Final Report Summary - ADAMANT (Antibody derivatives as molecular agents for neoplastic targeting)
Monoclonal antibodies represent the largest and fastest growing class of pharmaceutical biotechnology products measured in terms of cumulative sales and of number of products in clinical development. Antibodies can be used to deliver bioactive molecules to the tumour environment, thus enabling molecular imaging applications or therapeutic interventions which spare normal tissues. The antibody-based targeting of tumour neo-vasculature is particularly attractive because of the dependence of cancer on new blood vessels and because of the accessibility of these structures from the bloodstream. The formation of new blood vessels (angiogenesis) is a rare process in the healthy adult, mainly confined to the female reproductive system. In the ADAMANT project, researchers have developed and implemented state-of-the-art methodologies for the identification and validation of proteins which are preferentially expressed in the tumour neo-vasculature and stroma. These markers of angiogenesis have been used for the generation and extensive characterisation of human monoclonal antibodies, whose in vivo tumour-targeting properties have been investigated by quantitative biodistribution analysis and by imaging studies in tumour-bearing mice.
The human monoclonal antibody F16, labelled with the radionuclide iodine-124, has been studied in a phase 0 clinical trial for the positron emission tomography (PET) imaging of patients with cancer. The newly developed monoclonal antibodies have been chemically modified to serve as delivery vehicles for bioactive payloads (cytotoxic drugs, radionuclides or cytokines). The therapeutic potential of the resulting armed antibodies has been studied, both alone and in combination with other drugs, in rodent models of cancer. This led to the development of novel biopharmaceutical agents which could eradicate tumours that are not cured by conventional pharmacological interventions. Some of the newly developed products and combination treatments are now being studied in clinical trials. The project has led to an improved understanding of the molecular composition of the tumour neo-vasculature and of the tumour stroma. The project allowed the development of novel anti-cancer products, stimulating industrial activities in Europe and leading to the implementation of novel therapeutic concepts and of new clinical trials.
Project context and objectives
Within ADAMANT, the development of innovative antibody-based pharmaceutical products targeting the tumour neo-vasculature and the tumour stroma has required innovation in three main experimental parts. In the first part, researchers aimed at providing a comprehensive analysis of proteins which are preferentially expressed in the neo-vasculature of solid tumours and of lymphomas. Researchers have used a chemical proteomics methodology, invented by members of the consortium, for the identification of accessible vascular markers in normal tissues and at sites of disease. The technology relies on the terminal perfusion of animal models of pathology with a reactive derivative of biotin, which covalently modifies accessible proteins.
Alternatively, the biotinylation reaction can be performed by ex vivo perfusion of surgically resected human organs with cancer. Biotinylated proteins can be purified on streptavidin resin in the presence of strong detergents, digested and the resulting tryptic peptides separated by HPLC and analysed by mass spectrometry in the presence of internal standards for quantification purposes. This research activity led to the development of an 'Atlas' of vascular proteins expressed in various tumour types, such as solid and metastatic tumours or lymphomas. A transcriptomic study of gene expression in endothelial cells under different angiogenic stimuli has revealed a complementary set of candidate markers of angiogenesis. The most promising target proteins have been validated using immunochemical techniques, in part thanks to the development of new monoclonal antibodies.
In the second portion of experimental activities, human monoclonal antibodies have been isolated from large phage display libraries. These new reagents have been extensively validated using immunochemical techniques (e.g. immunohistochemistry, immunofluorescence) in tumour specimens and normal tissues. Many new antibodies have been studied using nuclear medicine techniques, such as imaging or biodistribution studies in tumour-bearing mice.
In the third set of experimental activities, novel biopharmaceuticals have been developed by arming the most promising recombinant antibodies with three main classes of bioactive payloads:
(i) drugs with cleavable linkers;
(ii) radionuclides;
(iii) cytokines.
Novel cytotoxic drugs have been developed (e.g. tubulysin or dolastatin analogues), which incorporate chemical handles for the site-specific coupling to recombinant antibodies. Novel procedures have been invented and implemented for the radiolabelling of antibodies for imaging and therapy applications. The consortium has systematically investigated the field of antibody-cytokine fusion proteins (immunocytokines), which has revealed impressive therapeutic activities, especially when used in combination with other judiciously selected pharmaceutical agents. Some of the most impressive curative effects, which have resulted in the complete eradication of aggressive tumour types in rodents, include the combination of IL2-based immunocytokines with taxanes or with monoclonal antibodies which are already used in clinical practice.
The project has included accompanying translational activities to facilitate the implementation of the newly developed products and methodologies in the clinical practice. An immuno-PET Phase 0 clinical trial for the imaging of cancer patients with the F16 antibody (specific to the alternatively spliced A1 domain of fibronectin) labelled with iodine-124 has been started. The consortium has collaborated with new biopharmaceuticals for the preparation of clinical trial protocols and for the evaluation of clinical results targeting the tumour neo-vasculature and stroma, thus contributing to an increased awareness of the opportunities and problems associated with this novel class of pharmaceutical agents.
Project results
Identification of new tumour-associated antigens, accessible and selectively present in human primary cancer lesions and metastases, by in vivo and ex vivo chemical proteomics:
- characterisation of accessible proteins in different primary malignancies and associated metastasis;
- development of chemical proteomic methods to study the accessible proteins in human tissues.
Overview
A significant limitation of the most currently applied anti-cancer treatments is related to their toxicity. Chemotherapy relies on the expectation that anti-cancer drugs will preferentially kill rapidly dividing tumour cells. Unfortunately, they also target normal cells, particularly cells from tissues with high turnover such as hematopoietic progenitors and epithelial intestinal cells. This toxicity limits both the treatment duration and the deliverable drug doses, which significantly restricts the therapeutic efficiency needed to obtain and maintain a complete remission. One of the most attractive strategies for the development of new, efficient and selective treatments is the delivery of targeted bioactive molecules selectively to the tumour microenvironment. This strategy is based on ligand binding with high affinity and specificity to target molecules, over-expressed at disease sites. Following this rationale, the discovery of protein targets readily accessible through the bloodstream and selectively over-expressed in pathological tissues has become a major research objective. This group of molecules has a high potential to serve as innovative tools for effective imaging and targeted cancer therapy approaches. Ideally, reachable cancer proteins (accessible biomarkers) of high therapeutic value should be expressed solely by the malignant tumour and be at least not accessible in normal tissues. In the frame of the project, researchers have embarked on the development of suitable pre-analytical methods to study accessible proteins in human tissues.
Following this, members of the consortium have applied the newly developed technologies in order to characterise a variety of primary malignancies and associated metastasis both in human clinical samples and suitable in vivo models. Along these lines, the original methodology was developed in a previous EU collaborative project (STROMA 2004-08) by ETHZ in collaboration with IRFMN who applied a chemical proteomic strategy, based on the terminal perfusion of tumour-bearing rodents with reactive ester biotin labels and subsequent labelled-protein recovery including mass spectrometric analysis. The major aim was to establish an atlas of vascular proteins in health and disease. This technology was further improved in collaboration between ETHZ and ULG, yielding novel methods where first human tumour-bearing organs were ex vivo perfused and latter tumour biopsies were in vitro labelled in order to gain access to potentially targetable proteins. In the frame of the project, the technique has improved, mostly in terms of being more quantitative. Recently, the technology has been further refined by involving glycosylated proteins, in addition to biotinylated ones, in order to complement the repertoire of potentially accessible biomarkers.
Once the suitable methodologies were established, the partners moved on to characterise a broad range of malignancies and to establish the map of potentially accessible tumour biomarkers suitable for targeted therapy. The research was primarily directed on colon, liver (metastasis), pancreas, breast, bone (metastasis) and glioblastoma tumours from patients. The method which combines biotinylation with glycoproteomics has been applied on the breast cancer tissue: a novel biomarker has been discovered (CD276) and validated in more than 30 individual cases. The investigation of a breast tumour-associated bone metastasis was finalised. This study had the unique opportunity to examine bone metastasis and the corresponding breast cancer primary lesion obtained simultaneously from a fresh autopsy performed on a patient who died from disseminated breast cancer. The accessible protein biomarkers were identified using the previously developed combinatory procedure of biotinylation and glycoprotein extraction. 29 of the potential biomarkers were found uniquely expressed in the primary breast cancer while 27 proteins were detected only in the bone metastasis lesions. Several proteins belonging to small leucine rich proteo-glycans, thrombospondin and integrin families were found upregulated in the primary breast tumour.
The proteomics technology has been applied to characterise pancreas ductal adenocarcinoma (PDAC) and its accessible protein biomarkers. Mass spectrometry (MS) analysis revealed 484 differentially expressed proteins, of which 84 were evaluated as potentially accessible. 11 selected candidates were confirmed by Western blot and multiple reaction monitoring (MRM)-MS analysis. Transforming growth factor beta-induced (TGFBI), latent transforming growth factor beta binding protein 2 (LTBP2) and asporin (ASPN) were further investigated employing immunohistochemistry. They were found to be significantly expressed in a large group of clinical PDAC samples compared to corresponding normal and inflammatory tissues. In conclusion, TGFBI, LTBP2 and ASPN are novel, overexpressed and potentially accessible proteins in human PDAC. Human glioblastoma specimens as well as tumours developed in nude mice (xenografted human glioblastoma U373 and T98G cells) along with normal mouse brains were biotinylated. These samples were analysed and mutually compared. An easy and cost-effective method to validate the potential accessibility of tumour antigens in vivo has been developed. This assay is conducted with xenografted U87 human glioblastoma cells on the chicken egg chorioallantoic membrane (CAM). Antibodies against suitable targets can be injected intravenously in the CAM vasculature and visualised ex vivo using immunofluorescence.
The current study has highlighted several known and novel proteins implicated in the human glioblastoma: tenascin-C, CD44, collagen-VI-A1, sparc-like-1 and prosaposin. Immunohistochemical validation experiments confirmed the expression of these proteins in large series of human glioblastoma cases. Specifically anti-CD44 and anti-collagen-VI-A1 have been successfully injected/ validated in the CAM/ U87 model. Ex vivo perfusions of surgically resected human colon cancer (CRC) using biotin derivative, followed by comparative MS-based proteomic analysis of the labelled proteins, revealed quantitative differences between normal and cancer colon. 67 of the total 367 proteins identified were found to be preferentially expressed at the tumour site. Human monoclonal antibodies against two potential tumour targets, NGAL and GW112, were generated and their selective expression in cancer colon and not or barely in healthy tissues proved. In over the half of CRC patients a metastatic disease is present at the time of diagnosis, reducing the five-year survival to less than 5 %. One of the most common sites of metastasis is liver, characterised as asymptomatic until the disease is well progressed. An effective treatment and early diagnosis modality for the CRC liver metastasis would make a significant difference for a large population of patients. Researchers aimed at investigating whether the metastasis displays a homogeneous distribution of certain biomarkers. Some biomarkers are only found in distinct regions of the lesion.
Consequently, they have ex vivo biotinylated accessible proteins from several CRC-liver metastases and divided the specimen in 4 zones: Normal, peri-tumoural, tumour-rim and centre. The proteins were affinity purified and analysed for each zone separately using MS. In total over 1500 unique proteins were statistically divided into six patterns of expression. Approximately 1/3 was expressed solely in one of the four zones. A further 1/3 was found in all zones. Remaining proteins were present in 2 or 3 regions studied. Several known and new markers have prompted our interest among which: periostin (POSTN), TGFBI and latent transforming growth factor beta binding protein 2 (LTBP2).
The found antigens were validated in a mouse model of CRC liver metastasis. Mice liver metastasis and corresponding normal tissues were biotinylated and analysed by MS. The investigation revealed 922 mouse proteins present solely in the normal liver, 704 human proteins present in the metastasis and 248 proteins (mouse and human origin) found in common between normal mouse liver and mouse liver metastasis. 70 of the in common proteins were up-regulated in the tumour. Researchers have assessed the potential accessibility of the identified proteins and found that 110 proteins qualify as differentially expressed and potentially accessible. The data were compared with results obtained from the analysis of human liver metastases and resulted in 64 proteins which were modulated in the mouse model as well as in the human liver metastases. Further validation studies are ongoing.
