Organizations involved: [just a few of the major early technology developers]

Stanford University – R&D; Tech.

Columbia University – R&D; Tech.

City of Hope National Medical Center – R&D; Tech.; Patent dispute

Genentech, Inc. – R&D; Tech.

Cross ref.: See the Monoclonal Antibodies entry (#300) for information about recombinant antibodies.

Description: The following section includes recombinant proteins (and glycoproteins). These products involve the splicing of exogenous DNA (gene sequences from another organism) into the genome of cells for expression (involving transcription, translation and, often, secretion) and production of a desired (glyco)protein. Recombinant proteins (and glycoproteins) are expressed by these transformed cells. Promoter and other gene sequences are often also inserted to along with the gene for the desired protein to control and attain high levels of protein expression. Transformed cells used for manufacture of recombinant proteins range from simple bacteria, e.g., Escherichia coli (E. coli) to transformed mammalian cells.

Biological.: Polypeptides and proteins can be chemically synthesized in vitro. However, chemical synthesis of proteins is only viable if a peptide/polypeptide/protein is relatively small, e.g., less than about 30 amino acids. Synthesis can be relatively quick, automated, and involves fewer issues concerning handling of complex biological materials and purification. However, larger peptides are prohibitively expensive and difficult to chemically synthesize. Synthesis of the correct amino acid sequence become less accurate as the number of amino acid residues (size) increases. In addition, post-translational processing (e.g., glycosylation, folding, cross-linking) of larger proteins, as occurs in humans and mammalian cells, is not performed by synthetic methods. Recombinant DNA techniques, such as those developed by Drs. Cohen and Boyer (Cohen-Boyer patents) exemplified by 4,237,224 and other U.S. patents, which expired in 1997 assigned to the University of California and Stanford University), can be used to express proteins from exogenous genes in various biological systems.

     Two relatively simple systems often employed for the production of recombinant proteins are bacteria, e.g., E. coli, and yeasts, e.g., Saccharomyces cerevisiae (S. cerevisiae). Bacteria efficiently synthesize and secrete proteins and enzymes, but without glycosylation (attachment of carbohydrate side chains to the protein), as is common in human and other mammals). E. coli may not be suited for the expression of eukaryotic genes, especially if the protein must be glycosylated and terminally processed, e.g. cross-linked, folded. Also, the recombinant protein may be toxic to the bacteria, may form insoluble misfolded proteins in inclusion bodies, or may be degraded by bacterial proteases. Yeast and other fungi produce recombinant glycoproteins that can be secreted into the culture medium, potentially offering simpler (less expensive) purification of the secreted proteins from fermentation broths. One of the major drawbacks to both bacterial and yeast systems is the large initial investment required, e.g., for fermentors/bioreactors, and associated infrastructure. Also, the post-translational processing of the proteins may not be accurate, e.g., protein folding, oxidation state, formation of dimers, etc. Also, proteins are not glycosylated in bacteria and are often hyper-glycosylated in yeast (and yeast glycosylation patterns differ from those of the human/mammalian protein).

     Recombinant proteins expressed by E. coli are generally held within inclusion bodies within the cells, which may complicate (or simplify) purification. Proteins expressed in bacteria such as E. coli, may be located in the cytoplasm, or be secreted through the cell membrane, where they may collect inside the periplasmic space, or may be truly extracellular. However, intracellularly expressed recombinant proteins in E. coli have a tendency to accumulate as insoluble aggregates (inclusion bodies), which must be solubilized and the proteins refolded to recover their native state (usually done in conjunction with chromatography). Purification of material is often immediately required upon harvest of high volumes of liquid fermentation broths, as the desired protein may be unstable in the fermentation broth. Also, contaminants, e.g., bacterial and yeast proteins, from the cultures may co-purify with the product.

