At the recent Cambridge Healthtech “Peptalk Conference”, a range of topics related to protein expression and characterization were covered, but the event was notable for its discussion of various approaches to the optimization of biologic therapeutics.
Genes for potential therapeutic proteins can be easily cloned and expressed, but frequently the results are disappointing. Yields may be paltry or nonexistent and the molecules often fail to fulfill their earlier promises. To improve the efficacy and robustness of biological drugs, many companies are turning to sophisticated revamping and engineering technologies.
“DNA is extremely uniform, and DNA production projects are more than 99% successful,” said James L. Hartley, Ph.D., of SAIC (www.saic.com), a contractor for the NCI. “On the other hand, protein expression projects may have a success rate of less than 10%, due to their tremendous variability.”
Dr. Hartley’s protein expression group carries out about 100 projects/year for the NCI/NIH, mostly mammalian proteins. Starting with cDNAs from vendors, the team targets improvements and modifications in order to improve the quality and quantity of the proteins. In engineering their proteins, Dr. Hartley prefers the use of viral expression elements, arguing that because viruses are selected by evolution for protein overexpression, they are the most logical choice for high yields.
Dr. Hartley is the inventor of Invitrogen’s(www.invitrogen.com) Gateway cloning system and favors its use in recombinational cloning for reliability and speed. His experiences with E. coli demonstrate that while it can work well for smaller proteins, above 30–40 kd the success rate falls off rapidly. For this reason, mammalian cells are the obvious choice for obtaining the most native product.
“We try different kits for optimizing DNA purification and test many different culture media, which can yield big improvements in cell performance,” Dr. Hartley stated. “While mammalian cells are very successful for producing secreted proteins, the baculovirus and/or insect cell platform usually outproduces mammalian cells for cytoplasmic proteins.”
Proteolysis is a frequent issue during purification, and Dr. Hartley recommends speed and increasing the amount of protein per cell, assuming the amount of protease is unchanged, to maximize the yield of the protein of interest.
“We are able to address deficiencies in therapeutic products through our mutational strategies,” states Manuel Vega, Ph.D., CEO of Nautilus Biotech (www.nautilusbiotech.com), “which allows us to move toward the holy grail of oral protein therapeutics.”
Protein therapeutics are a gigantic market—$57 billion in 2005—so there is tremendous incentive to pursue improvement strategies. The lack of sophistication of products derived from native proteins means that ample opportunity exists for a range of promising reconfigurations.
Nautilus seeks to address various deficiencies, the most obvious being the short half-life of native proteins. A brief residency time in the circulation requires numerous administrations of the drug, which means that proteins must be introduced through the invasive and labor intensive processes of either subcutaneous or intravenous injection.
“The decreased half-life is the result of proteolysis in blood and tissues in addition to kidney filtration,” Dr. Vega explained, “so we identified the entry points for proteolysis in the proteins and then mutated one or more amino acids, thereby decreasing the sensitivity of the proteins to degradation and extending their half-life in the bloodstream.”
The Nautilus scientists used a proprietary technology to screen hundreds of different amino acid substitutions for the most optimal choices. These were conservative mutations that blocked proteolysis without introducing alterations into the biological activity of the protein. A single mutation in the Interferon Alpha molecule could increase the half-life from 2 to 40 or more hours, and additional mutations can have an additive effect.
Another benefit of these protease-resistant proteins is that, following subcutaneous injection, the molecules survive much longer in the skin, compared to the native counterparts, and only gradually enter the circulation. The skin compartment thus becomes a source of protein that is sustainably released into blood over time.
The Nautilus scientists are now evaluating oral dosing in which the engineered proteins are lyophilized and packed into coated capsules that resist the acidic environment of the stomach. Once in the intestine, the proteins are released from the capsule and remain available for absorption directly through the intestinal mucosa. This approach would increase patient compliance and lower the costs of administration, especially for drugs requiring multiple injections. Animal studies have shown no toxicity or serious side effects, the company reported, and Phase I clinical studies on these orally delivered proteins are scheduled to begin later this year.
The Nautilus approach has the added benefit of creating new intellectual property, as many of the older biologicals are coming off patent soon, if they have not already done so. Of course, the disadvantage is that a new series of clinical trials will be required in order to gain approval, but invariably a modification of an already approved product will have easier sailing through the evaluation process.
With a number of products in the pipeline, including alpha and beta interferon and human growth hormone, Dr. Vega and his colleagues are considering a range of other possibilities, including single chain antibodies. There clearly is no limit to this strategy, he said, which could revolutionize biological drug development.
