Affibody Molecules for In vivo Characterization of HER2-Positive Tumors by Near-Infrared Imaging

Increased HER2 receptor expression, associated with resistance to therapy and with a poor prognosis, has been shown in 20% to 30% of breast cancers (1, 2). Therefore, HER2 has been an attractive target for molecular therapies (3). One of the most successful approaches is using HER2-specific antibodies, such as trastuzumab or pertuzumab, to prevent HER2 activation (4). Currently, the molecular therapies aimed at HER2 are often combined with conventional chemotherapy and/or radiotherapy (5). Extensive preclinical and clinical studies are required to optimize such combinations. One of the variables assessed in these studies is the HER2 expression level following different therapeutic interventions (6). Thus, there is a need for molecular probes to monitor in vivo the receptor down-regulation as an immediate response to these therapies because the current ex vivo analyses of tumor tissues obtained by surgery or biopsy from breast cancer patients are not always representative (7, 8). Molecular imaging using these probes could provide an ideal noninvasive, quantitative, repetitive imaging technique (9). Traditional approaches using antibodies or antibody fragments have been proposed to address this problem (10, 11). However, these probes should minimally affect the studied system, not interfere with the binding of the therapeutic agents, and minimally affect the target cells (12).

Recently, several successful tumor imaging studies have been reported using radiolabeled HER2-specific Affibody molecules for imaging of HER2-expressing tumors (13). These very stable and highly water-soluble α-helical proteins are relatively small (8.3 kDa) and can be readily expressed in bacterial systems or produced by peptide synthesis (14). Imaging agents based on those molecules will have a high affinity for HER2 and a smaller size compared with antibodies (20×) or antibody fragments (4×) and therefore should provide better results than the currently tested conjugates (15, 16). Furthermore, binding of the HER2-specific Affibody to an epitope that is not the target for trastuzumab enables monitoring of receptor expression following treatment with trastuzumab (17).

Optical imaging is the modality of choice for preclinical studies. It allows visualization of subcellular structures on a microscopic scale, even in vivo, and the macroscopic distribution of fluorescent labels in small animals (18, 19). However, high tissue autofluorescence and limited tissue penetration preclude the use of visible light for most in vivo imaging applications. Near-infrared (NIR) light solves these problems by reducing fluorescence background and enhancing tissue penetration (20). Over the past several years, there has been an explosion of reports describing successful in vivo NIR fluorescence imaging. Although most of these studies are qualitative, quantitative methods are beginning to emerge (21). Both antibodies and small molecules have been conjugated to NIR fluorophores to create tumor-specific NIR fluorescence contrast agents, although a longer washout time was required for an optimal signal-to-background ratio (22, 23). In this regard, HER2-specific Affibody molecules could be good candidates for NIR optical imaging due to their specific tumor targeting and fast body clearance, shown previously in several radionuclide imaging studies with HER2-positive tumor xenografts (24, 25).

In the present study, three different types of HER2-specific Affibody molecules were site-specifically conjugated at their COOH-terminal cysteine residues with Alexa Fluor dyes and compared with Alexa Fluor–labeled trastuzumab in terms of affinity and specificity to the HER2 receptor in vitro and in vivo. We also compared them using a quantitative NIR fluorescence imaging system to assess the in vivo application of these probes in HER2-positive tumor xenografts. Here, we report for the first time that the albumin-binding domain (ABD)–fused bivalent HER2-specific Affibody can be used for in vivo NIR optical imaging.

