T cells, adhesion molecules and modulation of apoptosis in visceral leishmaniasis glomerulonephritis

From a population of dogs presenting a positive serology for leishmaniasis during a survey by the Center for Control of Zoonosis of Teresina, Piauí, Brazil, performed from May 1996 through May 1998, 55 adult male and female dogs positive for anti-Leishmania antibodies were selected as previously described [14]. Briefly, the diagnosis of VL was confirmed by detecting Leishmania in smears of skin, spleen and popliteal lymph nodes, and/or culture of material from sternal bone marrow, spleen or popliteal lymph nodes. Eight dogs from the same endemic area without VL were used as controls. All Leishmania-infected dogs were routinely exterminated at the Center of Control of Zoonosis for the control of transmission of VL. The non-infected animals used as control in this study were street dogs collected to be exterminated for rabies control. Specimen sampling and euthanasia of the animals was performed under general anaesthesia using 25 mg/kg i.v. thiopental sodium (Sigma-Aldrich, USA) [23]. The kidneys were removed, renal tissues were fixed in 0.01 M, pH 7.4 phosphate-buffered 10% formalin and embedded in paraffin, and 3 μm thick sections of kidney were prepared and submitted to immunohistochemical staining and apoptosis analysis. All histological analysis was blind and done by two independent observers. The experimental protocol used in this study was approved by the Ethics Committees of all institutions involved in the study.

Formalin-fixed and paraffin-embedded kidney sections were deparaffinized in xylene, rehydrated in decreasing alcohol concentrations, and incubated with 0.03% hydrogen peroxide in methanol solution for 30 minutes in the dark to block endogenous peroxidase activity. Antigen retrieval was performed using 1.2 mg/ml Tris-HCl, pH 1.0, in a microwave oven (Sanyo, Brazil) on maximum power, in consecutive cycles of 10 and 5 minutes. After washing in 0.01 M phosphate-buffered saline, pH 7.2 (PBS), the sections were treated using a Blocking Kit (Vector Laboratories, Inc., Burlingame, USA), and a protein block (Dako Corporation). The tissues were then incubated overnight at 4°C in a humid atmosphere with the different antibodies diluted in PBS: mouse, polyclonal, anti-Leishmania amazonensis antibody [14], diluted 1:1600 (vol:vol); mouse, monoclonal, anti-canine CD4 (VMRD, cod DH29A, Pullman, USA) and CD8 (VMRD, cod CAD46A, Pullman, USA) antibodies, diluted 1:500 (vol:vol); goat, polyclonal anti-canine IgG, IgA, IgM and C3b antibodies (10 μg/ml) (Bethyl laboratories, Montgomery, USA); mouse, monoclonal, anti-canine ICAM-1 and anti- canine P-Selectin antibodies (kindly provided by Professor C. Wayne Smith, Baylor College of Medicine, Houston, Texas, U.S.A.) (10 μg/ml); goat, polyclonal, anti-human TNF-α (10 μg/ml) (cod-sc-1347, Santa Cruz Biotecnology Corporation, California, USA); mouse, monoclonal, anti-human IL-1α (10 μg/ml) (cod-sc-9983, Santa Cruz Biotecnology Corporation, California, USA); mouse, monoclonal, anti-Ki-67 (clone MiB-1, diluted 1:75 vol:vol) (code M 7240, Dako Corporation, USA); and mouse, monoclonal, anti-M30 CytoDeath antibody diluted 1:50 (vol:vol), (cat 2140349, Roche, Mannheim, Germany). When mouse antibody was used, the reaction proceeded using catalyzed signal amplification (CSA) system-peroxidase (Dako Corporation, code K 1500, Carpinteria, USA) following protocols provided by the manufacturer. When goat and rabbit antibodies were used the reaction proceeded using streptavidine-peroxidase system (Dako Corporation, cod K 1500, Carpinteria, USA). After each incubation step, the sections were washed three times in PBS. The reaction was developed using 0.06% hydrogen peroxide and 0.3 mg/ml 3,3'-diaminobenzidine (Sigma Chemical, USA) in PBS. Counterstaining was performed using Harry's haematoxylin (Sigma Chemical, USA).

