Richard H. J. Begent

Clinical evidence of efficient tumor targeting based on single-chain Fv antibody selected from a combinatorial library

We present a system for cancer targeting based on single–chain Fv (scFv) antibodies selected from combinatorial libraries, produced in bacteria and purified by using an engineered tag. Combinatorial libraries of scFv genes contain great diversity, and scFv antibodies with characteristics optimized for a particular task can be selected from them using filamentous bacteriophage. We illustrate the benefits of this system by imaging patients with carcinoembryonic antigen (CEA)–producing cancers using an iodine–123 labeled scFv anti–CEA selected for high affinity. All known tumor deposits were located, and advantages over current imaging technology are illustrated. ScFvs are produced in a cloned form and can be readily engineered to have localizing and therapeutic functions that will be applicable in cancer and other diseases.

PHAGE LIBRARIES FOR GENERATION OF CLINICALLY USEFUL ANTIBODIES

Insertion of antibody genes into filamentous bacteriophage makes it possible to generate and screen libraries of 10(7) or more antibodies. Each phage expresses an antibody on its surface and contains the corresponding antibody gene. Genes that encode antibodies with desired characteristics are readily selected and their antibodies expressed as soluble proteins in Escherichia coli. We used this system to produce an antibody to carcinoembryonic antigen with higher affinity and better tumour specificity than antibodies currently in use.

Purification of bacterially expressed single chain Fv antibodies for clinical applications using metal chelate chromatography

A new procedure is described for the purification of an anti-carcinoembryonic antigen (CEA) single chain Fv (scFv), referred to as MFE-23, from bacterial supernatant. A simple insertion of a hexa-histidine tail fused at the C-terminus (MFE-23 His) provides an affinity tag which selectively binds to transition metal ions immobilised on an iminodiacetic acid (IDA) derivitised solid phase matrix. This method proved to be superior to standard CEA antigen affinity chromatography in the following ways. (1) A higher yield was produced (10 mg/l as opposed to 2.2 mg/l of bacterial supernatant). The latter figure was largely affected by the limited availability (size of the column) of immobilised CEA antigen. (2) Scale-up was relatively simple and less costly. (3) The risk of tumour derived antigen leaching from the column is eliminated. Results showed that immobilised Cu2+ ions were more effective than Ni2+ and Zn2+ ions in retaining the His tagged product giving a 90% pure product on elution. Clinical grade material was generated using size exclusion chromatography to remove aggregated material, and Detoxi gel to remove bacterial endotoxins. Validation assays to measure DNA, copper and endotoxins were performed to assess the levels of contaminants. MFE-23 His retained 84% antigen binding after 6 months storage at 4 degrees C and > 75% after radiolabelling. Further experiments confirmed that the His tail did not affect biodistribution and tumour localisation in nude mice bearing human colorectal tumour xenografts.

Antibody-directed enzyme prodrug therapy: pharmacokinetics and plasma levels of prodrug and drug in a phase I clinical trial

Abstract Antibody-directed enzyme prodrug therapy (ADEPT) was administered to ten patients in a phase I clinical trial. The aim was to measure plasma levels of the prodrug 4-[(2-chloroethyl)(2-mesyloxyethyl) amino] benzoyl-l-glutamic acid (CMDA) and the bifunctional alkylating drug (CJS11) released from it by the action of tumour-localised carboxypeptidase G2 (CPG2) enzyme. New techniques were developed to extract the prodrug and drug from plasma by solid-phase adsorbtion and elution and to measure CPG2 activity in plasma and tissue. All extracts were analysed by high-performance liquid chromatography (HPLC) and liquid chromatography-mass spectrometry (LC-MS). CPG2 activity was found in metastatic tumour biopsies but not in normal tissue, indicating that localisation had been successful. The clearing agent SB43-gal, given at 46.5 mg/m2, achieved the aim of clearing non-tumour-localised enzyme in the circulation, indicating that conversion of prodrug to drug could take place only at the site of localised conjugate. Plasma prodrug did not always remain above its required threshold of 3 μM for the “therapeutic window” of 120 min after dosing, but the presence of residual prodrug after the first administration of each day indicated that this could be achieved during the remaining four doses over the following 8 h. Despite considerable inter-patient prodrug plasma concentration variability, the elimination half-life of the prodrug was remarkably reproducible at 18 ± 8 min. Rapid appearance of the drug in plasma indicated that successful conversion from the prodrug had taken place, but also undesirable leakback from the site of localisation into the bloodstream. However, drug plasma levels fell rapidly by at least 50% at between 10 and 60 min with a half-life of 36 ± 14 min. Analysis of the plasma extracts by LC/MS indicated that this technique might be used to confirm qualitatively the presence of prodrug, drug and their metabolites.

