Formatting antibody fragments to mediate specific therapeutic functions.

Monoclonal antibodies are increasingly being used as therapeutic agents in a wide range of indications, including oncology, inflammation and infectious disease. In most cases the basis of the therapeutic function is the high degree of specificity and affinity the antibody-based drug has for its target antigen. However, the mechanism of action (MOA), the way the drug takes advantage of this specificity to mediate a therapeutic effect, varies considerably from drug to drug. Three basic potential categories of MOAs exist: antagonists, agonists and specific delivery mechanisms to target an active function to a particular cell type. The latter functions include selective cell killing, based on Fc-mediated events, recruitment of effector cells, and drug or radioisotope delivery. The majority of these mechanisms are not necessarily optimally mediated by an IgG structure; clearly, in the case of antibody-dependent cellular cytotoxicity or complement-mediated lysis, Fc is required. However, Fab fragments (the fragment comprising one antigen-binding arm of the Y-shaped IgG molecule) can be formatted to mediate most mechanisms and have the advantage that valency and half-life can be controlled to simplify the drug and address only the mechanism required. Moreover, Fab fragments can be produced in microbial expression systems which address manufacturing issues such as scale of supply and cost of goods.

Prolonged in vivo residence times of antibody fragments associated with albumin.

Antibody fragments can be expressed at a high level in microbial systems, but they may have limited therapeutic value because they are rapidly eliminated from the body. We demonstrate here that site-specific conjugation or binding of bacterially derived Fab' to the long-lived protein serum albumin allows full retention of the antibody's binding characteristics while imparting the albumin's longevity in vivo. In rats the area under the curve for Fab' conjugated to rat serum albumin was 17-fold greater than for the control of Fab' conjugated to cysteine. Again, a bispecific F(ab')(2) with specificity for rat serum albumin showed an area under the curve about 8-fold greater than did a F(ab')(2) without specificity to albumin. Genetic fusions of scFv to albumin were similarly long-lived and could be expressed in yeast to provide the basis of a cost-effective production system.

Therapeutic antibody fragments with prolonged in vivo half-lives.

Antibody fragments can be isolated rapidly using techniques such as phage display and can be expressed to high levels in microbial systems. However, to date such antibody fragments have been of limited use for many therapeutic applications because they are rapidly cleared from the body. We present a strategy for the site-specific chemical modification of antibody fragments with polyethylene glycol, which results in the production of antibody fragments with long in vivo half-lives and full retention of antigen-binding properties. This technology should allow more rapid and economical production of therapeutic antibodies for chronic disease therapy.

F(ab')2 molecules made from Escherichia coli produced Fab' with hinge sequences conferring increased serum survival in an animal model.

Fab's with hinges based on the human gamma1 sequence containing 1, 2, or 4 cysteines have been produced by high level Escherichia coli periplasmic secretion, and coupled in vitro by reduction/oxidation to form F(ab')2. We find that the F(ab')2 made with hinges containing 2 or 4 cysteines have a high level (approximately 70%) of multiple disulphide bonds. These F(ab')2 molecules have an increased pharmacokinetic stability as measured by area under the curve compared to those made by direct coupling through a single disulphide bond. One particular molecule containing 4 hinge cysteines has a greater pharmacokinetic stability than a F(ab')2 formed by chemical cross-linking. F(ab')2 made from the Fab' with 4 hinge cysteines is also relatively resistant to chemical reduction in vitro allowing partial reduction to expose reactive hinge thiols. These hinge sequences provide a simple method for producing robust F(ab')2 in vitro, obviating the need to use chemical cross-linkers, and provide a route to hinge specific chemical modification with thiol-reactive conjugates.

Radioimmunotherapy of colorectal carcinoma xenografts in nude mice with yttrium-90 A33 IgG and Tri-Fab (TFM).