     Transformed animal, primarily mammalian, cell cultures are often used for producing recombinant proteins because of their ability to perform glycosylation better mimicking human protein glycosylation patterns, and for their other abilities to process the recombinant protein in a manner similar to that of human cells. However, mammalian cell culture systems and bioreactors are more complex and expensive than bacterial culture systems. Many mammalian cell lines, e.g., hybridomas, VERO and BHK cells, are anchorage-dependent, i.e., the cells prefer to be attached to a surface to grow in culture. Roller bottles, i.e., round cylindrical bottles constantly slowly or periodically turned, were initially used for mammalian cell culture. Subsequently, porous microcarriers (porous beads) suspended in culture medium, have largely replaced roller bottles for large-scale manufacture. Microcarriers can also be used with non-anchorage-dependent cells, with the addition of binding agents, e.g., fibronectin or serum proteins, or charged microcarriers. Hollow fiber bioreactors are also commonly used. Suspension culture methods are generally preferred for non-anchorage-dependent cells such as Chinese hamster ovary (CHO) cells. These cells may also be encapsulated and used for suspension culture.

     In contrast to recombinant protein expression, it is much easier to manufacture secondary metabolites (chemical substances not expressed from an organism’s genome) from microbial systems. manufacture of secondary metabolites, in contrast with recombinant methods, generally involves cultured microorganisms metabolizing relatively small molecules (compared to proteins), often with the addition of inducers and precursors. For example, 30-50 gram/Liter per week of penicillin G can be manufactured by culture of Penicillium species of fungi; 0.5 gram/Liter of rapamycin can be obtained from Streptomyces hygroscopius; and 4 gram/Liter of brefeldin A can be obtained from Penicillium species.

     Actual industrial yields of recombinant protein vary greatly with the methods, materials, conditions, technology, and equipment used. (e.g., see “Bioprocess Tutorial: Fermentation Technologies for Bioproduction,” by J. Davies, et al., Genetic Engineering News, vol. 20, no. 17, Oct. 1, 2000, p. 40). Generally, recombinant mammalian systems can yield about 1-4 grams/Liter of media per 2-3 weeks, while recombinant E. coli yield 1-4 grams/Liter per 1-2 days. Recombinant monoclonal antibody culture in Chinese hamster ovary (CHO) cells generally yields 0.5-1 gram/Liter. Mammalian cell perfusion bioreactor systems yield about 0.3 gram/Liter per day.

     “Current estimates for the production by cell culture of a gram of a relatively simple protein/Mab [monoclonal antibody] run between $200 and $300. For more complex molecules expressed in, e.g., CHO cells, it may reach $1,000 or more per gram,” as reported in “Industry Focus: Bioprocessing,” BioProcess International, June 2003, p. 16-17. With many CHO and other mammalian cell-expressed proteins, particularly monoclonal antibodies, requiring dosages of 100 mg or more (e.g., Remicade use for Crohn's disease involving 5 mg/kg dosage, with a 50 kg patient receiving 250 mg or .25 gram), it is easy to see why monoclonal antibodies and other products requiring relatively high doses can be rather expensive. Also, these products, particularly when successful, require manufacture at world-class scale, making them sensitive to any supply and manufacturing problems (and prone to related shortages).

Tech. transfer: Cohen-Boyer Patents: Fundamental aspects of modern recombinant DNA techniques, e.g., use of restriction enzymes for cutting and splicing nucleic acid sequences, were developed in the mid-1970s by Drs. Cohen and Boyer. Related patents (now expired) were co-assigned to the University of California and Stanford University, e.g., U.S. 4,237,224 issued on Dec. 2, 1980. U.S. 4,237,224 includes claims for the basic use of plasmid or viral vectors to transform or splice exogenous (foreign) DNA into unicellular organisms. Until their expiration, essentially all manufacturers of commercial recombinant proteins were required to obtain a nonexclusive license to these patents. The two universities collected over $200 million from Cohen-Boyer patent licensing. This generally involved a large number of relatively low-priced nonexclusive licensing fees plus royalties on sales. Stanford Univ. has reported a total of 369 companies took licenses, encompassing just about all companies active in commercial recombinant protein R&D and manufacturing in the 1980s through the mid-1990s.

     Stanford Univ. took the lead in licensing the Cohen-Boyer patents, and classed licenses (and royalty base) as involving “Basic genetic product, Process improvement product, Bulk product, and End product.” Royalty rates ranged from 10% for the basic genetic product, e.g., use of genetic constructs, to .5% for the end product. For example, use of a transformed organism for manufacture of insulin is a basic genetic product with a royalty of 10%. Bulk products were products processed further by a manufacturer and not used or consumed by the end user, with royalty ranging from 3% for annual sales volume under $5 million; 2% for sales of $5 to $10 million; and 1% for annual sales over $10 million. End product was a product for use by the ‘ultimate consumer’, such as an insulin injection, with royalties ranging from 1% for annual sales volume under $5 million; .75% for sales of $5 to $10 million; and .5% for annual sales over $10 million. Process improvement product was a material developed for or by a manufacturer to improve an existing process, e.g., a recombinant enzyme to catalyze a reaction, with royalties of 10% of the costs savings or other economic benefit. Since the end of Sept. 1996 (through the life of the relevant patents), the new license end-product royalty rate was a flat 1% based on sales volume.