A phase I study of single administration of antibody-directed enzyme prodrug therapy with the recombinant anti-carcinoembryonic antigen antibody-enzyme fusion protein MFECP1 and a bis-iodo phenol mustard prodrug

Purpose: Antibody-directed enzyme prodrug therapy is a two-stage treatment whereby a tumor-targeted antibody-enzyme complex localizes in tumor for selective conversion of prodrug. The purpose of this study was to establish optimal variables for single administration of MFECP1, a recombinant antibody-enzyme fusion protein of an anti–carcinoembryonic antigen single-chain Fv antibody and the bacterial enzyme carboxypeptidase G2 followed by a bis-iodo phenol mustard prodrug. MFECP1 is manufactured in mannosylated form to facilitate normal tissue elimination. Experimental Design: Pharmacokinetic, biodistribution, and tumor localization studies were used to test the hypothesis that MFECP1 localizes in tumor and clears from normal tissue via the liver. Firstly, safety of MFECP1 and a blood concentration of MFECP1 that would avoid systemic prodrug activation were tested. Secondly, dose escalation of prodrug was done. Thirdly, the dose of MFECP1 and timing of prodrug administration were optimized. Results: MFECP1 was safe and well tolerated, cleared rapidly via the liver, and was less immunogenic than previously used products. Eighty-fold dose escalation from the starting dose of prodrug was carried out before dose-limiting toxicity occurred. Confirmation of the presence of enzyme in tumor and DNA interstrand cross-links indicating prodrug activation were obtained for the optimal dose and time point. A total of 28 of 31 patients was evaluable for response, the best response being a 10% reduction of tumor diameter, and 11 of 28 patients had stable disease. Conclusions: Optimal conditions for effective therapy were established. A study testing repeat treatment is currently being undertaken.

CD44 is associated with proliferation in normal and neoplastic human colorectal epithelial cells

Flash-frozen biopsies obtained from surgical specimens of three adenomatous polyps and 22 colorectal adenocarcinomas (19 primary and three metastatic) were tested by immunohistochemistry for CD44 expression using F10-44-2 monoclonal antibody. CD44 positivity was correlated with proliferative status defined by Ki-67 monoclonal antibody reactivity. In normal colonic mucosa, CD44 was expressed in the proliferative zone of crypts. In tumours, CD44 expression was associated with proliferative areas irrespective of tumour stage or differentiation. Non-proliferating areas of the carcinomatous epithelium did not express CD44 although non-proliferating stromal lymphoid tissue did. There was no apparent association with tumour progression. F10-44-2-defined CD44 is consistently expressed during proliferation by normal colorectal epithelial cells and by both benign and malignant colorectal tumour cells.

Catalytic activity of an in vivo tumor targeted anti-CEA scFv::carboxypeptidase G2 fusion protein.

Antibody‐directed enzyme prodrug therapy (ADEPT) targets an enzyme selectively to a tumor where it converts a relatively non‐toxic prodrug to a potent cytotoxic drug. Previous clinical work using antibody‐enzyme chemical conjugates has been limited by the moderate efficiency of tumor targeting of these molecules. To address this a recombinant fusion protein composed of MFE‐23, an anti‐carcinoembryonic antigen (CEA) single chain Fv (scFv) antibody, fused to the amino‐terminus of the enzyme carboxypeptidase G2 (CPG2) has been constructed to achieve ADEPT in CEA‐producing tumors. MFE‐23::CPG2 fusion protein was overexpressed in Escherichia coli and purified using CEA affinity chromatography. Efficacy of MFE‐23::CPG2 delivery to tumors in vivo was assessed by measuring catalytic activity after intravenous injection of purified MFE‐23::CPG2 into nude mice bearing CEA‐positive LS174T human colon adenocarcinoma xenografts. Recombinant MFE‐23::CPG2 cleared rapidly from circulation and catalytic activity in extracted tissues showed tumor to plasma ratios of 1.5:1 (6 hr), 10:1 (24 hr), 19:1 (48 hr) and 12:1 (72 hr). 125I‐MFE‐23::CPG2 was retained in kidney, liver and spleen but MFE‐23::CPG2 catalytic activity was not, resulting in excellent tumor to normal tissue enzyme ratios 48 hr after injection. These were 371:1 (tumor to liver), 450:1 (tumor to lung), 562:1 (tumor to kidney), 1,477:1 (tumor to colon) and 1,618:1 (tumor to spleen). Favorable tumor : normal tissue ratios occurred at early time points when there was still 21% (24 hr) and 9.5% (48 hr) of the injected activity present per gram of tumor tissue. The high tumor concentrations and selective tumor retention of active enzyme delivered by MFE‐23::CPG2 establish that this recombinant fusion protein has potential to give improved clinical efficiency for ADEPT. Int. J. Cancer 85:571–577, 2000.

