|Home | About | Journals | Submit | Contact Us | Français|
The conference, which was organized by Visiongain and held at the BSG Conference Center in London, provided an excellent opportunity for participants to exchange views on the development, production and marketing of therapeutic antibodies, and discuss the current business environment. The conference included numerous interactive panel and group discussions on topics such as isotyping for therapeutic antibodies (panel chair: Nick Pullen, Pfizer), prospects for fully human monoclonal antibodies (chair: Christian Rohlff, Oxford BioTherapeutics), perspectives on antibody manufacturing and development (chair: Bo Kara, Avecia), market impact and post-marketing issues (chair: Keith Rodgers, Bodiam Consulting) and angiogenesis inhibitors (chair: David Blakey, AstraZeneca).
The first day was dedicated to discussion of antibody development and engineering, as well as debate on use of various types of antibodies. The session was chaired by David Blakey (AstraZeneca). The day opened with an overview on global trends in the antibody development and probabilities of approval success for human and humanized monoclonal antibodies (mAbs). The speakers then provided insights into the engineering and development of new therapeutic antibodies. Prospects for novel antibody formats, and assessment of immunogenicity, stability and aggregation risks in the development of therapeutic antibodies through use of in vivo and in silico methods were reviewed.
Global trends in antibody development were discussed by Janice Reichert (Tufts Center for the Study of Drug Development and Editor-in-Chief, mAbs). Dr. Reichert emphasized the increased focus on mAbs as therapeutic agents. Of the therapeutic proteins entering clinical study each year, the majority are mAbs. Major pharmaceutical firms are acquiring biotechnology companies to enter this market and new solutions to problems of immunogenicity, stability, affinity, specificity and production are being developed. The research on clinical pipelines undertaken at Tufts CSDD allows calculation of metrics such as clinical development and approval times and probabilities of approval success. Insights gained from these results are important for strategic planning.
The cumulative approval success rate for humanized mAbs was 16% for candidates entering clinical study during 1988 and 2008, and 29% for candidates entering clinical study during 1988 and 1997.1 A conservative estimate of the success rate for humanized monoclonal antibodies would be somewhere in between, at approximately 20%. The trend, however, is toward fully human monoclonal antibodies. There are currently two marketed human mAbs, with another four in regulatory review. The cumulative US approval success rate for human antibodies is currently low, but will rise to 18% if the four in regulatory review are approved.
In terms of therapeutic categories, oncology mAbs comprises approximately 50% of the total. Of 228 oncology mAbs that have entered clinical study since 1988, 56% are currently in clinical development. By comparison, 125 immunological mAb therapeutics have entered clinical study since 1990, of which 54% are currently in clinical development. The cumulative success rate for humanized oncology and immunological mAbs is 15% and 20%, respectively. Other therapeutic categories are being considered, including infectious disease. Sixteen anti-infective mAbs are currently in clinical study and one anti-infective mAb (palivizumab) has been approved to date. Oncology and immunology mAbs exhibit similar patterns for phase lengths and transition probabilities. The phase transition probability for phase 1 to 2 is high, followed by a lower phase 2–3 transition probability due to a proof-of-concept barrier. The transition probability for phase 3 to approval is comparable to that of phase 1 to 2.1
Other interesting trends include an increasing emphasis on antibody fragments.2 Fragments may be easier and less costly to produce, but have shorter circulating half-life compared to full size antibodies and no effector functions unless this is added. Also worth noting is the growing prevalence of modified versions of mAbs (glycosylation and Fc region engineering) and improvements on circulation half-life through PEGylation.3,4 Production methods as well as development and approval pathways for mAbs are well established and marketing approvals are set to increase if success rates are consistent with previous rates. This, together with competitive R&D times and potentially large markets, makes mAbs attractive for development as therapeutics.
Julian Burke (Genetix) presented a clinical update on the selection of cell lines for antibody expression and protein production. A hybridoma is a hybrid cell that has been engineered to produce a desired antibody in large amounts. ClonePix FL is an antigen based system for in vitro detection and selection of hybridomas. The system incorporates plating hybridomas into a 3D cell matrix-a method which was first described 25 years ago.5 Whilst this method is not new, the novel aspect of the ClonePix system lies in the screening and collection of only those clones secreting a specific antibody. There are two options for screening hybridomas: immunoglobin G (IgG) secretion assays and antigenspecific assays. Unlike IgG secretion assays, antigen-specific assays isolate only antigen-specific clones with the desired IgG isotype. The system can also optimize production through detection of the highest producing cell lines. This approach allows the production of 10,000 clones in three weeks compared to the conventional approach which produces approximately 1,000 clones in two months. After a few days growth post-selection, isolated clones can be rapidly re-screened for cell-line stability. This stability test can be run in parallel with the scaling up of clones, thus making the process highly time efficient. To summarize, ClonePix minimizes the labor requirement, shortens the process timeline and permits parallel interrogation of multiple antigens.
Masa Fujiwara (Chiome Bioscience) described the generation of antibodies using a novel antibody-generation technology called the ADLib (Autonomously Diversifying Library) System. This is a selection technology system based on cell-cell interactions and surface-displayed antigens in their native conformations. The system provides high-affinity antibody generation against difficult antigens such as self/human/homologous antigens, GPCRs, sugars/lipids, haptens and pathogens.
Dr. Fujiwara explained how the ADLib system can be used to generate specific monoclonal antibodies using a chicken B-cell line (DT40) that undergoes gene conversion at immunoglobulin loci. This gene conversion is enhanced by treatment of the cells with trichostatin A, a histone deacetylase inhibitor. DT40 cells that are specific to the target antigen are obtained through ‘fishing’ the ADLib library with antigens conjugated with magnetic beads. This selection process and the subsequent screening for specificity can be completed in approximately one week.6,7 As such, one of ADLib's attractive features is the system's ability to develop diverse monoclonal antibodies within weeks, not months.
