It’s been a little over a year since the first ￼International NASH Day, and we still aren’t close to having a NASH drug on the market. The projections of this epidemic are staggering, as the worldwide ￼prevalence of NASH￼ is expected to increase by 63% between 2015 and 2030. It’s not that we don’t have a drug because pharmaceutical companies have given up on this complex disease. The trouble is promising preclinical drugs have not performed as expected once tested in humans.
Large trials to date, such as GOLDEN-505, STELLAR3/4, and FLINT, have demonstrated only slight resolution of different stages of the disease – not enough to meet primary endpoints. A quick scan through the NIH clinicaltrials.gov database shows that there have been a staggering 772 trials to date relating to NAFLD and NASH.
How did so many of these drugs make it to extremely expensive, lengthy clinical trials in the first place? Unfortunately, preclinical models for accurately predicting human NASH have been rather limited to date. In vivo mouse models, including diet-induced models, the Methionine Choline Deficient (MCD) model, and genetically modified mouse models, are the most frequently used preclinical models. These typically require 20 weeks to show the first hallmarks of NASH disease progression, while the appearance of severe fibrosis can take 30 weeks or longer.
One major issue with most in vivo mouse models is that, in attempts to best mimic the major human risk factors associated with NASH, most models are at least partially diet-based, typically with some variation of the “Western Diet” (high fat, high sugar, and 0.1-2.0% cholesterol). One might assume that all lab mice of a particular strain would eat the same amount, but this is not actually the case. So much so that confirmation by surgery and liver biopsy is typically required prior to the initiation of drug treatment (~week 20-30, depending on the model) to prove that these mice truly have NASH. Subsequently, a significant percentage of mice are excluded from the drug treatment afterwards.
I say we break the bottleneck in NASH R&D today and move away from mouse models to human-based, scalable options!
Many of these dietary mouse models only progress as far as steatosis with or without mild inflammation. Others develop as far as fatty liver hepatitis with significant fibrosis.
To achieve fibrosis, these mice require much stimulation beyond the “western diet” (for example, genetic mutation of genes not directly relevant to human NASH such as bd/bd, ob/ob, or PPARa-/- mouse models).
Therefore, we cannot assume the causality to be the same – and that drugs designed for humans will work for mice (and vice versa). There is really not a “one-size-fits-all” with mouse models. That said, one model that does nicely recapitulate several attributes of human NASH is the severely metabolically deficient MCD model. It is known for its simplicity and more rapid onset of fibrosis (~5 weeks) and thus is the model most commonly used. Unfortunately (and ironically), these mice lose ~40% of their weight by 8 weeks, most likely because of a hypermetabolic state.
Something else to consider is that most of the more promising drugs in clinical trials target earlier stages of NASH much further upstream than fibrosis. Some examples include the promising THRb agonists from Viking Therapeutics and Madrigal Pharmaceuticals. These drugs directly affect de novo lipogenesis (DNL), and the mouse DNL pathway is markedly different than that of humans
But let’s say a perfect world existed where an in vivo mouse model is reproducible and biologically translatable. Even if this were to be achievable, scalability will always be a problem. A wonderful recent Hepatology review article by Farrell et al compares all current mouse models used for NASH.
On average, these ~20 different types of mouse models take 12-52 weeks to develop the different stages of NASH. Granted, it is estimated that NASH takes 5-50 years to develop in humans, but if it takes over half a year to test a single drug in a preclinical trial, no wonder we haven’t found the right drug/drug combination yet! This challenge gets even bigger when considering combination therapy treatment to address the multitude of pathways involved in NASH.
While mouse models are valuable for offering a systemic perspective on drug-organism interactions, their limitations have led to a growing interest in more scalable and, ideally, human-based in vitro models. Recent progress, both in academia and by biotechs, such as InSphero, has pushed complex 3D co-cultures towards highly reliable and predictive drug-testing models. Our own technology combines multiple liver cell types in a scaffold-free 3D microtissue to recapitulate different stages of NASH in a screening-ready 96-well plate.
I say we break the bottleneck in NASH R&D today and move away from mouse models to human-based scalable options!
Watch this video to learn how we applied our scalable Akura™ technology and 10 years of experience in perfecting 3D in vitro human liver models to develop the first automation-compatible 3D in vitro human liver disease platform for NAFLD and NASH.