Mitochondrial-triggered immune responses mechanistically connect drug-induced steatohepatitis and cardiomyopathy associated with nonalcoholic steatohepatitis

In our prior publication [1], we elucidated IL-33-dependent mechanisms that regulate drug-induced liver injury. In this commentary, we mechanistically connected IL-33-dependent drug-induced steatohepatitis with nonalcoholic steatohepatitis. Drug-induced steatohepatitis (DISH) is a rare form of drug-induced liver injury caused by drugs that are capable of inducing metabolic injury, steatosis or steatohepatitis [2]. DISH is distinct from nonalcoholic steatohepatitis (NASH) because the inciting agent is known and the disease generally improves upon drug withdrawal. Although <2% of NASH cases have been attributed to DISH, evidence suggests that DISH may share pathogenic mechanisms with NASH.

NASH is an advanced form of nonalcoholic fatty liver disease (NAFLD) and is a systemic metabolic disease that is histologically characterized by male sex, steatosis, hepatitis, hepatocyte ballooning and glycogen deposition. The most common causes of death in male patients with NASH and NAFLD are cardiomyopathy, heart failure and hypertrophy [3]. However, the exact mechanisms associating NASH or NAFLD with cardiac disease have not been completely elucidated. NAFLD is the most common cause of chronic liver disease and liver transplantation worldwide [4]. Thus, successful management of NAFLD/NASH is both economically and socially significant.

Mitochondrial injury, which is a major mechanism of drug-induced liver injury, requires cytochrome P450-derived reactive metabolites or the parent drug [5]. Mitochondrial injury and subsequent dysregulation promote the progression from NAFLD to NASH via the overproduction of reactive oxygen species through CYP2E1, increased levels of mitochondrial fatty acid oxidation triggered by cytokines such as IL-6 and reduced fatty acid clearance [6]. We developed a drug-induced model of steatohepatitis in mice using an epitope of mitochondrial CYP2E1 that induces oxidative stress via the inhibition of mitochondrial complex I [7]. We followed steatohepatitis and cardiomyopathy development for 15 weeks. Our results suggest that immune responses to a mitochondrial epitope in a susceptible host (IL-33−/− mice) induce steatohepatitis and cardiomyopathy reminiscent of NAFLD, NASH and associated cardiomyopathy (Fig. 1, Supplementary Figs. 1 and 2) and recapitulate the increased cardiovascular risk noted in human subjects. Specifically, this epitope induces proinflammatory signals (Il1b, Il17, and Il22), triggering CD4 + T cells (CD4) that likely promote Il6 and fatty acid synthesis and reducing fatty acid clearance, as well as heart failure (Bnp) and cardiomyopathy (Myh7, lox) (Supplementary Fig. 1A). We propose that this model provides an opportunity to discover the relationships between NAFLD/NASH, cardiac hypertrophy and cardiomyopathy and decipher the mechanism(s) associated with NASH and DISH.

Fig. 1.

