Cross ref: See the entries for recombinant monoclonal antibodies in the Recombinant DNA Products section, and the entries in this section below for non-recombinant monoclonal antibodies.
Organizations involved: (just a few of the major ones -involved in technology development):
Medical Research Council (U.K.) – R&D
Genentech, Inc. – R&D; Tech.; Patent dispute
Protein Design Labs, Inc. (PDL). – R&D; Tech.; Patent dispute
Description: Currently marketed monoclonal antibody-based biopharmaceuticals are manufactured using both classic/conventional hybridoma and newer recombinant technologies. Monoclonal antibodies expressed by classic murine or other rodent hybridomas, using either in vivo culture (ascites method) or in vitro culture in bioreactors, are, as expected, composed entirely of rodent/murine antibody sequences. Recombinant monoclonal antibodies generally combine human and murine antibody sequences/fragments in a hybrid antibody molecule, forming a chimeric or humanized antibody; or involve fully human sequences/fragments. The structure of antibodies (immunoglobulins; immune globulins) is discussed further in the “Biological” section.
Monoclonal antibody production can be carried out by either using in vitro animal (murine-human hybridoma or other recombinant) cell culture or by original in vivo murine ascites production methods. These methods are further discussed in the “Manufacturing” section below. Ascites production involves culture of the hybridoma cell line (an immortalized antibody-producing cell line) in the peritoneal cavity (gut) of mice. The hybridoma cells are injected into the peritoneal cavity and the mouse tissues provide all the nutrients for cell growth. The proliferating cells secrete monoclonal antibody into the ascitic fluid of the mouse for subsequent extraction. Hybridoma cell culture methods offer a higher level of control and less chance of viral or other contamination, but provide much lower yields (lower titers) than the in vivo ascites method. Most monoclonal antibodies are now manufactured by in vitro animal cell culture. The fermentors/bioreactors and methods used include a wide range of mammalian cell culture hollow fiber and capillary bioreactors, air lift fermentors, stirred tanks, and others. Reactor sizes can vary from desktop bioreactors to 1,000-10,000 liter or larger vessels.
Biological.: Immunoglobulin (antibody) molecules are composed of four polypeptide chains, two identical heavy (H) chains and two identical light (L) chains. Antibodies, or immunoglobulins comprise two heavy chains linked together by disulfide bonds and two light chains, each light chain being linked to a respective heavy chain by disulfide bonds. Both heavy and light chains are divided into variable (V) and constant (C) regions, with each heavy chain having at one end a variable domain followed by a number of constant domains. Each light chain has a variable domain at one end and a constant domain at its other end, the light chain variable domain being aligned with the variable domain of the heavy chain and the light chain constant domain being aligned with the first constant domain of the heavy chain. The constant domains in the light and heavy chains are not involved directly in binding the antibody to the antigen.
Each pair of light and heavy chain’s variable domains forms an antigen binding site. The variable domains of the light and heavy chains have the same general structure and each domain comprises four framework regions, whose sequences are relatively conserved, connected by three hypervariable or complementarity determining regions (CDRs). The CDRs are held in close proximity by the framework regions and, with the CDRs from the other variable domain, contribute to the formation of the antigen binding site. The VH-VL pairs confer specificity for antigen while the constant -region of the heavy chain is responsible for effector functions such as complement fixation and antibody dependent cellular cytotoxicity (ADCC). The variable and constant regions were so named because amino acid sequencing showed that the amino terminal regions of heavy and light chains from different antibodies had different sequences while the carboxy terminal region amino acid sequences were the same within a given isotype (class or subclass). Subsequent analysis of variable region amino acid sequences defined three hypervariable regions (also known as complementarity determining regions or CDRs) in each of the VH and VL regions that form the antigen binding site of the molecule.
The basic immunoglobulin (antibody; Ig; immune glo-bulin) protein structure involves two identical light poly-pep-tide chains of molecular weight approximately 23,000 Daltons (23 kDa), and two identical heavy chains of molecular weight 53,000-70,000 Daltons (52-70 kDa). The four chains are joined by disulfide bonds in a double-chain “Y” configuration. The “branch” or bifurcated portion is designated the Fab region. Heavy chains are classified as gamma, mu, alpha, delta, or epsilon, with some subclasses among them. The nature of the antibody/immunoglobulin heavy chain, with a long constant (not varying among different antibodies) region, determines the “class” of the antibody as IgG, IgM, IgA, IgD, or IgE, with each of these involved in different types of immune responses. The constant region and associated class (IgG, IgM, IgA, IgD, and IgE) determines the effector function of the antibody (type of immune response), including activation of complement or other cellular responses. The variable region determines the epitope (antigen) with which it can react (tightly bind), i.e., the variable portion is binds to specific immunogen(s)/antigen(s). Light chains are classified as either kappa or lambda. Each heavy chain class can be associated with either a kappa or lambda light chain. The light and heavy chains are covalently bonded to each other, and the “tail” portions of the two heavy chains are bonded to each other by covalent disulfide linkages. The amino acid sequence runs from the N-terminal end at the top of the Y to the C-terminal end at the bottom of each chain. The variable region specific for the epitope (antigen) which elicited it, is located at the N-terminal end, and is approximately 100 amino acids in length.
The variable region is divided into complementarity determining regions (CDR1, CDR2 and CDR3) and framework regions (FR1, FR2, and FR3). CDRs 1, 2, and 3 are delineated by amino acid positions 31-35, 50-65, and 95-102 for heavy chains and amino acid positions 24-34, 50-56, and 89-97 for light chains. While these amino acid positions define the boundaries of each CDR, the lengths of the CDRs can vary. The CDRs of the variable region form the antigen binding pocket of the molecule through the interaction between heavy and light chain variable regions, while the framework regions provide the scaffolding on which the antigen binding pocket sits. The constant region is responsible for antibody effector functions, but has little influence on antibody specificity or affinity. The variable region is linked in each chain to the constant (unchanging among different antibodies) region that extends the remaining length of the chain. Linkage occurs through a linking sequence known as the “J” region in the light chain gene, which encodes about 12 amino acids, and as a combination of “D” region and “J” region in the heavy chain gene, which together encode approximately 25 amino acids. The remaining portions of the chain are referred to as constant regions and within a particular antibody class do not vary with the specificity of the antibody (i.e., the antigen eliciting it).
