Organovo's Breakthrough In Hepatotoxicity Testing

In late April 2013, Organovo (NASDAQ:ONVO) presented data from its all cellular (scaffold free) 3D liver model at the Experimental Biology 2013 conference in Boston. Based on the wide range of in vivo like properties reported at the meeting, and the flexibility and power of bioprinting technology, we expect this product to be a game-changer in the critically important field of drug hepatotoxicity testing.

In an era of ballooning R&D budgets and declining R&D productivity, the inability to accurately predict liver toxicity is a critical issue for the pharmaceutical industry. Drug-induced hepatotoxicity accounts for about one-fourth of withdrawals of approved drugs. Most of these cases are "idiosyncratic," occurring in one of every 1,000 to 250,000 patients after 5 weeks or more of therapy. Insufficient statistical power makes it nearly impossible to identify such rare events during a typical drug development program.

Hypothetical situation: A large pharmaceutical company has spent hundreds of millions or even billions in R&D dollars to bring a compound into late-stage clinical trials, only to see that compound fail (or have restricted use / dose) based on hepatotoxicity. Well, it's not hypothetical, because it's happened many times before. In fact, liver toxicity is the second leading reasons why drugs are pulled from the market.

Recent examples include Pharmasset's Hepatitis C polymerase inhibitor PSI-938, which was terminated from clinical development only weeks after Gilead (GILD) agreed to buy the company for over $11B, and the identification of increased ALT levels as the dose-limiting toxicity in three of seven hepatitis C protease inhibitors now in late stage clinical development. Another example, in March 2000 the U.S. FDA forced Warner-Lambert to withdraw Rezulin (troglitazone) from the market due to severe adverse liver toxicity. Rezulin was approved in January 1997. Warner-Lambert spent hundreds of millions to develop the drug. One month before the withdrawal, Pfizer (PFE) agreed to acquire Warner-Lambert. The withdrawal cost Pfizer $136 million to institute.

It's a big problem despite the fact that the FDA and EMA each require extensive testing in animal models prior to and in parallel with clinical development. Typically, these liver tox studies include at least one rodent and one non-rodent species, examine a range of doses, and last up to one year. In general, these studies seem to be effective at identifying drugs that would produce severe hepatotoxicity in a large percentage of patients if a clinical trial were performed. However, they are relatively ineffective at identifying drugs, which will produce a low level of toxicity in a smaller percentage of patients. The poor predictive power of preclinical animal studies likely results from a combination of factors, including species specific metabolic routes and levels of serum protein binding; and in some cases, inability to dose high enough due to non-hepatic dose-limiting toxicities.

Cell based models are used both in broad screening in the early discovery process and as mechanistic tools for understanding toxicity. Unfortunately both the range of studies that can be performed and its in vivo relevance is reduced because hepatocytes have limited viability and undergo dramatic phenotype changes when grown in a two dimensional culture. Specifically, they lose their cubic shape, become flattened, and become unable to produce key drug metabolizing proteins and other characteristic proteins within hours to days of isolation. The ability to study idiosyncratic hepatotoxicity is especially limited as such reactions normally require weeks to months to develop and are thought to involve multiple liver cell types.

In efforts to develop more sophisticated in vitro systems that better recapitulate the in vivo situation a variety of modified assays have been developed to maintain liver-like function for longer periods. These systems typically include modification of materials to better mimic the liver extracellular matrix, use of other cell types, use of 3-dimensional scaffolds, and/or use of continuous perfusion to provide nutrients and remove waste materials. Some examples include

