Making a Drug
By Sharon Hesterlee, Ph.D., Sr Director Research and Advocacy
February 8, 2010
This is going to be a short primer on the process required for a good idea to go from the laboratory to the clinic (if only the Schoolhouse Rock people had been focused more on science).
This is usually what I call the “how does that work?” stage, in other words, figuring out how normal muscles operate, understanding what mutations cause disease and how those mutations interfere with normal muscle function. Representing years of investigation, the results are typically layered on top of one another and collectively build-up a detailed picture of biological processes. This kind of science is almost always “hypothesis-driven” using the scientific method that you learned about in school. You have an idea, you think of ways to test it (ideally giving you a clear yes or no answer) and you really want other people to be able to test the idea in different ways and still get the same answer.
At this point, scientists take what they know about a disease and try to intervene in some biological pathway or function to affect the disease process, by either slowing, stopping or reversing the symptoms. These studies may start in a culture dish, just studying the affect of some compound or manipulation on isolated cells, but eventually almost all of these types of studies are conducted in animal models. In the case of Duchenne, the “go-to” animal model is mice that lack dystrophin, although many different kinds of animals are susceptible to the disease including dogs, cats and chickens. The mouse model of Duchenne, called the “mdx mouse” is typically used in several ways: you can “knock-out” a gene that you think might be playing an important role in the disease process to see if mdx mice that lack that gene show improvements. You can supply compounds in the food or by injection, and you can also use biological approaches, such as a modified virus carrying a healthy copy of the dystrophin gene. Larger animal models like dogs are used primarily to test therapeutics that are complicated to deliver such as gene or stem cell therapies, but are not always required to move a therapeutic into the clinical testing pathway. Ideally, one tries to show that you can achieve “proof-of-concept” that a particular therapy is likely to modify the disease through a combination of different types of animal testing.
At the point that it has been thoroughly demonstrated that a particular drug or therapeutic can affect the disease process in an animal model, preclinical development is begun. This stage is much less hypothesis-drive and more goal-oriented. Scientists consider this the “unsexy” part of the process, a long, hard slog that is often more grunt work than brain power. Companies, on the other hand, tend to be very good at process-driven activities like getting a new compound into the clinical testing (although some do very good hypothesis-driven research as well). Before a drug can be tested in man for the first time (literally referred to as “first in man” drugs), the Food and Drug Administration, which is tasked with protecting people from unethical or unnecessarily dangerous research must review the clinical testing protocol.
The investigators who wish to do the clinical trial must submit an IND or “Investigational New Drug Application” that contains the rationale for the treatment, the “proof-of-concept” evidence suggesting it may work, the results of both short (1-3 months) and long term (6-12 months) formal toxicology testing in three different species, documentation on the manufacturing process, the detailed clinical testing plan, copies of informed consent documents and so on. Anything that will ultimately go into people must be manufactured according to a series of regulations called “Good Manufacturing Practices” (GMP) that cover everything from sterilization techniques to shelf life to sample labeling. Everything must be accompanied by Standard Operating Procedures (SOPs). It’s not for the faint of heart and most companies have regulatory advisers onboard or use Contract Research Organizations (CROs) to help them meet these regulatory requirements.
Typically investigators will undertake a series of meetings with the FDA, starting with a pre-IND meeting and followed up with subsequent meetings and communications if there are any surprises or questions about toxicology data, for example. The FDA encourages investigators to interact with it early so that they can help troubleshoot issues, although the FDA is not bound to abide by any of the advice it gives at this stage. After the IND is submitted the FDA has 30 days to either not respond to the application or to issue a clinical hold. Holds may be temporary, until simple questions are answered or much longer if the FDA feels there are fundamental problems with the protocol.
In addition to FDA approval, the investigators must also have approval from the Institutional Review Board at the clinical trial site. If the trial is designed to be a multi-center trial, the IRB at each institute must review and approve the protocol and informed consent process. The IRBs have the task of protecting human subjects. If any changes are made in the protocol or consent form at any time, all of the IRBs must approve the change. It can be a very time-consuming task since some of the IRBs are so overburdened that it can take months to review a protocol. The IRB approvals are usually only good for one year and then the IRB application must be renewed if the trial is not complete.