Optimization of antibodies, with a focus on structure-function relationships, was discussed by Bryan Edwards (MedImmune). Complementary determining regions (CDRs) are found in the variable domains of antibodies and confer the antigen specificity of the molecule. The CDR regions show high levels of natural sequence variability and are targeted during somatic hypermutation to generate higher affinity antibodies to the antigen. When considering strategies for in vitro optimization of antibodies, the CDR regions are typically targeted for mutagenesis. Three CDRs (CDR1, CDR2 and CDR3) are found on both the heavy and light chain regions of an antibody, and the highest level of natural sequence variability is found in the heavy chain CDR3 domain.
Dr. Edwards described the optimization of two lead antibody candidates, where the heavy and light chain CDR3 domains were randomized at all amino acid positions and higher affinity variants isolated by both phage and ribosome display. Further gains in antibody potency were then obtained by combining the beneficial amino changes introduced into the heavy and light chain CDR3 domains. Additional sequence space was also explored outside of the CDR3 regions by the generation of error-prone libraries, and subsequent selection by ribosome display. Lead antibody candidates were improved several thousand-fold in potency through a combination of these approaches.
By studying the solved crystal structures of Fab:antigen complexes, Dr. Edwards explained that not all CDR regions make direct contact with the target antigen and that amino acids in the framework regions can also contribute to antibody specificity. Moreover, beneficial changes introduced during optimization are not always due to the introduction of new contacts with the target antigen. Several amino acid changes can indirectly improve the potency of the antibody despite being some distance from the antigen binding site. It has been postulated that these amino acid changes improved affinity by reducing the free energy of the antibody:antigen complex, for example by stabilizing CDR loop conformations or by improving the stability of the VH-VL interface.
Pavel Bondarenko (Amgen) presented data on the structure and function of disulfide isoforms of the human IgG2 subclass. There are five different known human antibody isotypes; IgA, IgD, IgE, IgM and IgG, of which IgG is the isotype that provides the majority of immune responses against pathogens. IgG antibodies have predictable properties, controlled function and long circulation half-life. Due to these properties, IgG antibodies are the most common therapeutic modalities. There are four human IgG subclasses, IgG1, IgG2, IgG3 and IgG4, of which IgG1 is the most abundant.
One therapeutic function of monoclonal IgG antibodies involves binding Fab regions to target receptors, which blocks ligand-receptor interaction. Additional functions include initiating cell destruction through the attraction of immune complexes by the Fc and hinge region on the antibody. Due to their lower affinity for Fc receptors than IgG1s, IgG2s show reduced propensity for activating immune responses. This may be beneficial in some therapeutic aspects.
Scientists at Amgen have recently discovered that structural heterogeneity is a naturally occurring feature of human IgG2 antibodies.8,9 These distinct IgG2 forms are due to differences in the disulfide connectivity at the hinge region. There are three human IgG2 isoforms; IgG2-A, IgG2-B and IgG2-A/B. IgG2-A is defined by structurally independent Fab domains and hinge region. In IgG2-B, on the other hand, both Fab regions are covalently linked to the hinge. IgG2-A/B is an arrangement in which only one Fab arm is covalently linked to the hinge through disulfide bonds.
The disulfide isoforms may show differences in potency. Against a cell-surface receptor, IgG1 and IgG2-A forms of a mAb were shown to have approximately similar potency and both had greater potency that IgG2-B. The difference between the IgG2 isotopes was attributed to the greater flexibility of IgG2-A and its ability to bind with both Fab regions. IgG2 disulfide exchange is facilitated by the close proximity of cysteine residues at the hinge region of IgG2. The mutation of a single cysteine residue in the IgG2 hinge region resulted in a loss of disulfide heterogeneity.10 Also, redox treatments with a cysteine/cystamine mixture have been shown to cause enrichment of both IgG2-A and IgG2-B.9 Human IgG2 isoforms are dynamic and exhibit disulfide rearrangement in both blood and cell culture.11 Initially, IgG2 exists as IgG2-A, and is then rapidly converted to the asymmetric IgG2-A/B, followed by a slower conversion to IgG2-B. The biological relevance of IgG2 isoforms and in vivo conversion is currently being studied.9
Andrew Popplewell (UCB New Medicines) provided an introduction to the prospects for therapeutic antibody fragments. Out of the 22 currently FDA approved monoclonal antibody therapeutics, three are antibody fragment therapeutics. Antibody fragment formats include Fab regions, single chain variable domains (scFv) and variable loops on the heavy and light chain (dAbs).
Antibody fragments are highly flexible formats that may be combined to form new multivalent or multi-specific structures. Compared to full-length IgGs, fragments may have improved biodistribution, tissue penetration,12,13 target access, or potentially better safety profiles due to the lack of Fc regions. Antibody fragments can also be expressed in microbial systems. They also show shorter serum persistence, which, depending on use, can be either disadvantageous or advantageous. However, antibody fragment circulation time can be modified either through PEGylation,3,4,14 or through use of serum proteins as carriers.15 Either strategy may alter the pharmacokinetic properties of fragments, allowing the infrequent therapeutic dosing commonly used for full-length IgG therapy.
Stability is a major factor for the successful commercialization of antibody-based drugs. Fabs are generally more stable to thermal or physical stress compared to IgGs, and this stability is not affected by PEGylation or by antibody species origin (e.g. mouse, rat, human). ScFvs and dAbs typically exhibit reduced stability compared to Fabs, however advances in technology may contribute to improve stability in these fragments.
