Development and Evaluation of Catch-and-Release Anti-Carcinoembryonic Antigen Monoclonal Antibodies
Engler, Frank A.
MetadataShow full item record
“Catch-and-release” (CAR) antibodies exhibit pH-dependent binding to target ligands and receptors, with high affinity at neutral or slightly basic pH, and with markedly reduced binding affinity at acidic pH. This pH dependency allows CAR antibodies to bind (i.e., “catch”) target proteins in extracellular fluid, and to dissociate (i.e., “release”) from target within acidified endosomes. Although this is a rapidly emerging area with few published reports, prior work has demonstrated that some CAR monoclonal antibodies (mAb) exhibit the desirable pharmacokinetic characteristics of prolonged systemic exposure and decreased clearance, which has been attributed to decreased target-mediated elimination. After considering the impact of target-mediated elimination on the distribution of mAb in solid tumors, and considering the possible advantages of liberation of mAb from membrane binding in endosomes, this research group has proposed that anti-cancer CAR mAb will exhibit increased tumor selectivity and that anti-cancer CAR immunoconjugates may be employed to facilitate the delivery of macromolecular toxins to the cytoplasm of targeted cells. This dissertation has pursued this area of research through the development and evaluation of a novel catch-and-release monoclonal anti-carcinoembryonic antigen antibody. Monoclonal antibodies (mAb) were developed via standard hybridoma technology, following immunization of mice with purified human carcinoembryonic antigen (CEA). Several anti-CEA immune gamma globulin (IgG) mAb were produced, purified, and evaluated for pH-dependent binding with use of newly developed enzyme-linked immunosorbent assays (ELISA) and with the use of surface plasmon resonance (SPR) analyses. 10H6, a murine IgG1κ mAb, exhibited dramatically reduced CEA binding as pH was reduced from 7.4 to 5.5. The antibody, which was shown to bind to extracellular epitopes of CEA via flow cytometry, was found to exhibit reduced clearance in mice bearing tumors that express CEA (i.e., when compared to pharmacokinetic data for T84.66, a murine IgG1? mAb that exhibits pH-independent binding, with high affinity for CEA at pH 7.4 and 5.5). Using data collected for T84.66 in mice bearing tumors with (MC38 CEA+ ) or without (MC38 CEA− ) CEA expression, a new physiologically-based pharmacokinetic model was developed and applied to predict the disposition of anti-CEA CAR mAb. The model, which well-predicted the plasma pharmacokinetics of 10H6, also predicted increased tumor selectivity (i.e., increased cumulative tumor exposure relative to cumulative plasma exposure) for anti-CEA CAR mAb. The proposed use of anti-cancer CAR mAb to facilitate the delivery of macromolecules to the cytoplasm of targeted cells is based on the hypothesis that immunoconjugates of CAR mAb with endosomal escape peptides (EEP) will allow highly efficient pore formation and/or lysis of endosomes. Combined administration of CAR-EEP conjugates with catch-and-release immunotoxins (CAR-toxin) is anticipated to allow simultaneous receptor-mediated endocytosis of the conjugates within targeted cells. Following acidification of the endosomes, CAR-EEP and CAR-toxin conjugates will release from binding at the endosomal membrane, allowing efficient EEP-mediated lysis of endosomal membranes with delivery of CAR-toxin conjugates to the cytoplasm. We refer to this approach as a “ CAR-Bomb ” strategy, owing to the use of C atch A nd R elease mAb to allow for highly selective delivery of potent macromolecular toxins ( Bomb ). To pursue the CAR-Bomb strategy, gelonin, a type-1 ribosomal inactivating protein, was selected as a model toxin, and several cell penetrating peptides were considered for allowing endosomal escape, including TAT, GALA, EB1, and poly-arginine (RRR). The pharmacokinetics of unconjugated gelonin were studied through the use of newly developed anti-gelonin mAb and newly developed and validated ELISA. Conjugates of 10H6 with TAT, GALA, EB1, and RRR were prepared and evaluated for use in mediating membrane lysis with a model system that assessed the release of hemoglobin from red blood cells. Cell-cytotoxicity experiments were performed, using CEA-expressing tumor cells in culture, to evaluate cell killing mediated by unconjugated gelonin, anti-CEA mAb, gelonin immunoconjugates, EEP, conjugates of anti-CEA mAb and EEP, and various combinations. 10H6-EEP conjugates were able to permeabilize red blood cell membranes for release of hemoglobin at pH 5.5, but no lytic activity was seen at physiologic pH 7.4. Cell cytotoxicity studies demonstrated a 2.4-fold decrease in the IC50 of gelonin when coadministered with 10H6-GALA. Conjugates of anti-CEA mAb-gelonin provided a 13-fold decrease in IC50 (relative to molar concentrations of unconjugated gelonin). Pharmacokinetic investigations were then conducted to evaluate the influence of selected cell-penetrating peptides (TAT, GALA) on 10H6 disposition. Results showed that TAT-conjugation had little effect on either the plasma or the tissue pharmacokinetics of 10H6. GALA-conjugation substantially decreased exposure in the plasma and tissues, and increased relative exposure (i.e., the ratio of cumulative tissue exposure to cumulative plasma exposure) in the kidney and liver. These findings may allow rational development of strategies for 10H6-GALA dosing for future evaluation of the CAR-Bomb delivery approach. Although the CAR-Bomb approach is expected to allow for high selectivity with low potential for off-target toxicity, antibody drug conjugates have generally been associated with narrow therapeutic indicies. As such, there is considerable interest in the sources of interindividual variability in mAb disposition, as such variability clearly impacts the safe and effective use of immunoconjugates. Based on recent reports regarding the influence of diabetic nephropathy on the urinary elimination of IgG antibodies, we speculated that mAb clearance may be significantly increased in this disease state. To pursue this hypothesis, experiments were performed in a mouse model of diabetic nephropathy (streptozotocin model) to assess the influence of the disease state on the pharmacokinetics of a model mAb. This work demonstrated a substantial increase of mAb clearance in the streptozotocin-treated mouse model, with high correlation between the systemic clearance of mAb with the urinary albumin excretion rate (which may be readily measured in the clinic). This dissertation has led to the development of new reagents (e.g., anti-gelonin mAb, anti-CEA mAb, CAR-EEP and CAR-gelonin conjugates) and new methodologies to pursue new pharmacokinetic hypotheses. Our experimental data and mathematical modeling support the use of CAR mAb for increased tumor selectivity, and for possible use in enhancing the activity of macromolecular toxins. The conjugation methods, assays, PBPK model, and pharmacokinetic data will be employed in subsequent work to further the development of the CAR-Bomb strategy and to guide the design of in vivo studies.