IBC's 23rd Annual Antibody Engineering, 10th Annual Antibody Therapeutics international conferences and the 2012 Annual Meeting of The Antibody Society: December 3-6, 2012, San Diego, CA.

Session chair James D. Marks (University of California, San Francisco) gave the keynote presentation on a systematic approach to generate tumor targeting and internalizing mAbs. Such mAbs can deliver payloads into cells to achieve a therapeutic effect. He compared differences between ADCs and antibody-targeted immunoliposomes with respect to the amount of drug delivered, the potency of drugs used, and the size of the molecule. He noted specific features of immunoliposomes, including the impact of targeting on tumor deposition and micro-distribution. 1 Professor Marks then reviewed the generation of an internalizing anti-HER2 antibody F5 by selection of a naïve phage display antibody library on live cells, the construction of F5-targeted immunoliposomal (IL) doxorubicin, and its efficacy in preclinical studies. He discussed the factors that affect IL tumor localization and cellular uptake, including mAb affinity, valency, and epitope. 2,3 He also described a systematic approach to isolate mAbs from phage libraries to any cancer type using the basal subtype of breast cancer as a model. Methods were described to: (1) select internalizing mAbs; (2) identify the antigen bound by the mAb by using immunoprecipitation and LC/MS-MS; (3) predict the mAbs cognate antigen in silico when immunoprecipitation failed by cell line profiling; and (4) direct selections to a specific tumor antigen by combining cell selection followed by selection on yeast displaying the target antigen on their surface. Finally, Professor Marks introduced a project to develop a high-throughput platform to generate mAbs to type 1 and 2 membrane proteins using yeast display of the membrane protein for phage antibody library selection. Ulrik Nielsen (Merrimack Pharmaceuticals) provided an update on the development of HER2-targeted nanoliposomal doxorubicin (MM-302) using systems optimization of nanotherapeutics. In the Phase 1 clinical study of MM-302, complete regression of tumor lesion was observed after 6-mo treatment of HER2-positive breast cancer, and the drug was well tolerated. Dr. Ulrik Nielsen then reviewed the systems used to optimize nanoliposomal drugs, which include optimization of multiple factors such as the antibody copy number, antibody affinity, the degree of PEGylation, and encapsulated drug concentration. He discussed the computational models used to understand the therapeutic window (e.g., to avoid effects on the heart compared with tumor), explained how to predict the optimal affinity and copy number of antibodies for such therapeutic window, and how the low doxorubicin exposure to the heart DNA from MM-302 compared with the free doxorubicin was achieved. Using the computational simulation, the pharmacokinetics (PK) of liposomal drug can be predicted and carefully controlled, and the drug release rate and the drug responsive patients can also be identified. In addition to the computational modeling, Dr. Ulrik Nielsen introduced an experimental monitor and prediction of nanotherapeutic activity by nanoliposomal imaging reagent MM-Dx-929, ILs- 64 Cu, to show tumor deposition of immunoliposomes. Such an imaging method showed a high degree of variability in tumor deposition in human tumors and different tumor models. In a mouse model of BT474 tumor, MM-302 responses correlated with MM-Dx-929 deposition in tumors. Both computational and experimental results indicated that micro-distribution and cellular uptake of nanoliposomal drugs play critical roles in nanotherapeutics. The session co-chair, Dario Neri (Swiss Federal Institute of Technology Zurich) presented progress in the drug discovery using DNA-encoded chemical libraries. He first described the principle of the DNA-encoded chemical library, which consists of compounds covalently attached to coding DNA fragments, allowing identification of small molecule ligands to a target in a selection mode rather than screening individual compounds, one molecule at a time. The unique DNA sequences attached serve as amplifiable identification bar codes, which made it possible to identify the binding small molecule after the selection. He also reviewed the history of DNA-encoded chemical library from single compound to multiple compound complexes in the Encoded Self-Assembling Chemical (ESAC) library. In the ESAC library, more than one small molecule can be assembled as a multi-head binding compound complex, and encoded by corresponding DNA sequences incorporated in tandem by complimentary spacing of common DNA fragments. In such manner, the resulting multi-compound molecules that exhibit synergistic binding activity can be identified with enhanced binding specificity and affinity. 4 Professor Neri showed that multiple potent inhibitors with affinities ranging from 0.2 to 5 µM were identified for the tumor-associated antigen carbonic anhydrase IX (CA IX) and the pro-inflammatory cytokine interleukin-2 (IL-2). 5 In addition to the de novo identification of ligand-like small molecules to interrupt protein–protein interactions, DNA-encoded chemical approach can also be used for “affinity maturation” of sub-optimal ligand binding. By such means, several trypsin inhibitors with over 10,000-fold potency compared with the parental compound, and enhanced selectivity toward other serine proteases, were isolated from a DNA-encoded library based on benzamidine analogs. 6 In summary, the DNA-encoded chemical library offered a powerful and yet flexible approach to identify potent ligand-like small molecules for protein interaction interventions, and potential therapeutic drugs. Guoqing Chen (Igenica, Inc.) presented an innovative approach to develop anti-tumor antibodies. Dr. Chen began with a discussion of target “crisis” in cancer antibody drug discovery and the lack of reliable methods to identify effective target antigen/antibody pairs. He introduced Igenica’s technology platforms that enable a function-orientated approach for target antigen and antibody drug discovery. Igenica’s approach consists of both antigen and antibody discovery platforms, sTAg and iTAb. The sTAg (surface-tagged-antigen) discovery platform identified tumor associated cell surface antigens by coupling the pull-down and LC-MS/MS analysis of surface tagged protein from tumor samples. By comparing the surface proteome profile and quantities of multiple tumor samples, the relevant tumor antigen can be identified and used for animal immunization. The iTAb (in vivo anti-tumor antibody) screening platform uses in vivo tumor rejection as a criterion to identify anti-tumor antibody repertoire among the immunized mice, followed by functional antibody isolation and validation. Using the sTAg and iTAb platforms, both “Herceptin-like” and “Perjeta-like” anti-Her2 mAbs were isolated, indicating the feasibility of this novel approach. Dr. Chen also showed that a novel antigen/antibody pair, AML-01/IGN523, was discovered to treat AML, CLL, MM, and other cancers with stronger potency than existing antibodies. AML-01 is a single pass, type II transmembrane protein that overexpresses on multiple tumors; such overexpression leads to tumor formation and correlates with tumor progression and metastasis. This antigen, however, has no prior investigational reports, which suggests that Igenica’s approach is a unique and powerful tool for tumor associated antigen discovery and potent anti-tumor antibody identification. IGN523 is a humanized IgG1 mAb and Igenica plans to initiate a Phase 1 clinical study for this therapeutic candidate in 2013. Susanne Gräslund (Structural Genomics Consortium) discussed an effort to provide open access to all data and reagents associated with epigenetic processes. This protein-based consortium currently includes over 1300 PDB entries of 3D structures, protein expression clones, protocols for expression and crystallization, other protein-based protocols, as well as chemical probes and antibodies. First, she discussed the critical questions in epigenetics, such as which proteins and protein domains are responsible to read, write and erase histone marks; how they recognize histone marks specifically and how they assemble into functional multi-protein complexes; which genes are regulated at the transcriptional level by the complexes; how these complexes and functions differ between normal and disease states; and whether these epigenetic targets can be modulated for therapeutics. Then, she presented their progress in identification of epigenetic targets, which include 72 for acetylation, 266 for methylation, and 32 for poly ADP ribosylation. According to Dr. Gräslund, the specific aim of the project she presented is to generate renewable recombinant antibodies to the identified epigenetic targets. She described the project organization, which consists of multiple research institutes, including Structural Genomics Consortium, University of Toronto, University of Chicago, and Technical University, Braunschweig. The central goal is to generate renewable recombinant antibodies to the selected epigenetic targets. She also described a multi-step workflow that carries antigen production, Fab selection and production, initial validation, full-length IgG production, further validation, to final delivery of antibodies via Life Technology Corporation. Technically, by cloning the target antigen into vector pNIC-Bio2, which contains a biotinylation sequence and co-expresses BirA ligase, the expressed antigens will be biotinylated in vivo for easy immobilization during selection from the phage-displayed Fab library. A high throughput antigen production system was established using a two-step method (IMAC+Gel filtration) on AKTA Xpress, which allows efficient production of antigen protein in large quantities. The identified Fab sequence was cloned from phage vector into pSFV4 vector, which contains the optimized triple tac promoter, PelB leader for Fab secretion, and heavy chain C-terminal biotinylation tag. The IgG conversion and production use transient transfection of HEK293F cells and pFUSE vector system. The efficient production of target antigen, antibody fragment, and full-length IgG allows for antibody generation in a relative short time period. The identified antibodies were used to precipitate the target antigens; then mass spectrometry analysis was used to validate the binding specificities and identify the complex proteins. In addition, siRNA knockdown was used to confirm the antigen identity of the target antibody. Furthermore, the chromatin IP/ChIP was used to investigate genome wide protein-DNA interactions. The crystal structure of Fab/antigen complex was resolved to provide insight into structure and function relationships.

