In Vivo Imaging of Xenograft Tumors Using an Epidermal Growth Factor Receptor-Specific Affibody Molecule Labeled with a Near-infrared Fluorophore¹

Chemicals and Reagents

The Affibody molecules were provided by Affibody AB (Bromma, Sweden). The Eaff molecule is in a head-to-tail dimeric form with a molecular weight of 13.7 kDa. The HER2-specific Affibody (Haff) molecule is a fusion protein of the HER2 monomer with the albumin-binding domain (ABD) and has a molecular weight of 12.1 kDa. IRDye800CW maleimide dye (Ex/Em: 774/789 nm) and IRDye800CW-labeled EGF (EGF800) were from LI-COR Biosciences (Lincoln, NE). TO-PRO-3 and SYTOX Green nucleic acid stain reagents were purchased from Invitrogen (Carlsbad, CA). DY-682 maleimide dye, Recombinant human EGF and TCEP·HCl was purchased from Dyomics (Jena, Germany), R&D Systems (Minneapolis, MN), and Fisher Scientific (Pittsburgh, PA), respectively.

Conjugation of Affibody Molecules with NIR Dyes

The Affibody used in this study contains a unique C-terminal cysteine residue for thiol-reactive maleimide dye labeling. To reduce the spontaneously formed disulfide bond between the cysteine residues, the Affibody molecules were incubated with 5 mM TCEP·HCl. The excess TCEP·HCl was removed by passing the reaction mixture through a Zeba Spin Desalting Column (Fisher Scientific). The maleimide dye was reconstituted in DMSO to a concentration of 10 mM and added to the Affibody solution. The reaction mixture was incubated at room temperature for 3 hours before passing through another Zeba Spin Desalting Column to remove free dye. The concentrations of labeled Affibody molecules and the labeling efficiency were calculated based on the absorbance at 280 nm and 778 nm (IRDye800CW) or 688 nm (DY-682), respectively.

Cell Culture

The human skin epidermoid carcinoma cell line A431, ovarian adenocarcinoma cell line SK-OV-3 (SKOV3), and breast adenocarcinoma cell lines MDA-MB-231 (MDA231) and SK-BR-3 (SKBR3) were obtained from the American Type Culture Collection (ATCC, Manassas, VA). A431 and MDA231 cells were maintained in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal calf serum and 1% penicillin-streptomycin (complete DMEM). SKOV3 and SKBR3 cells were maintained in McCoy 5A medium (McCoy) supplemented with 10% fetal calf serum and 1% penicillin-streptomycin (complete McCoy).

To collect cell lysates for Western blot analysis, cells were rinsed with PBS and lysed with radioimmunoprecipitation assay buffer. When treatment with Eaff or EGF was necessary, the agents were diluted in serum-free medium to the desired concentrations and added to the cells. After incubation at 37°C for 2 hours, cells were rinsed with PBS and lysed with radioimmunoprecipitation assay buffer.

Western Blot

The Western blot analysis was performed as previously described with modifications [35]. In brief, protein samples were denatured by boiling for 5 minutes and loaded onto 10% Bis-Tris gels. After electrophoresis, proteins were transferred to nitrocellulose membranes. Membranes were blocked for 2 hours with Odyssey Blocking Buffer, then incubated for 1 hour with primary antibodies diluted in Odyssey Blocking Buffer. The mouse EGFR antibody, mouse P44/42 mitogen-activated protein kinase (MAPK) (ERK1/2) antibody, rabbit phospho-EGFR (Tyr1045) antibody, rabbit phospho-P44/42 MAPK (ERK1/2) antibody, and rabbit HER2/ErbB2 antibody were purchased from Cell Signaling Technology (Danvers, MA). The mouse pan-Actin antibody was purchased from NeoMarkers (Fremont, CA). Two primary antibodies from different species (such as mouse anti-P44/42 and rabbit anti-phospho-P44/42) were added together.

After incubating with primary antibodies, the membranes were washed with PBS containing 0.1% Tween 20 (PBST) three times. Then the membranes were incubated for 1 hour with IRDye800CW-conjugated goat antirabbit IgG and IRDye680-conjugated goat antimouse IgG secondary antibodies (LI-COR Biosciences) diluted in Odyssey Blocking Buffer. The blots were then washed three times with PBST and rinsed with PBS. Proteins were visualized by scanning the membrane on an Odyssey Infrared Imaging System (LI-COR Biosciences) with both 700- and 800-nm channels.

