September 15, 2011 (Vol. 31, No. 16)

Kathy Liszewski

Late-Stage Drug Failures Prompt Push for More Effective Approaches to Designing Therapies

Cell-based assays continue to provide powerful tools to fuel drug discovery. However, costly late-stage failures are driving the industry’s search for improved approaches and novel tools for interrogating cells. Throughput, reliability, cost, and physiologic relevance of cellular models for predictive toxicology remain central issues in the field.

Visiongain’s “Cell-Based Assays” conference next month will feature topics ranging from technological advances in cellular models to challenges to traditional drug discovery paradigms. Additionally, presentations will provide the latest developments in stem cell biology and demonstrate how applying engineering/computing principles to living systems could revolutionize the design of new therapies.

“In many ways, progress in the drug discovery industry has been quite disappointing in the last 10 years,” reports Mark Slack, Ph.D., group leader of cellular assays, Evotec.

“The inertia and lack of productivity relate to failures in efficacy, poor clinical translation, and stagnant methodologies. The old one-gene, one-drug paradigm is no longer relevant, giving way to phenotypic approaches, capturing targets in their functional background. The landscape is beginning to change with an emerging paradigm that considers the physiology of diseases and the whole organism.

“We are seeing that high-throughput screening formats are employing more physiologically relevant cellular models including use of primary cells and label-free, nonintrusive methods to interrogate new compounds. One rapidly emerging area is the use of stem cell technologies.

“Because of difficult ethical issues in the generation of such cell lines, landmark work has studied and now demonstrated that it is possible to induce the reversal of somatic tissues into stem cell populations (i.e., induced pluripotent stem cells, or iPSCs).

“We can perform assays from affected patients via iPSC technologies and generate cells expressing the disease phenotype. These will be very useful for drug discovery for specific diseases such as Parkinson and Huntington diseases.”

Another area that has benefitted from technological advances is automated electrophysiology. Tiny patch clamp electrodes are now applied to individual cells in a 384-well format, allowing the screening of tens of thousands of compounds.

“This has been an underexploited target class,” Dr. Slack says. “With the emergence of automated patch-clamp assays, a major bottleneck was overcome. However, we now need to deal with the high cost per data point and the relatively low throughput and efficiency of the devices.”

According to Dr. Slack, one of the more exciting recent developments slowly establishing itself in mainstream drug development is atomic force microscopy (AFM). Developed as an add-on to enhance the scanning tunneling microscope, AFM adds an atomically sharp tip.

“By monitoring the surface of biological specimens with the sensitive tip, one can actually see the cellular membranes and even their subcellular components at high resolution. Ultimately, these and other emerging technologies are providing more physiologically relevant models.”

Hepatotoxicity Testing

Unanticipated toxicity and adverse drug reactions after licensing a drug are the leading causes of late-stage attrition and withdrawal of drugs from the market. “Up to 30% of compound failures happen due to toxicity and issues of clinical safety,” notes Neil A. Hanley, M.D., Ph.D., professor of medicine at the University of Manchester.

Liver toxicity tops the list of drug-induced injury. “The underlying mechanisms of damage are quite complex and not completely understood. A major issue is that such toxicity is not reliably testable in a cell culture system, nor completely accurate in animal models.” Dr. Hanley suggests that, not only are better cellular models for hepatotoxicity needed, but that generating hepatocyte-like cells from human embryonic stem cells (hESCs) or induced pluripotent stem cells (hIPSCs) may hold the answer as an assay of choice.

“In theory, these stem cells could be utilized to generate all the different cell types of the adult liver. This is a key advantage that may allow better mirroring of the complexity of an intact organ. There have been a number of advances aimed at differentiating hESC into liver cell types. Three components needing to be addressed are soluble growth factors, support cells, and an extracellular matrix medium.”

Dr. Hanley indicates that there are many challenges that remain before this system is ready for more widespread use. “Many academic laboratories can reliably generate progenitor-like cells. That is about three-fourths of the journey. The last one-fourth, getting to the mature liver cell, is most likely contingent on the precision of preceding steps plus optimizing conditions for the final differentiation.

“An example of the last step is optimizing extracellular matrix composition to provide a scaffold for growth and differentiation.”

Before big pharma can utilize hESCs for hepatotoxicity, testing the industry must also tackle issues of scale-up and consistency of production, suggests Dr. Hanley. “Scale-up, the bioprocessing step, is a huge science in itself.”

Noninvasive Analysis of Live Cells

Being able to noninvasively monitor the same live cells in culture over extended time periods could have important implications for monitoring of cell phenotype and drug screening, suggests Molly M. Stevens, Ph.D., professor of biomedical materials and regenerative medicine at Imperial College London.