1. Addition of materials that mimic the liver extracellular matrix: Including materials that can serve as a more realistic extracellular matrix helps prevent flattening of the cells, improves viability (to several weeks), and delays loss of hepatocyte-specific functions such as the production of drug-metabolizing enzymes. This is especially true if the cells are "sandwiched" between two layers of extracellular matrix-like material.
2. Co-culture systems: Typically also two dimensional, these include the simple addition of epithelial cells that produce an extracellular matrix. In other cases the co-culture may include hepatocytes and other types of liver cells so that cell-cell interactions can be examined
3. Simple 3-D scaffolds: The Alvetex scaffold (AmsBio) is a fenestrated polystyrene scaffold with 40 uM pores for 3-D cell culture. It demonstrates superior CYP activity and inducibility compared to 2D cultures and can be used for co-culture experiments. A similar system produced by Insphero (3D InSightTM) is shipped as a co-culture of hepatocytes and a representative distribution of non-hepatocyte liver cells. It has been used in experiments demonstrating the importance of Kupffer cells in the hepatotoxicity of trovafloxacin, an exemplar of idiosyncratic hepatotoxocity. These systems generally depend on the tendency of cells of different types to self-organize, and offer the experimentalist little control over tissue architecture.
4. Microfluidic perfusion arrays: The Perfusion Array Liver System (PEARLTM, Celasic) incorporates a 3-D scaffold each of which includes an array designed to mimic the liver acinus, which is the smallest scale repeating organization of hepatocytes in the liver. Continuous perfusion guarantees delivery of nutrients and removal of waste products. Hepatocytes cultured under these conditions retain liver functions such as CYP metabolic activity and induction potential, albumin expression and drug metabolism over 28 days. Most studies to date have involved hepatocyte monocultures
5. Micropattern perfusion system: The HepatoPacTM (Hepragen) platform is another continuous perfusion system. Hepatocytes are co-cultured with stromal cells to provide an appropriate extracellular matrix. A variant, which includes Kupffer cells, has been reported.

Organovo's Scaffold-Free 3D Bioprinting Solution

Organovo is developing 3D bioprinting as a tissue engineering method that allows finer control over both the composition and architecture of the created tissue than is possible with scaffolding. We have described the technique in detail in a previous article, and so will only mention key features here.

Compared to scaffold-based approaches, bioprinting provides a method of tissue construction with exceptional architectural and compositional control. It also provides a 3D tissue construct that is optionally composed entirely of cells. In contrast to scaffold approaches, in which one attempts to control the composition and relative geometry of cell growth by selection of a suitable support, bioprinting directly controls the placement of multiple cell types in a manner similar to an inkjet printer. The inkjet simply moves back and forth, left and right, up and down to place cell types in the precise position specified by the operator.

The "bio-ink" is primarily composed of stem cells extracted from adult bone marrow and fat as the precursors, but any mammalian primary cell will work in the printer. Bio-ink production begins with the creation of a thick cell paste comprised of a slurry of cells and the other necessary components required to be part of the final tissue composition. After a maturation period, the bio-ink is loaded into the bioprinter, which then dispenses these building blocks in the geometry specified by the user through a printing head. A second printing head is used to deposit a sugar-based hydrogel bio-compatible scaffolding. Once the printing is complete, the structure is left for a day or two to allow the droplets to fuse together. In the final tissue construct the cells build their own extracellular matrix for support, providing a more natural structural support that more closely mimics the in vivo environment.

Using the NovoGen Bioprinter (shown above), Organovo scientists prepared morphologically correct replicas of the basic structural units of the liver. Figure A below shows the layout of the classical structural unit of the hepatic lobule. When viewed in cross section it appears as a hexagon, with the hepatic artery, bile duct, and portal vein triads at the hexagon vertices. The large oval in the middle of each hexagon represents the central vein. The shape and arrangement of these hepatic lobules can be compared to the bioprinted 3D liver tissues prepared with the NovaGen Bioprinter. These tissues were created from hepatocytes or hepatocyte like cells and non-hepatocyte liver cells. They exhibited tissue-like cellular density and tight intercellular junctions.

Source: LeCluyse EL (2012)

Source: Organovo

A more detailed view is provided in the Figure below. Bioprinting not only provided hepatic lobule-like structures with the correct geometry and aggregation pattern, but even allowed for the correct placement of stellate cells and endothelial cells at the interface between lobules.

The Organovo researchers demonstrated that structures of correct geometry could be prepared from a variety of different mixtures of cell types. Constructs generated from iPSC-derived hepatocyte-like cells were demonstrated to produce physiological albumin at a more physiological rate compared to 2D controls. The bioprinted cells were well-organized and produced microvascular networks over time. Constructs generated from different cell types (including hepatocytes, transformed cell lines, hepatic stellate cells, and endothelial cells) appropriately produced cholesterol, albumin, fibrinogen, transferrin, and CYP450 enzymes. The CYP450 activity was inducible, as seen in vivo, something that has historically been difficult to achieve in vitro. In most cases, protein production appeared to increase over time rather than tapering off and disappearing as seen with 2D cultures.

