What are the challenges of drug development?
Drug development is a very challenging and often arduous process with many associated challenges. We must work with a therapeutic to ensure that it is efficacious and safe, so effective translation is important throughout the drug development cycle.
Translation in this context represents predictability between preclinical methods and human response therapeutics.
Data from in vitro and in vivo studies are required to translate to the whole human system. This is where discrepancies creep in. In vitro systems are often segregated into tissue types or cell types, so they do not always offer strong translation to a whole system. In vivo studies and systems, on the other hand, represent a whole animal system, with all its organs and tissues integrated, but ultimately, they have physiological differences that do not necessarily offer the translatability required to predict the human response.
When we eventually get to clinical trials, we require translation to patient populations and eventual practice. Therefore, we must consider translation—essentially prediction—throughout the whole drug development cycle.
Image Credit: Gorodenkoff/Shutterstock.com
What models can we use during the drug development process to monitor safety?
We use various models throughout drug development, particularly to assess safety. At the beginning of the process, in silico models are frequently used to screen compounds whittling them down to just a few candidates to progress. These computational models simulate biological processes to predict how a drug will behave in the body, helping to identify potential safety issues and optimize drug candidates more efficiently.
We also use in vitro and in vivo models to confirm and validate safety and efficacy predictions. No single model is perfect, but when used together they build a risk assessment to guide decision-making before clinical trials.
What are the main reasons that drugs fail, and why is it important to prioritize safety and toxicology during the development process?
Most drugs fail at some point in the development pipeline. This can often be attributed to the poor translation of pre-clinical models to humans. This issue is particularly prevalent in safety and toxicology, where numerous factors can affect data translation.
When picking a model, it is essential to consider things like the expression of the target and the target organ and whether the models selected express the target required. It is also important to consider whether the target is expressed in other organs to monitor off-target toxicology or toxicity within a system.
Metabolic activity should also be considered. For example, whether the candidate is metabolized by different organs (e.g. for orally administered drugs the gut and/or liver) and whether the resulting metabolites are potentially toxic to other organs. It is also important to rule out activation of the immune system. This becomes even more important when developing human-specific drug modalities, such as cell therapies, DNA therapies, antibodies, or PROTACs. These drugs are specifically designed for use with a human target, making them both efficacious and useful. However, it can be challenging to understand these new drug modalities because of the lack of physiological, or human relevance associated with in vitro and in vivo models, respectively.
We often use in vitro 2D and 3D cell cultures for decision-making, which can lack the physiological relevance and complexity of the human body. These cell cultures offer high throughput and can screen many compounds simultaneously, but the data output is limited and often simplified.
In vivo models provide information from an entire system, so their data output is rich by comparison. However, these models present a translatability challenge between human systems, they are low throughput, expensive, and resource intensive.
Can you talk about new alternative methodologies and microphysiological systems?
New alternative methodologies (NAMs) are advantageous because they are often human (although not always). Microphysiological systems (MPS), also known as Organ-on-a-chip (OOC) are a type of NAM. They bridge the gap between traditional in vitro approaches and the human by providing greater physiological relevance and excellent translatability.
MPS typically provide high-content data (although, again, not always), to derive mechanistic insights. The challenge is that these are medium throughput compared to traditional approaches and are therefore best suited to phases where fewer drugs are tested in more detail. Additionally, MPSs are isolated systems, compared to animals, although linked organs can be useful in understanding specific interplays between different organs.
MPS come in many different shapes and sizes, but they share common features like human cells and fluidic flow. The cells used are generally primary or iPSC-derived and grown in 3D patterning, representing the required human organ tissue structure. Fluidic flow mimics blood flow, providing oxygen and nutrient supply to the tissues, while biomechanical stimuli keep the tissues viable and functional for longer than static cultures.
MPS offer several other advantages; for example, they can include mechanical cues and electrical stimulation, or control gas exchange and growth factors.
Before starting a study, it is important to consider the right MPS for your application, or context of use. Considerations include the size of the tissue, and the media volume required for analysis. For example, if you need high-content mechanistic detail, choosing a system that has provides high sample volume is important, but what you might gain in size, you may lose in throughput. It is also important to consider the number of compounds you want to screen, as well as the concentration range and the different conditions that must be accommodated.
The ease of onboarding and using the system is also key, as is the type of chip material. Many MPS or OOC are made of PDMS, a very high drug-binding material. Others, such as PhysioMimix® Multi-chip plates, are made of lower-binding materials which enable more straight forward prediction of drug properties.
The type of perfusion should also be explored. Various different types of OOC/MPS are available in the field; for example, circulating perfusion moves media around each of the wells, while single-pass perfusion flows the media past the tissue just once. Higher throughput system plates can also be placed on rockers, which move the media backward and forward across the tissues, which is less physiologically relevant but more amenable to automation.