The monoclonal antibody A33 recognises a tumour-associated antigen on human colorectal carcinoma, and has undergone preliminary evaluation in the clinic where selective localisation to hepatic metastases has been demonstrated [Welt et al. (1994) J. Clin. Oncol. 12, 1561-1571]. A33 and an A33 tri-fab fragment (TFM) were labelled with 90Y via a stable macrocyclic ligand for biodistribution and therapy studies in nude mice bearing SW1222 colon carcinoma xenografts. Biodistribution studies demonstrated tumour localisation for both A33 IgG and TFM with low bone, liver and kidney levels. Clearance of TFM from the blood was much faster than IgG and this led to lower tumour accumulation for TFM but superior tumour-blood ratios. The maximum per cent injected dose per g localised to tumour was 35.9% +/- 5.3% for A33 IgG and 12.9% +/- 4.6% for A33 TFM with tumour-blood ratios at 48 h after administration of 5.6 +/- 1.8 and 29.2 +/- 9.8 respectively. Autoradiography studies with 125I-labelled A33 IgG and TFM demonstrated a homogeneous distribution within tumour tissue which was not observed with other anti-colorectal tumour antibodies. TFM penetrated into the tumour tissue more rapidly than IgG. In therapy studies, a single dose of 90Y-A33 IgG (250 microCi per mouse) or 90Y-A33 TFM (300 microCi per mouse) led to complete regression of 2-week-old tumour xenografts with long-term tumour-free survivors. A transient drop in white blood cell count was observed with both IgG and TFM but was significantly more pronounced with IgG. The cell count fell to 8.4% of control for IgG, whereas with TFM cell counts fell to 51% of control before recovery. These results indicate that the more rapid blood clearance of 90Y-TFM confers reduced toxicity compared with 90Y-IgG although similar therapeutic effects are achieved. When the dose of 90Y-IgG was adjusted to give the same dose to tumour achieved with 300 microCi 90Y-TFM, a lesser therapeutic effect was observed. This may be owing to more rapid tumour penetration achieved with TFM. Both A33 IgG and TFM demonstrated potent anti-tumour effects against human tumour xenografts in this mouse model system. The stability of these 90Y-labelled conjugates and their effective tumour penetration are promising for the development of humanised reagents for clinical studies.

Preparation and preclinical evaluation of humanised A33 immunoconjugates for radioimmunotherapy.

A humanised IgG1/k version of A33 (hA33) has been constructed and expressed with yields up to 700 mg l-1 in mouse myeloma NS0 cells in suspension culture. The equilibrium dissociation constant of hA33 (KD = 1.3 nM) was shown to be equivalent to that of the murine antibody in a cell-binding assay. hA33 labelled with yttrium-90 using the macrocyclic chelator 12N4 (DOTA) was shown to localise very effectively to human colon tumour xenografts in nude mice, with tumour levels increasing as blood concentration fell up to 144 h. A Fab' variant of hA33 with a single hinge thiol group to facilitate chemical cross-linking has also been constructed and expressed with yields of 500 mg l-1. Trimaleimide cross-linkers have been used to produce a trivalent Fab fragment (hA33 TFM) that binds antigen on tumour cells with greater avidity than hA33 IgG. Cross-linkers incorporating 12N4 or 9N3 macrocycles have been used to produce hA33 TFM labelled stably and site specifically with yttrium-90 or indium-111 respectively. These molecules have been used to demonstrate that hA33 TFM is cleared more rapidly than hA33 IgG from the circulation of animals but does not lead to accumulation of these metallic radionuclides in the kidney. 90Y-labelled hA33 TFM therefore appears to be the optimal form of the antibody for radioimmunotherapy of colorectal carcinoma.

An in vivo model for screening peptidomimetic inhibitors of gelatinase A.

