Maximizing the safety profile of a novel cancer therapeutic via the de-immunization of an engineered mammalian asparaginase

Enzyme By Design, Inc., Chicago IL

Investigators

Abstract

Project Summary: The goal of this proposal is to develop a significantly deimmunized variant of the anti-cancer biologic drug L- asparaginase (ASNase). ASNases are enzyme drugs that systematically deplete L-asparagine from the blood and are a mainstay of treatment for acute lymphoblastic leukemia (ALL), a cancer of the white blood cells and the most prevalent pediatric cancer. Despite being highly effective in the treatment of pediatric ALL, ASNases are associated with a multitude of toxic side effects, some so severe as to be fatal. Their high toxicity precludes routine use to treat adult ALL, contributes to the much lower cure rate of <40% for adult ALL, and prevents their use in other hematological malignancies (e.g. acute myeloid leukemia) and solid tumors (e.g. pancreatic cancer), despite strong preclinical evidence for efficacy. There are two main sources for the toxicity: L- glutaminase (GLNase) coactivity and immunogenicity. All current FDA-approved ASNases have intrinsic GLNase coactivity AND are immunogenic due to their bacterial origins from E. coli (Elspar) or Erwinia chrysanthemi (Erwinaze). A PEGylated version of Elspar called Oncaspar was developed and is now the 1st- line treatment for pediatric ALL in the US. However, there is now evidence that antibodies against the PEG moiety are being generated. Moreover, some patients even have pre-existing PEG antibodies prior to treatment due to the increasing prevalence of PEG in everyday products. Hypersensitivity can abolish the drug's efficacy and halt ASNase treatment, which correlates with poorer outcomes. With these toxic side effects that can prevent or terminate treatment of this life-saving drug, there is a clear unmet clinical need for an ASNase that lacks GLNase coactivity and has reduced immunogenicity. We are currently developing an ASNase variant that addresses both problems as it is devoid of GLNase coactivity and predicted to be less immunogenic since it is mammalian in origin, specifically guinea pig (GpA), sharing 70% sequence identity to the human homolog. This is in sharp contrast to Elspar and Erwinaze, sharing only ~25% identity to the human enzyme. To further reduce the risk of immunogenicity, we used a structure-guided approach to truncate and further humanize GpA, resulting in our lead called GpA369hum, which is >80% identical to the human homolog. However, since even fully human proteins can be immunogenic, in the proposed work we will identify those regions in GpA369hum that are efficiently presented by MHCII molecules (Aim 1). We will then incorporate in silico predictors and our recently obtained crystal structure of GpA369hum to design deimmunized variants - GpA369hum-DI (Aim 2). Finally, we will confirm the reduced immunogenicity of GpA369hum-DI using an ex vivo dendritic cell/T (DC-T) cell proliferation assay (Aim 3). Whereas clinical immunogenicity is extremely complex and is impacted by many intrinsic and extrinsic factors, the proposed work will allow us to mitigate this risk by assessing potential immunogenicity issues at this early preclinical stage to avoid possible failure in the clinic due to immunogenicity, where costs are much higher, and more importantly, patients' lives are at risk.