Hybridoma technology – Human B cells can, theoretically, generate more than one thousand billion (one trillion; 1012) different antibodies with specificity for a wide variety of antigens/epitopes. For hybridoma cell-based production of a particular monoclonal antibody, a hybridoma (essentially, a fusion of two cells) cell line is typically derived from a single hybridoma cell expressing a single (mono-clonal) desired antibody. A hybridoma is formed by the fusion of B cells (the type of cells that respond to immunogens and produce antibodies) exposed to a target antigen and expressing a desired antibody with myeloma (type of tumor) cells, usually mouse (murine), in the presence of propy-lene glycol. Such fused hybridoma cells essentially contain the full genotype from both source cells. The hybridoma cells can secrete the immunoglobulin (monoclonal antibody) that was produced by the immunized donor B cell, while the myeloma genome confers “immortality” or the ability to replicate and be stably cultured in vitro. A particular fused cell clone (hybridoma cell) expressing the desired monoclonal antibody is selected, and is further cultured to form a master cell line for long-term storage. The fused hybridoma cells may be cultured in vitro much like other mammalian cells, with the monoclonal antibodies secreted into the culture medium, or the hybridoma may cultured in the peritoneal cavity of mice (ascites method), with the monoclonal antibody secreted into the ascites fluid.
Drs. Kohler and Milstein, U.K. Medical Research Council (MRC), are generally credited with having devised the hybridoma-forming cell fusion techniques that successfully resulted in the formation of the first monoclonal antibody-producing hybridomas [Nature, 256:495-497 (1975); Eur. J. Immunol. 6:511-519 (1976)]. The inventors/MRC did not patent their hybridoma and monoclonal antibody technologies. In the early 1970s, much was known about antibodies and their production by B lymphocytes (B cells), but there were no methods available for culture of B cells and the manufacture of a single, specific antibody (e.g., from a single B cell). By fusing antibody-forming human B cells (spleen-derived lymphocytes; splenocytes) with murine myeloma cells (reproducing malignant cells of bone marrow primary tumors), Kohler and Milstein created hybridoma cell lines, arising from a single fused cell hybrid (called a hybridoma or clone) with certain characteristics of both the source lymphocyte and myeloma cell lines. Each hybridoma secretes a single type of immuno-globulin specific to the antigen to which the source B-cell/splenocyte was immunized. The hybridoma cells retain the capacity of the source myeloma cells for indefinite cell division, allowing the cells to be readily cultured.
This combination of features offered considerable advantages over the prior method of polyclonal antibody production involving vaccination of animals and obtaining antisera from animals’ plasma or serum. These mixtures of diverse (poly-clonal) antibodies are variable, having diverse immunoglobulin structures and binding specificities. Monoclonal antibodies are immunoglobulins of a single type, ultimately from a single immortalized B-cell, and are much more easily characterized, while polyclonal antibodies have diverse structures and antigen specificities, with these difficult to ever fully characterize. The single type of immunoglobulin secreted by a hybridoma is specific to one and only one antigenic determinant, or epitope, on an antigen.
Fusion procedures, e.g., using propylene glycol, usually produce viable hybrid(oma) cells at very low frequency, about 1 x 10-6 to 1 x 10-8. Because of this, it is essential to have a means to select fused cell hybrids from the remaining unfused cells, particularly unfused myeloma cells. A means of detecting the desired antibody-producing hybridomas among the other hybridomas is also necessary. To facilitate screening of the antibodies secreted by the hybridomas and to prevent individual hybrids from overgrowing others, the mixture of fused myeloma and B lymphocytes is generally serially diluted in medium and cultured in multiple wells of microtiter plates. Fused cells are cultured in selective media, e.g., HAT medium containing hypoxanthine, aminopterin and thymidine, which permit the proliferation of hybrid cells and prevent growth of unfused myeloma cells, which normally would continue to divide indefinitely. The source myeloma cells used are generally mutants lacking the gene for hypoxanthine phosphoribosyl transferase (HPRT) and, thus, cannot grow in the HAT medium. In the successfully fused hybridoma cells, the human B lymphocyte supplies genetic information for production of the HPRT enzyme, so only successfully fused hybridoma cells can proliferate. Since B lymphocytes have a limited life span in culture (about two weeks), the only cells left proliferating in HAT media are hybridomas formed from the myeloma and B cells. After two to three weeks, the hybridomas become visible microscopically and the supernatant fluid of the individual wells is assayed for specific monoclonal antibody, e.g., using radioimmunoassays, enzyme immunoassays, cytotoxicity assays, plaque assays, dot immunobinding assays, and the like. Once a desired hybridoma clone secreting a desired antibody is identified, the hybridoma is further cultured, and cells are preserved, e.g., by cryopreservation for a master cell line.
Recombinant Monoclonal Antibodies: Monoclonal anti-bodies can also be developed and manufactured using more modern and controlled recombinant methods. The gene sequences for expression of a single antibody/immunoglobulin molecule or fragment may be inserted for expression into another type of cell, e.g., bacteria, mammalian, or other host cells. Using recombinant methods, monoclonal antibodies may be “chimeric” or composed of portions of immunoglobulin molecules from two different species. Most chimeric recombinant monoclonal antibodies are human-murine, involving the swapping of a human variable region, including the complementarity determining region (CDR; which determines epitope specificity) with a murine (mouse) antibody variable region, e.g., derived from a murine monoclonal antibody. “Humanized” recombinant monoclonal antibodies are composed of a human fragment/region with only a minimal antigen-binding CDR regions derived swapping in an antibody of another, usually rodent, species. “Fully human” recombinant monoclonal antibodies are constructed fully from human antigen-specific variable regions (e.g., isolated derived using phage display technology) and human constant regions. Distinctions between chimeric and humanized monoclonal antibodies are not fully clear, with different terminology having been used by different organizations with a stake in licensing of their related patents. Some marketed products involve only recombinant monoclonal antibody fragments, e.g., Fab fragments, which may be produced directly by recombinant methods or derived enzymatically by catalysis of cultured hybridoma monoclonal antibodies.
