Better Drug Discovery Through Better Target Identification

Is HTS a breakthrough or a bust? Second-generation technology holds promise for identifying the best agents for targets earlier in the game.

Abstract

High-throughput screening has the potential to shave hundreds of millions of dollars from the cost of research and development. Hailed as a major technological breakthrough 10 years ago, HTS has yet to live up to the hype. But scientists involved in second-generation HTS techniques say we’re getting close.

Out of every 6,000 births, one baby enters the world lacking a key biological ingredient: a critical protein essential for the survival of its motor neurons, which send nerve signals from the spinal cord to muscles. Infants lacking this protein develop spinal muscular atrophy (SMA), in which their muscles progressively weaken, resulting in paralysis and respiratory failure. SMA is the leading genetic cause of death of children under the age of 2 in the United States. It’s a methodical killer — with no cure or serious hope for a treatment.

Until now.

Using high-throughput screening (HTS) and a new software program designed in a lab at Columbia University, researchers from Columbia and other institutions recently went looking for a way to spur production of the missing survival motor neuron protein (SMN protein). The reason for its deficiency is the absence of the SMN1 gene, which directs the biosynthesis of most of the body’s SMN protein. Another gene, SMN2, also produces some SMN protein — but in quantities too small to stave off SMA.

Running a cell-based assay, the researchers analyzed a mountain of data covering more than 47,000 compounds in an effort to find a substance that might induce SMN2 to produce more SMN protein. The assay employed cells modified to report the presence of SMN protein by becoming luminescent.

Their search yielded two hits. One of those compounds, aclarubicin, already was known to be involved in SMN2 splicing. The other is indoprofen, an old nonsteroidal anti-inflammatory drug that could play a role in boosting SMN protein production and battling the disease.¹

By using a host of new technological tools like those developed at Columbia, HTS is helping to identify the best agents for specific targets earlier in the game. And despite a burst of criticism leveled at HTS over the past year, advanced screening programs are improving, catching the eye of big pharma, and appear to be here to stay.

For researchers, target identification is just one step leading to the long slog that goes into developing a new drug (or, in this case, a new use for an old drug) and then proving its effectiveness in clinical trials.

“Screening has been adopted by all major pharma-biotechnology companies,” says Brent Stockwell, PhD, the Columbia University assistant professor in biological sciences and chemistry who led the two-year effort to pinpoint indoprofen as the possible foundation of the first therapy for SMA. Researchers at Ohio State University, the Centre for Inherited Neuromuscular Disease RJAH Orthopaedic Hospital (United Kingdom), the University of Massachusetts Medical School, and the National Institute of Neurological Diseases and Stroke also were involved in the study.

“I would say that the net output of HTS will be the discovery of more and better lead molecules, better targeting of unmet needs, with fewer side effects,” Stockwell says. But, he adds, HTS is silent on in vivo toxicity — an issue only now in the beginning, painstaking stages of being sorted out.

For Stockwell, his work in HTS offers a chance to advance the science. His software package, the Small Laboratory Information Management System (SLIMS), is freely available for academic use.²

Part of the criticism of HTS is justified, says Kalypsys President and Chief Scientific Officer John McKearn, PhD: “Big pharma has been pouring money into HTS, often with little planning and getting little in return.” But, he adds, HTS is an industry standard because it has provided a faster, more precise tool for finding promising compounds. McKearn says Kalypsys isn’t far from proving the value of HTS — it expects to file its first investigational new drug application this year.

PHOTOGRAPH BY ROBERT BURROUGHS

Ultimately, Stockwell adds, it will likely take the deep pockets of a biopharmaceutical company to bring any indoprofen-based drug to market. But plenty of companies are working hard to develop proprietary technologies — for a price. Out-sourcing HTS has become one of the hottest tickets in biotech and is followed by more and more drug makers looking to compress the amount of time and money that goes into the discovery process.

THE NUMBERS GAME

At one point, scientists manually screened compounds in a test tube, testing hundreds or thousands a week. To get to bigger numbers, robots now arrange compounds in wells that contain a disease-linked target. A positive hit allows researchers to refine the compound into preclinical candidates.

