November 1, 2007 (Vol. 27, No. 19)
Techniques Improve Optimization, Discovery, and Production Speeds
Antibody engineering is entering a new phase, moving from mouse to humanized and fully human antibodies. Novel technologies such as de novo generation of human antibodies, engineering techniques to increase antigen affinity, and production of antibody fragments are replacing standard methodologies.
There are currently more than 150 companies developing about 300 antibody-based oncology drugs, according to a May 2007 report by the BioSeeker Group, “Antibodies in Oncology: Drug Pipeline Update 2007.” In addition, according to the report the antibody market is expected to double in value over the next five years to $29.7 billion. At the upcoming IBC “Antibody Engineering” conference in December in San Diego, companies will be presenting their latest antibody engineering technologies.
Inhibiting DNA Repair
Morphotek (www.morphotek.com) utilizes morphogenics to create human antibodies. Licensed from Johns Hopkins University, and developed by company CEO Nick Nicolaides, Ph.D., the technology is based on inhibiting the DNA mismatch repair process that occurs during DNA replication to correct mistakes in the molecule’s helix.
“By blocking this process via a modified gene or a chemical, we’re able to accumulate mutations in cells that have been exposed to these blocking tools,” explains Luigi Grasso, Ph.D., senior vp, R&D. The mutations are screened for newly acquired phenotypes that are of interest in the biotech arena.
The company leverages morphogenics in two applications. The first is to enhance hybridomas in order to produce human antibodies. “Making human hybridomas has been attempted over the past 20 years, but the problem is that they tend to be unstable, only secrete small amounts of antibody, or have low affinity. This is where Human Morphodoma® comes in. It uses morphogenics to make human B-cell hybridomas that produce human antibodies targeted against disease-associated antigens,” says Dr. Grasso.
This process avoids licensing third-party royalties required by recombinant antibody technologies such as phage display libraries or transgenic mice. “Our technology allows one to produce human antibodies from scratch, so to speak,” he explains.
The company currently has two human antibodies developed from its Human Morphodoma platform: MORAb-22, which is indicated for treating inflammatory disease, and MORAb-28, which is scheduled to start clinical trials next year for metastatic melanoma.
The second application for morphogenics is Libridoma™, which generates libraries of hybridomas that can be screened to identify fully human mAbs to known antigens and novel targets for diagnosis and treatment of disease.
From Discovery to the Clinic
XOMA (www.xoma.com) offers two validated therapeutic-antibody discovery platforms: antibody phage display and hybridomas. “We have multiple libraries including seven commercial antibody phage-display libraries, each with a unique repertoire of more than 10 billion human antibodies, and we screen two or three in parallel to speed the discovery process,” explains Mary Haak-Frendscho, Ph.D., vp, preclinical R&D.
“We also use two panning strategies, so this translates into four to six simultaneous campaigns. This increases the probability of success for finding rare therapeutic antibodies that have both the required binding specificity and desired function,” adds Dr. Haak-Frendscho.
Its humanization technology, called Human Engineering™, is based on conserved structure-function relationships among antibodies and defines which residues in a nonhuman variable region are candidates for substitution. Human Engineering provides four variants of humanized antibodies within three months, versus six months for most other methods, according to the company. Its track record to date is 100% success with 11 parental antibodies, Dr. Haak-Frendscho says.
To illustrate the efficiency of these platforms, the company will present data at the conference on XOMA 052. “It has some unique characteristics, it targets the receptor instead of the ligand and it is potent with 0.3 picomolar affinity,” explains Dr. Haak-Frendscho. Since affinity is related to potency, this offers potentially infrequent dosing, most likely a monthly dose, she adds.
XOMA 052 is a human antibody that targets IL-1 beta, which plays an important role in initiating the cytokine cascade that induces inflammation, so the idea is to dampen the response early, reports Dr. Haak-Frendscho. “With diabetes, there is inflammation of the pancreas with the beta cells that modulate insulin. We’re trying to reduce inflammation that damages these cells,” she explains. Phase I trials in the U.S. and Europe have been initiated for XONA 052 in type 2 diabetes. The company is also exploring its potential use in rheumatoid arthritis and systemic idiopathic juvenile arthritis.
Stem Cell Surface Receptors
Raven Biotechnologies (www.ravenbio.com) is generating antibodies that bind to cell surface receptors on human tissue fetal stem cells. “The argument is that tumors turn on fetal genes that they need in order to become tumors and then metastasize,” says Jennie Mather, Ph.D., founder, president, and CSO. “We have a technique that allows more than 50 percent of hybridomas to bind to the outside of the cell. That means we can generate tens of thousands of antibodies to saturate what’s on the outside of the cell and screen these for inhibition of cancer cells.”
