Fig 1.
Effects of oral antibiotic treatment on outcomes in a mouse model of non-progressing visceral leishmaniasis.
(A) Experimental design; (B) qPCR analysis of bacterial 16S rDNA content in mouse faecal pellets at different times before (n = 21) and during oral antibiotic regimens (n = 3–9), including a germ-free control (n = 1). Lines show medians and dashed line shows the detection limit (mean + 2 standard deviations of negative extraction controls); (C, D, E, F) Caecum (C), spleen (D) and liver (E) weights, and parasite loads (F) measured at necropsy at the indicated times after infection with Leishmania donovani strain 1S (Ld1S) and maintained on drinking water containing antibiotic cocktail (ABX) compared with naïve controls and no ABX controls, dpi, days post-infection, n = 2–5 per group.
Fig 2.
Oral antibiotic treated hamsters have delayed VL disease progression and reduced mortality.
(A) Experimental design; (B) qPCR analysis of bacterial 16S rDNA content in hamster faecal pellets at different times (x-axis) before and during oral antibiotic treatment (ABX); Leishmania donovani (Ld) infections were initiated at 4 weeks of ABX treatment. Lines show medians. (C) Representative images of viscera at humane end-point and caecum weights at 12 weeks post-infection in Leishmania donovani (Ld)-infected hamsters maintained on drinking water containing antibiotic cocktail (ABX) compared with naïve controls and no ABX controls; arrows indicate caecum. (D) L. donovani whole cell lysate-reactive IgG antibody titres measure by ELISA (OD, optical density), compared by one-way ANOVA. (E,F) Weight change over time, data shown for means (E) and individual animals (F) where dashed red line marks humane end-point. (G) Survival curve for Ld-infected hamsters treated with ABX or not, Gehan-Breslow-Wilcoxon test p = 0.019. (H) Pairwise comparison and Pearson correlation tests of individual animal’s caecum weight and post-infection change in body weight. Experiments using VL onset end point (B, C, D, E, H) n = 5–16 per group; experiments using humane/maximum end-points (F, G), n = 10 per group.
Fig 3.
Prolonged survival in antibiotic treated hamsters is associated with less severe splenomegaly despite equivalent or higher parasite loads.
(A) Spleen weights of hamsters at 12 weeks post-infection (p.i.) with L. donovani and given drinking water containing antibiotic cocktail (ABX) compared with uninfected naïve and untreated controls; p-values are from one way ANOVA. (B) Pairwise comparison and Pearson correlation test of splenomegaly severity and change in body weight (proxy for disease progression) at 12 weeks p.i. (C) Pairwise comparison between splenomegaly severity and time to humane end-point (fatal progression); p-value is from comparison of linear regression slope elevations. (D,E) Total parasite loads (D) and infection intensities (E) in spleens after 12 weeks of infection, measured by limiting dilution ex vivo culture assay. (F) Pairwise comparison and Pearson correlation test of spleen parasite densities and change in body weight (proxy for disease progression) at 12 weeks p.i. (G) Pairwise comparison between spleen parasite density and time to humane end-point (fatal progression); p-value is from comparison of linear regression slope elevations. In B-G data points are for individual hamsters treated with oral antibiotics (filled circles) and untreated controls (open circles). Analysis using VL onset end point (A,D,E), n = 5–27 per group; analysis using humane/maximum end-points (C, G), n = 9–10 per group; analysis using all endpoints (B,F), n = 22–25 per group.
Fig 4.
Prolonged survival in antibiotic treated hamsters is associated with less severe hepatomegaly despite equivalent parasite loads.
(A) Liver weights of hamsters at 12 weeks post-infection (p.i.) with L. donovani and given drinking water containing antibiotic cocktail (ABX) compared with uninfected naïve and untreated controls; p-values are from one way ANOVA. (B) Pairwise comparison and Pearson correlation test of hepatomegaly severity and change in body weight (proxy for disease progression) at 12 weeks p.i. (C) Pairwise comparison between hepatomegaly severity and time to humane end-point (fatal progression); p-value is from comparison of linear regression slope elevations. (D,E) Total parasite loads (D) and infection intensities (E) in livers after 12 weeks of infection, measured by limiting dilution ex vivo culture assay. (F) Pairwise comparison and Pearson correlation test of liver parasite densities and change in body weight (proxy for disease progression) at 12 weeks p.i. (G) Pairwise comparison between liver parasite density and time to humane end-point (fatal progression). In B-G data points are for individual hamsters treated with oral antibiotics (filled circles) and untreated controls (open circles). Analysis using VL onset end point (A,D,E), n = 5–27 per group; analysis using humane/maximum end-points (C, G), n = 9–10 per group; analysis using all endpoints (B,F), n = 22–25 per group.