The allure is obvious. Tufts Center for the Study of Drug Development estimates that for every 1,000 compounds that are tested, only 5 make it to trial and only 1 is ultimately approved by the U.S. Food and Drug Administration. If better drug leads could be discovered faster, then drug companies could save money and precious time. And in the race to develop new drugs, specialists at HTS have been steadily pressing the envelope to gain ever-larger amounts of data.

In Stockwell’s lab, researchers can hunt through 40,000 compounds a day. Add advanced robotics in a state-of-the-art biopharma facility, and the screens can run up to 200,000 a day.

And that’s just average. A company like Kalypsys, a drug-discovery company based in San Diego, talks about its ultra high-throughput technology that can screen upward of 1 million compounds a day in a variety of assays. Toolmakers like Caliper Life Sciences and Molecular Devices have been steadily upping the speed at which researchers can analyze samples through automation. And there are plenty more high-tech companies engaged in the race for the fastest and most effective way to screen. Factor in the rate at which HTS is developing, Stockwell adds, and it’s easy to imagine a day, later in the decade, that will see millions of compounds.

At issue is whether HTS has led the science of drug discovery down the wrong path: substituting the big numbers offered by robots for the genius and the inspired guesswork that takes place in the lab.

Along with combinatorial chemistry, those ever-growing numbers made HTS one of the most promising technologies of the ’90s, the Economist noted last March in an overview of problems plaguing pharmaceutical companies’ pipelines. The development of HTS is one of the reasons spending on drug-discovery work zoomed from $15 billion in 1996 to $32 billion in 2003.

REINVENTING HTS

But that’s where critics claim the numbers fail to add up. Some leaders in the field still wince at mention of an article in the Wall Street Journal last February that outlined how spending on drug research more than doubled while the number of new molecular entities approved by the FDA fell from 53 in 1996 to 21 in 2003 and attributed the failure to automation and new technologies. And analysts have been growing doubtful of big pharma’s ability to deliver on major new drug discoveries to replace the blockbusters that lose their patent protection every year. In the process, HTS has been held up for some harsh criticism as well.

At issue is whether HTS has led the science of drug discovery down the wrong path: substituting the big numbers offered by robots for the genius and inspired guesswork in the lab that went into identifying most of the big drugs that benefit people today.

“The approaches [HTS and combinatorial chemistry] looked promising, in that they generated lots of hits,” the Economist article noted. “But while the quantity improved, the quality did not. The number of new leads going into clinical testing did not increase, and enthusiasm for the new technologies waned.”

Part of the criticism of HTS is justified, says John McKearn, PhD, president and chief scientific officer at Kalypsys. Big pharma has been pouring money into HTS, often with little planning and getting little in return.

“It can be done much cheaper and faster and smarter than most big companies are currently doing,” says McKearn. “You don’t have to spend hundreds of millions in developing screening, but that’s what Pfizer, Glaxo, and others have done.”

Many of the current systems aren’t that good, says, McKearn — who is quick to add that he has seen it firsthand, once helping a major pharmaceutical manufacturer spend $20 million on a system that moved a chemical from “one end of the laboratory to the other.” Additionally, drug discoverers suffer from compound libraries that are far too old and inadequate to deliver substantial leads, he adds, with or without HTS. Says McKearn: “Few have replaced their compound collections with chemicals that are serious, drug-like molecules.”

Brent Stockwell, PhD, and colleagues at Columbia combined an analysis tool developed in Stockwell’s lab with HTS technology to find that indoprofen could increase production of a protein lacking in patients with spinal muscular atrophy. “The net output of HTS will be the discovery of more and better lead molecules and better targeting of unmet needs with fewer side effects,” says Stockwell.

PHOTOGRAPH BY CAROL SEITZ

But, he adds, the overall criticism is just not valid. “It’s like saying, ‘Advances in computer chips haven’t improved computing, and therefore, computers haven’t improved our lives,’” he says. Moore’s law (the exponential growth every few years in the capacity of computer chips) has worked a revolution in computer science, he notes, and the same principles are at work in HTS. And HTS is an industry standard today because it has provided a much faster, much more precise tool for finding promising compounds. However, very few drug-discovery organizations have seamlessly integrated HTS into their workstream, which is akin to a great computer chip not being used properly inside a computer.