Finally, if therapy to be tested involves manipulation of genes or genetic material, the NIH Recombinant DNA Advisory Committee or “RAC” may ask to review the protocol as well. Gene therapy trials, for example, are often reviewed by the RAC, but may not be if the protocol in question isn’t very different from other trials under way. The RAC doesn’t have any legal regulatory authority and it’s not affiliated with the FDA. The RAC was o developed in the 1970’s because a lot of people then felt that anything that involved DNA manipulation required a higher level of scrutiny. For example, in gene therapy studies the RAC may be concerned about modified genes being inadvertently inserted into the germ lines (eggs or sperm) and genetically modifying future generations. The RAC meets quarterly and reviews projects publically, unlike the FDA review process, which is not public. Read more about the RAC and its objectives here: http://oba.od.nih.gov/rdna_rac/rac_about.html.
Review by the FDA, IRB(s) and RAC can, and typically do, take place in parallel when the applicants have planned the submission process well.
Phase I: The “normal human” study
Most clinical trials of first-in-man drugs start with what’s often called a “normal human” study, meaning that healthy volunteers (often paid) are recruited to test the drug for safety and how the drug is broken down by the body (PK or pharmacokinetics). These are usually small, short-term studies. For certain types of studies, however, it’s considered unethical to test a drug or procedure in health people, possibly because the therapy is invasive or because the therapy could potentially make a healthy person sick just by doing what it’s designed to do. For example, exon-skipping would never be tested on healthy volunteers because a dystrophin gene with a few exons skipped is likely to be less functional than the dystrophin gene with which a healthy person is born. If a healthy volunteer study can’t be conducted because of ethical reasons, investigators may start with a phase I/II study that uses affected individuals and tests primarily for safety but also looks for some efficacy (whether or not the therapy is working). Another alternative is to do a short safety/dosing trial in affected individuals without looking for efficacy. The standard normal human study may take about six months from start to finish, give or take a few months. Even if the study subjects only take the drug for a month, it typically takes time to recruit people into the study (people don’t start all at once, but rather are recruited on a “rolling basis” as they are identified) and it also takes time to analyze the resulting data. Also, sometimes the FDA will require a “waiting period” between each group of participants who receive a different dose of the therapeutic, just in case problems develop.
Data generated by all levels of clinical studies is usually reviewed by an external “Data Safety Monitoring Board” (DSMB), which is made up of experts from relevant fields like toxicology. The DSMB, which is not affiliated with the company or investigators doing the study, typically has access to “unblinded” information-in other words, they know who is taking drug rather than placebo. Based on what it’s seeing in the data, the DSMB can recommend protocol changes, can stop the study for significant concerns, or even recommend that the study be stopped if there is an overwhelmingly positive response to a particular treatment and it becomes unethical to withhold the treatment from those on placebo.
All “adverse events” or “AEs” (anything that goes physically wrong with a person participating in a study) must be reported to the FDA even if it’s not likely that the AE was related to the treatment. The FDA distinguishes between AEs and serious AEs (SAEs)—an AE might be something like a headache or nausea without vomiting, while an SAE could range from severe vomiting to death. When any type of AE occurs, but especially an SAE, the event will be evaluated carefully to determine if it was likely to be related to the study drug. The clinical trial may be placed on hold temporarily during the course of this review. A clinical trial may be stopped if there is significant concern that the SAE or an unacceptable level of AEs are related to the study drug . This often sends drugs back to the drawing board where investigators will try to determine what caused the bad reaction and whether or not the therapeutic can be modified to reduce the likelihood of it happening again. A modified drug would start over at the proof-of-concept stage. For more on AEs see http://www.mda.org/research/trac/meeting_slides/endpoints/LMiller2.pdf).