Reagents. The HER2-specific Affibody proteins, Affibody monomer Z_HER2:342, and COOH-terminal cysteine added forms of Affibody His₆-Z_HER2:342-Cys (M_r 8,318), dimer His₆-(Z_HER2:477)₂-Cys (M_r 14,862), and ABD added ABD-(Z_HER2:342)₂-Cys (M_r 19,213) were kindly provided by Affibody AB. All other reagents were purchased from the following commercial sources: trastuzumab (Herceptin) was obtained from Genentech, neutral Tris(2-carboxyethyl)phosphine solution from Pierce, and NAP 5 columns (Sephadex G-25 DNA grade) from Amersham Biosciences. Alexa Fluor 488 C₅-maleimide (M_r 721), Alexa Fluor 680 C₂-maleimide (M_r ∼1,000), Alexa Fluor 750 C₅-maleimide (M_r ∼1,350), and SAIVI Alexa Fluor 680 and 750 Antibody/Protein 1 mg labeling kits were purchased from Invitrogen. Centricon YM-3 and YM-30 and Microcon YM-3 and YM-30 were obtained from Millipore. All media and supplements for cell culture were from Invitrogen.

Fluorescence conjugation of the HER2-specific Affibody molecules. The Affibody molecules, His₆-Z_HER2:342-Cys, His₆-(Z_HER2:477)₂-Cys, and ABD-(Z_HER2:342)₂-Cys, contain a unique COOH-terminal cysteine residue that allows for site-specific labeling (12). This cysteine was used to fluorescently label the Affibody molecules with thiol-reactive Alexa Fluor-maleimide dyes. To reduce oxidized cysteines before conjugation, the Affibody molecules were incubated with 5 mmol/L neutral TCEP at pH 7.4 for 20 min at room temperature. Immediately after reduction, Alexa Fluor-maleimide dyes dissolved in DMSO were added and conjugation was completed according to the manufacturer's protocol. Unreacted Alexa Fluor-maleimide dyes were removed by passage through a NAP-5 desalting column. Final traces were removed by concentrating and washing with a Centricon YM3 filter unit. All procedures were done under dimmed light. The molecular weights of the sample proteins were analyzed by SDS-PAGE and the degree of labeling was calculated based on the company's instructions.

Trastuzumab was also conjugated with Alexa Fluor dye according to a protocol from the SAIVI Alexa Fluor 680 and 750 Antibody/Protein 1 mg labeling kit. Briefly, Alexa Fluor NHS esters were incubated with the protein in a basic medium (pH 9.3). Labeled protein (trastuzumab) was isolated and purified by gel filtration. The final dye-to-protein ratio (number of Alexa Fluor molecules coupled to each protein molecule) was determined to be between 2.5 and 3.5 according to a protocol from Invitrogen. All of the fluorescently labeled protein samples were aliquoted and stored in the dark at −70°C.

In vitro binding affinity measurement of the Affibody conjugates using surface plasmon resonance. Surface plasmon resonance experiments were done on a BIACORE T100 instrument (Biacore) by long (33 min) injection. For HER2/FC (R&D Systems; 10 μg/mL in HBS-EP buffer), 1,500 RU protein was captured on the protein A/G modified CM5 sensor chip via an immobilization capture wizard. A concentration series of Z_HER2:342 were injected for 33 min over the HER2/FC capture surface at 5 μL/min (26). To verify the consistency of the kinetic data, we plotted the RU response against the Affibody concentration and fit the data to the four-variable logistics equation in BIA evaluation software version 3.1. The surface plasmon resonance instruments show a number of instrument effects, as well as binding activity, in the recovered signals. These effects are often removed through a double referencing protocol (27). The data were fitted using a simple one-to-one interaction model with nonzero initial conditions, and apparent K_D estimates were recovered.

Cell cultures. The human breast adenocarcinoma SKBR-3, the human breast ductal carcinoma BT-474, and the human ovarian adenocarcinoma SKOV-3 were obtained from the American Type Culture Collection. Human glioblastoma U251 MG cells were kindly provided by Dr. Kevin Camphausen (Radiation Oncology Branch, National Cancer Institute). The selection of cell lines was based on (a) the possibility and ease of generating xenografts and (b) the HER2 expression levels. BT474, SKBR-3, and SKOV-3 cells were chosen as cell lines with a high HER2 expression level (+4 to +5) and the U251 MG cell line was chosen as a negative control. All cells were maintained in culture medium at 37°C at 5% CO₂ in a humidified environment. SKBR-3 and SKOV-3 cells were grown in McCoy's 5A supplemented with 10% fetal bovine serum and penicillin/streptomycin (10,000 units penicillin, 10 mg streptomycin). BT-474 and U251 MG were grown in DMEM supplemented as described above. The cells were detached from culture plates with a solution of 0.05% trypsin and 0.02% EDTA in PBS (Invitrogen).