Dr. Popplewell emphasized that a more complete understanding of biophysical properties and improved stability engineering are required before antibody fragments can reach their full potential. Full-length IgG1s offer active immune cell recruitment, and may thus be better suited for certain therapeutic uses, such as oncology treatments. The development of products with enhanced Fc functionality, the option of using inactive Fcs, and improvements in yields from mammalian cell expression systems are providing further options for the full length IgG format. Dr. Popplewell summarized by pointing out that the intended mechanism of action should guide the choice of format.
The challenge of predicting immunogenicity in a potential drug was discussed by Phillipe Stas (Algonomics). This is a difficult, yet important, process because early and precise immunogenicity assessments can reduce the number of drugs that fail to demonstrate efficacy in clinical trials. In order to generate an accurate risk profile of the potential immunogenicity for a given drug, two questions in particular must be addressed. First, the probability of observing an immunogenic response must be analyzed. Second, the severity of the observed immunogenicity needs to be considered.
Neutralizing antibody responses can neutralize not only the therapeutic protein, but also its endogenous counterpart; the latter may induce severe side-effects in the patient.16 Risk factors for immunogenicity include the degree of ‘non-self’ of the antibody, the dosing of the drug (acute versus repeated), route of administration (intravenous versus subcutaneous) and other drug characteristics such as clearance rate of the drug. The patient's immune status and the properties of the disease (i.e., severity and availability of concomitant immunosuppressants) should be included in a risk analysis.
The different drug development stages offer opportunities to use different strategies for immunogenicity assessment. In the clinical phase, the general approach is to conduct antidrug antibodies (ADA) screening on individuals exposed to the drug. However, efforts are now geared toward assessing immunogenicity at earlier stages. The generation of ADA is dependent on the presence of T-cell epitopes. These can be measured in the preclinical setting using in-vitro T-cell assays. Prior to this stage, in silico methods may be used to identify T-cell epitopes. One such T-cell epitope screening tool is Algonomics' platform Epibase®, which can rapidly analyze and predict the potential immunogenicity of therapeutic protein leads. An in silico approach to T-cell identification can offer relatively inexpensive mapping of epitopes from a wide genetic background.17 Also, perhaps more importantly, the combined use of in vitro and in silico tools allows for a much more accurate and less time-consuming assessment of expected immunogenicity in a drug.
In silico methods in the development of therapeutic antibodies were reviewed by Jesús Zurdo (Lonza Biologics). Dr. Zurdo focused on assessment of stability and aggregation risks. In addition to increased production costs, aggregation reduces product stability, increases immunogenicity and may also elevate the toxicity. AggreSolve is an in silico protein analysis platform that can be applied to predict and overcome protein stability and aggregation issues. The Aggresolve platform assesses protein aggregation propensity and identifies aggregation ‘hot-spots.’ The platform can also predict sequence changes that are likely to reduce aggregation propensity. This library of potential substitutions can then be used to re-engineer antibodies with elevated stability and fewer aggregation problems. Compared to wild-type, selected re-engineered molecules achieve significantly reduced aggregation levels whilst retaining biological activity. Furthermore, protein stabilisation using re-engineering methods can also translate into elevated antibody productivity.
Previously published online as a mAbs E-publication: www.landesbioscience.com/journals/mabs/article/8956
The second day's proceedings opened with remarks from David Palella (BioScience Ventures) who focused on the near-term future of the industry. Mr. Palella suggested that the biggest influence on the industry was likely to be merger and acquisition activity by big pharmaceutical firms that need to fill pipelines efficiently and effectively. He also argued that the current recession could take the form of a double-dip, i.e., two periods of recessions with a shortlived recovery between them, and potentially last until 2011 or 2012. To survive in this environment, companies should do more business development, marketing and sales. He urged companies to consider all deals since any deal is a good deal right now and to take funding opportunities where they could. He also remarked that there is now a reverse trend toward in-sourcing (rather than outsourcing), which protects jobs and drives companies' internal development.
Mr. Palella predicted that the dollar index would go down dramatically in the next 18 months, perhaps declining by 20–40%, thereby making assets and products in Europe very expensive. With a strengthening of the yen, this would put Japanese and UK companies selling to the US at a disadvantage because these companies would find it harder to sell products or partner drugs. To mitigate this, UK and Japanese companies could set up manufacturing or other operations in the US, and keep their costs in dollars.
While discussing reimbursement and cost effectiveness issues, Mr. Palella suggested that the US was likely to establish a body similar to the UK's National Institute for Clinical Excellence (NICE), which decides the cost benefit of certain medicines. He also speculated that personalized medicine could cut mAbs sales. In concluding, he remarked that the news was not all bad and that the discovery of novel targets would continue to expand the possibilities for mAb development. Despite challenges, and given the enormous opportunities, therapeutic antibody development is the still the best healthcare area to be in.
A primer to successfully patenting antibodies was provided by Louise Holliday (D. Young & Co.). Dr. Holliday explained that protection can be sought for antibodies using four main types of claim: target, sequence, deposit and activity. If claiming by sequence, either the sequence of the whole antibody, heavy chain (VH), light chain (VL), complementarity-determining regions (CDRs) or mutations relative to the parent can be included. However, problems involving use of antibody fragments can arise if too great a portion of the sequence is included. When using a sequence definition, the best approach it to include a multitude of definitions of different breadth as fall-back positions, to obtain the maximum possible protection in each jurisdiction. When the sequence is unknown, innovators can claim by hybridoma, and deposit a sample at a recognized institution. If claiming by target, companies can define their patent by target antigen, e.g., a polypeptide comprising a certain sequence; an antibody capable of binding specifically to a polypeptide, or by qualitative activity, e.g., the antibody binds specifically to one target, but not another. If claiming by activity, patents can be defined, for example, by affinity constants, or comparable binding activity.
