The Collateral Benefits of India's Open Source Drug Discovery Programme

On a computer screen, UC Jaleel, a pharmaceutical researcher, shows me a complicated blue structure suspended against a black background. It looks like the outline of a giant lump of quartz. Jaleel informs me that this structure is actually a tiny protein called beta lactamase, secreted by the infamous mycobacterium tuberculosis, the germ that causes tuberculosis. Beta lactamase is one of the foes in our war against tuberculosis, because it makes tuberculosis resistant to a range of powerful antibiotics called beta lactams, leaving us with a smaller arsenal of drugs to fight the disease with. Pharmaceutical researchers are therefore searching for new drugs to block beta lactamase.

Jaleel and his team, a part of India’s ambitious Open Source Drug Discovery programme (OSDD), are among these researchers. Each day, they evaluate almost five lakh chemical compounds, in an attempt to find those that can paralyse targets such as beta lactamase. When they do manage to find such ‘hits’, as they are called in pharmaceutical parlance, it is only a prelude to a long and painstaking process of drug development. Several years of lab work, pre-clinical and clinical trials later, there is a small chance that one of these molecules will turn into a much-awaited new drug against tuberculosis.

But for that to happen, OSDD must train its own phalanx of drug discovery experts, who are currently scarce in India.

When India’s Council for Scientific and Industrial Research (CSIR) launched OSDD in 2008, the idea promised to disrupt the current drug-discovery paradigm, which is failing in diseases of the developing world. No pharmaceutical company wants to develop drugs for tuberculosis, malaria, or leishmaniasis, because the patients, typically poor, cannot afford to pay.

So, instead of investing their hopes in profit-driven corporations, OSDD threw the doors open to academic, government and private-sector researchers all over India and the world. The collaborative process of drug discovery meant an almost unlimited talent pool. Theoretically, the potential was immense.

But six years later, the challenges of this model are becoming apparent. As OSDD progresses beyond early stage discovery to late stage development, the lack of a critical mass of drug-discovery professionals in the Indian academic world is hampering progress. Also, the academic world does not have the system of training and incentives that make pharmaceutical companies tick.

“Typically the discovery timeline should be about two or two-and-a-half years. After that you should be talking a different language,” says Ram Vishwakarma, director of the Indian Institute of Integrative Medicine, a CSIR lab that partners with OSDD. “Now, we need to bring international-level expertise, and some of it will cost money.”

Addressing the Talent Crunch
One area of expertise that is missing in the Indian research landscape is cheminformatics.

Cheminformatics experts have a role to play at the crucial hit-identification stage of drug discovery. They use software to screen lakhs of compounds for tell-tale signs of effectiveness against a particular disease. For example, Jaleel’s team uses a training database of molecules known to be effective against tuberculosis to glean out structural properties common to all of them. Jaleel compares the process to the spam filter of Gmail. “How does Gmail know that certain emails, promotions are spam? It uses artificial intelligence. Cheminformatics is also like that,” he explains. Next, Jaleel’s team combs through an open database of 166 billion drug-like compounds, known as GDB-17, to find those that have properties similar to the training database.

But the lack of enough good cheminformatics experts in India is slowing things down for OSDD. In 2013, the Sir Dorabji Tata Trust gave OSDD a Rs 2.85 crore grant to create fellowships for PhD students. These students are now being trained in the discipline, even as they screen molecules. Jaleel says that while the enthusiasm and youth of these students is a distinct advantage, there are conceptual gaps in their understanding, and it will take time to turn them into experts.

OSDD is doing everything it can to address the gap. In February, it conducted a cheminformatics workshop in Pune where its 95 principal investigators got together with members of the UK’s Royal Society of Chemistry and developed their own cheminformatics algorithms for hit identification. Such in-house tools are necessary for OSDD, because the private sector does not make its proprietary tools available to others. Also, these proprietary tools are not designed with neglected diseases in mind.

The second talent gap is medicinal chemistry. When scientists identify a promising molecule after virtually screening compound databases, the hit they are left with is something of a blunt weapon. While it is capable of attacking the disease-causing bacteria, it often has too many side effects to become a viable drug.

For example, OSDD currently has a molecule developed in Lucknow’s Central Drug Research Institute (CDRI), known as the CDRI 830. “From a perspective of TB activity, this molecule is ready to go. But the chemical groups on it have been seen to cause cardiac problems,” says Tanjore S Balganesh, who heads OSDD. “So we have to detoxify it.”

