Using recombinant DNA technology to produce drugs in simple organisms like bacteria and yeast is nothing new. Animal-derived insulin was phased out in the early 1980s when Eli Lilly offered the first insulin drug created using recombinant techniques. Using an expression vector for the protein and injecting it into mammalian cells allowed for cheaper, more reproducible production of drugs. Other examples include erythropoietin, interferon, and the anti-Her2 monoclonal antibody.
Synthetic biology, sometimes considered the little brother of recombinant DNA technology, has come a long way — and fast. It's been only a decade or so since researchers at MIT introduced the field, but considering the inroads it's already made into basic research, applying it to the clinical realm isn't too far off. But as with any rapidly growing field, there's always the need for a reality check to separate out the hype. Harvard's Pam Silver, whose own lab covers a range of interests from creating chimeric proteins to engineering new pathways using synthetic biology constructs, says, "Synthetic biology makes a lot of promises. The question is, what can it deliver on in terms of health?" Silver, among others, thinks that within the next five years, advancements in metabolic engineering of cells to create bulk chemicals, protein engineering, and creating smart cells are sure to light the way for this nascent field into full-blown use in both drug discovery and production.
Microbes doing the work
One of the first to tap microbes for the scaled-up production of a common therapeutic is Amyris Biotechnologies, whose work focuses on using E. coli to chug out massive quantities of the antimalarial drug artemisinin, which up until now has been derived from plants. Co-founder Jack Newman thought of the idea during a postdoc in Jay Keasling's lab at the University of California, Berkeley. "We had all come to an engineering lab as a postdoc because we wanted to take basic research and do something with it that would be impactful," Newman says. His project took the metabolic genes that encoded the production of artemisinin in the plant Artemisia annua and put them into a microbe that could produce them in much higher quantities. The end goal was an economic one — to "provide lots of this drug at very low cost, in a scalable fashion," he says.
With funding from the Gates Foundation, the company formed as a partnership between Keasling's Berkeley lab and One World Health. The five-year project is now entering its final year, and about six months ago, the members decided to partner with Sanofi-Aventis to take production to "the multi-ton scale," Newman says. "Sanofi-Aventis has manufacturing capabilities, and they also have a regulatory and a pharmaceutical distribution pipeline, which made them a very attractive partner."
But researchers will still have to exercise caution as they ramp up programs like these. Oliver Rackham, group leader of the synthetic biology and drug discovery research program at the Western Australian Institute for Medical Research, notes that underestimating the complexity of living cells could be troublesome. "Standardization can be challenging as minor differences between cell types or growth conditions can result in quite profound changes in cellular metabolism, potentially resulting in variability in production efficiency or contamination of the product with unwanted secondary metabolites," says Rackham, whose own work focuses on creating bacterial and yeast cells that can make diverse unnatural polymers.
Pam Silver at Harvard is a big fan of using organisms to do things that chemists can't do as easily. But, she says, "There's all kinds of problems associated with scale-up. How do you separate the molecules from the cells, [and] is that more expensive than simply doing regular chemistry?"
Building better proteins
The area of protein engineering — making use of modified metabolic pathways to create new compounds or new drug analogs, or creating altered molecules like fusion proteins — is also in pharma's crosshairs. As Silver puts it, "I think protein-based drugs take up a very small fraction of the total market, but — at least this was true a year or two ago — it was the fastest growing segment of the drug market."
Newman agrees that creating analogs of current molecules will add weight to synthetic biology's toehold in drug discovery. "If you have something you know is an active part of the molecule, and you can make that using synthetic biology, you open up a lot of activity space around that pharmacophore."
Biotica Technology, a Cambridge, UK-based biotech company, uses genetic engineering to tweak the structure of natural polyketide synthase enzymes, which allows for swapping in and out of diverse modules, or substrates, along the synthesis path. Many important drugs are polyketides, including the antibiotic erythromycin. Analogs can offer increased efficacy, reduced toxicity, or simply be a legitimate way for drug companies to get around patents on drugs with similar, but not identical, chemical structures.
By creating hybrid, genetically altered PKS enzymes, Biotica's chemists can generate novel structural analogs of the resulting polyketide product. They can also do things like glycosylate the end product, making for an actively different compound. In the case of erythromycin, for instance, modifying the bacterial production strain by replacing the genes for the usual propionyl starter unit with genes that will encode for one that selects more complex acids, and then feeding the modified strain with novel acids, results in a set of more diverse analogs. Some had activity against gram-negative bacteria, which normal erythromycin does not.
"We have made a collection of the genes encoding the enzymes that are pre-PKS, PKS, or post-PKS enzymes," says Ming-Qiang Zhang, senior vice president of R&D at Biotica. "When we want to do a structure change, we can pull from this toolbox. Instead of being one genetic change, we can do several gene changes in order to get a product of interest."
Biotica is also tweaking the chemistry of erythromycin to selectively take advantage of its additional anti-inflammatory abilities. "Once erythromycin is at the site of the inflammation, it can act as an anti-inflammatory agent," Zhang says. "So what we try to do is to maintain that anti-inflammatory activity, but in the meantime, get rid of the antibiotic activity so that the new drug can be used as an anti-inflammatory agent without the fear of causing unnecessary bacterial resistance." Biotica has recently partnered with GlaxoSmithKline to advance the project. While Biotica will focus on drug discovery and lead optimization, GSK will lend its expertise in pharmacology, clinical development, marketing, regulations, and scale-up.