Cell Binding and Uptake Assay

Cells were seeded at approximately 3 × 10⁴ (A431), 2 × 10⁴ (MDA231 and SKBR3), or 1 × 10⁴ (SKOV3) cells per well in 96-well plates and cultured overnight before the assay. The cell density was approximately 70% to 80% confluent at the time of assay. The NIR fluorophore-labeled targeting agents were diluted in complete cell culture medium to designated concentrations and incubated in the 37°C incubator for 2 hours except where otherwise stated. For competition study, unlabeled agents were diluted in complete medium and incubated with the cells for 1 hour. The competitors were then removed before adding targeting agents. After incubating with targeting agents, cells were fixed with 3.7% formaldehyde/PBS and washed with PBST. The cells were then incubated in TO-PRO-3 stain agent (1:5000 in PBS) to normalize for cell numbers. After three additional washes with PBST, the plate was scanned and signal intensity quantified [36].

Microscopic Analysis

Cells were seeded in eight-well chamber slides (Nalge Nunc International Corp, Naperville, IL) and cultured overnight. The cells were incubated with 20 nM targeting agent at 37°C for 2 hours, fixed, and washed as previously described. Instead of TO-PRO-3, SYTOX Green nucleic acid stain agent was used to visualize the nuclei by microscopy. After the final wash, the slides were mounted with Fluoromount reagent (Sigma, St Louis, MO). The images were acquired using a Leica DM6000b Bio/Med microscope system (Leica Microsystems, Inc, Bannockburn, IL) equipped with a xenon lamp. The excitation and emission filters used for NIR imaging are HQ760/40 and HQ830/50 (Chroma Technology Corp, Rockingham, VT), respectively. The images were deconvolved using the accompanying software.

Xenograft Mouse Model

All animals were cared for and maintained under the supervision and guidelines of the University of Nebraska-Lincoln Institutional Animal Care and Use Committee. Mice were maintained on a purified maintenance diet (AIN-93M) from Harlan Teklad (Madison, WI). The xenograft tumors were established as previously described with modifications [37]. In brief, athymic nude (nu/nu) mice, obtained from Charles River Laboratories, Inc (Cambridge, MA) at 4 weeks of age, were subcutaneously injected with suspension cells in 0.1 ml of serum-free medium. The numbers of cells injected was 10⁶ or 5 × 10⁶ for A431 or SKOV3, respectively. To establish both A431 and SKOV3 tumors on the same mouse, SKOV3 cells were inoculated 1 week before inoculating A431 cells to achieve similar tumor growth.

In Vivo Animal Imaging

Mice were anesthetized with 2% isoflurane throughout the procedures. For imaging experiments, the contrast agent was diluted in 100 µl of PBS and injected through the tail vein. The images were acquired at the indicated time points with a Pearl Imager (LI-COR Biosciences). The Ex/Em settings for the 700- and 800-nm channels were 685/720 and 785/820 nm, respectively. A cooled charge-coupled device camera was used to collect the images. The images were analyzed using the accompanying software. Regions of interest (ROIs) for both tumor and contralateral areas (background) were selected from equivalent-sized areas containing the same number of pixels. ROIs were quantified for mean pixel values. TBR was derived by dividing mean tumor signal by mean background signal [28].

Organ and Tissue Analysis

Mice were killed 1 day after imaging agent injection and dissected to collect the organs. The excised organs were rinsed in PBS, and imaged using a Pearl Imager. The organs were then snap-frozen in OCT compound for cryosectioning. Sections (8-µm thickness) were scanned on an Odyssey Infrared Imaging System for the targeted fluorescence signal. To compare the fluorescence signal on tissue sections, ROIs were drawn on the tissue sections and blank regions of the slides. The signal intensity was calculated as the average signal intensity on the tissue region minus the average signal intensity on the blank region.

The Effect of Eaff on EGFR and ERK1/2 Phosphorylation

It is known that activation of the EGFR tyrosine kinase activity leads to the phosphorylation of a variety of target proteins, including ERK1/2 and EGFR itself [38]. To evaluate whether Eaff stimulates EGFR-mediated signaling pathways, A431 cells were treated with 5 or 20 nM Eaff. In contrast to EGF, which stimulated the phosphorylation of both proteins in a dose-dependent manner, Eaff treatment did not change the phosphorylation level of EGFR and ERK1/2 (p44/p42 MAPK). However, a high concentration of Eaff (100 nM), when applied together with 5 nM EGF, compromised the stimulatory effect of EGF on EGFR and ERK1/2 phosphorylation (Figure 1).

