post inoculation. ns: not significant. *P < 0.05; **P < 0.01; ***P < 0.001. www.nature.com/scientificreports/ alterations compatible with leishmaniasis and L. infantum infection was confirmed by biopsy. ELISA results showed that the highest relative increase over pre-infection values corresponded to specific IgG 1 (over 100% after 10 weeks of infection) followed by IgG 2 > IgG 3 > IgG 4 (Fig. 2B). The early increase of IgG 1 (95% positive on week 10 pi) ( Fig. 2A) suggests that, using ELISA and SLA as antigen, it could be a serum marker of active (recent) L. infantum infections whereas that of IgG 2 would indicate a more advanced infection. The IgG 4 :IgG 1 ratio was a good marker of the clinical course although the limited functional characterization of canine IgG subclasses 51 and the lack of stoichiometry in our indirect ELISA preclude any speculation before more studies are performed. The determination of IgG response to L. infantum using flow cytometry is less common, although some studies have been carried out 55 . Given the lack of subclass specificity of marketed Ab, the apparently different response found in vaccinated dogs with Leishvaccine (IgG 1 ) or Leishmune (IgG 2 ) must be taken cautiously 56 . Our results with FACS showed a general pi time-related increase of all subclasses and ELISA and flow cytometry values were correlated. This supports the value of the MAb for both assays and the consistency of the immune response, although IgG 1 was detected earlier by cytometry. Earlier IgG 1 response detected by FACS could be compatible with the recognition of membrane-bound antigen (Ag) and hence its potential value as infection marker. Interestingly, no correlation was found, at any sampling point, between the clinical score (CS) and IgG subclasses pattern. Contrarily to the ELISA results, the relative increase of IgG 4 along the follow-up was higher than that of IgG 2 and IgG 3 (Fig. 2D); IgG 4 values of experimental dogs (12 wpi) were also higher (ca. two fold) than those determined in dogs with chronic infection. Results seem factual and the possible significance (significance in acute phase, chronic phase) of the high relative levels of anti-Leishmania IgG 4 requires further research. The differential reactivity found in ELISA and FACS points towards variable sets of Ag being exposed and detected by the techniques and their degree of denaturalization (SLA vs. promastigotes). Scarce reactivity of IgG 2 with FACS, compared to ELISA, suggests that target Ag for IgG 2 response should be intracellular (soluble/ cytosolic). Further analyses with individual Ag (purified, WB) could confirm this hypothesis.
Results found in the follow-up of experimentally infected dogs with L. infantum support that the mousebased Th1 (IgG 2a : resistance)/Th2 (IgG 1 : susceptibility) framework does not represent the actual events in canine leishmaniasis. Marketed anti-IgG 1 and anti-IgG 2 are not specific of dog IgG subclasses 43,44,48 and it is very likely that variable and inconsistent results obtained with them 28,33,35,57 could be biased.
Our study was carried out under controlled conditions and dogs were exposed to a single infection whereas under field conditions polyparasitism (along with other pathogens) is the rule. Moreover, animals were genetically close, the infective dose was very high, with amastigotes inoculated by IV route. The experimental design was efficient achieving well established infections but these factors could affect the immune response elicited, particularly in the early phases of parasite dissemination. Therefore, further research is needed to confirm our findings in naturally-acquired infections and dog breeds.
Despite these limitations, we can conclude that MAb employed allowed the detection of specific canine IgG subclasses against L. infantum by ELISA or cytometry. Cross reactivity found between MAb recognizing IgG 1 and IgG 3 in capture ELISA possibly is not highly significant given their differential abundance in serum (ca. 20 fold) and the different patterns found for both subclasses along the experimental infection. The concentrations of all subclasses were increased after inoculation and IgG 1 was the predominant Ab and also the earliest to appear 45 . Differential subclasses pattern found with ELISA and FACS and the lack of correlation between CS and FACS-determined response could be related to the different antigen sets and epitopes detected. Careful analyses of these Ag could give significant clues on the role of IgG subclasses in infected animals. This is critical to get an accurate picture of the immune events in ZVL and could facilitate the diagnosis and prognosis of the disease as well as provide a useful tool for better therapeutics and vaccine development.

