July 1, 2005 (Vol. 25, No. 13)
Engineering and Expressing Novel Protein Therapeutics
Driven by technology, protein and peptide drug exploration has been fed by the success of the human genome project, striking improvements in technologies such as mass spectrometry, advances in molecular genetics, and progress in production technologies, including advances in establishing authentic glycosylation patterns.
“Protein drugs have moved from the fringes to the center of pharmaceutical R&D,” stated Jrgen Drews, M.D., managing partner of Bear Stearns Health Innoventures Management at the recent Cambridge Healthtech Institute conference, “Phage Display for Engineering Protein Therapeutics,” held in Cambridge, MA.
“In the latter half of the 20th century, innovations in recombinant DNA technology including phage display have allowed the development of increasingly sophisticated antibody and protein-based pharmaceuticals.”
In terms of their potential as drugs, proteins offer unique advantages, resulting from their great versatility and the fact they define the phenotype through their ability to carry and process information. This wealth of possibilities has not been lost to the industry. Indeed, understanding their structure provides a basis for the rational design of small drugs targeted at specific functions.
Since there are a vast number of “druglike” compounds that can be constructed from the basic elemental building blocks, an intimate knowledge of the proteome will be essential to sort out effective low molecular weight compounds from the vast array of possibilities.
While rational design has yet to make a substantial impact on products arising from Big Pharma, the elaboration of drugs based on recombinant proteins is a rapidly expanding field, even while the introduction of new low molecular weight drugs into the marketplace has slowed to a crawl.
There are already 75 protein therapeutics in the marketplace, but the total pipeline of preclinical to market candidates is 180, much larger, which bodes well for the future of the industry. This flood of products led to combined sales of $31.7 billion in 2003.
With regard to engineered antibodies, predictions are even brighter, based on the number of companies, greater than 200, with 21 products approved and 100s in the pipeline. Though the leading therapeutic area is oncology, the fastest growing indication is for autoimmune disorders. Sales of $5.4 billion in 2002 are predicted to reach $16.7 billion by 2008.
Dr. Drews described the recombinant protein field as a maturing industry, now reaching the status of a fully integrated sector. In such a climate it is anticipated that “blockbuster dependency,” the domination of the market by a few massive revenue generators, will decline as many products, some of them targeted at relatively small markets, are disseminated throughout the marketplace.
One of the most active areas today is the field of peptide mimetics, in which engineered peptide sequences can be used to substitute for the entire molecule. The widespread use of sophisticated crystallographic analysis and 3-dimensional molecular modeling makes possible the assembly of peptides which act as potent therapeutics.
There is no absence of hopeful entrants in the global peptide market today; out of 720 candidates in various stages of development, 56% have reached the level of advanced clinical trials. This appears to be a particularly fruitful line of attack, allowing companies to address targets that are inaccessible to small molecules as well as novel targets.
For instance, Affymax (Palo Alto, CA) developed Hematide, an erythropoiesis stimulating agent for chronic renal disease and cancer, both of which represent multibillion dollar markets. Hematide is a peptide developed as an initial weak binder from a phage library with no agonist activity against erythropoetin.
It was subsequently improved by mutagenesis. Since its amino acid sequence is unique and very different from erythropoietin, it provided the company with patentable drug candidates that are not blocked by recombinant protein intellectual property.
Peptide drugs have both advantages and disadvantages (Table 1).
Dr. Drews stresses a trend in pharmaceutical discovery that paints a bright future for biopharmaceuticals. Since 1997 the number of approvals for biopharmaceuticals has increased year by year while at the same time approvals for new chemical entities (NCEs) has dropped.
By 2003, the comparative U.S. success rates (approval of new products) for new chemical entities hovered around 1517%, whereas various recombinant products were twice as likely, on average, to gain approval. In the case of antineoplastics, success rates were as high as 60%. These numbers perhaps reflect the philosophical change in drug research paradigms.
Alternative to Phage Display
Phage display, invented in 1985 by George Smith, has given rise to a whole realm of research, with thousands of reports in the literature over the last 20 years.
However, there are alternative approaches to selecting variant proteins, including bacterial display, as discussed by Patrick Daugherty, Ph.D., and Hyongsok Soh, Ph.D., of the departments of chemical and mechanical engineering, respectively, at the University of California, Santa Barbara. Here the number of papers is much smaller, with a few dozen reports in the literature.
Nonetheless, this approach has much to recommend itself, according to Dr. Daugherty. Phage display ordinarily employs panning, allowing the phage carrying the gene for a desired protein sequence to bind to a surface coated with an antibody. Unbound phage are washed away, and the phage carrying the desired epitope are eluted.