A specific kit for apoptosis detection (Boehringer Manheimm, Germany) was used with the tissue sections, and the assay was performed following the protocols provided by the manufacturer. Formalin-fixed and paraffin-embedded sections were deparaffinized, hydrated and the endogenous peroxides blocked as stated above. The sections were washed in PBS, incubated sequentially with 0.1% Triton X-100 (Merck; Darmastadt, Germany) in 0.1% sodium citrate for 2 minutes on ice, with 20 μg/ml Proteinase-K in PBS for 15 min at 37°C, with 3% bovine serum albumin and 20% foetal bovine serum (Cultilab, Brazil) in PBS for 30 minutes, and then with the TUNEL mix [terminal deoxynucleotidyl transferase (TdT) and fluorescein isothiocyanate (FITC)-conjugated dUTP] in humidified chamber for 60 min at 37°C. The reaction proceeded with incubation with horse-radish peroxidase-conjugated anti-FITC antibody Fab fragment for 30 min at 37°C, and the reaction developed using 0.06% hydrogen peroxide and 0.3 mg/ml 3,3'- diaminobenzidine tetrahydrochloride (Sigma Chemical, USA) in PBS, and counterstained with Harris' hematoxylin. After each incubation step, the sections were washed three times in PBS. A negative control was performed omitting TdT in the reaction. As a positive control, the section was incubated with 1 mg/ml Deoxyribonuclease I (Gibco BRL, USA) in 50 mM Tris-HCl pH 7.5, 1 mM MgCl2, 1 mg/ml BSA for 10 minutes at room temperature.

Morphometric analysis were performed on selected sections stained for distinct markers using an automatic image analyser employing Bioscan Optimas software (Optimas, Edmonds, CA, USA, Version 4.10) on a total of 50 glomeruli per animal; in a minority of samples with not enough glomeruli, cells were counted in at least 20 glomeruli. The following parameters were evaluated: cells positive for Leishmania antigen, CD4⁺ and CD8⁺ T cells, apoptotic markers M30 and TUNEL, proliferative marker Ki-67 and cells expressing TNF-α and IL-1α.

The morphometric parameters were analysed using the Kruskal-Wallis and Dunnett's or Dunn tests to compare multiple groups, and the Mann-Whitney or Student t-test to compare two groups, employing Sigma Stat software (Jandel Corporation, USA). The semi-quantitative parameters were analysed by One-way analysis of variance and Newman-Keuls tests for the comparison of multiple groups, using GraphPad Prisma V.3 statistical software (USA).

Immunoglobulins IgG, IgM and IgA and C₃b were probed in 26 infected dogs, and in 5 non-infected, control dogs. The antigens were present in all the patterns of glomerulonephritis, and in the non-infected, control dogs. Semi-quantitative analysis of immunoglobulins and C₃b deposits in the glomerular capillary wall showed no significant difference when a group of infected dogs was compared to the non-infected, control group (Figure 1).

In dogs with VL, Leishmania antigen (Figure 2A) was detected in glomerular cells in 54 (98%) of 55 infected dogs. It was absent in one infected dog presenting chronic glomerulonephritis and in all eight non-infected control dogs from the same area (Figure 2D).

Of the 55 dogs with VL, CD4⁺ T cells (Figure 2B) were observed in the glomeruli in 44 cases (80%), and CD8⁺ T cells (Figure 2C) in 17 cases (31%), but were both absent/scarse in the non-infected control dogs (Figure 2E and 2F). In all cases exhibiting CD8⁺ T cells, CD4⁺ T cells were also present. CD4⁺ and CD8⁺ T cells were present in all patterns of glomerulonephritis except chronic glomerulonephritis.