     Stanford Univ. reports having received significant Cohen-Boyer licensing royalty income from sales of the following recombinant human biopharmaceuticals, presumably all classed as basic genetic product licenses (≥10% royalties), with company date of taking a license: a) Epogen (erythropoietin), Amgen (12/2/80); b) Procrit, Amgen partnered with Ortho/Johnson & Johnson (12/2/80); c) Neupogen (granulocyte-colony stimulating factor), Amgen (12/2/80); d) Activase (tissue plasminogen activator), Genentech (12/2/80); e) Nutropin (human growth hormone), Genentech (12/2/80); f) Pulmozyme, Genentech (12/2/80); g) Humulin (insulin), Genentech partnered with Eli Lilly & Co. (12/2/80); h) Protropin (human growth hormone), Genentech (12/2/80); i) Nutropin AQ, Genentech (12/2/80); j) Roferon A, Hoffmann-La Roche Inc. (12/2/80); and Leukine (granulocyte-macrophage colony stimulating factor), Immunex Corp. (9/1/86). Posilac (recombinant bovine growth hormone) used in cattle, licensed by Monsanto (12/2/80), also provided significant licensing income. Presumably, other recombinant products marketed in the U.S., particularly those manufactured in the U.S., during the life of the patents also involved licensing of the Cohen-Boyer patents. [From Stanford Univ., May 2004].

     Through the life of the patents, the universities received over $200 million in income. At patents’ expiry, each university lost about $12 million in annual income.

     Cotransformation Patents: Another early, broad recombinant DNA patent involved “cotransformation” processing assigned to Columbia University in 1983. In 1979, Dr. Richard Axel and collaborators published a landmark article in Cell. The article described "co-transformation", where a mammalian cell is transformed with the gene of interest (introduced into the cell on one DNA construct), and, at the same time, is transformed with a second gene known as a "marker gene" (introduced on a second DNA construct). The marker gene code for a protein necessary for the cell's survival under certain conditions. The experiments reported by Dr. Axel in 1979 demonstrated that, after co-transformation with two separate DNA constructs, a cell that had successfully incorporated a marker gene was also likely to have incorporated the second gene of interest. On June 1, 1980, Dr. Axel applied for a patent based on this work. U.S. 4,399,216 ultimately issued on February 2, 1985. In his patent, Dr. Axel specifically included a prophetic example stating that his technique could be used to transform CHO cells, e.g., with the beta interferon gene, and a marker known, e.g., dihydrofolate reductase (DHFR), rendering cells resistant to methotrexate.

     U.S. 4,399,216, Axel, et al., August 16, 1983, “Processes for inserting DNA into eucaryotic cells and for producing proteinaceous materials,” assigned to Columbia Univ., filed on Feb. 25, 1980, expired on Aug. 16, 2000. This patent covers cotransforming eucaryotic cells, involving inserting multiple DNA sequences–a sequence for expression of the desired protein along with an unlinked DNA sequence coding for a selectable phenotype not expressed by the host cell and permitting survival or identification of transformed eucaryotic cells. This included cotransformation with thymidine kinase (rendering cells resistant to acyclovir exposure), dihydrofolate reductase (rendering cells resistant to methotrexate), and other selectable phenotypic markers allowing the ready identification and isolation of transformed cells. Two other patents, 4,634,665 and 5,179,017, both originating from 4,399,216 and expiring on the same day, include additional/extended claims, e.g., adding Chinese hamster ovary (CHO) expression. U.S. 6,455,275, issued in 2002, is further discussed below.

