64Cu-Labeled Gp2 Domain for PET Imaging of Epidermal Growth Factor Receptor

Predicting and Interpreting Protein Developability Via Transfer of Convolutional Sequence Representation

Engineered proteins have emerged as novel diagnostics, therapeutics, and catalysts. Often, poor protein developability─quantified by expression, solubility, and stability─hinders utility. The ability to predict protein developability from amino acid sequence would reduce the experimental burden when selecting candidates. Recent advances in screening technologies enabled a high-throughput (HT) developability dataset for 105 of 1020 possible variants of protein ligand scaffold Gp2. In this work, we evaluate the ability of neural networks to learn a developability representation from a HT dataset and transfer this knowledge to predict recombinant expression beyond observed sequences. The model convolves learned amino acid properties to predict expression levels 44% closer to the experimental variance compared to a non-embedded control. Analysis of learned amino acid embeddings highlights the uniqueness of cysteine, the importance of hydrophobicity and charge, and the unimportance of aromaticity, when aiming to improve the developability of small proteins. We identify clusters of similar sequences with increased recombinant expression through nonlinear dimensionality reduction and we explore the inferred expression landscape via nested sampling. The analysis enables the first direct visualization of the fitness landscape and highlights the existence of evolutionary bottlenecks in sequence space giving rise to competing subpopulations of sequences with different developability. The work advances applied protein engineering efforts by predicting and interpreting protein scaffold expression from a limited dataset. Furthermore, our statistical mechanical treatment of the problem advances foundational efforts to characterize the structure of the protein fitness landscape and the amino acid characteristics that influence protein developability.

Molecular evolution of peptides by yeast surface display technology

Genetically encoded peptides possess unique properties, such as a small molecular weight and ease of synthesis and modification, that make them suitable to a large variety of applications. However, despite these favorable qualities, naturally occurring peptides are often limited by intrinsic weak binding affinities, poor selectivity and low stability that ultimately restrain their final use. To overcome these limitations, a large variety of in vitro display methodologies have been developed over the past few decades to evolve genetically encoded peptide molecules with superior properties. Phage display, mRNA display, ribosome display, bacteria display, and yeast display are among the most commonly used methods to engineer peptides. While most of these in vitro methodologies have already been described in detail elsewhere, this review describes solely the yeast surface display technology and its valuable use for the evolution of a wide range of peptide formats.

Targeting HER2-positive gastric cancer with a novel 18F-labeled ZHER2:342 probe

To realize the diagnosis of HER2-positive gastric cancer via PET imaging, herein, a new kind of 18F-labeled HER2 affibody probe was created; the bifunctional maleimide derivative 1,4,7-triazacyclononane-1,4,7-triacetic acid (NOTA-MAL) was first coupled to a polypeptide, and the resulting compound was subsequently labeled with the 18FAl complex. The binding characteristics of the probe were assessed using both in vitro studies and in vivo microPET imaging and biodistribution experiments. Immunohistochemical staining was performed to confirm the expression level of HER2 in the studied cell lines and tumors. The probe was successfully produced with the radiochemical purity of more than 95%. The NCI N87 cell-associated radioactivity was 19.31 ± 1.01% AD, and it decreased to 0.83 ± 0.04% AD per 106 cells after blocking HER2 as early as 15 minutes post-incubation (p < 0.05). A competition binding assay between radiolabeled and non-radioactive affibody molecules with NCI N87 indicated that the IC50 was 8.10 nM. The microPET imaging and biodistribution of human gastric cancer xenografts demonstrated that the probe could specifically accumulate in tumors at early time points. Protein detection confirmed a strong HER2 expression in NCIN87 and a weak HER2 expression in SGC7901. In conclusion, 18FAl-NOTA-MAL-Cys-GGGRDN(M0)-ZHER2:342 was successfully prepared via a one-step method. The favorable preclinical data showed specific and effective tumor targeting capacity of the proposed probe; this revealed that the probe proposed herein might have potential application in gastric cancer imaging.

