Targeted drug conjugates: principles and progress

Antibody targeted drugs as cancer therapeutics

Treatment of cancer is a double-edged sword: it should be as aggressive as possible to completely destroy the tumour, but it is precisely this aggressiveness which often causes severe side effects - a reason why some promising therapeutics can not be applied systemically. In addition, therapeutics such as cytokines that physiologically function in a para- or autocrine fashion require a locally enhanced level to exert their effect appropriately. An elegant way to accumulate therapeutic agents at the tumour site is their conjugation/fusion to tumour-specific antibodies. Here, we discuss recent preclinical and clinical data for antibody-drug conjugates and fusion proteins with a special focus on drug components that exert their antitumour effects through normal biological processes.

A comparison of mineral affinity of bisphosphonate–protein conjugates constructed with disulfide and thioether linkages

Chemical conjugation of bisphosphonates (BPs) to therapeutic proteins is an effective means to impart mineral affinity to proteins. Such conjugates can be implanted with mineral-based matrices to control the local delivery kinetics of the proteins. BPs linked to proteins with reversible (i.e., cleavable) linkages are desirable over conjugates with stable linkages to release the protein in free form. This study conducted a direct comparison of mineral affinity of BP-protein conjugates linked together with cleavable disulfide and non-cleavable thioether linkages. Bovine serum albumin (BSA) was used as a model protein and the desired conjugates were created with N-succinimidyl-3-(2-pyridyldithio)propionate (disulfide) and succinimidyl-4-(N-maleimido-methyl)cyclohexane-1-carboxylate (thioether) linkers. The disulfide-linked conjugates were cleaved in the presence of a major thiol constituent of serum, cysteine. The imparted mineral affinity, as assessed by hydroxyapatite binding in vitro, was lost upon the cleavage of the disulfide-linked aminoBP. The presence of the serum did not accelerate the cleavage of disulfide-linked conjugates. The aminoBP-BSA conjugates formed with disulfide and thioether linkages were subcutaneously implanted in rats with two different mineral-based matrices to assess protein loss from the matrices. All conjugates exhibited a higher retention in mineral matrices as compared to unmodified BSA. However, no significant differences in in situ pharmacokinetics of the disulfide- and thioether-linked conjugates were observed. We conclude that disulfide-linked BP conjugates were readily cleavable by the amino acid cysteine in vitro, but in vivo cleavage of the disulfide-linked conjugates was not evident when the proteins were implanted adsorbed to mineral-based matrices. BP-protein conjugates with faster-cleaving tethers might be required to significantly influence the release of the BP conjugates from the mineral matrices.

Targeted delivery of methotrexate to epidermal growth factor receptor–positive brain tumors by means of cetuximab (IMC-C225) dendrimer bioconjugates

We have constructed a drug delivery vehicle that targets the epidermal growth factor receptor (EGFR) and its mutant isoform EGFRvIII. The monoclonal antibody, cetuximab, previously known as C225, which binds to both EGFR and EGFRvIII, was covalently linked via its Fc region to a fifth-generation (G5) polyamidoamine dendrimer containing the cytotoxic drug methotrexate. As measured by mass spectrometry and UV/vis spectroscopy, the resulting bioconjugate, designated C225-G5-MTX, contained 12.6 molecules of methotrexate per unit of dendrimer. Specific binding and cytotoxicity of the bioconjugate was evaluated against the EGFR-expressing rat glioma cell line F98(EGFR). Using a competitive binding assay, it was shown that the bioconjugate retained its affinity for F98(EGFR) cells, with a 0.8 log unit reduction in its EC(50). Only cetuximab completely inhibited binding of the bioconjugate, which was unaffected by methotrexate or dendrimer. Cetuximab alone was not cytotoxic to F98(EGFR) cells at the concentration tested, whereas the IC(50) of the bioconjugate was 220 nmol/L, which was a 2.7 log unit decrease in toxicity over that of free methotrexate. The biodistribution of C225-G5-MTX in rats bearing i.c. implants of either F98(EGFR) or F98(WT) gliomas was determined 24 hours following convection enhanced delivery of (125)I-labeled bioconjugate. At this time, 62.9 +/- 14.7% ID/g tumor was localized in rats bearing F98(EGFR) gliomas versus 11.3 +/- 3.6% ID/g tumor in animals bearing F98(WT) gliomas, thereby showing specific molecular targeting of the tumor. The corresponding radioactivity of normal brain from the F98(EGFR) tumor-bearing right and non-tumor-bearing left cerebral hemisphere were 5.8 +/- 3.4% and 0.8 +/- 0.6% ID/g, respectively. Based on these results, therapy studies were initiated in F98(EGFR) glioma-bearing rats. Animals that received C225-G5-MTX, cetuximab, or free methotrexate had median survival times of 15, 17, and 19.5 days, respectively, which were not statistically different from each other or untreated control animals. Our results, which are both positive and negative, show that specific molecular targeting is but one of several requirements that must be fulfilled if an antibody-drug bioconjugate will be therapeutically useful.