Cell proliferation assay. The Cell Counting Kit-8 (CCK-8) Assay (Dojindo) was used to analyze the effect of Affibody molecules and/or trastuzumab on cell viability in vitro. Cells were plated in 96-well flat-bottomed plates at a density of 1 × 10³ to 2 × 10³ per 100 μL growth medium. Affibody molecules and trastuzumab were dissolved in PBS at 1 and 20 mg/mL, respectively. These stocks were further serially diluted in tissue culture medium and 100 μL of each were added directly to cell cultures plated previously into 96-well plates. Cells were incubated for 2 days after treatments at increasing concentrations from 1 nmol/L to 10 μmol/L. The results were expressed as the relative percentage of absorbance compared with untreated controls.

Clonogenic survival assay. Cells were seeded at 200 to 500 per well of six-well plates and left to attach for 2 days. Affibody or trastuzumab was diluted with the growth medium to concentrations of 0.1 and 1 μmol/L, respectively, and 2 mL of each were added to each well. Control cells were not treated with either drug. Three days after treatments, the medium was replaced by a drug-free medium. Following the addition of the drug-free medium, the cells were left to grow until colonies were visible. The cells were cultured for 14 days, with a medium change every 3 days. Colonies containing more than 50 cells were counted. Plating efficiency was calculated as number of colonies per well divided by number of cells plated and survival fraction was obtained by normalization to the untreated controls.

Confocal microscopy and flow cytometry analysis. Cultured cells were incubated with fluorescently labeled conjugates at 37°C for the indicated times. The detection of labeled cells was verified using established in vitro fluorescence methods. Flow cytometry was done using a FACSCalibur instrument (BD Biosciences) equipped with an argon-ion laser and an optional second red diode laser (source energy, 15 mW; detection time, 500 counts per second). Cell Quest software was used for data acquisition and analysis. Ten thousand cells were gated. Data were live gated by FL1 (blue laser, 488 nm) and FL4 (red laser 635 nm).

The different groups were also investigated qualitatively by laser scanning microscopy. The cells were analyzed using the 633 nm He-Ne Laser (source energy, 15 mW; LASOS) at 100% for excitation and the LP650 filter of the LSM510 (Zeiss). All images were recorded using identical settings.

Animals. Female athymic nude mice (nu/nu genotype, BALB/c background) ∼5 to 8 weeks old were used. This study was approved by Animal Safety and Use Committee of NIH (Animal Study Proposal ROB 117). Mice were cared for and treated in accordance with the Department of Health and Human Services Guide for the Care and Use of Laboratory Animals.

Subcutaneous tumor xenografts. Tumor cells were implanted into the right flanks of athymic mice. Briefly, 5 × 10⁶ to 10 × 10⁶ cells in 0.2 mL of 50% Matrigel (BD Biosciences) were injected s.c. into the right flanks. Tumor size was measured periodically using calipers and the tumors allowed to grow to ∼7 mm in diameter. The typical weight of the xenografts used for the imaging study was 100 to 300 mg.

In vivo NIR optical imaging. Fluorescence intensity was quantified using a NIR fluorescence small-animal imager; the details of which have been described elsewhere (28). Briefly, the system was based on a time-domain technique, where an advanced time-correlated single-photon counting device was used in conjunction with a high-speed repetition-rate tunable laser to detect individual photons. The imager had a laser source for fluorescence excitation (λ = 750 nm), an emission filter (λ = 780 nm) for fluorescence detection, and a computer for data analysis. A cooled, charge-coupled device camera was also used to guide the scan to the region of interest and to measure the fluorescence intensity distribution, which helped to locate the tumor inside the tissue.

