CASE Study:

EPOGEN--COMMERCIALIZING THE FIRST BIOTECH BLOCKBUSTER DRUG

Amgen Inc.'s Epogen-Commercializing the First Biotech Blockbuster Drug Thirty years after its inception, the biotechnology industry continued to dazzle scientists and investors. Beginning in 2003, biotechnology companies had started to receive a larger number of new- drug approvals from the U.S. Food and Drug Administration (FDA) than big pharmaceutical firms had. The biotechnology industry now generated $43 billion in revenues from the sale of more than 200 drugs. Another 55 biotech drugs awaited approval in the United States alone. These included therapies for cancer, congestive heart failure, pain, and diabetes. Biotechnology firms, most of them privately owned, raised almost $17 billion in capital from public and private markets in 2004. Of all the stars among the biotechnology firms, none shined brighter than Amgen Inc., the producer of blockbuster drugs such as Epogen, Neupogen, and Aranesp. With a market capitalization in excess of $100 billion, Amgen had matured from a promising start-up to one of the most valuable U.S. companies. In 2004, Amgen offered more than a dozen different drugs and had another five in phase three clinical trials. Despite its recent successes, Amgen remained heavily dependent on its early blockbuster drugs. Epogen, Amgen's first product, was still responsible for 26% of the company's $10 billion in product sales. Neupogen contributed another 29%.2 Even in 2004, Fu-Kuen Lin, the Amgen scientist who had discovered the gene coding for human erythropoietin (EPO), the breakthrough that would lead to the production of Epogen, graced the cover of Amgen's annual report. While many saw Amgen as the quintessential biotech success story, the commercialization of Epogen had been a difficult struggle-to say the least. EPO-the Early Days In 1980, when George Rathmann, a former director of research at Abbot Laboratories, accepted the job as Amgen's CEO, he and the company's founders, Franklin "Pitch" Johnson and Bill Bowes, understood biotechnology to be an exciting but unproven business opportunity. In a significant strategic decision that set Amgen apart from most other biotech start-ups, which developed improved versions of existing drugs, Rathmann, Johnson, and Bowes set their eyes on producing something new, human erythropoietin (EPO). EPO was a hormone that prevented anemia, a lowcount of red blood cells. Because almost all the body's EPO was made in the kidneys, anemia was a significant problem for people with chronic kidney disease." Anemia led to severe fatigue and increased risk of cardiovascular disease and even death. At the time, the market for EPO, which included dialysis patients and people with cancer undergoing chemotherapy, was estimated to be a $1 billion opportunity." However, isolating the gene for EPO proved to be a staggering task. It took Fu-Kuen Lin and his assistant Chi-Hwei Lin two years to identify the EPO gene on a single fragment of DNA. Success came in October of 1983 when the researchers used a novel method involving multiple short strands of DNA as "probes" to fish for the gene. This technique allowed Lin to isolate EPO on one of the 1.5 million fragments of the human genome." In February of 1984, Amgen successfully transformed Chinese hamster ovary cells to produce biologically active EPO." Protecting Amgen's Intellectual Property As soon as Amgen had achieved its scientific breakthrough, the company's lawyers sought ways to protect the intellectual property. Biotechnological products and processes could be patented if they met three requirements. Lisa J. Raines, vice president for government affairs at the Industrial Biotechnology Association, explained: First, you have to show it's novel-nobody ever made the product before. Second, you have to show it's nonobvious in light of the prior art. For example, if nobody had ever put a pencil and eraser together before, it's novel. But if pencils and erasers both existed in the past, it didn't take a genius to put the two together, and so it might be barred from patentability as obvious. The third criteria, the easiest to meet generally, is utility-it has to somehow be useful. For EPO, Amgen hoped to win three types of patents. First, the company applied for a product patent. In 1980, the U.S. Supreme Court had granted General Electric the first product patent for a living organism, arguing, in the words of Thomas Jefferson, that "ingenuity should receive a liberal encouragement. . . . Anything under the sun that is made by man," declared the Court, could be patented by man." Product patents afforded the best protection for new drugs. Armed with a product patent, companies had the right to exclude competitors from making, using, selling, offering to sell, or importing the patented invention." However, product patents were difficult to obtain for many biotechnology drugs because they were often highly purified proteins that had previously been discovered in their naturally occurring forms. The mere mention of their discovery