Near infrared imaging of epidermal growth factor receptor positive xenografts in mice with domain I/II specific antibody fragments

Abstract

Epidermal growth factor receptor (EGFR) is a transmembrane cell surface receptor that is frequently overexpressed and/or mutated in many cancers. Therapies targeting EGFR have poor outcomes due to the lack of reliable diagnostic tests to monitor EGFR. Current in vitro EGFR diagnostic methods are invasive, requiring biopsies, which limits tumor sampling and availability. EGFR molecular imaging provides non-invasive whole-body images capable of detecting primary tumors and metastases, which can be used to diagnose and monitor response to therapy. Methods: We evaluated properties of two anti-EGFR fragments, 8708 and 8709, as molecular-targeted imaging probes. 8708 and 8709 are anti-EGFR antigen binding fragments (Fabs) that recognize domain I/II of EGFR, which is distinct from epitopes recognized by current anti-EGFR therapeutic antibodies. We used complementarity determining region sequences from 8708 and 8709 Fabs to generate an anti-EGFR IgG and (scFv)₂ and scFv-Fc antibody fragments. We expressed, purified, and labeled the IgG and fragments with IRDye800CW and used them to image EGFR-positive and -negative xenografts in CD-1 nude mice. 8709 scFv-Fc was also tested for competitive binding with the therapeutic anti-EGFR antibody nimotuzumab and for quantifying ratios of EGFR and EGFRvIII deletion mutant. Results: IRDye800CW-labeled 8708 (scFv)₂ and 8709 scFv-Fc imaging probes showed high levels of accumulation and good retention in EGFR-positive xenografts, with peak accumulation occurring at 24 and 48 hours post injection, respectively. IRDye680RD-labeled 8709 scFv-Fc did not compete with IRDye800CW-labeled nimotuzumab for EGFR binding as assayed by flow cytometry using an EGFR-positive cell line. IRDye680RD-labeled 8709 scFv-Fc and IRDye800CW-labeled nimotuzumab used in combination were able to determine the ratio of cells expressing EGFR and a deletion mutant EGFRvIII. Conclusion: IRDye800CW-labeled 8708 (scFv)₂ and 8709 scFv-Fc had desirable binding affinities, clearance times, and tumor accumulation to be used for imaging in combination with current EGFR targeted therapies. This study highlights the potential for using 8708 (scFv)₂ and 8709 scFv-Fc as EGFR diagnostic and therapy monitoring tools.

Keywords: EGFR; IRDye800CW; antibody fragments; near infrared fluorescence imaging.

Figure 1

Antibody and fragments. The IgG and scFv-Fc are divalent and are 150 kDa and 110 kDa. The Fab and (scFv)₂ are divalent and are both about 55 kDa.

Figure 2

Protein expression yield of 8708 and 8709 fragments. Proteins were expressed using a mammalian system (scFv-Fc and IgG) or a bacterial expression system (Fab and (scFv)₂). Average amount of protein obtained is shown in mg/L. There was no expression (NE) for the 8708 IgG.

Figure 3

8708 and 8709 fragment flow cytometry. 8708 and 8709 fragments were titrated with A-431 or MDA-MB-435 cells to obtain saturation binding curves. A) Representative partial overlaid histogram of 8708 and 8709 Fab titration binding to A-431 cells. The concentration of Fabs shown on y-axis is in nanomolar (nM). B) Flow cytometry saturation binding curves of unlabeled and IRDye800CW-labeled 8708 Fab, (scFv)₂, scFv-Fc, and 8709 Fab, (scFv)₂, scFv-Fc, and IgG binding to A-431 and MDA-MB-435 cells showing the percent of cells bound ± SEM.

Figure 4

Near infrared images of IRDye800CW-labeled 8708 and 8709 fragments in A-431 xenograft bearing mice in the dorsal position. Images shown were taken at 6, 24, 48, and 72 hours post injection (hpi) for the Fab and (scFv)₂ with the addition of 168 hpi for the scFv-Fc and IgG. Near infrared images of the 8708 and 8709: A) Fab, B) (scFv) ₂, and C) scFv-Fc and 8709 IgG. Scale bars are shown on the right. Images of at least three different mice were taken. X = xenograft; K = kidney.

Figure 5

Near infrared image analysis of the A-431 xenograft. IRDye800CW-labeled fragments: A) Fab, B) (scFv)₂, and C) scFv-Fc and IgG. Imaging probes were injected into A-431 xenograft tumor bearing mice, imaged, and analyzed over time (hours post injection (hpi)). Data represents the mean of the normalized signal or TBR (tumor-to-background ratio) of the three different regions of interest in the xenograft at least three different mice ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001 and **** p < 0.0001.

Figure 6

Near infrared imaging of 8708 (scFv)₂ and 8709 scFv-Fc in MDA-MB-435 EGFR-negative control cell line. Representative images from A) 8708 (scFv)₂ and B) 8709 scFv-Fc imaging probes injected into MDA-MB-435 xenograft bearing mice. Images shown were taken at 6, 24, 48, and 72 hours post injection (hpi). A minimum of three mice were imaged for each probe. X = xenograft; K = kidney.

Figure 7

Near infrared image analysis of 8708 (scFv)₂ and 8709 scFv-Fc in A-431 and MDA-MB-435 cell lines. Fluorescent signal and TBR analysis of A) 8708 (scFv)₂ and B)8709 scFv-Fc mouse images. Images shown were taken at 6, 24, 48, 72, and 168 hpi. Normalized fluorescent signal is expressed in arbitrary units (au) and TBR (tumor-to-background ratio) in A-431 and MDA-MB-435 tumor bearing mice. A minimum of three mice were analyzed to determine the SEM. * p < 0.05, ** p < 0.01, *** p < 0.001 and **** p < 0.0001.

Figure 8

8709 scFv-Fc as a tool to monitor EGFRvIII mutation status. A) 8709 scFv-Fc saturation binding curve with or without a saturating dose of 8709 scFv-Fc or nimotuzumab. B) Nimotuzumab saturation binding curve with or without a saturating dose of 8709 scFv-Fc or nimotuzumab. C) Analysis of different ratios of wildtype EGFR and mutant EGFRvIII expressing cells using IRDye680RD-8709 scFv-Fc and/or IRDye800CW-nimotuzumab. D) Expected (% of wildtype EGFR) and Observed (% of cells bound by 8709 scFv-Fc to HEK293T cells expressing wildtype EGFR and mutant EGFRvIII. Data from A and B represents mean (n=3). Data from C and D represents mean (n = 8). mt = mutant; wt = wildtype.