November 1, 2005 (Vol. 25, No. 19)
Roadblocks Serve as Powerful Drivers for Complex Technology
Despite the fact that monoclonal antibody technology celebrates its 30th birthday this year, its movement into adulthood has not brought a serene and settled maturity.
“Going forward is difficult,” according to Richard Siegel, Ph.D., vp for pharmaceutical development at Centocor (Malvern, PA).
The introduction of OKT3 in 1985, a murine antibody used in transplantation rejection, appeared to herald the dawn of a new treatment paradigm. But progress was at a snail’s pace into the 1990s, until a host of new engineered antibodies achieved FDA approval.
As Dr. Siegel outlined in a presentation at the recent “SRI World Antibody Summit,” successful introduction of antibody products has been in the area of therapeutics, with little application to diagnostics.
The complexity of the technology, high costs, and long development time constrain the search for recombinant diagnostic antibodies, as biotech companies have aggressively pursued blockbuster antibodies for treatment of cancer and autoimmune diseases.
“Timing is critical, and we’re forced into risky spending propositions by the necessity to move a program forward without knowledge of its efficacy,” Dr. Siegel continued. Not only are the development costs high, but Mabs are expensive to produce.
Regulatory Concerns
While there is great interest in alternative production hosts, such as yeast, transgenic animals, and bacteria, the only FDA-approved antibody products are generated in mammalian cells grown in bioreactors. Options discussed at the meeting included Lemna, or duckweed, and the use of transgenic goats, corn, and chicken eggs.
In Centocor’s case, the commitment to build a large-scale production facility was made years ago, during Phase I evaluation. “Even in the best of situations there always is an element of risk,” Dr. Siegel said. “However in this instance it turned out to be a good choice.”
The regulatory picture today is especially complex because of the need to meet both European and American criteria. “While the U.S. regulatory situation is streamlined, the European regulatory climate can be challenging due to the inconsistent implementation of new regulations among all member states.”
Dr. Siegel expressed praise for the FDA as a good business partner, and feels FDA officials have gone out of their way to make the process as efficient as possible. The FDA will face the challenge of evaluating alternative antibody production technologies in the coming years, as transgenic animals and plants are considered.
“Prions are an issue of great concern in proteins produced in transgenic mammals,” he said. “There probably won’t be an acceptance by the public until a transgenic protein has moved through the entire regulatory process to the market place.”
Dr. Siegel’s remarks were reinforced by comments made by Jim Cornett, Ph.D., vp for business development at Medarex (Princeton, NJ). “When it comes to antibody targets, the low hanging fruit is gone,” he stated, “companies need to be more innovative in identifying novel targets for new antibody-based therapeutic products.”
By developing simpler technologies, such as disposable inserts for bioreactors, antibody expenses may be driven down considerably. In this respect, transgenic plants may prove to be of particular value in squeezing the cost of antibody production.
“We’re also concerned with purification costs,” Dr. Cornett added. “Protein A is costly, and alternate chromatographic methods could greatly lower downstream purification outlays.”
Dr. Cornett is particularly intrigued over the prospect of polyclonal recombinant antibodies. “The technology under development by Symphogen and Merus is especially appealing, as it combines the virtues of polyclonal antibody’s potential efficacy with the ability to engineer recombinant antibodies,” he said.
“This may be a promising technology for the area of infectious diseases as well as cancer. In the latter area, a cure is not required if the anticancer therapy can extend the patient’s life to a normal lifespan.”
One of the main barriers to effective antibody therapy is the lack of specific antigenic targets in cancer cells. Whereas cancer cells express a variety of proteins not seen in normal cells, these differences are mainly quantitative, and obtaining truly cancer-specific markers is virtually impossible.
An unusual approach to new anticancer therapeutic antibodies was discussed by Jon Weidanz, Ph.D., chief scientist at Receptor Logic (Amarillo, TX).
“We have approached the problem of developing antibodies that behave as T-cell receptor mimics,” stated Dr. Weidanz. The major histocompatibility complex proteins, located on many cell types, display specific peptides, forming a unique complex that has the potential to behave as a specific cancer marker.