Folate-mediated targeting of polymeric conjugates of gemcitabine

The synthesis of two new macromolecular prodrugs for active tumor targeting was set up. Gemcitabine (2'-deoxy-2',2'-difluorocytidine) was conjugated to alpha,beta-poly(N-2-hydroxyethyl)-DL-aspartamide (PHEA) through succinyl or diglycolyl hydrolysable spacers. The targeting agent folic acid was attached to the macromolecular backbone through the aminocaproic spacer. The two conjugates [PHEA-(5'-succinylgemcitabine)-1'-carboxypentyl-folamide and PHEA-(5'-diglycolyl-gemcitabine)-1'-carboxypentyl-folamide], were purified and extensively characterised by spectroscopic (UV, IR and NMR) and chromatographic analyses to determine the correct chemical structure, the purity degree and the reaction yield. In vitro studies demonstrated that the drug release depends on the spacer arm (diglycolyil or succinyl) and incubation pH. After 30 h incubation at pH 7.4, mimicking the plasma and extracellular compartments, the gemcitabine release from the succinyl and diglycolyl derivatives was 28 and 31%, respectively. After 30 h incubation at pH 5.5, mimicking the lisosomial compartment, the drug released from both bioconjugates was lower than 13%. In plasma, the polymer conjugation increased the drug stability and provided for a sustained drug release. In vitro citotoxicity studies performed using human nasopharyngeal epidermal carcinoma KB cells demonstrated that PHEA-(5'-succinylgemcitabine)-1'-carboxypentyl-folamide displays an higher dose dependent cytotoxic effect with respect to PHEA-(5'-diglycolyl-gemcitabine)-1'-carboxypentyl-folamide.

Blood Pharmacokinetics of Various Monoclonal Antibodies Labeled with a New Trifunctional Chelating Reagent for Simultaneous Conjugation with 1,4,7,10-Tetraazacyclododecane-N,N′,N″,N‴-Tetraacetic Acid and Biotin before Radiolabeling

PURPOSE

Knowledge of the blood pharmacokinetics of monoclonal antibodies is crucial in deciding the optimal time for starting the administration of a "clearing agent" or using a "clearing device." The primary purpose was to investigate whether the pharmacokinetics of various antibodies labeled with the same chelator and (111)In differed significantly after i.v. injection in immunocompetent rats. A new trifunctional chelator called "1033" containing a biotin and a radiometal chelation moiety is introduced, making it possible to use only one conjugation procedure for the antibody.

EXPERIMENTAL DESIGN

Sixty-five non-tumor-bearing rats were included and divided into four groups (I-IV). The blood pharmacokinetics was investigated for rituximab, BR96, and trastuzumab labeled with 1033 and (111)In (I-III). The whole-body activity and activity uptake in muscle, liver, and kidney, which might explain differences in the early pharmacokinetics in blood, were also measured. hMN14 labeled with another chelator [1,4,7,10-tetraazacyclododecane-N,N',N'',N'''-tetraacetic acid (DOTA)], but with the same radionuclide ((111)In-biotin-DOTA-hMN14), was studied (IV). The blood pharmacokinetics from another 15 tumor-bearing rats was compared with those of non-tumor-bearing rats (III) by injection of (111)In-1033-BR96.

