A new grammar for drug discovery

When Mark Fishman took the helm at the Novartis Institutes for BioMedical Research (NIBR) in May 2002, his focus was on changing the way Novartis discovers new medicines. Striving to push the frontiers of science at the intersection of chemistry, biology and medicine, the 56-year-old has combined traditional approaches to pharmaceutical discovery with emerging tools and concepts from the chemical and biological sciences. Pathway biology and genetics are at the heart of the NIBR’s early discovery efforts.

Close to modern biology and with an expertise in genetics, Fishman brought to the job just what Novartis was looking for. Since he took over as President of NIBR, the cardiologist has been using the latest advances in genetics to discover how diseases work and how best to treat them. He also refocused Novartis drug discovery efforts on diseases where there was clear unmet medical need and scientific tractability rather than calculating the size of the potential disease market. The strategy behind his new grammar for drug discovery, Fishman says, is quite simple: “Our goal is to get into patients and treat patients that have real unmet needs. Therefore, we will do best if we go for patient populations where we understand the mechanistic basis of the disease. Sometimes this can be a rare disease, or sometimes it can be a subset of a disease that has a more broad-based population. That’s the essence of how we get into the clinic.”

The new grammar speaks to how NIBR has organized its drug discovery efforts- which is to look at the shared molecular pathways that underlie fundamental disease processes. “The genome has 22,000 or so genes,” he explains. “That’s not really a tractable approach to discovering drugs – to go one-by-one in hosts of different diseases. The real way biology works is by using modules of interacting genes and proteins. From fruit fly to man, the same genes and proteins interact in the same way. These modules are what’s conserved as much or more than even the individual proteins. So it’s a collection of proteins or a collection of genes.”

Molecular pathways

Scientists at NIBR look at these pathways and decide how they can be used for drug discovery, as Fishman explains. “We ask ourselves whether these pathways are revealing of a particular key node – a key protein that, if hit by a drug, would either enhance or reduce the pathway. The pathways may have hundreds of proteins in them, but which one or two are key? And can we hit them with a drug? And when we have drugs, we’ll say ‘Okay, but then what diseases are they for?’ Sometimes we will have these drugs before the world at large even defines how diseases use these pathways.”

Fishman hopes to be able to divide diseases by the pathways that go wrong instead of, or in addition to the organ systems they affect. “We know that it is quite likely that some diseases that involve the vessel wall, transplantation, cancer and some rare diseases all affect the same single pathway of cell growth. So one drug in principle could be effective in treating all of those different diseases.”

In 2004, Novartis entered into a pioneering public-private collaboration known as the Diabetes Genetics Initiative (DGI), which has been aimed at deciphering the genetic causes of type 2 diabetes. The collaboration brought together diverse expertise in diabetes and metabolic disease, human genetics, genomics, statistical analysis and drug development. The researchers include groups from Novartis, the Broad Institute of Harvard and MIT, and Lund University in Sweden.

So far, the DGI has identified three new gene regions linked to type 2 diabetes. What is more, rather than keep this information to itself, as many drug firms have done, Novartis is making it available for free on the web. “First,” Fishman explains, “we thought it was the right thing to do. Second, the information still needs a lot more digestion. We felt that if we could get the input of the world at large, this could be fruitfully done and would benefit not only us and the patients that we’re hoping to treat but a much broader population. And third, I was hoping that others would do the same, and we’ll see.”

Target validation

For Fishman, the collaboration proved constructive. “At the end of the day, only drug companies are going to discover drugs. And if drug companies understand well the fundamental discoveries, they will do that more quickly. The other parties each bring to the table their own expertise, excitement and initiatives. It’s tremendously beneficial to have one party who is pushing hard to discover and publish, another one who is taking care of patients, and a third who is interested in turning this into drugs.”

How can the discoveries be turned into new drugs? “The process is what we call target validation. In other words, you find a gene that you think is related to the disease. And then you look at how it works in other populations of patients, and see if it plays a role in other contexts. And then, you set up models in animals to see if you can use this to mimic the disease,” says Fishman.

The first marketed drugs to result from uncovering genetic information are already out there. “Gleevec, for example, was based on an understanding of how the oncogene ABL was tied to the gene BCR. In fact, the statins were first tested in populations with genetic hypercholesterolemia. There are diagnostic tests, for example, for Herceptin that are used now to categorize patients with breast cancer. We use genetics essentially every day in the drug discovery process.”

Pharmacogenomics aims at optimizing drug therapy with respect to the patients’ genotype to ensure maximum efficacy with minimal side effects. It holds the promise of personalized medicine: drugs that are tailor-made for individuals and adapted to each person’s own genetic makeup. Fishman, however, is not comfortable with the terminology. “As a physician, we always practice personalized medicine. It might be based on a physical exam, or on looking for a swollen knee, or it might be an x-ray, or a urine test. Now we’re just adding another level of refinement through genetic testing. That doesn’t change the essential premise of how we deliver healthcare.”