Specific Cellular Binding and Uptake of Labeled Affibody Molecules

Because the Affibody monomer molecule contains five to six lysine residues [19,39], conjugation of dye molecules to lysine may result in heterogeneous multiple labeling. It is also possible to disturb the binding motif by attaching a dye molecule to its internal lysine residues. To avoid this problem, the Affibody molecules were conjugated with the maleimide dye to its C-terminal cysteine. The incorporation of dye into the Affibody molecules was monitored by gel electrophoresis (Figure W1). The NIR fluorophore-labeled Affibody molecules were designated as Eaff800 (Eaff labeled with IRDye800CW), Eaff682 (Eaff labeled with DY-682), Haff800 (Haff labeled with IRDye800CW), and Haff682 (Haff labeled with DY-682), respectively.

To compare the protein expression levels of EGFR and HER2 in various cell lines, cell lysates of MDA231, A431, SKOV3, and SKBR3 cells were analyzed by Western blot. Both EGFR and HER2 proteins were detected in these cell lines. However, the EGFR level was much higher in A431 cells than in any other cell lines, whereas HER2 was highly expressed in SKOV3 and SKBR3 cells (Figure 2A). This finding was consistent with previous reports on EGFR and HER2 expression [40]. These cell lines were also used to investigate the binding specificity of EGFR and HER2-specific Affibody molecules. The binding and uptake assay was performed by incubating 5 nM Eaff800 or Haff800 with MDA231, A431, SKOV3, and SKBR3 cells. Results demonstrated that whereas A431 cells contained high Eaff800 signal, there was only minimal Eaff800 binding and uptake for MDA231, SKOV3, or SKBR3 cells. On the contrary, the HER2-specific Haff800 showed the strongest signal in SKOV3 and SKBR3 cells, and the signal in MDA231 and A431 cells was low (Figure 2B). On the basis of these results, the A431 cells and SKOV3 cells were selected as EGFR- and HER2-overexpressing cells, respectively, for subsequent studies.

The binding and uptake of Eaff800 by A431 cells were dependent on Eaff800 concentration. With the increase of Eaff800 concentration, the signal intensity also increased (Figure 2C), with the maximum signal obtained at approximately 20 to 50 nM. As a comparison, the same concentrations of Haff800 and free IRDye800CW dye were incubated with A431 cells at the same condition. The signals from these agents were minimal, even at a concentration as high as 100 nM (Figure 2C).

Comparison of Eaff800 with EGF800

NIR fluorophore-labeled EGF has been used as an imaging agent [13,14]. So it was of interest to compare Eaff and EGF labeled with the same NIR fluorophore. Both Eaff800 (5 nM) and EGF800 (5 nM) were diluted in complete DMEM and incubated with A431 cells for different periods. As shown in Figure 3A, the binding and uptake of EGF800 was faster than that of Eaff800, and the signal intensity of EGF800 was higher in the early stage. Before 1 hour, the binding and uptake of both agents increased over time. At the 1-hour time point, binding and uptake of EGF800 decreased, whereas the Eaff800 signal continued to increase. The Eaff800 signal reached its highest level at 4 to 6 hours and declined thereafter (Figure 3A).

Competition analysis was performed by incubating different concentrations of unlabeled Eaff or EGF with A431 cells before adding targeting agents. Both Eaff and EGF blocked the binding and uptake of Eaff800. The blocking effect of Eaff and EGF was observed at a concentration as low as 5 nM, which was the same concentration as Eaff800 used in this assay. With the increase of competitor concentrations, the Eaff800 signal declined. However, the competition effect of Eaff was more prominent than that of EGF, especially at concentrations higher than 20 nM (Figure 3B). It was also noted that both Eaff and EGF blocked the binding and uptake of EGF800 by A431 cells (data not shown).

The localization of Eaff800 and EGF800 in A431 cells were examined by microscopy. For both Eaff800 and EGF800, intense signals were associated with the cell membrane. However, foci probably representative of internalized agents were also observed inside the cells (Figure 3C).