RESULTS

No statistical difference was detected between the groups regarding the blood pharmacokinetics of rituximab, BR96, or trastuzumab. The pharmacokinetics and biodistribution of (111)In-biotin-DOTA-hMN14 exhibited a clear difference compared with others. There were no significant differences in the blood pharmacokinetics of (111)In-1033-BR96 between tumor-bearing rats and non-tumor-bearing rats.

CONCLUSIONS

Different antibodies labeled with the trifunctional chelator 1033 and (111)In did not exhibit different blood pharmacokinetics, which means that the pharmacokinetics could be predicted irrespective of the IgG1 antibody chosen. A small tumor burden did not change the pharmacokinetics of the radioimmunoconjugates.

NANOMEDICINES: DELIVERING DRUGS USING BOTTOM UP NANOTECHNOLOGY

The use of nanosized materials changes the way in which drugs are handled by the body and offers opportunities to improve drug delivery. The physiological mechanisms controlling the distribution of nanosized materials (enhanced permeability and retention effect, cellular uptake pathways and opsonisation/elimination of nanoparticles) are described. Two different nanosized drug delivery systems are considered; drug delivery and DNA delivery. The deficiencies of currently available biodegradable polymers for preparation of drug containing nanoparticles are mainly the amount of drug that can be incorporated and the rapid rate of drug release. The development of new biodegradable polymers which can interact with the drug and so significantly increase drug loading and decrease the rate of drug release are outlined. DNA delivery necessitates overcoming a variety of biological barriers. We are developing polyelectrolyte complexes of DNA with cationic polyamidoamines (PAA) as a delivery system. Complexing PAA with DNA results in good transfection of cells in vitro. However, in vivo, a more complex arrangement of PAA, Polyethylene glycol-PAA copolymers, DNA and the use of ligands will be required. Despite these efforts, further developments will be needed in nanotechnology for both drug and DNA nanoparticle delivery systems to achieve our clinical objectives.

Methoxypoly(ethylene glycol)-conjugated carboxypeptidase A for solid tumor targeting

In vivo efficacy of novel anticancer agents has been hindered by the inability to deliver effective concentrations of drugs to tumors. The use of macromolecules such as antibodies and polymers for enzyme delivery to tumors has revealed that catalyzing the conversion of a nontoxic prodrug into its cytotoxic form can generate an effective level of cytotoxic agents at tumor sites. This study primarily focuses on the synthesis and characterization of methoxypoly(ethylene glycol)-modified carboxypeptidase A (CPA) for solid tumor targeting. The molecular weight of CPA has been successfully altered from 35 to 40-50 kDa via attachment of a defined number of mPEG moieties. Relatively pure mPEG-CPA conjugates containing one, two, and three mPEG chains were obtained at preparative scale quantities through controlled PEGylation followed by fractionation that involved size-exclusion chromatography. An enhancement in kinetic properties including k(cat) and k(cat)/K(m) towards hippuryl-L-phenylalanine (hipp-L-phe) was observed in mPEG-CPA conjugates. An increase in the V(m) appeared to be responsible for this enhancement. The attachment of mPEG to CPA substantially improved the stability of the enzyme with respect to the specific peptidase activity toward the model substrate. This finding is particularly important in the development of a novel CPA/methotrexate-alpha-peptide system in solid tumor chemotherapy.