Not that many such compounds are currently FDA-approved — duly noted by the Economist article — but McKearn’s point is that there has been a rapid increase in the discovery of compounds with promising pharmaceutical potential. McKearn believes Kalypsys and other top HTS outfits will be vindicated as compounds find their way through clinical trials. In fact, McKearn says, Kalypsys isn’t far away from proving it — the group expects to file its first investigational new drug application this year, and he holds out the possibility of more than one before the end of the year.

For HTS to prove itself — and quell the critics once and for all, McKearn says — will require “outrageous departures from the norm.” Which is another way of saying that McKearn isn’t setting the bar low.

More specifically, he says, Kalypsys has to demonstrate that you can take an experimental compound to the stage of an investigational new drug application — the iNDA being the starting gun for clinical trials — at a cost of $10 million to $15 million while improving overall quality or success rates. And $10 million, he believes, is an achievable figure.

Right now, the industry spends about $70 million per investigational NDA. If 1 in 10 makes it to the finish line, that represents a cost of $700 million. If Kalypsys can show that a drug-discovery company can cut that to $10 million and improve its shot at an NDA (submitted after phase 3 testing), then “That now decreases the portion of the iceberg that hits the Titanic.”

That’s a bold assertion, but if it is ultimately validated, McKearn says Kalypsys can help spearhead a whole new, slimmed-down drug-discovery approach. Dramatically reduced discovery budgets will allow not only a new generation of drugs at a lower price, but a whole new set of drugs targeting smaller populations — such as infants with SMA — that haven’t offered a big enough return to warrant the investment.

BETTER ASSAYS, BETTER LEADS

That’s all stellar potential, but analysts note that HTS, as an industry, is experiencing a modest slump.

Vikram Wadhwani, an analyst at Frost & Sullivan, a research firm, says the once-torrid, double-digit growth of HTS has slowed to a mere 5 to 6 percent. But Wadhwani isn’t close to counting HTS out. As more drug companies looking to beef up their pipelines shift R&D money out of development and into research, HTS is likely to bull ahead at a respectable pace of 9.5 to 10.5 percent a year, growing the business in the United States from a total of $3.1 billion in 2003 to $6 billion by 2010. And most of the HTS money will stay within the country’s borders, even as Europe and Japan push for more HTS.

The technology is still quite compelling, notes Wadhwani. Outsourcing groups have improved capacity exponentially as they continue to drive bigger and bigger numbers of assays for screening. HTS outfits that can handle 100,000 to 300,000 compounds a day are headed for 1 million as advances in miniaturization and microfluidics³ continue to allow greater volumes. And the quality of the screens is rising as well.

Instead of biochemical assays, says Wadhwani, cellular assays will move to the forefront. With cellular assays like the one used to identify indoprofen as a potential treatment for SMA, researchers can see in real time the protein expression that results from a screen.

“The real advantage is being able to do a lot more experiments for the same money and time, with a higher likelihood that downstream resources are put on the right compound.”

— Pratik Shah

“When you expose chemicals to a cell, you will see first hand how the cell reacts to the chemical,” says the analyst. “With a biochemical response, you can’t monitor the cell.”

That way, he explains, researchers are much more likely to see just how good a lead they have. And the sooner you make it to market, the sooner you stop spending money on R&D and start making money.

That drive to cell-based assays is behind a big push to match better biologics earlier in the game with the economies to be had by HTS technologies like the one developed by Kalypsys.

“People hear Kalypsys talk about ‘screening a million compounds a day’ and they think of the old high-throughput screening — poor libraries, mixtures of compounds and less relevant biochemical assays, and unidimensional screening,” says Pratik Shah, who recently set up shop as a venture capitalist with Thomas, Mc-Nerney, & Partners after seeing Kalypsys through its start-up phase as cofounder. “So they say, ‘Oh, it’s just numbers, and numbers are not everything.’ I don’t think people realize more broadly that with these technologies, you can do excellent cell-based assays in 1536 [wells] and get advantage of speed and low operating cost and bring relevant biology to the table and not be restricted to running tens of thousand of compounds over days and weeks. In addition, you get the ability to then profile your best several hundred compounds against a large number of relevant assays in replicate dose responses, quickly and inexpensively. It’s more like high-throughput profiling.”