While making analogs of drugs in prokaryotic cells has seen much success, metabolically engineering pathways in plants is cutting-edge research. MIT's Sarah O'Connor has used enzyme swapping of a re-engineered alkaloid biosynthetic gene to modify the monoterpene indole alkaloid pathway in the Madagascar periwinkle plant, Catharanthus roseus. "We took a very complex metabolic pathway that's very poorly understood and we were able to rationally re-engineer one step of the pathway to be able to take a substrate that the natural enzyme would not recognize, and we could show that that enzyme functions in vivo," she says. "Our goal is to try to explore the production of these compounds in plants and try to see if we can make new analogs, or perhaps make more of them, or perhaps reconstitute them in heterologous organisms." Ongoing work in her lab, as well as through a partnership with Vanderbilt University to screen these compounds, is hoping to explore whether some of these new "unnatural" natural products could serve as therapeutics.
"For plant pathways, it's definitely in the early stages," O'Connor says. One of the biggest challenges to reconstituting the pathways in simpler organisms like E. coli and yeast is getting the genes to encode all the steps. "It's tough to do," she says. "It's time consuming and very painstaking work" that led her to metabolically re-engineer the plant itself. While putting these genetic networks into basic organisms is simpler, more robust, and offers an easier growing environment, the advantage to working with plants is that "the metabolic pathway has already evolved in this particular organism, so in terms of tweaking the pathway or reprogramming [it], we can just focus in on one step rather than be faced with the challenge of reconstituting, say, all 20 steps," O'Connor adds.
At Harvard, Silver is also doing a form of protein engineering by using synthetic biology to create chimeric, or fusion, proteins. "The general idea is to take advantage of the high modularity of proteins and to design proteins that will specifically target to your cell of choice," she says. In recent work, she genetically mutated interferon to increase its specificity. As a drug, interferon is routinely administered to treat cancer and multiple sclerosis, but can have unwanted side effects, such as flu-like symptoms. Silver designed chimeric proteins in which the activation of the IFN-α receptor in HeLa, A431, and engineered Daudi cells depends on the presence of EGF receptor on the same cell. The mutant chimeric proteins inhibited proliferation of cells sensitive to IFN- α dependent on the EGF receptor. "What we've done that's novel is to reduce the affinity of the killer molecules for their receptor," she says. "By mutating interferon, it can't bind to cells except for the cells it's specifically targeted to," thereby reducing unwanted side effects.
Engineered to kill
Another major focus of synthetic biology drug development is the goal of shutting down tumor cells. UC Berkeley's Chris Anderson leads the Tumor Killing Bacterium Project, which is part of synBERC and aims to genetically reprogram E. coli cells to recognize tumor cells and then invade and kill them. Anderson uses a nonpathogenic strain of E. coli and adds genes to create a "smart cell," one that can recognize tumor cells and react to them. Input information comes in the form of cues like variable concentrations of oxygen, glucose, or cholesterol, or the presence of specific sugars on the surface of cancer cells, for instance.
At the moment, Anderson is working on the last step of the invasion process, called the payload delivery device. Once a bacterium is inside a cellular vacuole, it has to be programmed to do something detrimental to the cell, like roam freely in the cytoplasm or lyse both itself and the vacuole, Anderson says. "So we're trying to build both of those right now, and each of those have a lot of bits and pieces."
DNA synthesis is the biggest cost- and time-limiting factor to his work, and since last April, Anderson has been trying to figure out a way to automate the DNA construction and systhesis process. His students won last year's International Genetically Engineered Machine competition with their Clotho software, which is an open source platform that he'll be using to make his experiments faster and more efficient. A Java-based client, the software allows users to computationally create and then physically construct the bits of DNA that will be inserted into the bacterium. "The idea is that it's a design tool, a management tool, and also the assembly tool, so the thing that would actually generate the code that runs the robots that do the molecular biology operation to make your DNA," Anderson says. "It just makes it more scalable. We can do more data points, and we also have a lower failure rate."
Even with automation and large-scale experimental data sets — typically the team does two or three of these 5 kb to 10 kb DNA devices per experiment, but actually should be doing as many as 100, says Anderson — the problem of tuning gene expression remains. "A lot of synthetic biology has to do with just getting expression levels at the right level," he says. "We have no real good principles yet as to how to design a DNA that's going to have a particular level of activity."
In the pipeline
Most experts have high hopes for synthetic biology moving into, or at least very close to, the clinic in the next five years. Zhang at Biotica hopes to see some of the company's drug candidates in phase 2 clinical trials by then. And Silver sees a relatively short time frame as well, noting that one advantage to the work she's doing with proteins is that "some of proteins we're using are already FDA-approved, like interferon. If you use things that are already approved, that sometimes cuts the barriers to getting your molecule approved."
Zhang and others hope to see more drugs going the way of arteminisin — that is, being produced cheaply and reliably in microbes. "In the next five years, you may find a lot of drugs being produced like that," he says. "I would like to see, for example, Taxol being produced in that way."
Overall, Jack Newman at Amyris sees synthetic biology as having a big impact on large-scale drug production. "I think pharma is definitely starting to embrace the idea of using synthetic biology for new drug production and even drug discovery," he says. Drug production is a great way to take advantage of drugs that were originally extracted from plants — the anti-cancer drug Taxol is a good example of one that could be made more efficiently using microbes and large-scale fermentation processes, Newman adds.
As the developing world gains traction, there will be more incentive for pharma companies to cut production costs. "Pharma is very much focused on discovering new drugs that do new things, and the production process doesn't need to be very efficient because for the wealthy part of the world, cost isn't the biggest issue when it comes to healthcare," Newman says. "As cost does become more of an issue, especially with generic drugs, this will become more popular."