Sequence and deposit-type claims are sometimes called species claims, while target and activity claims are sometimes called genus claims. Of the 148 European antibody patents granted in 2008, approximately 60% were of the narrower species type and about 40% were based on genus-type definitions.
To be patentable, the claims in a patent application must satisfy a number of requirements. The subject-matter of the claims should be novel, have an inventive step, be sufficiently described in the application and, for US claims, comply with the “written description” requirement. In Europe, the test for inventive step (equivalent to obviousness in the US) is would the steps leading to the invention be obvious to try, with a reasonable expectation of success? To overcome an inventive step objection, the best approach is to focus on any unexpected or unpredictable advantages of the antibody, e.g., differential binding to different antigens, cross reactivity. In order to comply with the sufficiency requirement, the patent application must contain enough information for a person to be able to work the invention. The written description element requires an applicant to show that, at the time of filing, the inventor had possession of the full scope of the invention claimed.
Dr. Holliday remarked that it was important to include as much characterizing information on the antibody, and as much supporting data, as possible when the patent application is filed. It is likely that attention will be drawn to specific advantages of the antibody in order to establish that it is inventive, and so evidence or data should be included in the text to illustrate these advantages. In summarizing, Dr. Holliday emphasized that companies should consider potential obviousness and written description issues before filing, include multiple definition types for the antibody, focus on activity-based definitions to maximize claim scope, and include the necessary supporting data to increase the probability of patenting success.
In a lively round-table discussion on market impact and postmarketing issues for monoclonal antibodies, the panel argued that antibodies could become victims of their own success. Rituxan, the first blockbuster (sales over US $1 billion) mAb, was launched in 1997 and mAb products already account for 10% of the annual US healthcare budget. This achievement in such a relatively short time has been driven by the high selectivity for disease targets, which makes mAbs very efficacious, and their relatively good safety profile. However, in a climate where governments are aiming to slash their healthcare budgets, the panel concluded that the industry could see pricing pressure for this expensive drug category. The panel also highlighted the fact that mAb products may face generic competition in both the EU and US because EMEA is examining the question of requirements for approval of mAb biosimilars and the US government is examining legislative pathways for biosimilar drugs.
Simon Russell (Novartis) warned that as companies transition from focusing on cancer and rheumatoid arthritis to developing therapies for the trickier and much larger indications of Alzheimer disease and pain, the cost of the drugs, which often cost tens to hundreds of thousands of dollars a year, could alarm already costconscious payers. At the moment Medicare is the largest payer in the US for mAbs. However, Janice Reichert (Tufts Center for the Study of Drug Development) cited research showing that 85% of approved mAbs have restrictions in the part D payment system, meaning that there are conditions imposed on reimbursement.1 If more products are approved, the increased cost may seriously conflict with the government's stated aim of reduced spending. Tim van Hauwermeiren (Argen-x) argued that lack of innovation in the industry might play a part in reducing prices. He noted that more and more ‘me-too’ products are coming onto the market, particularly in oncology and they will not be able to command the prices that innovative, first-approved products achieve.
The panel agreed that the advent of biosimilars in the US might have a marked effect on pricing. Two legislative acts that differ in their terms of data exclusivity are currently being considered in the US. The panelists suggested that a seven year exclusivity period is likely. One reason for the growing interest in biosimilars, as demonstrated by the recent acquisitions of biosimilar assets by both Merck & Co and Teva, is the size of the biologics market, which recorded sales of $108 billion in 2008. At present, some believe that the price differential between innovator protein therapeutics and biosimilars will be in the 25–30% range, but Dr. Russell argued that the reductions could be as great as 50%. However, he added that generic companies could see push-back due to rapid adoption of new drugs from physicians concerned about the lack of interchangeability with biosimilars. With the current threat of biosimilars, Julian Burke (Genetix) emphasized that companies will need to speed up their routes to market to ensure that products have as much patent life as possible once they are approved.
On a more positive note, Alasdair Stamps (Oxford Bio-Therapeutics) noted that advances in diagnostics, and the subsequent selection of patients to fit new and exciting mAb therapies, could benefit the mAb industry. There are already trends toward tailoring specific treatments to specific patient groups. While this could have the effect of reducing the use of some drugs, Dr. Stamps argued that pricing levels could be maintained because the product would be more likely to be efficacious in the targeted patient population. For example, in Europe cetuximab's use as a first line therapy is only in patients who test positive for the KRASwild type gene.
Despite challenges that face the industry, the panel agreed that mAb development was still an attractive one, as demonstrated by the continuing success of drugs such as bevacizumab. The area could easily expand with the growing understanding of how mAb therapies work and the possibility of use of multiple biologics in patients. They also noted that innovative payment structures, e.g., Johnson & Johnson's agreement to refund payment for bortezomib according to its efficacy, could overcome some of the concerns over drug costs.
Chris Gregory (Immunosolv) discussed novel opportunities for antibody-based therapies that target cell death. He spoke about the strong correlation between levels of apoptosis and levels of macrophage infiltration. The macrophage-mediated clearance pathway is not available in vitro and apoptosis occurs in cell cultures despite steps to minimize the generation of non-viable cells. Due to the absence of phagocyte cells, cultures can become necrotic, and can also degenerate due to sheer stress. The presence of non-viable and dead cells also has inhibitory effects on cultures because apoptosis affects by-stander cells. Cell membrane changes can be used to detecting dying cells starting from a very early stage of the process. A common method involves detection of phosphatidylserine on the exterior of the cell. This molecule is normally found inside the cell, but is shifted to the exterior during apoptosis. Currently, annexin V is commonly used to detect apoptosis, but high levels of calcium are required in the binding buffer. Immunosolv has developed an antibody, Dead-Cert™ imab6, that does not need a special binding agent to detect phosphatidylserine, and can be used when calcium levels are too low for use of annexin V.
