April 15, 2014 (Vol. 34, No. 8)

Cliff Rowe, Ph.D. Principal Scientist CN Bio Innovations

Accurately Predicting In Vivo Hepatotoxicity

In 2013, the cost of identifying and developing a new drug was on average between $350 million and $5 billion. One major reason for such high development costs is that 95% of the drugs entering clinical trials fail to be both effective and safe.

Hepatotoxicity accounts for a substantial number of late-phase drug failures and post-marketing withdrawals, despite extensive preclinical safety assessment screens. Even though many mechanisms causing drug toxicity have been identified, the prediction of the extent of toxicity damage, and the long-term implications remain a great challenge.  There is clearly an unmet need for in vitro models capable of predicting the effects seen in clinical trials.

The toxicity of many drugs observed in the clinic is due to poor drug metabolism or the acute formation of toxic reactive metabolites. This is difficult to reproduce in vitro using primary hepatocyte cultures as those metabolic liver functions are lost early on in the culturing process. Mitochondrial damage and cholestasis, a disturbance in bile formation or retention of bile acids in the liver, are known to cause serious—and in some cases fatal—adverse reactions in patients and have proven equally difficult to predict.

Current In Vitro Model Systems

Until today, primary hepatocyte two-dimensional (2D) cultures have been used as the gold standard for testing in vitro hepatotoxicity. However, the availability of freshly isolated hepatocytes is very limited. Cryopreservation techniques can overcome this problem but other phenomena such as cell de-differentiation can limit the usability to short-term applications.

Furthermore, in vivo toxicity is accumulative, developing over weeks or months. Standard in vitro models can be applied to toxicity screening for only a few days at most. This has stimulated the development of 3D cell culture models, where the more physiologically relevant microenvironment results in a prolonged lifespan, and increases the potential to gain a better understanding of the long-term effects of sustained drug administration.

There are a large number of commercially available culturing tools for growing 3D cell cultures, such as rotating bioreactors, hanging drop plates, extracellular matrices, and magnetic levitation. One of the great challenges is to ensure adequate exposure to nutrients and sufficient gas exchange to maintain full cellular function throughout the cell colonies. This is particularly important in hepatocyte cultures as the liver is a highly active and therefore highly vascularized organ, thus hepatocytes are regularly bathed in the nutrients, hormones, and growth factors.

To meet these requirements, CN Bio Innovations developed and has fully commercialized a perfused 3D microtissue culture system (LiverChip™), which effectively mimics the liver sinusoid. This system provides shear stress and flow rates that are more alike to in vivo conditions (Figure 1). Cultures retain their primary hepatocyte phenotype and full metabolic activity for several weeks providing the potential to greatly improve hepatotoxicity prediction during drug development.


Figure 1. The LiverChip system

Distinguishing Harmful from Safe

As one of the most advanced 3D cell culturing systems, LiverChip was tested for its capacity to differentially analyze the extent of damage caused by hepatotoxic drugs and known safer alternatives with closely related structures in vitro. Three mechanisms of drug-induced liver injury were explored using Clozapine, a drug known to induce hepatic necrosis via reactive metabolite formation, Tolcapone, which causes mitochondrial damage, and Bosentan, which inhibits the bile salt export pump.

Hepatotoxicity through Reactive Metabolites

Application of the antipsychotic drug Clozapine, in multiples of the clinical maximum plasma concentration (Cmax), to LiverChip cultured human hepatocytes caused severe cytotoxicity after 48 hours of drug exposure. Lactate dehydrogenase (LDH) release, a marker for cytotoxicity, rose strongly in a dose and time-dependent fashion (Figure 2A). In contrast, the alternative drug with a favorable safety profile, Olanzapine caused only a minor increase in LDH release over control hepatocyte cultures and only after sustained exposure.

The markers for hepatocyte health and functionality including albumin secretion and mitochondrial activity via turnover of the dye WST-1 (Water Soluble Tetrazolium salts) also reveal the different impact of the two drugs on cell viability. The data show a significant reduction in albumin secretion and mitochondrial activity after one dose of Clozapine but not Olanzapine (Figure 2B). Cell survival and tissue integrity at the end of the experimental period is comparable to control tissues when treated with Olanzapine but extremely poor for the tissues exposed to Clozapine (Figure 2C).

Therefore, drugs metabolized to reactive chemicals that lead to hepatic necrosis can be detected in vitro using the LiverChip culture system and used to help select safer alternatives before in vivo tests are initiated.


Figure 2. (A) Hepatocyte necrosis test. Day 5 (D5), 7 and 9 of cell culture. (B) Albumin secretion and WST-1 turnover. (C) Liver tissue survival.

Mitochondrial Toxicity

Similar results were found with two anti-Parkinson’s drugs, Tolcapone and Entacapone. Tolcapone is known to be a mitochondrial toxin demonstrated by the markedly reduced turnover of the mitochondrial substrate WST-1 after six days of tenfold or more Cmax drug exposure (Figure 3A). In contrast, Entacapone has rather mild effects on mitochondrial activity and no effects on LDH levels or albumin secretion.

A marker commonly used for identifying the existence of hepatocellular injury in vivo is an elevated level of alanine transaminase (ALT) in plasma. Significantly increased levels measured in international units per liter  (U/L) often suggest liver malfunction that could be caused by viral hepatitis, diabetes, liver damage, or cholestasis. Hepatocytes exposed to Tolcapone show significantly elevated levels of ALT after three days of drug exposure but no change compared to control cultures when exposed to equal amounts of Entacapone (Figure 3B).

Mitochondrial toxicity is difficult to predict based on in vitro tests as the decline in mitochondria numbers does not instantly disrupt cell function. However, it can cause severe problems in patients and has been responsible for the withdrawal of several highly potent drugs from the market. The ability to distinguish between toxic Tolcapone and Entacapone using LiverChip shows that deselecting mitochondrial toxicity-inducing drug candidates at the in vitro testing stage is possible.

Cholestasis—Affecting Bile Formation

The identification of cholestasis-inducing hepatotoxins such as Bosentan is challenging because there is no observable perturbation of hepatocyte wellbeing for several days even at 100-fold Cmax exposure. After four days of drug exposure, a slight decrease in albumin secretion can be measured followed by significant hepatotoxicity effects after six days of the highest drug dose. No hepatotoxic effects were found within that time frame and dosage range with Macitentan, the safer drug.

LiverChip is one of the few in vitro testing systems maintaining fully functional liver cell cultures for long enough to detect slowly developing hepatotoxic adverse reactions such as cholestasis.


Figure 3. (A) Tolcapone severely impairs mitochondrial activity. (B) Tolcapone causes liver injury.

Platforms for ADMET Testing

Perfused 3D cell culture models are the most promising platforms for more effective drug absorption, distribution, metabolism, and excretion (ADME) testing and hepatotoxicity prediction by better mimicking liver complexity. They provide a more human physiologically relevant model for pre-clinical lead optimization of drug candidates. Most importantly, it is possible to distinguish earlier hepatotoxins from less-hepatotoxic structurally analogous compounds, with differing mechanisms of toxicity. Additionally, given the increased viability and prolonged maintenance of functionality of hepatocytes, in systems like LiverChip they are also providing a high value, extended time window for testing low clearance drug compounds.

Cliff Rowe, Ph.D. (cliff.rowe@cn-bio.com), is principal scientist at CN Bio Innovations.