A study released by Drug and Market Development Publishing notes that peptide drugs represent a small but increasing number of pharmaceutical molecules. The report (Peptides 2006: New Applications in Discovery, Manufacturing, and Therapeutics) points out that peptide therapeutics are being developed for such disorders as allergy/asthma, arthritis, cancer, diabetes, cardiovascular and CNS diseases, inflammation, and others.
At the recent IBC “Drug Discovery & Development of Innovative Therapeutics” conference in Boston, Illana Gozes, Ph.D., CSO of Allon Therapeutics presented a case study of AL-108. The drug is an intranasal formulation of a neuroprotective peptide shown in a Phase IIa trial to enhance memory and cognitive function in patients with amnestic mild cognitive impairment, a precursor to Alzheimer’s disease.
Dr. Gozes and Donald Schmechel, M.D., Ph.D., professor of geriatrics at Duke University School of Medicine (medschool.duke.edu), realized statistically significant, dose-dependent efficacy data for AL-108. This provided conceptual validation of the therapeutic potential of addressing the tangles component of the classic Alzheimer’s “plaques and tangles” pathology, according to Dr. Gozes.
The function of early drugs was to stop the progression of Alzheimer’s disease and restore cognitive function by preventing or eliminating beta-amyloid plaques found in the brains of these patients. Dr. Gozes’ strategy involves targeting the neurofibrillary tangles comprising tau protein.
The tau protein normally functions to stabilize microtubules, which are the main structural elements responsible for maintaining the integrity of a cell’s cytoskeleton. When the tau protein undergoes hyperphosphorylation, it converts from a soluble to an insoluble form, resulting in degeneration of microtubules and the formation of neurofibrillary tangles composed of insoluble tau protein.
Dr. Gozes’ research provided evidence that AL-108, a peptide component of activity-dependent neuroprotective protein (ADNP) that can mimic the neuroprotective function of ADNP, accelerates the process of cellular microtubule assembly and interferes with the phosphorylation and toxic aggregation of tau. This helps to maintain the cytoskeleton, viability, and functionality of nerve cells.
The concept of identifying a functional epitope of a neuroprotective protein for development as a peptide or a peptide mimetic/small molecule therapeutic can serve as a model for developing drugs that cross the blood-brain barrier and represent alternatives to potential cell- or gene-based therapies.
As an alternative to mAbs, Phylogica is focusing its drug discovery efforts on phylomer libraries, which are enriched for highly structured peptides derived from a diverse range of bacteria.
These naturally derived peptides comprise combinations of prototypic structural elements called antecedent domain segments, which are thought to be the evolutionary building blocks of protein exons. Many of these basic elements contain supersecondary elements or subdomains, which have evolved to enhance the biological properties of peptides or proteins, improving their stability or their binding ability, for example.
Evolution has produced a limited repertoire of these secondary structures and protein folds. By focusing on these naturally derived sequences including those from microorganisms living at extreme temperatures, Phylogica hopes to identify peptides with optimized properties such as enhanced thermal stability or increased protease resistance. By compiling large libraries of hundreds of millions of individual phylomers, Phylogica screens for peptides that interact specifically and with high affinity to human proteins and that can inhibit targeted protein-protein interactions.
“This approach differs from other protein scaffold companies,” explained Paul Watt, D.Phil., vp of drug discovery at Phylogica, “in that thousands of structural families represented in phylomer libraries are screened in parallel for target binding before affinity maturation in contrast to screening multiple variants of a single structural fold family to generate hits. The end result is increased rates of functional hits from phylomer libraries, from which Phylogica has the luxury of choosing shorter sequences for synthetic production.
“We also use a peptidomimetic approach,” Dr. Watt continued. “All of the peptides we use in vivo we also make backwards, using D-amino acids. The backbone is reversed in these retro-inverso peptide mimetics, but with the exception of a couple of amino acids that have a second chiral center, the side chains are all present in the same configuration as in the original peptides. Phylogica has found that retro-inverso peptide analogues not only retain the activity of the original peptide but sometimes have increased activity.
“We have found that this approach also yields peptides with exquisite stability,” Dr. Watt added. Once the company identifies protein fragments that bind with good specificity and are biologically active it applies an affinity maturation approach consisting of homolog scanning, mutagenesis, directed evolution, and synthetic dimerization.