“An antibody generated against a peptide-MHC complex will behave as a mimic of the T-cell receptor (TCR). This is because in the course of a normal immune response, protein antigens are broken into peptides that bind to the major histocompatibility complex proteins.
“This complex, in turn, is recognized by one of myriad different T-cell receptors, and the association of the MHC-Peptide with its matched T-cell receptor sets in motion a cascade that will result in the immune response.
“We reasoned that antibody mimics to the T-cell receptor could be used in a wide variety of therapeutic functions, including cancer cell targeting for therapy and imaging, and presentation of drugs to tumor cells,” Dr. Weidanz stated.
It had previously been shown that a translation initiation factor known as eIF4G was overexpressed in many cancer cells, with the result that peptides from its breakdown traveled to the cell surface bound to the major histocompatibility complex proteins. Dr. Weidanz and his coworkers were able to engineer similar complexes and employ them as antigens.
In the past, the approach of targeting the MHC-receptor complex in cancer cells has not been particularly successful, so the Receptor Logic group was determined to use a new technique. The Receptor Logic strategy of selecting promising peptides, such as the eIF4G peptide, provides a unique way for isolating responsive epitopes.
The Receptor Logic team generated antibodies against the complexes using a proprietary method, which included a specially formulated peptide-MHC immunogen, an optimized number of immunizations, and development of highly specific and sensitive screening assays.
While over production of the MHC-eIF4G complex is a recognizable feature of HIV infection, antibodies to the MHC-eIF4G complexes bound to only a limited number of cancer samples, as demonstrated by flow cytometry, using the antibody mimic.
Another well-known marker, Her2/neu, present in a wide range of breast cancer cells, was selected as a candidate. A TCR mimic antibody was generated against one of the Her2/neu peptides-MHC complexes, and the antibody was used to predict the presence of Her2/neu epitopes in a sample of cancer cell lines. In theory, a number of overexpressed cancer-related antigens could behave as markers for the T-cell receptor.
The antibodies that mimic the T-cell receptor are novel tools for target validation, and they have the potential to perform as a unique therapeutic tool, by arming them with drugs that when internalized, kill the appropriate cancer cell. The Receptor Logic group is pursuing this tactic with both cell culture and animal model investigations.
Biospecific Antibody Therapy
Combination therapy targeting two or more markers is a widely used approach, providing backup for attacking cancer cells that may escape through mutation and become refractory to a single anticancer agent.
In the field of antibody therapy this approach has been hindered by a number of factors, including limited availability of satisfactory antibodies, high costs, and regulatory complexity derived from simultaneous evaluation of two distinct antibodies.
Bispecific antibodies are not a new concept, and their description in the literature goes back 20 years or more. The Y-shaped structure of the classical antibody molecule is composed of two light and two heavy chains, held together with disulfide bonds.
Whereas the natural structure consists of two identical heavy and two identical light chains, bispecific antibodies are engineered so the two antibodies’ arms possess different specificities. Bispecific antibodies can be produced using recombinant technology, and there are a number of variations on this theme.
For instance a diabody can be produced, resembling two different variable regions joined together, back to back, with their antibody combining regions facing outward. Other permutations include diabodies joined to heavy chains for greater stability.
Significant progress has been achieved in the past decade in the production of various bispecific antibody fragments, such as the diabody and minibody constructs, via various recombinant technologies.
On the other hand, the successful design and production of full length IgG-like bispecific antibodies has been generally limited. The benefits of an IgG-like bispecific antibodies over smaller fragments include longer half-life (because of the Fc domain and the large size) and the capability in mediating immune effector function such as ADCC and CMC.
For these reasons the use of polyclonal and bispecific antibodies, as described by Zhenping Zhu, M.D., Ph.D., vp of antibody technology, ImClone Systems (New York City), are particularly noteworthy. In principle this line of attack should combine the power of a two-pronged strategy with the level of control associated with use of a single reagent.
The ImClone team elected to pursue antibodies against the insulin-like growth factor receptor and epidermal growth factor receptor, two markers widely expressed in cancer cells. Indeed, the successful anticancer antibody Erbitux is directed against the epidermal growth factor receptor. The goal was thus to build a tetravalent bispecific diabody that would bind simultaneously to both targets.