The company has developed a novel tool, Dead-Cert™ Nanoparticles, for removal of non-viable cells. These paramagnetic nanoparticles bind to membrane structures that are not accessible on viable cells, and thereby bind selectively to apoptotic and necrotic cells, and to cell debris. A simple magnetic separation removes non-viable cells and debris bound to Dead Cert™ Nanoparticles from the desired viable cells. The technique enhances the viability of cells and can increase the productivity of cultures. Other benefits include increased sensitivity of cell-based assays and improved transfection efficiency.
Scientists at Immunosolv are also investigating the effects of dead cells on tumor growth. Previous studies have indicated that irradiated cells aid tumor growth. Immunosolv reasoned that ‘altruistic’ apoptosis of tumor cells might recruit macrophages, but inhibit granulocytes. The company is targeting the cell surface of apoptotic cells to change receptor usage by macrophages, leading optimally to pro-inflammatory, immunostimulatory responses. Dr. Gregory noted that Immunosolv is also targeting lactoferrin, which is released by apoptotic cells. The presence of lactoferrin could explain why granulocytes fail to phagocytose apoptotic cells.2 Removal of lactoferrin might allow granulocytes to enter tumors and induce tumor destruction.
In the panel discussion about current and future prospects for angiogenesis inhibitors, Simon Barry (AstraZeneca) pointed to wide-spread evidence of the effectiveness of targeting vascular endothelial growth factor (VEGF) to inhibit angiogenesis, but also noted that the benefits are not long-lasting. Of the approved anti- VEGF agents, single therapy use is effective in a limited number of tumors, e.g., liver and renal cancers. For other cancers, VEGF inhibitors work better in combination therapy. Dr. Barry pointed out that VEGF inhibition was also linked with toxicities including hypertension, gastro-intestinal issues, hypothyroidism and neurotoxicity. When used in combination therapies, additional adverse events can include fatigue, diarrhea and bone marrow effects. Investigations into the limitations of VEGF inhibitors are currently focusing on changes in tumor cells; pharmacological changes, including how increases in VEGF and placenta growth factor (PIGF) levels could limit treatment efficacy and the role of stromal factors, particularly inflammatory cell infiltrates such as CD11b and GR-1 positive cells. The latter is especially of concern with bevacizumab treatment. Other issues being investigated include tolerability of VEGF inhibitors given in combination, differences in response between individuals, and the difference in the structure of blood vessels, particularly between those in animals used in preclinical models and those in humans with disease. Dr. Barry concluded by arguing that single use therapy for VEGFs would be limited to renal and liver cancer, and that the broader opportunity for VEGF agents was use in combination with cytotoxins. This would allow flexibility in altering drug regimens to overcome resistance mechanisms. Combination therapies might eventually achieve comprehensive suppression of the tumor vasculature by including inflammation therapies, stromal and cytotoxic agents. However, above all, tolerance profiles must be acceptable to keep patients on therapies.
After noting that the majority of angiogenesis treatments have focused on VEGF inhibition, Mats Hellström (BioInvent), discussed the need for treatments that overlap with, or are complementary to anti-VEGF drugs. BioInvent is examining the role of PIGF, which is widely expressed in tumors, but not adult tissues. Both vascular cells and tumor macrophages appear to be dependent on PIGF for survival, therefore blocking PIGF should cause tumor shrinkage. Anti-PIGF treatments have been shown to enhance the anti-tumor effects of membrane-bound VEGFR-2 and soluble VEGFR-2. As both PIGF and VEGF are involved in angiogenesis, a combination of two drugs targeting these pathways might show higher efficacy compared to single use. As an example, he mentioned that treatment with anti-VEGFR2 and anti-PlGF resulted in improved efficacy as assessed by reduction of tumor volume in a study of colon carcinoma CT26.
David Blakey (AstraZeneca) spoke about the combination of vascular disrupting agents (VDA) with anti-angiogenesis agents. Studies have shown that certain tublin binding agents can have selective activity on tumor endothelium due to its increased reliance on a tublin cytoskeleton rather than an actin cytoskeleton for maintainng endothelial cell shape. Normal vasculature has a more mature actin cytoskeleton, as well as an established basement membrane. Tublin-based VDAs destabilize the tublin in the tumor endothelium, resulting in the endothelial cells rounding up, leading to leakage, vessel congestion and subsequent tumor cell death because of loss of blood flow and nutrient starvation. In studies, tublin binding agent ZD6126, a small molecule inhibitor, has been shown to induce profound tumor necrosis within 24 hrs when administered at a 200 mg/kg dose. Tumor regrowth can re-occur from cells surviving in the tumor rim that are thought to receive nutrients from adjacent normal vessels unaffected by the VDA. This limits the use of VDA as single agent treatments. Inhibiting tumor regrowth could be achieved by treatment with both a VDA and anti-angiogenic agents in preclinical models. In the clinic, the tubulin binding VDA fosbretabulin given in combination with bevacizumab has shown sustained tumor blood flow inhibition in a phase Ib trial. Patients with advanced solid malignancies received fosbretabulin at 45 mg/m2, 54 mg/m2 or 63 mg/m2 every 14 days followed by bevacizumab at 10 mg/kg 4 hours later. Out of 16 patients, nine showed significant disease control.