An in vivo biotinylation of vascular structures by terminal perfusion was performed in three different syngeneic mouse models of CRC liver metastasis. The recovery and mass spectrometric analysis of accessible biotinylated proteins in liver metastasis revealed quantitative differences in the expression of vascular proteins compared to the host organ. The preferential expression of 10 markers at metastatic sites was confirmed by immunofluorescence analysis and, in the case of periostin, oncofoetal fibronectin and angiopoietin-like 2 proteins, by in vivo targeting experiments using radiolabelled antibody preparations [Borgia et al., Cancer Research, Cancer Research 2010].
Antibodies production and target validation:
- generation and production of new recombinant human antibodies against tumour specific antigens;
- in situ validation of antibody reactivity, specificity and target distribution in human tissue.
Overview
Tumour vessel and tumour environment specific molecules can serve for therapeutic purpose and can be developed as targets for antibody-based pharmacodelivery. An important precondition for their possible future clinical exploitation is the assessment of the 'novel' target distribution and of the specificity/reactivity of the newly produced antibody against it. Following the discovery of novel tumour specific accessible antigen(s), a key task is the generation of antibodies and their deep investigation. Immunohistochemical staining methods were optimised and the antibody reactivity tested in a wide range of normal, tumorous and non-tumorous pathologies and the distribution of the accessible antigen evaluated. As a result several newly produced human recombinant antibodies could be characterised and new putative tumour specific targets could be validated, as a prerequisite for the preclinical studies aimed at assessing therapeutic efficacy of the conjugated antibodies.
Target selection and first level of validation
Antibody production
Philochem has implemented the transient gene expression (TGE) as a standard methodology for the simple, versatile and fast (from months to weeks) generation of proteins against which antibodies should be raised. Using this method, several antigens could be expressed. For recombinant antibody production, two new synthetic human antibody libraries (named PHILO1 and PHILO2) were cloned to broaden the epitope recognition in respect to the previously characterised antibody library (ETH2-GOLD). A fully murine antibody library (PhiloTOP) was produced allowing the production of murine antibodies against markers which are not immunogenic in the mouse (e.g. EDB). Such antibodies serve as research tools in order to test long-term therapy effects of fully murine proteins in immunocompetent mice. With the aid of all these technologies/methodologies, recombinant antibodies against the selected stromal and vascular antigens were successfully produced. Among them are high affinity antibodies against human ED-A and ED-B fibronectin, MG50, periostin, uPA, human carbonic anhydrase IX (CAIX) and the vascular marker of B-cell lymphoma BST-2, as well as against murine matrix metalloproteinases 1, 2, and 3 (MMP1, 2, and 3) and EDB domain of fibronectin.
Target distribution and antibody reactivity validation
The tissue distribution of these putative target structures was investigated in detail. For a variety of the markers, a nice tumour or tumour vessel restricted occurrence could be proven. The tissue reactivity of the recombinant antibodies L19, F8, F16, G11 and D11 directed against fibronectin and tenascin-C variants were comparatively examined. For histological analysis a thyramid based amplification system was introduced leading to sensitive and reproducible immunofluorescence staining results for the biotinylated recombinant antibodies. Fluorescence double and triple staining procedures combining different recombinant antibodies were designed to investigate putative target structures in relation to vessels, stromal components, tumour cells and basement membrane structures as well as in relation to tumour type, subtype, malignancy grade and progression. It could be demonstrated that fibronectin and tenscin-C variants show different expression patterns discriminating tumours with wide stromal deposition from tumours with a restricted vascular deposition. The incorporation of fibronectins or tenascins in the blood vessel wall of the tumour was different too. These observations are of special interest for the selection of candidate tumours for treatment with fibronectin or tenascin-C antibodies and for individualised therapy strategies.
Clinically valuable expression patterns could be described for melanoma, lymphoma, lung carcinoma and renal cell carcinoma. For the first time, the spatial relation of vascular fibronectin and tenascin-C deposition in relation to vascular BM was defined by laser scanning microscopy colocalisation studies showing a stratified organisation of the vessel wall matrix. With respect to the preclinical studies aimed at assessing the efficacy of the antibodies as therapeutics, the reactivity of several recombinant antibodies against fibronectin and tenascin-C isoforms as well as against MMPs, uPA and CA IX was investigated in human tumour xenografts and in syngeneic mouse tumours, with particular focus on renal cell carcinoma, ovarian cancer, melanoma and lymphoma. Tumours strongly vary in their stromal and vascular expression of antigens recognised by the recombinant antibodies depending on the tumour type, the mouse strain (immunological background) and the species reactivity of the antibodies. Experimental tumours reflecting the results obtained with the clinical specimens were candidate to undergo experimental targeted therapy.
Target distribution in non-neoplastic pathologies
Targeted therapy of human neoplastic diseases using the recombinant antibodies F16, F8 and L19 (the best candidates for clinical development) might cause systemic side effects in healthy organs or in other non-neoplastic pathologies if there is an expression of the antigen/target molecules. Since no reactivity could be appreciated by analysing the collection of healthy tissue available commercially, it was a central task to analyse non-neoplastic pathologies. Because of the high incidences of both neoplastic and cardiovascular diseases there are high co-morbidity rates, the reactivity of the antibodies was assessed in different cardiac diseases to evaluate the risk of side effects and to possible identify new fields of potential diagnostic or therapeutic use. Studies were focused on evaluating the antigens recognised by the F16, F8 and L19 antibodies in pathologies of the cardiovascular system. F16 antibody stained areas of active tissue remodelling in atherosclerotic plaques (in both human and animal models) and may thus deserve to be exploited as a suitable building block for the development of radiopharmaceuticals for plaque imaging or for an antibody-based targeted delivery of therapeutic agents to atherosclerotic lesions.
Diseased human cardiac tissue of the right atrial auricle (RAA) and left ventricular septum (LVS) derived during cardiac surgery from patients with aortic valve stenosis (AVS) and/or coronary artery disease (CAD) were systematically analysed. A re-expression of F8 and F16 antigens could be observed, together with quantitative differences in the expression levels between AVS and CAD as well as in association to the grade of histological damage. Thus, F8 and F16 were suggested to be of potential interest for an antibody-mediated targeted delivery of diagnostic or therapeutic agents in human cardiac diseases with respect to anti-inflammatory or anti-fibrotic therapy strategies. The reactivity of the recombinant antibodies was investigated in a heterotopic rat heart transplantation model of chronic cardiac allograft rejection. A relevant reactivity was demonstrated for F8 but not for L19, F16, D11 and G11. Tissue reactivity revealed an extensive antigen expression (ED-A fibronectin) in heart allografts exhibiting signs of chronic rejection compared to healthy controls with clear spatial association to vessels showing cardiac allograft vasculopathy (CAV) and to areas of cardiac interstitial fibrosis (CIF), confirming ED-A fibronectin as a promising marker of chronic rejection (CAV and CIF) and F8 antibody as a vehicle for targeted therapy strategies.
Radioimmunotherapy (RIT)
- Highly selective tumour localisation of novel radio- or fluorescently labelled antibody conjugate.
- Development of effective radioconjugates for preclinical PET.
- Preclinical antitumor activity of a radioconjugate causing significant growth inhibition.
Overview
The use of tumour-targeting antibodies to deliver therapeutics to the tumour site is a highly selective form of cancer therapy, reducing the side effects frequently seen after conventional treatments while increasing the specificity. The aim is to develop novel and effective agents, using antibodies that target either the tumour vasculature or the tumour cells to deliver radiation to the tumour (radioimmunotherapy (RIT)). The antibodies are delivered into the circulation and can therefore reach sites in the body which are inaccessible to external beam radiation. Initial investigations, to confirm tumour specificity, are performed with fluorescently labelled antibodies. The antibody F8-SIP directed against ED-A fibronectin selectively targets tumour blood vessels, as can be demonstrated by multifluorescence microscopy. The antibody shows the same localisation as the vasculature over the whole tumour, indicating that the vast majority of blood vessels will be targeted for therapy, and when looked at in high power it has a highly specific perivascular distribution in colorectal tumours. The related antibody L19-SIP, when conjugated to radioactivity (125I-L19-SIP), produces a significant therapeutic effect. Researchers have been investigating ways to further increase this by using a combination of therapies, which treat different regions of the tumour. The vascular disrupting agent (VDA) combretastatin, currently in clinical trial, produces massive destruction of the centre of the tumour, but leaves a viable vascular rim which continues to grow. Combined therapy is therefore required to kill the whole tumour mass. We combined these 2 treatments, giving either radiolabelled antibody alone, or with combretastatin administered at 24 hours earlier and imaged the effect of the VDA on radiolabelled antibody distribution across the tumour. In the antibody alone group the tumour was viable and the antibody was distributed heterogeneously throughout, with the hottest areas shown in red. When combretastatin was also administered, the tumours developed massive central necrosis and the antibody was concentrated within the remaining viable rim where it could be most efficacious. As the total radioactive dose to the tumour was the same in both cases, researchers have achieved a complementary combined treatment with great therapeutic potential, which will form the basis of future radioimmunotherapy trials.
Development of effective radioconjugates for preclinical PET (immuno-PET)
There is a nuclear medicine requirement for non-invasive imaging of cancer patients in order to determine whether radiolabelled antibodies (radioimmunoconjugates) are successfully targeting, and being retained in, the tumour while clearing from normal organs and to determine target status. Procedures for development of these conjugates for single-photon emission computed tomography (SPECT) imaging are routinely available, but this is not the case for PET imaging. Researchers have therefore concentrated on developing standard procedures for producing suitable PET isotopes (positron emitters: 68Ga, 124I and 89Zr) and conjugating them to tumour-targeting antibodies. New linkers have been developed and logistics established for worldwide distribution of both isotopes and linkers, while labelling protocols were published for dissemination of this so-called immuno-PET technology.
Several preclinical and clinical immuno-PET studies for the non-invasive in vivo assessment of target status and antibody distribution have been successfully carried out. The potential of immuno-PET for quantitation of antibody distribution in patients has been evaluated. These achievements are of interest when using immuno-PET for selection of RIT candidates. This immuno-PET technology is disseminated worldwide and many leading academic centres and pharmaceutical companies use our technology in the (pre)clinical development of biotech products. While nuclear imaging like SPECT and PET is particularly well suited for quantitative imaging of deep-seated tumours, optical imaging using fluorescent dyes can provide complementary clinical potential for imaging superficial tissues. For this purpose a dual modal nuclear / optical imaging probe is being developed, in which nuclear imaging is used for whole body imaging and optical imaging for local imaging. Since angiogenesis is an early event in tumour invasion, radioactively-fluorescently labelled dual modal ligands directed against vascular tumour targets might be of diagnostic value and a guide for RIT in future trials.
Radioconjugate producing significant tumour growth inhibition
The major cause of death from colorectal cancer is the spread of disease to form liver metastases. Researchers have therefore investigated the effectiveness of RIT in this clinically relevant site, using an antibody that targets a common marker of colorectal tumour cells called carcinoembryonic antigen (CEA). They initially attached a red fluorescent label to the antibody, and were able to demonstrate highly specific uptake in all tumour metastases, regardless of size, with none in the surrounding normal liver. Radiation was attached to the antibody for a therapy experiment and researchers found that they could significantly inhibit the growth of tumours in the liver, without concomitant toxicity. They found that smaller liver metastases had far greater uptake of the therapeutic antibody than larger deposits. This system is being optimised for treatment of small liver metastases in the clinic, where the prognosis is still very poor.
Antibody drug conjugates
Generation of novel and effective antibody-drug conjugates, composed of either anti-tumour cell, anti-vascular or anti-stroma antibodies conjugated to tubulysins or dolastatin-15 analogues.