Phase I study of sequence-selective minor groove DNA binding agent SJG-136 in patients with advanced solid tumors.

Purpose: This phase I dose-escalation study was undertaken to establish the maximum tolerated dose of the sequence-selective minor groove DNA binding agent SJG-136 in patients with advanced solid tumors. The study also investigated antitumor activity and provided pharmacokinetic and pharmacodynamic data. Experimental Design: Sixteen patients were assigned sequentially to escalating doses of SJG-136 (15-240 μg/m2) given as a 10-minute i.v. infusion every 21 days. The dose was subsequently reduced in incremental steps to 45 μg/m2 due to unexpected toxicity. Results: The maximum tolerated dose of SJG-136 was 45 μg/m2. The main drug-related adverse event was vascular leak syndrome (VLS) characterized by hypoalbuminemia, pleural effusions, ascites, and peripheral edema. Other unexpected adverse events included elevated liver function tests and fatigue. The VLS and liver toxicity had delayed onset and increased in severity with subsequent cycles. Disease stabilization was achieved for >6 weeks in 10 patients; in 2 patients this was maintained for >12 weeks. There was no evidence of DNA interstrand cross-linking in human blood lymphocytes with the use of the comet assay. Evidence of DNA interaction in lymphocytes and tumor cells was shown through a sensitive γ-H2AX assay. SJG-136 had linear pharmacokinetics across the dose range tested. Conclusions: SJG-136 was associated with dose-limiting VLS and hepatotoxicity when administered by short injection every 21 days. DNA damage was noted, at all dose levels studied, in circulating lymphocytes. The etiology of the observed toxicities is unclear and is the subject of further preclinical research. Alternative clinical dosing strategies are being evaluated.

Nanoparticles functionalised with recombinant single chain Fv antibody fragments (scFv) for the magnetic resonance imaging of cancer cells

Superparamagnetic iron oxide nanoparticles (SPIONs) can substantially improve the sensitivity of magnetic resonance imaging (MRI). We propose that SPIONs could be used to target and image cancer cells if functionalized with recombinant single chain Fv antibody fragments (scFv). We tested our hypothesis by generating antibody-functionalized (abf) SPIONs using a scFv specific for carcinoembryonic antigen (CEA), an oncofoetal cell surface protein. SPIONs of different hydrodynamic diameter and surface chemistry were investigated and targeting was confirmed by ELISA, cellular iron uptake, confocal laser scanning microscopy (CLSM) and MRI. Results demonstrated that abf-SPIONs bound specifically to CEA-expressing human tumour cells, generating selective image contrast on MRI. In addition, we observed that the cellular interaction of the abf-SPIONs was influenced by hydrodynamic size and surface coating. The results indicate that abf-SPIONs have potential for cancer-specific MRI.

Accurate quantification of 131I distribution by gamma camera imaging.

The development of targeted therapy requires that the concentration of the therapeutic agent can be estimated in target and normal tissues. Single photon emission tomography (SPET), with and without scatter correction, and planar imaging using131I have been compared to develop a method for investigation of targeted therapy. Compton scatter was investigated using line spread functions in air and water, these data were used to set a second peak, adjacent to the photopeak, for scatter correction. The system was calibrated with an eliptical phantom containing sources in background activity of various intensities. Scatter corrected reconstructions gave accurate estimates of activity in the sources regardless of background activity. For planar scanning and SPET without scatter correction there was an overestimate of activity in the source of 290% and 40% respectively. The validity of this method was confirmed in patients by comparing activity in the cardiac ventricles measured by SPET with scatter correction with that in a simultaneous blood sample. A coefficient of correlation of 0.955 was achieved with 25 data points. SPET with scatter correction was compared with planar imaging in measuring activity in the liver and spleen of patients receiving 75 mCi131I-antibody to CEA intravenously. Planar imaging gave significantly higher values than SPET for the spleen (t=5.4,P<0.001 by the pairedt-test) but no significant difference for the liver. SPET with scatter correction forms a basis for an improved technique of quantifying the targeting efficiency.