This is where medicinal chemists come in. They sculpt the molecule, knocking off toxic structures and retaining useful ones. They ensure the molecule is not too complex, in order to keep costs in check.  It is a long and painstaking process. But, according to Balganesh, medicinal chemists are few in number in academia across the world. A handful of universities teach medicinal chemistry and typically, researchers learn on the job in large pharmaceutical companies.

Balganesh says he has seen a clear move by medicinal chemists from academia into the industry over the last five years. But there isn’t a reciprocal move back, because academia doesn’t offer any incentives for this. “The only incentive OSDD can give them is emotional—I can’t give them high salaries or glamour.”

Meanwhile, it is hard to train the medicinal chemists in OSDD because of the distributed nature of the project, says Bheemarao G Ugarkar, a principal investigator for medicinal chemistry in OSDD. Close collaborations between various researchers is critical for drug-discovery research. “A lot of the training in the industry happens in a group environment with biologists, computational chemists, synthetic chemists, etc. sitting in one room day in and day out, discussing the project,” explains the researcher, who was previously with AstraZeneca India. Unless OSDD finds a way to mimic such a collaborative environment, Ugarkar says, the cycle of developing and testing a drug will remain that much slower.

A talent gaps exists in the discipline of biology too. The biggest problem, Balganesh says, is that Indians “are too domain oriented. People do mathematics because they don’t like biology; they do biology because they don’t like math. But in medicinal chemistry, you have to make math and biology talk to each other.”

One area where OSDD has a distinct strength is synthetic chemistry. “Indian synthetic chemists are the best in the world. Most of the synthetic chemists at CSIR come from Nobel laureate laboratories,” says Balganesh. Moreover, those who work in the academic environment, unconstrained by the pressing demands of corporations, have the ability to develop more complex molecules. “In the corporate world, if you cannot synthesise molecules in three-four steps, you are asked to leave it. But these scientists want to make complex ones because it is useful for publishing.”

The challenge is to get these chemists to create more and more diverse compounds. The libraries of the pharmaceutical companies are currently geared towards profitable diseases such as cancer and heart disease. On the other hand, any new tuberculosis drug will be an antibiotic, which typically has a different, more complex structure than a cancer drug. “It is like a square peg in a round hole,” Balganesh says about trying to find a TB drug in existing compound libraries.

According to Vishwakarma, though, the solution lies in bypassing open-access databases altogether and gaining access to the databases of large pharmaceutical companies by forging stronger ties with them. He believes open-access databases aren’t of great quality. “Pharmaceutical companies are the only ones with real drug-like compounds. These have evolved over 60-70 years of hard work,” he explains. “If Pfizer, for example, decides to open all its compound libraries, I am hundred percent sure I can find a drug for malaria.”

Getting the private sector to co-operate will be the next big challenge for OSDD. India’s patent regime has traditionally been unfavourable for innovation. Even today, compulsory licences make big pharma firms jittery. “We need to build a trustworthy ecosystem,” says Vishwakarma, “We cannot overhype or underplay our strengths. And we need humility.”

By far, the most important challenge in open source drug discovery is going to be fostering a conducive environment for results-based research. Currently, Indian academicians are evaluated by the number of publications they put out each year. Once a researcher publishes a result, he has no incentive to work on it further. But product development requires several years of sustained efforts and, currently, such efforts are not rewarded. “Every biotech institute in India judges only by publications. I argue with these people: It doesn’t work. Why would I get in and work on a product if 15 years later I am nowhere?” asks Balganesh.

It is going to be an uphill task for OSDD to create what Balganesh calls “a cultural change”. While OSDD’s core team has an industry-like incentive system, it will be hard to change the way OSDD’s partner organisations function.

Reason for Optimism
In 2010, OSDD had a major success when it annotated the genome of mycobacterium tuberculosis. This was an important first step for its drug-discovery efforts. According to Anshu Bharadwaj, a principal investigator at OSDD, the tuberculosis bacteria has 4,000 proteins and 950 metabolites. The interactions among these proteins need to be understood to find the most crucial ones among them for the tuberculosis bacterium’s survival. But this information is scattered across the literature on TB. “One paper will say protein A interacts with B in certain conditions. Another will say protein C interacts with D in certain conditions. But this information is in two different places,” she explains.