Organovo's 3D bioprinting process is clearly game changing. In comparisons with two dimensional in vitro systems, the bioprinted liver model provides a dramatic enhancement in the levels and durability of liver specific functions such as the synthesis of albumin, fibrogen, transferrin, and CYP450 activity. It further provides unparalleled control over both the architecture and cell type content of tissue constructs, which gradually form tight junctions and microvasculature reminiscent of the in vivo state. The construct is created without the presence of scaffolding or other supporting material or cells that are not native to liver tissue. We believe that it will prove to be a powerful tool for the evaluation of hepatotoxicity of novel drug candidates, for the investigation of the mechanism of toxicity of specific compounds, and possibly for the investigation into general mechanisms of poorly understood adverse events such as idiosyncratic hepatotoxicity.

If all this were not enough, in December 2012, the company partnered with Autodesk research to create the first 3D design software for bioprinting. The software, which will be used to control Organovo's NovoGen MMX bioprinter, will enhance the usability and functionality of the bioprinter, and will open up bioprinting to wider group of users, allowing them to create their tissues to their own specifications.

Revenue Estimates for Liver Toxicity Screening

Wouldn't this have been great technology for both Warner-Lambert or Pharmasset to have prior to wasting hundreds of millions on their respective drugs. Clearly this is an enormous opportunity for Organovo.

We estimate sales as shown in the table below. We begin with an estimate of approximately 500 small molecule drugs entering Phase I testing each year (source: Clinicaltrials.gov). Although companies are not required to register Phase I trial, there were approximately 1500 Phase II trials of small molecules initiated in 2012, and we assume that approximately 2/3 of Phase I drugs are suspended from development, in accord with published estimates. We assume that at peak sales, 20% of new Phase I candidates will be tested with Organovo's bioprinted liver technology, and that completion of these studies will require an average of three 24 well plates. We estimate an additional 10,000 compounds per year are potential candidates for testing based on being of high interest (preclinical development compounds) or belonging to a class with a high suspicion of hepatotoxicity.

We expect about 10% peak market penetration of this group. We expect the largest use of the product will be for projects not associated with a single lead compound, but instead basic toxicology research directed toward the idiosyncratic toxicity issue and studies performed in support of marketed drugs. We estimate approximately 2,000 plates per year for this type of work, which is 20% of an estimated addressable market of 10,000. We estimate meaningful annual sales can be achieved rapidly after commercialization.

Use	Addressable Market (Units)	Penetration	Units
Clinical Development Candidate Characterization 500 Candidates @ 3 Plates per Compound	1,500	20%	300
Characterization of Earlier Stage Discovery Compounds of Special Interest, 20,000 compounds @ 0.5 Plates per Compound	10,000	10%	1,000
Basic Toxicology Research and Research Support for Marketed Drugs	10,000	20%	2,000
Total Number of Plates			3,300
@$9000/plate			$29,700,000

Click to enlarge

Conclusion

Idiosyncratic hepatotoxicity is a major problem for pharmaceutical and biotechnology companies. Behind cardiovascular toxicity, it is the second leading reason why a drug may be pulled from the market or discontinued in late-stage clinical trials for adverse events. Pharmaceutical companies have spent millions (see Pharmasset or Warner-Lambert) to develop or commercialize drugs, only to get stymied by previously unseen liver toxicity issues. It is no surprised that Pfizer, a victim of this unforeseen liver tox issue with Rezulin, has a research relationship with Organovo.

Organovo has come up with a way to build 3D liver-like structures in ELISA wells that better represent the in vivo environment. These constructs are up to 20 cells high, stable at room-temperature, viable for several days, and are far superior to 2D, co-culture, or scaffolding-based liver systems. Biological assay or protein production testing shows remarkable viability and function for these constructs. We believe this is a $30 million opportunity for the company upon commercialization, and this is only the beginning. After liver, we expect Organovo to push forward with development of kidney, lung, and even heart tissue systems.

We've been bullish on the Organovo story for some time now. That said, we have spoken with some investors skeptical of the emergence of 3D bioprinting, believing it's more "interesting science" than developed application to deliver real revenues. It's clear that Organovo is not going to print patients a new liver - at least not in the next decade! But the data presented at the Experimental Biology conference in April 2013 is the first to clearly show investors an application for 3D bioprinting that can and will generate meaningful revenues in the coming years. Organovo is pioneering a potential game-changing technology for liver toxicity testing. That is what turns "interesting science" into revenues, and thus an "interesting investment."

Co-Authored by Jason Napodano, CFA & John Tucker, PhD

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.