or isolation in the scientific literature was enough to qualify as prior art. " Raines pointed to insulin as one famous example: "In 1951, Frederick Sanger succeeded in identifying the chemical structure and precise molecular weight of human insulin. This discovery won him the Nobel Prize, but it couldn't win him a patent" because physicians at the University of Toronto had isolated a miniscule amount of insulin in 1921.12 Under the "product of nature" doctrine, human proteins that had previously been identified were considered neither new nor novel. Process patents afforded the next best level of protection. These patents covered the method to make the final product. Raines explained: "Since genetic engineering is the only commercially feasible method for manufacturing human proteins, a patent on the recombinant manufacturing process can be tantamount to a product patent for biotechnology products, and many have been granted. "14 In practice, the dividing line between product and process patents was often difficult to discern. John R. Thomas, associate professor of law at George Washington University, explained:Even the most novice claims drafter would encounter scant difficulty in converting a patent claim from artifact [i.e., a product claim] to technique [i.e., a process claim] and back again. For example, consider the following artifact claim: An apparatus for measuring activity of the autonomic nervous system of a patient, comprising: means for obtaining ECG signals from said patient whilst said patient is at rest; [and} means for measuring the R-R intervals for adjacent PORS portions of said signals . .. A few simple changes to the claim transforms it to one concerned with technique, in the following way: A method of measuring activity of the autonomic nervous system of a patient, comprising the steps of: obtaining ECG signals from said patient whilst said patient is at rest; [and] measuring the R-R intervals for adjacent PORS portions of said signals . . .15 To enforce a process patent, a company needed to show that its competitor was actually using the patented process. Simply owning or selling a machine that was capable of performing the process did not generally infringe a technique claim, even if the only practical use of the machine was to perform the process."However, holders of process patents could pursue a charge of indirect infringement if a competitor "actively induced infringement," for instance by selling the machine along with instructions on how to perform the patented process. Although process patents generally afforded weaker protection than product patents, protecting a process also had its advantages. For instance, customers who had legally purchased a patented product were free to resell the invention. This "first-sale doctrine" did not apply to process patents. Thomas explained: "The result is that the holder of a patent with technique claims can more extensively control the use of his technology in the marketplace, while the proprietor of an artifact claim must employ contractual or technical mechanisms to maintain downstream control of his technology." 19 Another advantage of process patents was their greater level of abstraction. Many process patents were worded in quite general terms, covering in fact several and sometimes unknown methods to achieve the desired outcome. For example, Alexander Bell's method for "transmitting vocal sounds telegraphically by using electrical undulations" covered the use of variable resistance to achieve the "undulations," even though Bell had relied on electromagnetism, an inferior method, and had not considered using variable resistance. As with product patents, goods that were manufactured abroad using a patented process could not be imported into the United States.20 A final form of protecting Amgen's invention was to patent the starting materials, the host cell and its vector. Raines explained: "If an end product is not patentable because it lacks novelty (as in the case of insulin) . . . the inventor may nevertheless patent the starting materials. Obtaining a patent on a new DNA molecule or on the genetically engineered cell containing the inserted cell is relatively simple."21 A weakness of patents on starting materials was that they did not prevent foreign manufacturers from using these starting materials and exporting the final drug to the United States. "Recombinant biotechnology involved the alteration of a host cell so that the cell would replicate a desirable protein. In this process, the vector acted as a vehicle for inserting the gene into the host cell. When a vector successfully incorporated the isolated gene into the host cell, the cell was said to be transformed. Transformed host cells could be patented because they constituted a type of "machine" capable of producing a desirable protein (Thomas, 1998: p. 237).Surprise! Amgen applied for all three types of patents: starting material, process, and final product. The company heard back from the U.S. Patent and Trademark Office (PTO) in October 1987. The news