Strategies for targeting protein therapeutics to selected tissues and cells

The advent of recombinant biotechnology and the recent sequencing of the human genome now allow for identification of scores of potential protein therapeutics along with the capacity to produce them in quantities and purities required for clinical application. Thus, clinical development of potential protein therapeutics has become as commonplace as development efforts of classical small molecule therapeutics. Whereas small molecule therapeutic lead candidates are identified through screens of large sets of possibilities, therapeutic protein candidates are defined by genetic information as a single composition (or a limited set of isoforms). Small molecule leads are identified through the combined assessment of desired selectivity, biodistribution and pharmacokinetic properties. In essence, these selection parameters emulate the actions of protein therapeutics that function as systemic hormones through their ability to target selective cells and tissues of the body via selective receptor interaction with minimal actions elsewhere. However, many, if not most, potential protein therapeutics do not normally circulate through the body to reach their target cell or tissue; rather, they are frequently synthesised at local sites, act at that site and are degraded without reaching appreciable systemic levels. Dose-limiting adverse events are associated with systemic administration of many of these proteins, restricting their clinical potential. This review examines current strategies to reduce these dose-limiting events by possibly focusing the delivery of potential protein therapeutics to discrete tissues and cells.

Intracellular Signal-Responsive Gene Carrier for Cell-Specific Gene Expression

We designed a peptide-polymer conjugate (CPCCtat) as a novel gene carrier that could control gene expression responding to the intracellular caspase-3 signal. This carrier consists of an uncharged main polymer chain and a cationic peptide side chain, which includes the substrate sequence of caspase-3 and the protein transduction domain sequence of HIV-1 Tat. In the present study, CPCCtat formed a tight complex with DNA through an electrostatic interaction, and in this state the gene expression was totally suppressed. In contrast, the complex disintegrated in the presence of caspase-3 due to cleavage of the cationic portion from CPCCtat. This event led to an activation of gene expression. Our results also indicate that the complex can be delivered into living cells due to the cell-permeable peptide side chain of CPCCtat. This intracellular signal-responsive system with CPCCtat will be useful for the cell-specific gene expression system.

Synthesis, physico-chemical and biological characterization of a paclitaxel macromolecular prodrug

Paclitaxel was attached to poly(hydroxyethylaspartamide) via a succinic spacer arm by a two-step protocol: (1) synthesis of 2'-O-succinyl-paclitaxel; (2) synthesis of PHEA-2'-O-succinyl-paclitaxel. The 2'-O-succinyl-paclitaxel derivative and the macromolecular conjugate were characterized by UV, IR, NMR and mass spectrometry analysis. The reaction yields were over 95% and the purity of products over 98%. Paclitaxel release and degradation from 2'-O-succinyl-paclitaxel occurred at a faster rate at pH 5.5 than 7.4. After 30 h of incubation at pH 5.5 and 7.4 the released free paclitaxel was about 40 and 20%, respectively. In plasma both drug release and degradation were found to occur at a higher rate than in buffer at pH 7.4 suggesting that an enzymatic mechanism could be involved. The paclitaxel release and degradation from PHEA-2'-O-succinyl-paclitaxel were negligible at pH 5.5 and 7.4 and very slow in plasma. Investigation carried out using murine myeloid cell line showed that the polymeric prodrug maintains partial pharmacological activity of paclitaxel. The DL50 of the conjugate (over 40 ng/ml) as compared to free paclitaxel (about 1 ng/ml) was correlated to the slow drug release. Finally a pharmacokinetic study carried out by intravenous inoculation of the macromolecular prodrug to mice demonstrated that the polymer conjugation modify dramatically the in vivo fate of the drug. The conjugate disappeared from the bloodstream much more quickly as compared to both free drug and naked polymer. Massive accumulation of bioconjugate in the liver (80% of the dose) was found to persist throughout 1 week.