Gelatinase A, a matrix metalloproteinase, is frequently associated with human solid tumors, and its secretion and activation in the tumor milieu is considered important in the process of angiogenesis, invasion, and metastasis. Consequently, metalloproteinase inhibitors may be of value in the therapy of cancer as well as other disease states involving tissue remodeling and release of biologically active peptide/protein by proteolytic cleavage. Here we describe the development of a rapid screening assay for in vivo activity of peptidomimetic inhibitors of gelatinase A that involves assessment of inhibition of an enzyme-substrate reaction in a circumscribed body compartment, the mouse pleural cavity. As examples of the utility of this assay, in vivo activity of the aryl sulfonamide, sulfamyl urea, morpholino and carboxylic acid functionality at the P3' position of a series of hydroxamic acid inhibitors was examined after administration both intraperitoneally (ip) (to approximate systemic administration) and orally. For up to 2 h after ip administration, all inhibitors tested showed marked activity (> 90% inhibition) at 17 mumol/kg (approximately 10 mg/kg). This activity declined in a dose-responsive manner to insignificant levels at 0.67 mumol/kg (approximately 0.4 mg/kg). Aryl sulfonamides showed significant inhibition (> 50%) for up to 7 h after administration. A higher dosage (136 mumol/kg, approximately 80 mg/kg) was required to reveal oral activity, which was observed only with morpholino compounds (> 50% inhibition). Thus, the model described may be of value in the identification of orally active gelatinase A inhibitors.

Comparison of half-lives and cytotoxicity of N-chloroethyl-4-amino and N-mesyloxyethyl-benzoyl compounds, products of prodrugs in antibody-directed enzyme prodrug therapy (ADEPT).

The synthesis of two novel drugs, 4-[bis[2-(mesyloxy)ethyl]amino]benzoic acid (7) and 4-[(2-chloroethyl)[2-(mesyloxy)ethyl]amino]benzoic acid (8) is described here. They are the active drugs of two prodrugs (9 and 10) designed for use as anti-cancer agents. The prodrugs (9, 10 and 11) were made as a series of compounds which are bifunctional alkylating agents in which the activating effect of the ionized carboxyl function is masked through an amide bond to a glutamic acid residue. These relatively inactive prodrugs were designed to be activated to their corresponding alkylating agent active drugs (7, 8 and 12 respectively) at a tumour site by prior administration of a monoclonal antibody conjugated to a bacterial enzyme. This system is called antibody-directed enzyme prodrug therapy (ADEPT). The chemical half-lives of the prodrugs and their active drugs were measured in order to determine their relative reactivities. The half-lives ranged from 21 to 324 min for the active drugs and from 42 to 1158 min for the prodrugs. The viability of two different tumour cell lines was monitored with each active drug and prodrug. The IC50 values varied from 65 to 625 microM for the active drugs: no IC50 values could be obtained for the prodrugs, using a rapid incubation procedure. Each in vitro technique demonstrated the ability of the glutamic acid moiety to deactivate the drugs, forming effective prodrugs.

Antibody directed enzyme prodrug therapy (ADEPT): a three phase system.

Monoclonal anti-CEA antibody, A5B7, and its fragments conjugated to CPG2 localize to a peak concentration in the LS174T xenografts within 24 h after injection, but enzyme activity persists in plasma such that prodrug injection has to be delayed for 5-6 days in order to avoid toxicity. Injection of prodrug at this time did not result in growth delay of this tumour. A three-phase system has been developed in which residual plasma enzyme was inactivated and cleared by a galactosylated anti-CPG2 antibody, SB43gal, allowing prodrug administration within 24 h after the conjugate. Using this three-phase system, a marked growth delay of this tumour was achieved after a single course of treatment consisting of conjugate injection followed by SB43gal, 19 h later and three doses of the prodrug.

Antibody directed enzyme prodrug therapy (ADEPT): clinical report.