     Columbia Univ. collected nearly $370 million in royalties on its original cotransformation patents between 1983 and 2000. Through the 17 year life of the patent (expired in Aug. 2000), the university collected 1% royalty from 30 manufacturers, including royalties related to 12-14 (undisclosed) recombinant proteins. Licensees and the royalties cumulatively paid to Columbia (from available recent court filings; see below) include: Genentech estimating it paid Columbia over $70 million in royalties; Biogen Idec, $35 million; Genzyme, almost $25 million; Wyeth (originally Genetics Inst.), over $70 million; Baxter, over $5 million (for Factor VIII supplied to Wyeth); Ares/Serono, over $6 million. Genentech took a license on Oct. 12, 1987; Amgen, now including Immunex, on Oct. 1, 1991; BASF, now Abbott Bioresearch, on June 1, 1995; Biogen (now Biogen Idec), in 1993; Wyeth (Genetics Institute), in 1990; Ares/Serono S.A., on May 1, 1992; Johnson & Johnson (Ortho), N.A; and Baxter (for its own Factor VIII), in 1997.

     The university initiated public policy and political controversies in early 2000 when it began lobbying Congress to grant it a special 15-month patent extension for 4,399,216, which would have brought the university an extra $70-100 million in royalty income. The university argued that it was unfair that the Hatch-Waxman Act, which allows extension of U.S. pharmaceutical composition-of-matter patents for the amount of time the product was under review by FDA (e.g., from BLA or NDA filing until approval), but which does not allow for comparable extension of process patents. As can be seen from the technology transfer information presented in the monographs below, it is very common for process, rather than composition, patents to provide critical protection for recombinant and other products.

     On Sept. 24, 2002, Columbia Univ. received its fourth, patent, 6,455,275, a divisional of the original 1980 cotransformation patent, with a 17 year life (expiry in 2019). This patent, “DNA construct for producing proteinaceous materials in eucaryotic cells,” by Axel, et al., has claims explicitly covering cotransformation of Chinese hamster ovary (CHO) cell lines (not all eucaryotic cells, as claimed by the 4,399,216). Simply stated, the original patent primarily claimed the process of inserting foreign genes into another cell, while the new patent claims rights to the transformed cells. This patent was quickly challenged in court after Columbia started requesting new royalty payments from relevant recombinant protein manufacturers. For example, Biogen (now Biogen Idec), Genzyme and Abbott jointly filed a suit in federal court in Boston, and Genentech and Amgen jointly filed a federal suit in San Francisco to have the patent declared invalid (as an illegal extension of the original patent). Other companies have also filed suits. The companies allege that Columbia used “bait and switch” and other questionable legal tactics to obtain the new patent that is “substantially the same” as the original. The suit filed in Boston cites 6,455,275 as a “submarine patent,” because its application lay dormant and hidden for so long. In response, Columbia noted that the patent office properly reviewed this patent application after its filing in 1995 and concluded that it was a new invention. With the new and original patents being very similar in content, much legal wrangling will involve interpreting the semantics of claims and supporting documentation. In May 2004, the U.S. patent office granted the Public Patent Foundation (PUBPAT; New York, NY), acting on behalf of companies having filed suits, a Request for Reexamination of 6,455,275, citing “a substantial new question of patentability” regarding every claim of the patent.

     On Nov. 5, 2004, a federal district judge in Boston dismissed most of the companies claims against Columbia University, with both sides claiming victory. During hearings, Columbia Univ. stated it would not assert claims or seek royalties from alleged patent infringers, and the case was dismissed. The companies portrayed this as the university abandoning its patents. However, a patent office reexamination requested by the university was then in progress, and the ruling did not exclude the possibility of the university reasserting infringement at a later date.

     In Aug. 2005, Genzyme Corp. and Biogen Idec Inc. settled their dispute with Columbia Univ. and withdrew from their lawsuit regarding the university allegedly illegaly extending its original patent. The companies reported that their settlement with the university allows them to continue selling affected products. Terms were not disclosed, but Biogen Idec reported it was “very pleased” by the settlement. At the time, Genentech and Amgen were reported to be making progress in settling their lawsuit against the university. Other companies will likely or may already have settled with the university.

     The U.S. Patent and Trademark Office (PTO) has yet to rule regarding the university’s more recent patent; and Genentech, Amgen and other companies are still in dispute with the university.