Herceptin-geldanamycin immunoconjugates: pharmacokinetics, biodistribution, and enhanced antitumor activity

The efficacy of monoclonal antibodies (mAbs) as single agents in targeted cancer therapy has proven to be limited. Arming mAbs with a potent toxic drug could enhance their activity. Here we report that conjugating geldanamycin (GA) to the anti-HER2 mAb Herceptin improved the activity of Herceptin. The IC(50)s of the immunoconjugate H-GA were 10-200-fold lower than that of Herceptin in antiproliferative assays, depending on the cell line. The H-GA mode of action involved HER2 degradation, which was partially lactacystin sensitive and thus proteasome dependent. The linkage between GA and Herceptin remained stable in the circulation, as suggested by the pharmacokinetics of Herceptin and conjugated GA, which were almost identical and significantly different from that of free GA. Tumor uptake of Herceptin and H-GA were similar (52 +/- 7 and 43 +/- 7% of the initial injected dose per gram tissue, respectively; P = 0.077), indicating no apparent damage attributable to conjugation. Therapy experiments in xenograft-bearing mice consisted of weekly i.p. doses, 4 mg/kg for 4 months. H-GA showed a greater antitumor effect than Herceptin because it induced tumor regression in 69% of the recipients compared with 7% by Herceptin alone. Median survival time was 145 days as opposed to 78 days, and 31% of the recipients remained tumor free 2 months after therapy was terminated versus 0% in the Herceptin group. Enhancement of Herceptin activity could be of significant clinical value. In addition, the chemical linkage and the considerations in therapeutic regimen described here could be applied to other immunoconjugates for targeted therapy of a broad spectrum of cancers.

The lectin-cell interaction and its implications to intestinal lectin-mediated drug delivery

Based on the fact that oligosaccharides encode biological information, the biorecognition between lectinised drug delivery systems and glycosylated structures in the intestine can be exploited for improved peroral therapy. Basic research revealed that some lectins can mediate mucoadhesion, cytoadhesion, and cytoinvasion of drugs. Entering the vesicular pathway by receptor mediated endocytosis, part of the conjugated drug is accumulated within the lysosomes. Additionally, part of the drug is supposed to be transported across the epithelium. Moreover, factors probably adversely influencing feasibility of the concept such as toxicity, immunogenicity, and intestinal stability of plant lectins are discussed. As exemplified by lectin-grafted prodrug and carrier systems, this strategy is expected to improve absorption and probably bioavailability of poorly absorbable drugs, peptides and proteins as well as therapeutic DNA.

Enabling ScFvs as multi-drug carriers: a dendritic approach

To enable scFvs as multi-drug carriers, we designed and synthesized dendritic linker molecules bearing up to nine chlorambucil residues at the branch ends. A maleimide group was used at the focal point of the dendron for easy linkage to the scFv. Originally designed molecules showed poor water solubility. To address this problem, a lysine residue with an unprotected carboxylic acid group was inserted into the dendron branches. The new molecules showed excellent water solubility and are now suitable for conjugation. Such dendritic molecules will allow studies to understand the relationship between the drug/antibody ratio and the potency of the immunoconjugates. The dendritic approach could also be applied to drugs other than chlorambucil and carriers other than scFvs to greatly increase the drug/carrier ratio.

Recombinant antibodies for cancer diagnosis and therapy

Recombinant antibodies currently represent over 30% of biopharmaceuticals in clinical trials, highlighted by the recent Food and Drug Administration (FDA) approvals of Zevalin(TM) (ibritumomab-tiuxetan; IDEC Pharmaceuticals, San Dieago, CA, USA) for cancer radioimmunotherapy and Humira(TM) (adalimumab; Abbott Laboratories, IL, USA) for rheumatoid arthritis. Together, these FDA approvals have excited the biotechnology industry, particularly since sales of recombinant antibodies are increasing rapidly to a predicted US dollar 4 billion per annum worldwide in 2003. To date, 10 engineered therapeutic antibodies have gained FDA approval and many others are in Phase III trials. Many recent FDA-approved antibodies are simple molecular designs that have taken 10 years to be developed into effective therapeutic reagents. Emerging new technologies have created a vast range of recombinant, antibody-based reagents, which specifically target clinical biomarkers of disease. Radiolabelling of antibodies has increased their potential for cancer imaging and targeting. Recombinant antibodies have also been reduced in size and rebuilt into multivalent molecules for higher affinity. In addition, antibodies have been fused with many molecules, including toxins, enzymes, drugs and viruses, for prodrug therapy, cancer treatment and gene delivery. Recombinant antibody technology has enabled clever manipulations in the construction of complex in vitro libraries for the selection of high-affinity reagents against refractory targets. Furthermore, innovative affinity maturation methods have been developed which enable rapid selection of extremely high-affinity reagents. This review focuses on developments in the last 12 months and describes the latest developments in the design, production and clinical use of recombinant antibodies for cancer diagnosis and therapy.