Biological therapeutics

NIBR has gone through rapid growth. Fishman now calls it a phase of refinement rather than revolution, and a continuous process of improvement. The productivity of his approach can be measured by the ever-expanding pipeline of new molecular entities (NME) that are getting ready to and are going into the clinic. Novartis has 138 projects in clinical development, amongst which are 50 NMEs, and 104 projects in phase 2 and beyond. Last year, 20 new projects were added to the pipeline. “What’s particularly positive,” Fishman emphasizes, “is the fact that we have been able to bring in additional biological therapeutics quite quickly. Today, 25 percent of our pipeline includes antibodies and other biologicals, as opposed to the past when it was all low molecular weight entities. That’s been an extremely positive approach.”

Fishman believes that it has been an absorbing time working at NIBR. His team has been able to combine the low molecular weight approaches with the biologicals approach. “That’s been an extremely exciting kind of combination, and we use the talents that we have for both. Some of the antibodies that we have coming along are particularly well targeted. Some of the low molecular weight drugs that we have will be very important in treating, for example, particular types of cancer and autoimmune diseases. We have others that are coming along in metabolism and cardiovascular disease that are quite interesting. There’s a lot that’s coming out. And although not everything works, often it comes out better in the clinic than we had even anticipated.”

Fishman won’t pick a favorite, but one thing he finds particularly exciting is the antibody for interleukin-1, shown to work well in Muckle-Wells syndrome. “That one will expand its indications to include others, including pediatric illnesses of a certain inflammatory nature, and some subsets of those with rheumatoid arthritis.”

Another project on Fishman’s agenda is the integrated biomedical research and development center in Shanghai, with an initial focus on the infectious causes of cancer. China’s pharmaceutical market is positioned to double to $25 billion by 2010, with an opportunity for Novartis to increase its market share considerably. But Fishman isn’t particularly interested in market share. His goal in research is to find medicines for which there’s a need. One of the reasons he favored Shanghai was talent. “There is a huge talent pool of scientists who have graduated in China. There are many who have trained in Europe or the United States and want to go back. That’s a cadre of scientists that we felt could contribute greatly to the process of drug discovery,” says Fishman.

Another reason was the opportunity to understand diseases that affect a large part of the population, in this case the one in China. “They have a particular propensity to certain disorders, such as cancers that are caused or related to infectious agents. This for us was an opportunity to learn about the fundamental issues underlying these diseases and to help those patients. And at the end of the day, we anticipate that whatever drugs will come out of that might be applicable to a subset of patients elsewhere in the world.”

Diabetes duel

One drug that won’t be coming out just yet is Galvus. The FDA's request for an additional safety study delayed the market entry of Novartis’ potential diabetes blockbuster. Galvus, Fishman insists, is a good drug. “It’s simply a matter of answering some of the FDA’s questions, especially about how the drug might accumulate in patients with renal failure, which is a reasonable question to ask. Then we fully anticipate that it will be out there.”

In what was dubbed the Diabetes Duel, Galvus lost the first round to Merck’s Januvia. Both pills were expected to reach the US market in 2007 and to bring in combined sales of $5 billion by 2010. With the FDA’s request for more data on the DPP-4 inhibitor, Novartis is extremely unlikely to see approval this year. Fishman laughs at the idea of losing the battle. “There are several phases to how drugs come to the market. The first is, you’d like to get there first if you can. I give Merck credit for having gotten there quickly. The second is that you make a medicine that is used by a lot of people, presumably is differentiated and distinct, and builds it’s own franchise over time. So there are different opportunities if you’re there first or if you are there second.”

Confident, Fishman hopes that in the near future his division will achieve two major goals: “One is that we have a large number of new medicines that will help a lot of patients with currently unmet medical needs. At NIBR, I have separate groups that serve as platforms. Some of these are looking at devising the new grammar, at doing the fundamental research that we believe will change how drugs are discovered in the next few years. And so secondly, I hope that soon we will have really begun to change the whole grammar of drug discovery.”

Mark Fishman: “Mark Fishman came to Novartis in 2002 from Harvard and the Massachusetts General Hospital, where he was Chief of Cardiology, Professor of Medicine, and Head of a research center. His research interest over the years has been in the biology of development and genetics.. His combination of interests in a variety of clinical and in scienctific field helped to generate the core strategy followed at the Novartis Institutes for BioMedical Research (NIBR), which is to focus on patients and to attack tractable medical problems.”