Patents for the first wave of blockbuster protein drugs have already expired and dozens more — worth billions in potential generic sales — are due to expire by 2010. However, FDA has yet to issue any comprehensive guidance on generic biopharmaceuticals. Barring an act of Congress, it will be at least three years before follow-on biologicals are approved in the U.S., most experts agree.
Ironically, some follow-on biologics are already made and sold in the U.S., in fact if not in name. Approved years ago under procedures established for small molecules, human growth hormone, insulin, interleukins and interferons are available, as are versions of therapeutic proteins modified with polyethylene glycol (PEG) residues to improve pharmacokinetics.
Part of the challenge that FDA faces is the sheer complexity of biomolecules. Creating a generic small-molecule drug is straightforward. Even when structural or process route data aren’t available, a competent chemist can usually reverse-engineer chemical drugs in a matter of days. Once a GMP manufacturing process is in place, generics houses apply for Abbreviated New Drug Applications (ANDAs), which means no clinical trials.
Proteins, however, are nowhere near that easy to duplicate. In addition to primary structure (amino acid sequence), biochemists must worry about secondary, tertiary and quaternary structures which may coexist in any combination. Even their primary structures are not straightforward, since they encompass post-translational modifications (PTMs) such as glycosylation, acetylation and several dozen other chemical transformations unique to the cell line in which each protein is manufactured.
Speaking at last fall’s ISPE annual meeting in Scottsdale, Ariz., Dennis Fenton, VP for manufacturing at Amgen (Thousand Oaks, Calif.), observed, “Each manufacturer must make its own cell line, and each manufactured cell line is unique.” Even when biogenerics developers are lucky enough to find that they generate PTMs in similar quantities and familiar patterns, almost any process variable—nutrient composition, culture time, downstream separations—can affect the locations and distributions of PTMs.
Until FDA and Congress act — probably some time this year — manufacturers of generic versions of approved biopharmaceuticals must, like Sandoz, wait in regulatory limbo. Last September, the company sued FDA when the Agency was unable to reach a decision, after two years, on a new drug application for Omnitrope, the Sandoz version of human growth hormone. According to Sandoz, Omnitrope is almost an exact copy of similar HGH products already marketed in the U.S. FDA argued that it had simply not completed its review and is not purposely delaying.
The stakes for follow-on proteins are high. Generic challenges to some of biotech’s biggest blockbusters — such as Amgen’s anemia drugs Epogen and Aranesp, which combine for more than $5 billion in annual sales — will have the same impact as generic small-molecule drugs for branded medicines. A recent study by URCH Publishing (London) estimated that the global market for “biogenerics” could be as high as $5.4 billion, or as low as $0.8 billion, by 2010, depending on how quickly FDA moves to issue guidance. According to DataMonitor (London), “at-risk” biologicals currently enjoy $20 billion in annual sales.
Science is improving, but far from perfect
A session held at the Scottsdale ISPE meeting featured Fenton of Amgen, Suzanne Sensabaugh, senior director for global biogenerics for generic giant Teva Pharmaceuticals (North Wales, Pa.), and Kathleen Clouse, Ph.D., acting director of FDA’s division of monoclonal antibodies in CDER’s Office of Biotechnology Products. The discussion promised to be contentious, but ended with the participants reaching common ground. They agreed that the science for establishing the bioequivalence of follow-on biologicals is improving, as are controls for verifying safety and efficacy, but that they are far from perfect. Bottom line: NDAs for follow-ons must be considered individually.
The ultimate issue is not whether chemical equivalence is possible, but rather what type and degree of chemical equivalence FDA will demand as a first-pass judgement on similarity. As Fenton noted, detailed process characterization will be critical to product quality for biosimilars, but the complexity of biologics and the structural impact of process steps makes characterization difficult. “Is it possible to produce the same product by two different methods?” questioned FDA’s Clouse. “In theory, yes. But it’s very hard.”
Assuming that a follow-on is deemed by some criteria to be “similar” to the originator product, there remains the question of the types of bioassays for assuring regulators that the proteins are behaving similarly. Assuming that a biosimilar arrives at this point in the regulatory process, nobody knows how much clinical testing, if anything, FDA will require. Estimates range from small safety/efficacy studies (ideal) to full-blown Phase III trials.
Good analytical methods, from chromatography to peptide mapping to nuclear magnetic resonance (NMR), are available for characterization, Clouse noted, but these “give average data, and may not be able to prove bioequivalence.” Thus, there will likely be a need for additional clinical trials to prove the safety and efficacy of follow-on products, she said.
