Targeting of Epithelial Cell Adhesion Molecule-Expressing Malignant Tumors Using an Albumin-Binding Domain-Fused Designed Ankyrin Repeat Protein: Effect of the Molecular Architecture

Abstract

Designed ankyrin repeat protein (DARPin) Ec1, a small scaffold protein (18 kDa), binds with high affinity the epithelial cell adhesion molecule (EpCAM) that is overexpressed in several carcinomas. To enhance the targeted delivery of cytotoxic drugs using Ec1, we investigated the potential of fusing Ec1 with an albumin-binding domain (ABD) to improve its circulation time and decrease renal uptake. Two fusion proteins were created, Ec1-ABD, with the ABD at the C-terminus, and ABD-Ec1, with the ABD at the N-terminus. Both variants were labeled with ¹¹¹In. ABD-fused variants bound specifically to EpCAM-expressing cells with picomolar affinity. Adding human albumin reduced the affinity. This effect was more pronounced for Ec1-ABD; however, the affinity remained in the subnanomolar range. The position of the ABD did not influence the internalization rate of both variants by human cancer cells. In mouse models with human cancer xenografts, both variants demonstrated over 10-fold lower renal uptake compared to the Ec1. Tumor uptake of the ABD-fused variants was higher than the uptake of Ec1. ABD-Ec1 provided two-fold higher tumor uptake, indicating fusion with an ABD as a promising way to modulate the targeting properties of an Ec1-based construct. However, the effect of fusion depends on the order of the domains.

Keywords: albumin-binding domain (ABD); designed ankyrin repeat protein; epithelial cell adhesion molecule (EpCAM); fusion protein; nude mice; xenograft.

Figure 1

Schematic structures of (A) ABD-Ec1, (B) Ec1-ABD, and (C) non-ABD-fused Ec1 (as a control). All three variants contain a DOTA chelator at the C-terminus for labeling with In-111.

Figure 2

In vitro binding specificity of [¹¹¹In]In-Ec1-ABD (A,B) and [¹¹¹In]In-ABD-Ec1 (C,D) on SKOV-3 (A,C) and OVCAR3 (B,D) cells. For the blocking of EpCAM receptors, a 100-fold molar excess of a non-labeled Ec1 was added before adding the radiolabeled conjugate. The data are presented as an average value from three samples ± SD.

Figure 3

Cellular processing of [¹¹¹In]In-Ec1-ABD (A–D) and [¹¹¹In]In-ABD-Ec1 (E–H) by SKOV-3 cells (A,B,E,F) and OVCAR3 cells (C,D,G,H) without addition (A,C,E,G) and with addition (B,D,F,H) of HSA (100 nM). Cells were incubated with 2 nM of radioconjugate. The data are presented as an average (n = 3) and SD.

Figure 4

Blood kinetics of ABD-fused Ec1 variants labeled with ¹¹¹In. An amount of 10 µg/30 kBq in 100 µL of 1% BSA in PBS was injected per mouse. Uptake is expressed as %ID/g and presented as the average value from 4 mice ± SD.

Figure 5

Effect of ABD fusion and construct design on biodistribution of ¹¹¹In-labeled Ec1 derivatives in nude mice bearing SKOV-3 xenografts 48 h after injection. p-values are calculated by one-way ANOVA with Tukey correction for multiple comparisons.

Figure 6

Uptake of (A) [¹¹¹In]In-Ec1-ABD and (B) [¹¹¹In]In-ABD-Ec1 in SKOV-3 (EpCAM-positive) and Ramos (EpCAM-negative) xenografts at 48 h after injection. Data are expressed as %ID/g and represent averages from four mice ± SD. p-value was obtained in an unpaired t-test.

Figure 7

Imaging of distribution of (A) [¹¹¹In]In-Ec1-ABD, (B) [¹¹¹In]In-ABD-Ec1, and (C) [¹¹¹In]In-Ec1 in mice bearing EpCAM-expressing SKOV-3 xenografts at 48 h after injection. An amount of 10 µg of labeled compound (1.5 MBq, 100 µL in 1% BSA in PBS) was injected into the tail vein. The arrow points to the tumor (T).