Overview
The availability of potent cytotoxic drugs having low-nanomolar or sub-nanomolar activity that can be conjugated to a monoclonal antibody through suitable chemical functions was the key for this part of the project. It was decided to focus on three different classes of potent drugs, all having a natural origin: Tubulysins and dolastatins, which are both peptides, and vinblastine, which is an alkaloid. These molecules, which are unsuitable anti-cancer drugs as such because of their high toxicity and narrow therapeutic window, were synthetically modified to:
i) achieve the appropriate level of chemical and metabolic stability, and
ii) install chemical functions allowing for the attachment to the antibody via cleavable chemical linkers.
The drugs might therefore be selectively released exerting a potent anti-tumour cell, anti-vascular or anti-stroma activity.
Almost 50 novel derivatives of highly potent compounds were synthesised, some in large quantities and high purity. These compounds were made available for biological studies on cancer cell lines and tumour animal models and for experiments of antibody-functionalisation. Some of the most potent tubulysins, dolastatins and vinblastine analogues were linked to human monoclonal antibodies or tumour homing peptides through disulfide or peptide linkers, thus obtaining several different conjugates. These conjugates, which were shown to have antibody-dependent anti-tumour effects in vitro, with IC50 values in the nanomolar range, were tested in vivo. Although the results are still preliminary, moderate retardation of tumour growth was observed with some of these conjugates, thus reinforcing vascular targeting antibody-drug conjugates as promising cancer therapy strategy. General recombinant antibody formats, which can be chemically modified with drugs and cleavable linkers at unique molecular positions, were successfully engineered.
Immunocytokines
- Cloning, expression and in vitro characterisation of novel immunocytokines
- In vivo testing of the therapeutic activity of novel immunocytokines.
Overview
Many pro-inflammatory cytokines have been used for cancer therapy applications in clinical trials, but severe side effects have often prevented dose escalation to concentrations needed to induce cures. The fusion of cytokines with antibodies or antibody fragments, capable of selective accumulation at the tumour site, yields a new class of biopharmaceuticals (termed immunocytokines), which often displays superior therapeutic properties. In the frame of the project, researchers have produced and characterised several novel immunocytokines, based on IL2, IL7, IL10, IL12, IL17, IL18 and interferon-alpha as potent stimulator of the immune system. The resulting fusion proteins have been characterised in vitro (in terms of their pharmaceutical quality and activity) and in vivo (in terms of their tumour-targeting performance and therapeutic activity). Some of the tested immunocytokines (most notably those based on IL2, IL12 and IL15) have exhibited an impressive ability to inhibit tumour growth rate when used as single agent and to display superior performance compared to immunocytokine based on antibodies of irrelevant specificity in the mouse. The antibodies used as fusion partners for the production of vascular tumour-targeting immunocytokines were F8 and L19 (specific to the alternatively spliced EDA and EDB domain of fibronectin) and F16 (specific to the alternatively spliced A1 domain of tenascin-C). As negative control, monoclonal antibodies specific to hen egg lysozyme, an irrelevant antigen in the mouse, were used.
The fusion proteins that were cloned, expressed and characterised (in vitro and in vivo) were many more than the four mentioned in the milestones. The fusion protein F16-IL2 has exhibited a strong anti-tumour effect both in a subcutaneous and in an orthotopic model of glioblastoma multiform. A microscopic analysis of tumour sections in mice treated with saline, temozolomide, F16-IL2 or a combination of temozolomide plus F16-IL2 has evidenced that the immunocytokine promotes a strong infiltration of leukocytes into the tumour mass. Interleukin-12-based immunocytokines has been investigated for many years by the ETH Zurich and Philochem scientists. This heterodimeric cytokine has traditionally provided challenges for protein expression (the N-terminus of the p40 subunit needs to be free for maximal biological activity), but also opportunities for the design of new formats. A novel disulfide-linked format of an immunocytokine consisting of the anti-EDA antibody fragment scFv (F8) and of human IL12 has recently been developed by the Philochem group and has been moved to GMP manufacturing. The protein displayed excellent pharmaceutical quality and impressive tumour-targeting performance in xenograft tumour models in rodents.
Clinical significance
The clinical significance of research in the field of vascular tumour-targeting immunocytokines is reinforced by the clinical results recently observed by the Philogen group, who has brought three immunocytokines (L19-IL2, L19-TNF, and F16-IL2) to Phase II clinical trials in patients with various cancer types. The most advanced of these products is L19-IL2, which has exhibited numerous objective responses in a Phase IIa study with 32 patients and which is currently being investigated in a controlled phase IIb in 90 patients with metastatic melanoma, in combination with dacarbazine. Even though these clinical activities were not part of the project, the positive results observed so far provide a strong rationale for research in the immunocytokine field.
Optimisation of combination therapy with antibody conjugates (immunocytokines)
- Preclinical antitumor activity of combination treatments for cytokine conjugates
- Anti-metastatic activity of immunoconjugates in combination with chemotherapy in one preclinical model.
Overview
The development of antibody derivatives in clinical trials relies on their use in combination with other anti-cancer treatments. In the frame of project, members of the consortium have investigated the immunoconjugates described above (i.e. F16-IL2, F8-IL2, F9-IL2) in combination with standard-of-care chemotherapy or targeted drugs. A great effort to optimise the best treatment conditions on ad hoc mouse tumour models reflecting the tumour patient's characteristic was made. Representative treatment regimens have yield cures in tumour-bearing mice and represent promising candidates for clinical development.
F16-IL2 immunocytokine in combination with Temozolomide on glioblastoma
Temozolomide, the standard of care for newly diagnosed glioblastoma patients, in combination with F16-IL2 - a clinical-stage immunocytokine consisting of human interleukin (IL)-2 fused to the human antibody F16, specific to the A1 domain of tenascin-C- was investigated on a model of glioblastoma. Immunohistochemical analysis with human glioblastoma surgical specimens and with U87 xenografts showed a strong and selective tumour staining using the F16 antibody. A quantitative biodistribution in nude mice confirmed the preferential accumulation of the radiolabelled F16-IL2 at the tumour site. In the therapy study, the combination of F16-IL2 with temozolomide induced a complete tumour remission in the animals. The same treatment led to a substantial size reduction of the U87 tumours growing orthotopically in the brain, compared to temozolomide or the immunocytokine administered alone. The combined use of temozolomide and F16-IL2 may deserves clinical investigations, which will be facilitated by the excellent safety profile in Cynomolgus monkeys and by the fact that F16-IL2 is currently being investigated in clinical trials in patients with metastatic cancer.
F8-IL2-immunocytokine in combination with taxanes on melanoma model
Human melanoma is characterised by strong vascular and stromal positivity for the EDA domain of fibronectin (EDA-Fn) and the pattern of EDA-Fn expression in melanoma xenograft models fully replicated that of the patient lesion. We found that F8 -an antibody directed against EDA fibronectin- conjugated to interleukin 2 (F8-IL2) and combined with Paclitaxel (PTX), but not decarbazine (DTIC), completely inhibited the growth of human melanoma xenografts transplanted intradermally, with 80 % complete regressions. Researchers showed the particular role of Paclitaxel in this combination regimen. PTX increased tumour vessel perfusion and permeability in melanoma xenografted tumours, favouring the uptake of F8 antibody and PTX prones the infiltration of NK cells and macrophages in melanoma xenograft. These preclinical studies, showing complete tumour regressions and anti-metastatic activity, endorse the use of chemo-immunotherapy in the treatment of malignant melanoma and suggest that PTX is endowed with properties affecting the tumour stroma that deserve attention for the design of treatment regimens.
L19-IL2 immunocytokine in combination with rituximab on non-Hodgkin lymphoma (NHL)
The antibody-mediated delivery of therapeutic agents was investigated in hematologic malignancies. The EDB domain of fibronectin, is expressed in B-cell NHL and the human monoclonal anti-EDB antibody L19 can selectively localise to the lymphoma-associated subendothelial extracellular matrix. In vivo, the preferential accumulation of the antibody at the tumour site was confirmed by quantitative biodistribution analyses with radio-iodinated antibody preparations. The fusion protein L19-IL2, which mediates the delivery of interleukin-2 to the neovasculature, displayed a superior antilymphoma activity compared with unconjugated IL2 in localised and systemic xenograft models of NHL. When co-administered with rituximab, L19-IL2 induced complete remissions of established localised lymphomas and provided long-lasting protection from disseminated lymphoma. The combined use of rituximab and L19-IL2, which dramatically increases the infiltration of immune effector cells in lymphomas, may deserve clinical investigations, facilitated by the fact that L19-IL2 is currently being studied in phase II clinical trials in patients with solid tumours [Schliemann et al, Blood 2009].
PET clinical validation (microdosing)
The aim is the clinical evaluation of a promising anti-vascular antibody in immuno-PET microdose phase 0 trial.
Overview
Immuno-PET has the potential to become a powerful imaging tool for efficient clinical antibody development because this technique allows accurate quantification. With relatively few patients the antibody can be characterised for tumour-targeting capacity, and homogenicity of tumour-targeting, possible cross-reactivity with normal tissues, residence time in blood, tumour and normal tissues, possible complex formation in blood, catabolic routing and inter-patient variation. Microdosing (phase 0) studies as described by the European Agency for the Evaluation of Medicinal Products (EMEA), could be used to compare upfront compounds with similar intended Mode of Action but potentially different PK/PD characteristics, thus weeding out unsuitable agents for clinical development prior to extensive clinical investigations in patients. These studies do not need the full set of toxicity and other safety studies. Researchers anticipated that for antibodies the microdosing principles will be applicable, especially when there is evidence that MAb biodistribution will not be strongly dependent on MAb dose. The aim of was the clinical evaluation of a promising anti-vascular antibody in an immuno-PET microdose phase 0 trial.
124I-F16SIP: aphase 0 single microdose study
Based on available preclinical data, F16SIP directed against the alternatively spliced domain A1 of tenascin was selected for phase 0 studies in operable head and neck cancer patients. All consortium partners were involved in the design of the study. The following EMEA-released guidelines and position papers were considered: 'Note for guidance on the preclinical evaluation of anti-cancer medicinal products', 'Note for guidance on the preclinical evaluation of biotechnology-derived products', 'Note for guidance on non-clinical safety studies for the conduct of human clinical trials for pharmaceuticals', and 'Position paper on non-clinical safety studies to support clinical trials with a single microdose'. Since the clinical trial was planned to be performed at VU University Medical Center in the Netherlands, discussions were performed with Dutch regulatory authorities about the non-clinical safety studies needed to allow a clinical immuno-PET study with a single low dose of antibody. It became clear that 30 nMol MAb is allowed for phase 0 clinical trials and that toxicity studies performed with cold F16SIP are also valid for studies with radiolabelled F16SIP, as long as the binding and pharmacokinetic properties of the MAb are not altered by the labelling procedure.
Based on this information, the following steps were made:
i. GMP produced clinical-grade batch of F16SIP was made available with the consortium;
ii. non-clinical safety studies were performed according to aforementioned regulation;
iii. a tissue cross reactivity test was performed according to FDA recommendations;
iv. procedures and a manufacturing infrastructure for GMP production of 124I-F16SIP were set up. An Investigation Medicinal Product Dossier (IMPD) of 124I-F16SIP was written;
v. a clinical protocol entitled: 'A phase 0 single microdose study to evaluate the pharmacokinetics / -dynamics and specific tumour-targeting of 124I-F16SIP in head and neck cancer patients', was developed and approved by METC;
vi. three head and neck cancer patients were enrolled and treated in this trial and two other patients are planned.
The first patient had a T2N0M0 lateral tongue carcinoma on the left side. The first PET-CT scan at 30 minutes after injection showed mainly blood pool. The second PET-CT scan at 24 h after injection showed clear uptake of 124I-F16SIP in the left lateral tongue, corresponding to the localisation of the primary tumour. During surgery, at 168 h after injection samples were taken from blood, tumour, skin, healthy mucosa, fatty tissue and muscle. The highest uptake was found in tumour tissue with a tumour-to-blood ratio of 8.2. The second patient had a T2N0M0 check carcinoma. The second PET-CT scan at 24 h after injection showed some uptake of 124I-F16SIP in the tumour area, while biopsies at 120 h after injection revealed a tumour-to-blood ratio of 6.6. Because of slow patient accrual, the original study protocol was amended in a way to limit the number of PET scans to two in order to avoid that patients have to travel to the clinic on several days in a row. In the meantime a third patient has been included in this phase 0 trial, and the trial is expected to be finished within a couple of months. Results from this phase 0 trial thus far show that F16SIP is a promising antibody for targeting antigen-positive tumours like head and neck tumours. This antibody is candidate to be used for tumour detection and when coupled to a toxic payload for selective tumour eradication.