“Stem cell research and regenerative medicine rely on understanding and manipulating cells that can differentiate and interact with their environment. Primary stem progenitor and lineage-specific cells are the gold standards.”

However, Dr. Stevens also notes that these can be unreliable due to their variable sources or loss of phenotype when maintained in long-term culture. Raman spectroscopy may help overcome these limitations. “Raman microspectrometry is a way to noninvasively characterize live cells. The Raman spectrum provides an intrinsic, global biochemical fingerprint with molecular-level data on cells.”

Raman spectroscopy is a laser-based analytical optical technique that measures photon scattering from chemical bond vibrations. The energy differences allow one to identify and characterize specific chemical bonds present. The Stevens group recently reported in Nature Materials on the use of the technique to distinguish how different stem cells can form bone tissue with a different chemical make-up.

According to Dr. Stevens, Raman spectroscopy can yield some important advantages over conventional cytochemical methods because it provides a rapid, noninvasive means to analyze cells without the need for fixatives or labels.

“An additional advantage is that, while most biological assays characterize only one marker, Raman spectroscopy can provide a cell-specific biochemical signature. We interpret the signatures with the help of our own sophisticated multivariate statistical analyses and can then detect subtle yet reliable changes in cell phenotype. Applications include toxicity testing of drugs, identification of cancerous cells, and as a tool to characterize cellular processes.”

Automating 3-D Construction

Increasing numbers of scientists are turning to 3-D substrates on which cells can be grown. Such substrates more closely mimic the natural in vivo physiological state. However, building 3-D tissues can be a slow and inconsistent process. Automating such construction to enhance throughput faces several challenges.

“Ideally, tissues should be created in a quick, consistent, and simple way. This hasn’t been practical before,” reports Rosemary Drake, Ph.D., CSO, TAP Biosystems.

“It’s been a costly and inconvenient process with poor reproducibility. We addressed these issues in collaboration with leading tissue-engineering academic scientists and developed the RAFT system (Real Architecture For 3-D Tissue). This includes a workstation, consumables, and reagents for making a range of multicellular 3-D tissue models.”

The process can take less than an hour. “Extracellular matrix is largely composed of collagen. In the RAFT system, we mix collagen with cells (such as from epithelium, endothelium, nerve, smooth muscle, tendon, and bone) in a 12-, 24-, or 96-well format to form a cell-seeded hydrogel.

“Next, absorbent plungers simultaneously apply gentle compression and absorb some of the liquid from the gel. This results in a 50–100 fold increase in the concentration of the cells and collagen, giving a consistent transparent tissue model in the bottom of the well. During culture, matrix-rich tissue is created.”

Dr. Drake said that “the density of the collagen matrix is the closest we can get to a tissue-like environment, and cells respond to this in a similar manner to cells in vivo. Therefore, these tissue models have broad applicability in cell-based screening, target validation, lead optimization, and toxicity testing.”

She cited a practical example of the technology. “Our academic collaborators seeded human limbal epithelial stem cells onto a layer of fibroblasts in compressed collagen. After three weeks of culture, the cells formed tissue strikingly similar to the human central cornea.”

Although not all assays need to be done in 3-D, Dr. Drake explained that many applications would benefit from such an approach.

“In particular, it would be useful in improving our understanding of how cancer cells invade and move through tissues. Automating such processes provides a consistent way to interrogate more complex cellular processes.”


TAP Biosystems’ RAFT™ (Real Architecture for 3D Tissue™) system is used for the generation of consistent multicellular 3-D tissue models.

Executable Cell Biology

Are living processes logical and predictable? Can we develop biological models to more fully understand the complexities of biological systems?

Yes, says Jasmin Fisher, Ph.D., researcher, executable biology, Microsoft Research Cambridge. “Over the past decade, we have generated and accumulated so much data that it exceeds the human capacity to analyze it. Data from microarrays, genome sequencing, and other large-scale technologies requires sophisticated analysis by computational methods.”

According to Dr. Fisher, “Executable cell biology is a concept gaining momentum that suggests we can develop techniques for creating dynamic models that capture time- and space-dependent processes and can automate reasoning and analysis. Such large-scale models, based on formal methods from engineering and computer science, could revolutionize biology and medicine and enable the design of new therapies.”

Dr. Fisher’s group is studying how cells make decisions, with a particular focus on the process where stem cells commit to a single lineage (leading to a single-cell type) while having the ability to undergo multilineage differentiation.