Regardless of process, product can equal product, Sensabaugh argued. In other words, products that are slightly different analytically can behave essentially the same in patients, despite the fact that they were manufactured by different companies, through different processes, at different locations. Sensabaugh noted that six structurally identical and equivalent HGH products — from Eli Lilly’s Humatrope to Teva’s Tev-Tropin—are on the market today, having been approved in an earlier era under regulations for small-molecule drugs (see "Product Equals Product," below).
In fact, HGH will probably serve as the classic case study for follow-on protein product approval. FDA considers the six products “similar” chemically and therapeutically equivalent despite their diverse manufacturing pedigrees. The critical regulatory issues for biosimilars, Fenton said, will be ensuring patient safety through expanded human clinical testing, and protecting innovators’ intellectual property.
While FDA expects to offer a guidance document on follow-on biologics some time in 2006, manufacturers should not expect blanket guidance for all biosimilar candidates. “Demonstrating ‘generic’ interchangeability for follow-on proteins may be impossible,” Clouse said. “They will have to be considered on a case-by-case basis.”
Roadblocks
As occurred with small-molecule generics after enactment of the Hatch-Waxman Act of 1984, which essentially established the generics industry, original patent-holders will not meekly relinquish control over franchises they developed over many years. Originator firms will attack biogenerics from every angle: legal, intellectual property, medical and scientific. Through established lifecycle-management strategies, large biotech companies will create versions of their products incorporating chemical modifications (such as PEGylation), or changes in dosage form or delivery vehicle.
At the same time, they will try to raise scientific questions, for example on the similarity — or lack thereof — between their proteins and follow-on products. Typically, these battles will be fought over higher-order structure or post-translational modification patterns, which will necessarily differ for proteins produced through different processes. Although nobody can predict the impact of such differences on safety and efficacy, the shadow of doubt may suffice to cause regulators to ask for clinical trials, which will delay introduction of follow-on proteins for months or years. More than anything, FDA needs to step in with one or more guidances on how to deal with subtle structural discrepancies.
In addition to validating the very notion of generic pharmaceuticals, Hatch-Waxman provided a legal framework for fast-tracking small-molecule generics. Under section 505(j) of the legislation, follow-on companies may apply for Abbreviated New Drug Applications (ANDAs), which uses chemistry rather than clinical testing to demonstrate equivalence. European regulators have recently promulgated a regulation along these lines, but no products have yet been approved under it. In the U.S., at least one large biotech company has objected to FDA’s attempt to create a 505(b)(2) guidance. Omnitrope, whose approval is still in limbo, was submitted under 505(b)(2).
Since precise chemical equivalence is next to impossible to establish for biopharmaceuticals manufactured by different processes, in different locations, by different companies, regulators must define chemical similarity differently for biologics, and while they’re at it, follow-on firms should probably gird their loins for clinical trials. Which raises a philosophic question regarding another unresolved biogenerics issue: If the molecules differ enough chemically to warrant clinical testing, how can creating a follow-on product, even before patent expiration, violate an originator’s intellectual property? Such are the conundrums of this business.
Exquisite science and quality by design
The lack of FDA guidance on biogenerics, and its effect on stakeholders, was evident at “Follow-On Biologics Workshop,” a three-day conference held in New York in December and organized by FDA and the National Institute of Standards and Technology (NIST; Gaithersburg, Md.). The panel consisted almost entirely of members of government and academia, with no significant input from would-be follow-on developers or large biotech companies (although these constituencies were present in the audience).
Three-quarters of the presentations covered some aspect of protein analytics, including optical spectroscopy, NMR, light scattering, ultracentrifugation, mass spectrometry (MS), and chromatography plus bioassays and surface plasmon resonance (SPR). One by one, speakers argued for their particular protein analytic technology, with hardly a mention of clinical relevance. The notable exceptions were treatments of quaternary structure, primarily aggregation. Protein aggregates are almost always immunogenic.
Scientists can spend years testing proteins for similarity to other proteins, and never arrive at an “answer” unless the comparison is anchored in some combination of biochemistry and clinical experience.
Speakers spent a good deal of time on post-translational modifications, secondary and tertiary structure, and protein aggregation. From these presentations, it became apparent that analytical science can now scrutinize and compare the most esoteric of protein properties. What is less than clear is the relevance of these differences on what matters most: safety and efficacy.
