Giulia Agnello

Engineering Reduced-Immunogenicity Enzymes for Amino Acid Depletion Therapy in Cancer

Cancer has become the leading cause of death in the developed world and has remained one of the most difficult diseases to treat. One of the difficulties in treating cancer is that conventional chemotherapies often have unacceptable toxicities toward normal cells at the doses required to kill tumor cells. Thus, the demand for new and improved tumor specific therapeutics for the treatment of cancer remains high. Alterations to cellular metabolism constitute a nearly universal feature of many types of cancer cells. In particular, many tumors exhibit deficiencies in one or more amino acid synthesis or salvage pathways forcing a reliance on the extracellular pool of these amino acids to satisfy protein biosynthesis demands. Therefore, one treatment modality that satisfies the objective of developing cancer cell-selective therapeutics is the systemic depletion of that tumor-essential amino acid, which can result in tumor apoptosis with minimal side effects to normal cells. While this strategy was initially suggested over 50 years ago, it has been recently experiencing a renaissance owing to advances in protein engineering technology, and more sophisticated approaches to studying the metabolic differences between tumorigenic and normal cells. Dietary restriction is typically not sufficient to achieve a therapeutically relevant level of amino acid depletion for cancer treatment. Therefore, intravenous administration of enzymes is used to mediate the degradation of such amino acids for therapeutic purposes. Unfortunately, the human genome does not encode enzymes with the requisite catalytic or pharmacological properties necessary for therapeutic purposes. The use of heterologous enzymes has been explored extensively both in animal studies and in clinical trials. However, heterologous enzymes are immunogenic and elicit adverse responses ranging from anaphylactic shock to antibody-mediated enzyme inactivation, and therefore have had limited utility. The one notable exception is Escherichia colil-asparaginase II (EcAII), which has been FDA-approved for the treatment of childhood acute lymphoblastic leukemia. The use of engineered human enzymes, to which natural tolerance is likely to prevent recognition by the adaptive immune system, offers a novel approach for capitalizing on the promising strategy of systemic depletion of tumor-essential amino acids. In this work, we review several strategies that we have developed to: (i) reduce the immunogenicity of a nonhuman enzyme, (ii) engineer human enzymes for novel catalytic specificities, and (iii) improve the pharmacological characteristics of a human enzyme that exhibits the requisite substrate specificity for amino acid degradation but exhibits low activity and stability under physiological conditions.

Uncoupling Intramolecular Processing and Substrate Hydrolysis in the N-Terminal Nucleophile Hydrolase hASRGL1 by Circular Permutation.

The human asparaginase-like protein 1 (hASRGL1) catalyzes the hydrolysis of l-asparagine and isoaspartyl-dipeptides. As an N-terminal nucleophile (Ntn) hydrolase superfamily member, the active form of hASRGL1 is generated by an intramolecular cleavage step with Thr168 as the catalytic residue. However, in vitro, autoprocessing is incomplete (~50%), fettering the biophysical characterization of hASRGL1. We circumvented this obstacle by constructing a circularly permuted hASRGL1 that uncoupled the autoprocessing reaction, allowing us to kinetically and structurally characterize this enzyme and the precursor-like hASRGL1-Thr168Ala variant. Crystallographic and biochemical evidence suggest an activation mechanism where a torsional restraint on the Thr168 side chain helps drive the intramolecular processing reaction. Cleavage and formation of the active site releases the torsional restriction on Thr168, which is facilitated by a small conserved Gly-rich loop near the active site that allows the conformational changes necessary for activation.

Discovery of a substrate selectivity motif in amino acid decarboxylases unveils a taurine biosynthesis pathway in prokaryotes.