But there are other, real limits in terms of the number of targets any biopharma company can screen against. In reality, he says, most large R&D operations can run 20 to 30 screens in an entire year. For those working in four or five therapeutic areas, it can be no more than a handful of targets for a given therapeutic area like cancer or inflammation. “So if you think about how many shots on goal most people have with high-throughput screening today,” he adds, “it’s not that many.”

The objective is to get better biologic information faster.

“That occurs in two ways: better assays up front and high-throughput profiling. I only want compounds that will work in vivo and have the right profile,” theorizes Shah. “Therefore, your initial readout needs to be a lot more biologically sophisticated and accurate. Does it have a relevant biological effect in a living system? At the end of the day, that’s what counts. Bring that filter up front and screen a million diverse compounds, and you’re more likely to get lead compound starting points that make sense to follow-up on.”

GROUNDS FOR COOPERATION

A number of HTS outsourcers are making this argument to pharmaceutical companies. Part of the daily flow of biotech news almost always includes some new collaboration between the developers of high-throughput technologies and the drug companies eager to get good targets into clinical trials. The same day Germany’s BioVision inked a deal with Novartis to use its tech to discover the most clinically relevant peptides, Applied Biosystems was touting a new rat genome survey microarray that could be used to “interrogate” 27,088 genes.

Biogen Idec, which has been making exceptional headway in oncology and inflammation, recently signed a throughput deal with Dyax, a small company with headquarters in Cambridge, Mass., that uses its phage display libraries to develop biologic leads. “Biogen Idec provides funding and up to 30 protein targets a year, and we’ll apply our libraries to deliver back to them antibody leads,” says David Buckler, PhD, Dyax’s director of selections and screening for lead discovery. For Dyax, it’s a chance to partner with bigger companies in the drug-discovery process — a pairing that holds the promise of future milestone payments and royalties if their hits turn into patented products.

Dyax isn’t just a hired gun. Its researchers are also developing therapeutic drug candidates of their own, often with partners. Today, two of its candidates are in phase 2 clinical trial for three different indications. The FDA recently gave one of its experimental drugs — DX-88 — fast-track status. Dyax and its partner, Genzyme, have successfully completed two phase 2 trials of DX-88 for the treatment of a hereditary angioedema (HAE). The recombinant small protein inhibits human plasma kallikrein, a key enzyme in the inflammatory cascade. That mechanism may create a treatment for HAE, a rare and sometimes fatal inflammatory condition for which there is no cure. Patients with hereditary angioedema have a deficiency in the gene that produces the molecule that naturally inhibits kallikrein.

Dyax has inked deals with a number of other collaborators interested in their phage display technology that can identify compounds that bind to therapeutic targets. The company says it can gain an edge over others by focusing on antibodies, in addition to small proteins like DX-88.

“A fully human antibody is something that is naturally accepted by our bodies,” says Buckler. And, with fewer toxicity issues and exquisite specificity, biologics have a natural edge over the typical small molecules that are screened.

But in this field, the ultimate goal is the therapeutic that makes it all the way to the end of the pipeline with a green flag from the FDA.

“It’s really about trying to pick the right molecules,” says Shah, the venture investor. “The real advantage [of HTS] is being able to do a lot more experiments for the same money and time, with a higher likelihood that downstream resources are put on the right compound.”

In that sense, the future of HTS is currently in clinical trials.

There is a lot riding on the outcome.

FURTHER READING.

Background on the topic of HTS

Dove A. Screening for content — the evolution of high throughput. Nat Biotechnol. 2003;21:859–864.

Mullin R. Drug discovery: as high-throughput screening draws fire, researchers leverage science to put automation into perspective. Chem Eng News. 2004;82:23–32.

Details of the indoprofen study and the software described early in the article

Kelley BP, Lunn MR, Root DE, et al. A flexible data analysis tool for chemical genetic screens. Chem Biol. 2004;11:1495–1503.

Lunn MR, Root DE, Martino AM, et al. Indoprofen upregulates the survival motor neuron protein through a cyclooxygenase-independent mechanism. Chem Biol. 2004;11:1489–1493.