Regulatory affairs and strategic monitoring
- Design of clinical development plan and protocol for phase 1 first-in-man (FIM) study
- Micro-dosing study protocol
- Organisation of a multidisciplinary oriented focus group (FG).
Overview
The mission of ADAMANT was to discover new potentially 'druggable' anti-cancer targets in solid tumours and generate highly effective new monoclonal antibodies against these targets, minimising collateral damage to the normal tissues by homing the delivery of toxic agents (chemical or natural products molecules, radionuclides or immunostimulatory cytokines) specifically to the cancer cells or the cells of the cancer-supporting scaffold (stroma) to the cancer cells or the cells of the cancer-supporting scaffold (stroma). The challenge of the project was to identify these new potentially druggable targets in a preclinical framework mimicking the ultimate setting in which these antibodies would be used (i.e. the cancer patient) as much as possible.
Since the inception, it was recognised that the traditional 'bench to bedside' framework could be converted to a more innovative 'bed to bench-side' approach retrofitting the constructs of the preclinical experiments from the actual clinical settings. With this challenge in mind a task was devised to support the project and indeed relayed mainly on the collaboration with the more 'clinical oriented' tasks: Optimisation of the combined therapy with other agents; PET clinical validation and RIT. Southern European New Drug Organisation (SENDO) was involved in the project to provide, in connection with the performance of the preclinical activities, the expertise in clinical drug development and specifically in preparing dossiers, designing pharmacodynamic and pharmacokinetics based phase 0 and I trials and running them in cancer patients at selected and qualified clinical centres.
The organisational model of SENDO is that of an academia / industry go-between geared at understanding the needs and fulfilling the missions of both stakeholders. During the project, SENDO promoted and supported innovative anti-cancer drug research stimulating interaction with industry in the early stages of drug development. Working through its SME, SENDO TECH, the foundation is committed to guarantee the highest level of quality research, validity and reliability of clinical data collected. They have in place established standard operating procedures (SOPs) for data collection, logistical organisation of the trial, monitoring activities and medical writing. SENDO TECH is a certified CRO in Italy and France since 2009 and 2010, respectively.
Within ADAMANT, the full quality-controlled radiological assessment of clinical response in the patients included in the phase I first in man studies (including the blinded revision of all the radiological assessments of patients treated with L19-IL2) was provided; in addition the protocol templates for phase 0 and I FIM studies and all kinds of quality controlled study report templates were designed and finalised. Based on that, the study report of the first FIM concluded was produced and clinical data have been presented at two international conferences during 2010: ASCO and ASTRO. Regulatory assistance to define controversial ethical and legal aspects of phase 0 clinical trials in Europe was provided. Within this framework the first PET phase Zero IMPD was finalised and a submission for a microdosing study was done by VUA. An Immuno-PET clinical study has been designed and activated. To facilitate the strategic monitoring of the results, multidisciplinary oriented focus groups have been organised to discuss and analyse the open questions on PET in translational oncology and on pharmacodelivery of cleavable linker's drugs. The discussions helped the creation of a 'platform of concepts' stemming from the project collective experiences and specific expertise. This platform could become both a conceptual signature of the project and a springboard for future scientific explorations of the participants. Some of the focus debated issues have been subjects of publications.
Exploitation of experimental results
To assure the visibility of the results, the consortium has disseminated information through various channels involving different targets. There are communications by the coordinator such as the website addressing the general public and two ads appearing in the magazine 'The Parliament', addressing European decision-makers, and dissemination activities that are done by individual participants. The project website has been available since the start of the project at http://www.adamant-fp7.eu and has been periodically updated, providing news of major events of the consortium and the end of the project. The website will be maintained online after the end of the project. All participating groups have published their research results either alone or together with other beneficiaries in peer-reviewed scientific journals. A total of 50 papers in peer-reviewed journals have been published (of which 14 open access), and more than 70 congress presentations given. More manuscripts (n = 14) are under preparation and submitted or accepted for publication. In several instances beneficiaries' scientific representatives have been invited as speakers at national and international meetings. These activities will continue as expected. A workshop on 'Tumor microenvironment and metatases' was organised by the Belgian Society for Cell and Developmental Biology in Liege, 27 March 2010 in collaboration with three Seventh Framework Programme (FP7) projects.
Personal training programmes were facilitated among the partners of the ADAMANT consortium:
- Fabia Doll, Master Student at ETH Zurich (B2), performed her Master thesis at UCL (B7) from May to July 2008;
- Dr Andrei Turtoi, scientist at ULG (B8), spent a four-month training period (from June to September 2009) at UCL (B7) in the frame of his postdoctoral training to learn the technique of orthotopic liver metastasis in murine models grafting colorectal tumour cells into the liver via splenic injection;
- Antonietta Silini performed her PhD thesis in cooperation between IRFMN (B1) and UCL (B7) and spent one month (July-August 2009) at UCL;
- Dr Sreejith P Shankar, postgraduate student at KemoTech (B4), spent a three-months secondment period from February to April 2009 at VUA (B10) during his PhD training;
- Francesca Pretto is a PhD student of ETH Zurich (B2), from January 2010 she is performing part of her research at IRFMN (B1).
Some 20 doctoral, PhD and graduation theses have been promoted by the beneficiaries on subjects connected with the project, 11 have been completed and 9 are still in progress. A total of six patent applications were filed with EPO and two more are under evaluation. On three occasions a research valorisation roundtable was held with the experienced support of the advisory board. The aim of this roundtable was to assure the assessment of the societal and industrial exploitation of results and best return on investment of the project. Progress presented by the ADAMANT consortium was analysed in detail by members of the advisory board, resulting in suggestions from people not directly involved in the project. Analyses were discussed in plenum and the advice allowed a focused continuation of the work until the next major meeting.
Potential impact
In spite of their commercial success, conventional monoclonal antibodies which are currently used for the treatment of solid tumours offer only an incremental benefit to cancer patients. For this reason, there is an urgent need for superior anti-cancer biopharmaceuticals. Virtually all large pharmaceutical companies are heavily investing into biomedical strategies for the arming of monoclonal antibodies with bioactive payloads.
The potential impact of the project can be analysed in terms of four main milestones achieved over the last three years:
i. the discovery and validation of novel markers of tumour neo-vasculature and stroma, followed by the generation and characterisation of human monoclonal antibodies;
ii. the development and implementation of novel antibody-based therapeutic strategies;
iii. the stimulation of innovative clinical trials for the treatment of patients with incurable forms of cancer;
iv. the creation of new jobs, in terms of formation of new companies and growth of existing companies.
Traditionally, markers of tumour angiogenesis and of the tumour stroma have been discovered by serendipity. Out of hundreds of different antibodies which are routinely analysed by immunohistochemistry, it would occasionally happen that the study of a new monoclonal antibody would reveal a staining pattern coinciding with tumour neo-vascular structures. Markers of angiogenesis, such as the EDB domain of fibronectin, endoglin, PSMA, have all been discovered this way. The project has contributed to the development and implementation of methods for the systematic characterisation of vascular proteins in health and disease. There now are atlases of vascular markers for various tumour types, which allow choosing a target for antibody product development based on an unbiased comparative evaluation of expression patterns for the first time. We are confident that the quality of the tumour-associated antigen used for product development will ultimately determine the quality of the cognate antibody-based product.
On the second aspect of ADAMANT's potential impact (novel antibody strategies), it is still not clear what the most successful strategy will be for arming antibodies. Different strategies have distinctive favourable features and carry distinctive liabilities. For example, radiolabelled antibodies may be ideally suited for the therapy of radiosensitive tumours (e.g. lymphomas), but are less effective for radioresistant tumours or for tumours which do not exhibit an adequate antibody uptake compared to normal tissues. For this reason, researchers have chosen to focus on three main antibody functionalisation strategies (radiolabelling, fusion to cytokines, coupling to drugs with cleavable linkers) and have made an impact in all these areas. The Philogen group (one of the partners in the consortium) has brought seven armed vascular targeting antibodies to the phase I and to phase II clinical trials. At present, the most promising clinical results with armed vascular targeting antibodies have been reported for the:
(i) radioimmunotherapy of patients with incurable forms of lymphomas; and
(ii) therapy of metastatic melanoma patients with IL2-based or TNF-based immunocytokines.
In the first case, multiple major responses have been observed in patients with Hodgkin's lymphoma, who had failed all other therapeutic options.
In the case of IL2-based immunocytokines, the treatment of metastatic melanoma patients with L19-IL2 plus dacarbazine has led to the doubling of median overall survival compared to historical control patients, treated with dacarbazine alone. This treatment modality is currently being investigated in phase IIb clinical trials in Germany, Italy and Austria. Importantly, the recent discovery that taxanes may be better suited for the combined use with IL2-based immunocytokines and that curative effects can be observed when Paclitaxel is administered before the immunocytokine (not after) will stimulate the execution of novel combination trials for the treatment of patients with metastatic melanoma. The implementation of the first phase 0 immuno-PET clinical trial in Europe (i.e. the use of the F16 antibody labelled with iodine-124 for the imaging of cancer patients) is another important milestone of the project. We expect that in the future targeted antibody drugs will be developed in combination with imaging modalities, which may guide patients' selection and may reveal a correlation between responding patients and patients with a good antibody uptake in the tumour. The companies which have participated in the project have grown substantially during the activities of the project. PHC has opened new laboratories in Otelfingen and has created new jobs (Philochem AG (and its mother company, Philogen) now have approximately 80 employees, between the Siena and Otelfingen site). KTECH has made substantial contributions to the development of tubulysin derivatives as an innovative and general class of cytotoxic agents for the coupling to antibodies. Moreover, Kemothech srl has recruited two employees thanks to the results achieved. Finally, Targetome is a spin-off company of the University of Liege, which has been founded as a result of innovative technologies in target discovery and validation developed in the frame of the project.
Website address:
http://www.adamant-fp7.eu/
E-mail address: adamant@marionegri.it
List of beneficiaries
B1 IRFMN, Coordinator, WPs 6 and 10 leader
Raffaella Giavazzi
Laboratory of the Biology and Treatment of Metastasis, Department of Oncology,
Istituto di Ricerche Farmacologiche 'Mario Negri'
Via Giuseppe La Masa, 19, Milano 20156, Italy
B.2 ETHZ, Scientific Deputy, WP 5 leader
Dario Neri
Institute of Pharmaceutical Sciences, Swiss Federal Institute of Technology
Wolfgang-Pauli-Str. 10, ETH Honggerberg, HCI G396, 8093 Zurich, Switzerland
B3 FSU Jena, WP 2 leader
Alexander Berndt
Institute of Pathology, Friedrich Schiller University
Ziegelmuhlenweg 1, 07740 Jena, Germany
B4 KTECH WP 4 leader
Matteo Zanda
Scientific Direction, KemoTech s.r.l.
Via Roma,72 Cagliari 01923, Italy
B5 PHC WP 8 leader
Eveline Trachsel
Therapeutic Antibody Research, Philochem AG c/o ETH Zurich, Institute of Pharmaceutical Sciences
Wolfgang-Pauli-Str. 10 HCI E250 CH-8093 Zurich
B6 SENDO WP9 leader
Silvia Marsoni
SENDO Foundation
Via Visconti di Modrone, 12; 20122 Milan, Italy
B7 UCL WP3 leader
Barbara R. Pedley
Department of Oncology, UCL Cancer Institute, Paul O'Gorman Building, University College London
72 Huntley St, London WC1E 6BT, United Kingdom
B8 ULG WP1 leader
Vincent Castronovo
Faculte de Medecine, Laboratoire de Recherche sur les Metastases, Universite de Liege
Avenue de l'Hopital 3, 4000 Liege, Belgium
B10 VUA WP7 leader
Guus van Dongen
Laboratory for Tumor Biology, Department of Otolaryngology, VU University Medical Center,
Vereniging voor Christelijk Hoger Onderwijs Wetenschappelijk Onderzoek en Patientenzorg De Boelelaan 1081 HV, Amsterdam, the Netherlands.