Development of ESBA105, a single chain antibody fragment for the treatment of acute anterior uveitis (AU) was described by David Urech (ESBATech). Current steroid-based treatments for uveitis can cause multiple problems, including lens opacification and elevation in intraocular pressure. Steroids are also confined to the anterior chamber of the eye, but inflammation migrates to the back of the eye. Systemic TNFα inhibitors have shown potential as treatments for uveitis, but they are expensive and difficult to withdraw quickly when adverse events occur because the agents have long half-lives.
ESBA105 is a topical TNFα inhibitor currently in clinical studies. Preclinical data in rabbits showed that ten drops of ESBA105 given over one day resulted in concentration levels about 1,000-fold higher than those for the anti-TNFα control. The study also revealed higher levels of penetration in the back of the eye, suggesting ESBA105 might be effective in treating diseases of the retina. A study of laser-induced choroidal neovascularization in monkeys using anti-VEGF bevacizumab, anti-TNFα adalimumab, and saline as controls, and both topical and IV ESBA105 as treatment, showed that topical ESBA105 was statistically different from saline, but not significantly different from adalimumab, which reduced the incident of grade 4 lesions by 80%. ESBA105 demonstrated good systemic and local tolerability with only very mild adverse events in a Phase 1 study of 27 volunteers who received up to 16 daily topical applications. The candidate is currently in a Phase 1b/2a study to assess intraocular levels, biodistribution and reduction of post-surgical inflammation following cataract surgery and a Phase 2a study in acute anterior uveitis patients.
An overview of the development of ranibizumab (Lucentis) with particular reference to pharmacokinetics (PK) was provided by Peter Kuebler (Genentech). Ranibizumab was derived from murine A.4.6.1, an anti-VEGF mAb. The binding site sequence was inserted in a Fab framework that was then humanized and affinity maturated (140x) to yield ranibizumab as a 48 kDa Fab V2.3,4 The drug inhibits rhVEGF-induced HUVEC proliferation across subtypes, and inhibits VEGF-mediated vascular permeability.5 Ranibizumab was selected for age-related, wet macular degeneration due to its high affinity for multiple biologically active forms of VEGF, inhibition of VEGF-induced angiogenesis and vascular leakage and inhibition of choroidal neovascularization and leakage in a primate model.6
Preclinical ocular and systemic PK studies focused on retinal exposure for efficacy data and systemic exposure for safety data. One single dose IV, and five single dose intravitreous (ITV) studies, were conducted in rabbits; single and multiple IV and ITV studies were also conducted in cynomolgous monkeys. Ranibizumab demonstrated bi-phasic elimination after IV dosing (single bolus of either 1 or 4 mg) in cynomolgous monkeys, with rapid decreases in serum concentration (to less than 0.5 µg/mL) observed within 12 hours of injection.7 In contrast, a single 2.0 mg/eye ITV dose resulted in vitreous humor, aqueous humor and retina ranibizumab concentrations that remained greater than 10 µg/mL for 10 days. Data from animal studies indicated that the drug penetrated all layers of the retina to reach the choroid, bound to VEGF in the receptor binding region, and inhibited angiogenesis and permeability with minimal systemic exposure after ITV injection. The PK animal studies identified the advantage of low systemic exposure, but high site-of-action exposure, and established the relationship between drug concentrations in the retina, and aqueous and vitreous humors relative to each other, which later allowed estimation of all three concentrations from a single sample from humans.
Results from the animal studies formed the foundation of the population PK model used for clinical studies in humans. In developing this model, co-variates such as demographics, pathophysiological variations, blood panel values, use of concomitant verteporfin PDT and use of IOP-lowering agents were identified. Creatinine clearance was found to be the most significant covariate for explaining inter-individual variability for ranibizumab clearance. PDT was statistically significant in explaining the intersubject variability in the rate of ranibizumab elimination from the eye and absorption into the systemic circulation. The population PK model was used to support use of monthly dosing by predicting that this dosing regimen would maintain vitreous ranibizumab above IC95 and helped contextualize systemic exposure. In concluding, Dr. Kuebler noted that translational pharmacology can enhance the understanding of novel therapeutic platforms, and robust PK and PK/PD modeling of Fab therapeutics are useful tools throughout development.
In his presentation, Dr. Mark De Souza (Dyax) first mentioned Dyax's protease inhibitor pipeline, which has already produced DX-88. This candidate was based on human tissue factor pathway inhibitor scaffold, and is a selective inhibitor of plasma kallikrein. DX-88 has been studied in two Phase 3 studies of hereditary angioedema and is currently undergoing regulatory review in the US for this indication.
Dyax is also investigating matrix metalloproteinase (MMP) inhibitors. MMPs have been studied extensively in cancer and inflammation, but dose limiting toxicities have blocked development of hydroxamate-based MMP inhibitors. Dyax is specifically examining MMP-14 as a target. This 65.9 kDa protein is a membrane-bound, zinc-endopeptidase that regulates pericellular proteolysis and tumor functions such as invasion, migration and metastases. MMP-14 also has a role in vascularization and radioresistance of tumors. High expression of MMP-14 and MMP-2 is associated with poor prognosis in a number of cancers, including small-cell lung, bladder and ovarian cancer. Low expression of MMP-14 is a good indicator of survival in advanced colorectal and breast carcinoma, and metastatic gastric and breast cancers. Dyax's DX-2400, a highly selective, human anti-MMP-14 IgG1 antibody, inhibits MMP14 and thereby blocks ProMMP-2 activation in tumor cells, and reducing migration and invasion of endothelial cells. DX-2400 was selected due to its ability to bind to human MMP14, but also to mouse, rat and cynomolgous monkey versions, simplifying pharmacological and toxicology studies.