Bacterial display, on the contrary, uses cell sorting instrumentation, with fluorescently labeled antibodies that bind to the appropriate bacterial cell epitopes. The proteins are displayed on the surface of the cells, but in some cases displayed proteins have not been sufficiently accessible to extracellularly added antigens, thus precluding screening. This may make them less accessible to the proteins that serve to indicate their presence.
Dr. Daugherty’s group has focused on improving bacterial display technology, which has suffered from numerous technical drawbacks, no doubt contributing to its lack of popularity with the biotechnology community.
In order that the proteins (or peptides) be clearly displayed, Dr. Daugherty and co-workers generated libraries of target sequences fused to a surface-exposed N-terminus of a circularly permuted outer membrane protein (Omp) variant which serves as an anchor, with the displayed protein projecting far out from the cell surface.
The Omp protein forms a circular basket-shaped configuration, firmly tethered to the displayed protein. This technology allows for a high affinity screening in which peptides can be recovered with 20-fold greater affinities than the best values reported for phage display (Kd in the range of 4 nM).
The other problematic component of bacterial display attacked by Dr. Daugherty and Dr. Soh is the detection screening system. Cell sorter (FACS) machines are expensive, running to half a million dollars. Dr. Daugherty and Dr. Soh have sought to design an inexpensive chip-based flow sorter, which separates particles according to their size and dielectric properties.
The researchers attached polystyrene spheres of 3 m size to antibodies which bound to the proteins on the cell surface. This principle, quite different from the FACS technology, has achieved 180-fold enrichment in the space of a few hours and can be carried out by an unattended instrument.
Most FDA-approved therapeutic antibodies for cancer treatment are naked antibodies, which activate the host defenses to react against the cancer cell. Although these agents produce clinically significant responses in patients, they rarely result in a permanent remission of the tumor and the patients invariably relapse.
While unconjugated antibodies are a more “natural” approach to treatment, there is increasing interest in the use of conjugated antibodies for situations in which the natural host defenses mobilized by the naked antibody are insufficient to overcome the malignancy.
In principle, the concept is simple: the antibody targets a cancer cell, delivering a drug payload. This approach has proved difficult to realize in actuality, yet there are a number of conjugates in development or in clinical trials and one (Mylotarg) that has gained approval (Table 2).
Paul Carter, Ph.D., vp for antibody technologies of Seattle Genetics (Bothell, WA), discussed his company’s efforts to target CD30 using a conjugated antibody. This attractive target for Hodgkin’s disease and T and B cell lymphomas is expressed in up to 600,000 copies per cell, while it shows very limited expression on normal tissues. Some monoclonal antibodies against this marker are efficiently internalized in tissue culture model systems.
The Seattle Genetics team conjugated monomethyl auristatin E, a powerful dolastatin 10 derivative to their antibodies, through protease cleavable linkers. The auristatin derivative is released inside the cell where it inhibits tubulin formation.
In animal experiments, the anti-CD30 auristatin conjugate proved to be highly efficacious against Hodgkins disease and anaplastic large cell lymphoma tumor cell lines. The antibody-drug conjugated was well tolerated by the animals and cleared slowly, in some cases still detectable after 50 days in the circulation, with a half life of ~6 days in cynomolgus monkeys. The drug conjugates displayed significant antitumor activity with a broad therapeutic window in mice.
Seattle Genetics scientists embarked upon a program to produce an optimized antibody-drug conjugate. They used chemical modifications to reduce different cysteines and generated a family of molecules with different levels of conjugation.
These forms could be separated by hydrophobic interaction chromatography, reducing heterogeneity and yielding families of molecules with varying numbers of auristatin molecules conjugated to them. The in vivo therapeutic window was increased by ~twofold by tailoring the drug loading stoichiometry.
In addition, Carter and colleagues modified the amino acid sequences of the antibodies, replacing cysteines in the hinge region with serines, which do not form disulfide bonds. These mutational modifications did not alter the antigen binding of the antibody. This antibody engineering strategy was used to define the stoichiometry and site of drug loading to reduce conjugate heterogeneity and/or improve yield. In vivo antitumor efficacy and toxicity of engineered antibody-drug conjugates are comparable to partially loaded parent conugates.
The strides that have been made in the past year in engineering protein therapeutics have addressed issues of immunogenicity, toxicity, and aggregation. Variations on display technology including yeast and ribosomal display have allowed optimization of protein drugs by improving their stability and ease of delivery.
With the present flourishing pipeline and many more items in the laboratory the field will continue to overshadow low molecular weight drug development.