The human monoclonal antibody F16, labelled with the radionuclide iodine-124, has been studied in a phase 0 clinical trial for the positron emission tomography (PET) imaging of patients with cancer. The newly developed monoclonal antibodies have been chemically modified to serve as delivery vehicles for bioactive payloads (cytotoxic drugs, radionuclides or cytokines). The therapeutic potential of the resulting armed antibodies has been studied, both alone and in combination with other drugs, in rodent models of cancer. This led to the development of novel biopharmaceutical agents which could eradicate tumours that are not cured by conventional pharmacological interventions. Some of the newly developed products and combination treatments are now being studied in clinical trials. The project has led to an improved understanding of the molecular composition of the tumour neo-vasculature and of the tumour stroma. The project allowed the development of novel anti-cancer products, stimulating industrial activities in Europe and leading to the implementation of novel therapeutic concepts and of new clinical trials.
Project context and objectives
Within ADAMANT, the development of innovative antibody-based pharmaceutical products targeting the tumour neo-vasculature and the tumour stroma has required innovation in three main experimental parts. In the first part, researchers aimed at providing a comprehensive analysis of proteins which are preferentially expressed in the neo-vasculature of solid tumours and of lymphomas. Researchers have used a chemical proteomics methodology, invented by members of the consortium, for the identification of accessible vascular markers in normal tissues and at sites of disease. The technology relies on the terminal perfusion of animal models of pathology with a reactive derivative of biotin, which covalently modifies accessible proteins.
Alternatively, the biotinylation reaction can be performed by ex vivo perfusion of surgically resected human organs with cancer. Biotinylated proteins can be purified on streptavidin resin in the presence of strong detergents, digested and the resulting tryptic peptides separated by HPLC and analysed by mass spectrometry in the presence of internal standards for quantification purposes. This research activity led to the development of an 'Atlas' of vascular proteins expressed in various tumour types, such as solid and metastatic tumours or lymphomas. A transcriptomic study of gene expression in endothelial cells under different angiogenic stimuli has revealed a complementary set of candidate markers of angiogenesis. The most promising target proteins have been validated using immunochemical techniques, in part thanks to the development of new monoclonal antibodies.
In the second portion of experimental activities, human monoclonal antibodies have been isolated from large phage display libraries. These new reagents have been extensively validated using immunochemical techniques (e.g. immunohistochemistry, immunofluorescence) in tumour specimens and normal tissues. Many new antibodies have been studied using nuclear medicine techniques, such as imaging or biodistribution studies in tumour-bearing mice.
In the third set of experimental activities, novel biopharmaceuticals have been developed by arming the most promising recombinant antibodies with three main classes of bioactive payloads:
(i) drugs with cleavable linkers;
(ii) radionuclides;
(iii) cytokines.
Novel cytotoxic drugs have been developed (e.g. tubulysin or dolastatin analogues), which incorporate chemical handles for the site-specific coupling to recombinant antibodies. Novel procedures have been invented and implemented for the radiolabelling of antibodies for imaging and therapy applications. The consortium has systematically investigated the field of antibody-cytokine fusion proteins (immunocytokines), which has revealed impressive therapeutic activities, especially when used in combination with other judiciously selected pharmaceutical agents. Some of the most impressive curative effects, which have resulted in the complete eradication of aggressive tumour types in rodents, include the combination of IL2-based immunocytokines with taxanes or with monoclonal antibodies which are already used in clinical practice.
The project has included accompanying translational activities to facilitate the implementation of the newly developed products and methodologies in the clinical practice. An immuno-PET Phase 0 clinical trial for the imaging of cancer patients with the F16 antibody (specific to the alternatively spliced A1 domain of fibronectin) labelled with iodine-124 has been started. The consortium has collaborated with new biopharmaceuticals for the preparation of clinical trial protocols and for the evaluation of clinical results targeting the tumour neo-vasculature and stroma, thus contributing to an increased awareness of the opportunities and problems associated with this novel class of pharmaceutical agents.
Project results
Identification of new tumour-associated antigens, accessible and selectively present in human primary cancer lesions and metastases, by in vivo and ex vivo chemical proteomics:
- characterisation of accessible proteins in different primary malignancies and associated metastasis;
- development of chemical proteomic methods to study the accessible proteins in human tissues.
Overview
A significant limitation of the most currently applied anti-cancer treatments is related to their toxicity. Chemotherapy relies on the expectation that anti-cancer drugs will preferentially kill rapidly dividing tumour cells. Unfortunately, they also target normal cells, particularly cells from tissues with high turnover such as hematopoietic progenitors and epithelial intestinal cells. This toxicity limits both the treatment duration and the deliverable drug doses, which significantly restricts the therapeutic efficiency needed to obtain and maintain a complete remission. One of the most attractive strategies for the development of new, efficient and selective treatments is the delivery of targeted bioactive molecules selectively to the tumour microenvironment. This strategy is based on ligand binding with high affinity and specificity to target molecules, over-expressed at disease sites. Following this rationale, the discovery of protein targets readily accessible through the bloodstream and selectively over-expressed in pathological tissues has become a major research objective. This group of molecules has a high potential to serve as innovative tools for effective imaging and targeted cancer therapy approaches. Ideally, reachable cancer proteins (accessible biomarkers) of high therapeutic value should be expressed solely by the malignant tumour and be at least not accessible in normal tissues. In the frame of the project, researchers have embarked on the development of suitable pre-analytical methods to study accessible proteins in human tissues.
Following this, members of the consortium have applied the newly developed technologies in order to characterise a variety of primary malignancies and associated metastasis both in human clinical samples and suitable in vivo models. Along these lines, the original methodology was developed in a previous EU collaborative project (STROMA 2004-08) by ETHZ in collaboration with IRFMN who applied a chemical proteomic strategy, based on the terminal perfusion of tumour-bearing rodents with reactive ester biotin labels and subsequent labelled-protein recovery including mass spectrometric analysis. The major aim was to establish an atlas of vascular proteins in health and disease. This technology was further improved in collaboration between ETHZ and ULG, yielding novel methods where first human tumour-bearing organs were ex vivo perfused and latter tumour biopsies were in vitro labelled in order to gain access to potentially targetable proteins. In the frame of the project, the technique has improved, mostly in terms of being more quantitative. Recently, the technology has been further refined by involving glycosylated proteins, in addition to biotinylated ones, in order to complement the repertoire of potentially accessible biomarkers.
Once the suitable methodologies were established, the partners moved on to characterise a broad range of malignancies and to establish the map of potentially accessible tumour biomarkers suitable for targeted therapy. The research was primarily directed on colon, liver (metastasis), pancreas, breast, bone (metastasis) and glioblastoma tumours from patients. The method which combines biotinylation with glycoproteomics has been applied on the breast cancer tissue: a novel biomarker has been discovered (CD276) and validated in more than 30 individual cases. The investigation of a breast tumour-associated bone metastasis was finalised. This study had the unique opportunity to examine bone metastasis and the corresponding breast cancer primary lesion obtained simultaneously from a fresh autopsy performed on a patient who died from disseminated breast cancer. The accessible protein biomarkers were identified using the previously developed combinatory procedure of biotinylation and glycoprotein extraction. 29 of the potential biomarkers were found uniquely expressed in the primary breast cancer while 27 proteins were detected only in the bone metastasis lesions. Several proteins belonging to small leucine rich proteo-glycans, thrombospondin and integrin families were found upregulated in the primary breast tumour.
The proteomics technology has been applied to characterise pancreas ductal adenocarcinoma (PDAC) and its accessible protein biomarkers. Mass spectrometry (MS) analysis revealed 484 differentially expressed proteins, of which 84 were evaluated as potentially accessible. 11 selected candidates were confirmed by Western blot and multiple reaction monitoring (MRM)-MS analysis. Transforming growth factor beta-induced (TGFBI), latent transforming growth factor beta binding protein 2 (LTBP2) and asporin (ASPN) were further investigated employing immunohistochemistry. They were found to be significantly expressed in a large group of clinical PDAC samples compared to corresponding normal and inflammatory tissues. In conclusion, TGFBI, LTBP2 and ASPN are novel, overexpressed and potentially accessible proteins in human PDAC. Human glioblastoma specimens as well as tumours developed in nude mice (xenografted human glioblastoma U373 and T98G cells) along with normal mouse brains were biotinylated. These samples were analysed and mutually compared. An easy and cost-effective method to validate the potential accessibility of tumour antigens in vivo has been developed. This assay is conducted with xenografted U87 human glioblastoma cells on the chicken egg chorioallantoic membrane (CAM). Antibodies against suitable targets can be injected intravenously in the CAM vasculature and visualised ex vivo using immunofluorescence.
The current study has highlighted several known and novel proteins implicated in the human glioblastoma: tenascin-C, CD44, collagen-VI-A1, sparc-like-1 and prosaposin. Immunohistochemical validation experiments confirmed the expression of these proteins in large series of human glioblastoma cases. Specifically anti-CD44 and anti-collagen-VI-A1 have been successfully injected/ validated in the CAM/ U87 model. Ex vivo perfusions of surgically resected human colon cancer (CRC) using biotin derivative, followed by comparative MS-based proteomic analysis of the labelled proteins, revealed quantitative differences between normal and cancer colon. 67 of the total 367 proteins identified were found to be preferentially expressed at the tumour site. Human monoclonal antibodies against two potential tumour targets, NGAL and GW112, were generated and their selective expression in cancer colon and not or barely in healthy tissues proved. In over the half of CRC patients a metastatic disease is present at the time of diagnosis, reducing the five-year survival to less than 5 %. One of the most common sites of metastasis is liver, characterised as asymptomatic until the disease is well progressed. An effective treatment and early diagnosis modality for the CRC liver metastasis would make a significant difference for a large population of patients. Researchers aimed at investigating whether the metastasis displays a homogeneous distribution of certain biomarkers. Some biomarkers are only found in distinct regions of the lesion.
Consequently, they have ex vivo biotinylated accessible proteins from several CRC-liver metastases and divided the specimen in 4 zones: Normal, peri-tumoural, tumour-rim and centre. The proteins were affinity purified and analysed for each zone separately using MS. In total over 1500 unique proteins were statistically divided into six patterns of expression. Approximately 1/3 was expressed solely in one of the four zones. A further 1/3 was found in all zones. Remaining proteins were present in 2 or 3 regions studied. Several known and new markers have prompted our interest among which: periostin (POSTN), TGFBI and latent transforming growth factor beta binding protein 2 (LTBP2).
The found antigens were validated in a mouse model of CRC liver metastasis. Mice liver metastasis and corresponding normal tissues were biotinylated and analysed by MS. The investigation revealed 922 mouse proteins present solely in the normal liver, 704 human proteins present in the metastasis and 248 proteins (mouse and human origin) found in common between normal mouse liver and mouse liver metastasis. 70 of the in common proteins were up-regulated in the tumour. Researchers have assessed the potential accessibility of the identified proteins and found that 110 proteins qualify as differentially expressed and potentially accessible. The data were compared with results obtained from the analysis of human liver metastases and resulted in 64 proteins which were modulated in the mouse model as well as in the human liver metastases. Further validation studies are ongoing.
An in vivo biotinylation of vascular structures by terminal perfusion was performed in three different syngeneic mouse models of CRC liver metastasis. The recovery and mass spectrometric analysis of accessible biotinylated proteins in liver metastasis revealed quantitative differences in the expression of vascular proteins compared to the host organ. The preferential expression of 10 markers at metastatic sites was confirmed by immunofluorescence analysis and, in the case of periostin, oncofoetal fibronectin and angiopoietin-like 2 proteins, by in vivo targeting experiments using radiolabelled antibody preparations [Borgia et al., Cancer Research, Cancer Research 2010].