Drug delivery strategy utilizing conjugation via reversible disulfide linkages: role and site of cellular reducing activities

The first disulfide linkage-employing drug conjugate that exploits the reversible nature of this unique covalent bond was recently approved for human use. Increasing numbers of drug formulations that incorporate disulfide bonds have been reported, particularly in the next generation macromolecular pharmaceuticals. These are designed to exploit differences in the reduction potential at different locations within and upon cells. The recent characterization of a novel redox enzyme in endosomes and lysosomes adds more excitement to this approach. This review focuses on understanding where and how the disulfide bond in the bioconjugate is reduced upon contact with biological milieu, which affects delivery design and the interpretation of the delivery strategies.

Cellular handling of a dexamethasone-anti-E-selectin immunoconjugate by activated endothelial cells: comparison with free dexamethasone

PURPOSE

For selective inhibition of endothelial cell activation in chronic inflammation, we have developed a dexamethasone-anti-E-selectin immunoconjugate. The present study was performed to evaluate the cellular handling of this immunoconjugate by activated primary endothelial cells and to compare its drug delivery capacity with free dexamethasone.

METHODS

The binding, uptake, and degradation of 125I-radiolabeled dexamethasone-anti-E-selectin immunoconjugate by TNFalpha-activated endothelial cells were studied for different time periods and at different concentrations, as well as in the presence of inhibitors for E-selectin binding and lysosomal degradation. Its drug delivery capacity was compared with the uptake of unconjugated 3H-labeled dexamethasone.

RESULTS

The immunoconjugate was internalized by E-selectin expressing activated endothelial cells and degraded in the lysosomal compartment. The receptor-mediated binding and uptake was saturable, implying a maximal attainable intracellular concentration of the drug. In contrast, free dexamethasone entered both resting and activated endothelial cells by passive diffusion.

CONCLUSIONS

The dexamethasone-anti-E-selectin immunoconjugate is capable of selective delivering the coupled drug into activated endothelial cells. This targeting concept enables disease-induced drug delivery in which intracellular concentrations can be reached comparable with those obtained after incubation with 3 FM dexamethasone.

Ligand-targeted therapeutics in anticancer therapy

Cytotoxic chemotherapy or radiotherapy of cancer is limited by serious, sometimes life-threatening, side effects that arise from toxicities to sensitive normal cells because the therapies are not selective for malignant cells. So how can selectivity be improved? One strategy is to couple the therapeutics to antibodies or other ligands that recognize tumour-associated antigens. This increases the exposure of the malignant cells, and reduces the exposure of normal cells, to the ligand-targeted therapeutics.

Intracellular signal-responsive artificial gene regulation for novel gene delivery

We describe two types of artificial gene-regulation systems responding to cyclic AMP-dependent protein kinase (PKA) or caspase-3. These molecular systems use newly synthesized cationic polymers, PAK and PAC. The PAK polymer includes substrate oligopeptide for PKA, ARRASLG, as receptor of PKA signal, while the PAC polymer possesses oligopeptide that is comprised of a substrate sequence of caspase-3, DEVD, and a cationic oligolysine, KKKKKK. These polymers formed stable complexes with DNA to totally suppress the gene expression. However, PKA or caspase-3 signal disintegrates the PAK-DNA or the PAC-DNA complex, respectively. This liberates the DNA and activated the gene expression. These systems are the first concept of an intracellular signal-responsive gene-regulation system using artificial polymer. We expect that these systems can be applied to the novel highly cell specific gene delivery strategy that is involved in our previously proposed new drug delivery concept, the drug delivery system based on responses to cellular signals.