Building a better drug
Out of the plant, into the lab
In 1962, a botanist named Arthur Barclay hiked into a forest of towering pine and Pacific yew trees in western Washington state to gather bagfuls of bark, leaves and needles. These specimens — a few of the roughly 200 plant samples he collected that year — were key ingredients in a multiyear effort by the National Cancer Institute to search for plant-based sources of anti-cancer therapies.
Eventually, this program would analyze some 35,000 plant samples, yielding, among other things, a promising new cancer-fighting chemical isolated from Pacific yew bark collected by Barclay and his team. That drug, called taxol, blocked cells from dividing and, in early trials, looked promising for treating a variety of cancers.
“Taxol at that time was a special opportunity. It was the first of its kind for a new way of stopping cell division,” says Paul Wender, PhD, a Stanford professor of chemistry. The drug was effective in treating more than 30 percent of women with ovarian cancer, a rate unheard of at that time.
Wender was one of many researchers who recognized the limitations of using a natural source for the drug — harvesting the bark killed the yew tree — and began trying to synthesize taxol in the lab. He was successful, as were others who developed an even more efficient method used to make the drug today.
Although taxol is now produced without destroying the yew tree, not all drugs derived from plants have an alternate source. More than half the drugs people take today were originally isolated from plants. We are still reliant on those plants, the farmers who grow them and the environment that sustains them for many of our critical drugs.
Concern over this precarious drug sourcing from plants drove Stanford’s Christina Smolke, PhD, and Elizabeth Sattely, PhD, to investigate ways of using other organisms to act as factories for the chemicals. In addition to preserving the environment and creating a more stable source of the drug, this alternate method makes it easier to modify the drug to be safer or more effective.
Smolke, an associate professor of bioengineering, recently engineered yeast to produce hydrocodone, a painkiller derived from the opium poppy. Sattely, an assistant professor of chemical engineering, engineered tobacco, a common laboratory plant, to produce the immediate chemical precursor to the cancer drug etoposide, originally isolated from a leafy Himalayan shrub called the mayapple.
The drug-making pathways Smolke and Sattely reproduced are two of only three that have been transferred intact into another plant or yeast. Researchers have partially reconstructed other pathways, meaning that the plant or yeast yields a chemical that can then be transformed into a drug in the lab. Although the successes so far are limited, Smolke says recent advances in technology will make it easier for more drugs to follow.
“I think in 10 or 20 years we will absolutely see more of our medicines produced through this type of biotechnology,” Smolke says. “This is the way to get around a lot of the challenges and the headaches that we have faced with trying to extract compounds from natural sources that don’t really scale.”
Bond by bond
When taxol was first found in the bark of Pacific yew trees, Wender says, desperate patients would sneak into forests to collect the bark. The process killed trees and destroyed habitat that housed, among other animals, the endangered spotted owl.
Even collected legally, it took the bark of two to 10 mature yew trees, which can take hundreds of years to reach full height, to provide enough taxol to treat a single person. The NCI estimated that keeping up with taxol demand would require harvesting 360,000 mature yew trees per year.
“Its lack of availability was having a huge impact on science and on clinical studies,” says Wender. He chaired the National Institutes of Health Taxol Study Section, tasked with balancing the environmental concerns of harvesting the trees against patients’ needs for the drug.
“I had a winemaker call me to ask if he should plant yew trees or wine grapes; that’s how visible the issue was,” he says.
At the time, there were essentially two paths for producing a drug that had been identified in a plant. One was the route Wender took — synthesizing it bond by bond in the lab. This approach isn’t always possible for large, complex molecules (which many drugs are), is often slow and may produce yields too low to supply patients.
“In the case of morphine, chemists worked out methods to build that molecule through chemistry,” says Smolke. “But the process is very inefficient and, even through decades of research, they’ve never been able to increase the efficiency to make it competitive to poppy.”
If chemistry can’t re-create the drug, another option is a hybrid approach in which scientists look for the drug’s chemical cousin in other plant species, extract it and use laboratory chemistry to finish the job. That approach, called semi-synthesis, is what’s used today to turn a compound found in the needles and twigs of a yew species that is more common than the original source (and can be harvested without damaging the plants) into the taxol taken by patients.
A third option for producing plant-based drugs has recently become more feasible, thanks in part to improvements in genome data, sequencing technology and DNA synthesis. Whereas it was once laborious to sequence and identify genes, many tools now exist to sift through an organism’s genetic code to identify bits that are important for any biological process — like making a particular chemical. These advances in technology are in part what inspired the recent creation of the interdisciplinary institute Stanford ChEM-H, of which Sattely, Smolke and Wender are members. ChEM-H intends to leverage chemistry, engineering and medicine to improve human health.
Plants normally produce complex chemicals through the work of protein assembly lines. They start with a chemical from the environment or from within the plant itself. Then proteins called enzymes sequentially nip a bit here and add a bit there until the final chemical — our drug — emerges.