To analyze the specific target accumulation of the imaging probes, mice were anesthetized by inhalation of isofluorane. Affibody conjugates (10 μg) or trastuzumab conjugates (50 μg) were injected into the tail vein and imaged at predetermined time points after injection (n = 3 in each group). For the blocking experiment, 1 mg unlabeled Affibody or 5 mg unlabeled trastuzumab was injected into the tail vein 1 h before the injection of labeled conjugates. The fluorescence signal mean and SD were calculated by averaging the maximum pixel values over the tumor area and corresponding contralateral site. Unless otherwise indicated, three independent experiments were done.

Cellular toxicity of trastuzumab and HER2-specific Affibody molecules. To evaluate the toxicity of HER2-specific Affibody proteins, their effects on cell proliferation and clonogenic survival were assessed. Two cancer cell lines known to overexpress the HER2 receptor, SKBR-3 and SKOV-3, were used for this study. Incubation of these cells with the Affibody for 48 hours did not significantly inhibit cell growth and proliferation in any concentration range tested (1 nmol/L-10 μmol/L); however, trastuzumab did inhibit cell growth and proliferation in a dose-dependent manner, up to 20% during a 48-hour incubation period in both cell lines. Simultaneous exposure to Affibody molecules and trastuzumab did not change the effect of the latter on the proliferation of the cells (Fig. 1A). Clonogenic survival assays showed similar results, suggesting that the Affibody did not significantly affect the clonogenic survival of these cells 2 weeks after the treatment, whereas trastuzumab did decrease clonogenic survival to 50% (Fig. 1B). Affibody as well as trastuzumab did not induce any significant effect on HER2-negative U251 cells as expected (Data not shown).

Fluorescence conjugation of HER2-specific Affibody molecules. Thiol-reactive conjugation with maleimide-containing dyes was adopted for our study because a unique COOH-terminal cysteine residue had been incorporated into the Affibody protein by the manufacturer to allow site-specific labeling (Supplementary Fig. S1A). This cysteine residue was used to fluorescently label the Affibody proteins with thiol-reactive Alexa Fluor-maleimide dyes. Before conjugation, the Affibody proteins had to be reduced to expose their free sulfhydryl groups because some of the proteins had formed dimers during storage (Supplementary Fig. S1B). Estimated labeling efficiencies of the final, labeled Affibody conjugates ranged from 60% to 80% based on assessments by silver staining and spectroscopic absorbance measurements, assuming a 1:1 labeling reaction. Trastuzumab was also conjugated with Alexa Fluor 680 and 750 using commercially available kits. The degree of labeling for the trastuzumab conjugates was estimated at 2.5 to 3.5 based on the manufacturer's instruction.

Measurement of in vitro binding affinities of the Affibody conjugates. Biacore analysis was done without regeneration of the HER2 capture surface due to surface instability. Because the reported K_D of Z_HER2:342 was 22 pmol/L (24), the Affibody was tested at a concentration range of 7.8 pmol/L to 0.5 nmol/L, in 2-fold dilutions, to encompass the equilibrium constant. Representative data are shown in Supplementary Fig. S1C. The titration series of the Alexa Fluor–labeled Affibody worked well enough to see saturation binding of the surface, but the binding affinity was somewhat less than the unlabeled Affibody Z_HER2:342. The calculated K_D values and rate constants are presented in Table 1.