Monoclonal antibodies may be generated from a non-human source, usually murine (mouse), e.g., a hybridoma developed from mice immunized with a target antigen. When administered to patients, such fully murine antibodies are often recognized as foreign by the patient’s immune system, resulting in an immune rejection-type response, e.g., human anti-mouse antibodies (HAMA) for murine antibodies, which may both hamper the action of the antibody and produce serious side effects. Recombinant chimeric antibody design and humanization seek to avoid HAMA responses by creating antibodies which retain the murine-originating antigen specificity (variable or CDR regions), but in which a large proportion or essentially all of the non-antigen-binding parts (constant or framework regions) of the murine protein are replaced with human-derived sequences. Chimeric monoclonal antibodies generally retain variable domains of the original, e.g., murine, antibody’s heavy and light chains, and may result in immune response (HAMA) in patients. In contrast, humanized antibodies have a higher proportion of human-derived sequences, sometimes including largely human-derived variable domain sequences, and are generally recognized as self, i.e., not responded to as foreign or immunogenic.
The humanization of antibodies is made possible by retaining the minimal non-human gene sequence required to retain the original murine/rodent antibody’s binding properties. Within each variable domain are three stretches of amino acids called the complementarity determining regions (CDRs), which are responsible for the antibody’s substrate or epitope binding specificity. The CDRs lie in a conserved protein framework which holds them in position. All parts of a murine or certain other non-human antibodies except for the CDRs can be replaced by equivalent human sequences to generate a humanized antibody. However, such antibodies often have greatly reduced substrate binding affinity, sometimes to the point of non-utility. Retention of further specific amino acid residues from the variable domain framework of the original, e.g., murine, antibody is often required to create humanized antibodies which retain full substrate binding affinity.
Terminology applied to recombinant antibodies can be complex, including in patents where definitions can vary, depending on the interest/needs of the applicant. The term “CDR-grafted antibody molecule” refers to an antibody molecule where the heavy and/or light chain contains one or more complementarity determining regions (CDRs) from a donor antibody grafted into a heavy and/or light chain variable region framework of an acceptor antibody. A “chimeric antibody” is an antibody having unmodified non-human variable domains fused to the constant regions of the light and heavy chains of a human antibody. A humanized antibody is an antibody that includes not only the fused variable domains (usually murine) to human constant regions, but also altered selected variable domain framework residues that more closely match the most related available human variable template sequence. A “CDR grafted antibody” is an antibody having non-human CDR domains from an antibody grafted into the most closely related human antibody sequence available so that only the CDR domains are non-human in origin. CDR grafted antibodies have reduced immunogenicity and a reduced human-anti-mouse (HAMA) immune response. Terminology related to “humanization” can similarly be complex.
Marketed Products: Monoclonal antibody (Mab)-based products having received FDA approval (but not necessarily currently marketed) may be roughly classified as:
a) in vivo (ascites) cultured murine monoclonal antibodies (classic ascites method), e.g., CEA Mab (arcitumomab; CEA-Scan) from Immu-no-medics, Inc.; and TAG-72 Mab radioimmune conjugate -(satu-momab) from Cytogen Corp. (EUSA Pharma); and CD3 Mab (Orthoclone OKT-3) from Ortho/J&J. Note, few of these are currently marketed.
b) in vitro cultured murine hybridoma monoclonal antibodies, e.g., NR-LU-10 Mab radio-conj. (nofetumomab; Verluma) from NeoRx Corp.; Prostate Mab radioconj. (capromab pendetide; ProstaScint) from Cytogen Corp. (EUSA Pharma); and the Stem Cell Concentration System (Isolex 300 & 300i) from Nexell using a CD34 monoclonal antibody;
c) in vitro cultured recombinant chimeric (murine-human) monoclonal antibodies, e.g., Platelet Mab, rDNA (Reopro) from Centocor/Lilly; CD20 Mab, rDNA (rituximab; Rituxan) from Biogen Idec/Genentech; TNF Mab, rDNA (infliximab; Remicade) from Centocor/J&J; and IL-2R Mab, rDNA/Novartis (basi-liximab; Simulect) from Novartis.
d) in vitro cultured recombinant humanized monoclonal antibodies, e.g., RSV Mab, rDNA (palivizumab; Synagis) from MedImmune; IL-2R Mab, rDNA/Roche (dacliximab; Zenapax) from Protein Design Labs/Roche; HER2 Receptor Mab, rDNA (trastuzumab; Herceptin) from Genentech/Roche; and Gemtuzumab Ozogamicin (Mylotarg), an immunotoxin (antibody-toxin conjugate) involving a humanized monoclonal antibody, from Wyeth; and
e) in vitro cultured recombinant fully human monoclonal antibodies, e.g., adalimumab (Humira; D2E7) from Abbott Labs.; RSV Mab2, rDNA/MEDI-524 (NuMax), the next-generation version of Synagis from MedImmune; and Vectibix (panitumumab) from Amgen.
Nomenclature: The U.S. Adopted Names Council (USAN; affiliated with the U.S. Pharmacopeia or USP) has established guidelines for the systematic generation of compendial or generic names for monoclonal antibodies and monoclonal antibody fragments. Four or more syllables are concatenated as a linear notation to describe the product and provide a unique and unambiguous name, e.g., rituximab or satumomab.
All monoclonal antibody names end with “mab.” Ending syllables are preceded by one or two characters designating the source organism (species) for the antibody (antigen-specific portion). These include: “a” for rat, “e” for hamster, “i” for primate/monkeys, “o” for murine (mouse), “u” for human, and “xi” for chimeric (multi-species mixed source) antibodies. An additional syllable, preceding the species designation and termed an “infix,” designates the target of the antibody activity. Examples include: “bac” for bacteria, “ci(r)” for cardiovascular, “col” for colon tumors, “gov” for genital (gonad, ovary) tumors, “lim” for immune modulators, “mar” for mammary (breast) tumors, “pr(o)” for prostate tumors, “tum” for other, miscellaneous tumors, and “vir” for viral. The starting syllable is a USAN-assigned unique and compatible syllable to form a complete compendial monoclonal antibody name.
For monoclonal antibodies bound to linker or chelator molecules, such as those used to link radioisotopes to monoclonal antibodies, another separate word is added after the monoclonal antibody descriptor, e.g., pendetide, to designate the linker-chelator group.
