Modeling Hemolytic-Uremic Syndrome: In-Depth Characterization of Distinct Murine Models Reflecting Different Features of Human Disease

Abstract

Diarrhea-positive hemolytic-uremic syndrome (HUS) is a renal disorder that results from infections with Shiga-toxin (Stx)-producing Escherichia coli. The aim of this study was to establish well-defined refined murine models of HUS that can serve as preclinical tools to elucidate molecular mechanisms of disease development. C57BL/6J mice were subjected to different doses of Stx2 purified from an E. coli O157:H7 patient isolate. Animals received 300 ng/kg Stx2 and were sacrificed on day 3 to establish an acute model with fast disease progression. Alternatively, mice received 25 ng/kg Stx2 on days 0, 3, and 6, and were sacrificed on day 7 to establish a subacute model with moderate disease progression. Indicated by a rise in hematocrit, we observed dehydration despite volume substitution in both models, which was less pronounced in mice that underwent the 7-day regime. Compared with sham-treated animals, mice subjected to Stx2 developed profound weight loss, kidney dysfunction (elevation of plasma urea, creatinine, and neutrophil gelatinase-associated lipocalin), kidney injury (tubular injury and loss of endothelial cells), thrombotic microangiopathy (arteriolar microthrombi), and hemolysis (elevation of plasma bilirubin, lactate dehydrogenase, and free hemoglobin). The degree of complement activation (C3c deposition), immune cell invasion (macrophages and T lymphocytes), apoptosis, and proliferation were significantly increased in kidneys of mice subjected to the 7-day but not in kidneys of mice subjected to the 3-day regime. However, glomerular and kidney volume remained mainly unchanged, as assessed by 3D analysis of whole mount kidneys using CD31 staining with light sheet fluorescence microscopy. Gene expression analysis of kidneys revealed a total of only 91 overlapping genes altered in both Stx2 models. In conclusion, we have developed two refined mouse models with different disease progression, both leading to hemolysis, thrombotic microangiopathy, and acute kidney dysfunction and damage as key clinical features of human HUS. While intrarenal changes (apoptosis, proliferation, complement deposition, and immune cell invasion) mainly contribute to the pathophysiology of the subacute model, prerenal pathomechanisms (hypovolemia) play a predominant role in the acute model. Both models allow the further study of the pathomechanisms of most aspects of human HUS and the testing of distinct novel treatment strategies.

Keywords: Escherichia coli; Shiga toxin; Shiga-toxin-producing Escherichia coli; acute kidney injury; enterohemorrhagic E. coli; hemolytic-uremic syndrome; mouse models.

Figure 1

Dose-response study for Stx2 in C57BL/6J mice. (A–F) Mice were subjected to different Stx2 concentrations. The effect was assessed after 72 h. Data are expressed as mean ± SD for n observations, *p < 0.05 vs. sham. (A) Final score and overall weight loss at 72 h (score: Kruskal–Wallis test with Dunn’s correction, weight loss: one-way ANOVA with Bonferroni post hoc test; sham n = 6, 50 ng/kg BW Stx2 n = 4, 100 ng/kg BW Stx2 n = 5, 150 ng/kg BW Stx2 n = 7, 300 ng/kg BW Stx2 n = 8). Kidney function was evaluated by (B) plasma urea and (C) plasma creatinine. (D) Hematocrit and (E) hemoglobin were measured as indicators of red blood cell count. Immune response was monitored by (F) white blood cell count. (B–F) Groups: sham n = 6, 50 ng/kg Stx2 n = 4, 100 ng/kg Stx2 n = 6, 150 ng/kg Stx2 n = 7, 300 ng/kg Stx2 n = 8 (one-way ANOVA with Bonferroni post hoc test).

Figure 2

Clinical presentation of C57BL/6J mice subjected to different Stx2 regimens. Clinical presentation was assessed by a score ranging from 1 = very active to 7 = dead for the (A) acute model (sham n = 6, Stx2 n = 11) and (B) subacute model (sham n = 8, Stx2 n = 10). (A,B) Data are expressed as median ± interquartile range for n observations. *p < 0.05 for sham vs. Stx2 at each respective time point (Mann–Whitney U-test). Weight loss of C57BL/6J wild-type mice was assessed every 24 h for the (C) acute model (sham n = 6, Stx2 n = 11) and the (D) subacute model (sham n = 8, Stx2 n = 10). (C,D) Data are expressed as mean ± SD for n observations. *p < 0.05 for sham vs. Stx2 at each respective time point (two-way ANOVA with Bonferroni post hoc test).

Figure 3

Hemoconcentration and thrombocytes in C57BL/6J mice subjected to different Stx2 regimens. Percentage change of (A) hematocrit, (B) hemoglobin, and (C) thrombocytes for the acute (sham n = 6, Stx2 n = 11) and subacute model (sham n = 5, Stx2 n = 6) compared with the respective sham group. (A–C) Data are expressed as mean ± SEM for n observations.

