Drug candidates often fail in the clinic for two reasons: they lack efficacy or possess an adverse toxicity profile. These failures can come at a high cost, especially when drug candidates have already progressed along the development pipeline to clinical trial.

Advances in three-dimensional (3D) cell culture systems are helping address the above challenges and improve the likelihood of success for pipeline assets. From disease modelling and validation of novel targets to screening for safety and efficacy, 3D cell cultures offer the exciting potential for the development of novel medicines and increased productivity.

This article describes the benefit of 3D cell culture on the drug discovery workflow, with a focus on lead optimisation in which compounds are assessed for their potential metabolic liability and off-target toxicities. The powerful combination of human-induced pluripotent stem cells (HiPSCs) and 3D cell culture systems are also highlighted.

The 3D Cell Culture Advantage

A survey was recently conducted of researchers in academia and industry to explore the opportunities, challenges and future of 3D cell culture. The need for greater biological relevance was cited by many survey respondents as the driving force behind adoption of 3D cell culture systems with the most common applications being disease modelling, oncology and toxicity screening.

The advantage of 3D cell cultures for delivering greater biological relevance is clear. When grown in two-dimensional (2D) monolayers, cells reside on a continuous layer of matrix, are not exposed to soluble gradients and possess an unphysiological apicalbasal polarities. With use of 3D cell culture systems, along with matrices and scaffolds that simulate the extracellular matrix (ECM), tissue microenvironmental cues are retained. Local soluble chemicals and the ECM environment facilitate correct cellular differentiation, function and communication. As a result, cells grown in 3D cultures more closely mimic in vivo physiology in terms of morphology, structural complexity and phenotype.

When applied to drug discovery and safety profiling, the advantages of human 3D model systems are clear:

– 3D cell cultures provide more architecturally relevant barriers for compounds to traverse than present in 2D cell cultures. Tissue absorption is a major factor in determining efficacy and many compounds have similar access to all cells in flat 2D cultures, resulting in overestimation of efficacy.

– Mixed cell type 3D culture systems allow better modelling of cell-cell interactions than is possible with 2D cell culture systems and can even maintain neuronal inputs to the organ or organoid.

– 3D cell cultures can be long-lived, in some cases for months. This temporal aspect of cell culture is important when modelling diseases that are slow to develop, such as neurodegenerative diseases. The relatively long life of the culture provides a system to identify pathological defects emerging in the tissue over time, as well as monitor the prolonged exposure of cells to pharmacologically active compounds.

Drug-Induced Liver Injury (DILI) assessment using 3D hepatocyte cultures

A comparison of 2D and 3D primary human hepatocytes for assessment of drug compounds to cause liver injury demonstrates the advantages of 3D approaches. Drug-induced liver injury (DILI) is a leading cause of drug attrition and clinical failure and can manifest as mitochondrial dysfunction, apoptosis, necrosis, hepatocyte hypertrophy and hyperplasia, fibrosis and cholestasis. The impact of DILI on drug development is clear – among the drugs withdrawn due to toxicity during the period of 1990-2010, 26% were attributed to DILI.

While 2D cell culture systems play a pivotal role in research, classical systems do not reflect the complexity of liver tissue. Although hepatocytes are often the in vivo targets of DILI, primary human hepatocytes are not suitable for hepatotoxicity tests under conventional 2D monolayer culture conditions, due to the rapid loss of their hepatic phenotypes, functions and cell viability.

In contrast, 3D cell culture systems sustain cell viability, maintain in vivo phenotypes as well as genomic and proteomic expression profiles. Compared with other 3D systems, spheroid hepatocyte cultures require fewer cells, are technically easier to establish, and are adaptable to high throughput screening.

There are several well-established protocols, tools and reagents to support primary 3D liver spheroids and toxicity assays. Single spheroids can be generated using seeding densities with less than 5,000 cells per well on 96-well spheroid plates. Morphology and sizes can be routinely monitored and bioluminescent ATP assays for viability can be performed directly in the spheroid microplate. The primary human hepatocytes maintain Cytochrome P450 (CYP) drug metabolic activity in long-term cultures of up to four weeks. In contrast, 2D human hepatocyte cultures are short-lived, typically lasting only one week.

As shown in Figure 1, primary human hepatocyte spheroids are more sensitive than 2D cultures for detecting liver toxicity.

Amiodarone, an approved drug with a FDA black box warning used to treat life-threatening arrhythmia, was added to 2D and 3D primary human hepatocyte cultures. Dose response curves were generated, measuring cell viability via an ATP assay. The IC50 reported for the 3D cell cultures (26μM) was significantly lower than that from the 2D cultures (209μM).