Material and methods
Leishmania infantum strain. L. infantum strain was obtained from the spleen of a naturally infected dog from southern Spain (Órgiva, Granada), clinically and serologically diagnosed. Details on the isolation of parasites to inoculate the dogs have been previously published 49 . Briefly, the spleen was cut into small pieces and homogenized in a glass-in-glass tissue grinder. Suspensions were centrifuged and cell pellets were treated with cell lysis buffer. Isolation was performed under sterile conditions and amastigotes were kept at 4 °C for 24 h and used to inoculate dogs. www.nature.com/scientificreports/ supplemented with 10% heat inactivated fetal calf serum, 1% l-Glutamine 200 mM (BioWhittaker, Lonza), and 100 U/mL penicillin plus 100 μg/mL streptomycin (BioWhittaker, Lonza). Several subcultures were performed to obtain supernatants. Isotyping of MAb was carried out by capture ELISA. Briefly, 96-well high binding microtiter plates (NUNC, Thermo Fisher) were coated with 3 μg/mL (50 µL/well) of goat polyclonal anti-mouse IgG (SouthernBiotech) in HCO 3 − /CO 3 − (0.1 M pH 9.7), overnight at 4 ºC. After blocking (2% BSA, 75 μL/well, 30 min, 37 ºC) and three washes with PBS-Tween, supernatants from hybridoma cultures (8 replica/hybridoma line) were added (50 μL/ well) and incubated at 37 ºC, 2 h. Plates were washed as above and peroxidase labeled antibody (Ab) against murine isotypes (SouthernBiotech) and light chains were added: anti-IgA, anti-IgM, anti-IgG 1 , anti-IgG 2a , anti-IgG 2b , anti-IgG 3 , anti-kappa and anti-lambda at 1/2000 dilution and incubated at room temperature (RT) for 30 min. Plates were extensively washed (5×) and color reaction was developed with 100 μL/well OPD (1 mg/ mL) in citrate-phosphate buffer (0.1 M citric acid-0.2 M sodium phosphate, pH 5.3) with hydrogen peroxide (1/1000). Reaction was stopped with H 2 SO 4 3 N (50 μL) and absorbance read at 492 nm with Varioskan LUX (Thermo Fisher Scientific). It was confirmed that hybridoma lines B6 and A5 secreted IgG 1 -kappa Ab whereas IgG 2b -kappa Ab were produced by lines E5 and A3G4. Anti-Leishmania IgG subclasses of experimentally infected dogs determined by ELISA. Procedures to propagate promastigotes by back transformation of amastigotes from the original isolate and to obtain soluble leishmanial antigen (SLA) have been described previously 49 . Briefly, promastigotes were cultured in RPMI 1640 medium, supplemented with antibiotic-antimycotic mixture (BioWhittaker), l-glutamine (Bio-Whittaker), 10% fetal calf serum (Gibco) and human urine (1%). SLA was obtained by centrifugation of mid-log phase promastigotes subjected to freezing-and-thawing cycles.
Levels of specific IgG and subclasses in the infected dogs were determined by indirect ELISA using as positive control the serum from a dog with an immunologically and parasitologically confirmed chronic natural infection by L. infantum; negative control serum came from a healthy dog. Microplates were coated with 10 μg/mL SLA (50 μL/well) overnight at 4 ºC. After blocking (2% BSA, 30 min), dog sera were added (50 μL/well) at 1/400 (total IgG), 1/100 (IgG 1 , IgG 3  No. 1090-05) was employed using the same conditions. Color was developed and OD read as above. Results were expressed as % of the positive control sera (OD of the sample/OD positive control × 100). Mean % OD + 3 standard deviations-SD-of the pre-infection values was used as the cut-off value. Relative increases (%) over pre-infection values were also estimated.

Determination of anti-Leishmania IgG subclasses by flow cytometry (FACS). The same negative
and positive controls used for ELISA were employed in the evaluation of total IgG and IgG subclasses. Mid-log phase promastigotes of the homologous isolate of L. infantum were diluted in PBS (5 × 10 8 cells/mL) and divided in 50 μL aliquots in 1.5 mL Eppendorf tubes. An equal volume of serum was added for each experimental animal and sampling time. Tubes were kept ice-cooled for 1 h and reaction was stopped by adding 1 mL/tube 1% formaldehyde in PBS. Contents of the eppendorf tubes was divided onto 5 tubes (200 μL/tube), replenished with BD FACSFlow buffer and centrifuged (15, www.nature.com/scientificreports/ Fragment Specific, Jackson ImmunoResearch, Cat. No. 115-545-008) was added at 1/500 (100 μL/tube). To determine total IgG, cells were resuspended with fluorescein labeled anti-IgG (FITC-conjugated AffiniPure Rabbit anti-Dog IgG H + L, Jackson ImmunoResearch, Cat. No. 304-095-003) (100 μL/tube) at 1/100. All tubes were incubated (in the dark, 30 min, RT), washed and their contents transferred onto 5 mL cytometry tubes (Corning Falcon) using FACSFlow buffer to get 0.5-1 mL/tube. Cellular analysis was performed with a flow cytometer BD FACSCalibur (Becton Dickinson) and the programme CellQuest Pro v4.0.2. Fluorescent cells were counted (10,000 events, FL-1 filter) and results, mean fluorescence intensity (MFI), were expressed as above (see ELISA).