Engineering an EGFR-binding Gp2 domain for increased hydrophilicity

The Gp2 domain is a 45 amino-acid scaffold that has been evolved for specific, high-affinity binding towards multiple targets and was proven useful in molecular imaging and biological antagonism. It was hypothesized that Gp2 may benefit from increased hydrophilicity for improved physiological distribution as well as for physicochemical robustness. We identified seven exposed hydrophobic sites for hydrophilic mutations and experimentally evaluated single mutants, which yielded six mutations that do not substantially hinder expression, binding affinity or specificity (to epidermal growth factor receptor), and thermal stability. Eight combinations of these mutations improved hydrophilicity relative to the parental Gp2 clone as assessed by reverse-phase high-performance liquid chromatography (p < 0.05). Secondary structures and refolding abilities of the selected single mutants and all multimutants were unchanged relative to the parental ligand. A variant with five hydrophobic-to-hydrophilic mutations was identified with enhanced solubility as well as reasonable binding affinity ( K d  = 53-63 nM), recombinant yield (1.3 ± 0.8 mg/L), and thermal stability ( T m  = 53 ± 3°C). An alternative variant with a cluster of three leucine-to-hydrophilic mutations was identified with increased solubility, nominally increased binding affinity ( K d  = 13-28 nM) and reasonable thermal stability ( T m  = 54.0 ± 0.6°C) but reduced yield (0.4 ± 0.3 mg/L). In addition, a ≥7°C increase in the midpoint of thermal denaturation was observed in one of the single mutants (T21N). These mutants highlight the physicochemical tradeoffs associated with hydrophobic-to-hydrophilic mutation within a small protein, improve the solubility and hydrophilicity of an existent molecular imaging probe, and provide a more hydrophilic starting point for discovery of new Gp2 ligands towards additional targets.

Constrained Combinatorial Libraries of Gp2 Proteins Enhance Discovery of PD-L1 Binders

Engineered protein ligands are used for molecular therapy, diagnostics, and industrial biotechnology. The Gp2 domain is a 45-amino acid scaffold that has been evolved for specific, high-affinity binding to multiple targets by diversification of two solvent-exposed loops. Inspired by sitewise enrichment of select amino acids, including cysteine pairs, in earlier Gp2 discovery campaigns, we hypothesized that the breadth and efficiency of de novo Gp2 discovery will be aided by sitewise amino acid constraint within combinatorial library design. We systematically constructed eight libraries and comparatively evaluated their efficacy for binder discovery via yeast display against a panel of targets. Conservation of a cysteine pair at the termini of the first diversified paratope loop increased binder discovery 16-fold ( p < 0.001). Yet two other libraries with conserved cysteine pairs, within the second loop or an interloop pair, did not aid discovery thereby indicating site-specific impact. Via a yeast display protease resistance assay, Gp2 variants from the loop one cysteine pair library were 3.3 ± 2.1-fold ( p = 0.005) more stable than nonconstrained variants. Sitewise constraint of noncysteine residues-guided by previously evolved binders, natural Gp2 homology, computed stability, and structural analysis-did not aid discovery. A panel of binders to programmed death ligand 1 (PD-L1), a key target in cancer immunotherapy, were discovered from the loop 1 cysteine constraint library. Affinity maturation via loop walking resulted in strong, specific cellular PD-L1 affinity ( K_d = 6-9 nM).

Engineered Charge Redistribution of Gp2 Proteins through Guided Diversity for Improved PET Imaging of Epidermal Growth Factor Receptor