Following an extensive series of studies in nude mice with human xenografts a pilot scale clinical trial of antibody directed enzyme prodrug therapy has been initiated. The principle is to activate a relatively inert prodrug to an active cytotoxin by a tumour located enzyme. In the first stage of the study a prodrug para-N-(mono-2-chloroethyl monomesyl)-aminobenzoyl glutamic acid was administered to six patients with advanced colorectal cancer in a dose escalating protocol. Nausea and vomiting occurred as the only discernible toxic effect at the higher dose levels. Three of these patients and two other patients with advanced disease have proceeded to the second stage of the study in which an antibody-enzyme conjugate was given IV, followed after 36-48 h by a galactosylated anti-enzyme antibody. When plasma enzyme levels had become undetectable the patients received multiple doses of the prodrug. At the lower doses toxicity was minimal as were clinical responses. Two patients received higher doses which resulted in myelosuppression and temporary regression of advanced disease. No complications resulted from administration of the antibody-enzyme complex or enzyme inactivating antibody. The myelosuppression is attributable to the relatively long half-life of the active drug formed from the prodrug used in the present study.

Selective uptake of toxic nucleoside (125IUdR) by resistant cancer.

We report uptake of a thymidine analogue 125-Iodine-5-iodo-2'-deoxyuridine (125IUdR) by nude mice bearing human xenografts of choriocarcinoma or colonic cancer. When 125IUdR was given alone, uptake by intestinal tissues was 5-10 times greater than by the tumours as measured by tissue gamma counting. This ratio was reversed when hydroxyurea or cytosine arabinoside were used as inhibitors of ribonucleotide reductase and were given in combination with 5-fluorouracil or methotrexate to inhibit thymidine synthesis shortly before injecting 125IUdR. Counting the radioactivity in tissues removed 24 hours after 125IUdR gave tumour to highest normal tissue ratios of up to 15:1, but the corresponding nuclear grain counts, which is probably a more reliable indicator of selective uptake into DNA, were in excess of 100:1. The addition of unlabelled IUdR to the regimen only reduced the uptake of 125IUdR when given in relatively large amounts. For this approach to be exploited it is concluded that the tumour must be resistant at the cell level to the inhibitor of DNA synthesis either de novo or as a result of prior exposure to it. This inhibitor can then be used to block uptake of the potentially toxic nucleoside analogue by normal renewal tissues while it is taken up by the resistant cancer cells. By inhibiting synthesis of the corresponding normal nucleosides with inhibitors to which the cancer cells are not resistant, incorporation of the toxic analogues into tumour DNA was enhanced. Although 125IUdR is a convenient agent for exploring this approach and is highly cytotoxic when incorporated in DNA, the clinical potential of reverse role chemotherapy probably lies with the development of toxic non-radioactive nucleoside analogues.

Ablation of human choriocarcinoma xenografts in nude mice by antibody-directed enzyme prodrug therapy (ADEPT) with three novel compounds.

Three novel prodrugs have been designed for use as anticancer agents. Each is a bifunctional alkylating agent which has been protected to form a relatively inactive prodrug. They are designed to be activated to their corresponding alkylating agents at a tumour site by prior administration of an antitumour antibody conjugated to the bacterial enzyme carboxypeptidase G2 (CPG2) in a two-phase system called antibody-directed enzyme prodrug therapy (ADEPT). The Km and Vmax values for three different antibody-CPG2 conjugates were determined in relation to each prodrug. The Km values ranged from 4.5-12 mumol/l and the Vmax from 0.5-1.6 mumol/U/min. Athymic Nu/Nu mice with palpable transplanted human choriocarcinoma xenografts, which are resistant to conventional chemotherapy, were treated with anti-human chorionic gonadotropin antibodies conjugated to CPG2. This was followed by each of the three novel prodrugs. Significant increase in survival was obtained in three of the regimens tested using only one course of treatment. This demonstrates the potential of a tumour-localised bacterial enzyme to activate protected alkylating agents in order to eradicate an established human xenograft.

Disposition of the prodrug 4-(bis (2-chloroethyl) amino) benzoyl-L-glutamic acid and its active parent drug in mice.