Antibodies production and target validation:
- generation and production of new recombinant human antibodies against tumour specific antigens;
- in situ validation of antibody reactivity, specificity and target distribution in human tissue.
Overview
Tumour vessel and tumour environment specific molecules can serve for therapeutic purpose and can be developed as targets for antibody-based pharmacodelivery. An important precondition for their possible future clinical exploitation is the assessment of the 'novel' target distribution and of the specificity/reactivity of the newly produced antibody against it. Following the discovery of novel tumour specific accessible antigen(s), a key task is the generation of antibodies and their deep investigation. Immunohistochemical staining methods were optimised and the antibody reactivity tested in a wide range of normal, tumorous and non-tumorous pathologies and the distribution of the accessible antigen evaluated. As a result several newly produced human recombinant antibodies could be characterised and new putative tumour specific targets could be validated, as a prerequisite for the preclinical studies aimed at assessing therapeutic efficacy of the conjugated antibodies.
Target selection and first level of validation
Antibody production
Philochem has implemented the transient gene expression (TGE) as a standard methodology for the simple, versatile and fast (from months to weeks) generation of proteins against which antibodies should be raised. Using this method, several antigens could be expressed. For recombinant antibody production, two new synthetic human antibody libraries (named PHILO1 and PHILO2) were cloned to broaden the epitope recognition in respect to the previously characterised antibody library (ETH2-GOLD). A fully murine antibody library (PhiloTOP) was produced allowing the production of murine antibodies against markers which are not immunogenic in the mouse (e.g. EDB). Such antibodies serve as research tools in order to test long-term therapy effects of fully murine proteins in immunocompetent mice. With the aid of all these technologies/methodologies, recombinant antibodies against the selected stromal and vascular antigens were successfully produced. Among them are high affinity antibodies against human ED-A and ED-B fibronectin, MG50, periostin, uPA, human carbonic anhydrase IX (CAIX) and the vascular marker of B-cell lymphoma BST-2, as well as against murine matrix metalloproteinases 1, 2, and 3 (MMP1, 2, and 3) and EDB domain of fibronectin.
Target distribution and antibody reactivity validation
The tissue distribution of these putative target structures was investigated in detail. For a variety of the markers, a nice tumour or tumour vessel restricted occurrence could be proven. The tissue reactivity of the recombinant antibodies L19, F8, F16, G11 and D11 directed against fibronectin and tenascin-C variants were comparatively examined. For histological analysis a thyramid based amplification system was introduced leading to sensitive and reproducible immunofluorescence staining results for the biotinylated recombinant antibodies. Fluorescence double and triple staining procedures combining different recombinant antibodies were designed to investigate putative target structures in relation to vessels, stromal components, tumour cells and basement membrane structures as well as in relation to tumour type, subtype, malignancy grade and progression. It could be demonstrated that fibronectin and tenscin-C variants show different expression patterns discriminating tumours with wide stromal deposition from tumours with a restricted vascular deposition. The incorporation of fibronectins or tenascins in the blood vessel wall of the tumour was different too. These observations are of special interest for the selection of candidate tumours for treatment with fibronectin or tenascin-C antibodies and for individualised therapy strategies.
Clinically valuable expression patterns could be described for melanoma, lymphoma, lung carcinoma and renal cell carcinoma. For the first time, the spatial relation of vascular fibronectin and tenascin-C deposition in relation to vascular BM was defined by laser scanning microscopy colocalisation studies showing a stratified organisation of the vessel wall matrix. With respect to the preclinical studies aimed at assessing the efficacy of the antibodies as therapeutics, the reactivity of several recombinant antibodies against fibronectin and tenascin-C isoforms as well as against MMPs, uPA and CA IX was investigated in human tumour xenografts and in syngeneic mouse tumours, with particular focus on renal cell carcinoma, ovarian cancer, melanoma and lymphoma. Tumours strongly vary in their stromal and vascular expression of antigens recognised by the recombinant antibodies depending on the tumour type, the mouse strain (immunological background) and the species reactivity of the antibodies. Experimental tumours reflecting the results obtained with the clinical specimens were candidate to undergo experimental targeted therapy.
Target distribution in non-neoplastic pathologies
Targeted therapy of human neoplastic diseases using the recombinant antibodies F16, F8 and L19 (the best candidates for clinical development) might cause systemic side effects in healthy organs or in other non-neoplastic pathologies if there is an expression of the antigen/target molecules. Since no reactivity could be appreciated by analysing the collection of healthy tissue available commercially, it was a central task to analyse non-neoplastic pathologies. Because of the high incidences of both neoplastic and cardiovascular diseases there are high co-morbidity rates, the reactivity of the antibodies was assessed in different cardiac diseases to evaluate the risk of side effects and to possible identify new fields of potential diagnostic or therapeutic use. Studies were focused on evaluating the antigens recognised by the F16, F8 and L19 antibodies in pathologies of the cardiovascular system. F16 antibody stained areas of active tissue remodelling in atherosclerotic plaques (in both human and animal models) and may thus deserve to be exploited as a suitable building block for the development of radiopharmaceuticals for plaque imaging or for an antibody-based targeted delivery of therapeutic agents to atherosclerotic lesions.
Diseased human cardiac tissue of the right atrial auricle (RAA) and left ventricular septum (LVS) derived during cardiac surgery from patients with aortic valve stenosis (AVS) and/or coronary artery disease (CAD) were systematically analysed. A re-expression of F8 and F16 antigens could be observed, together with quantitative differences in the expression levels between AVS and CAD as well as in association to the grade of histological damage. Thus, F8 and F16 were suggested to be of potential interest for an antibody-mediated targeted delivery of diagnostic or therapeutic agents in human cardiac diseases with respect to anti-inflammatory or anti-fibrotic therapy strategies. The reactivity of the recombinant antibodies was investigated in a heterotopic rat heart transplantation model of chronic cardiac allograft rejection. A relevant reactivity was demonstrated for F8 but not for L19, F16, D11 and G11. Tissue reactivity revealed an extensive antigen expression (ED-A fibronectin) in heart allografts exhibiting signs of chronic rejection compared to healthy controls with clear spatial association to vessels showing cardiac allograft vasculopathy (CAV) and to areas of cardiac interstitial fibrosis (CIF), confirming ED-A fibronectin as a promising marker of chronic rejection (CAV and CIF) and F8 antibody as a vehicle for targeted therapy strategies.
Radioimmunotherapy (RIT)
- Highly selective tumour localisation of novel radio- or fluorescently labelled antibody conjugate.
- Development of effective radioconjugates for preclinical PET.
- Preclinical antitumor activity of a radioconjugate causing significant growth inhibition.
Overview
The use of tumour-targeting antibodies to deliver therapeutics to the tumour site is a highly selective form of cancer therapy, reducing the side effects frequently seen after conventional treatments while increasing the specificity. The aim is to develop novel and effective agents, using antibodies that target either the tumour vasculature or the tumour cells to deliver radiation to the tumour (radioimmunotherapy (RIT)). The antibodies are delivered into the circulation and can therefore reach sites in the body which are inaccessible to external beam radiation. Initial investigations, to confirm tumour specificity, are performed with fluorescently labelled antibodies. The antibody F8-SIP directed against ED-A fibronectin selectively targets tumour blood vessels, as can be demonstrated by multifluorescence microscopy. The antibody shows the same localisation as the vasculature over the whole tumour, indicating that the vast majority of blood vessels will be targeted for therapy, and when looked at in high power it has a highly specific perivascular distribution in colorectal tumours. The related antibody L19-SIP, when conjugated to radioactivity (125I-L19-SIP), produces a significant therapeutic effect. Researchers have been investigating ways to further increase this by using a combination of therapies, which treat different regions of the tumour. The vascular disrupting agent (VDA) combretastatin, currently in clinical trial, produces massive destruction of the centre of the tumour, but leaves a viable vascular rim which continues to grow. Combined therapy is therefore required to kill the whole tumour mass. We combined these 2 treatments, giving either radiolabelled antibody alone, or with combretastatin administered at 24 hours earlier and imaged the effect of the VDA on radiolabelled antibody distribution across the tumour. In the antibody alone group the tumour was viable and the antibody was distributed heterogeneously throughout, with the hottest areas shown in red. When combretastatin was also administered, the tumours developed massive central necrosis and the antibody was concentrated within the remaining viable rim where it could be most efficacious. As the total radioactive dose to the tumour was the same in both cases, researchers have achieved a complementary combined treatment with great therapeutic potential, which will form the basis of future radioimmunotherapy trials.
Development of effective radioconjugates for preclinical PET (immuno-PET)
There is a nuclear medicine requirement for non-invasive imaging of cancer patients in order to determine whether radiolabelled antibodies (radioimmunoconjugates) are successfully targeting, and being retained in, the tumour while clearing from normal organs and to determine target status. Procedures for development of these conjugates for single-photon emission computed tomography (SPECT) imaging are routinely available, but this is not the case for PET imaging. Researchers have therefore concentrated on developing standard procedures for producing suitable PET isotopes (positron emitters: 68Ga, 124I and 89Zr) and conjugating them to tumour-targeting antibodies. New linkers have been developed and logistics established for worldwide distribution of both isotopes and linkers, while labelling protocols were published for dissemination of this so-called immuno-PET technology.
Several preclinical and clinical immuno-PET studies for the non-invasive in vivo assessment of target status and antibody distribution have been successfully carried out. The potential of immuno-PET for quantitation of antibody distribution in patients has been evaluated. These achievements are of interest when using immuno-PET for selection of RIT candidates. This immuno-PET technology is disseminated worldwide and many leading academic centres and pharmaceutical companies use our technology in the (pre)clinical development of biotech products. While nuclear imaging like SPECT and PET is particularly well suited for quantitative imaging of deep-seated tumours, optical imaging using fluorescent dyes can provide complementary clinical potential for imaging superficial tissues. For this purpose a dual modal nuclear / optical imaging probe is being developed, in which nuclear imaging is used for whole body imaging and optical imaging for local imaging. Since angiogenesis is an early event in tumour invasion, radioactively-fluorescently labelled dual modal ligands directed against vascular tumour targets might be of diagnostic value and a guide for RIT in future trials.
Radioconjugate producing significant tumour growth inhibition
The major cause of death from colorectal cancer is the spread of disease to form liver metastases. Researchers have therefore investigated the effectiveness of RIT in this clinically relevant site, using an antibody that targets a common marker of colorectal tumour cells called carcinoembryonic antigen (CEA). They initially attached a red fluorescent label to the antibody, and were able to demonstrate highly specific uptake in all tumour metastases, regardless of size, with none in the surrounding normal liver. Radiation was attached to the antibody for a therapy experiment and researchers found that they could significantly inhibit the growth of tumours in the liver, without concomitant toxicity. They found that smaller liver metastases had far greater uptake of the therapeutic antibody than larger deposits. This system is being optimised for treatment of small liver metastases in the clinic, where the prognosis is still very poor.
Antibody drug conjugates
Generation of novel and effective antibody-drug conjugates, composed of either anti-tumour cell, anti-vascular or anti-stroma antibodies conjugated to tubulysins or dolastatin-15 analogues.
Overview
The availability of potent cytotoxic drugs having low-nanomolar or sub-nanomolar activity that can be conjugated to a monoclonal antibody through suitable chemical functions was the key for this part of the project. It was decided to focus on three different classes of potent drugs, all having a natural origin: Tubulysins and dolastatins, which are both peptides, and vinblastine, which is an alkaloid. These molecules, which are unsuitable anti-cancer drugs as such because of their high toxicity and narrow therapeutic window, were synthetically modified to:
i) achieve the appropriate level of chemical and metabolic stability, and
ii) install chemical functions allowing for the attachment to the antibody via cleavable chemical linkers.
The drugs might therefore be selectively released exerting a potent anti-tumour cell, anti-vascular or anti-stroma activity.
