Hyderabad: Vasant Narasimhan, MD, global head of drug development and chief medical officer of Novartis International AG, is as much at ease with technology as he is with medicine. Narasimhan joined the Switzerland-based pharmaceutical company in 2005, prior to which he worked at McKinsey & Co. Narasimhan received his medical degree from Harvard Medical School and a master’s degree in public policy from Harvard’s John F. Kennedy School of Government. In an interview during a recent visit to the Hyderabad campus of Novartis, he spoke about a range of topics including clinical trials, gene-editing tools, the role of sensors and machine learning, and deep learning algorithms in medicine. Edited excerpts:
What kind of research does Novartis conduct out of its India labs? How different is it from the work done at your research centres in the US and Switzerland?
In both the US and Switzerland, we do a lot of basic science research—a lot of exploration into new technologies, new therapeutics and new biology. In order to develop a medicine, we have to generate a lot of data and get insights from that data. Hyderabad is one of our core centres to capture that data, analyse it and generate insights from that data. This allows us to take a medicine and find out if it works, then register it and administer it to people, and give it to physicians.
The site in Hyderabad also supports clinical research across all of our medicines. Consider the case of the drug Entresto (used to treat people with chronic heart failure)—a very important one for Novartis. All the studies and analysis for this drug came from Hyderabad. The regulatory files for this important medicine and registration packages were created in Hyderabad. Even some of the work on how the pill (Entresto) itself is made was done in the Genome Valley in Hyderabad (If Bengaluru is known as the Silicon Valley of India, Hyderabad is referred to as the Genome Valley for its research and development work in biotechnology).
One of our biggest efforts in clinical research happens here in India. Over 50 clinical trials are running in India today, and we are working on medicines—from oncology to heart disease, lungs—currently being evaluated involving hundreds of investigators or physicians who work in Indian hospitals and thousands of patients in the country today.
Give us some examples of the science-based innovations that you are referring to.
We are one of the world’s largest investors in basic science research. One example involves using CRISPR CAS-9, which is a gene-editing technology. CRISPR CAS-9 was discovered by studying bacteria. We learnt that these are enzymes that allow bacteria to cut viruses out of the DNA. You can use that same enzyme to cut mammalian cells in DNA including human DNA. This opened up the possibility to remove a specific gene from an animal, or potentially a person, and replace it with another gene. Mechanisms other than CRISPR CAS-9 cut DNA in many places and lead to many side effects. Think about genetic diseases or chronic diseases in which specific genes play a role. The question is whether you can actually correct a genome. That is the hope that CRISPR CAS-9 offer. You can take a genetic disease and correct the genome for those cells and potentially cure the condition.
Have you achieved any significant breakthroughs?
These are very early days. Right now all the work with CRISPR CAS-9 is happening in animals since we first need to establish the safety of doing this DNA cutting. I expect work in humans to begin in the next three years. Another area is cell therapy—where you take cells out of a person’s body, reprogramme and put them back into the body. We are in the midst of filing a drug (application) in the US for (treating) cancer in children, and we will be one of the first companies to bring forward a cell therapy to tackle cancer.
Novartis does a lot of research work on neuroscience too. How easy is the task given that the brain is still little understood?
Neuroscience is a humbling space. It is perhaps one of the most difficult spaces for drug development in the industry. We are still at the very beginning of understanding the functions of the brain, which is incredibly complex. Neuroscience is an area where Novartis is involved in for over 50 years. Right now our focus is on Alzheimer’s disease where we have a major programme to prevent the onset of this disease. By 2050, it is expected that Alzheimer’s will become a major crisis.
We are also focusing on multiple sclerosis (a chronic disease that attacks the central nervous system—brain, spinal cord, and optic nerves). Other than that, we are working on a new drug for migraine, etc. even as we continue to do research in neuroscience. What’s heartening is that we are actually seeing drugs that are having a significant impact on diseases (linked to the brain). We have to invest through failures. Companies should not pull back from investing in (discovery of drugs to treat) neurodegenerative diseases just because the incidence of failure is much higher. Moreover, we actually grow neural networks in a petri dish and then try to test these medicines on these neural networks because animal models do not help much in neuroscience.
How are you advancing research with companies like Microsoft, Google and Qualcomm?
The science is getting so complex that we have to work with smaller and larger companies. Consider multiple sclerosis, where a lot of evaluation is done by the human eye. This implies that you are depending on the physician’s ability to look at the coordination and movement of patients. Technology can help make this task more quantitative and consistent. With Microsoft Kinect, we developed an algorithm and a tool (Assess MS) that allows us to perform this task. We then did a study to compare the results from doctors and this technology tool in which the latter proved to be more consistent. Now we are trying to put this into clinical trials. We are yet to see the results. Next, we have to go to regulators to get permission for the same.
We do quite a bit of work on machine learning and deep learning, not only in clinical trials but we also invest in understanding how to develop medicines more efficiently. We took 10 years of history in Novartis on how we develop drugs from clinical trials and worked with an external group to develop a machine learning algorithm to help us predict how we run our operations. We will be rolling out this technology this Spring—we call it “Nerve”. We hope to soon use Nerve in other areas too.
With Google, we have collaborated on the Google Lens. It is a complex project—adding a sensor to a contact lens, which could then allow near or distance correction and do away with the need to have reading glasses. The very interesting part of the sensor that Google is developing is that it can continuously monitor things like glucose or other blood parameters that you can find in tear ducts. The question now is: Could you connect that contact lens wirelessly to a device and explore organs like the human pancreas where the insulin is pumped in from a device?
With Qualcomm, we have invested in a range of technology companies—those looking at microelectronics to correct human disease. For instance, we have a company that proposes to take clinical trials to patients; other companies are trying to make apps therapeutic, especially for neurological diseases. It is a race to see who can scale these technologies.
Moreover, we have close to 100,000 patients in clinical trials and have numerous clinical trial sites in 52 countries. We can now use deep learning predictive algorithms to continuously monitor all those trial sites on 50 variables that help us predict that a quality issue may happen. This will help with regulatory bodies.
Are regulatory bodies able to keep pace with these innovations?
One of the most important things is to engage regulatory bodies very early in the process; to start to teach them about these therapies so that they can get comfortable, allowing us to move it to human clinical trials. We are obligated to do trials in a very step-wise way—a very controlled approach. A very big thing for regulatory agencies is how they will evolve the regulation to allow us to move forward with these technologies.
Is it easier to do clinical trials in emerging countries where the laws are not so stringent?
We use the same standards everywhere in the world. What is easier, or harder, is the amount of data and information that a regulator would ask for before permitting the trial to begin. In India, permissions are now given in about four-six months from almost a year earlier. India is aspiring to do this in the one-two month framework, which will then be world class.