“From a primary screen we rapidly get multiple hits in the low-to-mid nanomolar range that are biologically active,” Dr. Watt said. “We are not aware of another peptide approach that has such inherent structural diversity both at the secondary and super-secondary structural level or at the tertiary subdomain level.”
Intracellular Target Program
For its intracellular target program Phylogica uses protein transduction domains to get phylomers into cells. It has so far validated several intracellular phylomers in a number of in vivo models, including for the treatment of acute burn wounds, traumatic brain injury, and stroke.
The company’s extracellular target program focuses on blocking growth factors and receptors, with one of the most promising targets being CD40 signaling ligands involved in multiple chronic inflammatory diseases. Phylogica has validated several hits against this target and has matured these to the low-nanomolar range. Another area of interest is developing phylomers as antimicrobial agents.
Dr. Watt points to the small size and simple chemistry of phylomers as additional advantages of the technology, making phylomers relatively easy to synthesize. “We have made 17 grams of a 29-amino acid phylomer at 96 percent purity in partnership with Genzyme,” Dr. Watt stated.
In terms of noninjectable delivery, phylomers have been applied directly to acute burn sites in which the dermis is absent. The company has also demonstrated biological effects of phylomers following inhaled delivery in a lipopolysaccharide-induced acute respiratory distress model, blocking neutrophil influx into the lungs.
In rodent models, the in vivo half-life of phylomers is about 100 minutes without the use of pegylation or any type of controlled-release delivery method, “which is compatible with acute delivery regimes,” Dr. Watt explained. “We have also pegylated one of our phylomers for half-life extension and have shown that it retains biological activity.”
Escoublac is a start-up company that is part of the Biogen Idec. It was founded by Gerard Karsenty, M.D., Ph.D., Columbia University College of Physicians and Surgeons (www.cumc.columbia.edu), after he and his colleagues discovered that osteocalcin, a protein secreted by bone-forming osteoblasts, has a role in energy metabolism and related disorders. Dr. Karsenty’s group postulated, based on gene knock-out experiments, that osteocalcin is involved in cross-talk between osteoblasts, pancreatic beta cells, and adipocytes.
Osteocalcin is post-translationally modified to include three gamma-carboxy-glutamic acid residues. The majority of osteocalcin circulates in the fully carboxylated form, with a smaller fraction under-carboxylated or noncarboxylated; the under- or noncarboxylated forms bind less tightly to bone. Dr. Karsenty determined that ostocalcin’s role in mediating this crosstalk depends on the ratio of the noncarboxylated to carboxylated form.
When pancreatic islet cells and adipocytes grown in culture are exposed to osteocalcin, the beta-cell mass increases, beta cells express and secrete more insulin, and adipocytes express and secrete more adiponectin, resulting in improved insulin sensitivity. Giving noncarboxylated osteocalcin to mice produces the same effects as seen in tissue culture.
If these results translate to humans, increasing the ratio of noncarboxylated to carboxylated osteocalcin in the body, “could be a new way to approach type 2 diabetes and possibly obesity,” reported Nils Bergenhem, Ph.D., CSO of Escoublac, which licensed the osteocalcin technology from Columbia.
Escoublac is developing noncarboxylated osteocalcin as an injectable appetite suppressant. The company is exploring both biological and chemical synthesis routes to produce osteocalcin. Dr. Bergenhem envisions a once-daily or twice-daily dosing regimen possibly by modifying or derivatizing the endogenous form of the peptides.
HDL: From Good to Better
Novartis Institutes for Biomedical Research in-licensed a mimetic peptide of Apo-A1 from Bruin Pharmaceuticals and is evaluating it in clinical trials as a treatment for atherosclerosis and cardiovascular disease. ApoA1 is a major component of HDL.
“HDL is not one particle; rather, it is composed of multiple types of subparticles that exist in a highly dynamic system,” explained Gene Liau, Ph.D., director of cardiovascular and metabolism research at NIBR. Research is under way to characterize the various proteins and lipids in HDL particles. “This is a field that is rapidly moving forward,” Dr. Liau said.