CD4⁺ T and CD8⁺ T cells and cells stained for Leishmania antigen were quantified in part of samples showing different patterns of glomerulonephritis: focal segmental glomerulosclerosis (N = 8), mesangial proliferative glomerulonephritis (N = 8), membranoproliferative glomerulonephritis (N = 8), and minor glomerular abnormalities (N = 8). Significantly more T CD4⁺ cells were observed in all infected dogs when compared with non-infected animals. However no significant differences were observed among different patterns of GN in infected dogs (Kruskal Wallis and Dunnett's tests) (Figure 3A). CD8⁺ T cells tend to have more in infected than in non-infected dogs but the difference was not significant (Fig. 3B) Additionally, cells stained for Leishmania antigen were observed only in samples from infected animals independent of the pattern of glomerular alteration, even in minor glomerular abnormalities (Kruskal Wallis and Dunnett's tests) (Figure 3C). There was a positive correlation between the number of cells stained for Leishmania antigen and CD4⁺ T cells (R = 0.57, p < 0.001, Spearman test) (Figure 3D), and between the number of CD4⁺ and CD8⁺ T cells (R = 0.40, p < 0.001, Spearman test) (Figure 3E) in the glomeruli, but no correlation was found between the cells stained for Leishmania antigen and CD8⁺ T cells (data not shown).

Adhesion molecules were probed in 20 infected and 5 non-infected control animals. ICAM-1 and P-selectin were present in all. In the control animals ICAM-1 was absent in only one animal and in the other control animals it was present in minimal intensity, differently what was observed in the infected animals. P-selectin was found in one control animal. Both ICAM-1 (Figure 4A and 4C) and P-selectin (Figure 4B and 4D) were localized in the endothelial lining of the glomerular capillaries, the mesangium and Bowman's capsule. In all patterns of glomerulonephritis the expression of ICAM-1 was from minimum to moderate or severe intensity and P-selectin from absent to severe intensity. In the control animals the expression of these molecules ranged from absent to medium intensity. It was not observed any correlation between severity of disease and these markers. Where ICAM-1 and P-selectin were present, primarily CD4⁺ T cells were also detected, except in two cases of focal segmental glomerulosclerosis and in one case of membranoproliferative glomerulonephritis.

We analysed proliferation and apoptosis in samples from 28 dogs with VL and 7 non-infected control dogs.

Proliferative marker Ki-67 (Mib-1) antigen was detected in interstitial inflammatory infiltrate in some areas close to glomeruli but it was absent in glomerular cells both in dogs with and without VL (data not shown).

Apoptosis was detected in tissue samples using two different methods: detection of the M30 cytodeath marker (Figures 5A and 5B) and the TUNEL method (Figures 5C and 5D). Apoptosis was observed in glomeruli in all 28 dogs with VL as well as in seven control animals without Leishmania infection. However, fewer apoptotic cells were found in infected animals than in control animals (Figures 6A, B and 6C). The results using these two methods were very similar.

Apoptosis of glomerular cells was quantified in samples submitted to detection of the M30 cytodeath marker. Significantly fewer cells stained for the M30 marker were detected in all patterns of glomerulonephritis when compared with non-infected control samples (p < 0.05, ANOVA and Newman-Keuls tests) (Figure 6A). There was a negative correlation between the number of Leishmania antigen⁺ cells and the number of cells stained for the M30 cytodeath marker (R = -0.35, p < 0.001, Spearman test) (Figure 6B). There was also a negative correlation between the total number of cells per glomerulus and the number of cells stained for the M30 cytodeath marker (R = - 0.47, p < 0.001, Spearman test) (Figure 6C).

Apoptosis-related cytokines were investigated and TNF-α was detected in the endothelial lining of the glomerular capillaries, in the mesangium, and on mononuclear cells in glomeruli. TNF-α was expressed in all animals studied. The expression of TNF-α in 28 animals with naturally acquired VL was less than in the seven non-infected control animals (Figures 7A and 7B). IL-1α was detected in the endothelial lining of the glomerular capillaries, mesangium and in mononuclear cells in glomerulus (Figures 7C and 7D). However the intensity was similar in the 28 VL dogs when compared with the seven non-infected control dogs. When cells expressing TNF-α were quantified, fewer cells were detected in samples of VL dogs presenting different patterns of glomerulonephritis than in the non-infected control animals (p < 0.05, (Kruskal Wallis and Dunn's tests) (Figure 8A). Furthermore, there was a positive correlation between the number of cells expressing TNF-α and M30 cytodeath marker (R = - 0.44, p < 0.001, Spearman test) (Figure 8B).