Non-alcoholic fatty liver disease (NAFLD) constitutes a clinicopathological condition that encompasses a spectrum of liver disorders. NAFLD is the most common liver disease affecting between 20% and 44% of European adults and 43-70% of patients with type 2 diabetes [1]. It is one prime cause for chronic and end-stage liver disease, such as cirrhosis and primary hepatocellular carcinoma. Onset of NAFLD is hallmarked by the accumulation of lipids within hepatocytes (hepatic steatosis), which arises from an imbalance between triglyceride import, production and extrusion primarily caused by obesity and a hypercaloric diet. Furthermore, steatosis can be induced by a variety of drugs.

Steatosis can progress into non-alcoholic steatohepatitis (NASH), an inflammatory condition that can result in liver failure. In this process, Kupffer cells and the pro-inflammatory cytokines they secrete, are of fundamental importance in disease progression. Moreover, increased intracellular lipid levels is an important source of reactive oxygen species (ROS), which can activate stellate cells and, in turn, lead to collagen deposition and liver scarring (hepatic fibrosis) [2]. Furthermore, oxidative stress can result in pathological polyploidization, thus linking NASH to development of hepatocellular carcinomas [3].

Previous NAFLD and NASH models

The transition from steatosis to NASH, fibrosis and HCC is characterized by an intricate interplay of a multitude of nutritional and genetic factors, pathways and cell types. Primary human hepatocytes (PHH) are considered the gold standard in vitro model system to study human liver diseases [4]. However, when maintained in conventional 2D cultures, PHH dedifferentiate and lose hepatocyte-specific functions after few hours,which significantly limits their utility for long-term functional studies. In the absence of relevant in vitro systems, animal models constitute a cornerstone to study liver pathobiology. Yet, animal studies are expensive, time-consuming and not compatible with high-throughput screens of potential drug candidates. Moreover, the liver is an organ with pronounced species differences and no animal model is available that replicates the full spectrum of NAFLD manifestations observed in humans, limiting the predictive power and translatability of results to man [5].

The NAFLD and NASH model used by HepaPredict

3D spheroids are extensively characterized with regards to expression levels and activities of fatty acid uptake transporters, key transcription factors involved in lipogenesis, insulin signaling transducers as well as the lipid-metabolic enzymes on protein and transcript level [6, 7]. Importantly, primary human hepatocytes are susceptible to nutrient induced lipid accumulation, when exposed to elevated levels of fatty acids, fructose and insulin, as is the case in NAFLD in vivo [8]. Furthermore, the system has been optimized to support the co-culture of hepatocytes with other primary non-parenchymal cells [6], thus allowing to mimic the delicate interactions between hepatic cell types that are a prerequisite for the progression of steatosis to NASH.


  1. Blachier et al., J Hep, 2013
  2. Matsuzawa et al., Hepatology, 2007
  3. Gentric et al., JCI, 2015
  4. Gomez-Lechon et al., Exp Opin Drug Met, 2014
  5. Takahashi et al., WJG, 2012
  6. Bell et al., SciRep, 2016
  7. Bell et al., DMD, 2017 
  8. Feaver et al., JCI Insight, 2016