How Diabetes Drugs Get From Idea to You
Robots whirl about, day and night, in university and industry laboratories across the globe in a systematic search for new medicines that could benefit humanity. These tireless machines are a link in the chain that takes medications from a twinkle in some scientist's eye to the rescue of sick people.
The drug discovery process is not an easy one, nor is it cheap. It can take 10 to 15 years to develop a single drug at an average cost of between $800 million and $1 billion, according to the Pharmaceutical Research and Manufacturers of America (PhRMA). And for every drug that makes it to the consumer, there are hundreds or even thousands that don't. That's a lot of failure. But the rewards of success are great for both the maker, which benefits financially, and the user, who gets healthier. While the story of each drug's origins is unique, some common threads are woven through the discovery process.
|Yellow robots aid the drug discovery process at a Merck lab in Pennsylvania.|
A Targeted Approach
The hardest part of making a new drug, according to David Moller, MD, vice president of endocrine and cardiovascular research and clinical investigation at Eli Lilly & Co., is "picking the targets." A target, in pharmaceutical parlance, is a part of the body—usually a gene or protein—that a medication is meant to influence. In type 2 diabetes, for example, targets would typically be on the biological pathways that determine blood glucose levels. They might be genes and proteins involved in either insulin production or insulin resistance. (The hormone insulin is the "key" that opens cells to glucose to provide the body with energy.)
Where do the ideas for targets come from? Often, they emerge not from the private companies that eventually make and market a drug but from the nonprofit and public sectors: university and government laboratories. "Drug companies have concluded that universities have a large contribution to make, and they are increasingly trying to partner with basic scientists and universities," says Daniel Drucker, MD, FRCPC, a professor in the Division of Endocrinology and Metabolism at the University of Toronto. Much of this basic research is written up in scientific journals that are available to everyone.
Developing a new drug is expensive, so researchers must be highly selective about what targets to focus on. In a process called "target validation," drug companies winnow down the list of possible targets to just a few. "It's all toward the development of a project we can sink our teeth into," says Moller. Target validation usually involves some basic experiments, typically animal studies. One method of validation is to see what happens when a target is entirely removed from an animal. If the effect is dramatic, the target may be worth exploring. If something doesn't work in an animal, it probably won't work in a human.
Then again, just because obliterating a target cures a disease in a test animal, that result is by no means a guarantee that it will work in people. "The whole problem with productivity in the industry is that things that work in animal models don't work in humans," says Moller. "You may get a little efficacy, but nothing like what we see in mice or rats." To speed up drug discovery and control costs, the industry is trying to get better at predicting what drugs will work in humans.
Crank It Up
Once a target has been selected, it's time to bring in the robots and their human handlers. A staff of 20 to 30 scientists is common, Moller says, for a small-molecule discovery project that may take years.
Most drugs are small molecules, just a cluster of a few atoms strung together. Small molecules can have outsized effects on the human body, and they are excellent drug candidates, in part because they're easy to test and manufacture. In diabetes, most medications are small molecules and sold as pills. However, there are some key exceptions. Proteins such as insulin and exenatide (Byetta) are larger molecules and must be injected. "Protein engineering is an art in itself," says Moller.
Each company seeking small-molecule drugs has a library of chemicals, often millions of compounds. The hope is that this library will yield a new medication or at least a starting point for one. These chemicals come from various sources, including commercial catalogs. Some are proprietary molecules developed in-house.
In recent years, robots have become a standard part of drug discovery. Their job is simple: move around countless plastic trays, each the size of a smartphone and dimpled by hundreds, if not thousands, of tiny wells. The robots and trays allow researchers to quickly test all the compounds in their libraries. Each well in a tray is like a miniature test tube in which chemicals from the library are mixed, one at a time, with the target. A series of lab tests show whether a particular chemical interacts with the target. If it does, perhaps blocking or improving its function, that's a "hit." And that could be the basis for a new medicine.