To validate this approach, a 100-compound screen was performed using DILI and control compounds selected based on recent publications to examine the response of 2D monolayer versus 3D spheroid hepatocytes (1). A two-week hepatotoxicity assay with three repeated doses of testing compounds was carried out with 3D primary human hepatocyte spheroids made from cryopreserved primary human hepatocytes in Corning® 96-well spheroid microplates. For comparison, 2D monolayer cultures from the same primary human hepatocytes were used in a single-dose, short-term cytotoxicity assay.

Bioluminescent ATP assays were performed, and eight-point dose response curves were generated. IC50 values for each compound were calculated. Using clinical Cmax values as drug exposure references and the margin of safety (MOS) approach, quantitative analysis of assay specificity and sensitivity were performed (1). Results indicated that 3D primary human hepatocyte spheroids are two to three times more sensitive to DILI compound treatment than 2D cultures (Table 1).

Use of this 3D approach enables earlier and more accurate screening of compounds for potential toxicity in the drug development workflow and allows for derisking of compounds selected to advance in the pipeline. DILI assessments conducted in 2D culture systems may provide misleading results and allow compounds with potentially serious toxicity flags to progress to further development only to fail at a much later, more costly point.

A recently-published study assessed the predictive accuracy of primary human hepatocytes in 3D spheroid culture on a panel of 123 drugs with and without causative involvement in clinical DILI events (2). The 3D model achieved 69% sensitivity and 100% specificity. The authors concluded that their spheroid system “exceeded both the sensitivity and the specificity of all previously published in vitro assays at substantially lower exposure levels.”

Applying 3D Cell Culture to Induced Pluripotent Stem Cells (iPSCs)

The convergence of 3D cell culture systems and the ability to culture HiPSCs into functional cell types and tissues creates additional advantages for disease modelling, target identification and lead optimisation.

HiPSCs can be developed by genetic reprogramming of somatic cells such as skin fibroblasts to an embryonic-like state by viral introduction of transcription factors that induce pluripotency. The resulting iPSCs are self-renewable and can be directed to differentiate into a variety of cell types. HiPSCs exhibit disease phenotypes close to the human pathology, particularly when cultured under conditions that allow them to recapitulate tissue architecture in the form of multicellular spheroids or organoids.

When cultured under 3D conditions, human iPSCs clearly provide optimised systems that more accurately reflect disease-related target mutations, compound pharmacology and toxicology. HiPSCs and 3D cell culture systems can also be used to generate complex, multicellular systems in which cells are spatially arranged similarly to tissues in vivo. These organoids and more complex organ systems are being leveraged for lead optimisation, including estimations of preclinical toxicity and potential metabolic liability.

Below, we explore the use of HiPSC-derived 3D liver tissue for drug discovery and lead optimisation.

3D HiPSC-derived liver tissue

Drug discovery

HiPSC-derived human hepatocytes are being employed in drug discovery against molecular targets that cause liver disease. For example, HiPSC-derived hepatocytes generated from patients with homozygous familial hypercholesterolemia enabled identification of compounds that lower serum lowdensity lipoprotein C (LDL-C). Cayo et al have identified several cardiac glycosides that lowered expression of apolipoprotein B in the human hepatocytes in vitro (3). Patients treated with these same cardiac glycosides also had reduced serum LDL-C levels, collectively supporting the potential use of cardiac glycosides to treat hypercholesterolemia.

Lead optimisation – toxicity evaluation

Primary human hepatocytes can be used to profile compounds for their potential to induce liver toxicity and for their metabolic liability in patients. Because human hepatocytes may respond differently to drugs than rodent- or tumour-derived hepatocytes, their use provides major advantages in testing for drug safety at the preclinical discovery stage. However, the source of these cells is limited, restricting their widespread use in drug safety screens.

In contrast, human hepatocytes derived from HiPSCs can be produced in high abundance and reproducibility, with the desired genetic profile. 3D cell culture systems developed using HiPSC-derived hepatocytes offer significant value in evaluating the potential of drugs for toxicity.

Liver cells grown in organoid culture have advantages over cells grown in 2D as they develop relevant cell-cell interaction due to the polarity of the hepatocytes. This polarity is critical for cell metabolism which, in turn, is a feature essential for drug metabolism. Most importantly, human liver organoids respond differently to drugs than similar cells grown in 2D culture.