To create a comprehensive map of how proteins interact with each other and with DNA, RNA and metabolites in mycobacterium tuberculosis, OSDD launched a project called Connect2Decode in 2009. Under this, experienced researchers, as well as PhD students, read over 40,000 previously published research papers and pieced together the jigsaw of protein-protein interactions. Once they had this map in place, network analysis was used to find the crucial proteins without which the mycobacterium cannot grow or survive. Such proteins are obviously excellent drug targets for TB.  

But the Connect2Decode project ran into controversy in 2010, causing several experts to raise doubts about OSDD. When OSDD went public with its results before they were published in peer-reviewed journals, several other researchers, including Pushpa Mittra Bhargava, the founder director of the Centre for Cellular and Molecular Biology, Hyderabad, raised doubts about the validity of Connect2Decode’s results and the claims that students had carried out such a complicated project. Talking to the science magazine Nature, Bhargava called such claims “simply hilarious”.

The handling of the episode cast a cloud of suspicion on OSDD, says Jayant Singh, associate director health care practice at Frost & Sullivan, who tracks CSIR closely.

By 2011 though, OSDD laid such doubts to rest by publishing its results in several peer-reviewed journals, including PLOS One.

When I ask Bharadwaj about the scepticism over OSDD’s model of using students for drug discovery, she calls such scepticism unreasonable. While it does pose challenges, OSDD has several processes in place to ensure the quality of their work, she says. For the Connect2Decode project, for example, only researchers with subject-expertise were picked, after which they went through a substantial training period. “We had a very well-defined SOP [standard operating procedure]. We devised a couple of platforms that allowed us to semi automate. We also had a layer of quality control. That is the most important thing,” explains Bharadwaj.

While OSDD could have done without the controversy, the final publication of the results demonstrated that the open-source model could actually work on such complex questions.

The Road Ahead
Balganesh says a drug developed entirely though OSDD will not enter clinical trials for at least another two or three years. Within the next five years, starting 2014, OSDD is targeting at least two compounds against TB in late clinical trials, one molecule for malaria and one molecule for leishmaniasis.

Meanwhile, in January, OSDD received regulatory approval to carry out late stage (phase 2B) clinical trials on a promising new tuberculosis drug molecule called PA 824 in India. This molecule changed several hands before it came to the public-private partnership Global Alliance for TB Drug Development (GATB). GATB conducted phase 1 and 2 trials in collaboration with Novartis in Brazil and Africa, before handing it to OSDD for the next stage.  

Such clinical trials of TB drugs that are not developed internally are going to be a critical part of OSDD’s mission. Clinical testing is currently the biggest lacuna in tuberculosis drug discovery, because large pharmaceutical companies don’t want to pay the high cost of testing a drug on humans. “The most expensive part of drug discovery and most time consuming and risky are clinical trials. But the market is absent here,” says Zakir Thomas, an ex-commissioner of Income Tax who became project director at OSDD.  

The trial for PA 824 is going to cost Rs 25 crore. According to Samir Brahmachari, ex-director of CSIR and the brain behind OSDD, OSDD has applied for a grant of Rs 600 crore from the government, which has been cleared by the Finance Expenditure Committee. In the meantime, the trial will be funded with money from CSIR and by raising philanthropic funds.

Creating an Ecosystem
Will OSDD ever be able to create an entirely new drug against tuberculosis, malaria or leishmaniasis? There are doubts.

Frost & Sullivan’s Singh, for one, is not too optimistic. “The risk in producing a new drug is huge. They need expertise, and they lack significantly,” he says.

Molecular biologist Richard Jefferson, the founder of Cambia, an international non-profit that promotes innovation for social good, feels similarly. Even though Jefferson himself is a great believer in collaborative innovation, he thinks the open-source model may not work for drug discovery, because of the magnitude of the problem.

In its efforts to overhaul the research ecosystem, however, OSDD may see successes that have nothing to do with whether they create a drug or not. Singh believes the project will produce better-trained professionals in drug discovery, who can then be harvested by the private sector.

It has happened before.  

The National Institute of Pharmaceutical Education and Research (NIPER) in Ahmedabad, was created to raise the quality of pharmaceutical research in India. Among its objectives was the creation of novel medical devices, says Singh. Even if such medical devices were never created, today NIPER is a source of talent for the pharmaceutical industry. This could be OSDD’s contribution too. Just creating a strong research ecosystem is “a good endpoint”, Singh says. “After all, that is the role of the government: To act as a facilitator, not a provider.”