Real-time in vivo imaging reveals localised Nrf2 stress responses associated with direct and metabolism-dependent drug toxicity

Abstract

The transcription factor Nrf2 coordinates an adaptive response to chemical and oxidative stress characterised by the upregulated expression of cytoprotective target genes. In order to understand the mechanistic relevance of Nrf2 as a marker of drug-induced stress it is important to know if this adaptive response is truly localised in the context of organ-specific drug toxicity. Here, we address this knowledge gap through real-time bioluminescence imaging of transgenic Nrf2-luciferase (Nrf2-luc) reporter mice following administration of the metabolism-dependent hepatotoxin acetaminophen (APAP) or the direct nephrotoxin cisplatin. We detected localised bioluminescence in the liver (APAP) and kidneys (cisplatin) in vivo and ex vivo, whilst qPCR, Taqman low-density array and immunoblot analysis of these tissues further revealed increases in the expression level of several endogenous Nrf2-regulated genes/proteins, including heme oxygenase 1 (Hmox1). Consistent with the toxic effects of APAP in the liver and cisplatin in the kidney, immunohistochemical analysis revealed the elevated expression of luciferase and Hmox1 in centrilobular hepatocytes and in tubular epithelial cells, respectively. In keeping with the role of reactive metabolite formation in APAP-induced chemical stress, both the hepatotoxicity and localised Nrf2-luc response were ameliorated by the cytochrome P450 inhibitor aminobenzotriazole. Together, these findings show that Nrf2 can reflect highly-localised cellular perturbations associated with relevant toxicological mechanisms.

Figure 1

Acetaminophen activates hepatic Nrf2 signalling in vivo. Wild type C57Bl/6 J mice (n = 5 per group) were administered saline or 300 mg/kg APAP. (A) Total hepatic GSH content and (B) serum ALT levels in mice at the indicated times post-APAP administration. (C) qPCR analysis of Nrf2 target genes in the livers of mice at the indicated times. (D) Immunoblot analysis of Hmox1 and Nqo1 in livers of mice (two representative samples per group) at the indicated times. (E) Densitometric analysis of Hmox1 and Nqo1 protein levels in livers of mice (five per group) at the indicated times. Data represent mean + S.D. Statistical analysis was performed with (A) one-way ANOVA (Tukey’s multiple comparison) or (B,C,E) a Kruskal-Wallis (Conover-Inman pairwise comparison) test; *P ≤ 0.05, **P ≤ 0.001, ***P ≤ 0.0001 versus 0 h. ^@P ≤ 0.001 versus 6 h saline.

Figure 2

Acetaminophen provokes a localised hepatic Nrf2 stress response. Nrf2-luc mice (n = 5 per group) were administered saline or 300 mg/kg APAP. (A) In vivo bioluminescence imaging of the same representative mice at the indicated times post-APAP administration. See Fig. S2 for imaging data for all mice. (B) Ex vivo bioluminescence imaging of livers and kidneys of the mice shown in A, 24 h post-APAP administration. (C) Luminescence signals from in vivo and ex vivo imaging of all mice. (D) Correlation of bioluminescence signals in livers imaged ex vivo and serum ALT levels in the same mice. (E) qPCR analysis of Nrf2 target genes in the livers of mice 24 h post-APAP administration. (F) Immunoblot analysis of Hmox1 in livers of mice at 24 h. Data represent mean + S.D. Statistical analysis was performed with a (C) Mann-Whitney U test, (D) Pearson’s R test or (E) unpaired t-test; *P ≤ 0.05, **P ≤ 0.001, ***P ≤ 0.0001 versus saline.

Figure 3

Histological analysis of the hepatic stress response to acetaminophen in Nrf2-luc mice. Nrf2-luc mice were administered saline or 300 mg/kg APAP. At 24 h, in saline-treated mice there were no histological changes (HE stain), Hmox1 expression was restricted to Kupffer cells and erythrocytes within sinuses, and staining for luciferase yielded only a non-specific serum reaction. In APAP-treated mice, there was extensive centrilobular coagulative necrosis with hydropic degeneration of surrounding hepatocytes (HE stain). Hmox1 was expressed by the necrotic centrilobular hepatocytes as well as individual intact hepatocytes adjacent to the affected area (arrow), whilst Kupffer cells close to affected areas also exhibited strong Hmox1 expression. Luciferase was expressed by the necrotic and degenerate centrilobular hepatocytes. CV: central vein; P: portal vein. Scale bars = 20 µm.