How has CN Bio qualified your systems, and what applications does CN Bio assist with?
Over the years, CN Bio has employed various approaches to qualifying our systems. In 2020, we became the first organ-on-a-chip provider to co-publish with the FDA on the characterization of our liver MPS for looking at drug toxicology, metabolism and accumulation (Rubiano et al., 2020).
However, we have worked with various applications in disease modeling, ADME and toxicology. We continually add new applications to our PhysioMimix OOC Systems (Single-organ, Multi-organ and Higher Throughput) to expand their capabilities and we have been “packaging up” applications into all-in-one kits to enable end-users to replicate our assays and models in their laboratories more easily.
PhysioMimix OOC Systems feature a controller unit, a docking station, and a driver, the latter two of which sit inside a standard laboratory incubator. The culture of our predictive human models happens in Multi-chip consumable plates which are connected to the systems microfluidics via the driver and docking station and powered by the controller unit. Multiple plates can be run simultaneously to increase throughput.
Our Multi-chip Liver plates feature 12- or 48-individual organ-on-a-chip models per plate, with each well including a porous scaffold where the liver microtissues sit. The perfusion moves around these tissues, which recapitulate the liver sinusoidal structures.
Our Multi-chip Barrier plates also feature 12-wells per plate. They can support insert-based barrier models (e.g., Transwell®s), such as the gut and the lung. Our Multichip Dual-organ plate combines the power of the liver and barrier model in a six-well plate, whereby an interconnecting microchannel connects the two models. This enables users to develop powerful experiments that examine the interaction between different organs in detail.
Within disease modeling, we are especially focused on fibrotic liver disease (such as MASH), whilst our ADME applications look at metabolism, absorption, and the resulting bioavailability of drugs within the gut and the liver.
Our safety toxicology applications target Drug-induced liver injury (DILI). Liver toxicity remains a significant challenge, causing about 30% of drug failures, with hepatotoxicity accounting for 18% of these cases. Thus, there is a need for a better model to predict these events in drug discovery.
Can you please explain DILI further and how this can occur?
DILI can occur in several ways. A drug, or its metabolites, can directly affect the liver once it has passed through. Drugs can also impact the liver via interactions with other organ systems, such as the immune system. The liver is an organ often harmed during high inflammatory states in the body.
DILI can be sub categorized into types. The simplest is intrinsic DILI, which is often dose-dependent, relatively predictable, and has an early onset. These types of DILI events can often be predicted using standard approaches, such as 2D or 3D spheroid assays.
Indirect or idiosyncratic DILI is less predictable, however, and this often exhibits a latent onset linked to complexity in terms of the drug’s pharmacodynamics or even the person’s phenotype, genetics, disease profiles, lifestyle, and other factors. This complexity requires a more functional, robust, long-lived human model that can capture those latent onset events.
The best way to capture complex DILI mechanisms is by using human liver cells (hepatocytes), as well as immune system components (Kupffer cells) to capture immune-mediated DILI events. We also need to ensure our hepatocytes are functional and can produce albumin and urea, for example.
Hepatocytes must also be metabolically active to allow insight into drug metabolism. Plus, they must also express key transporters and damage markers we expect to see in the clinic, such as ALT and AST. Finally, the in vitro cultures must be robust and able to be used for around 28 days to allow understanding of more chronic DILI events.
How does the PhysioMimix culture liver microtissue?
PhysioMimix OOC Systems provide recirculating culture media flow through the scaffold and the liver tissues. It is important to provide oxygen, nutrients, and sheer stress to the liver tissues to allow their high metabolic activity.
These microtissues develop from four hours to maturation at day seven and beyond. The microtissues secrete albumin secretion, form bile caliculi, and tight junctions, demonstrating the polarity of the tissue.
When we compare these tissues to the Human Gene Atlas, we can see good correlation and, therefore, high physiological relevance.
Over the years, we have made considerable efforts to make these tissues robust and reliable. In one study, we analyzed 360 independent wells and demonstrated below 20% intra- and intra-CV.
A study completed by the FDA analyzed the PhysioMimix system for drug analysis. In terms of metabolic activity, traditional methods of hepatocyte culture of spheroids and sandwich cultures were compared to the Liver MPS.
The study confirmed that the high metabolic activity of the Liver MPS can be maintained over 28 days. In contrast, spheroid cultures had low metabolic activity, and sandwich cultures started high but quickly reduced their activity at around 12 days.
High metabolic activity is important for understanding the dynamics of drugs once they enter the body. We and others have qualified our liver MPS for applications in drug administration, to understand ADME, clearance and the bioavailability of drugs.
