Novel strategies towards engineering therapeutic enzymes with reduced immunogenicity for cancer therapy

Abstract

Heterologous enzymes have been investigated for a variety of therapeutic applications, including the treatment of a number of cancers that are sensitive to the systemic depletion of specific amino acids. One such example is acute lymphoblastic leukemia (ALL) for which enzyme-mediated L-Asparagine (L-Asn) depletion by the Escherichia coli L-Asparaginase II (EcAII) has been proven critical for treatment. However, the repeated or prolonged therapeutic administration of such enzymes is restricted by their immunogenicity, which frequently results in the generation of anti-enzyme antibodies that may in turn mediate a variety of adverse hypersensitivity reactions and neutralization of the enzymes themselves. Thus, while the therapeutic efficacy of asparaginase is well established, a significant number of patients still develop adverse immune responses to the enzyme. Here, we have developed and explored novel strategies towards engineering an asparaginase with reduced immunogenicity for ALL therapy. First, we identified and investigated human enzymes that putatively shared functional similarity to asparaginase with the long-term aim of engineering such enzymes to acquire biochemical and pharmacological properties requisite for eventual therapeutic application. In one study, we described the bacterial expression and characterization of the human asparaginase-like protein 1 (hASRGL1). We presented evidence that hASRGL1 exhibited an activity profile consistent with enzymes previously designated as [Beta]-aspartyl peptidases, which had only been previously identified in plants and bacteria. Similar to non-mammalian [Beta]-aspartyl peptidases, hASRGL1 was revealed to be an N-terminal nucleophile (Ntn) hydrolase whereby Thr168 serves as the essential Ntn for both intramolecular processing and catalysis. In a second study, we described the optimized bacterial expression and biochemical characterization of the human N-terminal asparagine amidohydrolase 1 (hNTAN1). We demonstrated that hNTAN1 catalysis is dependent upon direct involvement of a thiol group, and subsequently identified Cys75 as an essential residue that may act as the catalytic nucleophile. Further, we presented the first description of hNTAN1 kinetics, secondary structure composition, and thermal stability. Second, we devised and validated a novel therapeutic deimmunization approach by combinatorial T-cell epitope removal using neutral drift. We showed that combinatorial saturation mutagenesis coupled with a robust neutral drift screen enabled the isolation of engineered EcAII variants that contained multiple amino acid substitutions yet exhibited catalytic efficiencies nearly indistinguishable to that of the parent enzyme. Three regions of EcAII were computationally identified as putative T-cell epitopes and then subjected to saturation mutagenesis at 4 positions (per region) believed to be critical for MHC-II binding. The resulting libraries were then sequentially subjected to a neutral drift FACS screen in order to isolate EcAII mutants that retained wild-type function. Pools of neutral drift variants were then computationally evaluated for MHC-II binding and those that displayed scores indicative of compromised binding were purified and biochemically characterized. Finally, T-cell activation assays and antibody titers in HLA-transgenic mice were used to evaluate T-cell epitope removal and immunogenicity, respectively. Ultimately, we revealed that mice immunized with an EcAII neutral variant containing 8 amino acid substitutions -- 3 of which were non-phylogenetically conserved -- within computationally predicted T-cell epitopes, displayed a significant 10-fold reduction in serum anti-EcAII IgG titer relative to mice similarly immunized with the parent enzyme.