“The Baculovirus system is ideal for safe and rapid expression of viral protein vaccines,” says Clifton McPherson, Ph.D., a research scientist in process development at Protein Sciences (www.proteinsciences.com). Dr. McPherson presented the company’s studies of the Baculovirus expression vector system in the development of a vaccine against SARS.
The company uses a variant of the familiar Sf9 cell line, grown in serum-free medium under GMP conditions at a 500-L scale. Dr. McPherson and his colleagues have introduced a number of modifications and improvements into the system that optimize its performance.
The Baculovirus system has a number of significant advantages over other methods of recombinant protein production, including lack of adventitious viral agents that could replicate in mammalian cells. Moreover, the insect host is a member of the lepidopterin order, which is a non-biting insect harboring virtually no adventitious viruses that can also infect humans.
Three vaccines produced in insect cells are close to market, including Provenge™ a prostate cancer immunotherapy from Dendreon (www.dendreon.com); Ceravix™, a papilloma virus vaccine from GlaxoSmithKline (www.gsk.com); and FluBIOk™ from Protein Sciences, a non-egg based flu vaccine.
Protein Sciences’ SARS vaccine is based on the spike protein, which allows the virus to bind to a mammalian cell receptor and is absolutely essential for infection. Two versions were cloned, a full-length and a truncated version, the latter of which was also expressed in E. coli during early development. There are 20 N-linked glycosylation sites that are glycosylated only in the mature form. Since vaccine production is based on a cloned recombinant viral protein, the Baculovirus platform is much safer than vaccines that require handling and processing of live virus.
Because the protein is heavily glycosylated, it can be conveniently purified on a lentil lectin Sepharose column. “While the procedure is not elegant, we were able to isolate a pure product of excellent quality,” Dr. McPherson stated.
The purified, truncated viral protein was combined with Alhydrogel® and used to immunize rabbits, either with or without adjuvant. The animals raised antisera that neutralized the SARS virus without causing serious side effects, Dr. McPherson stated. The SARS vaccine will be scheduled for Phase I trials in the near future.
The Baculovirus technology initiated by the Protein Sciences group could become a viable alterative to mammalian cells. Although there currently are no FDA-approved proteins expressed in Baculovirus, several are on the verge of acceptance. While the first protein drug to break this approval barrier will receive much scrutiny, subsequent candidates will certainly move more rapidly. With an active pipeline, the future of this approach looks bright.
Expression in Pichia
The Research Center for Applied Biocatalysis and the Graz University of Technology are developing innovative protein expression tools. Anton Glieder, Ph.D., professor at Graz’s Institute of Molecular Biotechnology, presented his group’s work on high-throughput protein engineering strategies in Pichia pastoris.“We have optimized our expression strategies by screening in small (300-mL) volumes in 96-well plates and then upscaling to 5-L volumes,” Dr. Glieder explained. “We use rational design, classical genetic engineering, and directed evolution of genes linked with new promoters in linear DNA cassettes. In this way, we avoid the time-consuming cloning steps in E. coli and go directly into the Pichia strain with our mutated genes.”
Although Pichia, compared to E.coli, has the disadvantage of slow growth and poor transformation efficiency, it can be engineered to yield up to 20–30 g/L of expressed protein, much higher than that obtained in baker’s yeast or E. coli. Using their high-throughput platform, Dr. Glieder’s group evaluated the effects of culture volume, media composition, time of induction and harvest, and cell viability. By adjusting these parameters, they were able to optimize yields.
The researchers’ major goal is to discover and produce large quantities of improved industrial enzymes using accelerated expression strategies. One of the most important of these is hydroxyl nitrile lyase, an enzyme that converts precursor molecules derived from plants to active forms that form a component of the plant’s defensive mechanisms against predators.
On a commercial level, the enzyme is used in the pharmaceutical industry to catalyze intermediate steps in the production of key intermediates for the synthesis of platelet aggregation and angiotensin converting enzyme inhibitors.
One of the most popular promoters is the AOX1 promoter, which is induced by methanol and repressed by glucose and ethanol. The Glieder group has done an extensive analysis of this promoter, testing the inductive activities of hundreds of mutated variants. With their new promoters, they have observed a 10-fold range of variation in expression evaluated in a model induction system. With a multicopy promoter strain, they increased the yield to more than 50 times the level obtained with the wild type promoter.
Through these model systems the Glieder team has established the tools for fast and efficient directed evolution of many proteins, sidestepping laborious cloning in E. coli. The availability of extensive libraries of new synthetic promoters for protein expression in Pichia enables the isolation of optimal promoter/gene combinations for each individual protein in a single high-throughput experiment.