A novel therapy for improving selectivity in cancer chemotherapy aims to modify distribution of a cytotoxic drug by generating it selectively at tumour sites. In this approach an antibody-enzyme conjugate is allowed to localise at the tumour sites before injecting a prodrug which is converted to an active drug specifically by the targeted enzyme in the conjugate. We present here pharmacokinetic studies on the prodrug 4-(bis (2-chloroethyl) amino) benzoyl-L-glutamic acid and its activated derivative, benzoic acid mustard. The glutamic acid is cleaved from the prodrug to form the active drug by carboxypeptidase G2 (CPG2), an enzyme from Pseudomonas sp., which is not found in mammalian cells. The prodrug and its parent active drug were rapidly distributed in plasma and tissues after administration of prodrug or active drug (41 mumol kg-1 intraperitoneally) to mice bearing human choriocarcinoma xenografts. Prodrug and active drug both followed a two-compartment kinetic model. Prodrug was eliminated more rapidly (t1/2 alpha = 0.12 h, t1/2 beta = 0.70 h) than active drug (t1/2 alpha = 0.37 h, t1/2 beta = 1.61 h). Conversion of the prodrug to the activated parent drug was detected within 5 min of administration to mice which had previously received a F(ab')2-anti-human chorionic gonadotrophin antibody (W14A) conjugated to the enzyme, CPG2 (1,000 U kg-1). Tumour was the only tissue that activated all the prodrug reaching the site. It contained the highest concentration of targeted enzyme conjugate capable of catalysing the reaction of prodrug to drug. Plasma and other tissues were also capable of activating the prodrug but active drug production was limited by the amount of enzyme present. The active drug measured in plasma and tissues other than tumour was attributable to residual antibody-enzyme conjugate at non-tumour sites. Low levels of conjugate in tissues and plasma militate against the advantage of tumour localised enzyme therefore necessitating removal of non-localised enzyme.

Novel prodrugs which are activated to cytotoxic alkylating agents by carboxypeptidase G2.

The synthesis of three novel prodrugs, 4-[bis[2-(mesyloxy)ethyl]amino]benzoyl-L-glutamic acid (7), 4-[(2-chloroethyl)[2-(mesyloxy)ethyl]amino]benzoyl-L-glutamic acid (8), and 4-[bis(2-chloroethyl)amino]benzoyl-L-glutamic acid (9), for use as anticancer agents, is described here. Each is a bifunctional alkylating agent in which the activating effect of the ionized carboxyl function is masked through an amide bond to the glutamic acid residue. These relatively inactive prodrugs are designed to be activated to their corresponding nitrogen alkylating agents (10, 11, and 12, respectively) at a tumor site by prior administration of a monoclonal antibody conjugated to the bacterial enzyme carboxypeptidase G2 (CPG2). The viability of two different tumor cell lines was monitored with each prodrug in the presence of CPG2. All three compounds showed substantial prodrug activity--with conversion to the corresponding active drug leading to greatly increased cytotoxicity.

A cytotoxic agent can be generated selectively at cancer sites.

Attempts to improve the selectivity of anti-cancer agents by conjugating them to antibodies directed at tumour associated antigens have demonstrated tumour localisation but only limited therapeutic success. We report here the advantage of a 2-stage approach in which the first component combines the selective delivery of antibody with a capability to generate a cytotoxic agent from a second subsequently administered component. A bacterial enzyme, carboxypeptidase G2 (CPG2) was conjugated with F(ab')2 fragment of a monoclonal antibody directed at beta subunit of human chorionic gonadotrophin (beta-hCG) and injected into nude mice bearing hCG producing CC3 xenografts of human choriocarcinoma. Time was allowed for the conjugate to localise at tumour sites and clear from blood before injecting para-N-bis (2-chloroethyl) aminobenzoylglutamic acid. Cleavage of the glutamic acid moiety from this molecule by CPG2 released a benzoic acid mustard. Growth of the tumour which is resistant to conventional chemotherapy was markedly depressed by a single course of treatment. This demonstrates for the first time the potential of an antibody directed enzyme to activate an alkylating agent and to eradicate an established human cancer xenograft.