Janice Reichert (The Antibody Society, Reichert Biotechnology Consulting LLC) provided a brief update on the activities of The Antibody Society. Substantial contributions to the Antibody Engineering and Antibody Therapeutics meetings are made by Society members, who serve on the Scientific Advisory Boards, act as session chairs, give presentations, and contribute to preparation of the meeting report that is subsequently published in mAbs . She discussed the Society’s website ( www.antibodysociety.org ), which is regularly updated with antibody-related news (e.g., new marketing approvals for therapeutic antibodies), as a valuable resource for the antibody community and she invited participants to also receive news by joining the Society’s LinkedIn group or Liking mAbs on Facebook. The use of Guidelines on Information About Translational Experiments (GIATE) in the management of translational research data, an initiative promoted by the Society, was discussed by May Yong (University College London). Translational research is a process that involves turning innovative ideas into new health products. She noted that the presenters at the meeting each work on sections of translational research, including engineering new agents, determining the extent to which these agents interact with other molecules in a complex environment, testing of the agents for efficacy and safety issues, and measuring therapeutic effects in human. Large amounts of data with very diverse origins are generated during translational research. The data needs to be organized to enable complex information from every section of translational research to be linked and placed into context, and ultimately allow a coherent picture of an agent to develop. Data management also ensures that knowledge is not lost, resources are not wasted by unnecessary repetitions, and, most importantly, that data was understood and used as intended.1 The recording of data according to a community-agreed format (as exemplified by the uptake of MIAME 2 by the genomics researchers) gives scientists the ability to share data; it means that scientists would create data sets with consistent variable types, allowing them an equal platform with which to compare data. It is in this spirit that she, Richard Begent and other members of The Antibody Society published a list of minimum information types necessary to describe experiments to a useful level. 3 They proposed that these information types should be collected from every experiment, whether the experiment was conducted in the cell line or human tissue. They hope to introduce the practice of recording consistent data sets, which can be related to allow tracking of the decision-making rationale during the development process of an agent. Antibody Society members with different expertise, from physicists to molecular biologists and clinicians were consulted because the challenge in creating this list of minimum information lay in the diversity of data types to be recorded. Dr. Yong explained that their minimum information list is named Guidelines on Information About Translational Experiments (GIATE). According to GIATE, an experiment is described in terms of the molecular target, the therapy agent and the model in which it is performed. Different information types are collected in different models. For example, information about molecular interactions are collected in experiments in the molecular model, but not in the cell line, animal or human models. Pharmacodynamics data are collected in experiments using the animal and human model, but not in experiments conducted using cell lines. In cell model experiments, therapy effect data are recorded. There are many existing minimum information lists that describe specific areas such as diseases 4 or proteomic experiments 5 to great detail. GIATE is different because it covers many levels of investigations. The strength of this approach is not in the level of detail but in the way it allows investigation across models. In other words, it allows users to track what is known from cancer biology to the diagnosis and treatment. Dr. Yong noted that the aim is not to collect a high level of detail from each experiment. Instead, the goal is to provide researchers with summaries containing the same type of information. Researchers are also provided with the name and contact details of the data holder, from whom they can contact for more information. In addition to creating consistent data sets, GIATE aims to minimize ambiguity about what experimental data has been recorded. Society members are commonly involved in collaborations with researchers from different groups and countries. In addition, as the development process of a new drug can span many years, it is critical that experimental data and conclusions made by scientists who are no longer working on that project are included appropriately. In concluding, Dr. Yong noted that, because terminologies vary across groups, countries, expertise and timeframes, scientists who collect the data and the researchers who re-examine them must use a shared vocabulary. Therefore, the meaning of every information type requested by GIATE is clarified with semantics provided by well-known lexicons such as the National Cancer Institute Thesaurus. More information about GIATE can be found at the GIATE section of the Society’s website ( www.antibodysociety.org/data/datastandards.php ). An app for the Mac OSX platform can be downloaded from the site and used to assist in recording experimental data to the GIATE standard.

James S. Huston (The Antibody Society; Boston Biomedical Research Institute; Huston BioConsulting, LLC) chaired this session and introduced the speakers. William M. Pardridge (University of California, Los Angeles) gave the keynote presentation on targeted delivery of therapeutics across the blood brain barrier (BBB). 1,2 He provided a brief history of the failed attempts in the 1980s to deliver large molecules to the brain. All large molecules and most small molecules do not cross the BBB. He noted that experiments showing molecules uptake in the brain at 0.5% of the injected dose (I.D.) are not exhibiting transport across the BBB, but reflecting the volume of blood in the brain. The key to success for crossing the BBB is engineering a molecular Trojan horse. This approach takes advantage of endogenous receptors that transport across endothelial cells from the blood to the brain. These transcytosis receptors include the insulin receptor (IR) for mono-directional transport from the blood membrane to the brain membrane and the transferrin receptor (TfR) for bi-directional transport between the blood and brain membranes. Antibodies that bind to these receptors were fused to various therapeutic proteins to create the molecular Trojan horses. Professor Pardridge then presented three examples of approaches to crossing the BBB. In the first example, an antibody to human IR (HIRMAb) was fused to TNFR to treat pro-inflammatory aspects of many neurological diseases. 3 The uptake in monkey brains was shown to be 3% I.D., higher than TNFR-Fc, but similar to the uptake of neuroactive small molecules. The uptake in mouse brains was similar, although a surrogate TfR-TNFR was necessary because no mAbs to mouse IR were available. Neuroprotection with TfR-TNFR was observed in mouse models for Parkinson disease and stroke. In the second example, HIRMAb was also fused to a single chain variable fragment (scFv) anti-Aβ amyloid to treat Alzheimer disease (AD). Positive pharmacology was observed in transgenic AD mice. 4 In the third example, HIRMAb was fused to α-L-iduronidase (IDUA) to treat the lysosomal storage disorder Hurler syndrome. 5 The uptake in monkey brains was shown to be 1% I.D. A surrogate molecule TfR-IDUA was shown to be effective in a Hurler mouse model. All of these fusion molecules were also shown to be isolated to high purity and retain high binding affinity and/or enzyme activity. Ruben J. Boado (ArmaGen Technologies) continued the discussion of crossing the BBB by describing non-viral gene therapy to the brain via antibody-targeted nano-containers. 2 Danica Stanimirovic (National Research Council of Canada) presented the discovery and engineering of BBB-transmigrating antibodies targeting novel receptors that mediate transcytosis across brain endothelial cells. 6 Two workflows were used to find novel receptors: (1) a proteomics approach isolating the luminal membrane proteins of brain endothelial cells via gradient fractions or in situ capture and identification via mass spectrometry; and (2) functional panning of phage display single domain antibody (sdAb) libraries (llama VHH) against brain vs lung endothelial cells. 7 In addition to the transcytosis receptors mentioned in the prior talks, the membrane protein TMEM30A was identified from both workflows. FC5, a sdAb against TMEM30A, was isolated from workflow 2 and characterized. Pharmacological efficacy was demonstrated with systemically administered FC5 conjugated to different neuroactive peptides in rat chronic pain models. Enhancement in analgesia was observed when FC5 fused with Fc was used as a carrier for analgesic peptides. Most recently, a bispecific sdAb-sdAb was generated linking FC5 with an anti-amyloid- β peptide antibody with promising initial results of entry into the brain parenchyma. 8 The session concluded with discussion of antibody therapy for several neurodegenerative diseases. Cynthia A. Lemere (Brigham and Women’s Hospital; Harvard Medical School) gave an overview of pre-clinical and clinical results of passive amyloid-β (Aβ) immunotherapy for Alzheimer disease (AD) from her own and other academic laboratories or companies. 9,10 AD pathogenesis is associated with accrual of amyloid plaques of Aβ peptides that precede neurofibrillary tangles and cognitive decline. Administration of antibodies to different regions and forms of Aβ has been shown to decrease Aβ plaques in the brains of AD mouse models and AD patients in clinical trials; however, a lack of robust cognitive efficacy has been observed in Phase 2 or 3 trials of mild-to-moderate AD patients. Earlier administration may be necessary to prevent the disease. This hypothesis will be tested by treating individuals with dominantly-inherited AD-associated genes with passive Aβ immunotherapy this year in several international trials. 11 Passive immunotherapy against other targets, including pyroglutamate-3 amyloid-β protein (pE3-Aβ), BACE1, and Tau, has been shown to be effective in AD mouse models. In particular, recent work from Professor Lemere’s laboratory showed a reduction of amyloid plaques after administration of anti-pE3-Aβ in AD transgenic mice. 12 In addition, no microhemorrhage was observed, unlike prior anti-Aβ therapy in mouse models and humans. Anne Messer (Wadsworth Center, New York State Department of Health; State University of New York at Albany) presented progress developing intracellular antibody therapy for Parkinson disease. 13 Dr. Messer has worked on developing intrabody therapeutics (gene therapy of scFv or nanobodies against misfolded proteins) for numerous neurodegenerative diseases. For Parkinson disease, intrabodies were selected against peptides of α-synuclein using a human yeast surface display library. 14 The anti-α-synuclein scFvs were shown to bind α-synuclein and protect in an in vitro PD assay with efficacy correlated more with solubility of the scFvs than affinity. Binding alone, however, is not sufficient for a prolonged effect. Following work on anti-huntingtin intrabodies in Huntington disease, 15 the anti-α-synuclein intrabodies were fused to a proteasome targeting (PEST) motif to direct the antigens to the proteasome. 16 The PEST motif fusions not only reduced the amount of α-synuclein, but also increased the solubility of the anti-α-synuclein intrabodies.