In preclinical models DX-2400 has been shown slow tumor growth in MDA-MB-231 breast cancer xenografts. Anti-tumor activity was demonstrated in Her2-negative breast cancers at a DX-2400 dose regimen of 10 mg/kg every other day; doxorubicin administered once a week was used as a positive control. After 63 weeks following tumor cell implantation, inhibition of tumor growth of the DX-2400 and doxorubicin-treated animals was similar, and both DX-2400 and doxorubicin demonstrated significant (p<0.05) inhibition of tumor growth compared to negative controls (PBS and control IgG).
In an additional preclinical study, DX-2400 at 1–10mg/kg doses was shown to slow tumor growth and impair vascularization of a breast cancer xenograft (p < 0.05) compared with placebo. DX-2400 also decreased total MMP activity and MMP-2 activation in tumors, and reduced the incidence of metastasis. These studies have validated the role of MMP-14 in tumor progression, and suggest the potential of DX-2400 for treatment of solid tumors.
Paulo Fontoura (Hoffmann-La Roche) outlined how monoclonal antibodies have provided a paradigm shift in the treatment of multiple sclerosis (MS) due to their higher potency, selectivity and efficacy. MAb targets with potential roles in MS include immune cells involved in plaque pathogenesis, e.g., T cells, B cells and cytokines, as well as molecules involved in remyelination and tissue growth and repair, e.g., Nogo-A and LINGO. However, a number of mAbs have not been successful as treatments for MS—some candidates failed due to immunogenicity, and one, ustekinumab, did not show efficacy in an exploratory MRI trial.8 However, targeting α4/β1 integrins, which inhibit migration of self-reacting lymphocytes, has proven successful.9 This approach ultimately led to development of a humanized IgG4k mAb, natalizumab, that was approved for treatment of patients with relapsing forms of MS.
Natalizumab blocks lymphocyte adhesion to the CNS endothelin and inhibits migration of CD4 cells. It may also have other relevant biological effects which are still unclear. The concept of remission in MS patients was thought unlikely before natalizumab entered studies. However, following a two-year Phase 3 trial, approximately 30% of patients were in remission during the study. After stopping therapy, disease activity has been shown to return in the first three months. The product was approved in the US in November 2004, but was voluntarily withdrawn from the market in February 2005 because several patients involved in clinical studies of natalizumab developed PML.10 This disease is caused by the John Cunningham virus (°CV), which is present in 80% of healthy adults but normally held in check by the immune system.11 As none of affected patients had been treated with natalizumab only, it was concluded that occurrence of PML could be due to drug interactions, and that natalizumab could be used as monotherapy. Natalizumab was reintroduced to the market in June 2006, although with a strict pharmacovigilance program. Since then, additional cases of PML have occurred in natalizumab-treated patients, including one in a treatment naïve patient. At present there is no simple test for PML, so a tentative diagnosis is based on symptoms, with drug suspension recommended and the need of a lumbar puncture to confirm. There is no approved therapy for PML.
Dr. Fontoura also discussed rituximab and other anti-CD20 mAbs as possible MS treatments. CD20 is expressed in pre-, mature and malignant B cells. B lymphocytes are implicated in MS as they are present in demyelinating lesions and meningeal B cell follicles. The presence of such follicles has been associated with progressive forms of the disease, and correlated with a worse prognosis. Phase 1 and 2 trials of rituximab have confirmed a beneficial therapeutic effect in total number of new Gd+ lesions in relapsing remitting MS.12 A phase II trial in primary progressive MS has been completed, and data presented at meetings has hinted at potential therapeutic efficacy; this would be a major accomplishment, given the current lack of effective therapies for these forms of the disease. Newer generation anti-B cell therapies that are currently in development may have potential in MS, including ocrelizumab, veltuzumab, afutuzumab and ofatumumab. However, Dr. Fontoura warned that past experience with mAbs has shown their multiple biological effects extend beyond their primary one; therefore, safety concerns should still be present in developing these therapies, especially with regard to the potential to induce diseases such as PML. The risk of this disease may be related to the mode of action of specific mAbs, as well as the underlying disease pathophysiology. The question of how long mAbs can safely be used as treatments for chronic diseases such as MS remains open, as well as what are the advantageous or detrimental consequences of discontinuing their use. Interestingly, cessation of some mAb treatments, e.g., anti-CD52 or anti-CD20, has been shown to lead to an increase in regulatory lymphocytic cell populations, which might be of therapeutic benefit.
The development of novel, non-immunogenic mAbs for prostate cancer was reviewed by Dr. Matthew Baker (Antitope). He noted that T cell epitopes are a major determining factor in antitherapeutic immune response, and explained that the consequences of clinical immunogenicity could include reduced efficacy due to neutralization of therapies, altered pharmacokinetics, allergic reactions including anaphylaxis, and autoimmunity, although the latter is rare. He noted that, in addition to the presence of T cell epitopes in the sequence of protein therapeutics, causes of immunogenicity could include impurities or aggregates, route of administration (subcutaneous administration most frequently associated with immunogenicity) and frequent administration of mAbs, especially in immunosuppressed patients.