Other molecules, besides radioisotopes, may be linked to monoclonal antibodies or fragments, whether by chemical conjugation or by recombinant expression as a fusion protein. Immunotoxins are formed from conjugation or linkage of an anti-body or functional fragment with a toxin, usually a cytotoxin, to kill antibody-targeted cells. For example, gemtuzumab ozo-gamicin (Mylotarg) is an immunotoxin formed from a recombinant humanized monoclonal antibody with specificity for CD33 antigens on T-cells chemically conjugated to cali-cheamicin (a cytotoxin).
Manufacture: Monoclonal antibody manufacturing methods and underlying technologies are also discussed in the sections above and the Tech. transfer section below.
Antibody titers produced by in vitro hybridoma culture vary with the equipment and methods used. In vitro culture yields are relatively low compared to in vivo ascites production. In vitro culture may provide up to 100 grams of product per 1,000 Liters (0.1 gram or 100 mg/L). Ascitic fluids generally contain higher antibody titers, e.g., 15 mg/mL or 15 grams/Liter, than result from cell culture of the same hybridoma cells. A single mouse generally provides about 5 mL of ascitic fluid, e.g., providing 75 mg of monoclonal antibody. This ascites method provides higher yields, but overall is generally more expensive than cell culture. There is also considerable potential for the presence of murine contaminants, including murine retroviruses, and the presence of a wide variety of murine proteins generally makes downstream purification and other processing more complicated and expensive than cell culture methods. Ethical, regulatory, political, and public relations issues involving the use of laboratory animals are also concerns in the use of the ascites method, and have contributed to its essential abandonment for manufacture of more recent monoclonal antibody-based products.
After culture, whether by in vitro cell culture or the in vivo ascites method, the monoclonal antibodies are concentrated and purified from the culture medium or ascitic fluid. Generally, the culture medium/ascitic fluid is removed and concentrated, e.g., by centrifugation and/or filtration, followed by ultrafiltration and several chromatography steps. These separation and purification steps remove extraneous cell debris, media nutrients and fluid, other proteins, nucleic acids, endotoxins, etc. Various types of chromatography may be used to separate monoclonal antibodies from other materials based on molecular size/shape, electrostatic charge distribution, solubility, and other physical/chemical characteristics. Monoclonal antibody purification often includes affinity chromatography, e.g., with matrix-bound Protein A, which has an affinity for antibodies/immune globulin; cation and/or anion exchange chromatography; and often a final step of gel filtration chromatography. Affinity chromatography using Protein A (derived from Staphylococcus aureus bacteria), which selectively binds the Fc portion of most antibodies, has been used for antibody manufacture for over 25 years, and is the classic or standard method for antibody purification. However, use of Protein A is not without problems, e.g., Protein A fragments may leach into the product, and may need to be removed by further separation steps.
Commercial recombinant monoclonal antibody manufacture may be revolutionized in coming years, with the antibodies being expressed in vivo by transgenic animals, e.g., in the milk of transgenic mammals, and by expression in recombinant plants, potentially providing orally administrable monoclonal antibodies.
Status: Note, as of 2006, very few non-recombinant monoclonal antibody-based products remain on the market. Between the cost of their manufacture (apparently, more expensive than recombinant products), low sales (with radioimaging products losing market to PET and other newer and improved imaging methods), concerns about safety (particularly, for murine monoclonal antibodies), and general technological obsolescence, development, manufacturing and marketing of many of these products has been halted or stalled.
From an FDA perspective, “a monoclonal antibody is a clonal product defined as any intact antibody, antibody fragment, conjugate, fusion protein, or bispecific antibody that contains a VH-VL pair where the CDRs form the antigen binding site.” Antibody fragments or fusion proteins containing only the constant region domains (framework) are not considered monoclonal anti-bodies.
Tech. transfer (non-rDNA): The basic hybridoma (B-cell/myeloma cell fusion) technology invented by Kohler and Milstein, U.K. Medical Research Council (MRC), was published and not patented at the time of its discovery. In hindsight, the MRC missed an incredible opportunity to obtain considerable royalty income. Ascites production technology was also apparently not patented (and/or has long expired).
Tech. transfer (rDNA): Many aspects of the patents, licensing, and patent disputes concerning recombinant chimeric, humanized and other genetically engineered monoclonal antibodies are very complex, and some aspects remain unresolved. Some of the issued patents concerning recombinant monoclonal antibodies involve commonalities, con-flicting terminology, overlapping claims, etc. As with other basic biotechnologies, the ambiguities and legal uncertainties surrounding these patents often have been (are and will be) resolved by patent infringement suits, interference/priority proceedings, and/or cross-licensing. Some disputes have involved protracted, sometimes lasting over a decade, patent infringement and/or priority lawsuits. A few of the more important disputes have only recently been resolved. Currently Proteins Design Laboratories (PDL) and Genentech, Inc. appear to hold dominant positions in U.S. patents concerning recombinant chimeric/humanized monoclonal antibodies. However, as further discussed below, the situation is complex, evolving, and involves considerable uncertainties.
Some basic technologies for recombinant humanization of mo-no-clonal antibodies are disclosed in:
a) Queen, et al. (e.g., U.S. 5,530,101, Queen, “Improved Humanized; 5,585,089; 5,693,761; 5,693,762; and also EP 0451216B10), assigned to Protein Design Labs. (PDL);
b) Winter, et al., U.S. 5,225,539, originally assigned to the U.K. Medical Research Council (MRC), now assigned to PDL;
c) Boss, et al., U.S. 4,816,397 (later revoked, see below); and Adair, et al., e.g., U.S. 5,859,205 and WO 91/09967, assigned to Celltech Group [Note, Celltech’s technology is now essentially assigned to Genentech (see below); Boss is also an inventor of GS expression technology now held by Lonza; Celltech was acquired by UCB Bioproducts in May 2004]; and
d) Cabilly, et. al., U.S. 6,113,415 (New Cabilly patent), a continuation of 4,816,567, both now assigned to Genentech, Inc.
In Dec. 2010, the U.S. Patent and Trademark Office (the PTO) terminated its interference proceeding between certain PDL's Queen et al., U.S. Patent No. 5,585,089, (the '089 Patent) and certain pending claims of Adair et al., U.S. Application No. 08/846,658, (the '658 Application), assinged to UCB Pharma S.A., in favor of PDL '089 Patent. The PTO held that the involved claims in the '658 Application, were not patentable. See the Tech. transfer (rDNA) section of the Monoclonal Antibodies entry (#300) for further information. .