Targeting EGFR-Overexpressing Xenograft Tumors by Eaff800

Eaff800 probe was tested in mouse xenograft tumor models by in vivo imaging. Nude mice bearing A431 xenograft tumors were injected with 0.5 nmol of Eaff800 through the tail vain. Whole mouse images were acquired at different time points after agent administration. Figure 4A showed a series of these images (dorsal view). Intense signal was found in the region where the kidney is located. The belly view of the mouse also revealed high signal intensity at the liver region (data not shown). The accumulation of Eaff800 signal in the liver and kidney was confirmed by imaging dissected organs, as discussed later. The tumor could be identified at 1 hour after probe injection and it became most prominent after 1 day (approximately 24 hours). Quantification of the ROIs of equivalent-sized areas from the tumor and contralateral region showed that the signal intensities of both tumor and normal tissue peaked at 20 minutes after agent injection. Thereafter, the signal decreased over time (Figure 4B). The signal decline was more dramatic in normal tissues, especially during the period from 6 to 48 hours, resulting in an increased TBR, which reached the highest level at 24 hours (Figure 4C). At 96 hours, most of the agent was cleared out of the body. The residual signals in the tumor and normal tissue were 8.7 ± 0.5% and 5.7 ± 1.0% of those at the highest levels (20 minutes after agent injection), respectively.

To investigate the accumulation of imaging agent in different organs, mice were killed 1 day after agent injection. The organs were collected and imaged. Consistent with in vivo imaging observations, signal intensities were high in the liver, kidney, and tumor (Figure 5A). These organs were then snap-frozen in OCT compound, sectioned at 8-µm thickness, and scanned. Quantification of tissue section revealed the strongest signal in the liver, followed by the kidney and tumor. All other tissues contained only a low level of signal (Figure 5, B and C). Interestingly, unlike the even distribution of Eaff800 in the liver, a higher Eaff800 signal was located in the renal cortex of the kidney compared with other regions (Figure 5B).

Two-color In Vivo Imaging Using Eaff800 and Haff682

Haff conjugated with Alexa Fluor 750 was successfully used for optical imaging [18]. Before animal injection, Eaff800 and Haff682 were added together to A431 cells or SKOV3 cells to examine whether the Affibody molecules interfere with each other's binding. Compared with the binding and uptake of individual probes, no significant difference was observed when two probes were added together (data not shown), indicating the lack of cross reactivity.

To test the in vivo specificity of labeled Affibody molecules, 100 µl of imaging agent solution containing 0.5 nmol of Eaff800 and 0.5 nmol of Haff682 was injected into nude mice bearing both A431 tumors and SKOV3 tumors through the tail vain. Multiple-color images were acquired 1 day after agent injection. As expected, A431 tumors contained a higher Eaff800 signal (Figure 6A-a), whereas SKOV3 tumors contained a higher Haff682 signal (Figure 6A-b). The merged picture displayed a two-color image representing selective accumulation of two different probes (Figure 6A-d). The A431 tumor on the left side is green, which represented predominant Eaff800 signal, and the SKOV3 tumor on the right side is red, which represented predominant Haff682 signal. The TBRs for A431 tumor and SKOV3 tumor were 1.7 ± 0.2 and 2.1 ± 0.3, respectively. Because pseudocolored images can distinguish different intensities easier, the mouse images with Eaff800 signal or Haff682 signal were also presented in pseudo color (Figure W2). Images of tumor sections also demonstrated that Eaff800 and Haff682 accumulated preferably in A431 tumors and SKOV3 tumors, respectively (Figure 6B).

Next, we exchanged the labeling, that is, Eaff was labeled with DY-682 (Eaff682), and Haff was labeled with IRDye800CW(Haff800). As expected, Eaff682 accumulated in A431 tumor and Haff800 accumulated in SKOV3 tumor predominantly (Figure W3).

Whereas both probes specifically targeted their respective tumors, the organ distribution pattern and the clearance rate of Eaff800 and Haff682 were different. Whole body images showed predominant Eaff800 signal (green) in the liver region (Figure 6A-d). The accumulation of Eaff800 in the liver was confirmed by images of dissected organs (Figure W4A). Analysis of tissue sections revealed that the liver-to-kidney ratio of Eaff800 signal was approximately 3.8 (Figure W4, B and C), which is similar to the result when Eaff800 alone was injected (liver-to-kidney ratio = 3.5). However, the signal intensities of Haff682 in the liver and kidney were comparable (liver-to-kidney ratio = 1.1; Figure W4, B and C). The clearance rate of Eaff800 was faster than that of Haff682. At 1 day after probe administration, the signal intensity of Eaff800 as measured by whole body imaging decreased to 11.1 ± 0.6% of the highest level (20 minutes after agent injection), whereas residual signal of Haff682 still remained 57.9 ± 1.2% of the highest level.

When Eaff682 and Haff800 were used, more Eaff682 accumulated in the liver (Figure W5), revealing that the accumulation of Eaff in the liver was independent of the fluorophore conjugated.