“The elucidation of intricate mechanisms that govern stem cell decisions is essential for understanding normal development. Moreover, defects in these mechanisms play an important role in diseases such as cancer.”

“We first generate a detailed map of different interactions that increase or inhibit different cellular processes. We then add dynamicity by inducing from these interactions ‘state transitions’—when a specific signal or a gene gets turned on or off. This allows us to test the translated program to determine if it correlates with cell behavior. This also can allow us to capture dynamically such processes and find out what happens first, second, etc., and how feedback comes into play.”

For example, Dr. Fisher and colleagues adapted software originally designed to find errors in microcircuitry, only they used it to study C. elegans. They found a similar warning in a simulation of signaling pathways in the worm. They predicted and later experimentally verified the existence of a specific mutation that produced a functional defect in cell growth.

“One of the ultimate goals of executable biology is to simplify model building so that any scientist can use it. My personal vision is that within five years this will become a mainstream technique in biology.”

Stem Cell Advances

The use of stem cells for drug screening and predictive toxicology is a field that has begun to mature, according to Stephen Minger, Ph.D., global head of R&D, cellular technologies, GE Healthcare Life Sciences.

But, he points out, there is room for much-needed improvement. “Some of the drug discovery industry has an ingrained approach that hasn’t really changed much for more than 50 years. Yet traditional technologies remain poorly predictive.

“There are still many compounds that are pulled at late stages of clinical trials, or even after reaching the market, due to previously undetected toxicity issues. Vioxx is an example.”

GE Healthcare has worked with Genentech to assess Cytiva™ Cardiomyocytes, developed from human embryonic stem cells (hESCs), for use in cardiotoxicity testing. “We performed a retrospective blinded dose- and time-dependent study of more than 26 compounds with known preclinical and clinical cardiotoxic effects.

“We found that Cytiva cardiomyocytes showed a good correlation with previously reported clinical cardiotoxicity. Another advantage of using hESC-derived cells is that sometimes you can gain a greater understanding of the underlying mechanism of toxicity—for example detecting mitochondrial involvement.”

Dr. Minger sees the field continuing to progress both in the use of hESCs in drug discovery and also as a direct therapy.

“Already,” he says, “a variety of stem cells are being used therapeutically in more than 300 clinical trials worldwide, with hESCs currently being trialed for such indications as spinal cord injury and age-related macular degeneration. I feel that there will be many more applications on the horizon.”

Creating a Natural Environment

Tissue engineers have been defining new principles for developing functional substitutes for tissues that have been damaged, according to Tetsuro Wakatsuki, Ph.D., co-founder and CSO of InvivoSciences. He says his company applied those principles to develop 3-D tissue constructs for which drug developers can perform functional assays of cells grown in an arguably more natural 3-D environment.

Generally, he adds, stiff surfaces are used for 2-D cell culture. “The extracellular matrix (ECM) in which native cells reside, however, can be deformed, stiffened, or degraded to adjust their mechanical environment. The ECM stiffness can even specify stem cell lineages, suggesting the importance of the mechanical environment in development.

“In many diseases, including skin and cardiac fibrosis, mechanical properties of tissues are damaged,” he continues. “Cells in those diseased tissues lose their mechanical homeostasis. Therefore, candidate compounds for reconstituting mechanical homeostasis can be identified through the mechanical measurements of cells and ECMs and monitoring physiological changes. Cell and tissue mechanics are also responsible for regulating cancer metastasis, wound healing, and embryonic development.”

To achieve high-throughput profiling of cells and tissue mechanics, the company developed miniature hydrogel tissue constructs and an automatic device, the Palpator™, to quantify the resulting mechanical properties after exposure to compounds.

“The Palpator automatically inserts a probe into each well to assess and quantify tissue mechanics. Additionally, a conventional microplate reader measures intra- and extracellular biochemical activities using optical biological probes such as dyes that report mitochondrial membrane potential.

“Thus, we combine biomechanical and biochemical assays of cells and ECMs to comprehensively analyze the biological benefits of treatments. We are envisioning that this system will be a useful technology to evaluate efficacy, toxicity, and mechanism of action for early stages of drug development.”


InvivoSciences’ MC-8™ is capable of growing ET constructs without a support layer and is a reportedly reliable method for growing contractile cells in 3-D constructs suitable for measurement of contractile force or tissue stiffness using the company’s Palpator.

Previous article2011 Lasker Awards Honor Research on Protein Folding and Development of Artemisinin
Next articleScientists Complete Genome Sequencing and Analysis of 17 Laboratory Mouse Strains