Figure 4

Cisplatin provokes a localised renal Nrf2 stress response. Nrf2-luc mice (n = 4 per group) were administered saline or 20 mg/kg cisplatin. (A) In vivo bioluminescence imaging of the same representative mice at the indicated times post-cisplatin administration. See Fig. S5 for imaging data for all mice. (B) Ex vivo bioluminescence imaging of kidneys and livers of the mice shown in A, 96 h post-cisplatin administration. (C) Luminescence signals from in vivo and ex vivo imaging of all mice. (D) Correlation of bioluminescence signals in kidneys imaged ex vivo and serum BUN levels in the same mice. (E) Immunoblot analysis of Hmox1 in kidneys of mice at 96 h. Data represent mean + S.D. Statistical analysis was performed with a (C) Mann-Whitney U test or (D) Pearson’s R test; *P ≤ 0.05 versus saline.

Figure 5

Histological analysis of the renal stress response to cisplatin in Nrf2-luc mice. Nrf2-luc mice were administered saline or 20 mg/kg cisplatin. At 96 h, in saline-treated mice there were no histological changes (HE stain), Hmox1 expression was restricted to intravascular erythrocytes, and staining for luciferase yielded only a non-specific serum reaction. In cisplatin-treated mice, proximal tubules exhibited attenuated epithelium (arrow) or necrosis and loss of epithelial cells (arrowhead), whilst lumina were often filled with protein casts (HE stain). Hmox1 and luciferase expression were detected in viable, degenerating proximal tubular epithelial cells (arrow) and within the proteinaceous material in the lumen of proximal tubules with necrotic epithelial cells (arrowhead). Scale bars = 20 µm.

Figure 6

The Nrf2 stress response to acetaminophen requires reactive metabolite formation. Nrf2-luc mice (n = 3 per group) were administered saline or 100 mg/kg ABT, then 1 h later administered saline or 300 mg/kg APAP. (A) Overview of study design, with times of drug administration, imaging and serum ALT measurements indicated. (B) Serum ALT levels in mice treated as indicated, 24 h post-APAP administration. (C) In vivo and (D) ex vivo bioluminescence imaging of the same representative mice, 24 h post-APAP administration. See Fig. S8 for imaging data for all mice. (E) Luminescence signals from in vivo and ex vivo imaging of all mice. (F) qPCR analysis of Nrf2 target genes in the livers of mice treated as indicated, 24 h post-APAP administration. (G) Immunoblot analysis of Hmox1 protein levels in livers of mice at 24 h. (H) Densitometric analysis of Hmox1 proteins levels in G. Data represent mean + S.D. Statistical analysis was performed with (B,H) one-way ANOVA (Tukey’s multiple comparison) or (E,F) a Kruskal-Wallis (Conover-Inman pairwise comparison) test; *P ≤ 0.05, **P ≤ 0.001, ***P ≤ 0.001 versus saline + saline or as indicated; NS, non-significant.

Figure 7

Histological analysis of the effect of aminobenzotriazole on the hepatic stress response to acetaminophen in Nrf2-luc mice. Nrf2-luc mice were administered saline or 100 mg/kg ABT, then 1 h later administered saline or 300 mg/kg APAP. At 24 h, in mice treated with saline + saline, there were no histological changes (HE stain), Hmox1 expression was restricted to Kupffer cells and erythrocytes within sinuses, and staining for luciferase yielded only a non-specific serum reaction. In mice treated with saline + APAP, there was extensive centrilobular coagulative necrosis with glycogen loss (confirmed by PAS reaction, data not shown) of surrounding hepatocytes (HE stain). Hmox1 was expressed by the necrotic centrilobular hepatocytes as well as proximate Kupffer cells. Luciferase was expressed by the necrotic and degenerate centrilobular hepatocytes. The livers of mice treated with ABT + saline showed features identical to those observed in mice treated with saline + saline (see above). In mice treated with ABT + APAP, histological changes (HE stain) were restricted to a slight condensation of centrilobular hepatocytes (equivalent of reduced glycogen content; PAS reaction not shown). Hmox1 expression was seen in random individual and occasional smaller aggregates of morphologically unaltered hepatocytes (arrows). Kupffer cells close to positive hepatocytes also showed enhanced expression of Hmox1 (small arrows). Luciferase expression was detected in random individual and occasional smaller aggregates of morphologically unaltered hepatocytes (arrows). CV: central vein; P: portal vein. Scale bars = 20 µm.