Louis M. Weiner (Georgetown University Medical Center) organized and Kerry Chester chaired the last session of the conference, which was dedicated to discussion of immunomodulatory antibodies and ADCs as treatments for cancer. Mark Sliwkowski (Genentech) reviewed the development of anti-HER2 trastuzumab (Herceptin®), pertuzumab (Perjeta TM ) and trastuzumab emtansine, which is also known as T-DM1. HER2 is a member of the ErbB family, but, unlike the other members, it is always in an open conformation. Dr. Sliwkowski elucidated that this open conformation permits rapid interaction with other ErbB family members and explains why none of the 11 ErbB ligands binds to HER2. About 18–20% of all breast cancer patients have HER2 gene amplification. Before the approval of trastuzumab in 1998, HER2-positive breast cancer patients had about half the overall survival of HER2-negative patients. Dr. Sliwkowski pointed out that the treatment of early diagnosed breast cancer with trastuzumab increased the four-year disease-free survival (DFS) of patients from 67% to more than 85%. In 2006 the US Food and Drug Administration (FDA) expanded the approved use of trastuzumab to include treatment of early breast cancer. The antibody works by two fundamentally different mechanisms of action, inhibition of signaling and antibody-dependent cell-mediated cytotoxicity (ADCC). The signaling of HER2 is inhibited by blocking HER2 signal transduction and inhibition of HER2 shedding, which leads to a very active form of HER2. Although trastuzumab increased the survival of HER2-positive breast cancer patients, ~5,000 women die of HER2 positive breast cancer per year. Therefore, Genentech has developed second-generation anti-HER2 drugs. Pertuzumab was created to antagonize the ligand-dependent signaling of HER2 in complex with other ErbB receptors (like HER3) and to inhibit the formation of corresponding dimers. In xenograft models that are resistant to trastuzumab, a combination of pertuzumab and trastuzumab showed strong activity. Consequently, this antibody combination was tested in clinical trials. In the proof of concept trial ( NCT01674062 ; BO17929), the antibody combination was used to treat women that had tumor progression during treatment with trastuzumab plus chemotherapy. In this trial, the antibody combination achieved a 25% objective response rate. Another 25% of treated patients had stable disease after six month. Pertuzumab in combination with trastuzumab plus chemotherapy was then tested as a neoadjuvant treatment in the Phase 2 “NeoSphere” trial. The antibody combination with docetaxel was associated with a significantly higher pathological complete response rate (pCR) than trastuzumab plus docetaxel alone in patients with HER2-positive breast cancer. In this trial, pertuzumab plus docetaxel alone had a pCR of 24% that was statistically not different to the pCR of trastuzumab plus docetaxel (29%). As expected, given the short duration of therapy, the antibody doublet without docetaxel was less effective than the antibody-chemotherapy combinations. Pertuzumab in combination with trastuzumab was approved by the FDA for the treatment of HER2-positive metastatic breast cancer in 2012. Dr. Sliwkowski then discussed the development of trastuzumab emtansine (T-DM1), which was performed Genentech/Roche in cooperation with ImmunoGen. T-DM1 is an ADC consisting of the cytotoxic agent maytansine attached to trastuzumab. The stability of the linker between the cytotoxic drug and the antibody determines the ADC clearance in animal models. The clearance correlated well with the toxicity of the ADC. While trastuzumab is cytostatic, T-DM1 is cytotoxic and more effective than trastuzumab in xenograft models. In the Phase 2 TDM4450 study ( NCT00679341 ), the objective response rate of T-DM1 was similar to trastuzumab plus docetaxel, but patients treated with T-DM1 experienced much fewer adverse events than those who were treated with trastuzumab plus chemotherapy. Grade 3 or higher adverse events were reduced by almost 50%. In the Phase 3 trial EMILIA ( NCT00829166 ), T-DM1 achieved a significant improvement in progression-free survival (PFS) with an increased median from 6.4 mo in the control group to 9.6 mo in the T-DM1 group. T-DM1 was well tolerated with no unexpected safety signals. A statistically significant overall survival benefit was also observed (30.9 mo for T-DM1 vs. 25.1 mo for lapatinib and capecitabine). Based on this trial, Genentech has submitted a biologics license application for T-DM1 to the FDA; a first decision on the application by FDA is expected in February 2013. Charles Drake (Johns Hopkins University School of Medicine) detailed recent results of a neutralizing antibody against the receptor programmed death-1 (PD-1). PD-1 is an inhibitory receptor expressed on T cells, termed an immune ‘checkpoint’. Dr. Drake first described the importance of immune checkpoints and summarizing the clinical results of antibodies against CTLA-4, another immune checkpoint. While CTLA-4 knockout mice die at an age of 3–4 weeks due to massive lymphoproliferation, PD-1 knockout mice show no clinical symptoms. The CTLA-4 blocking antibody ipilimumab was approved by the FDA in 2011 for the treatment of patients with metastatic melanoma. Ipilimumab is the first drug that showed a survival benefit in patients with metastatic melanoma, indicating the potential of antibodies neutralizing immune checkpoints. Interestingly, blocking PD-1 with a monoclonal antibody does not increase the survival of mice in an established melanoma model. 1 The PD-1 blocking antibody MDX-1106, however, produced objective responses in patients with different solid tumors including renal cell carcinoma, melanoma, colorectal cancer and non-small cell lung cancer in Phase 1 studies. 2,3 In its first study, patients received a single intravenous infusion of MDX-1106 and a second and third dose only when a lack of progression was determined. 2 Because the antibody was well tolerated and associated with evidence of antitumor activity, it was next tested in a Phase 1b trial where the mAb was given every 2 weeks in 8 week (4 treatment) cycles. Patients received up to 12 cycles (over approximately 2 y) until disease progression or a complete response occurred. In this study, grade 3 or 4 drug-related adverse events occurred in 14% of patients, with three patients dying of pulmonary toxicity. 3 Objective responses were determined in ~20–25% of treated patients, with many of these responses durable in nature. In preclinical studies, anti-PD-1 antibodies in combination with radiotherapy increased the long-term survival of mice in a glioblastoma model compared with antibody treatment or radiotherapy alone. The combination of two neutralizing antibodies targeting two different immune checkpoints like PD-1 and LAG-3 (lymphocyte activation gene-3) or PD-1 and CTLA-4 have a synergistic anti-tumor effect in mice. The combination of MDX-1106 and ipilimumab has entered clinical studies for the treatment of unresectable stage III or stage IV metastatic melanoma. Taken together, these data support the notion of using blocking antibodies in combination treatment regimens. The clinical development of brentuximab vedotin was discussed by Megan O’Meara (Seattle Genetics, Inc.). CD30 is a member of the tumor necrosis factor (TNF) receptor family. It was originally discovered on Reed–Sternberg cells in 1982. Dr. O’Meara pointed out that Reed–Sternberg cells are the malignant driver cells for Hodgkin lymphoma (HL). The receptor is also associated with anaplastic large cell lymphoma (ALCL). CD30 has very limited expression in normal tissue, but is expressed on activated lymphocytes. Brentuximab vedotin (ADCETRIS ® ) is an ADC comprising an anti-CD30 monoclonal antibody attached by a protease-cleavable linker to the microtubule disrupting agent, monomethyl auristatin E (MMAE). The antibody and the cytotoxic agent are separated by an optimized linker that is stable in the serum, but releases MMAE after internalization and degradation of the antibody. MMAE blocks the polymerization of tubulin and inhibits cell division. Dr. O’Meara explained that MMAE is over 200 times more potent than, for example, doxorubicin and therefore cannot be used as a non-targeted chemotherapeutic agent. Brentuximab vedotin was tested in two Phase 1 studies where it induced durable responses with moderate adverse effects. Two Phase 2 pivotal studies were then initiated for relapsed or refractory HL and for relapsed or refractory systemic ALCL. Patients were treated with monotherapy at a concentration of 1.8 mg/kg every three weeks and received up to 16 cycles if they had stable disease or better. In HL a 75% objective response and 34% complete remission rate was achieved. In systemic ALCL, 86% of the patients achieved an objective response, 57% achieved a complete remission and 29% a partial remission. Across both trials, the most common adverse reactions (≥ 20%), regardless of causality, were neutropenia, peripheral sensory neuropathy, fatigue, nausea, anemia, upper respiratory tract infection, diarrhea, pyrexia, rash, thrombocytopenia, cough and vomiting. The product received approval from the FDA for the treatment of relapsed or refractory HL and ALCL in 2011. Brentuximab vedotin is currently in clinical trials for the treatment of other CD30-positive tumors, including solid tumors, and other non-Hodgkin lymphomas, including diffuse large B cell lymphoma (DLBCL). Brentuximab vedotin was recently tested in a Phase 1 dose-escalation study in newly-diagnosed advanced stage HL patients as a frontline therapy in combination with adriamycin, bleomycin, vinblastine, and dacarbazine (ABVD) or AVD, which omits bleomycin. The maximum tolerated dose was not reached; however, 11/25 patients in the ABVD cohorts had adverse events of pulmonary toxicity that led to discontinuation of bleomycin. No pulmonary toxicity was observed in the AVD cohorts. A 96% complete remission rate was observed at the end of frontline therapy with brentuximab vedotin plus AVD. A Phase 3 study is ongoing to assess treatment with brentuximab vedotin plus AVD vs. ABVD alone in treatment-naïve HL patients. In a recently conducted Phase 1 study for the frontline treatment of patients with mature T cell and NK cell neoplasms, brentuximab vedotin plus CHP showed manageable AE. This combination therapy has reached an objective response rate of 100% and a complete remission rate of 88%. A Phase 3 study comparing CHOP alone to brentuximab vedotin with CHP in the frontline treatment of CD30-positive mature T cell lymphomas will commence soon. In conclusion, brentuximab vedotin showed promising activity in clinical trials as a frontline treatment of different CD30 positive tumors with a manageable safety profile. Tibor Keler (Celldex Therapeutics, Inc.) discussed an agonist anti-CD27 monoclonal antibody. CD27 is member of the TNF receptor superfamily. It is constitutively expressed on most T cells and a subset of B and NK cells. The only known ligand is CD70. The engagement of CD27 by CD70 promotes a positive costimulatory signaling, resulting in T cell proliferation and survival. Soluble CD70 promotes strong CD8 + cytotoxic T cell responses. Blocking CD27 dependent costimulation of CD8 + T cell with a neutralizing anti-CD70 antibody completely abrogates antitumor activity of an agonistic anti-CD40 antibody in xenograft models; 4 however, an agonistic antibody against mouse CD27 has strong anti-tumor activity in xenograft models. 4,5 Fully human antibodies against human CD27 have been generated by immunizing transgenic mice (HuMab-Mouse TM ) that express human immunoglobulins. 6 The lead candidate 1F5 (CDX-1127) binds to CD27-expressing lymphoma cell lines and blocks CD70 binding. It has a high affinity to the human antigen (K D = 1.9 x 10 −10 M) and a similar binding characteristic to non-human primate CD27. The antibody has been evaluated with isolated human T cells and in transgenic mice expressing human CD27. Immobilized 1F5 has costimulatory activity on TCR stimulated human T cells and T cells from hCD27 transgenic mice. In these mice, 1F5 increases the number of antigen specific INF-γ producing T cells after immunization with a model antigen. In hCD27 transgenic mice, 1F5 has anti-tumor activity in disseminated lymphoma and colon carcinoma syngeneic models. The efficacy of 1F5 depends on the tumor burden whereby treatment of established tumors was less effective. The in vivo anti-tumor activity requires the engagement of mouse Fc-receptors presumably to induce cross-linking of CD27 on T cells. This has been demonstrated with an aglycosylated variant (N297S) that showed a complete loss of activity, but no loss in binding to CD27. In addition to T cell activation, 1F5 was shown to mediate ADCC activity against CD27 + human lymphoma cell lines and significant efficacy was observed in xenograft tumor models of human lymphoma cells in SCID mice. In non-human primates, however, no significant depletion of circulating CD27 + lymphocytes was observed. CDX-1127 has entered a 2-arm Phase 1 study; Arm 1 for the treatment of B cell hematologic malignancies known to express CD27 in which both T cell activation and direct effector function may contribute to the mechanism of action, and arm 2 for solid tumors (none of which express CD27) that may be more likely to be sensitive to immune therapy, including metastatic melanoma, renal cell carcinoma, hormone-refractory prostate adenocarcinoma, ovarian cancer, colorectal adenocarcinoma or non-small cell lung cancer.