Figure 4

Indicators of hemolysis in C57BL/6J mice subjected to different Stx2 regimens. Plasma bilirubin levels were measured for the (A) acute model (sham n = 5, Stx2 n = 7) and (B) subacute model (sham n = 7, Stx2 n = 6). Plasma lactate dehydrogenase (LDH) levels were measured for the (C) acute model (sham n = 6, Stx2 n = 7) and (D) subacute model (sham n = 7, Stx2 n = 7). (A–D) Data are expressed as mean ± SD for n observations. *p < 0.05 sham vs. Stx2 (t-test). Hemolysis was photometrically quantified in plasma samples of the (E) acute model (sham n = 6, Stx2 n = 8) and (F) subacute model (sham n = 6, Stx2 n = 10). (E,F) Data are expressed as mean ± SD for n observations. *p < 0.05 for sham vs. Stx2 (Mann–Whitney U-test).

Figure 5

Immune response in C57BL/6J mice subjected to different Stx2 regimens. (A) Percentage change of leukocyte counts and (B,C) ratio of neutrophils to lymphocytes for the acute and subacute model compared with the respective sham group. (A–C) Data are expressed as mean ± SEM for n observations (acute: sham n = 6, Stx2 n = 11; subacute: sham n = 5, Stx2 n = 6). Representative images (scale bar 100 µm) of immunohistochemical detection and quantitative data of (D) F4-80 (surface antigen of macrophages) and (E) CD3 (surface antigen of T cells) in renal sections of C57BL/6J wild-type mice are depicted (acute: sham n = 4, Stx2 n = 9; subacute: sham n = 5, Stx2 n = 4). *p < 0.05 for sham (white dots) vs. Stx2 (black dots; t-test).

Figure 6

Indicators of kidney injury in mice subjected to different Stx2 regimens. (A,D) Plasma urea, (B,E) plasma creatinine, and (C,F) plasma neutrophil gelatinase-associated lipocalin (NGAL) levels were measured for the acute model (sham n = 6, Stx2 n = 8) and subacute model (sham n = 7, Stx2 n = 7, NGAL Stx2 n = 6). Data are expressed as mean ± SD for n observations. *p < 0.05 for sham vs. Stx2 (t-test). Representative images (scale bar 100 µm) and quantitative data of (G) periodic acid Schiff (PAS) staining and (H) immunohistochemical detection of kidney injury molecule-1 (KIM-1) in renal sections of C57BL/6J wild-type mice are depicted (acute: sham n = 4, Stx2 n = 9; subacute: sham n = 5, Stx2 n = 4). *p < 0.05 for sham (white dots) vs. Stx2 (black dots; Mann–Whitney U-test).

Figure 7

Indicators of cell death, reactive proliferation, and endothelial damage in kidney tissue of C57BL/6J mice subjected to different Stx2 regimens. Representative images of immunohistochemical detection and quantitative data of (A) cleaved caspase 3 (acute: sham n = 4, Stx2 n = 9; subacute: sham n = 5, Stx2 n = 3), (B) Ki67, and (C) CD31 (acute: sham n = 4, Stx2 n = 9; subacute: sham n = 5, Stx2 n = 4) in renal sections of C57BL/6J wild-type mice are depicted. *p < 0.05 for sham (white dots) vs. Stx2 (black dots; t-test).

Figure 8

Indicators of thrombus formation and complement activation in C57BL/6J mice subjected to different Stx2 regimens. (A) Representative images (scale bar 100 µm) of SFOG staining in renal sections. Thrombus formation was observed in Stx2 groups of both models (see zoomed areas on the right, scale bar 50 µm). (B) Representative images (scale bar 100 µm) of immunohistochemical detection and quantification of C3 and C3b complement deposition by C3c staining in renal sections of C57BL/6J wild-type mice are depicted (acute: sham n = 3, Stx2 n = 8; subacute: sham n = 5, Stx2 n = 4). *p < 0.05 for sham (white dots) vs. Stx2 (black dots; Mann–Whitney U-test).

Figure 9

Electron microscopic analysis of kidney tissue from C57BL/6J mice subjected to different Stx2 regimens. Representative ultrastructural images of renal tubules show (A) no pathological changes in sham animals, (B) prominent vacuoles (v) in the acute and (C) pronounced tubular atrophy (d = detached tubular cell) in the subacute model. Magnification: 2,000×; ultrathin sections, n = 2 animals were studied per group.

Figure 10

Changes in renal gene expression in response to different Stx2 regimens. (A) Heat map of features obtained from ANOVA filtering adjusted p < 0.05 comprising z-score scaled values. In each group, n = 4 animals were studied. (B) Proportional Venn diagram of differentially expressed genes according to the limit fold change for acute (dark gray) vs. subacute model (light gray). Number of overlapping genes is highlighted in white. Gene expression data were filtered for candidate genes from complement pathway. Log2 signals for (C) complement C3a receptor 1 (C3ar1), (D) complement factor C3 (C3), (E) complement C1q B chain (C1qb), and (F) tissue factor (F3) are illustrated by box plots.