Errors Found in Patent for AIDS Gene, Scientists Say

Scientists have uncovered what they believe are glaring errors in a patent issued last month to Human Genome Sciences Inc. for a human gene that plays a crucial role in AIDS.

The potential setback comes amid concerns that the race to patent human genes could lead to shoddy science and profiteering. Indeed, Human Genome Sciences’ shares have fallen along with the rest of the biotech industry’s after a spate of negative news recently.

The company’s description of the chemical makeup of the gene contains at least four significant mistakes, according to research scientists, an allegation that legal experts say could allow the company’s competitors to attack the patent’s validity.

But Human Genome Sciences officials say that because the company isolated the gene first, any errors in describing it won’t matter--it is still entitled to royalties from anyone using the gene to discover new treatments.

The scientific challenge to the company’s claim is likely to be one of many in the industry, as the race to be first with genetic breakthroughs inevitably leads to challenges to U.S. patent policy.

“This is a perfect example of the rush to sequence [human genes],” said Christopher C. Broder, a former member of a National Institutes of Health team that did pioneering work on the role of the gene in AIDS. “They get it wrong. They don’t know the function [of the gene]. That is what I have problems with: the whole notion of the rush to patent genes.”

Academic scientists such as Broder are particularly angry that companies such as Human Genome Sciences have been able to isolate and analyze genes by the tens of thousands and have won hundreds of patents for their efforts.

When Human Genome Sciences announced last month that it won its patent on the AIDS gene, its stock price soared, though it has since dropped.

Company officials say they are confident the patent issued last month by the U.S. Patent and Trademark Office will withstand legal attack. Any errors in the company’s description of the gene and the protein it produces are of no consequence, they contend.

The description of the gene, also called its sequence, is spelled out in the patent as a series of letters representing the order of the chemical compounds in its DNA. However, the patent also refers to an actual copy of the gene in a living cell that Human Genome Sciences deposited in the American Type Culture Collection in Virginia.

“When we file a patent, we don’t claim the sequence as the invention,” said William A. Haseltine, chairman and chief executive of the Rockville, Md., company, which has filed about 7,500 gene patents. “The invention we claim is the gene we deposit with the ATCC. We know that our sequence and most sequences are not perfect.

“Anyone who wishes can go to the ATCC,” he said. “It’s the same as the olden days, when inventors used to deposit a little model of their inventions.”

Similar deposits are used to underpin patents for such things as a new strain of brewer’s yeast. But it is not clear that the same rules would apply to a gene that is spelled out by the order of chemical letters in its DNA.

Patent experts say significant errors in that sequence could loosen the company’s hold on what may well prove to be a valuable piece of intellectual property.

“A patent must claim the invention in enough detail so that the public knows what is and what is not protected,” said Stanford Law School professor John H. Barton. “If you don’t make the written patent final--subject, obviously, to some arrangement for correction if there really is a clerical error--then the rest of the world doesn’t know what is patented, and that isn’t right.”

Another patent specialist, Craig Jepson, a professor at Franklin Pierce Law Center in Concord, N.H., believes that the company could prevail but that the outcome is far from certain.

“It sounds like the patent may have problems,” he said, “but I couldn’t say they are fatal.”

The scientific challenge to the company’s claim is but the latest twist in a tale that illustrates both the hope of biotechnology to provide new understanding of disease and the raging controversy over commercial rights to the human genetic heritage.

Genomics companies such as Human Genome Sciences and its competitors Celera Genomics and Incyte Pharmaceuticals are able to cast a broad net and haul in genes and gene fragments by the tens of thousands, using automated machines to spell out genetic code and sophisticated software to make reasonable guesses about their function.

Haseltine insists that his company is different from its rivals--that it does not rely solely on software but does experiments to determine the role played by the genes it seeks to patent.

“You would think that all we did was find a gene and use a computer to find its function,” he said. “We made a real invention and we made it first.”

The controversy started quietly enough five years ago, when researchers at Human Genome Sciences isolated a particularly interesting gene and quickly determined it was a member of a class that produces protein receptors. These substances sit on the surface of cells like antennas, ready to pick up chemical signals from the body. Haseltine said company scientists showed that the gene was a receptor for chemokines, which appear to play a role in inflammatory diseases such as arthritis.

He acknowledged that the company had no idea that this particular gene and receptor played a role in AIDS.

In the months after the company filed its patent, however, other researchers at several academic centers, including the Aaron Diamond AIDS Research Center in New York and the NIH, made a remarkable series of discoveries about how HIV, the virus that causes AIDS, enters living cells, the first step in an infection.

Unaware of the company’s patent application, these scientists found and isolated a protein that the virus requires for entry--the CCR5 receptor. And they isolated the gene that carries the instructions for building the receptor. A drug that can block the protein could be a new weapon against AIDS.

Haseltine complains that several of the government-funded groups that made the discoveries have applied for patents of their own. “The fight going on now is not a disinterested fight about who gets credit for it, but who gets money for it,” Haseltine said.

The academic scientists counter that it is simply unfair to award ownership rights to a gene to a company that had no idea of its function in disease and that did not even spell out the gene and protein sequence correctly in its patent.

Broder, now on the faculty of the Uniform Services University in Bethesda, Md., discovered the errors in the Human Genome Sciences patent the day it was announced.

Broder said he did a quick comparison of the amino acid building blocks of the CCR5 protein described in the company’s patent and the protein that had been filed in a public database. He noted that four out of the 352 amino acids in the protein were incorrectly identified in the patent--each error corresponding to a mistake in the genetic code. The protein described in the patent, he said, would not fold in the same way as the CCR5 receptor identified by academic researchers and would most likely be useless as a tool for developing new AIDS therapies.

Broder pointed out the discrepancy to John P. Moore, whose team at the Aaron Diamond AIDS Research Center helped discover the role of the gene in AIDS.

Such fundamental errors show that the work should not earn the company a patent, Moore said in a recent interview. “It’s like patenting an airplane that doesn’t have a tail,” he said. “They know it won’t fly, but they’ll stop everyone who has an airplane with a tail.”

Moore talked to a former postdoctoral fellow in his lab, Tanya Dragic, who agreed to put the patent to the test. Using standard methods, Dragic, now at the Albert Einstein College of Medicine in New York, plans to reproduce the gene described in writing in the Human Genome Sciences patent. Then she will insert it into cells and determine if a working receptor protein appears on the cell surfaces and does indeed interact with HIV.

The process should take a few weeks, she said.

Why are these scientists devoting their energy to attacking the science in the company’s patent?

Dragic and the others argue that the errors are signs of how little is required of companies to win gene patents.

“They use sophisticated equipment and computers that analyze the utility of a gene sequence,” she said. “It isn’t the hard work. It isn’t the innovative work. It’s not fair for others to have to pay licensing fees just because they got lucky.”