The session on antibody-drug conjugates (ADCs), was chaired by Benjamin P. Chen (Ignatius Transaction Partners, LLC), and the keynote presentation was provided by Peter Senter (Seattle Genetics, Inc.), who discussed his thoughts on the past, present and future of ADCs. Paul Polakis (Genentech) then discussed ADCs for the treatment of cancer, and Hans-Peter Gerber (Pfizer) gave an update on the INO-VATE Phase 3 study of inotuzumab ozogamicin plus rituximab as a treatment for relapsed/refractory aggressive non-Hodgkin lymphoma in patients who are not candidates for intensive high-dose chemotherapy. Vedolizumab in patients with moderately to severely active ulcerative colitis and Crohn disease was discussed by Doina Roman (Takeda). William Olson (Progenics Pharmaceuticals) presented an update on PSMA ADC 1301, which is currently in Phase 2 clinical trials for the treatment of metastatic castration-resistant prostate cancer. Prostate-specific membrane-associated antigen (PSMA) is an internalizing membrane protein with a large extracellular domain that is expressed in nearly all cases of prostate cancer, but has limited expression on normal tissues, making it an attractive ADC target for cancer therapy. The Progenics PSMA ADC comprises a fully human IgG1 antibody linked to the tubulin inhibitor monomethylauristatin E (MMAE) by a valine-citrulline dipeptide linker. The conjugate demonstrated picomolar cytotoxic activity in vitro on PSMA-positive C4–2 cells and could selectively target cells with expression levels as low as 10,000 copies of PSMA per cell. Preclinical xenograft studies showed potent and selective anti-tumor activity in a number of human prostate cancer model systems with no evidence of target-related toxicities. In a Phase 1, open-label, dose-escalation study, PSMA ADC was administered every 3 weeks up to 4 doses in men with castration-resistant metastatic prostate cancer who had previously been treated with docetaxel. A total of 52 patients were treated at doses ranging from 0.4 mg/kg up to 2.8 mg/kg. Dose limiting toxicities, consisting primarily of neutropenia, were observed at 2.8 mg/kg; subsequent doses were reduced to the maximum tolerated dose of 2.5 mg/kg. Pharmacokinetic analysis indicated that the PSMA ADC had a half-life of approximately 2 d at doses over 2.0 mg/kg, and conjugate stability was observed with mean free MMAE concentrations less than 10 ng/mL. Waterfall plots of PSA levels demonstrate that many patients responded to PSMA ADC treatment. Biomarker analysis indicated that reductions in PSA levels usually correlated with reductions in circulating tumor cells (CTCs), particularly at dose levels above 1.8 mg/kg. With Phase 1 studies demonstrating an MTD of 2.5 mg/kg and durable coordinated reductions in PSA and CTCs, the PSMA ADC has advanced to a recently initiated Phase 2 clinical trial. The open-label, single-arm study will assess the anti-tumor activity and tolerability of PSMA ADC dosed at 2.5 mg/kg every 3 weeks for up to 8 doses in about 75 men with metastatic castration-resistant prostate cancer who have received one or two cycles of prior chemotherapy, including one with docetaxel. Dr. Olson concluded by discussing the use of PSMA itself as a potent predictor of patient responses and a biomarker for companion diagnostics to enrich responsive patient populations. The Dako 3E6 antibody can be used in IHC assays to measure neovascular expression of PSMA in non-prostatic formalin fixed paraffin-embedded tissues. Alternatively, a PSMA+ CTC assay can be used to quantify PSMA-expressing circulating tumors in patients. The PSMA CTC assay qualification and validation studies demonstrated good spike recoveries with low to moderate PSMA expression and no recoveries from non-PSMA expressing cells. The current Phase 2 PSMA ADC clinical trial includes an evaluation of the PSMA biomarker assays.

The session on antibodies in a complex environment was chaired by Kerry Chester (University College London), who also gave a presentation on developing dual-specific antibodies to tackle tumor heterogeneity. Roberto Polakiewicz (Cell Signaling Technology, Inc.) then presented data from a proteomics approach for rapid identification and cloning of monoclonal antibodies from circulation. Dr. Polakiewicz introduced the topic by pointing out that the isolation of monoclonal antibodies from a polyclonal antibody response is challenging mainly because of the complexity of the antibody mixture. A key limitation especially for the isolation of patient derived human antibodies is the availability of sufficiently large numbers of B cells that represent the full repertoire of circulating antibodies. In the proteomics based approach antibodies can be isolated and identified directly from blood plasma, which has the advantage that the circulating antibody pool is reflected and that antibodies can be enriched and selected by function, e.g., antigen binding. For bottom-up proteomics, however, the resolution of antibody diversity represents a tremendous challenge especially because variable regions and CDRs are not represented in genomic databases. The solution presented by the research teams of Cell Signaling Technology was the use of a custom reference database provided by deep sequencing of the B cell cDNA repertoire and the development of novel bioinformatics for peptide and VR identification by mass spectrometry. 1 Once peptide sequences of isolated antibodies are obtained via LC-MS/MS they can be matched with sequences derived from custom databases so that Antibody Sequence ID lists can be created. Then functional validation proves that the identified antibodies have the desired target binding properties. In a proof of concept study it was aimed to generate internalizing antibodies binding to RTK X protein. Therefore, rabbits were immunized with whole cells expressing RTK X. Enrichment and functional binding of the polyclonal serum could be demonstrated. Subsequently clusters of clonal families were identified. For selected antibodies cell binding and selective RTK X mediated internalization was shown. After presenting these proof-of-concept data, the question of whether the NG-XMT TM technology could be applied to isolate human antibodies from circulation was raised. This would be especially useful for the isolation of native human antibodies with desired properties, e.g., neutralizing or antibodies against auto antigens. In a first setting antibodies against the hepatitis B virus (HBV) antigen should be isolated from vaccinated healthy donors and in a second approach anti-human cytomegalovirus (HCMV) antibodies were isolated from naturally exposed healthy donors in order to identify neutralizing antibodies. In both studies donors could be identified that showed a good polyclonal response against the antigens of choice. In the HBV setting high affinity antibodies could be identified from a single vaccine recipient and the vaccine response could be tracked at the proteomic level over the course of the immunization schedule. In the HCMV setting in vitro neutralizing antibodies could be identified that demonstrated potencies even superior to the most potent currently known anti-HCMV neutralizing antibodies. 2 Dr. Polakiewicz summarized by noting that, by using NG-XMT TM technology, it is possible to identify antigen specific circulating antibodies from various sources (rabbit, mouse and human) based on functional properties through specific activity driven purification within a short period of time. John McCafferty (University of Cambridge) gave a presentation on blocking mediators of the tumor supportive microenvironment with antibodies. He pointed out the important role of tumor necrosis factor (TNF)-converting enzyme (TACE) proteins in the tumor microenvironment. TACE is a critical “sheddase” of many substrates, including epidermal growth factor (EGF) family ligands. It is upregulated in a wide range of cancers including colorectal, hepatic carcinomas and others. TACE expression correlates with disease severity and prognosis and several knockdown studies have indicated a role for TACE in tumor growth. The ectodomain of TACE consists of two distinct domains a highly conserved catalytic domain and a variable Dis-Cys Domain. A strategy was devised that allowed for the selection of inhibitory antibodies that bound to the Dis-Cys domain to achieve TACE specificity. Antibody D1 was shown to efficiently inhibit the proteolytic activity of TACE by binding to its Dis-Cys Domain. Binding of D1 was VH dependent. By performing a chain shuffling selection a VL domain could be identified that directed binding to the catalytic domain. The resulting cross domain inhibitor [D1-(A12)] demonstrated a higher affinity toward full length TACE (0.46 nM vs. 26 nM) and abrogated colorectal xenograft growth in vivo. 3 Dr. McCafferty then focused on antibodies binding to the Met receptor tyrosine kinase. This receptor is believed to play an important role in the hepatocyte growth factor (HGF)-mediated resistance to several anti-cancer kinase inhibitors. 