Other recombinant monoclonal antibody technology has been patented and licensed by Xoma. Note, as discussed below, the Boss and Cabilly U.S. patents have been combined into Cabilly II held by Genentech.
The situation concerning licensing of recombinant monoclonal antibody technologies is complex, and those involved are not very helpful in trying to clear up what covers what, who owns what and who has licensed what for what products. Several involved companies have cross-licensed their patents, further complicating determination of who owns and who has licensed what patents/technologies. Protein Design Labs., Celltech, and Genentech. have each granted exclusive and nonexclusive licenses for their recombinant monoclonal antibody technology to a number of companies. Various companies have developed humanized and fully human antibody technologies, often to get around existing patents, e.g., Medarex and Abgenix, which express human monoclonal antibodies in mice, and Cam-bridge Antibody Technology (CAT), Dyax, and Mor-phosys, which use phages to generate libraries of recombinant human antibodies for screening. However, among these other companies (other than Genentech, Celltech and PDL), only CAT has a product in the market involving its technology.
In recent years, additional U.S. patents have issued which further complicate the situtation concerning recombinant monoclonal antibody design and manufacture. U.S. 6,548,640, “Altered antibodies,” Winter, G.P., April 15, 2003, resulting from research at the Medical Research Council (MRC; U.K.), and assigned to BTG International Ltd. (which handles licensing for the MRC), is a continuation of an application with a priority filing date of May 3, 1988. This patent “relates to altered antibodies that have a heavy or light chain variable domain in which the framework regions differ from the framework regions naturally associated with the complementarity determining regions of the variable domain and in which the framework regions are derived from a source of framework regions that differs from the framework regions naturally associated with the complementarity determining regions of the variable regions,” with claims that may cover an unknown proportion of marketed recombinant antibodies and those in development. U.S. 6,632,927, Humanized antibodies, Adair, et al., Oct. 14, 2003, assigned to Celltech Therapeutics Ltd. (Celltech Biologics), has presumably been packaged into the Celltech/Boss portfolio transferred to Genentech.
U.S. patent 5,598,435, assigned to Xoma Corp., broadly describes aspects of recombinant immunoglobulins. The exemplary (no. 1) claim states, “A polynucleotide molecule encoding a secretable immunoglobulin chain, said polynucleotide molecule comprising DNA encoding a prokaryotic secretion signal peptide directly linked to DNA encoding a member of the group consisting of an immunoglobulin light chain, an immunoglobulin light chain fragment, and an immunoglobulin heavy chain fragment, wherein said fragment comprises the entire variable region of said immunoglobulin chain.” Xoma also has other recombinant mono-clonal antibody technologies and patents available for licensing.
PDL Biopharma/Protein Design Labs. tech. – Protein Design Labs.’ (PDL; now PDL Biopharma) patents largely cover the grafting of variable, complementarity determining regions (CDR; antigen-specific region) of murine antibodies into the constant or framework regions of human monoclonal antibodies, i.e., antibody humanization. The resulting antibodies are generally composed of about 90-95% human sequences, and are much less immunogenic than murine antibodies or lack immunogenicity in humans, unlike murine antibodies. In order to maintain potent binding affinity after CDR grafting, PDL developed methods for identifying and modifying critical amino acids affecting antibody affinity and specificity.
U.S. 5,530,101, Queen et al. (Protein Design Labs.) and related U.S. patents provide criteria for designing “humanized” immunoglobulins. Four basic criteria (rules) for design of humanized monoclonal antibodies are described. WO 90/07861 describes the preparation of a single CDR-grafted humanized antibody having specificity for the p55 Tac protein of the IL-2 receptor – dacliximab (Zenapax; see related entry). The four basic monoclonal antibody design criteria established by Queen, et al., were used to design this humanized antibody, with the variable region frameworks of the human antibody EU being used as acceptor (framework), the donor CDRs as defined by Kabat et al., with mouse donor residues used in place of the human acceptor residues at positions 27, 30, 48, 66, 67, 89, 91, 94, 103, 104, 105 and 107 in the heavy chain and at positions 48, 60 and 63 in the light chain of the variable region. This humanized monoclonal antibody has an affinity for p55 of 3 x 109 M-1, about one-third of that of the donor murine monoclonal antibody, along with significant reduction/elimination of human-antimouse antibody (HAMA) rejection-type adverse reactions, i.e., reduced immunogenicity.
PDL humanization patent royalties were $52.7 million in 2003, and $40.4 million in 2002 (and these have surely increased greatly in recent years including with new products entering the market). Major sources were from sales of Synagis by MedImmune, Inc.; Herceptin and Xolair by Genentech, Inc.; Mylotarg by Wyeth; and Zenapax by Hoffmann-La Roche. PDL also now receive royalties on sales of Raptiva and Avastin by Genentech and other products.
The two main PDL patents, including its first patent, EP 0 451216, issued by the European Patent Office (EPO) for antibody humanization methods are still being challenged. The EPO, but not the U.S. patent office, allows for an opposition period in which other parties may submit arguments as to why a patent was incorrectly granted and should be withdrawn or limited. A total of 18 notices of opposition to PDL’s European patents were filed during this period. The preliminary review issued by the EPO raised significant questions regarding the validity of PDL’s European patents. At an oral hearing in March 2000, the Opposition Division (OD) of the EPO decided to revoke the broad claims in PDL’s first European patent based on impermissible addition of subject matter after the filing of the original application. As expected, PDL appealed this decision. In Nov. 2003, the EPO upheld PDL’s appeals concerning EP 0451216, ruling that certain of the patent’s rejected broad claims would be reconsidered by the Opposition Division, EPO. The claims concern the production of humanized antibody light chains that contain amino acid substitutions made under PDL’s antibody humanization technology. PDL’s patents remain in force during the appeals process. In Feb. 2005, the EPO Opposition Division (OD) revoked the claims of PDL’s second European antibody humanization patent, EP 0 682 040. The claims involve production of humanized antibody light chains that contain amino acid substitutions using PDL’s antibody humanization technology. The OD based its decision on formal issues and did not consider substantive issues of patentability. PDL is appealing this to the Technical Board of Appeal, EPO. The appeal suspends the decision of the OD during the appeal process, which is likely to take several years. Opposition filings and proceedings regarding PDL’s recombinant monoclonal antibody humanization technology are also taking place in Japan.