Taurine, the most abundant free amino acid in mammals, with many critical roles such as neuronal development, had so far only been reported to be synthetized in eukaryotes. Taurine is the major product of cysteine metabolism in mammals, and its biosynthetic pathway consists of cysteine dioxygenase and cysteine sulfinic acid decarboxylase (hCSAD). Sequence, structural, and mutational analyses of the structurally and sequentially related hCSAD and human glutamic acid decarboxylase (hGAD) enzymes revealed a three residue substrate recognition motif (X1aa19X2aaX3), within the active site that is responsible for coordinating their respective preferred amino acid substrates. Introduction of the cysteine sulfinic acid (CSA) motif into hGAD (hGAD-S192F/N212S/F214Y) resulted in an enzyme with a >700 fold switch in selectivity toward the decarboxylation of CSA over its preferred substrate, l-glutamic acid. Surprisingly, we found this CSA recognition motif in the genome sequences of several marine bacteria, prompting us to evaluate the catalytic properties of bacterial amino acid decarboxylases that were predicted by sequence motif to decarboxylate CSA but had been annotated as GAD enzymes. We show that CSAD from Synechococcus sp. PCC 7335 specifically decarboxylated CSA and that the bacteria accumulated intracellular taurine. The fact that CSAD homologues exist in certain bacteria and are frequently found in operons containing the recently discovered bacterial cysteine dioxygenases that oxidize l-cysteine to CSA supports the idea that a bona fide bacterial taurine biosynthetic pathway exists in prokaryotes.

GFP Reporter Screens for the Engineering of Amino Acid Degrading Enzymes from Libraries Expressed in Bacteria

There is significant interest in engineering human amino acid degrading enzymes as non-immunogenic chemotherapeutic agents. We describe a high-throughput fluorescence activated cell sorting (FACS) assay for detecting the catalytic activity of amino acid degrading enzymes in bacteria, at the single cell level. This assay relies on coupling the synthesis of the GFP reporter to the catalytic activity of the desired amino acid degrading enzyme in an appropriate E. coli genetic background. The method described here allows facile screening of much larger libraries (10(6)-10(7)) than was previously possible. We demonstrate the application of this technique in the screening of libraries of bacterial and human asparaginases and also for the catalytic optimization of an engineered human methionine gamma lyase.

Abstract 1042: Development of AEB1102, an engineered human arginase 1 for patients with solid tumors

Introduction: Tumor dependence on specific amino acids for survival and proliferation is well recognized and has been exploited effectively with the use of Asparaginase for the treatment of Acute Lymphoblastic Leukemia. Decades of research have led to an understanding of tumor L-Arginine (L-Arg) dependence, with functional expression of the three enzymes of the L-Arg biosynthetic pathway: Ornithine Transcarbamylase (OTC), Argininosuccinate Synthase (ASS1) and Argininosuccinate Lyase (ASL) being required to convert ornithine to L-Arg. In a variety of tumor types, silencing of one or more of these enzymes disables endogenous L-Arg synthesis forcing a reliance on the extracellular pool of L-Arg for tumor survival and proliferation. This mechanism has been confirmed with Arginine Deiminase (ADI-PEG) and pegylated wild-type Arginase I (BCT-100). Aeglea Biotherapeutics Inc. has developed an alternative approach using a bioengineered human PEGylated Arginase I with enhanced pharmacological properties. Replacement of manganese, the natural metal co-factor in wild-type human Arginase I, with cobalt confers improved catalytic activity and serum stability. The resulting product candidate (AEB1102) displays highly favorable PK/PD properties, and is expected to be naturally tolerated by the human immune system as no protein modifications have been introduced. Experimental Procedures and Results: Non-clinical dose range finding and GLP toxicology studies were performed with AEB1102 in monkeys and mice. Bioanalytical assays detecting L-Arg and AEB1102 enzyme activity were used to monitor PD/PK in the dose range finding studies with these results subsequently being used to design the toxicology studies. Subsequent toxicology studies identified an NOAEL in both species at a dose that is predicted to translate to significant sustained reduction of L-Arg serum levels with weekly intravenous dosing. To determine tumor types most likely to respond to AEB1102 expression profiling of OTC, ASS1 and ASL was performed using in situ hybridization on multiple tumor histologies. Melanoma was identified as a tumor type likely to respond to AEB1102 owing to a significant reduction in ASS1 expression. Non-clinical in vivo studies using the A375 melanoma xenograft as well as melanoma PDx models confirm sensitivity to AEB1102, weekly dosing resulted in a significant delay in tumor growth and improved survival. Conclusion: Non-clinical activities required to support the IND submission of AEB1102 for solid tumors were successfully executed, enabling the Phase 1 study to be initiated in October 2015. Translational work profiling the expression of OTC, ASS1 and ASL has identified melanoma as a relevant tumor type to pursue in future clinical studies. Citation Format: Scott W. Rowlinson, Susan E. Alters, Giulia Agnello, Joseph Tyler, Ann Lowe, Mauri Okamoto-Kearney, Dale Johnson, Everett Stone, George Georgiou, David G. Lowe. Development of AEB1102, an engineered human arginase 1 for patients with solid tumors. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 1042.

