What is NASH?
Non-alcoholic fatty liver disease (NAFLD) is the most common form of chronic liver disease in the United States. Liver disease is often closely associated with obesity and metabolic syndrome. Current estimates suggest between 30 – 40% of the US population has NAFLD with about 20% of these individuals progressing to nonalcoholic steatohepatitis or NASH (Spengler et al., 2015). NASH is diagnosed when the liver has progressed from simple fat deposits to more serious inflammation, cell damage, and fibrosis, which can eventually lead to cirrhosis and eventually hepatocellular carcinoma.
Most patients who have NAFLD are unaware as this tends to be asymptomatic. However, there is an excess accumulation of fat in the liver, which can interfere with normal liver function. It is necessary to develop tools to diagnose NAFLD at an early stage to help prevent the transition from simple fatty liver into the more serious NASH, as well as to develop therapies to treat patients who have already progressed to liver fibrosis.
While there is no single mouse model that recapitulates all of the same symptoms and progression of the human disease, investigators have generated multiple dietary models that can mimic certain aspects of disease development and progression.
How can you decide which of these models will work best for your research purposes? Read on for some key tips on choosing the best model (or models!) for you.
What phenotype are you interested in studying?
One major challenge of finding the most appropriate model is to determine which phenotype of NASH you are looking to study. A single characteristic does not diagnose NASH; instead, it is a collection of conditions. In human patients, several pathologies will inform the correct NASH diagnosis. Histological examination of a patient’s liver is necessary to score for inflammation, steatosis, ballooning, and presence of fibrosis. All of these characteristics must be present for the diagnosis, but they may vary in severity.
Mouse models, however, may be able to recapitulate one or more of these phenotypes. A search through the literature will reveal many different tables that will provide a binary list of phenotypes observed in each model – usually along with a dietary, genetic, or chemical intervention. Liver fibrosis is not commonly observed in mouse models without one of these interventions; similarly, in humans, a ‘multi-hit’ hypothesis has been developed to explain why some patients progress to NASH from simple fatty liver while others do not. It is critical to look at the primary references for each of the models listed in the tables, as a simple ‘yes’ or ‘no’ does not provide details such as the severity of the phenotype. A mouse model that presents with mild fibrosis and a model that presents with severe fibrosis would both be marked as ‘yes.’
In addition, most human patients suffer from many comorbidities associated with liver disease such as obesity, hyperglycemia, and increased serum cholesterol and triglyceride levels. This will be important to consider when testing any potential therapeutics on mouse models of the disease – are you accurately mimicking the human condition that would be present in the disease?
High fat? What’s that?
You’ve looked through the literature and found publications showing that C57BL/6J mice develop fatty liver disease on a high-fat diet – great! But when you look up high-fat diets, you might be a bit shocked to find dozens of possibilities.
A “high fat” or “western” diet is often cited in the literature, but there isn’t a specific definition for either of these names. High fat is relative to what the mouse would normally eat; for example, if your facility maintains mice on a 6% fat diet, a 10% fat diet might be considered high fat by these standards.
Commonly, a high-fat diet can range between 40% - 70% in fat content and has an undefined amount of calories coming from other sources such as sugar. It is critically important to know what diet was used in any publication you are trying to reproduce as differences in nutrient content can have significant effects on metabolic outcomes.
One fascinating study looked at C57BL/6J mice maintained on either a high-fat diet or control diet, but the mice were pair-fed to ensure all groups had an isocaloric intake. Mice fed the high-fat diet - with 71% total calories from fat, 11% as carbs, and 18% from protein - developed metabolic dysfunction, liver steatosis, and increased blood glucose levels compared to the control diet. The control diet in this particular study contained 35% calories from fat, 47% calories from carbohydrates, and 18% of calories from protein. Another group, using the same strain of inbred C57BL/6J mice, placed mice on a 45% lard fat diet as their version of high fat, and a 10% fat diet as their control, but saw similar results. You may have noticed that 35% fat for a control group seems pretty high – a great example of why it is critical to pay attention to the details!
This goes for any diet you can think of: high fat, high sugar, or high cholesterol. Without determining the actual nutrient makeup in the diet, you may have a hard time reproducing experimental results. Always check the “methods” section of the literature. And, when you go to publish, be sure to include as much detailed information as you can about any diets that you may have used.
Some diets that are high in fat may also have other additives to mimic the ‘fast food’ diet eaten in many developed countries. The American Lifestyle-Induced Obesity Syndrome (ALIOS) diet involves high trans-fats and sugars, and investigators generated the Amylin liver NASH (AMLN) diet to also add cholesterol. Both of these diets have been used extensively to model liver dysfunction, though the availability has decreased due to the current FDA ban on trans-fats. New diets are currently being investigated to replace the AMLN diet, such as the GAN diet, which shows promise for inducing NASH in mouse models.
Another popular set of diets often used to recapitulate hallmark features of NASH are nutrient-deficient or nutrient-defined diets. Methionine and choline-deficient (MCD) diets are particularly popular as they are able to induce liver damage and fibrosis rapidly. Methionine and choline are important for hepatic mitochondrial beta-oxidation and production of very low density lipoprotein (VLDL), so by removing these from the diet, liver metabolism is disrupted. This particular diet has been shown menus oxidative damage mechanisms implicated in human NASH progression.
MCD diets also cause rapid weight loss and hypoglycemia – not seen in the human NASH population. Variants of this diet, such as the choline-deficient, reduced methionine diet (CD or CDAHFD in much of the literature), have been generated to avoid weight loss and maintain euglycemia in the mouse model. Investigators hope to build upon these nutrient defined diets to generate a model that closely mimics the pathogenesis observed in human patients.
As a side note, there is a growing body of literature that suggests gut microbiome may also play a role in NASH development (Kolodziejczyk et al. 2019). If you are interested in the microbiome, you might be interested in this blog post describing microbiome stability in JAX mice.
Overall, you may need to use multiple mouse models to recapitulate all of the phenotypes of NASH you are interested in studying. By using more than one model, results are likely to be more relevant and translatable. Having a deep understanding of the type of model (and diet!) you are working with will be critical in producing the most relevant data. JAX maintains a portfolio of Key NASH Models with links to each strain and curated publication lists that might be a good place to start your search!
If you are curious to learn more about some of the studies mentioned in this post or would like to learn more about some of the important phenotypes that can be modeled in mice, please see the JAX webinar on Choosing the Best Mouse Models for Your NASH Research. Also, be sure to check out the Metabolic Disease page to find out additional information on available metabolic models.