Fig 5.
Leukocyte profiles of dysbiotic hamsters infected with L. donovani.
Hamsters were maintained on drinking water with or without antibiotic cocktail (ABX) and infected with L. donovani (n = 15–19) and compared with naïve (Na) controls (n = 4–5). Data show individual component assays from complete blood counts at 12 weeks post-infection. Groups were compared by one way ANOVA and differences at p < 0.05 are indicated.
Fig 6.
Anaemia and iron sequestration in dysbiotic hamsters infected with L. donovani.
Hamsters were maintained on drinking water with or without antibiotic cocktail (ABX) and infected with L. donovani (n = 15–19) and compared with naïve (Na) controls (n = 4–5). The erythrocytic compartment was analysed with respect to total red cell count (A), haemoglobin levels (B), haematocrit (C) and mean corpuscular volume (D). (E) Representative images and quantification of ferric iron content in liver sections stained with Prussian blue (n = 4–12), magnification 200X, scale bars 50 μm. All data are from 12 weeks post-infection, with mean +SEM and individual hamster data points shown. Groups were compared by one way ANOVA (A-D) or Kruskal-Wallis test (E) and differences at p < 0.05 are indicated.
Fig 7.
Gene expression analysis in dysbiotic hamsters infected with L. donovani.
Hamsters were maintained on drinking water with or without antibiotic cocktail (ABX) and infected with L. donovani (n = 6) or kept naïve (n = 1). At 12 weeks post-infection, RNA transcript abundance for the indicated genes in spleen and liver tissue was measured by RT-qPCR with normalisation to Rpl18. Data are mean +SEM log2-transformed fold change in gene expression compared to untreated, uninfected control hamsters (n = 4). # indicates transcript abundance was below detection limit. Differences are indicated when p < 0.05 between groups in t-test.
Fig 8.
The GI tract is a site of chronic L. donovani infection in hamsters but not mice.
(A, C) qPCR analysis of L. donovani parasite loads in tissues from hamsters (A) and mice (C). Parallel reactions for a parasite-specific target (kDNA minicircle) and a host-specific gene (Actbl2) were performed on DNA extracted from the indicated tissues at 84 days post-infection (dpi) for hamsters (n = 6 per group) and 21 and 84 dpi for mice (n = 3–5 per group). Lines show medians and dashed line shows the detection limit (mean + 2 standard deviations of uninfected control samples, n = 4). (B) Histological sections for hamster small intestine (left, centre) and mesenteric lymph node (right) stained with haematoxylin and eosin; arrows indicate amastigote forms of L. donovani; magnification 400X, scale bars 20 μm.
Fig 9.
Evidence of bacterial co-infection in hamster visceral leishmaniasis.
(A) Naïve or L. donovani (Ld) infected hamsters maintained on drinking water with or without antibiotic cocktail (ABX) were fed fluorescently-labelled 4 kDa dextran at 12 weeks post-infection (wpi). After 4 hours the intensity of fluorescence in serum samples was measured as an indicator of intestinal barrier leakiness. (B) Histological sections stained with haematoxylin and eosin showing presence of Michaelis-Gutmann bodies (MGBs, white arrows) in L. donovani infected liver and duodenal lymph node tissue, black arrow indicates intracellular amastigotes. (C) Quantification of MGB abundance at 12 wpi. (D) Histological liver sections stained with haematoxylin and eosin (left), alizarin red for calcium detection (centre) and immunohistochemical (IHC) detection of bacterial endotoxin (right), white arrows indicated MGBs, black arrow indicates intracellular amastigotes, asterisks indicate anti-endotoxin IgG-reactive areas. Image magnifications and scale bars are 400X and 20 μm (C), 12.5X and 1 mm (D, H&E, upper), 40X and 200 μm (D, H&E, upper) 100X and 500 μm (alizarin), 400X and 30 μm (IHC). (E) Quantification of anti-endotoxin IgG-reactive areas at 12 wpi. (F,G) Pairwise comparison and Pearson correlation tests of liver endotoxin content and peripheral blood lymphocyte frequency (F, n = 10) and liver inflammation scores (G, n = 28). (H) Quantification of unique bacterial genera present in liver tissue assessed by 16S rDNA sequencing. Data are for male hamsters. Data on charts (A,C,E,H) are means +SEM, compared by one-way ANOVA, n = 3–12.
Table 1.
Frequency of live bacterial genera isolated from hamster liver homogenates.