The company’s platform is based on a series of steps that begin with its 30 cells lines as the source for novel antigens. An initial screen confirms whether the antibodies target something on the cell surface. If not, they are thrown out.
A second screen shows if there is significantly greater presence of targets on cancer versus normal cells. High-throughput tissue screening called Cell Array™ can assess the prevalence of a targeted antigen on a large number of cell lines and patient samples within 7–10 days, reports the company. Almost 90% of the antibodies fail and are discarded within the first week, which allows discovery to be focused on accessible, cell-surface cancer targets, according to Dr. Mather. This is much faster than the months required to express and purify a protein and then raise antibodies against a target identified in a genomics-based screen.
This discovery process makes target validation and drug-lead creation simultaneous events. “We have an advantage because the antibody is going to be the drug, all we do is humanize it,” says Dr. Mather. “Our approach finds a lot more potential targets because we use the whole cell.”
RAV12, the company’s lead antibody, has an unusual mechanism of action: oncotic killing not apoptotic. It is currently in Phase II trials for pancreatic cancer. “We have data that shows it can shrink primary tumors, but it also seems to stop metastasis,” adds Dr. Mather.
Increasing Protein Expression
“In protein therapeutics, there are often difficulties in fusing proteins together or adding linkers, including making the protein inexpressable in E. coli. We helped a customer express a fused antibody in E. coli. They had a construct with low expression, and our technology, Translational Engineering™, significantly increased the expression,” states Alan Carter, svp, business development at CODA Genomics (www.codagenomics.com).
The main thing that drives antibody research, adds Joseph Kiddel, Ph.D., vp of research, is to work with a protein in the lab that has some desired type of binding characteristics. “When you first start working with a gene, it’s far from optimization,” notes Dr. Kiddel. “The more work you do to optimize binding, the more you tend to wander away from the gene sequences that express well. We go back and reoptimize the gene by identifying the translation control signals that might be a problem in providing a high yield of the protein.”
The result is a domesticated gene that is easy to work with in the lab and expresses a lot of protein. If protein yield is increased in the early stages, Dr. Kiddel explains, this makes other processes like purification and fermentation easier. Also, since it’s known how the gene was created, it provides the customer with IP position on an optimized gene that may be superior to the original gene.
Developed at University of California, Irvine, Translational Engineering is based on ribosomal movement along mRNA. “As a ribosome moves along, it periodically reaches a signal that tells it to slow down,” explains Dr. Kiddel. “If your goal is to make a lot of protein, that may not be what you want it to do. These signals vary when you move from one organism to another. Particularly if you are moving a human antibody into E. coli, the place where the ribosome would normally pause has changed. This is why there are so many problems getting antibodies to express in E. coli.
“Our technology is aimed at adapting genes so they express better in E. coli,” says Dr. Kiddel, adding that the company is working to increase understanding of how the biology of translation control applies to each protein family.
Researchers at MacroGenics (www.macrogenics.com) are designing new generations of improved antibodies using Fc optimization to increase the specificity of activating Fc present on the surface of tumor-infiltrating effector cells. The company altered a functional genetics screen, yeast-surface display, to identify not only variants that bind to the activating Fc receptor, CD-16A, but also variants with decreased binding to inhibitory receptors.
“The idea was to use the yeast-surface display technology to have Fc versions tailored to meet the optimization expectations for specific antibodies for designated diseases or indications,” explains Jeffrey B. Stavenhagen, Ph.D., director, molecular immunology.
Fc receptors are on the surface of immunoeffector cells that act to trigger their effector cell function. “By selecting out specific Fc regions that are tailored to specifically hit the activating receptors and not target the inhibitory receptors, you have optimized the activity that antibody triggers against tumor cells,” notes Dr. Stavenhagen.
The yeast surface display allows fast screening of millions of mutants and enables viewing of larger libraries, explains Dr. Stavenhagen. “We look at changes in a biological functional assay rather than an in silico screen where you are relying on previous data,” he says.
The group was able to correlate the increased binding to the activating receptor with the increased tumor killing in a mouse model.
“We’ve been able to take it through identification and characterization in vitro and show in a tumor model mouse (both nude mice and transgenic mice) that those Fc-optimized antibodies enhance tumor-cell killing,” states Dr. Stavenhagen. In addition to focusing on cancer, the company also plans to modify Fc regions and tailor binding to other receptors such as those for autoimmune diseases.