Toan D. Nguyen

Anti-cancer antibody-drug conjugates (ADCs) aim to expand the therapeutic index of traditional chemotherapy by employing the targeting specificity of monoclonal antibodies (mAbs) to increase the efficiency of the delivery of potent cytotoxic agents to malignant cells. In the past three years, the number of ADCs approved by the Food and Drug Administration (FDA) has tripled. Although several ADCs have demonstrated sufficient efficacy and safety to warrant FDA approval, the clinical use of all ADCs leads to substantial toxicity in treated patients, and many ADCs have failed during clinical development due to their unacceptable toxicity profiles. Analysis of the clinical data has demonstrated that dose-limiting toxicities (DLTs) are often shared by different ADCs that deliver the same cytotoxic payload, independent of the antigen that is targeted and/or the type of cancer that is treated. DLTs are commonly associated with cells and tissues that do not express the targeted antigen (i.e., off-target toxicity), and often limit ADC dosage to levels below those required for optimal anti-cancer effects. In this manuscript, we review the fundamental mechanisms contributing to ADC toxicity, we summarize common ADC treatment-related adverse events, and we discuss several approaches to mitigating ADC toxicity.

Small therapeutic proteins are receiving increased interest as therapeutic drugs; however, their clinical success has been limited due to their rapid elimination. Here, we report a half-life extension strategy via strategy via red blood cell red blood cell (RBC) hitch-hiking. This manuscript details the development and characterization of novel anti-RBC single-domain antibodies (sdAbs), their genetic fusion to therapeutic antibody fragments (TAF) as bispecific fusion constructs, and their influence on TAF pharmacokinetics and biodistribution. Several sdAbs specific to the band 3 antigen were generated via phage-display technology. Binding affinity to RBCs was assessed via flow cytometry. Affinity maturation via random mutagenesis was carried out to improve the binding affinity of the sdAbs. Bi-specific constructs were generated by fusing the anti-RBC sdAbs with anti-tissue necrosis factor alpha (TNF-α) TAF via the use of a glycine-serine flexible linker, and assessments for binding were performed via enzyme-linked immunosorbent assay and flow cytometry. Pharmacokinetics of anti-RBC sdAbs and fusion constructs were evaluated following intravenous bolus dosing in mice at a 1 mg/kg dose. Two RBC-binding sdAbs, RB12 and RE8, were developed. These two clones showed high binding affinity to human RBC with an estimated K_D of 17.7 nM and 23.6 nM and low binding affinity to mouse RBC with an estimated K_D of 335 nM and 528 nM for RB12 and RE8, respectively. Two derivative sdAbs, RMA1, and RMC1, with higher affinities against mouse RBC, were generated via affinity maturation (K_D of 66.9 nM and 30.3 nM, respectively). Pharmacokinetic investigations in mice demonstrated prolonged circulation half-life of an anti-RBC-TNF-α bispecific construct (75 h) compared to a non-RBC binding control (1.3 h). In summary, the developed anti-RBC sdAbs and fusion constructs have demonstrated high affinity in vitro, and sufficient half-life extension in vivo.

Our group has developed and experimentally validated a strategy to increase antibody penetration in solid tumors through transient inhibition of antibody-antigen binding. In prior work, we demonstrated that 1HE, an anti-trastuzumab single domain antibody that transiently inhibits trastuzumab binding to HER2, increased the penetration of trastuzumab and increased the efficacy of ado-trastuzumab emtansine (T-DM1) in HER2+ xenograft bearing mice. In the present work, 1HE variants were developed using random mutagenesis and phage display to enable optimization of tumor penetration and efficacy of trastuzumab-based therapeutics. To guide the rational selection of a particular 1HE mutant for a specific trastuzumab-therapy, we developed a mechanistic pharmacokinetic (PK) model to predict within-tumor exposure of trastuzumab/T-DM1. A pharmacodynamic (PD) component was added to the model to predict the relationship between intratumor exposure to T-DM1 and the corresponding therapeutic effect in HER2+ xenografts. To demonstrate the utility of the competitive inhibition approach for immunotoxins, PK parameters specific for a recombinant immunotoxin were incorporated into the model structure. Dissociation half-lives for variants ranged from 1.1 h (for variant LG11) to 107.9 h (for variant HE10). Simulations predicted that 1HE co-administration can increase the tumor penetration of T-DM1, with inhibitors with longer trastuzumab binding half-lives relative to 1HE (15.5 h) further increasing T-DM1 penetration at the expense of total tumor uptake of T-DM1. The PK/PD model accurately predicted the response of NCI-N87 xenografts to treatment with T-DM1 or T-DM1 co-administered with 1HE. Model predictions indicate that the 1HE mutant HF9, with a trastuzumab binding half-life of 51.1 h, would be the optimal inhibitor for increasing T-DM1 efficacy with a modest extension in the median survival time relative to T-DM1 with 1HE. Model simulations predict that LG11 co-administration will dramatically increase immunotoxin penetration within all tumor regions. We expect that the mechanistic model structure and the wide range of inhibitors developed in this work will enable optimization of trastuzumab-cytotoxin penetration and efficacy in solid tumors.