The next step is called optimization: trying to get the hit chemical to form a stronger and more selective fit with the target. A chemical that sticks tightly to a target can be effective at doses low enough to be both safe for humans and easy to turn into a medication. The chemical must also bind to the target specifically but not to anything else in the body; that is a key to preventing unwanted side effects. Tinkering with hit chemicals used to be pursued blindly. Thanks to recent technological advances, though, researchers can now see what a target looks like. Using computer simulations, they can dock a drug candidate with the target and make educated guesses about how to modify its chemistry to make it more effective.
Once researchers think they may have a winner, they need to figure out how to manufacture the drug. The recipe for making the drug, including ingredients that hold the medicine together and release it at the right time, is established before testing begins and then held constant throughout clinical trials. The drug must be studied in what is essentially its final form.
With the finalized pill or vial in hand, it's time to think about getting the drug into humans. But first, studies in animals are required to test its safety. "That's the way it should be," says Drucker, the Toronto professor. "We wouldn't want to expose people to harm." At this point, the Food and Drug Administration (FDA) gets involved. If safety data from animal tests pass muster, the FDA grants "investigational new drug" status and allows testing in humans.
Universities and companies both participate in the earlier stages of drug development. But testing drugs in people, Drucker says, "is much more the domain of the pharmaceutical industry," because drug companies have the money and expertise to conduct clinical trials.
The FDA requires three levels of clinical trials, often called Phases I, II, and III. Typically, Phase I is a safety study that determines optimal doses and possible side effects; it is short in duration and includes just 50 to 100 healthy people. In Phase II, the drug is tested on between 100 and 500 people with a disease for a longer period of time to see if it actually works. Few drugs make it this far. Of every 250 drugs that reach animal studies, only 5 to 10 get to human trials, according to PhRMA. But if all goes well in Phases I and II, the company may move into larger, more expensive Phase III studies. They typically include from 1,500 to 4,000 people and take one to three years. In recent years, Phase III trials have become even more rigorous for diabetes medications (see "The FDA and Diabetes," left).
With clinical results in hand, the company submits a "new drug application" to the FDA. A drug application can total 100,000 pages or more, including all the data from trials, along with information about manufacturing and how the drug works. From this, the FDA determines whether the medication is safe and effective, the labeling is appropriate, and the manufacturing methods are adequate to ensure the drug's purity, identity, quality, and strength.
With guidance from an independent advisory committee, the FDA can either approve or reject a drug outright, or request more data. This typically takes between six months and two years. If the drug is approved, the trials don't stop there. There are Phase IV trials, done after the medication is on the market. These can address long-term side effects and specific safety concerns. They can also be used to see if the drug may have other uses, such as a diabetes medication that turns out to be helpful with weight loss.
The long and winding road of drug discovery and development, littered with false starts and outright failures, may appear unnecessarily bumpy. That is, unless you're the person who gets the benefits of a new lifesaving medication without needing to worry that the drug is harmful.
The FDA and Diabetes
In 2010, the Food and Drug Administration (FDA) changed its rules on approving diabetes drugs. The impetus for this ruling came after concerns that diabetes medications on the market, such as rosiglitazone (Avandia), raise the risk for heart disease, which is already high in people with diabetes. The FDA now requires that diabetes drugs undergo cardiovascular safety studies prior to approval. The agency recently requested that Novo Nordisk conduct a cardiovascular outcomes trial on its ultra-long-acting insulin degludec (Tresiba) before it would consider approval. The trial may take years to complete.
The additional safety testing has changed the way that pharmaceutical companies pursue new diabetes medications. "This is a mixed bag for patients and physicians working in the diabetes area," says Daniel Drucker, MD, an endocrinologist at the University of Toronto. "The old way of testing, as little as four to five years ago, may have involved 1,000 to 2,000 patients. … Now, we'd need 5,000 to 6,000 patients to have enough cardiovascular safety information." That translates into higher costs and longer studies.
The result is that pharmaceutical companies say they have the resources to work on only a few diabetes drugs at a time. "The size, length, and cost of studies have essentially doubled," says David Moller, MD, of Eli Lilly & Co. "We have to be more selective about what molecules we want to invest in as far as late-stage work." Even though the FDA rule means that people with diabetes may need to wait longer for new drugs, the medications should be safer when they do come along.