At the beginning of drug discovery, high-throughput screening is required. This takes the form of 2D or spheroid 96—or 384-well plate screening tests to screen out any obvious toxic compounds.
Once this has been performed, PhysioMimix becomes a very useful tool for further exploring promising compounds and testing them in a human-relevant system before they are tested in an in vivo animal models and then clinical trials.
When used in this process, PhysioMimix allows for an earlier understanding of human safety profiles, the refinement of in vivo experimental design, and the de-risking of the drug development pipeline to reduce the risk of unexpected adverse effects in the clinic. These added efficiencies also give developers more time to modify the drug design if needed.
What are the benefits of using MPS prior to preclinical testing?
Using these systems before preclinical testing enables fewer but better profiled candidates to be taken into animals, reducing animal numbers, cost, and time. This also contributes to the three R's efforts: replacement, reduction, and refinement.
Understanding the safety profiles of newer drug modalities represents a further challenge due to their highly human-specific targets. These targets are often either not expressed or differentially expressed in human in vitro models, and animals are less suited for their testing - meaning there is a lack of appropriate models for safety profiling.
Human MPS provide these otherwise missing test models, allowing more translatable data generation and enabling improved confidence to progress into the clinic.
Our human DILI assay is performed using our PhysioMimix Multi-chip Liver-12-or 48-plates. These plates are open well and, therefore, accessible at any time for drug dosing and sampling.
The Liver-48 plate is a miniaturized version of the Liver-12, ensuring that the cost of cells and media consumption is equivalent between plates, but the throughput increases four-fold. This allows for further testing of different compounds, concentrations, or conditions. The ratio of cells and media used in the Liver-48 is also maintained, meaning that we can detect the same clinical markers as the Liver-12. Both plates show good tissue formation and functionality over the length of the DILI assay.
For both plate types, the assay procedure and timeline remain the same. The liver's hepatocytes and macrophages are seeded on day zero, and the media is changed daily to optimize the process.
Cells are allowed to acclimatize until day four, when a quality check (QC) is performed to ensure that they are functioning as expected. This involves assaying a variety of cell health and functional markers.
Once the QC is complete, we proceed to dose, which typically occurs every one to two days. Due to the large volumes and ease of access of the liver plates, it is possible to assess relevant clinical markers found in standard clinical liver function tests such as ALT and AST levels within the culture media to support in vitro to in vivo data translatability.
On day eight, the experiment ends, tissues can be removed, and high-content analysis can be completed, such as microscopy, omics, or flow cytometry. These endpoints and the temporal data collected over the experiment can be used to form a detailed risk assessment and safety profile of the drug candidates.
The only markers that the assay cannot measure are blood markers such as bilirubin and prothrombin because it is impossible to replicate these without blood components.
The DILI assay shows excellent sensitivity when compounds with known DILI severities are tested, demonstrating that we can detect a range of DILI risks. For example, a classic DILI pair, highlighted by the IQ MPS consortium, are troglitazone and pioglitazone.
Troglitazone was the first marketed compound for type 2 diabetes. However, it was withdrawn from the market three years later due to many cases of liver failure and related deaths. Pioglitazone was the third marketed compound for type 2 diabetes and is still used as a therapy for insulin resistance today.
When tested in our DILI assay, cell health and functional biomarkers flagged troglitazone’s hepatotoxic risks, whereas pioglitazone showed no hepatotoxic effect.
Tools like PhysioMimix allows its users to enhance their preclinical safety studies, providing human-relevant outcomes that can be translated to the clinic. However, our team has been considering how to more effectively translate data between in vitro and in vivo studies. There is still a gap here, which can make a final assessment of drug candidates challenging.
In preclinical testing, the drug developer remains responsible for species choice. Target binding affinity and the toxicology target also impact the preclinical animal model of choice.
International guidelines specify the need for the relevant use of at least one rodent and one non-rodent species. Rats and dogs are the most used species in toxicology testing across modalities, particularly for small molecules.
We decided that developing rat and dog liver MPS models would be the most effective translational tools in the first instance. However, it is worth noting that in the new modalities space, non-human primates (NHPs) are more commonly used due to having closer target expression profiles to humans in many cases. We are exploring the utility of rat and dog MPS models by culturing primary canine and rat hepatocytes in our Liver-12 plate and dosing with drugs with known interspecies differences. The preclinical animal DILI assay is performed in a similar way to the human DILI assay with seeding on day zero, changing media on day one, a QC check on day four, and dosing through to day eight. This approach facilitates data discrepancy predictions between animal and human outcomes in the preclinical space, potentially reducing the number of drugs that are likely to be safe in humans from being dropped unnecessarily from the pipeline and derisking any that weren’t picked up by animals before entering the clinic. De-risking can be achieved through better informed clinical design for narrow safety margins, or redevelopment of the therapeutic entity via “re-engineering, or ceasing the program before trails altogether.