4 The aim was to select antibodies that would block HGF mediated Met receptor activation. After a primary selection of Met blocking antibodies antibody 7A2 was identified which specifically inhibited HGF mediated SKOV3 cell migration. In the following in vitro affinity maturation campaign antibodies could be identified that showed a 50-fold increased potency in the SKOV3 cell migration assay as compared with the parent antibody. By solving the crystal structure of the complex, it was shown that the lead antibody 107_A07 binds to the first Ig domain of the Met receptor. It does not directly bind to the HGF binding site but efficiently blocks HGF signaling demonstrated by the inhibition of cell migration and growth of tumor xenografts in vivo. As the first speaker after the networking and refreshment break in the afternoon, Ann L. White (University of Southampton) discussed the roles of isotype and Fcγ receptor interactions for immunostimulatory activities of monoclonal antibodies. She classified antibody-antibody Fc receptor interactions into activatory interactions like with the mouse FcγRI (CD64); mouse FcγRIII (CD16) and mouse FcγRIV receptor or inhibitory interactions like observed with mouse FcγRIIB. The anti-tumor activity of a given monoclonal antibody can be significantly influenced by the antibody subtype and its Fc receptor interactions. These interactions have been best studied for “direct binding antibodies” that specifically recognize tumor antigens, e.g., the anti-CD20 antibody rituximab. These antibodies are most effective when they demonstrate high activatory to inhibitory ratios (e.g., high affinity to activatory receptors and low affinity to inhibitory receptors). This leads to efficient tumor antigen specific cell clearance (ADCC or CDC). In humans, IgG1 and in mice IgG2a represent subtypes with high A/I ratios. For immunomodulatory antibodies that are supposed to stimulate anti-cancer immunity the role of Fc receptor interactions is much less well understood. Dr. White presented their studies to elucidate the influence of Fc receptor interactions for the bioactivity of an anti CD40 antibody. In preclinical mouse models this particular rat IgG2a anti CD40 antibody showed anti-tumor activity against CD40 positive B cell lymphomas as well as certain CD40 negative tumors. For their studies they reformatted this antibody into a mouse IgG1 and mouse IgG2a format and could show in a mouse tumor model that only the parent antibody and IgG1 subtype prolonged survival of mice in a dose dependent manner. When analyzing T cell proliferation after antibody administration it was found that the mIgG1 subtype induced proliferation of CD8 + and CD4 + T cells whereas the mIgG2a subtype had no effect. Subsequently they addressed the question of whether or not these effects depend on FcγRIIB interactions. By comparing CD8+ T cell responses and anti-ovalbumin antibody titers in FcγRIIB wild type and FcγRIIB knockout mice they could demonstrate that FcγRIIB is required for T cell expansion and increased anti-ovalbumin titers. Additional in vitro experiments proved that FcγRIIB interaction with mIgG1 is responsible for B cell proliferation. Finally, it was shown that the interaction of mIgG1 with FcγRIIB expressed on, e.g., antigen presenting cells leads to receptor clustering which then triggers the activation signal and proliferation in CD40 expressing cell lines. 5 In concluding, Dr. White noted that the data emphasize that isotype specific FcγR interactions can significantly influence the bioactivity of therapeutic antibodies. Qi Zeng (A*Star) provided an update of her work on antibody targeting intracellular oncoproteins for cancer therapy. She noted that antibodies have traditionally been thought of as too large to routinely enter the cytosol, and, therefore, therapeutic antibodies could only access molecules that appear on the cell surface. Research from her groups and others has indicated that targeting intracellular oncoproteins may become a new option for the development of anti-cancer antibodies. 6 Professor Zeng first discussed phosphatase of regenerating liver 3 (PRL-3) protein, which is a very unconventional target for antibody therapies because it is expressed in the cytoplasm of target cells. PRLs are unique phosphatases with a characteristic prenylation motif at their C-termini; they belong to the class of protein tyrosine phosphatases. Overexpression of PRL-3 is associated with multiple human cancers, and, in gastric cancer, the mortality of patients overexpressing PRL-3 has been shown to be higher at all tumor stages. She then presented data showing that anti-PRL3 antibodies could have a therapeutic effect in vivo even though the antibody target is expressed intracellularly. In an in vivo metastasis formation model used in studies by her group, immunocompromized mice were i.v. injected with GFP-transformed PRL-3 expressing cancer cells and then treated with anti PRL-3 antibody every third day for a period of 17 d. The lungs of animals treated with anti-PRL-3 antibody showed a significantly reduced number of metastasis compared with control groups, thereby demonstrating the anti-tumor activity of anti PRL-3 antibodies. Similar results were obtained in a syngeneic metastasis formation model in immunocompetent mice using an anti-mouse PRL-3 antibody and PRL-3 positive B16F0 cells. In addition, internalization studies revealed that a portion of PRL-3 mAb is taken up by PRL-3 expressing cells. 7 Three possible models for targeting intracellular oncoproteins were presented by Professor Zeng. First, antibodies could be endocytosed and released from the endosome into the cytoplasm of live cells by mechanisms that are currently not well-understood. Second, the intracellular antigen could be externalized by unconventional secretion mechanisms and subsequent antibody binding could then trigger ADCC or CDC that would kill cancer cells. Third, proteolytic processing could lead to PRL-3 fragments being displayed on MHC class I complexes that are then bound by the anti PRL-3 antibody and targeted for destruction. Combinations or other mechanisms are also possible. Professor Zeng also discussed data for another potential intracellular oncoprotein target, Polyomavirus middle T antigen (mT). She presented data showing that treatment of transgenic MMTV-PymT(+/−) mice with an anti-mT antibody inhibited the spontaneous formation of breast tumors. In addition, when mice were immunized with the mT protein, the endogenous antibody response would reduce spontaneous tumor formation, indicating that the antigen is accessible for antibody therapy. Her group has shown that intracellular proteins such as PRL-3 and mT could be used as antigens to challenge and stimulate immune responses in hosts to produce antibodies for their own anticancer therapies. This unconventional concept suggests that intracellular proteins with selective and high expression in cancer cells are useful targets for mAb-based or vaccination immunotherapies, thus challenging current understanding. In concluding, Professor Zeng noted that their studies provide a proof-of-concept for targeting intracellular oncoproteins using antibody therapy and vaccination, and this can be translated into new treatments for patients in many potential ways. If the myriad of previously unexplored candidate target proteins are investigated, a new era of cancer therapies may soon become a reality. The last speaker of the session, John Lambert (ImmunoGen, Inc.) presented results from their maytansinoid antibody-drug conjugate (ADC) programs. He first pointed out that the first-generation ADCs failed in clinical development mainly because of poor tumor penetration and moderate cytotoxicity of the delivered drug payload. Discovery of the highly potent maytansinoids, however, opened new avenues for ADC development. He exemplified the potential of ADCs by showing the very convincing results of the Phase 3 clinical study of trastuzumab emtansine (T-DM1) in metastatic breast cancer. Several other maytansinoid-based ADCs (also called antibody-maytansinoid conjugates; AMCs) are currently undergoing clinical testing. For each of these molecules, the linker chemistry used for ADC synthesis was tailored toward a specific indication and model system. The importance of the linker chemistry used for antibody conjugation was demonstrated by showing that different linkers (e.g., cleavable vs. non cleavable) often had the same potency in vitro but showed substantially different efficacies in in vivo xenograft models. Dr. Lambert presented data showing that a thioether-coupled immunoconjugate was inactive whereas the cleavable disulfide linked version showed potent anti-tumor activity in a COLO 205 xenograft model. 8 An explanation for these effects could be that differently linked maytansinoids are differently metabolized in vivo. For instance, the non-cleavable SMCC-DM1 conjugate produced only one prominent metabolite that had poor bystander killing potential whereas the disulfide linked conjugate SPDB-DM4 was metabolized into three components of which a lysine –N ε -SPDB-DM4 demonstrated excellent bystander killing potential. 9 Dr. Lambert noted that tumor heterogeneity represents a challenge for the development of new ADCs, and that, in contrast to a SMCC linked conjugate, the disulfide linked AMCs showed efficacy in a HT29 xenograft model in which the target antigen is expressed heterogeneously. He then discussed the development of a novel anti-folate receptor 1 antibody (FOLR1). For the initial discovery more than 2000 clones were screened for FOLR1 binding and cross-species reactivity (cynomolgus monkey). Approximately 100 clones were evaluated in a high throughput screen for their payload delivery properties. Finally, after humanization about 10 clones were characterized for their bioactivity in vitro and in vivo. In these assays, different mAbs were tested as SMCC-DM1 conjugates and the antibody M9346A was identified as the novel lead compound based on its superior efficacy in a KB xenograft model. Different linker chemistries were then tested, which showed that a sulfo-SPDB-DM4 conjugate was the most active linker-maytansinoid format for the M9346A antibody. This improved linker was thus incorporated into the novel AMC named IMGN853 which showed efficient dose dependent activity in an ovarian cancer model. Studies with several linker-maytansinoid formats in another antibody-target system utilizing a multidrug resistant (MDR) and non-multidrug resistant folate receptor in vivo model showed that in the MDR model, a sulfo-SPDB-DM4 conjugate was more active than an SPDB-DM4 AMC, whereas in the non-MDR model both linkers showed similar activity. This result could again be explained by differences in the intracellular processing and increased bystander cell killing activity of the hydrophilic linker Sulfo-SPDB-DM4. In concluding, Dr. Lambert emphasized that linkers have attributes beyond simply joining a payload to an antibody of choice, and that both the anti-tumor activity and safety of an ADC can be influenced by the linker chemistry. The linker chemistry can influence the product of release (payload charge or hydrophobicity), the rate and site of payload release and the intracellular transport of the released payload. For the development of novel well tolerated and effective ADCs the properties of the antibody, the linker and the payload need to be matched with the biology of the target.