In vivo, T cell epitopes are generated in a multi-step process that includes uptake of the antigen, antigen processing and T cell activation. Activated T cells support B cells, which generate antibodies. In the case of an immunogenic response to T cell epitopes that are part of the protein sequence of a therapeutic, B cells generate anti-therapeutic antibodies. T cell epitopes can be detected at the whole protein or peptide level by ex vivo T cell assays such as Antitope's EpiScreen™ technology platform. This technology involves synthesis of overlapping peptides derived from a therapeutic mAb or protein, and testing of these peptides individually against HLA-typed human blood samples to determine the precise location of T cell epitopes. Mapping the T cell epitopes indicates the frequency and location of T cell epitopes in the candidate mAb or protein. The magnitude of the T cell response to the T cell epitopes can be estimated, thus allowing determination of the immunogenicity potential of an antibody.
Dr. Baker discussed how deimmunization13 could prevent anti-therapeutic antibody responses. In an example of deimmunization, T cell epitopes in a humanized anti-epidermal growth factor receptor (EGFR) mAb were found using T cell assays, and specific anchor residues within the T cell epitopes were targeted for mutation. When injected into mice, the original humanized mAb generated an anti-therapeutic antibody response, but the response to the deimmunized mAb was similar to that of the control syngeneic mAb.
Several deimmunized mAbs or proteins are in clinical studies including MLN591 (anti-PSMA), ThromobView (anti-D dimer), Anthim (anti-anthrax protective antigen), and bouganin. A total of more than 550 patients have been treated with the candidates, and none have developed anti-therapeutic antibody responses. Dr. Baker reviewed Antitopes' progress on the development of the deimmunized anti-PSMA mAb J591, which was selected for clinical trials in hormone-refractory prostate cancers. The original mouse-based antibody was immunogenic in all patients. An epitope depleting mutation using human IgG1/kC was introduced to overcome immunogenicity while retaining full affinity. In a Phase 1 study, no immunogenicity was observed in more than 300 patients after up to four weekly IV doses with a maximum 343 mg/m2 of deimmunized J591.14 Phase 2 study data showed J591 localized to the original prostate cancer as well as metastasized cancers.15 In combination with a low dose IL-2, J591 stabilized PSA serum concentrations in prostate cancer patients participating in the Phase 2 study. Antitope is also developing an improved, second-generation anti-PSMA mAb, J415. The murine version shows improved binding to PSMA. In collaboration with Dr. Neil Bander (Cornell University), a humanized version was developed using composite human antibody humanization technology. This molecule is expected to have improved activity and stability compared to J519. J415 is due to go into clinical trials in the next 12 months.
The final presentation was from Robert Lutz (ImmunoGen), who provided updates on three mAbs in the company's clinical pipeline. These candidates, IMGN101, BT062 and trastuzumab- DM1 (T-DM1), are immunoconjugates, with maytansinoid used as the cell-killing agent. IMGN901 and BT062 are in Phase 1 studies as treatments for multiple myeloma (MM), and trastuzumab- DM1 is in Phase 3 breast cancer studies.
IMGN901 targets CD56, which was strongly expressed in about three quarters of MM patients tested (43/55 cases), and associated with poor prognosis. A Phase 1 trial of IMGN901 is being conducted in patients who had already had an average of six prior treatment failures and had been pre-screened for CD56 expression. IMGN901 was administered as a monotherapy, with weekly dosing for two weeks in a 21 day cycle. Dose escalation started at 40 mg/m2, and has been increased to 140 mg/m2 so far. The maximum tolerated dose (MTD) has not yet been established. Preliminary results indicate that eight patients' disease progressed and eight patients had stable disease. An additional three patients had minimal objective responses, with two of these achieving a drop in urine M component. IMGN901 also well tolerated, with patients remaining on the drug longer than other previous treatments.
In preclinical tests of IMGN901 (9 mg/kg IV) in combination with lenalidomide (100 mg/kg IP) and low-dose dexamethasone (1.5 mg/kg SQ), complete tumor regressions were observed in four of six mice, and all mice on triple combination therapy exhibited partial tumor regressions. Treatment with only IMG901, or a combination of lenalidomide and dexamethasone without IMGN901, did not result in tumor regressions. IMGN901 in combination with bortezomib was also highly effective in an OPM2 MM model; 6.6 mg/kg dose of IMG901 with bortezomib caused all the mice to be tumor-free.
BT062 targets CD138, the primary marker for MM. BT062 consists of chimerized mAb nBT062 linked to maytansinoid DM4, and is currently undergoing phase I dose escalation study in patients with relapsed or refractory MM. At present clinical data are not available, but the treatment appears to be well-tolerated. A preclinical single dose response study using MM xenographs indicated that a 24.8 mg/kg effectively controlled tumor growth for approximately 50 days. BT062 has also shown some ability to overcome the protective effects of growth factors IGF-1 and IL-6 and bone marrow stromal cells, and has shown efficacy against MM in the human bone marrow microenvironment as tested in a SCID-hu/INA-6 model.
Dr. Lutz also discussed progress in the clinical studies of T-DM1, which is in a clinical study program implemented by Genentech and Hoffmann-LaRoche. The candidate comprises trastuzumab (Herceptin) conjugated to DM1 through ImmunoGen's SMCC linker, and is being studied as a treatment for HER2-positive, metastatic breast cancer. Phase 1 data with a 3.6 mg/kg every three weeks dosing regimen showed progression free survival of 9.8 months (n = 15). Interim data from a Phase 2 study of metastatic breast cancer patients whose disease had progressed after trasuzumab and possibly lapatinib showed that 38% had a confirmed objective response rate. Final study results are due in the second half of the year. T-DM1 is also in a Phase 3 study currently enrolling up to 580 patients at 260 centers. Treatment with T-DM1 will be compared to treatment with the combination of lapatinib with capecitabine, and the study endpoint will be progression-free survival.
Previously published online as a mAbs E-publication: www.landesbioscience.com/journals/mabs/article/8956