Activity and distribution studies of etoposide and mitozolomide in vivo and in vitro against human choriocarcinoma cell lines.

The in vivo antitumor activity of etoposide and mitozolomide was assessed in nude mice bearing a xenograft (CC3) of human gestational choriocarcinoma. Both agents demonstrated, at best, marginal activity observed as a delay in tumour growth. This lack of sensitivity suggests that the CC3 xenograft is not a good model for selection of agents for clinical evaluation in gestational choriocarcinoma. Plasma and tissue concentrations of etoposide and mitozolomide were measured in nude mice. Drug concentrations found in tumour tissue were 60% and 30% of plasma levels for mitozolomide and etoposide respectively. Etoposide and mitozolomide activity was also evaluated in vitro with another choriocarcinoma cell line (JAR). Maximum cell-kill was achieved after exposure to etoposide 0.05-1 microgram/ml for 3-24 h. In vitro response to etoposide demonstrates the importance of exposure time in determining cytotoxicity. In contrast, mitozolomide at concentrations from 1-100 micrograms/ml did not have a marked effect against JAR after exposure for 3-24 h.

Plasma and tissue disposition of mitozolomide in mice.

When mitozolomide was administered i.p. to mice, drug disposition appeared to fit a simple, one-compartment kinetic model with an elimination half-life of less than 1 h. The disposition of mitozolomide in mice bearing the ROS osteosarcoma, also followed a first-order process but in this case the elimination of the drug was significantly faster from plasma, liver, lung and kidney tissue compared to the elimination half-life of the drug from the same tissues of mice without tumour (P less than 0.05). Mitozolomide was rapidly and extensively distributed into tissues, including the tumour. Mitozolomide was not concentrated in any particular tissue although the brain contained the lowest drug concentration compared to any tissue studied. After 4 h following administration, mitozolomide could not be measured in plasma or tissues. AUC values calculated from mitozolomide concentration versus time profiles in plasma, liver and kidney homogenates were 27-29% lower in mice pretreated with phenobarbitone compared to those values obtained from mice administered saline only, (P less than 0.02). Since phenobarbitone is known to induce liver microsomal enzymes, it is possible that hepatic metabolism is involved in the degredation of mitozolomide.

Pharmacokinetics and toxicity of the epipodophyllotoxin derivative etoposide (VP 16-213) in patients with gestational choriocarcinoma and malignant teratoma.

Serum levels of etoposide obtained 5 min after administration of 100 mg/m2 were between 11 and 30 micrograms/ml. By 24 h after drug administration, serum levels had fallen to between 0.19 and 1.11 micrograms/ml. Interpatient variation of etoposide serum concentrations obtained 5 min after drug administration was low, whereas interpatient variation 24 h later was noticeably higher. A significant correlation was observed (r = -0.698) between the WBC nadir and the mean etoposide serum concentrations, measured 24 h after drug administration, in patients receiving etoposide in combination with cyclophosphamide and actinomycin D. However, a relationship was not observed in those patients receiving etoposide alone. There was no observed difference in the efficacy or toxicity of 500 mg/m2 etoposide when the dose was administered either as 100 mg/m2 on each of 5 consecutive days or as 250 mg/m2 on days 1 and 3. There was no significant difference between AUC values calculated from etoposide concentration versus time profiles in patients receiving the drug on days 1 and 3 and those values obtained with the 5-day schedule. Patients resistant to a conventional dose of etoposide were given a higher dose of 1 g/m2/24 h, but this schedule did not cause an increase in efficacy despite an increase in serum levels of the drug. CSF levels in two of these patients receiving high-dose etoposide were 1.28% and 2.09% of the serum concentrations.