Utilizing Ancestral Sequence Reconstruction to Generate a Novel Recombinant Humanized L-Asparaginase with Enhanced Chemotherapeutic Properties

Abstract ID 100580 Poster Board 090 L-asparaginase (L-ASNase) has been a critical component of acute lymphoblastic leukemia (ALL) chemotherapy regimens for several decades, however, recent preclinical data has shown that L-ASNase may also be beneficial in several adult solid malignancies such as pancreatic cancer, colorectal cancer and metastatic breast cancer. Current clinical L-ASNases are bacterial in origin and are thus highly immunogenic, with reactions ranging from silent inactivation to severe anaphylaxis. Additionally, adults can have significant liver and pancreatic toxicity with L-ASNase, with studies showing this is at least partly related to the glutaminase co-activity seen in bacterial L-ASNases. Thus, development of a less immunogenic L-ASNase with reduced glutaminase activity is essential to overcome the current limitations of L-ASNase and expand its use beyond ALL. Human L-ASNase has evolved to have inferior catalytic properties which restricts its candidacy as a therapeutic enzyme. However, the utility of L-ASNase as a chemotherapeutic was first established in guinea pig (GP) L-ASNase, discovered first by J.G. Kidd in 1953 and then confirmed by J.D. Broome in 1963. GP L-ASNase has significantly superior enzyme kinetics compared to human L-ASNase and shares ∼70% sequence identity with human L-ASNase compared to ∼30% by bacterial L-ASNases and is thus predicted to be less immunogenic. Additionally, GP L-ASNase has no known glutaminase co-activity. Thus, GP L-ASNase is the ideal ortholog to serve as a template for optimization to create a more humanized less toxic L-ASNase variant. Ancestral Sequence Reconstruction (ASR) is an innovative protein drug discovery and optimization platform that can be leveraged to improve the pharmaceutical properties of L-ASNase. Analysis of the predicted molecular evolution of L-ASNase maps the functional divergence of extant orthologs by means of evolutionary intermediaries, enabling the identification of critical residues responsible for superior activity. ASR was performed utilizing 54 extant L- ASNase sequences, aligned using MUSCLE and an evolutionary tree inferred using MrBayes. 53 ancestral L-ASNase sequences were identified and ten ancestral variants spanning the ancient primate and GP lineage resurrected. E. coli codon optimized complementary DNA sequences were subcloned into an expression vector and transformed into E. coli BL21 (DE3) cells for protein expression. An-ASNase candidates were isolated through Ni2+ affinity chromatography, followed by purification with size exclusion chromatography. L-ASNase activity was assessed using a modified Nessler's reagent assay in a continuous spectroscopic enzyme-coupled assay. At an enzyme concentration of 0.1 mg/mL and an asparagine substrate concentration of 1 μM, An-88, An-104, and An-107 exhibited outstanding L-ASNase activity, comparable to clinically relevant E. coli and Erwinia L-ASNases. An-88 has 81% similarity, while both An-104 and An-107 ASNases shared an 88% identity with human L-ASNase. Preliminary cytotoxicity assessments of An-104 and An-107 on a T-ALL cell line, CCRF-CEM, demonstrated comparable anti-leukemia cytotoxicity to existing bacterial L-ASNases, with An-107 demonstrating the highest cytotoxicity. Thus, we have shown that ASR is a viable platform to bioengineer a less toxic humanized L-ASNase drug candidate. Lead candidate toxicity profile will be defined, and chemotherapeutic potential will be measured against hematologic and solid tumors.