The Gp2 domain is a protein scaffold for synthetic ligand engineering. However, the native protein function results in a heterogeneous distribution of charge on the conserved surface, which may hinder further development and utility. We aim to modulate charge, without diminishing function, which is challenging in small proteins where each mutation is a significant fraction of protein structure. We constructed rationally guided combinatorial libraries with charge-neutralizing or charge-flipping mutations and sorted them, via yeast display and flow cytometry, for stability and target binding. Deep sequencing of functional variants revealed effective mutations both in clone-dependent contexts and broadly across binders to epidermal growth factor receptor (EGFR), insulin receptor, and immunoglobulin G. Functional mutants averaged 4.3 charge neutralizing mutations per domain while maintaining net negative charge. We evolved an EGFR-targeted Gp2 mutant that reduced charge density by 33%, maintained net charge, and improved charge distribution homogeneity while elevating thermal stability ( Tm = 87 ± 1 °C), improving binding specificity, and maintaining affinity ( Kd = 8.8 ± 0.6 nM). This molecule was conjugated with 1,4,7-triazacyclononane,1-glutaric acid-4,7-acetic acid for 64Cu chelation and evaluated for physiological distribution in mice with xenografted A431 (EGFRhigh) and MDA-MB-435 (EGFRlow) tumors. Excised tissue gamma counting and positron emission tomography/computed tomography imaging revealed good EGFRhigh tumor signal (4.7 ± 0.5%ID/g) at 2 h post-injection and molecular specificity evidenced by low uptake in EGFRlow tumors (0.6 ± 0.1%ID/g, significantly lower than for non-charge-modified Gp2, p = 0.01). These results provide charge mutations for an improved Gp2 framework, validate an effective approach to charge engineering, and advance performance of physiological EGFR targeting for molecular imaging.

Fn3 proteins engineered to recognize tumor biomarker mesothelin internalize upon binding

Mesothelin is a cell surface protein that is overexpressed in numerous cancers, including breast, ovarian, lung, liver, and pancreatic tumors. Aberrant expression of mesothelin has been shown to promote tumor progression and metastasis through interaction with established tumor biomarker CA125. Therefore, molecules that specifically bind to mesothelin have potential therapeutic and diagnostic applications. However, no mesothelin-targeting molecules are currently approved for routine clinical use. While antibodies that target mesothelin are in development, some clinical applications may require a targeting molecule with an alternative protein fold. For example, non-antibody proteins are more suitable for molecular imaging and may facilitate diverse chemical conjugation strategies to create drug delivery complexes. In this work, we engineered variants of the fibronectin type III domain (Fn3) non-antibody protein scaffold to bind to mesothelin with high affinity, using directed evolution and yeast surface display. Lead engineered Fn3 variants were solubly produced and purified from bacterial culture at high yield. Upon specific binding to mesothelin on human cancer cell lines, the engineered Fn3 proteins internalized and co-localized to early endosomes. To our knowledge, this is the first report of non-antibody proteins engineered to bind mesothelin. The results validate that non-antibody proteins can be engineered to bind to tumor biomarker mesothelin, and encourage the continued development of engineered variants for applications such as targeted diagnostics and therapeutics.

Influence of composition of cysteine-containing peptide-based chelators on biodistribution of 99mTc-labeled anti-EGFR affibody molecules

Epidermal growth factor receptor (EGFR) is overexpressed in a number of cancers and is the molecular target for several anti-cancer therapeutics. Radionuclide molecular imaging of EGFR expression should enable personalization of anti-cancer treatment. Affibody molecule is a promising type of high-affinity imaging probes based on a non-immunoglobulin scaffold. A series of derivatives of the anti-EGFR affibody molecule ZEGFR:2377, having peptide-based cysteine-containing chelators for conjugation of ^99mTc, was designed and evaluated. It was found that glutamate-containing chelators Gly-Gly-Glu-Cys (GGEC), Gly-Glu-Glu-Cys (GEEC) and Glu-Glu-Glu-Cys (EEEC) provide the best labeling stability. The glutamate containing conjugates bound to EGFR-expressing cells specifically and with high affinity. Specific targeting of EGFR-expressing xenografts in mice was demonstrated. The number of glutamate residues in the chelator had strong influence on biodistribution of radiolabeled affibody molecules. Increase of glutamate content was associated with lower uptake in normal tissues. The ^99mTc-labeled variant containing the EEEC chelator provided the highest tumor-to-organ ratios. In conclusion, optimizing the composition of peptide-based chelators enhances contrast of imaging of EGFR-expression using affibody molecules.