Almost 50 novel derivatives of highly potent compounds were synthesised, some in large quantities and high purity. These compounds were made available for biological studies on cancer cell lines and tumour animal models and for experiments of antibody-functionalisation. Some of the most potent tubulysins, dolastatins and vinblastine analogues were linked to human monoclonal antibodies or tumour homing peptides through disulfide or peptide linkers, thus obtaining several different conjugates. These conjugates, which were shown to have antibody-dependent anti-tumour effects in vitro, with IC50 values in the nanomolar range, were tested in vivo. Although the results are still preliminary, moderate retardation of tumour growth was observed with some of these conjugates, thus reinforcing vascular targeting antibody-drug conjugates as promising cancer therapy strategy. General recombinant antibody formats, which can be chemically modified with drugs and cleavable linkers at unique molecular positions, were successfully engineered.
Immunocytokines
- Cloning, expression and in vitro characterisation of novel immunocytokines
- In vivo testing of the therapeutic activity of novel immunocytokines.
Overview
Many pro-inflammatory cytokines have been used for cancer therapy applications in clinical trials, but severe side effects have often prevented dose escalation to concentrations needed to induce cures. The fusion of cytokines with antibodies or antibody fragments, capable of selective accumulation at the tumour site, yields a new class of biopharmaceuticals (termed immunocytokines), which often displays superior therapeutic properties. In the frame of the project, researchers have produced and characterised several novel immunocytokines, based on IL2, IL7, IL10, IL12, IL17, IL18 and interferon-alpha as potent stimulator of the immune system. The resulting fusion proteins have been characterised in vitro (in terms of their pharmaceutical quality and activity) and in vivo (in terms of their tumour-targeting performance and therapeutic activity). Some of the tested immunocytokines (most notably those based on IL2, IL12 and IL15) have exhibited an impressive ability to inhibit tumour growth rate when used as single agent and to display superior performance compared to immunocytokine based on antibodies of irrelevant specificity in the mouse. The antibodies used as fusion partners for the production of vascular tumour-targeting immunocytokines were F8 and L19 (specific to the alternatively spliced EDA and EDB domain of fibronectin) and F16 (specific to the alternatively spliced A1 domain of tenascin-C). As negative control, monoclonal antibodies specific to hen egg lysozyme, an irrelevant antigen in the mouse, were used.
The fusion proteins that were cloned, expressed and characterised (in vitro and in vivo) were many more than the four mentioned in the milestones. The fusion protein F16-IL2 has exhibited a strong anti-tumour effect both in a subcutaneous and in an orthotopic model of glioblastoma multiform. A microscopic analysis of tumour sections in mice treated with saline, temozolomide, F16-IL2 or a combination of temozolomide plus F16-IL2 has evidenced that the immunocytokine promotes a strong infiltration of leukocytes into the tumour mass. Interleukin-12-based immunocytokines has been investigated for many years by the ETH Zurich and Philochem scientists. This heterodimeric cytokine has traditionally provided challenges for protein expression (the N-terminus of the p40 subunit needs to be free for maximal biological activity), but also opportunities for the design of new formats. A novel disulfide-linked format of an immunocytokine consisting of the anti-EDA antibody fragment scFv (F8) and of human IL12 has recently been developed by the Philochem group and has been moved to GMP manufacturing. The protein displayed excellent pharmaceutical quality and impressive tumour-targeting performance in xenograft tumour models in rodents.
Clinical significance
The clinical significance of research in the field of vascular tumour-targeting immunocytokines is reinforced by the clinical results recently observed by the Philogen group, who has brought three immunocytokines (L19-IL2, L19-TNF, and F16-IL2) to Phase II clinical trials in patients with various cancer types. The most advanced of these products is L19-IL2, which has exhibited numerous objective responses in a Phase IIa study with 32 patients and which is currently being investigated in a controlled phase IIb in 90 patients with metastatic melanoma, in combination with dacarbazine. Even though these clinical activities were not part of the project, the positive results observed so far provide a strong rationale for research in the immunocytokine field.
Optimisation of combination therapy with antibody conjugates (immunocytokines)
- Preclinical antitumor activity of combination treatments for cytokine conjugates
- Anti-metastatic activity of immunoconjugates in combination with chemotherapy in one preclinical model.
Overview
The development of antibody derivatives in clinical trials relies on their use in combination with other anti-cancer treatments. In the frame of project, members of the consortium have investigated the immunoconjugates described above (i.e. F16-IL2, F8-IL2, F9-IL2) in combination with standard-of-care chemotherapy or targeted drugs. A great effort to optimise the best treatment conditions on ad hoc mouse tumour models reflecting the tumour patient's characteristic was made. Representative treatment regimens have yield cures in tumour-bearing mice and represent promising candidates for clinical development.
F16-IL2 immunocytokine in combination with Temozolomide on glioblastoma
Temozolomide, the standard of care for newly diagnosed glioblastoma patients, in combination with F16-IL2 - a clinical-stage immunocytokine consisting of human interleukin (IL)-2 fused to the human antibody F16, specific to the A1 domain of tenascin-C- was investigated on a model of glioblastoma. Immunohistochemical analysis with human glioblastoma surgical specimens and with U87 xenografts showed a strong and selective tumour staining using the F16 antibody. A quantitative biodistribution in nude mice confirmed the preferential accumulation of the radiolabelled F16-IL2 at the tumour site. In the therapy study, the combination of F16-IL2 with temozolomide induced a complete tumour remission in the animals. The same treatment led to a substantial size reduction of the U87 tumours growing orthotopically in the brain, compared to temozolomide or the immunocytokine administered alone. The combined use of temozolomide and F16-IL2 may deserves clinical investigations, which will be facilitated by the excellent safety profile in Cynomolgus monkeys and by the fact that F16-IL2 is currently being investigated in clinical trials in patients with metastatic cancer.
F8-IL2-immunocytokine in combination with taxanes on melanoma model
Human melanoma is characterised by strong vascular and stromal positivity for the EDA domain of fibronectin (EDA-Fn) and the pattern of EDA-Fn expression in melanoma xenograft models fully replicated that of the patient lesion. We found that F8 -an antibody directed against EDA fibronectin- conjugated to interleukin 2 (F8-IL2) and combined with Paclitaxel (PTX), but not decarbazine (DTIC), completely inhibited the growth of human melanoma xenografts transplanted intradermally, with 80 % complete regressions. Researchers showed the particular role of Paclitaxel in this combination regimen. PTX increased tumour vessel perfusion and permeability in melanoma xenografted tumours, favouring the uptake of F8 antibody and PTX prones the infiltration of NK cells and macrophages in melanoma xenograft. These preclinical studies, showing complete tumour regressions and anti-metastatic activity, endorse the use of chemo-immunotherapy in the treatment of malignant melanoma and suggest that PTX is endowed with properties affecting the tumour stroma that deserve attention for the design of treatment regimens.
L19-IL2 immunocytokine in combination with rituximab on non-Hodgkin lymphoma (NHL)
The antibody-mediated delivery of therapeutic agents was investigated in hematologic malignancies. The EDB domain of fibronectin, is expressed in B-cell NHL and the human monoclonal anti-EDB antibody L19 can selectively localise to the lymphoma-associated subendothelial extracellular matrix. In vivo, the preferential accumulation of the antibody at the tumour site was confirmed by quantitative biodistribution analyses with radio-iodinated antibody preparations. The fusion protein L19-IL2, which mediates the delivery of interleukin-2 to the neovasculature, displayed a superior antilymphoma activity compared with unconjugated IL2 in localised and systemic xenograft models of NHL. When co-administered with rituximab, L19-IL2 induced complete remissions of established localised lymphomas and provided long-lasting protection from disseminated lymphoma. The combined use of rituximab and L19-IL2, which dramatically increases the infiltration of immune effector cells in lymphomas, may deserve clinical investigations, facilitated by the fact that L19-IL2 is currently being studied in phase II clinical trials in patients with solid tumours [Schliemann et al, Blood 2009].
PET clinical validation (microdosing)
The aim is the clinical evaluation of a promising anti-vascular antibody in immuno-PET microdose phase 0 trial.
Overview
Immuno-PET has the potential to become a powerful imaging tool for efficient clinical antibody development because this technique allows accurate quantification. With relatively few patients the antibody can be characterised for tumour-targeting capacity, and homogenicity of tumour-targeting, possible cross-reactivity with normal tissues, residence time in blood, tumour and normal tissues, possible complex formation in blood, catabolic routing and inter-patient variation. Microdosing (phase 0) studies as described by the European Agency for the Evaluation of Medicinal Products (EMEA), could be used to compare upfront compounds with similar intended Mode of Action but potentially different PK/PD characteristics, thus weeding out unsuitable agents for clinical development prior to extensive clinical investigations in patients. These studies do not need the full set of toxicity and other safety studies. Researchers anticipated that for antibodies the microdosing principles will be applicable, especially when there is evidence that MAb biodistribution will not be strongly dependent on MAb dose. The aim of was the clinical evaluation of a promising anti-vascular antibody in an immuno-PET microdose phase 0 trial.
124I-F16SIP: aphase 0 single microdose study
Based on available preclinical data, F16SIP directed against the alternatively spliced domain A1 of tenascin was selected for phase 0 studies in operable head and neck cancer patients. All consortium partners were involved in the design of the study. The following EMEA-released guidelines and position papers were considered: 'Note for guidance on the preclinical evaluation of anti-cancer medicinal products', 'Note for guidance on the preclinical evaluation of biotechnology-derived products', 'Note for guidance on non-clinical safety studies for the conduct of human clinical trials for pharmaceuticals', and 'Position paper on non-clinical safety studies to support clinical trials with a single microdose'. Since the clinical trial was planned to be performed at VU University Medical Center in the Netherlands, discussions were performed with Dutch regulatory authorities about the non-clinical safety studies needed to allow a clinical immuno-PET study with a single low dose of antibody. It became clear that 30 nMol MAb is allowed for phase 0 clinical trials and that toxicity studies performed with cold F16SIP are also valid for studies with radiolabelled F16SIP, as long as the binding and pharmacokinetic properties of the MAb are not altered by the labelling procedure.
Based on this information, the following steps were made:
i. GMP produced clinical-grade batch of F16SIP was made available with the consortium;
ii. non-clinical safety studies were performed according to aforementioned regulation;
iii. a tissue cross reactivity test was performed according to FDA recommendations;
iv. procedures and a manufacturing infrastructure for GMP production of 124I-F16SIP were set up. An Investigation Medicinal Product Dossier (IMPD) of 124I-F16SIP was written;
v. a clinical protocol entitled: 'A phase 0 single microdose study to evaluate the pharmacokinetics / -dynamics and specific tumour-targeting of 124I-F16SIP in head and neck cancer patients', was developed and approved by METC;
vi. three head and neck cancer patients were enrolled and treated in this trial and two other patients are planned.
The first patient had a T2N0M0 lateral tongue carcinoma on the left side. The first PET-CT scan at 30 minutes after injection showed mainly blood pool. The second PET-CT scan at 24 h after injection showed clear uptake of 124I-F16SIP in the left lateral tongue, corresponding to the localisation of the primary tumour. During surgery, at 168 h after injection samples were taken from blood, tumour, skin, healthy mucosa, fatty tissue and muscle. The highest uptake was found in tumour tissue with a tumour-to-blood ratio of 8.2. The second patient had a T2N0M0 check carcinoma. The second PET-CT scan at 24 h after injection showed some uptake of 124I-F16SIP in the tumour area, while biopsies at 120 h after injection revealed a tumour-to-blood ratio of 6.6. Because of slow patient accrual, the original study protocol was amended in a way to limit the number of PET scans to two in order to avoid that patients have to travel to the clinic on several days in a row. In the meantime a third patient has been included in this phase 0 trial, and the trial is expected to be finished within a couple of months. Results from this phase 0 trial thus far show that F16SIP is a promising antibody for targeting antigen-positive tumours like head and neck tumours. This antibody is candidate to be used for tumour detection and when coupled to a toxic payload for selective tumour eradication.