was decidedly mixed. The PTO granted Amgen a patent on the host cell and vector but rejected the company's process claim. The latter decision was not completely unexpected. A 1985 federal circuit decision, dubbed "the burden of Durden," had made it considerably more difficult for biotechnology firms to obtain process patents. The case, In re Durden, involved a known chemical process that relied on a novel starting material to produce a new type of carbamates. The process itself, however, was well-known, as Union Carbide readily conceded: "The claimed process, apart from the fact of employing a novel and unobvious starting material and apart from the fact of producing a new and unobvious product, is obvious."2? The PTO denied Union Carbide's application for a process patent, arguing that a new process could still be obvious even if "the specific starting material or resulting product, or both, is not to be found in the prior art."2 The PTO subsequently applied the Durden decision to hundreds of process patent applications. Genentech estimated that 60% of biotechnology product patents that were not accompanied by process patents could be directly linked to Durden.24 The most shocking news that Amgen received, however, concerned the product patent application. It was denied! EPO had already been patented. The holder of the patent was Genetics Institute (GI) of Cambridge, Massachusetts. GI had received a product patent for human EPO and the pharmaceutical compositions containing EPO.25 Genetics Institute Genetics Institute was founded in 1981 by two Harvard scientists. The chairman of Genetics Institute was Benno Schmidt, a financier who was also president of J.H. Whitney, the investment banking firm. GI succeeded in isolating EPO from human urine in July of 1984, five months after Amgen's breakthrough. Although it had isolated EPO, the company was not able to produce meaningful commercial quantities of the drug because its scientists did not know which gene produced the protein. Isolating EPO from urine on an industrial scale was not an option because it took almost 6 million gallons of human urine to treat one patient for a year. Legal Stalemate Learning of Amgen's patent, Genetics Institute asked for a cross-license, which would enable the company to produce EPO using Amgen's starting material. In response to the request, Amgen filed suit in Federal District Court in Boston charging patent infringement." In its patent covering the starting materials, Amgen had claimed all possible DNA sequences that could serve as functional substitutes or "analogs" for natural EPO. In Amgen's interpretation, all proteins with the biological properties of EPO were analogs, making it certain that Genetics Institute would have to infringe on Amgen's patent when it attempted to produce EPO.29 Genetics Institute countersued, charging Amgen with infringement of its product patent on EPO. In December 1989, the court ruled in favor of Genetics Institute, arguing its patent covered EPO irrespective of how the drug was produced." By taking up production of EPO, Amgen had in fact infringed Genetic Institute's patent." Amgen's lawsuit against Genetics Institute was only partially successful. The problem with Amgen's analogs claim, the court ruled, was that there were thousands of potential analogs for EPO whose effects Amgen did not understand. For example, replacing a single amino acid in the EPO created 3,600 potential analogs. Substituting three amino acids created over a million. In the course of its research, Amgen had produced 50 to 80 analogs, but the companycould not reliably predict whether or not these analogs would have the same effects as human EPO.32 Arguing that Amgen had only disclosed how to make "the gene and a handful of analogs whose activity has not been clearly ascertained," the court did not afford Amgen the broad protection it had sought. 33 However, the legal ruling did uphold Amgen's more specific patent claims related to the isolated DNA sequence and the Chinese hamster cell line. Genetics Institute had argued that Lin had failed to deposit the Chinese hamster cells with a publicly accessible repository as required by law. But the court did not follow GI's argument, ruling that Amgen had provided a sufficient description of the method it had used. A person skilled in the art, the court declared, would have no difficulty replicating Amgen's procedure. Genetics Institute had infringed Amgen's patent when it started producing EPO.34 The legal stalemate-neither company was able to produce EPO without infringing the other's rights-had come at considerable economic cost. Amgen's legal bills reportedly amounted to 10% of Epogen's development cost,35 and the stock price of both companies came under severe pressure.*6 A few weeks after the court decision, Amgen received a letter from Genetics Institute. Stating his demands, Schmidt was ready to play hardball: "You will execute a royalty-free license with us. I can enforce my patent against you, I can shut you down, I can take you off the market."37