PDL and Genentech were involved in an extended dispute over PDL’s recombinant antibody humanization patents. Genentech’s holds both the original and New Cabilly patents (discussed below), while PDL holds its broad Queen and Winter patents. The companies originally settled their disputes on Sept 25, 1998, with Genentech receiving options to nonexclusively license PDL technology for its own recombinant monoclonal antibody products, and PDL received options to nonexclusively license Genentech patents and patent applications covering the expression of recombinant antibodies and certain chimeric antibodies. Each party originally had the ability to select up to six antibodies to be covered in the Agreement. The agreement would expire in Sept. 2003. Genentech exercised one of its rights with respect to Herceptin (trastuzumab) in late 1998. However, PDL did not agree with Genentech’s interpretation that it has not needed to take a license for several of its other current and then upcoming products.
On Dec. 22, 2003, after expiration of their original cross-licensing agreement, Genentech and PDL “conclusively” resolved their disputes concerning PDL’s recombinant humanized monoclonal antibody design/construction patents (and licensing of three humanized antibodies from Genentech). This dispute had become particularly heated after Genentech claimed (after its June 2003 FDA approval), that its marketed product, omalizumab (Xolair), was not covered by PDL patents and the companies’ prior cross-licensing option agreement. This dispute enlarged to include efalizumab (Raptiva) and bevacizumab (Avastin), also manufactured and marketed by Genentech, with Genentech refusing the pay PDL royalties for either of these products. Under the new (Dec. 2003) agreement, PDL would receive licensing fees of $1 million for each the three disputed products (including Avastin after its approval); and PDL would receive royalties on Genentech’s sales of the three products, including for 2003 sales of Xolair and Raptiva. In exchange, PDL agreed to royalty rate reductions for certain high levels of aggregate sales by Genentech products. PDL also obtained additional rights for nonexclusive, royalty-bearing licenses under certain of Genentech’s antibody patents.
Genentech/Celltech tech. (Cabilly/Boss) and recent disputes – Genentech received over $133 million in royalty payments from its Cabilly-Boss patent in 2007.
Patent disputes between Celltech Group (acquired by UCB Bioproducts in May 2004) and Genentech over recombinant monoclonal antibody design/construction technologies began in 1983. Celltech Group filed its original U.K. application for methods of designing/constructing recombinant antibodies and antibody fragments together with vectors and host cells useful in these processes on March 25, 1983. Shortly thereafter, Celltech filed its “Boss” application in the U.S. On April 8, 1983, about two weeks after Celltech’s filing of its priority U.K. application, Genentech filed its “Cabilly” application for similar technology in the U.S.
Celltech’s filings, eventually granted as U.S. 4,816,397 (“Boss” patent), “Multichain polypeptides or proteins and processes for their production”, Boss, M.A., et al., and European Patent Application 120694, describe expression in E. coli of chimeric antibodies (but with few specific details). The Boss patents describe recombinant expression of monoclonal antibodies by cells transformed with nucleic acid sequences encoding the desired immunoglobulin chains. Granted Boss patents include: U.S. 4,816,397, expiry 2006 (later revoked); European patent: 0 120 694, expiry 2004; Japan: 2,594,900, expiry 2004; and UK: 2 137 631, expiry 2004.
The exemplary claim of 4,816,397 (Boss) states, “A process for producing an Ig molecule or an immunologically functional Ig fragment comprising at least the variable domains of the Ig heavy and light chains, in a single host cell, comprising the steps of: (i) transforming said single host cell with a first DNA sequence encoding at least the variable domain of the Ig heavy chain and a second DNA sequence encoding at least the variable domain of the Ig light chain, and (ii) independently expressing said first DNA sequence and said second DNA sequence so that said Ig heavy and light chains are produced as separate molecules in said transformed single host cell.” The patent’s coverage included, according to Celltech, “Any antibody, or antibody fragment, produced from cells into which the separate sequences for the antibody’s heavy and light chains have been introduced by transformation. The antibody producing cell may have been transformed with either a single or separate vectors. These cells may be: eukaryotic hosts or cell lines, e.g., NS0 cells, CHO cells, transgenic animals, yeast cells, or bacterial cells. The antibodies produced may be: whole antibodies, or Fab’ fragments, or di-Fab’ (F(ab’)2) fragments, or any antibody fragment incorporating both a heavy chain variable domain and a light chain variable region, or chi-me-ric antibodies in which the constant domains of either or both heavy and light chains have been derived from a different source to the variable domains, or humanized anti-bodies...[truncated]”
Celltech also received the “Adair” patents covering aspects of the design and engineering of recombinant humanized antibodies. Celltech’s Adair patents include U.S. 5,859,205, expiry date 2016, and European 0 460 167, expiry date 2010. The “Adair” patents claim a series of structural features necessary in order for most humanized antibodies to obtain full pharmacological activity.
Genentech’s original “Cabilly” patent application, granted as U.S. patent 4,816,567, “Recombinant immunoglobulin preparations,” Cabilly, et al., was issued on March 28, 1989, the same day as Celltech’s Boss patent. Claim no. 1 of this patent is, “A method comprising (a) preparing a DNA sequence encoding a chimeric immunoglobulin heavy or light chain having specificity for a particular known antigen wherein a constant region is homologous to the corresponding constant region of an antibody of a first mammalian species and a variable region thereof is homologous to the variable region of an antibody derived from a second, different mammalian species; (b) inserting the sequence into a replicable expression vector operably linked to a suitable promoter compatible with a host cell; (c) transforming the host cell with the vector of (b); (d) culturing the host cell; and (e) recovering the chimeric heavy or light chain from the host cell culture.” Note, with both of Celltech’s Boss and Genentech’s Cabilly patents issued on the same day, both were originally set to expire on March 28, 2006.