In addition to its role in cholesterol transport and metabolism, HDL is also involved in oxidation reactions. “If HDL is working properly it should have a positive role, removing or inactivating oxidation products,” Dr. Liau pointed out, “as a result, improving the state of the endothelial lining of blood vessels, blocking a variety of inflammatory processes, and exerting antithrombotic effects.”
NIBR’s developmental candidate, D-4F, is an 18-amino acid functional ApoA1 mimetic. It is attracted to lipid membrane surfaces as well as seeks out HDL in the body, and new evidence suggests that it has high binding affinity for oxidized lipids. The premise for developing D-4F as a therapeutic compound is that its incorporation into HDL will change how the particle functions and promote anti-inflammatory activity.
The mimetic is amenable to standard peptide synthesis processes, which NIBR is optimizing for improved efficiency. The initial iteration of D-4F comprises D-amino acids for enhanced protease resistance on oral delivery. Despite low levels of circulating peptide in early clinical trials of an oral formulation of D-4F, the results showed promising evidence of biological activity. Novartis is evaluating alternative delivery options to increase plasma drug levels.
According to Jeff Morhet, president and CEO of InNexus Biotechnology, Dynamic Cross Linking (DXL™) technology enhances antibody potency by promoting antibody aggregation at the target and can be applied to nearly any therapeutic antibody in development.
By conjugating an autophilic peptide to an antibody, DXL technology allows the antibody to remain monomeric in solution or serum while promoting cross-linking of the peptide-bound antibodies at or near the target site. The peptides essentially fit together, much like a zipper closing, enabling dynamic cross-linking of the antibodies; as one antibody dissociates from the target, another will be on hand ready to associate, thereby potentiating the overall effect of drug treatment.
“DXL does not change the binding constant or the affinity; rather, it significantly enhances the avidity—the complex structure built at the target,” explained Dr. Morhet.
InNexus developed a patented method for conjugating peptides to affinity sites on antibodies and has designed a platform technology for creating antibody-based therapeutics and diagnostics. The company currently has two preclinical candidates, one for non-Hodgkin’s lymphoma and one for breast cancer.
The company’s lead compound, DXL625 has undergone safety and efficacy testing in a rabbit model of non-Hodgkin’s lymphoma that expressed the CD20 target at only moderate to low levels. “The literature has shown that 48 to 49 percent of patients with non-Hodgkin’s lymphoma have moderate to low CD20 expression,” Dr. Morhet explained.
InNexus is also exploring the development of antibody-based therapeutics for immune modulation in diseases such as rheumatoid arthritis.
Mining the Knowledge Bank
Before PepBank, “there was no single source for peptide sequence data, especially for synthetic, nonbiological peptides,” noted Timur Shtatland, Ph.D., senior bioinformatics scientist at Agencourt Biosciences, a Beckman Coulter company. Researchers using phage display to develop peptide sequences did not have a peptide repository analogous to the large DNA repositories being developed.
While working at the Center for Molecular Imaging Research (CMIR) at Massachusetts General Hospital, Harvard Medical School, Dr. Shtatland was involved in developing molecular imaging and diagnostic agents, many of which were peptide-based and were attached to radiolabels, conjugated to magnetic nanoparticles, or used as enzyme substrates.
Dr. Shtatland and his colleagues at CMIR created PepBank and began the task of developing algorithms designed for automatic mining of peptide sequences in texts and entering them in a standardized format (using a one-letter amino acid code) with accompanying information and abstract citations retrieved from MEDLINE.
The ability to do a sequence similarity search on such a large peptide database “is what truly makes PepBank unique,” added Dr. Shtatland.
When he developed PepBank, Shtatland envisioned one type of application in particular: following phage display selection, users could enter their sequences in bulk and search PepBank for sequences that had already been found and that tend to pop up repeatedly in unrelated experiments. In this way, it would be possible to eliminate upfront peptide sequences known to be nonspecific and associated with either lack of efficacy or an increased risk of toxicity.
PepBank incorporates peptides from UniProt and from the Artificially Selected Proteins/Peptides Database (ASPD). Following the creation of PepBank, subsequent additions to MEDLINE were automatically downloaded and searched, and new peptide sequences and related information were added to the database.
PepBank’s improved search interface allows users to search for peptides related to particular diseases or biological areas, such as cardiovascular disease, cancer, diabetes, angiogenesis, or apoptosis.