Regulatory affairs and strategic monitoring
- Design of clinical development plan and protocol for phase 1 first-in-man (FIM) study
- Micro-dosing study protocol
- Organisation of a multidisciplinary oriented focus group (FG).
Overview
The mission of ADAMANT was to discover new potentially 'druggable' anti-cancer targets in solid tumours and generate highly effective new monoclonal antibodies against these targets, minimising collateral damage to the normal tissues by homing the delivery of toxic agents (chemical or natural products molecules, radionuclides or immunostimulatory cytokines) specifically to the cancer cells or the cells of the cancer-supporting scaffold (stroma) to the cancer cells or the cells of the cancer-supporting scaffold (stroma). The challenge of the project was to identify these new potentially druggable targets in a preclinical framework mimicking the ultimate setting in which these antibodies would be used (i.e. the cancer patient) as much as possible.
Since the inception, it was recognised that the traditional 'bench to bedside' framework could be converted to a more innovative 'bed to bench-side' approach retrofitting the constructs of the preclinical experiments from the actual clinical settings. With this challenge in mind a task was devised to support the project and indeed relayed mainly on the collaboration with the more 'clinical oriented' tasks: Optimisation of the combined therapy with other agents; PET clinical validation and RIT. Southern European New Drug Organisation (SENDO) was involved in the project to provide, in connection with the performance of the preclinical activities, the expertise in clinical drug development and specifically in preparing dossiers, designing pharmacodynamic and pharmacokinetics based phase 0 and I trials and running them in cancer patients at selected and qualified clinical centres.
The organisational model of SENDO is that of an academia / industry go-between geared at understanding the needs and fulfilling the missions of both stakeholders. During the project, SENDO promoted and supported innovative anti-cancer drug research stimulating interaction with industry in the early stages of drug development. Working through its SME, SENDO TECH, the foundation is committed to guarantee the highest level of quality research, validity and reliability of clinical data collected. They have in place established standard operating procedures (SOPs) for data collection, logistical organisation of the trial, monitoring activities and medical writing. SENDO TECH is a certified CRO in Italy and France since 2009 and 2010, respectively.
Within ADAMANT, the full quality-controlled radiological assessment of clinical response in the patients included in the phase I first in man studies (including the blinded revision of all the radiological assessments of patients treated with L19-IL2) was provided; in addition the protocol templates for phase 0 and I FIM studies and all kinds of quality controlled study report templates were designed and finalised. Based on that, the study report of the first FIM concluded was produced and clinical data have been presented at two international conferences during 2010: ASCO and ASTRO. Regulatory assistance to define controversial ethical and legal aspects of phase 0 clinical trials in Europe was provided. Within this framework the first PET phase Zero IMPD was finalised and a submission for a microdosing study was done by VUA. An Immuno-PET clinical study has been designed and activated. To facilitate the strategic monitoring of the results, multidisciplinary oriented focus groups have been organised to discuss and analyse the open questions on PET in translational oncology and on pharmacodelivery of cleavable linker's drugs. The discussions helped the creation of a 'platform of concepts' stemming from the project collective experiences and specific expertise. This platform could become both a conceptual signature of the project and a springboard for future scientific explorations of the participants. Some of the focus debated issues have been subjects of publications.
Exploitation of experimental results
To assure the visibility of the results, the consortium has disseminated information through various channels involving different targets. There are communications by the coordinator such as the website addressing the general public and two ads appearing in the magazine 'The Parliament', addressing European decision-makers, and dissemination activities that are done by individual participants. The project website has been available since the start of the project at http://www.adamant-fp7.eu and has been periodically updated, providing news of major events of the consortium and the end of the project. The website will be maintained online after the end of the project. All participating groups have published their research results either alone or together with other beneficiaries in peer-reviewed scientific journals. A total of 50 papers in peer-reviewed journals have been published (of which 14 open access), and more than 70 congress presentations given. More manuscripts (n = 14) are under preparation and submitted or accepted for publication. In several instances beneficiaries' scientific representatives have been invited as speakers at national and international meetings. These activities will continue as expected. A workshop on 'Tumor microenvironment and metatases' was organised by the Belgian Society for Cell and Developmental Biology in Liege, 27 March 2010 in collaboration with three Seventh Framework Programme (FP7) projects.
Personal training programmes were facilitated among the partners of the ADAMANT consortium:
- Fabia Doll, Master Student at ETH Zurich (B2), performed her Master thesis at UCL (B7) from May to July 2008;
- Dr Andrei Turtoi, scientist at ULG (B8), spent a four-month training period (from June to September 2009) at UCL (B7) in the frame of his postdoctoral training to learn the technique of orthotopic liver metastasis in murine models grafting colorectal tumour cells into the liver via splenic injection;
- Antonietta Silini performed her PhD thesis in cooperation between IRFMN (B1) and UCL (B7) and spent one month (July-August 2009) at UCL;
- Dr Sreejith P Shankar, postgraduate student at KemoTech (B4), spent a three-months secondment period from February to April 2009 at VUA (B10) during his PhD training;
- Francesca Pretto is a PhD student of ETH Zurich (B2), from January 2010 she is performing part of her research at IRFMN (B1).
Some 20 doctoral, PhD and graduation theses have been promoted by the beneficiaries on subjects connected with the project, 11 have been completed and 9 are still in progress. A total of six patent applications were filed with EPO and two more are under evaluation. On three occasions a research valorisation roundtable was held with the experienced support of the advisory board. The aim of this roundtable was to assure the assessment of the societal and industrial exploitation of results and best return on investment of the project. Progress presented by the ADAMANT consortium was analysed in detail by members of the advisory board, resulting in suggestions from people not directly involved in the project. Analyses were discussed in plenum and the advice allowed a focused continuation of the work until the next major meeting.
Potential impact
In spite of their commercial success, conventional monoclonal antibodies which are currently used for the treatment of solid tumours offer only an incremental benefit to cancer patients. For this reason, there is an urgent need for superior anti-cancer biopharmaceuticals. Virtually all large pharmaceutical companies are heavily investing into biomedical strategies for the arming of monoclonal antibodies with bioactive payloads.
The potential impact of the project can be analysed in terms of four main milestones achieved over the last three years:
i. the discovery and validation of novel markers of tumour neo-vasculature and stroma, followed by the generation and characterisation of human monoclonal antibodies;
ii. the development and implementation of novel antibody-based therapeutic strategies;
iii. the stimulation of innovative clinical trials for the treatment of patients with incurable forms of cancer;
iv. the creation of new jobs, in terms of formation of new companies and growth of existing companies.
Traditionally, markers of tumour angiogenesis and of the tumour stroma have been discovered by serendipity. Out of hundreds of different antibodies which are routinely analysed by immunohistochemistry, it would occasionally happen that the study of a new monoclonal antibody would reveal a staining pattern coinciding with tumour neo-vascular structures. Markers of angiogenesis, such as the EDB domain of fibronectin, endoglin, PSMA, have all been discovered this way. The project has contributed to the development and implementation of methods for the systematic characterisation of vascular proteins in health and disease. There now are atlases of vascular markers for various tumour types, which allow choosing a target for antibody product development based on an unbiased comparative evaluation of expression patterns for the first time. We are confident that the quality of the tumour-associated antigen used for product development will ultimately determine the quality of the cognate antibody-based product.
On the second aspect of ADAMANT's potential impact (novel antibody strategies), it is still not clear what the most successful strategy will be for arming antibodies. Different strategies have distinctive favourable features and carry distinctive liabilities. For example, radiolabelled antibodies may be ideally suited for the therapy of radiosensitive tumours (e.g. lymphomas), but are less effective for radioresistant tumours or for tumours which do not exhibit an adequate antibody uptake compared to normal tissues. For this reason, researchers have chosen to focus on three main antibody functionalisation strategies (radiolabelling, fusion to cytokines, coupling to drugs with cleavable linkers) and have made an impact in all these areas. The Philogen group (one of the partners in the consortium) has brought seven armed vascular targeting antibodies to the phase I and to phase II clinical trials. At present, the most promising clinical results with armed vascular targeting antibodies have been reported for the:
(i) radioimmunotherapy of patients with incurable forms of lymphomas; and
(ii) therapy of metastatic melanoma patients with IL2-based or TNF-based immunocytokines.
In the first case, multiple major responses have been observed in patients with Hodgkin's lymphoma, who had failed all other therapeutic options.
In the case of IL2-based immunocytokines, the treatment of metastatic melanoma patients with L19-IL2 plus dacarbazine has led to the doubling of median overall survival compared to historical control patients, treated with dacarbazine alone. This treatment modality is currently being investigated in phase IIb clinical trials in Germany, Italy and Austria. Importantly, the recent discovery that taxanes may be better suited for the combined use with IL2-based immunocytokines and that curative effects can be observed when Paclitaxel is administered before the immunocytokine (not after) will stimulate the execution of novel combination trials for the treatment of patients with metastatic melanoma. The implementation of the first phase 0 immuno-PET clinical trial in Europe (i.e. the use of the F16 antibody labelled with iodine-124 for the imaging of cancer patients) is another important milestone of the project. We expect that in the future targeted antibody drugs will be developed in combination with imaging modalities, which may guide patients' selection and may reveal a correlation between responding patients and patients with a good antibody uptake in the tumour. The companies which have participated in the project have grown substantially during the activities of the project. PHC has opened new laboratories in Otelfingen and has created new jobs (Philochem AG (and its mother company, Philogen) now have approximately 80 employees, between the Siena and Otelfingen site). KTECH has made substantial contributions to the development of tubulysin derivatives as an innovative and general class of cytotoxic agents for the coupling to antibodies. Moreover, Kemothech srl has recruited two employees thanks to the results achieved. Finally, Targetome is a spin-off company of the University of Liege, which has been founded as a result of innovative technologies in target discovery and validation developed in the frame of the project.
Website address:
http://www.adamant-fp7.eu/
E-mail address: adamant@marionegri.it
List of beneficiaries
B1 IRFMN, Coordinator, WPs 6 and 10 leader
Raffaella Giavazzi
Laboratory of the Biology and Treatment of Metastasis, Department of Oncology,
Istituto di Ricerche Farmacologiche 'Mario Negri'
Via Giuseppe La Masa, 19, Milano 20156, Italy
B.2 ETHZ, Scientific Deputy, WP 5 leader
Dario Neri
Institute of Pharmaceutical Sciences, Swiss Federal Institute of Technology
Wolfgang-Pauli-Str. 10, ETH Honggerberg, HCI G396, 8093 Zurich, Switzerland
B3 FSU Jena, WP 2 leader
Alexander Berndt
Institute of Pathology, Friedrich Schiller University
Ziegelmuhlenweg 1, 07740 Jena, Germany
B4 KTECH WP 4 leader
Matteo Zanda
Scientific Direction, KemoTech s.r.l.
Via Roma,72 Cagliari 01923, Italy
B5 PHC WP 8 leader
Eveline Trachsel
Therapeutic Antibody Research, Philochem AG c/o ETH Zurich, Institute of Pharmaceutical Sciences
Wolfgang-Pauli-Str. 10 HCI E250 CH-8093 Zurich
B6 SENDO WP9 leader
Silvia Marsoni
SENDO Foundation
Via Visconti di Modrone, 12; 20122 Milan, Italy
B7 UCL WP3 leader
Barbara R. Pedley
Department of Oncology, UCL Cancer Institute, Paul O'Gorman Building, University College London
72 Huntley St, London WC1E 6BT, United Kingdom
B8 ULG WP1 leader
Vincent Castronovo
Faculte de Medecine, Laboratoire de Recherche sur les Metastases, Universite de Liege
Avenue de l'Hopital 3, 4000 Liege, Belgium
B10 VUA WP7 leader
Guus van Dongen
Laboratory for Tumor Biology, Department of Otolaryngology, VU University Medical Center,
Vereniging voor Christelijk Hoger Onderwijs Wetenschappelijk Onderzoek en Patientenzorg De Boelelaan 1081 HV, Amsterdam, the Netherlands.