Each of the enzymes in that assembly line is coded for by a gene in the organism’s DNA. So, the trick is finding which genes code for the enzymes in the assembly line.
Discovering the enzymes’ codes is moving a lot faster now because of new strategies, says Smolke. “One strategy that we think is particularly powerful is looking at the plant’s DNA sequence and using bioinformatics tools to identify regions that code for candidate enzymes.” Her team can then put the DNA sequence for that enzyme in yeast to see if the enzyme carries out the expected role.
But there’s another hurdle after discovering the relevant enzymes, and that’s the matter of getting them to work in a foreign setting. “You are taking enzymes that have evolved to work in a particular organism and then you are asking them to work in a very different organism,” Smolke says. For example, the complex origami of how proteins fold into the correct shape occurs differently in plants versus yeast.
In recent years, more information has become available about how plants and yeast differ, and how to modify genes so that the proteins for which they code perform properly in the foreign environment.
In 2013, advances in technology and knowledge paid off when a team of researchers from the company Amyris, UC-Berkeley and the National Research Council of Canada isolated the six genes from sweet wormwood that code for the anti-malaria drug artemisinin — the drug whose discovery earned Chinese scientist Youyou Tu a Nobel Prize in 2015 — and put them into yeast. Today, about one-third of the world’s supply of artemisinin is produced by yeast in a lab, rather than extracted from the yellow-flowering herb.
The artemisinin example proved that the cellular machinery from a plant could be coaxed to operate in a completely different organism, but it’s a relatively simple six-part system compared with hydrocodone (23 genes) or etoposide (10 genes).
Smolke recently published in Science her success in producing hydrocodone, the precursor to morphine and other painkilling drugs, in yeast. The yield is currently small, but she has formed a company focused on scaling up production.
Given the addictive nature of opioids and hydrocodone’s close chemical relationship to heroin, Smolke’s work raises concerns about providing easier access to illegal as well as legal drugs, a problem that could arise with other legal drugs that have illegal chemical cousins. She says she supports an open, deliberative process that engages scientists, policymakers, regulators and doctors to discuss the concerns about the technology and its benefits, and develop options for governance as it becomes more widespread.
Improving on nature
Obtaining a more reliable and less expensive source for a drug would be reason enough to pursue this line of research, but there’s an added advantage: a chance to refine nature’s handiwork.
“When we extract a drug from a plant, we can only use what the plant gives us,” says Sattely.
Sattely chose to work with tobacco because it is a relatively common laboratory plant that grows well and is easy to genetically manipulate, and by using a plant she avoids many of the challenges of getting enzymes to work properly in yeast. From here, she may take the extra step of moving the genes into yeast, but she says it might be possible to scale up production in tobacco.
Sattely isn’t alone in trying to grow drugs in tobacco, rice, corn or other fast-growing plants. Since the 1990s scientists have realized that plants could be used to grow drugs, called “pharming,” but those efforts involved drugs that are the product of a single gene. It’s much harder to find multiple genes and get them to work in a coordinated fashion to produce a final product.
Sattely points out that many drugs we take aren’t as effective as they could be or have side effects. Once a drug is being produced in a controlled way, it might be possible to introduce a slightly different raw material into the molecular assembly line, which could ultimately produce a better drug.
Or, by mixing and matching genes that make up a variety of different molecular pathways, scientists could create entirely new classes of drugs.
“We don’t have to be limited to what nature gives,” says Smolke. “We can take inspiration from the basic structures, then improve them to either enhance their therapeutic properties or reduce their negative properties.”
For example, could opioids be altered to be as effective without being addictive? Or could an anti-cancer therapy be made less toxic? It’s much easier to explore those questions when scientists can tweak the molecular production line.
Wender has already begun chemically modifying taxol, manipulating the drug isolated from yews in such a way that cancer cells are less likely to expel it as they gain resistance.
He and Nelson Teng, MD, PhD, associate professor of obstetrics and gynecology, have shown that the modified drug is effective in animals and in tumor samples from women whose ovarian cancer has developed resistance to taxol. They plan to test the modified drug in women with ovarian cancer soon. These kinds of improvements on nature could be much simpler when a plant, rather than a chemist in the lab, is doing the work.
Teng helped carry out some of the earliest trials of taxol, producing the extraordinary results that drove desperate patients into the woods to harvest their own yew bark. He calls the work by Smolke and Sattely game changing.
“A lot of time, intuitively, we don’t think we can change life,” Teng says. “But it turns out that it’s not quite true; we can actually change yeast or plants at the genetic level such that their machinery can make molecules of almost any design.”
Teng says a grateful patient once gave him a package of taxol tea, made from the bark of Pacific yew trees. He never drank it, he says, because although it might contain beneficial compounds like taxol, it might also contain any number of other chemicals that are detrimental to human biology. Why risk it?
Advancing chemical engineering techniques could change that balance — biological organisms might be coaxed to produce chemicals with the effects we need without the effects we don’t.