However, such elaborate structures may assume three-dimensional folding that is incompatible with antibody function for both specificities. Indeed, when Dr. Zhu and his colleagues inserted cysteines into different sites, the initial choices were unsuccessful, and only one or the other of the two specificities was expressed.
It was not until they engineered a carefully designed insertion of cysteines into the molecule that they were able to generate a Di-diabody that acquired the biological activity of both parent molecules. The disulfide bond-stablized Di-diabody should be more stable in vivo, and thus potentially more potent as well.
When evaluated in human tumor xenografts in nude mice, the newly engineered molecule was equally active in cancer cells, expressing either growth factor receptor. Further the Di-diabody mediated effective antigen dependent cell cytotoxicity, demonstrating a functional Fc domain. In fact, the Di-diabody was as effective as the combination of the two monospecific parent antibodies. It should be effective in patients with a wide range of cancers.
Vaccinex, based in Rochester, NY, is a privately held company focusing on a novel production method for generating fully human Mabs through use of a vaccinia vector system.
Maurice Zauderer, Ph.D., CEO and president, presented his company’s technology. “We have the only large library technology that can express fully functional antibodies in mammalian cells,” he stated. By using separate human light chain and human heavy chain libraries, a vast number of possible antibodies can be generated, up to 1014. The intellectual property position of Vaccinex is consolidated through a number of issued and pending patents.
Humanizing Antibodies
Vaccinia has a number of favorable qualities as a mammalian expression vector, including the fact that it is a cytoplasmic virus packaged into many fully infectious particles that are easily recovered from an infected cell.
Screening is performed by coinfecting HeLa cells with the two libraries, plating out cells in microtiter plates, and performing ELISAs on the supernatants, in a fashion similar to conventional Mab screening. Positive wells are harvested, and the yield of viral particles is subcloned and rescreened until the appropriate antibody candidate sequences are isolated.
Virus encoding the specific heavy and light chain genes can then be recovered and amplified. The platform is quite robust, allowing recovery of highly recalcitrant types of antibodies.
In one example, an antigen that differed between mouse and human by a single amino acid was plugged into the Vaccinex system, generating antibodies with affinities as high as 10 nM. By titrating the antigen dose, lower affinity variants could be eliminated.
“Our technology essentially mimics the natural human immune system,” said Dr. Zauderer. “The initial antibodies are of moderate affinity, as is the case in the primary human immune response. Then we boost the affinity to a higher level by further rounds of selection.”
Dr. Zauderer described the use of a gene replacement strategy, which imitates the in vivo immune response. After isolation of the initial antibodies, a V gene replacement library is employed, either a heavy chain replacement or a light chain replacement.
The heavy chain library members are paired with the moderately performing light chain, and then the process is reversed using a light chain library, against the heavy chain. This allows adding sequences, which replace lower affinity sequences with higher ones, resulting in final products of much higher affinity. In addition, in vitro mutagenesis protocols can be employed, to further increase antibody diversity.
“One of the important advantages of our platform is that we express antibodies directly in the mammalian cell host,” Dr. Zauderer explained. “This allows us to avoid re-engineering an antibody sequence that may express well in a non-mammalian host, but require substantial modifications in order to make the transition to a fully mammalian system.”
The Vaccinex group has a large portfolio of antibodies available, and can reportedly generate new antibodies in a time frame of approximately four to six months. In addition to providing antibody selection services to generate near-term revenue, the company actively participates in co-development partnerships, of which a number are currently ongoing.
Where Are We Headed?
Antibody technology is perhaps the most active area of pharmaceutical development today, making prognostications particularly difficult as the picture changes from day to day. Controlling the astronomical cost of these therapies is critical if the healthcare delivery system is to avoid a major train wreck.
While price controls may be advocated, there is neither the political will nor the legislative mechanisms to bring this about. Improvements in production efficiency will come incrementally, but major cost savings and improvement in patient outcome could be realized if antibodies, like all proteins, did not have to be injected.
According to Dr. Siegel, the push to develop alternative routes of administration is a pressing as well as a daunting challenge. One interesting alternative is the inhalation route, currently being considered for delivery of insulin. It is not obvious how this could be adopted to antibody delivery, but these challenges will be powerful drivers in the antibody industry for the foreseeable future.