Genentech subsequently exactly copied word-for-word all of Celltech’s Boss patent claims into a previously-filed continuation filing of its Cabilly patent (eventually granted and termed the “New Cabilly Patent”). This resulted in the U.S. patent office declaring an interference between the Celltech Boss patent and Genen-tech’s continuation filing (with the first-to-invent the technology to receive patent protection). In Feb. 1991, Genentech initiated an interference with the patent office to determine which party was the first to invent (rather than file). In Aug. 1998, a panel of patent office judges ruled in favor of Celltech, finding that Genentech had deficient laboratory notebooks and corroborating evidence, i.e., that Celltech’s Boss patent was valid and Genentech’s newer Cabilly application (incorporating all of the Boss patent claims) was invalid.
Genentech filed a suit in U.S. District Court appealing the granting of Celltech’s Boss patent. In March 2001, the court ruled that Genentech, not Celltech, had priority (was the first to invent); vacated the 1998 patent office decision in favor of Celltech/Boss; withdrew Celltech’s Boss U.S. patent; and cleared the path for Genentech to eventually receive its continuation (later granted as 6,331,415, the “New Cabilly” patent) of its original Cabilly patent.
On Dec. 18, 2001, as a result of earlier interference proceeding, the patent office revoked Celltech’s Boss patent, and granted Genentech (and coassigned to the City of Hope National Medical Center) its “New Cabilly Patent,” (Cabilly II), 6,331,415, entitled “ Methods of producing immunoglobulins, vectors and transformed host cells for use therein,” with an expiration date of Dec. 18, 2018. This essentially combined both Genentech’s Cabilly and Celltech’s Boss original patents’ claims regarding methods for design/construction of recombinant humanized antibodies. This new patent also extended the expiration date for essentially the same technology by over a decade.
Also on Dec. 18, 2001, as the result of mediation ordered by the U.S. District Court, Genentech and Celltech Group settled their disputes concerning recombinant humanized antibodies, including the New Cabilly Patent. Both companies cross-licensed their technologies and Genentech took the lead in licensing the combined technologies. Genentech would also compensate Celltech for any lost income it would have received from licensing of its Boss patent or its own sales of products until the patent’s expiration in 2006. For Celltech’s prior licensees, e.g., MedImmune, this effectively meant over a decade extension of their need to pay royalties for the same technology (repackaged into U.S. 6,331,415, the New Cabilly patent), into 2018! Genentech now holds patents broadly covering methods for preparing recombinant chimeric antibodies in diverse expression systems, including co-expression of immunoglobulin heavy and light chains and their refolding into immunologically active forms. Many companies are developing recombinant monoclonal antibodies, and many of these companies had presumed that U.S. patent protection Genentech and Celltech technologies would expire in 2006 (now extended to 2018)/
MedImmune, Inc. responded to the Genentech-Celltech cross-licensing agreement by filing suits in federal court alleging antitrust violations between Genentech and Celltech, and challenging 6,331,415. MedImmune alleged that Genentech and Celltech conspired to gain a monopoly, with Celltech agreeing to abandon defense of its 2006-expiring Boss patent in exchange for continuing (through 2018) licensing revenue from Genentech’s licensing of its New Cabilly patent and access (through cross-licensing) to the New Cabilly patent. MedImmune asserted that the Genentech-Celltech agreement resulted in 29 years of patent protection for the same technology.
In May 2004, the U.S. District Court dismissed MedImmune’s suit, with the Genentech/Celltech patents and agreements remaining intact. The judge found in favor of Genentech, because a U.S. District Court judge had previously approved the Genentech-Celltech agreement. See the RSV Mab, rDNA (Synagis) entry.
Genentech is estimated to have received about $220 million in Cabilly licensing royalties in 2005 (according to UBC analysts), with MedImmune paying about $30 million in royalties related to sales of Synagis. Some have estimated Genentech’s royalties to be as high as $300 million/year. The most significant Cabilly-related royalties on U.S. sales are paid to Genentech by MedImmune for Synagis; Abbott Labs. for Humira; Centocor/Johnson & Johnson for Remicade; and ImClone Inc. for Erbitux. However, the largest portion, probably the majority, of New Cabilly licensing fees received by Genentech come from Hoffmann-La Roche (the majority owner of Genentech) for its ext-U.S. sales of Herceptin, Rituxan, and Avastin. Genentech refuses to disclose how many companies pay royalties for licensing of New Cabilly, but in Feb. 2007 reported that Abbott Laboratories, Centocor Inc. and Imclone Systems Inc. paid Genentech a combined $105 million in royalties related to the patent in 2006. In Jan. 2007, Genentech reported that revenue from the Cabilly patent license with MedImmune totaled 6 cents a share in 2006 and would increase to 7 cents in 2007. With the New Cabilly patent potentially around for another decade (until Genentech gives up its appeals, which is highly unlikely, or the patent validity is ruled on by the Supreme Court), Genentech stands to make billions from related licensing in coming years.
On Sept. 13, 2005, in a reexamination case brought by MedImmune Inc. (see the Synagis entry), the U.S. patent office rejected an extension of the New Cabilly patent (6,331,415) on the grounds that it covered overlapping material to the first Cabilly patent. The patent office rejected the patent under the doctrine of “obviousness-type double patenting” and said New Cabilly is unpatentable over claims of U.S. 4,816,567, the first Cabilly patent issued in 1989. MedImmune’s counsel had argued that claims 1-36 of the Cabilly II patent were obvious variants of claims 1-7 of the Cabilly I patent and, thus invalid for obviousness-type double patenting. In the meantime, despite this ruling, the patent and Genentech’s licenses remained fully enforceable. Genentech described this ruling as “a routine and expected next step in the reexamination procedure” and filed an appeal. By rejecting all of the existing New Cabilly claims, the patent office would force Genentech to make its case for each one explicitly and in a manner that addresses the arguments that the patent office has laid out in its judgment.
On Oct. 18, 2005, the U.S. Court of Appeals for the Federal Circuit (CAFC) in Washington, DC, ruled against MedImmune in its appeal to have Genentech’s Cabilly II patent ruled invalid along with alleged violations of antitrust, patent and unfair competition laws. MedImmune, a licensee in good standing, having kept up with its royalties paid to Genentech from sales of Synagis, had filed a declaratory judgment action challenging the patent’s validity and enforceability. The court ruled that MedImmune lacked standing and could not bring suit against Genentech, since there was no legal dispute, since MedImmune had not violated its license agreement (contract), e.g., by refusing to pay royalties due. This and the lower court ruling this affirmed based their decisions on the Federal Circuit decision in Gen-Probe Inc. v. Vysis. Inc., holding that a patent licensee in good standing cannot establish the requisite case or controversy under Article III to invoke the Declaratory Judgment Act because the license agreement removes any reasonable apprehension of a suit for infringement. MedImmune had not stopped paying royalties, because this would have resulted in immediate revocation of its license, threatening marketing of Synagis (an RSV monoclonal antibody; see related entry), it’s main product, and exposed the company to severe finanacial penalties. The court also upheld the dismissal of MedImmune’s antitrust claims.