Evaluation of affibody charge modification identified by synthetic consensus design in molecular PET imaging of epidermal growth factor receptor

Tumor overexpression of epidermal growth factor receptor (EGFR) correlates to therapeutic response in select patient populations. Thus, molecular positron emission tomography (PET) imaging of EGFR could stratify responders versus non-responders. We previously demonstrated effectiveness of a "synthetic consensus" design principle to identify six neutralizing mutations within a 58-amino acid EGFR-targeted affibody domain. Herein, we extend the approach to identify additional neutralized variants that vary net charge from -2 to either -4 or +4 while retaining high affinity (1.6 ± 1.2 nM and 2.5 ± 0.7 nM), specific binding to EGFR, secondary structure, and stability (T_m = 68 °C and 59 °C). We radiolabeled the resultant collection of five charge variants with ⁶⁴Cu and evaluated PET imaging performance in murine models with subcutaneously xenografted EGFR^high and EGFR^low tumors. All variants exhibited good EGFR^high tumor imaging as early as 1 h, with EA35S (+3/-5) achieving 7.7 ± 1.4 %ID/g tumor at 4 h with 1.5 ± 0.3%ID/g EGFR^low tumor, 34 ± 5 tumor:muscle and 12 ± 3 tumor:blood ratios. The positively charged EA62S mutant (+6/-2) exhibited 2.2-3.3-fold higher liver signal than the other variants (p<0.01). The EA68 variant with higher charge density was more stable to human and mouse serum than neutralized variants. In a comparison of radiometal chelators, 1,4,7-triazacyclononane,1-glutaric acid-4,7-acetic acid (NODAGA) exhibited superior physiological specificity to 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA). In total, these studies comparatively evaluated a set of EGFR-targeted affibodies varying in net charge and charge density, which revealed functional variations that are useful in engineering an ideal probe for translational studies.

Beyond antibody engineering: directed evolution of alternative binding scaffolds and enzymes using yeast surface display

Pioneered exactly 20 years ago, yeast surface display (YSD) continues to take a major role in protein engineering among the high-throughput display methodologies that have been developed to date. The classical yeast display technology relies on tethering an engineered protein to the cell wall by genetic fusion to one subunit of a dimeric yeast-mating agglutination receptor complex. This method enables an efficient genotype-phenotype linkage while exploiting the benefits of a eukaryotic expression machinery. Over the past two decades, a plethora of protein engineering efforts encompassing conventional antibody Fab and scFv fragments have been reported. In this review, we will focus on the versatility of YSD beyond conventional antibody engineering and, instead, place the focus on alternative scaffold proteins and enzymes which have successfully been tailored for purpose with regard to improving binding, activity or specificity.

Targeting Insulin Receptor in Breast Cancer Using Small Engineered Protein Scaffolds

Insulin receptor (InsR) and the type I insulin-like growth factor (IGF1R) are homologous receptors necessary for signal transduction by their cognate ligands insulin, IGF-I and IGF-II. IGF1R mAbs, intended to inhibit malignant phenotypic signaling, failed to show benefit in patients with endocrine-resistant tumors in phase III clinical trials. Our previous work showed that in tamoxifen-resistant cells, IGF1R expression was lacking, but InsR inhibition effectively blocked growth. In endocrine-sensitive breast cancer cells, insulin was not growth stimulatory, likely due to the presence of hybrid InsR/IGF1R, which has high affinity for IGF-I, but not insulin. Combination inhibition of InsR and IGF1R showed complete suppression of the system in endocrine-sensitive breast cancer cells. To develop InsR-binding agents, we employed a small protein scaffold, T7 phage gene 2 protein (Gp2) with the long-term goal of creating effective InsR inhibitors and diagnostics. Using yeast display and directed evolution, we identified three Gp2 variants (Gp2 #1, #5, and #10) with low nanomolar affinity and specific binding to cell surface InsR. These Gp2 variants inhibited insulin-mediated monolayer proliferation in both endocrine-sensitive and resistant breast cancer, but did not downregulate InsR expression. Gp2 #5 and Gp2 #10 disrupted InsR function by inhibiting ligand-induced receptor activation. In contrast, Gp2 #1 did not block InsR phosphorylation. Notably, Gp2 #1 binding was enhanced by pretreatment of cells with insulin, suggesting a unique receptor-ligand-binding mode. These Gp2 variants are the first nonimmunoglobulin protein scaffolds to target insulin receptor and present compelling opportunity for modulation of InsR signaling. Mol Cancer Ther; 16(7); 1324-34. ©2017 AACR.