Phase II: Safety and Efficacy
In a phase II study, investigators continue to explore the safety aspects of a potential treatment, but also look at efficacy (whether or not the drug works) and try to optimize the dose and the dosing schedule. They may also look at a variety of different “endpoints”—the things that are actually measured to determine if the treatment is working. Some endpoints for muscle disease include, but are not limited to, the six minute timed walk, manual muscle testing and creatine kinase levels. One of the endpoints measured will be identified as the “primary outcome measure” and it will be the one used to determine if the study has a positive outcome or not. Typically, in a phase II study, this is the endpoint that investigators feel is most important for determining efficacy. In a phase I study the primary outcome tends to be safety-related.
It’s important to know that more than one phase II study usually takes place before moving on to phase III—often what you learn in the first phase II study suggests protocol changes to improve the action of the drug or limit side effects and then the changes are tested in a second phase II study. This is particularly likely to be true if, for ethical reasons, there was no phase I normal human study. These studies may be called Phase IIa, Phase IIb or even Phase IIc etc. The final phase II is often referred to as the “Phase III enabling study.”
Phase III: Efficacy
The phase III study is usually much larger in size than a phase II and is designed to definitively establish safety and efficacy in preparation for seeking approval to market the drug for a give purpose. If there are any existing treatments, for example prednisone in the case of DMD, the phase III aims to demonstrate that the new therapy works better than prednisone. Some phase III studies involve thousands of participants, although this isn’t likely to ever be a requirement for a new DMD drug. You should be aware that the FDA frequently requires two positive phase III studies before a drug can be approved in order that substantial evidence be obtained that it works. Again, this requirement is less likely to be enforced for a rare disease like DMD, but two phase III studies could be required.
The FDA does have a mechanism for “Accelerated Approval” for serious or life-threatening diseases that can reduce the number of studies required or allow an endpoint (measurement of how well the drug works) that is an indirect measurement of the effect, but there will be a requirement for post-approval studies to continue to try to gain more evidence that the therapeutic is actually working.
If a phase III study meets its endpoints and a second phase III study is not required, the study sponsor will likely submit a New Drug Application (NDA) to gain approval from the FDA to market the drug. All of the data that has been gathered throughout the testing process under the Investigational New Drug application becomes part of the NDA . Regulators will decide if the drug is safe and effective, if the labeling is appropriate and if the manufacturing process meets appropriate standards. In a standard review the FDA has about nine months to make a ruling, but some therapeutics can receive an accelerated review that is less than 9 months.
Phase IV: Post-marketing surveillance
Often even after a drug is approved and is being prescribed, the FDA requires additional studies monitoring safety and efficacy. It’s through this type of post-marketing surveillance that less common, but severe side effects can be identified, sometimes leading to a drug’s withdrawal (Fen-Phen and Vioxx, for example).
Companies will sometimes make available on a voluntary basis experimental drugs for severe or life threatening conditions to qualified individuals under an “Expanded Access” program (EAP). The conditions are that there must be some evidence that the therapy might be beneficial and some information on safety issues available. The participants in the program must not be eligible to participate in any ongoing clinical trial for the therapy. EAPs are often controversial—they can be very expensive for companies to run because of the cost of the therapeutic and the requirement that people are monitored as stringently for safety as those participating in trials. The FDA must approve an EAP, but it can’t require one. Any adverse events that occur in the EAP may negatively impact ongoing trials or even cause them to be halted so the EAP program represents an additional risk that the drug may fail.
1. What is the role of the FDA Office of Orphan Products? http://www.mda.org/research/trac/meeting_slides/endpoints/SLinde-Feucht.pdf
2. General on clinical trials: http://en.wikipedia.org/wiki/Clinical_trial
3. General on clinical trials: http://www.clinicaltrials.gov/ct2/info
4. Expanded Access Programs: http://www.nlm.nih.gov/services/ctexpaccess.html
5. Expanded Access Programs: http://www.fda.gov/ForConsumers/ByAudience/ForPatientAdvocates/AccesstoInvestigationalDrugs/ucm176098.htm