Prokaryotic Genomic Libraries
“The genomes of exotic microbial species are fertile sources for unusual biocatalysts with new industrial applications,” says Emmanuel Maille, general manager for business development at Proteus (www.proteus.fr).
As Maille explained, the company searches out prokaryotic enzymes with unique commercial applications by employing high-throughput screening and polishes them using directed evolution protocols. The genomes of various extremophiles have been collected through worldwide collaborations.
Proteus’ protein-optimization technology uses cell-free expression systems to rapidly scan large numbers of candidates. Combining biodiversity with functional genomics allows Proteus access to these novel biomolecules.
Gilles Ravot, Ph.D., CSO of Proteus, presented the firm’s strategy for identifying esterase activities by high-throughput in vitro screening of a genomic DNA library from Pseudoalteromonas. Phenomics™ is the cell-free expression platform developed by his company for this purpose. Dr. Ravot stated that it is one of the few approaches for producing toxic or unstable proteins and for labeling target proteins specifically. Finally, the in vitro translational machinery is able to initiate at ORF located far away from the 5´- and 3´- transcript ends, thereby expediting the screening procedure. The high-throughput technology used 96-well plates with different representatives of their library in each well to which a cell-free transcription/translational mix and substrate are added.
The lipase/esterase screening platform presented by Dr. Ravot uses the synthetic C2-CLIPS-OTM ester substrate, which closely mimics the chemical structure and energy state of the natural substrate. Wells with functional enzyme molecules produced a colored product; these wells are retested and their status confirmed.
Because of their ability to catalyze reactions in organic media in a highly selective, stereospecific way, lipases/esterases are important industrial classes of biocatalysts. Their high temperature and pH endows them with stability in the presence of solvents and caustic agents making them highly valuable for bioindustrial applications, Dr. Ravot said.
“We seek to radically improve gene translation of transfected genes by reengineering translational speeds,” explained Joseph D. Kittle, Ph.D., senior vp of market development for CODA Genomics (www.codagenomics.com). The company, founded by Professors Wes Hatfield, Ph.D., and Richard Lathrop, Ph.D., of the University of California, Irvine, is based on the power of computation to design and build novel biologics and studies from the 90s that demonstrated that translational elongation rates are greatly influenced by over represented codon pairs in the context of the genetic message. From these investigations into the basic molecular biology of gene translation, commercial applications were developed with important implications for enhanced protein yield and improved functionality of protein drugs.
Transcription's Bumpy Road
Dr. Hatfield observed that certain nucleotide sequences of codon pairs occur in E.coli gene transcripts more frequently than would be predicted by chance and developed the hypothesis that these codon pairs relationships had evolved over evolutionary time spans in order to control rates of transcription. Furthermore, he reasoned that these frequencies were related to the structure of transfer RNAs, which had co-evolved in order to provide the control mechanisms for the rate of translational steps, without affecting the selection of amino acids for a particular site.
It stood to reason that by eliminating these translational pause sites, it should be possible to greatly improve the yield of proteins in engineered cell lines. However, these pauses have been observed to be conserved across species and may be important for controlling proper folding, solubility, and secretion of the proteins. Therefore, CODA builds codon pause maps for different genes, utilizing their importance for proper translation of the proteins.
In other cases, variation in pauses among species suggests that these signals are scrambled when a protein is expressed in a heterologous host. In order to enable heterologous genes to function optimally, much more than simple codon optimization is required. According to D. Adams, Ph.D., director of pharmaceutical account development, “CODA has developed patented technologies that optimize codon pair usage within an open reading frame and identify aspects of the sequences likely to reduce protein yield. It takes into account the thermodynamic requirements of the molecule in order to generate a protein product that will engage in proper folding.”
The CODA technology has been widely utilized to great advantage in the production of important engineered biological proteins. Researchers modified a translational pause to affect expression of a single chain antibody that was poorly expressed. By altering a single nucleotide in the flexible linker connecting a variable heavy chain with a constant heavy chain, they were able to achieve significant protein expression from a virtually silent clone. The targeted change was from GGA GGC to GGT GGC and resulted in a 30-fold increase in recombinant antibody expression. These high-level expression genes, in which the pause sites are removed, are referred to as hot rod™ genes by the company. “We believe that our technology can provide improved expression, solubility, and preservation of activity,” Dr. Adams continued.
Dr. Kittle asserts that CODA’s gene-design engine also optimizes the gene-assembly process, thereby streamlining synthetic gene production, and allows the simultaneous optimization of multiple genes and interchangeability of gene segments.