Impact of single-chain Fv antibody fragment affinity on nanoparticle targeting of epidermal growth factor receptor-expressing tumor cells

Abstract

To determine the importance of single-chain Fv (scFv) affinity on binding, uptake, and cytotoxicity of tumor-targeting nanoparticles, the affinity of the epidermal growth factor receptor (EGFR) scFv antibody C10 was increased using molecular evolution and yeast display. A library containing scFv mutants was created by error-prone PCR, displayed on the surface of yeast, and higher affinity clones selected by fluorescence activated cell sorting. Ten mutant scFv were identified that had a 3-18-fold improvement in affinity (KD=15-88 nM) for EGFR-expressing A431 tumor cells compared to C10 scFv (KD=264 nM). By combining mutations, higher affinity scFv were generated with KD ranging from 0.9 nM to 10 nM. The highest affinity scFv had a 280-fold higher affinity compared to that of the parental C10 scFv. Immunoliposome nanoparticles (ILs) were prepared using EGFR scFv with a 280-fold range of affinities, and their binding and uptake into EGFR-expressing tumor cells was quantified. At scFv densities greater than 148 scFv/IL, there was no effect of scFv affinity on IL binding and uptake into tumor cells, or on cytotoxicity. At lower scFv densities, there was less uptake and binding for ILs constructed from the very low affinity C10 scFv. The results show the importance of antibody fragment density on nanoparticle uptake, and suggest that engineering ultrahigh affinity scFv may be unnecessary for optimal nanoparticle targeting.

Figure 1. Selection of higher affinity scFv by fluorescent activated cell sorting

Yeast displaying scFv were stained with biotinylated EGFR-ECD at the indicated concentrations. The sort gate was set to capture yeast cells with higher binding affinity (gate P2) and better scFv expression (gate P3). The number inside the P2 and P3 gates indicates the percentage of cells within that gate.

Figure 2. Binding of yeast-displayed scFv to 1μM biotinylated EGFR-ECD

(a) Parental C10 clone. (b) Polyclonal yeast (P2 poly) and monoclonal yeast (P2/1, P2/2, P2/3, P2/4, P2/5) from the P2 gate sorting. (c) Polyclonal yeast (P3 poly) and monoclonal yeast (P3/1, P3/2, P3/3, P3/4, P3/5) from the P3 gate sorting.

Figure 3

Deduced amino acid sequences of C10 scFv and the affinity-matured mutants.

Figure 4. Differential binding of scFv to EGFR positive and negative cell lines as determined by flow cytometry

C10 scFv and C10 mutant scFv stained both EGFR and EGFR vIII positive cells (A431, MDAMB468, and NR6M) but did not stain EGFR negative cells (MDAMB453 and NR6).

Figure 5. Effect of intrinsic antibody affinity on internalization of EGFR-targeted ILs

(a) Internalization of EGFR-targeted ILs with different affinity compared to nontargeted liposomes in EGFR-overexpressing cell MDAMB468 as determined by fluorescent microscopy. (b) Uptake of EGFR ILs into MDAMB468 cells as determined by flow cytometry.

Figure 6. Effect of scFv affinity and scFv surface density on internalization of EGFR-targeted ILs

(a) Uptake of EGFR ILs with different scFv surface densities into MDAMB468 cells as determined by flow cytometry. Each data point represents the mean of three independent measurements. (b) Apparent K_D of ILs with a surface density of 74 scFv/liposome for MDAMB468 cells. (c) Uptake of EGFR ILs with different scFv surface densities into MDAMB231 cells compared to A431 cells as determined by flow cytometry.

Figure 7. Effect of EGF on the binding and uptake of EGFR scFv antibodies and ILs, C10 (◆), P2/4 (■) and 2224 (▲))

(a) Effect of increasing EGF concentration on the binding of EGFR scFv to MDAMB468 cells. (b) Effect of increasing EGF concentration on the binding of EGFR ILs to MDAMB468 cells. (c) Effect of increasing EGF concentration on the uptake of EGFR ILs into MDAMB468 cells. (d) Effect of increasing EGF concentration on the binding of EGFR ILs to U87vIII cells.

Figure 8. Effect of intrinsic antibody affinity on EGFR ILs cytotoxicity

Cytotoxicity of anti-EGFR immunoliposomal topotecan in (a) EGFR-overexpressing MDAMB468 breast carcinoma and (b) U87vIII glioblastoma. Immunoliposomes constructed with the P2/4 (◆) and 2224 (■) mutants were compared to those prepared using the parental C10 scFv (▲), nontargeted liposomal topotecan (×), and free topotecan controls (●). Data indicate mean; bars, ± SD