Because of differences in expression of metabolising P450 enzymes between individuals, one can envision drugs screened against large populations of HiPSC-derived hepatocytes in 3D to provide a better preclinical assessment of the potential to induce toxicity than presently available with any other in vitro screening system.

Compound screening of potentially hepatotoxic compounds against HiPSC-derived hepatocytes have yielded similar results as when the same compounds were tested against primary human hepatocytes.

These data and others validated the use of HiPSC-derived hepatocytes for preclinical lead optimisation and showed that the cells had a high degree of clinical predictability.

Lead optimisation – compound metabolism screening

Genetic variability may cause differences from one individual to another in P450 enzyme expression, which can result in individuals being either ‘slow’ or ‘fast’ metabolisers of many commonly prescribed drugs. HiPSC-derived hepatocytes can be used to evaluate the metabolic liability of compounds and have the potential to be used to assess the influence of patient genetic variability on drug metabolism.

The utility of the HiPSC-derived hepatocytes in this regard was reported by Takayama et al, who used HiPSC-derived hepatocytes isolated from individuals with or without a single-nucleotide polymorphism (SNP) in the P450 enzyme, CYP2D6 (4). The SNP in the CYP2D6 genes reduces enzymatic activity and individuals with this SNP are slow metabolisers of drugs, including tamoxifen.

Normally, CYP2D6 metabolises tamoxifen to a cytotoxic metabolite in breast cancer cells. Co-cultures of MCF7 human breast cancer cells with HiPSC-derived hepatocytes from individuals with the normal CYP2D6 enzyme showed substantial loss of cancer cells after tamoxifen treatment. As expected, cancer cells co-cultured with hepatocytes from individuals with the CYP2D6 SNP showed resistance to tamoxifen, because less of the cytotoxic metabolite was produced, and ultimately diminished breast cancer cell toxicity.

Humans vary in the expression levels of P450 enzymes, and therefore metabolise drugs differently. HiPSC-derived hepatocytes therefore provide better models of the diversity of metabolic responses existing in patient populations. Therefore, they may provide better predictors of clinical efficacy and safety than many other models of liver function.

These properties of HiPSC-derived hepatocytes may also facilitate the development of unique multi-tissue/organ systems to determine human drug toxicity. End-organ toxicity is a major limiting factor in drug development. Consequently, approaches that predict preclinical drug safety may provide the means to reduce the attrition rate of novel leads in the clinic.

In this respect, microfluidic systems have proven to be very useful. Here, different organoids derived from HiPSCs – such as hepatocytes, cardiac myocytes, gastrointestinal and kidney cells or neurons – are grown in separate, but interconnected, chambers, and can be used to profile drugs and their metabolites simultaneously for potential toxicity in different tissues in the body (5).

The way forward

While 3D cell cultures present many advantages in drug discovery, hurdles remain before their widespread adoption can be achieved. Our recent survey highlighted key areas which must be addressed. The first is the need for advances in imaging and detection techniques to effectively monitor tissues in 3D cell culture.

Despite increasing adoption of 3D approaches, challenges remain as to the way spheroid cultures are characterised and analysed. This analysis is particularly important when screening compounds in spheroid culture in which cells can organise into a proliferating zone, a quiescent region of viable cells and a hypoxic, necrotic core. All regions of the spheroid must be analysed to properly assess the impact of drug candidates.

The thickness of 3D cell cultures (typically >100m) causes light scattering and limits imaging to the surface-layer cells, preventing complete characterisation of the cell population within wholemount 3D cell cultures. This introduces a sampling bias in imaging analysis, since only the exterior cells can be imaged where concentrations of nutrients, oxygen and drug compound are greatest.

While the high cellular density of 3D cell cultures scatters light and limits the ability to characterise the entire spheroid, several approaches are available for rendering tissue transparent to enable analysis. Commonly-used clearing techniques have a number of disadvantages when being considered for high throughput, cost-efficient screening, however. Protein hyperhydration can require days to weeks for clearing and the process severely reduces tissue integrity. This approach is also not compatible with immunolabelling techniques.

Use of highly lipophilic organic solvents are not compatible with plastic well plates. Hydrogel embedding techniques are slow, requiring days to weeks to clear even small tissue samples. This approach uses a complex processing technique that is not amenable to a well plate format and thus not amenable to medium and high throughput applications.