Abstract 3964: Reducing systemic arginine with arginase (AEB1102) therapy does not suppress the immune response induced by anti-PD-1 and anti-PD-L1, and exerts an additive anti-tumor and synergistic survival benefit

Tumor dependence on specific amino acids for survival and proliferation is well recognized and has been exploited effectively with the use of asparaginases for the treatment of acute lymphoblastic leukemia. Sensitivity of tumors to L-Arginine (L-Arg) deprivation results from an impaired ability to make L-Arg as a result of decreased functional expression of one or more of the three enzymes of the L-Arg biosynthetic pathway: ornithine transcarbamylase (OTC), argininosuccinate synthase (ASS1) and argininosuccinate lyase (ASL). Native human arginase I is not a viable drug candidate due to low activity and low stability in serum. We have developed an alternative approach using a bioengineered human PEGylated arginase I (AEB1102) with enhanced pharmacological properties. We and others have successfully utilized arginase I to impart a direct tumor cell killing effect through L-Arg starvation in multiple tumor types e.g. AEB1102 single agent efficacy in melanoma, small cell lung cancer (SCLC) and sarcoma PDx models. However, the compatibility of AEB1102 with checkpoint inhibitors is unclear as arginase I has been shown to be both immune suppressive and immune neutral (PMID: 23717444), or immune promoting (PMID: 27043409). Because of these conflicting reports we decided to investigate the impact of systemic L-Arg removal on the anti-tumor efficacy of check point inhibitors. Murine syngeneic models (e.g. CT26, MC38 and LL2) were dosed with AEB1102 alone and in combination with anti-PD-L1 and anti-PD-1 monoclonal antibodies (mAbs). Depending on the model, in vivo treatment with AEB1102 monotherapy resulted in tumor growth inhibition (TGI) ranging from 52% to 72% compared to the untreated control group, whereas standard monotherapy using immunomodulatory mAbs that target PD-1 and PD-L1 resulted in TGI ranging from 12% to 60%. Of significance, combination therapy of AEB1102 with anti-PD-1 or anti-PD-L1 resulted in additive anti-tumor effect with TGI ranging from 60% to 86%. In the CT26 model, when AEB1102 was administered in combination with anti-PD-L1 for at least 6 weeks, a 33% frequency of complete tumor regression (non-palpable tumors) was observed, indicating that synergy occurs with combination therapy. Collectively these results demonstrate that disrupting the L-Arg physiological balance in the tumor microenvironment inhibits tumor growth and further sensitizes the tumor to immunotherapy when AEB1102 is combined with anti-PD-1 and anti-PD-L1. These data open the possibility of further improving outcomes in L-Arg dependent tumors through combination of AEB1102 with anti-PD-1 and anti-PD-L1 inhibitors. Citation Format: Giulia Agnello, Susan E. Alters, David G. Lowe, Scott W. Rowlinson. Reducing systemic arginine with arginase (AEB1102) therapy does not suppress the immune response induced by anti-PD-1 and anti-PD-L1, and exerts an additive anti-tumor and synergistic survival benefit [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3964. doi:10.1158/1538-7445.AM2017-3964