(272c) Methods and Application of Ligand Charge Engineering Towards Improved Tumor Detection

Methods and Application of Ligand Charge Engineering towards Improved Tumor Detection

Brett A. Case, Nicole Olson, Benjamin J. Hackel

Small, engineerable protein scaffolds – such as affibodies and fibronectin domains – have been demonstrated as effective targeting agents for molecular imaging and therapy.  These affinity ligands, modifiable through directed evolution towards numerous targets, have favorable pharmacokinetic properties for select applications.  Small size drives rapid clearance from the blood and healthy tissue making them suitable for imaging with short-lived positron emitting isotopes including ⁶⁸Ga (t_1/2 = 68 min.) and ¹⁸F (t_1/2= 110 min.).  However off-target uptake and retention, particularly in the clearance organs, hinders the effectiveness of many scaffolds.  Such signal prevents comprehensive imaging proximal to these organs and presents toxicity concerns for therapeutic and diagnostic applications.  Anecdotal data suggests that there may be a correlation between excretory retention and protein charge, necessitating comparative analysis across ligands of differing charge. However, significant complications can arise when modulating protein charge. As charged residues are exchanged for neutral amino acids, or vice-versa, the protein’s native fold may be altered leading to the formation of insoluble aggregates. Additionally, loss of intramolecular structure reduces thermal stability and may be detrimental to target binding dependent upon paratope positioning.

This study aimed to create and analyze protein ligands with reduced charge using an affibody targeted to epidermal growth factor receptor (EGFR) as a basis. The affibody scaffold is a 58 amino acid, three helical bundle based on the B domain of staphylococcal protein A, typically diversified for binding at 13 positions on the first and second helices. The EGFR affibody (EA68), a hybrid of an evolved paratope and an evolved framework, exhibited good affinity (K_d = 5.3±1.7 nM) and stability (T_m = 73°C). Analysis of charge mutants emphasized maintaining or increasing thermal stability, recombinant production in the soluble fraction of E. coli, and target affinity. Three experimental models were used create mutants based on EA68 and with 3 positive and 3 negative residues replaced by neutral amino acids. Natural homolog sequence frequency guided the design of charge-reduced ligands covering narrow sequence space (11 clones purified from bacterial productions) as well as broad (>650,000 clones via a yeast surface display library). In addition, a consensus design approach was used based on sequence tolerance observed within the functional members of the library.

Charge reduction was generally well tolerated among all models. Soluble recovery yield equaled or surpassed that of parental EA68 in all cases; thermal stability ranged between 48°C and 71°C with a mean of 62°C; and EGFR affinity remained below 20 nM for all but two clones with 8 mutants retaining single nanomolar dissociation constants. Narrow and broad models each provided a preeminent ligand with yields in excess of 5 mg/L, affinities of 6.9±1.4 nM and 1.3±0.3 nM, and stabilities of 72°C and 69°C respectively.  

Methods for modifying charge while retaining or improving desirable ligand characteristics including bacterial production yield, thermal stability, and target affinity, particularly in the context of an EGFR-targeting affibody will be discussed. Moreover, application of these results to additional scaffolds will be shared.  In addition to providing a thorough data set elucidating the ability to modulate charge within this hybrid affibody, the functional mutants provide a diverse family of charge distributions with which to evaluate biodistribution. Comparison of physiological distribution, plasma clearance, and tumor targeting -- via PET/CT imaging and excised tissue gamma counting -- in a xenograft mouse model will be covered.