We have collaborated with Visikol to develop an organic solvent approach (Visikol® HISTO-M™) that replaces the water in the cytosol and makes the cytosol match the refractive index of lipids and proteins causing the spheroid to become transparent. Untreated, each cell acts as a lens and scatters light through the spheroid because the proteins and lipids have a high refractive index while the cytosol has a low refractive index. The reagent rapidly renders tissue transparent, dramatically improving both wide field imaging and confocal microscopy and visualisation of the spheroid interior.

A second challenge identified via our survey is the need for novel tools and technologies to support development of spheroid and organoid cultures. To address this challenge, many new and highly innovative products are reaching the lab bench and screening laboratories and enabling critical activities such as bulk spheroid production, sorting and dispensing. For example, microcavity vessels enable growth of more than 3,000 uniformsized spheroids per T-25 flask. The spheroids are confined to <400μm and maintain their shape up to 30 days in culture. Additional advances are making use of fluorescence to sort and dispense spheroids.

A final challenge that must be addressed to enable widespread adoption of 3D culture systems is the need for greater expertise combined with more training and support in establishing and optimising 3D cell cultures.

The number of publications citing use of 3D cell culture systems in the past eight years has grown five-fold and provides a robust resource for the research and discovery communities. At the same time, it is incumbent on companies such as Corning to continue development of methods and technologies such as those for 3D spheroid cultures of human hepatocytes and offer expert technical and applications support. We also listen to the needs of customers as they progress towards adoption of 3D cell culture systems and have developed products and protocols based on their input including those for immune cell invasion, the blood brain barrier and intestinal organoids.

Summary – Increasing Drug Discovery Success With 3D Cell Culture

In summary, remarkable advances have been made in the use of 3D cell culture systems and offer important advantages across the drug discovery workflow. As remaining hurdles are addressed, more widespread adoption of 3D approaches will improve productivity and lead to fewer costly late-stage failures. Better disease models will become available, including those for oncology, neuroscience and cardiovascular diseases.

With incorporation of HiPSCs, these models will allow researchers to discern differences among patients and guide development of more personalised therapeutics. Further along the workflow, 3D systems will enable more effective compound screening and deliver better hits. Compound libraries originally screened using 2D systems will be rescreened to find active hits; approved drugs will be screened in 3D to potentially reposition or expand indications.

Finally, 3D systems will allow more accurate assessments of toxicity and metabolic liabilities. The end result of incorporation of 3D systems across the discovery workflow will be compounds entering the clinic with a better prediction of efficacy and tolerance. DDW

—

This article originally featured in the DDW Winter 2018/19 Issue

—

Dr Richard M. Eglen is Vice President and General Manager of Corning Life Sciences. Eglen joined Corning in 2011 with more than 35 years’ experience in the life sciences industry. He has authored more than 325 publications, book chapters and patents, and serves on numerous industry, academic advisory and journal editorial boards.

Dr Feng Li is a Senior Development Scientist of Corning Life Sciences and expert cell biologist with more than 12 years’ industry experience. In recent years, he has focused on spheroid models using primary human hepatocytes and Corning 3D technologies for applications such as liver toxicity testing and disease modelling.

Dr Anthony G. Frutos is the Business Technology Director for Corning Life Sciences. He has 19 US patents and is an author on more than 39 technical publications. He holds a Bachelor’s degree in chemistry from Brigham Young University and a PhD in analytical chemistry from the University of Wisconsin, Madison.

References

1 Li, Feng et al. 3D Primary Human Hepatocytes (PHH) Spheroids Demonstrate Increased Sensitivity to Drug- Induced Liver Injury in Comparison to 2D PHH Monolayer Culture. www.corning.com/catalog/cls/documents/applicationnotes/ CLS-AN-514.pdf.

2 Vorrink, Sabine et al. Prediction of Drug-Induced Hepatotoxicity Using Long-Term Stable Primary Hepatic 3D Spheroid Cultures in Chemically Defined Conditions. Toxicological Sciences 163,2 (2018): 655-665.

3 Tulloch, Nathaniel L et al. Growth of engineered human myocardium with mechanical loading and vascular coculture. Circulation research vol. 109,1 (2011): 47-59.

4 Takayama, Kazuo et al. Prediction of inter-individual differences in hepatic functions and drug sensitivity by using human iPS-derived hepatocytes. Proceedings of the National Academy of Sciences of the United States of America vol. 111,47 (2014): 16772-7.

5 Oleaga, Carlota et al. Multi-Organ toxicity demonstration in a functional human in vitro system composed of four organs. Scientific reports vol. 6 20030. 3 Feb. 2016, doi:10.1038/srep20030.