Ancestral L-Asparaginases Are Cytotoxic Against ALL Cell Lines with Predicted Reduced Immunogenicity

L-asparaginase (L-ASNase) is a critical component of the chemotherapy regimen used to treat acute lymphoblastic leukemia (ALL). Current clinical L-ASNases are bacterial in origin, derived from either Escherichia coli or Erwinia chrysanthemi. Microbial L-ASNases share only ~30% sequence identity to human L-ASNase. Therefore, it is not surprising that they possess high allergenic and immunogenic risk profiles. Immune reactions to clinical L-ASNases have been reported in 5-30% of patients and ASNase discontinuation leads to significantly inferior disease-free survival (Gupta et al, JCO 2020), thus highlighting the need for a less toxic alternative. Herein, we describe the utilization of ancestral sequence reconstruction (ASR) as a protein drug discovery and optimization platform to create novel humanized L-ASNase protein drug candidates. Selection of the ASR approach was based on the foundational knowledge that guinea pig (GP) L-ASNase possesses favorable anti-leukemic and enzymatic properties while also sharing ~70% sequence identity to human L-ASNase.

The amino acid sequences of 10 ancestral (An) L-asparaginase candidates spanning the phylogeny connecting guinea pig to human were inferred, reverse translated into codon-optimized cDNAs, de novo synthesized, and expressed using a bacterial protein expression platform. An-L-ASNase variants activity were determined via a modified Nessler’s reagent assay. An-104 and An-107, both of which whilst having 88% sequence identity to human L-ASNase, demonstrated comparable asparaginase activity to clinically relevant L-ASNases when tested at an enzyme concentration of 0.1 mg/mL and an asparagine substrate concentration of 1 μM. Next, An-L-ASNases were tested for cytotoxic activity against two ALL cell lines known to be highly sensitive to L-ASNase: CCRF-CEM and MOLT-4. Utilizing the colorimetric MTT assay to measure the half-maximal inhibitory concentration (IC₅₀), by correlating metabolic activity with relative number of viable cells, An-104 and An-107 displayed IC₅₀ values comparable to published E. coli L-ASNase data. We also assessed the activity of An-104 against the MOLT-4 cell line, using an alternative method via trypan blue staining to determine cell counts and viability. An-104 had impressive killing against these cells with the calculated IC₅₀ in a similar range to that reported of Rylaze®, the recently FDA approved Erwinia L-ASNase.

As severe toxicity from the clinically available L-ASNases can result in discontinuation, in addition to precluding widespread utilization amongst adult population, bioengineering an equally efficacious, less immunogenic alternative remains paramount. The HLA-DRB1*07:01 allele has been associated with the highest rate of hypersensitivity reactions to L-asparaginase, based upon analysis of 1870 patients enrolled on leukemia trials at St. Jude Children’s Research Hospital and the Children’s Oncology Group (Fernandez et al. Blood 2014). Using the IEDB analysis resource NetMHCIIpan (ver. 4.1) tool, the number of immunogenic epitopes in the functional domain of An-L- ASNases was assessed in the context of HLA-DRB1*07:01 using 15-mer peptides and defining percentile rank < 10 as the immunogenic threshold. After accounting for human epitopes, all ancestral proteins were predicted to have lower immunogenicity than the bacterial L-ASNases, with An-104 and An-107 predicted to have 2 to 3-fold lower immunogenicity. Notably several immunogenic peptides were identified in the C-terminal ankyrin repeat domain of the ancestral proteins, the analogous domain of which in humans is known to be non-functional. Therefore, we have created ancestral-human hybrid L-ASNases incorporating the non-functional human C-terminal domain resulting in 90-99% sequence identity to human L-ASNase.

In summary using ASR as a protein discovery and optimization platform, we have generated An-L-ASNase candidates possessing i) comparable ASNase-specific activity, ii) similar cytotoxic activity against ALL cell lines, and iii) lower predicted immunogenicity risk compared to existing commercial ASNase products. Based on these results and ongoing optimization studies, we expect to advance a bio-better next-generation humanized ASNase candidate into clinical development.