The morning session of the final day of the meeting, chaired by Rathin Das (Synergys Biotherapeutics, Inc.), was dedicated to discussion of early stage therapeutic antibodies. Alexey A. Lugovskoy (Merrimack Pharmaceuticals) presented the case study of MM-141, a fully human tetravalent antibody for the treatment of advanced solid tumors. MM-141 targets the PI3K/AKT/mTOR signaling pathway that is activated through IGF-1R and ErbB3 (HER3). This pathway is believed to promote tumor cell survival and is often activated in cancers in response to cytotoxic and targeted therapies. It has been proven clinically that antibody blockade of IGF-1R is ineffective, and Merrimacks data suggest that it is because heregulin (HRG) and ErbB3 provide strong compensation for IGF-1R blockade. Therefore only the dual IGF-1R/ErbB3 antibody co-inhibitor can completely block the IGF-induced activation of PI3K/AKT/mTOR. Merrimack used yeast antibody module display approach to select MM-141modules with optimal affinities and stabilities from structure-focused scFv libaries. The clinical molecules, MM-141, binds to IGF-1R and ErbB3 with subnanomolar affinities, block IGF-and heregulin- induced signaling, and induces strong receptor downregulation. MM-141 is stable in solution and manufacturable. While MM-141 showed monotherapy activity in xenograft models of pancreatic cancer, prostate cancer, breast cancer and Ewing’s sarcoma, it is likely to be most effective in combination with chemotherapeutics and targeted therapeutics. In fact, everolimus, gemcitabine and docetaxel all activated PI3K/AKT/mTOR and MM-141 inhibited this acquired resistance thus potentiating the activities of these drugs in xenograft models. These results suggest that MM-141 has great potential as an effective therapeutic for treatment of advanced solid tumors. David P. Humphreys (UCB) discussed improved antibody therapies for the treatment of Clostridium difficile infection (CDI). He first explained the disease basics of CDI, which is a serious, but common, nosocomial infection that is costly to treat. CDI is difficult to eradicate from hospitals and affects mostly people older than 65. The current antibiotic therapies target the C. difficile organism and only indirectly affect the toxin that causes the symptoms, which include diarrhea, fever and gut inflammation. The symptoms of CDI lead to extended hospital stays and transfers to intensive care units, both of which are major healthcare cost drivers. The mAbs developed by UCB are intended to reduce duration and severity of diarrhea, death rate and recurrence of infection. Dr. Humphreys emphasized the importance of antibodies in CDI because a correlation exists between the serum concentration of mAbs targeting the toxins TcdA and TcdB and lower disease occurrence, reduced recurrence and asymptomatic carriage. 1,2 Dr. Humphreys then referred to anti-toxin mAbs in clinical development. MBL and Medarex generated neutralizing human antibodies (IgG1) against TcdA (CDA1) and TcdB (MDX-1388). CDA1 alone showed a trend toward delay in recurrence in a Phase 2 trial, and together with MDX-1388 reduced recurrence rates but did not affect the duration and severity of diarrhea. Phase 3 trials of two combinations of mAbs (MODIFY I and II; NCT01241552 and NCT01513239 ) sponsored by Merck are ongoing. The UCB drug composition is a mixture of three human IgG1, one anti-TcdA mAb and two anti-TcdB mAbs that all target the C-terminal ‘cell-binding domains’ of the toxins. The total amount and the ratio of the drug mixture were tested in a hamster primary model of infection that matches that of the Medarex / Merck studies. The mAb mixture was dosed daily via intraperitoneal (i.p.) injection from day -3 to infection (day 0) and survival and body weight was observed for 28 d. UCB mAbs showed better protection of hamsters (higher levels, more durable) compared with MDX. They also demonstrated higher affinity in the low picomolar range, higher potency (low ng/ml) and the UCB mAb mixture was more effective at higher [TcdA] concentration. Additionally the trans-epithelial electrical resistance (TEER) assay was used as an in vitro surrogate for diarrhea due to loss of tight junctions. One UCB anti-TcdA mAb was protective in the TEER assay, CDA-1 was not. Furthermore, the UCB mAbs were potently neutralizing of TcdA derived from strains of ribotype 027 and 078, whereas the MBL/MDX mAbs were substantially non-neutralizing. 3 These results suggest the potential for UCB mAbs to improve patient outcomes by affecting duration and severity of diarrhea, death in addition to recurrence. The use of human anti-CCR4 mAb immunotherapy for the treatment of cutaneous T cell lymphoma was discussed by Wayne A. Marasco (Dana Faber Cancer Institute). The chemokine receptor CCR4 is expressed on skin-homing malignant T cells that form cutaneous T cell lymphoma (CTCL) and is associated with their skin-homing capacity. CCR4 is also expressed on certain T cell subsets like Tregs that can infiltrate the tumor and suppress cellular immunity. For these reasons, CCR4 might be a valuable target for antibody-based therapy for CTLC or other solid tumors. Dr. Marasco presented a murine anti-CCR4 antibody (mAb1567). Epitope mapping using N-terminal tagged CCR4 with extracellular loops swapped to CCR8 showed that binding occurs N-terminal, but the extracellular loops contribute or interact with the binding antibody as well. MAb1567 inhibits chemotaxis of CCR4-positive CTCL cells and showed potent antitumor effects in a mouse CTCL tumor model. The anti-tumor response was found to be complement-dependent cytotoxicity (CDC) and antibody-dependent cell-mediated cytotoxicity (ADCC) mediated by neutrophils. In vitro, NK cell mediated ADCC was also observed. Moreover, mAb1567 effectively inhibits chemotaxis of CD4( + )CD25(high) Tregs via CCL22 and abrogates Treg suppression activity in vitro. Humanization of mAb1567 with 22 changes in the heavy chain frameworks and 12 changes in the light chain framework resulted in 50% loss of affinity, but the effector functions of the humanized antibody were improved. The humanized antibody caused significant tumor shrinkage in vivo and was superior compared with the chimeric version. Affinity maturation focused on light chain CDR3 (variant mAb2–3) and further increased the biological activity. MAb2–3 was selected for further preclinical development. CCR4 is expressed on T cell subsets, and mAb2–3 increases proliferation of those subsets, but not in the Treg subset. Therefore, the antibody might have signaling function, probably because it binds N-terminal where CCR4 is signaling, but, to use it to block CCR4, the signaling properties need to be carefully observed. Taken together, results suggest that this high-affinity humanized mAb2–3 with potent antitumor effect and a broad range of mechanisms of action may provide a novel immunotherapy for CTCL and other solid tumors. 4 Stephen F. Carroll (Synergys Biotherapeutics) discussed the development of multifunctional anti-angiogenic antibody therapeutics for cancer and other angiogenesis-dependent diseases. Synergys has in-licensed a therapeutic fully human anti-Ang-1/2 IgG1 from NeoPharm, Inc. Currently marketed anti-angiogenic drugs provide only modest survival benefits for cancer patients and have raised some safety concerns. Therefore, Synergys aims to generate more efficacious and less toxic therapeutic products by targeting more than one angiogenic factor simultaneously. That is achieved by neutralizing both angiopoietin-1 (Ang-1) and angiopoietin-2 (Ang-2) through human antibodies where each arm can bind and neutralize Ang-1 or Ang-2 (dual-targeted, not bispecific). Ang-1 and Ang-2 are autocrine factors secreted by endothelial cells that bind the same receptor (Tie-2). Both angiopoietins are often overexpressed in cancer tissues and both are involved in tumor-induced vascular remodeling. Ang-1 functions as a Tie-2 agonist and promotes interaction of endothelial cells and pericytes and supports vascular maturation. Simultaneous inhibition of Ang-1 and Ang-2 by AMG-386, a Peptibody targeting Ang-1 and Ang-2 from Amgen, results in blocking of tumor xenograft growth and ovarian follicular angiogenesis. AMG-386 is currently undergoing clinical trials and has shown that the Ang-1/Ang-2/Tie-2 pathway is a suitable target for abrogating tumor angiogenesis. The anti-Ang-1/2 antibodies being developed by Synergys were isolated from a human scFv phage display library, affinity matured and converted to human IgG. They show significant in vivo activity in colon and pancreatic cancer models and, in combination with gemcitabine, reduced tumor size and blood vessel formation. Dr. Carroll concluded that human anti-Ang-1/2 antibodies would provide similar efficacy and safety profiles as AMG-386 and hold promise as new therapeutics for cancer and other angiogenic diseases. David Hilbert (Zyngenia) discussed the preclinical development of a TNF/Ang-2 bispecific Zybody for treatment of inflammatory diseases. Zyngenia specializes in the development of modular mAb-based fusion proteins called Zybodies. These consist of a conventional mAb as scaffold and various modular recognition domains (MRDs) fused N- or C-terminal to the heavy or the light chain to generate up to penta-specificity and 2–10 valencies per molecule. Because the mAb scaffold is not changed the molecules retain affinity, binding specificity and valency, as well as FcRn binding and Fc-mediated effector functions. Zybodies have drug-like properties like production and purification yields comparable to mAbs, no post-translational modifications, high stability and low immunogenic potential. Dr. Hilbert showed simultaneous binding of a Zybody to five independent targets and discussed the diverse therapeutic opportunities that are enabled by multi-specific, multi-valent and simultaneous target binding. The modulation of multiple cellular receptors could enhance cytotoxicity or inhibition of disease-promoting intracellular signaling. Enhanced target cross-linking could lead to enhanced clearance of multiple cytokines, enhanced receptor internalization or agonistic receptor signaling. In inflammatory bowel disease (IBD; Crohn disease, ulcerative colitis) the antagonism of multiple immune modulators (e.g., TNF, IL-1, IL-6, CD20, CTLA4) often results in increased incidence of opportunistic infections. TNF and Ang-2 are sequential, but functionally distinct inflammatory signals. TNF is a pro-inflammatory cytokine that mediates the release of Ang-2 from vascular endothelial cells. Ang-2 then destabilizes cellular junctions to allow cellular infiltrates and inflammation. The rationale for development of a TNF/Ang-2 dual antagonist is based on the fact that TNF antagonists have been approved