Charles Cooney, Ph.D., professor of biochemical engineering at the Massachusetts Institute of Technology (Cambridge, Mass.), opened the session on manufacturing by challenging a long-held tenet of biotechnology that “the process is the product.” According to Cooney, processing primarily affects product quality, which may be maximized through an exercise he called “quality by design.” The idea is simple enough: Know your goal, understand the capabilities of analytics to measure progress towards the goal, determine how unit operations affect protein properties, and plan accordingly. In an abrupt departure from the exquisite science of the previous 15 presentations, Cooney argued for focusing not on every imaginable protein property, but “the ones that are important.”
While recognizing the difficulties and risks involved in follow-on protein approvals, Cooney urged rapid resolution of the regulatory quandary facing biogenerics, and called for establishment of clear ground rules for moving forward. These include a common vocabulary for discussing protein properties, real-time analytics, alignment of process and manufacturing science with clinical needs, reference points for purity, structure and efficacy, a science- and risk-based approach to uncertainty, building on converging analytic and manufacturing technologies and an atmosphere that encourages process improvement.
On the last day of the conference Adrian Bristow (NIST), who discussed reference standards for biogenerics (summary: good ones do not exist), remarked on the apparent disconnect between sophisticated analytics and safety/efficacy, which should be the goal of comparability studies. Commenting on the previous two days of presentations, he said, “We’ve seen some really fantastic science, but I’m not sure we haven’t asked these guys to solve the wrong problems. They’re coming up with increasingly minor modifications in the molecule where the clinical relevance is going to be very difficult to prove or, let’s face it, trivial or non-existent.”
Bristow also mentioned the sophistication of what passes for protein analytics compared with the simplicity of proteomics approaches based on chips that detect picograms of material in a milliliter of serum. “Why can’t they do that in a pharmaceutical preparation?” he asked.
PRODUCT EQUALS PRODUCTThe following is excerpted from an interview with Suzanne M. Sensabaugh, senior director for global biogenerics at Teva Pharmaceuticals. Teva currently distributes biosimilar medicines in over 17 countries, says Sensabaugh. These include Interferon alpha-2b, Granulocyte-Colony Stimulating Factor and Human Growth Hormone. Teva plans to market “biotechnology-derived” drugs in the U.S. once a clear regulatory pathway is introduced. Pharmaceutical Manufacturing: There seems to be agreement that the analytical means we have today are more than sufficient to show comparability of biosimilars to the original product. Would you agree? S.S.: Yes, analytical methods are available today to adequately characterize certain protein products and demonstrate comparability. In fact, this was acknowledged by FDA in 1996 when draft guidance was issued on establishing comparability for biotechnology- derived products. P.M.: What exactly do you mean when you emphasize "Product = Product"? S.S.: Historically, biological products have been complex mixtures with low purity that were difficult to characterize. Because of this limited ability to characterize, a biological product was defined by its manufacturing process. However, as the biotechnology industry gained experience and knowledge, improvements were made to the production and testing methods, and controls, that enable manufacture of high-quality and high-purity product that can be adequately characterized. P.M.: What is the significance of the six human growth hormone (HGH) products already on the market? S.S.: HGH serves as a good case study of how a follow-on protein product can be approved. In the U.S., the six approved HGH products are considered identical in structure and therapeutically equivalent even though they have all been manufactured using different manufacturing processes, including cell lines, in different facilities. An FDA Advisory Committee recommended, and FDA accepted the opinion, that limited clinical trials of 50 to 100 patients followed for six months would be adequate for approval of these products. These products were approved under the FD&C Act, Section 505(b)(1) — the same regulatory mechanism used for new drug applications (NDAs). It is of interest that the E.U., where a legal pathway exists that allows follow-on protein products (termed "Similar Biological Medicinal Products") to be marketed with abbreviated preclinical or clinical studies, considers HGH to be such a product. P.M.: To what degree should current E.U. draft guidelines for biosimilars serve as a model or reference for FDA? S.S.: The E.U. draft guidelines serve as a good starting place for FDA. Although the draft guidelines describe drug development programs that are more conservative than they need be, as the E.U. regulatory authorities gain more experience and confidence with these products, we expect to see the requirements become more in-line with current scientific knowledge. In order to bring affordable follow-on protein products to the globe, it is very important that the major markets harmonize their requirements for drug development. |