On Jan. 9, 2007, the U.S. Supreme Court (in MedImmune, Inc., v. Genentech, Inc., et al.; Case No. 05-608) reversed the U.S. Court of Appeals for the Federal Circuit finding that MedImmune had no standing to contest its Cabilly license with Genentech and to challenge the validity of the New Cabilly patent. The Supreme Court ruled that MedImmune’s Declaratory Judgement Act complaint pled a contract dispute, that it had no obligation to pay royalties on an invalid patent and that was sufficient to confer subject-matter jurisdiction under the Act. The Supreme Court articulated the specifics of the case as follows: “the question in each case is whether the facts alleged, under all the circumstances, show that there is a substantial controversy, between parties having adverse legal interests, of sufficient immediacy and reality to warrant the issuance of a declaratory judgment.” The issue was whether MedImmune, having made royalty payments “under protest and with reservation of all of [its] rights,” elilmianted the existence a legal case or controversy. Justice Scalia summarized the decision as holding that “a licensee’s failure to stop its royalty payments did not render a dispute over the validity of the patent non-justiceable,” i.e., just because a patent licensee is in good standing regarding royalty payments and other license/contractual requirements, it can still challenge the validity of the patents involved. Note, the Supreme Court made no rulings regarding the New Cabilly patent or its licensing, just that MedImmune (and other licensees) could challenge the patent. Those having supported MedImmune in its case by filing amicue briefs included the U.S. Solicitor General; the General Counsel of the U.S. Patent & Trademark Office; the Generic Pharmaceuticals Association; the Natural Resources Defense Counsel; and Medtronic, Inc.
On Feb. 21, 2007, the U.S. patent office issued a final Office action in its reexamination of Cabilly, et al., U.S. 6,331,415 (New Cabilly) and rejected the patentability of all of the claims of the patent, effectively revoking/invalidating the patent. Genentech is appealing the decision. In the meantime, the patent remains valid and enforceable through the appeals process. Genentech has estimated that the entire appeals process may take one to two years, or longer.
On Feb. 25, 2008, the U.S. patent office, upon appeal issued its final office action, upholding its Feb. 2007 rejection of all Cabilly-Boss (6,331,415) patent claims. This is being appealed by Genentech to the Board of Appeals within the patent office.
In June 2008, MedImmune and Genentech (also acting for the City of Hope Medical Center) came to a settlement, ending their particular litigation. Financial terms of the settlement were not disclosed. The settlement permits MedImmune to obtain licenses for certain additional pipeline products under the Cabilly patent family.
Genentech continued its appeal the U.S. patent office regarding revocation of the patent. In the meantime, it was still receiving royalties and probably even seeking to sign up new licensees.
On Feb. 24, 2009, the U.S. Patent and Trademark Office issued a Notice of Intent to Issue a Reexamination Certificate (NIRC) confirming the patentability of all claims of the Cabilly et al. patent, U.S. 6,331,415, of which claims 21 through 32 were amended in a manner that did not affect the commercial importance of the patent. This favorable decision by the Patent Office is final and unappealable. The decision resolved the reexamination proceedings involving the Cabilly patent, which the Patent Office conducted on the basis of two different reexamination requests, an Inter Partes Reexamination (RE Appl. No. 90/007,542) and an Ex Parte Reexamination (RE Appl. No. 90/007,859), filed by third parties in 2005 that were later merged into one proceeding. Genentech expects the reexamination certificate will formally issue later in 2009, formally concluding the current reexamination proceedings.
Note, Genentech’s intellectual property concerning humanized and other recombinant monoclonal antibodies is broader than just the New Cabilly patent. Besides its own relevant patents, Genentech has used its strong/dominant position to obtain rights to other patents. For example, in Jan .2006, Genentech and Amgen cross-licensed to each other multiple patents related to recombinant antibodies and their manufacture, including Amgen receiving rights to New Cabilly. Amgen refused to provide any information regarding what it licensed to Genentech in return, and neither party would disclose what other Genentech patents, besides New Cabilly, were licensed to Amgen, or whether Genentech and/or Amgen have rights to further license their cross-licensed patents to other companies.
This dispute will likely take years to fully resolve, and the Supreme Court may well end up deciding the validity of New Cabilly. Most analysts expect that Genentech has a good chance of successfully appealing and retainng New Cabilly. Earlier challenges to the original Cabilly patent were unsuccessful, and the Genentech has a history at aggressively defending itself in such cases.
Market: As can be seen from review of this section, relatively few non-recombinant monoclonal antibody products (in this section), have been developed and even fewer remain on the market.
The author estimates the total 2006 worldwide market for therapeutic and immunodiagnostic monoclonal antibodies worldwide, primarily recombinant products, to be $19.9 billion, based on summing of sales reported in this publication.
In May 2006, a Research and Markets study reported the global market for therapeutic monoclonal antibodies was $13.6 billion in 2005. A May 2005 report had estimated the global market was $11.2 billion in 2004, and that “The market has been growing at an impressive compound average annual growth rate of 42% over the previous five years... Industry participants forecast of the market vary widely, however, a consensus is emerging that the market should reach US$34 billion by the end of the decade [2010]”
A D&MD study in Jan. 2002 reported that total worldwide sales of approved monoclonal antibodies were over $2 billion in 2000. More current and accurate estimates of worldwide (and often U.S.) sales may be derived from examining the entries for recombinant monoclonal antibodies (in the Recombinant DNA Products section) and the entries below for non-recombinant monoclonal antibodies.
Index Terms:
Companies involvement:
Full monograph
300 Monoclonal Antibodies, including Recombinant Mabs
antibodies (see also immune globulins; monoclonal antibodies)
biopharmaceutical products
monoclonal antibodies
NA
NA
NA
NA
Copyright© 2020, Biotechnology Information Institute