From computer simulation to stem-cell cultivation, research into new drug treatments is being transformed by innovative technologies and an improved understanding of biology. All this is helping to provide earlier and better predictions of safety and efficacy before products are even tested in patients.
Medicine development has changed dramatically since the 1950s, when inadequate testing of the pregnancy morning-sickness treatment Thalidomide led to the birth of thousands of children with shortened limbs. Yet more recent withdrawals of “blockbuster” drugs, from the diet treatment Fen-Phen in 1997 to the painkiller Vioxx in 2004, show that even modern drug testing gives an incomplete picture of the risks and benefits.
The development of more reliable ways to simulate diseases and their responses to drugs has many benefits. It minimises the potential risks of proceeding with experimental treatments in healthy volunteers and then patients, and reduces the unnecessary use of animals. (Testing on animals is controversial, and often has only a limited value in demonstrating how humans will respond.) In the process, it also allows manufacturers to kill off dangerous or ineffective drugs more rapidly, thus reducing the costs of innovation and freeing up money for a larger number of researchers to reinvest in more promising treatments.
The techniques employed range from “in silico” analysis by computer through “in vitro” techniques in test tubes to improved “in vivo” testing in animals and enhanced feedback on drugs used in humans. Here are four of the most interesting current experiments in drug testing.
A human on a chip…
Jim Powers, CEO of Virginia-based company HemoShear, points to the need to marry engineering knowhow with biology in simulating human systems in cells.
“We are restoring the biological properties of cells in the lab,” he says. “We are replicating the biology of human disease on the bench, by recreating organ function.”
HemoShear offers testing that reproduces the shear forces resulting when blood flows through organs. “Without restoring critical functions to the cell, the biology is instantly changed once it is taken from the human body and put into a dish,” says Powers.
Harvard University’s Wyss Institute is among other groups exploring microphysiological systems that simulate interactions between cells and circulating fluid. The aim is to mimic the complexity of organs including the kidney, liver, lungs and gastrointestinal tract.
By integrating cells from 10 or more interconnected organs, this will allow a better understanding of toxicity, treatments and disease states. Scientists are even dubbing it a “human on a chip”. It has received funding from a variety of sources, including People for the Ethical Treatment of Animals (Peta), the campaign group which seeks alternative forms of testing.
…and a heart in a dish
From his laboratory in Cardiff, Stephen Minger, chief scientist at GE Healthcare Life Sciences, oversees the production of billions of identical versions of an item that takes a month to make. “It’s an artificial heart in a dish,” he says.
His team produces cardiomyocytes – beating heart cells derived from human embryonic stem cells, which are harvested, frozen in liquid nitrogen and sold to other industry labs for testing. These are helping to address one of the greatest concerns with experimental drugs: the risk of toxicity in heart cells, causing arrhythmias that trigger heart attacks and strokes. Traditionally, tests used animals and donated human adult heart cells from the recently deceased. Stem cells allow more consistent production of uniform, healthy cells. “It’s like making a jet engine,” says Minger. “We want a very standardised, reproducible result. We can control the process.”
Efforts are under way to expand the use of stem cells, and one avenue is to extend the variety of the adult cells they can mimic. David Hay, principal investigator at the University of Edinburgh’s MRC Centre for Regenerative Medicine, has developed a technique to produce hepatocytes, or liver cells, from induced pluripotent stem cells – embryo-like stem cells artificially derived from adult cells. He commercialises these through his spin-off company FibromEd and also sees potential for stem cell-derived brain cells and those for other organs.
Hay has also developed a synthetic surface to allow stem cells to last up to four weeks – which is four times longer than adult cells and current stem cells last. “Where some of the gold-standard models are failing is the length of time you can culture in a dish,” he says.
A final area being explored is the use of different stem cells to better reflect genetic variations between ethnic groups. At present, most stem cells derive from a line developed in Wisconsin in the 1990s from a single donor.
Pooling the data
Pharmaceutical companies may be used to competing with each other – and benefiting from each other’s failures – but more than a dozen of them have now joined forces with academic institutions under the umbrella of the joint industry and EU-funded Innovation Medicines Initiative on the “eTox” project. By sharing data on compounds they have tested in previous years, they are developing computer models to predict a drug’s toxicity, or how damaging it would be on a human, based on its chemical structure.
“There is a strong spirit of co-operation through public-private partnerships,” says Patrick Wier, GlaxoSmithKline’s vice-president for clinical safety. “Everyone appreciates the complexity of drug safety and recognises that we need to work together.”
With the Hamner Institutes in North Carolina, the company and others are also developing a “virtual liver” to predict injury to the organ from medicines.
Meanwhile Mark van Mierle, managing director of Elsevier Life Science Solutions, says his company uses semantics, statistics and scientific expertise to trawl through published academic articles, regulatory submissions and reports on side effects in order to identify mechanisms, drug “pathways” and “off-target” effects to help clients make better research decisions.
Such approaches will intensify with the advent of “big data” in healthcare, fuelled by growing transparency over clinical trial results and the accelerating processing power to assess it. Analysis of past information cannot allow for all the “unknown unknowns” of biological systems and their interactions with new compounds. But the more information that is collected, shared and analysed, the more chance there is of reducing the risks of repeat mistakes.
Genetic modifications in mice – and fish
At his laboratory at the Roslin Institute in Edinburgh University, Peter Hohenstein is developing new techniques to better understand kidney disease. His approach uses transgenic mice – animals with additional, artificially introduced genetic material.
“There’s a problem of limited predictability with many animal models,” Hohenstein says. “Better transgenic tools and techniques now becoming available could make better models.”
His work includes the use of “knockout” mice, in which some genes have been removed and replaced, in order to mimic Wilms’ tumours, a form of kidney cancer in children. “The best way to better drugs is better understanding of disease. A driving force is the quality of the model.”
Hohenstein argues that disease research will always require the use of animals to simulate biological systems, but that there are ways to improve efficiency. The three or more modifications he needs have required three generations of mice, resulting in many that do not carry the correct mutations and cannot be used for his experiments or others. He is developing approaches that reduce the high proportion of “surplus” mice by placing three mutations in a single stem cell, designed to immediately make the mutant mice needed for his experiments.
The longest track record for genetic modification is in mice, but there is now greater use of zebra fish, which have advantages including their transparency – which allows for easier observation – and the fact that their young develop outside the mother. In future, Hohenstein sees scope for the use of rats and even pigs, whose characteristics, including their size, provide greater parallels with humans.