for use in both Crohn disease (CD) and ulcerative colitis (UC), and that combination therapies targeting multiple immuno-regulators have resulted in high incidences of infection. The bispecific Zybody targeting TNF and Ang-2 (HUM-ang2 = Z2.1) consists of adalimumab and the Ang-2 binding MRD fused C-terminal to the heavy chain. In vitro the Zybody inhibits both NF-κB and AKT pathways that are activated by TNF-α and Ang-2, respectively. Z2.1 is able to bind TNF and Ang-2 (and not to Ang-1) simultaneously and independently with low nanomolar affinities. TNF induced cytotoxicity is inhibited with similar IC 50 as adalimumab and the Fc effector functions of adalimumab are retained. In a hTNF transgenic mouse model of autoimmune polyarthritis, the lead Hum-ang2 Zybody is superior to adalimumab in both arthritic score and histopathology score. Dr. Hilbert then presented results from the development of a TNBS-induced model of colitis in hTNF transgenic mice. In this model, the mice were pre-sensitized with TNBS at day -7 and challenged at day 0. The mice were dosed i.p. on days 0, 1, and 2 and assessed histopathologically for preservation of the mucosal epithelium. HUM-ang2 Zybody treated mice showed significantly reduced disease severity compared with adalimumab treated mice. PK/PD studies with HUM-ang2 Zybody were performed in cynomolgus monkeys and PK parameters were measured to understand clearance, exposure and in vivo stability of HUM-ang2, and pharmacokinetic properties for a human mAb were maintained for both specificities. The serum levels of Ang-2 were measured to assess the in vivo biological activity of HUM-ang2 Zybody. In monkeys treated with a single i.v. dose of HUM-ang2, increased serum levels of bound Ang-2 were found, while monkeys treated with multiple doses showed dose dependence and saturation of bound serum Ang-2. Dr. Hilbert concluded with an overview of the clinical program planned for HUM-Ang2. A Phase 1 study may include evaluation of safety, PK, exploratory biomarkers, as well as response indicators (improved mucosal healing, evidence of biomarker pharmacodynamics changes, improved efficacy compared with placebo and adalimumab). In the last talk of the session, Peter Ulrichts (arGEN-X) presented ARGX-110, an afucosylated, human IgG1 that binds and neutralizes CD70, as an immune checkpoint inhibitor and proliferation blocker. CD70 is the ligand for CD27 and, upon binding, activates the NF-κB pathway and results in shedding of sCD27. In healthy individuals, CD70 is transiently expressed on activated B and T cells and mature dendritic cells, but not on resting lymphocytes and non-hematopoietic cells. CD70 is, however, overexpressed on a variety of solid and hematological tumors and is thought to promote proliferation and survival of tumor cells. CD70 also induces Treg activation and concomitant immune escape. Dr. Ulrichts presented a detailed biochemical and functional characterization of ARGX-110 and explained the dual mode of action with very potent blocking of CD27 signaling by high affinity binding of CD70 and the Fc-mediated killing of CD70-expressing cells by CDC/ADCP/ADCC. The therapeutic potential of ARGX-110 was demonstrated in a Burkitt lymphoma xenograft model where the fucosylated, parental antibody of ARGX-110 significantly increased survival time at low doses. ARGX-110 also showed increased depletion in a whole blood ex vivo depletion assay compared with standard of care mAbs. Pharmacokinetic studies in cynomolgus monkeys resulted in a half-life of 13 d, suggesting a half-life of 23 d in humans according to human PK modeling. Importantly, the antibody showed no sign of toxicity and had no effect on B, T and NK cell numbers. ARGX-110 is scheduled to enter a Phase 1 study in patients with CD70-positive malignancies in January 2013.

The chair of the Antibody Therapeutics session on modulating the half-life of antibody therapeutics was Trudi Feldman (Abbott Laboratories). The first speaker, Ray Camphausen (Adnexus, a Bristol-Myers Squibb R&D Company) discussed approaches to extend the half-life of small modular antibody scaffolds using the example of the Adnectins. Three methods, pegylation, albumin-based approaches and Fc fusion, were presented. He discussed the Adnectin™ molecule CT-322, which is designed to block binding of vascular endothelial growth factor (VEGF) A, C and D to their receptor, VEGFR-2. They showed that, by adding a 40-kDa branched PEG, the half-life was increased to 70 h without losing the molecule’s high affinity to human VEGFR-2. In an alternative approach, the albumin binding domain (ABD) of protein G was fused to CT-322. When ABD was linked at the N-terminus, a half-life of 133–199 h was achieved in cynomolgus; however, this was only 67–87 h when ABD was added at the C-terminus. The third approach, Fc-fusion (N-terminal or C-terminal) gave a half-life of ~80 h in cynomolgus. Additional approaches include use of Adnectins with dual or multiple binding sites. Dr. Camphausen presented the example of FGF21-PKE Adnectin, which uses the Adnectin to increase half-life of FGF21. FGF21 is 20 kD and normally has a half-life of 1–4 h in cynomolgus. When fused to the albumin-targeted Adnectin, the half-life is increased to 96 h in cynomolgus. Adnexus is collaborating with Novozymes in the use of human serum albumin (HSA), optimizing binding to FcRn and reaching 180 h half-life in cynomolgus. Roland Kontermann (University of Stuttgart) presented concepts for half-life extension strategies. He reviewed the FcRn recycling pathway, showing this can be exploited for half-life extension approaches using albumin or the Fc domain. For single chain diabodies (scDb), current half-life extensions approaches include pegylation (PEG), glycosylation, albumin (HSA) fusion, fusion with the ABD from streptococcal protein G, Fc fusion, and immunoglobulin-binding domains (IgBD). He showed successful half-life extension with scDB-ABD, scDB-HSA and scDB-PEG 40K . For scDB-ABD, Dr. Kontermann also presented data showing that the half-life extension is FcRn-dependent, leading to selective and improved accumulation in antigen-positive xenograft tumors, and more effective than scDB PEG 40K . He described a novel approach using IgBDs. They tested domains from Staphylococcus protein A (SpA B , SpA B, SpA B ), Staphylococcus protein G (SpG c3 ) and Staphylococcus protein L (Ppl C4+ ). SpG c3 was the most promising module for half-life extension when testing scDbs and single-chain variable fragments (scFvs), and gave similar results as ABD. Dr. Kontermann’s overall conclusion was that there is further room for improvement in half-life extension. Javier Chaparro-Riggers (Rinat/Pfizer) discussed the concept of pH-sensitive antibodies, focusing on the anti-PCSK9 antibody J16 and it pH-sensitive version J17. The limitation of conventional antibodies is that they only have one binding cycle, and therefore have a short half-life due to target-mediated degradation. To circumvent this, pH-sensitive antibodies have been developed that will release the target at low pH and can then be rescued by the FcRn pathway. The J16 antibody reduces LDL-C levels, and thereby lowers coronary heart disease events. J16 was made pH sensitive by targeting the histidine residues H310, H433 and H435, creating the novel antibody J17. J17 showed prolonged half-life and better efficacy in mouse and cynomolgus. In summary, pH-sensitive antibodies are expanding the therapeutic space and are more effective especially when targeting soluble proteins. Syd Johnson (MacroGenics) discussed the half-life extension of bispecific dual-affinity re-targeting (DART) proteins. He showed data using ABD-fusion, pegylation and Fc-bearing domains (MP3 containing three gamma chains and MP4). The MP3-DART showed ~2-fold longer half-life than the ABD fusion, and is currently the most favorable format for this system. Gabriele Proetzel (The Jackson Laboratory) presented data on two humanized FcRn mouse models, Tg276 [B6.Cg-Fcgrt Tg(CAG-FCGRT)276Dcr/DcrJ; JR004919 ] and Tg32 [B6.Cg-Fcgrt Tg(FCGRT)32Dcr/DcrJ; JR014565 ]. Both models express human FcRn (hFcRn) while lacking murine FcRn, with Tg276 expressing hFcRn from a ubiquitous promoter and Tg32 from the human FcRn promoter. Expression studies using a monoclonal antibody directed against human FcRn (ADM31) confirmed the overall ubiquitous expression pattern for Tg276. In the case of Tg32, the expression pattern is similar to that described for human. Overall, only low expression levels were detectable in endothelial cells, which was surprising. Both models have been shown to be useful for pharmacokinetic studies and correlate well with cynomolgus data. In Tg276 models, the half-lives measured were shorter compared with Tg32; however, the Tg276 model has proven valuable when comparing antibodies half-life to each other. Dr. Proetzel discussed an example in which use of standard mice could have potentially led to dismissal of antibody candidates that looked inferior in the C57BL/6J wild-type mouse model. In contrast, the humanized FcRn mouse model showed the improvement in PK of the antibody candidate. Both humanized models have been backcrossed to immunodeficient backgrounds (Rag1-null or scid), which allows xenograft studies in the correct hFcRn context. Patrik Forrer (Molecular Partners) gave an update on DARPin drugs. He discussed the best-in-class VEGF antagonist MP0112, which is licensed to Allergan. MP0112 (AGN150998) is in a Phase 2 study ( NCT01397409 ) of patients with wet age-related macular degeneration (AMD); the active comparator is ranibizumab (Lucentis®; Genentech). The estimated study completion date is September 2013. MP0112 is stable at room temperature for 12 mo, the IC 50 is 6 pM (ranibizumab IC 50 = 300 pM), and the half-life is 6.2 d in rabbits. Dosing frequency for MP0112 is 120 d vs. 40 d for ranibizumab. Molecular Partners is developing therapeutic multispecific DARPins inhibiting VEGF and PDGF (MP060). By blocking PDGF in addition to VEGF, the recruitment of pericytes can be stopped and the maturation of neovessels blocked. Proof-of-concept in vitro and in vivo has been established using a HUVEC spheroid xenograft model in scid mice. To prolong half-life of DARPins, pegylation and serum albumin binding approaches were compared. A naked DARPin has a half-life of 10 min in cynomolgus monkeys. Using PEG20, this can be prolonged to 12 h, and, with PEG60, to 2 d. In an alternative approach, DARPins binding serum albumin were developed. Dr. Forrer showed data for three different variants binding different epitopes that achieved half-lives of 1 d, 8 d and 16 d in cynomolgus primates. This result is promising and suggests that it might be possible to for DARPins to achieve the natural half-life of IgGs.