Development of steatohepatitis and cardiomyopathy using a drug-induced murine model. A Sex differences in weight were seen within the experimental groups (immunized (IMM)-males (M, N = 20) vs. females (F, N = 22, p < 0.0001); unimmunized (UI)-M (N = 9) vs. F (N = 10) (p > 0.01) but not between IMM and UI IL-33−/− M (34.9 ± 5.2 g vs. 33.8 ± 4.0 g) and BALB/c IMM-M (30.5 ± 2.9 g, N = 4), or IMM and UI IL-33−/−F and BALB/c F. (26.7 ± 2.2 g vs. 27.4 ± 2.1 g vs. 24.5 ± 1.7 g). B Splenic proinflammatory signals (Il-1b, Il17, Il-22) and cellular (Cd4) mRNA were significantly upregulated in IL-33−/− IMM-M (N = 7) and F (N = 5) compared to UI (N = 3 and N = 5, respectively). C Representative hematoxylin and eosin (H&E), trichome and periodic acid Schiff (PAS) staining of liver sections (5 μm) demonstrated steatosis/steatohepatitis in IL-33−/− IMM-M and hepatitis in IL-33−/− IMM-F; minimal fibrosis but macrocytic steatosis in M compared to F; and glycogen deposition and ballooning in M compared to F. 40× magnification (inset 160×). D NASH CRN scoring using the methods described by Puri and Sanyal demonstrated higher severity in IL-33−/− IMM-M (N = 13) than in IL-33−/− IMM-F (N = 15) (p < 0.01) and IL-33−/− UI-M (N = 4, p < 0.01). IL-33−/− IMM-F severity scores were higher than IL-33−/− UI-F severity scores (N = 5, p < 0.001). E Hepatic Il-6 expression was upregulated in IL-33−/− IMM-M (N = 7, p < 0.05) compared to UI mice (N = 3). F Total liver cholesterol levels were similar in all groups (Abcam Cholesterol Assay Kit). G Targeted proteomics analysis of a single mouse liver from each treatment group showed upregulated cellular remodeling (Krt13, Col1a2) in IL-33−/− IMM-M compared to F (13.404, 3.151) and UI-M (9.669, 2.094), as well as fatty acid metabolism (Cyp4a12a (13.364 vs. F; 1.293 vs. UI-M)), Cyp4a12b (4.176 vs F; 1.424 vs UI-M)) and fatty acid synthesis (Cyp4v2 (2.246 vs. F; 1.585 vs UI M). However, fatty acid elongation was elevated in IMM-M compared to IMM-F but reduced compared to UI-M (Elov1, 2.165 vs. F; 0.439 vs. UI-M). Additionally, fatty acid transport, fabp3, glycolysis, Pygm, Pgam2, Eno3 and Ldhb were reduced in IL-33−/− IMM-M compared to F. H Heart weight (HW) and tibial length (TL) analyses confirmed enlarged hearts in IL-33−/− IMM-M (N = 11, 10.4 ± 1.3 mg/mm) compared to IL-33−/− IMM-F (N = 13, 6.8 ± 0.9 mg/mm, p < 0.0001), IL-33−/− UI-M (N = 9, 8.9 ± 0.9 mg/mm, p < 0.01) and BALB/c IMM-M (N = 4, 7.9 ± 0.4 mg/mm, p < 0.01). IL-33−/− UI-M hearts were also larger than IL-33−/− UI-F hearts (N = 10, 6.9 ± 0.9 mg/mm, p < 0.001). IL-33−/− IMM- and UI-F were similar to BALB/c IMM-F (N = 4, 6.1 ± 0.4 mg/mm). I Representative H&E staining of heart cross-sections (5 μm) showing size differences between IL-33−/− IMM-M and F. J Myocardial proinflammatory signals (IL1b, IL6, IL17 and Ifng) in IL-33−/− IMM-M were elevated compared to those in UI-M (p < 0.05). K Representative M mode ECHOs demonstrated an increased LV diameter in IL-33−/− IMM-M compared to IMM-F (Supplementary Fig. 2). L Upregulated BNP and Myh7 mRNA confirmed myocardial injury in IL-33−/− IMM-M compared to UI mice (p < 0.05). Inos mRNA was elevated in both IL-33−/− IMM-M and F hearts compared with UI hearts.

Investigations using mouse models are dependent on the method utilized to induce disease [8]. The complex roles of sex and diet in disease pathogenesis could also produce differences in outcomes when investigating sex hormones and NAFLD/NASH. Thus, investigating the pathogenesis of NAFLD/NASH using a drug-induced murine model in which diet is not a factor and sex effects can be investigated will yield important mechanistic insights regarding the link between the liver and the cardiovascular system

Funding

The work was supported in part by the Johns Hopkins School of Medicine Stimulating and Advancing ACCM Research (StAAR) award program.

Competing interests

D.B.N. is the inventor of Recognition of CYP2E1 epitopes, Patent number: 9339531, date of Patent: May 17, 2016, Assignee: The Johns Hopkins University. No funds have been received in response to this patent.

Supplementary information

The online version contains supplementary material available at 10.1038/s41423-021-00724-6.