This approach also lends itself to ethical considerations, including the length of testing required, cost, and availability of the models (especially for NHPs). It promotes responsible use of animals too by safeguarding their use. One animal is sacrificed for hundreds to thousands of in vitro tests. Where toxicity is flagged using MPS, this would better inform go/no-go decisions about proceeding in vivo with that species.
Could you please discuss the proof-of-concept data showing that in vitro MPS models can be translated to in vivo?
A range of compounds were tested in rats and dogs, with no toxicity reported. On entering the clinic, however, elevations in ALT were seen showing liver toxicity. Some of these drugs were subsequently labeled with a box warning, while others were withdrawn from the market.
A good example of this is sitaxentan, a medication used to treat pulmonary hypertension which was removed from the market in 2010 due to concerns over liver toxicity. A review examined studies performed in multiple preclinical animal species where the severity of the toxicity was missed.
The authors highlighted that it was not a deficiency in the preclinical package that resulted in the missed toxicity and translational errors; rather, it was a shortcoming of the preclinical species to predict human outcomes, or the mechanistic details of toxicity.
When we tested this drug in both our animal DILI models, using the PhysioMimix OOC System, it was shown that the rat or dog models did not pick up the toxicity. When tested in the human MPS model, however, toxicity was flagged by all the cell health functional biomarkers. This demonstrates the translatability of both models’ in vivo outcomes.
The IQ MPS consortium also highlighted tolcapone and entacapone. Both compounds were used as therapies for Parkinson's; however, tolcapone was discovered to be fatally hepatotoxic, while entacapone is considered safe. However, there has been an associated increase with ALT in some patients.
Preclinical tests did not identify the heightened risk of tolcapone. However, numerous retrospective rat and dog tests have detected tolcapone-mediated toxicity in this case.
When tested in our model, the human liver MPS captured tolcapone-mediated toxicity. Interestingly, entacapone toxicity was also detected in higher concentrations, demonstrating the sensitivity of our human model.
Rat and dog models were also able to capture the potential risk of tolcapone in DILI patients across different markers, demonstrating the successful translation of data between in vitro assays and in vivo outcomes.
We are continuing to qualify these models using different DILI compounds in the lab to gain a deeper understanding of the full translatability potential of these models.
Finally, please provide an overview of using MPS to enhance preclinical safety and toxicity testing.
To conclude, MPS enhances preclinical safety toxicity testing in several ways. When utilized prior to regulatory-enabling preclinical studies, our preclinical animal Liver MPS models provide an understanding of inter-species differences in toxicity. This allows for more effective in vivo experiment design.
Our human liver MPS assays allow an earlier understanding of human-specific toxicity, and a deeper mechanistic understanding of why toxic events occur due to the high content analysis capability of the approach. Insights from MPS can support stop/go decisions and potentially enable modification of drug design before proceeding candidates to preclinical testing. The translation of human outcomes can be directly linked to clinical observations, which allows greater confidence in progressing a lead candidate into clinical trials.
Watch the Full Webinar
About Emily Richardson
Dr. Emily Richardson is a Lead Scientist in the R&D team at CN Bio. Joining the team in 2020 as a Senior Scientist, she has since led the development of the PhysioMimix® lung and lung-liver microphysiological systems (MPS).
About CN-Bio
CN Bio is a leading organ-on-a-chip (OOC) company that offers a portfolio of products and contract research services to optimise the accuracy and efficiency of bringing new medicines to market. With more than a decade of research and development experience, we aim to transform the way human-relevant pre-clinical data is generated through the development of advanced in vitro human organ models.
CN-Bio's PhysioMimix® OOC range of microphysiological systems (MPS) enable researchers to recreate human biology in the lab. The technology bridges the gap between traditional cell culture and human studies, to support the development of safer and more efficacious therapeutics, whilst reducing the dependence on animal model usage.
CN Bio’s portfolio of products (MPS, 3D validated cells, consumable plates) and services support researchers that require reliable, data-rich, in vitro studies, to uncover novel mechanistic insights into drug or disease mechanism of action.
Sponsored Content Policy: News-Medical.net publishes articles and related content that may be derived from sources where we have existing commercial relationships, provided such content adds value to the core editorial ethos of News-Medical.Net which is to educate and inform site visitors interested in medical research, science, medical devices and treatments.
CN Bio wins ‘Most Impactful Industry Collaboration of the Year’ at the OBN Awards 2023 for ongoing research with the FDA