The morning session of the Antibody Therapeutics track was chaired by Philip Thorpe (University of Texas Southwestern). Dario Neri (Swiss Federal Institute of Technology Zurich) gave a compelling keynote lecture about the delivery of immune modulators to vascular structures of disease sites. Angiogenesis, the formation of new blood vessels, is a hallmark of tumors and critically required to support tumor growth. To specifically and effectively deliver drugs into tumors, Professor Neri’s group aimed to identify unique tumor markers of the vasculature. Tumor marker discovery was driven by mass spectroscopic analysis of in situ biotinylated vascular structures of tumor-bearing mice and led to the identification of tumor-specific protein domains of oncofetal fibronectin and tenascin-C. 1 Antibodies against identified domains of fibronectin (EDA, EDB) and tenascin-C (A1) further attested disease specificity of these splice isoforms and confirmed that EDA was strongly expressed in the vasculature of human lung and liver metastases, and in a large panel of other human tumors with negligible expression in tissues of healthy patients. 1 In normal tissue, oncofetal fibronectin expression is restricted to the placenta and the endometrium in the proliferative phase. 2 Human antibodies in industrial development include L19 and F8 against the fibronectin domains EDB and EDA, respectively, and F16 against the tenascin-C domain A1. Of note, the EDA domain of fibronectin is also expressed in various inflammatory diseases, thus providing the potential for treatment in non-oncological indications. 2,3 After establishing a platform for tumor-specific drug delivery, Professor Neri’s group developed antibody-drug fusion proteins for the immune modulation of tumors. The L19-tumor necrosis factor (TNF)-fusion protein forms stable trimeric structures and was tested in clinical settings where recombinant TNF was already in use. Patients destined to undergo amputation because of large tumors in a limb can be treated by TNF therapy in a precarious procedure termed isolated limb perfusion (ILP). After isolation of the limb with a compressing device the patient is perfused with TNF at a concentration that would be lethal when given systemically. The targeted version of TNF (L19-TNF) proved therapeutically efficient at 20-fold lower concentrations compared with the non-targeted version and with no requirement for ILP. Intriguingly, large tumors in patients were resolved within six weeks. F8-TNF was effective in a sarcoma mouse model in combination with doxorubicin. Here, doxorubicin showed no effect, F8-TNF alone cured two out of five mice, whereas the combination of both drugs was curative. Surprisingly, it turned out that mice were vaccinated after F8-TNF treatment, since subsequent challenges with cancerous cells did not give rise to tumors. A combination treatment with L19-TNF and doxorubicin has now entered clinical trials. L19-IL2 (darleukin), which consists of interleukin (IL)-2 (Novartis) fused to the L19 antibody, has now moved to Phase 2a and Phase 2b clinical trials in metastatic melanoma patients. The EDA domain of fibronectin is also expressed in various inflammatory diseases, including endometriosis, arthritis and atherosclerosis. To target anti-inflammatory drugs to affected areas, fusion proteins of the F8 antibody and IL10 were generated. In arthritis mouse models F8-IL10 was curative in combination with methotrexate. 2,4 Professor Neri concluded that vascular targeting of drugs has great potential in cancer treatment and for applications beyond oncology. Bruce Cree (University of California, San Francisco) focused his presentation on the emerging role of B cells in demyelinating diseases. Strong evidence for a contribution of B cell mediated effects have been reported for a rare inflammatory disease of the central nervous system (CNS), termed neuromyelitis optica (NMO). This disease is characterized by severe attacks of optic neuritis and myelitis, but, unlike multiple sclerosis (MS), the brain is most commonly spared in early disease stages. In more than 90% of cases, NMO is a relapsing disease with severe cumulative disabilities including paraplegia and blindness due to optic neuritis. Remarkable advances were made by the discovery of autoantibodies in NMO patients against aquaporin 4 (AQP4), 5 a water channel that is widely expressed in the optic nerves, the spinal cord and the periventricular regions. This association with a B cell-mediated activity led to the hypothesis that B cell depleting therapies might be effective in reducing relapse occurrence in NMO patients. An open-label preliminary study showed that rituximab was well-tolerated and 7 of 8 patients experienced substantial recovery of neurologic function over one year. 6 In the second part of his talk, Dr. Cree reviewed the current state of therapy in MS and suggested that the B cell-specific antibodies ofatumumab and ocrelizumab in Phase 3 clinical trials would mark a shift of paradigm in the field from T to B cell-targeted treatment. Eric Fedyk (Millennium Pharmaceuticals Inc.) discussed the mechanism of action of vedolizumab, a promising gut-selective anti-inflammatory biologic in development for the treatment of inflammatory bowel diseases (IBD). 7,8 Crohn disease (CD) and ulcerative colitis are chronic inflammatory conditions of the gastrointestinal tract that typically emerge in young adults and are characterized by symptoms of diarrhea, incontinence, pain, fever and weight loss, leading to a lifetime of disability. IBD is associated with an accumulation of activated helper T (T H ) cells in the gastrointestinal tract. Blocking T H cell infiltration therefore constitutes a therapeutic paradigm for the treatment of IBD. Memory T H cells exhibit preferential routes of migration while surveying tissues for the presence of antigen. A distinct subset of gut-homing memory T H cells exists whose infiltration of the gastrointestinal tract requires binding of their α4β7 integrin to MAdCAM-1 expressed on the vascular lumen. Natalizumab is a dual α4β1 and α4β7 integrin antagonist that is approved for treatment of multiple sclerosis (MS) and CD; however, use of this antibody is restricted due to an association with an often fatal opportunistic viral infection of the CNS, termed progressive multifocal leukoencephalopathy (PML). It is believed that antagonizing the α4β1 integrin predisposes patients to PML by impairing immune surveillance of the CNS for reactivated JC virus, the cause of PML. Millennium Pharmaceuticals is consequently developing the α4β7 integrin antagonist vedolizumab on the premise that it will block gut inflammation without predisposing patients to PML. Vedolizumab binds exclusively to α4β7 integrin and not to the α4β1 and αEβ7 integrins. 9 In human whole blood, vedolizumab binds to a discrete T H cell subset with α4 high β1 low phenotype, but not to the majority of memory T H cells, neutrophils, and most monocytes. 9 It is a selective antagonist of the α4β7 integrin, which blocks cellular adhesion to MAdCAM-1, but not VCAM-1. 9 In experiments with colitic cotton-top tamarins, a monkey model of human IBD, blocking the α4β7 integrin alleviated chronic diarrhea and gastrointestinal inflammation. 10 A decrease in the frequency of lymphocytes in gastrointestinal tissues corresponded to a significant increase in α4β7 memory T H cells in peripheral blood. 11 No cases of PML have been reported in clinical trials with vedolizumab to date and the potential effects of vedolizumab on immune surveillance of the CNS was investigated in rhesus experimental autoimmune encephalomyelitis (EAE). Unlike natalizumab which blocks entry of T lymphocytes into the CNS and delays development of EAE, vedolizumab did not inhibit immune surveillance and inflammation of brains of Rhesus macaques and did not inhibit development of EAE. Vedolizumab therefore appears to be a gut-selective anti-inflammatory biologic that may be efficacious in IBD without predisposing patients to PML. Alan Epstein (University of Southern California Keck School of Medicine) discussed current approaches in cancer immunotherapy. While concluding that the majority of immunotherapeutic approaches aim to improve immunogenicity of tumors to the host immune system, he advocated a more integrated therapeutic strategy tailored towards reversal of tumor-mediated immunosuppression in addition to antigen-specific activation. Because tumors have evolved several ways to escape immunity (e.g., cellular suppression by regulatory T cells and myeloid-derived suppressor cells, molecular suppression by loss of HLA class I molecules or increased expression of HLA-G), better understanding of the molecular mechanisms of immunosuppression is imperative for an effective tumor therapy. Professor Epstein suggested that mechanisms of immunosuppression in a tumor are similar to those of early immune development where the fetus is protected against immune-destruction in the womb. As a therapeutic strategy, Professor Epstein’s laboratory aims to alter the tumor microenvironment by targeting immunocompetent cells. The chemokine (C-C motif) ligand 16 (CCL16)/LEC, a molecule that interacts with chemokine receptors, attracts leukocytes, monocytes, dendritic cells and lymphocytes. To target tumors, the LEC molecule was fused to an antibody, Tumor Necrosis Therapy targeting (abbreviated TNT), that binds to necrotic areas of tumors. The LEC-TNT fusion protein significantly reduced tumor volumes in several mouse tumor models, including colon 26, Renca renal and MAD 109 lung carcinomas. Surprisingly, a complete cure of tumors in the colon 26 model at an advanced stage was observed with the LEC-TNT fusion protein in addition to CD4 depletion. In contrast, CD4 depletion alone did not show any effects on the tumor volume. LEC-TNT furthermore showed synergistic effects when combined with an IL2-Fc fusion protein. To address toxic side effect of IL2 injections, the gene was mutated at codon 38, an alteration that reduced vasopermeability by 96% compared with unmodified IL2, whereas IL2 activity is marginally affected. To integrate vaccine technology with immunomodulatory reagents, studies are now in progress to test combination regimens. In preliminary studies, LEC-TNT in combination with 5-fluorouracil and cyclophosphamide completely abolished tumor growth in the 4T1 breast carcinoma mouse model. The LEC-TNT fusion protein is currently being tested in spontaneous tumors of dogs and may be tested in patients in the near future, pending support.