     City of Hope/Genentech Dispute: Genentech and the City of Hope Medical Center (COH) have been involved in a long-running dispute arising from contract work performed by the City of Hope on basic recombinant expression methods that largely formed the basis for Genentech. Genentech was founded on April 7, 1976, and in May began negotiation of a $300,000, two-year contract with COH for Drs. Riggs and Itakura, COH, to collaborate on development of methods for bacterial expression of insulin and somatostatin, with COH to receive a royalty of 1.5% on “gross sales of Genentech itself and of any and all Genentech subsidiaries and licensees or joint venturers.” Dr. Riggs and Itakua, COH, developed basic expression vector methods and constructs, and received U.S. patents Method and means for microbial polypeptide expression for which they received U.S. patents 5,583,013 and 5,221,619, both entitlted “Method and means for microbial polypeptide expression,” and 4,704,362, “Recombinant cloning vehicle microbial polypeptide expression.”

     These patents were exclusively licensed to Genentech, and were widely used and licensed by Genentech, including for essentially all early Escherichia coli (E. coli)-expressed products. Genentech reported concluded 35 licensing agreements for this technology with 22 companies.

     Under Genentech’s interpretation of its 1976 agreement with COH, Genentech paid COH royalty payments on sales of products made using methods from COH and the sales of its (sub)licensees having explicitly (sub)licensed these patents, with COH receiving ~$300 million over 20 years. However, COH filed a contract dispute suit against Genentech in 1999. In Oct. 2001, the first trial resulted in a hung jury, 7-5 in Genentech’s favor, essentially claiming that the original contract called for royalties on all of product sales by Genentech and its (sub)licensees, irrespective of whether the COH patents were actually used or licensed, or not. Illustrating the importance of contract language, in June 2002, a jury retrial directed Genentech to pay ~$300 million in additional royalties, including royalties on products for which Genentech itself and its (sub)licensees had not used the patents or Genentech received royalties, plus punitive damages of $200 million (total, slightly over $500 million). In Oct. 2004, Genentech lost an appeal of this ruling before the California Court of Appeal. The court cited “substantial evidence of fraud and malice” by Genentech, which failed to pay royalties on its own products and those of its various licensees during the time period covered. This included Genentech allegedly hiding nonpayment of licensing fees due from sales of recombinant hepatitis B virus vaccines and other products using COH inventions over a 15-year period. Genentech appealed to the California Supreme Court. The company had set aside $600 million to cover court-ordered payments, if the company lost its appeal.

     In April 1998, the California Supreme Court unanimously upheld the prior jury's verdict that Genentech had breached a 1976 contract with City of Hope National Medical Center to pay it 2% of all income Genentech received from licensing technology at issue to other companies. The court also strukck down the jury's earlier $200 million punitive damage award and reduced the damages to $300 million in lost royalties, down from the original total damages of $500 million. The court fond that a company that markets another firm's technology s in exchange for royalties has no special obligation to protect the other's interests, apart from its duty to adhere to its contract with the technology source. Without any such obligation, the justices said, punitive damages cannot be awarded for a breach of contract. With Genentech having previously taken a charge to write off the entire prior verdict as a loss in 2002, the new verdict was a large win for the company, paying out only $300 million plus ~$175 million interest (~$475 million total), but not having to pay an additional $200 million in punitive damages plus ~$115 million in interest.

Market: The current (2008/2009) world market for recombinant proteins is estimated to be at least $70 billion, out of total worldwide biopharamceutical of slightly over $100 billion. This includes nearly 80 recombinant protein products for which reasonably good or actual recent (2008 or 2007) sales data are reported in this database. Annual sales for other products included in this section are estimated to be $2-$3 billion annually, plus $5-$8 billion for miscellaneous recombinant proteins, mostly biogenerics (copies/equivalents), manufactured and marketed in other, mostly lesser-developed, countries (e.g., where lack of current patents or their enforcement allows, and not covered by this publication).

Ref.: a) Advances in Large-Scale Biopharmaceutical Manufacturing and Scale-Up Production, American Society for Microbiology Press and the Institute for Science and Technology Management, by Langer, Eric S., Washington, DC, 2008.

b) “Manufacturers Must Cooperate to Compete: The Biomanufacturing Industry Faces a Catch-22,” E. Langer (BioPlan Associates), Contract Pharma, Oct. 2002, p. 66-70.

c) “Biopharmaceutical Benchmarks,” by Walsh, G., Nature Biotechnology, Aug. 2003, p. 865-870 (with fold-out chart).