Fluorescence-activated cell sorting analysis. To assess the binding specificities and patterns of the fluorescence-labeled Affibody conjugates, flow cytometry was done with two different types of human cancer cell lines. The breast adenocarcinoma cell line SKBR-3 overexpresses the HER2 receptor and the brain glioblastoma cell line U251 MG does not express the HER2 receptor. After labeling SKBR-3 cells with Affibody-Alexa Fluor 488 conjugate, cell sorting revealed HER2 receptor-specific and high-affinity binding of the conjugate. As shown in Fig. 2A and Supplementary Table S1, 5 nmol/L of the conjugate bound to 94.6% of cells and 50 nmol/L of the conjugate bound to 98.9% of cells. These results, together with the results of separate experiments using 1, 10, and 100 nmol/L of the conjugate (data not shown), indicated high-affinity binding with saturation at concentrations lower than 10 nmol/L. Alexa Fluor dyes themselves did not bind to any of the cells (1.1% for SKBR-3 cells and 0.9% for U251 MG cells). A blocking experiment that included preincubation with a 100-fold excess of unlabeled Affibody showed almost complete binding inhibition (0.8% with 5 nmol/L and 1.1% with 50 nmol/L), suggesting that the conjugate competes with unlabeled Affibody and is specific to the HER2 receptor. Non-HER2-expressing U251 MG cells failed to show any significant binding to the Affibody conjugate under all conditions tested, confirming the specificity of the HER2 receptor.

As shown in Fig. 2B and Supplementary Table S2, double labeling of cells with the Affibody-Alexa Fluor 488 and trastuzumab-Alexa Fluor 680 conjugates further showed that their binding is HER2 receptor specific and independent of each other, indicating that the binding epitopes are separate and different. Blocking the binding with excess unlabeled Affibody or trastuzumab showed almost complete binding inhibition of the corresponding conjugate (97.0% when blocked with Affibody and 98.3% when blocked with trastuzumab), whereas the binding of the other conjugate was not blocked. The binding of both conjugates was also completely inhibited in the presence of the two blockers (97.4%). It also showed that the non-HER2-expressing U251 MG cells did not bind the conjugates in any of the treatments.

Confocal microscopy. Confocal microscopy experiments were done to further investigate the binding specificities and patterns of the Affibody and trastuzumab conjugates. After double-labeling cells with the Affibody-Alexa Fluor 488 and trastuzumab-Alexa Fluor 680 conjugates, confocal images showed an almost exclusive binding of the conjugates to cell surface membranes. The images were perfectly merged on the same field, showing that their binding occurred in the same region (Fig. 3A). Blocking with 100-fold excess amounts of unlabeled Affibody or trastuzumab resulted in an almost complete absence of fluorescence from the corresponding conjugate, whereas the fluorescence from other conjugate did not change. The binding of both conjugates was also completely inhibited in the presence of the two blockers (unconjugated Affibody and trastuzumab), suggesting that the conjugates are specific to HER2 receptors and the binding of the two conjugates occurs independently. Because it is known that trastuzumab can be internalized into the cells, the study was further extended to see whether the Affibody conjugate can also enter the cells. To address this question, the labeling incubation time was prolonged, up to 6 hours, and confocal images were generated. As shown in Fig. 3B, the images suggested that both conjugates could be internalized by the cells and that the internalization was time dependent because more fluorescence was seen inside cells at 6 hours compared with the fluorescence seen at 1 and 3 hours. In contrast, non-HER2-expressing U251 MG cells failed to show any significant fluorescence from the two conjugates in any cellular regions.

In vivo NIR optical imaging. Finally, Alexa Fluor 750–conjugated Affibody molecules and trastuzumab were tested as NIR probes for in vivo optical imaging using HER2-overexpressing tumor mouse xenografts. We investigated which form of HER2-specific Affibody molecules could be best candidate among three different forms available from the company (monomer, dimer, and ABD-fused dimer). Examples of mouse whole-body images measured by charge-coupled device camera at different time points after tail-vein injection of the trastuzumab-Alexa Fluor 750, Affibody ABD-(Z_HER2:342)₂-Alexa Fluor 750, and Affibody Z_HER2:342 monomer-Alexa Fluor 750 are shown in Fig. 4A. The pharmacokinetics of each probe in the tumor area and contralateral normal tissue area were quantified using a fiber optic device and averaged. Corresponding graphs are shown in Fig. 4B. Mice injected with the trastuzumab-Alexa Fluor 750 conjugate showed moderate fluorescence intensities at first around the tumor and liver area, until the 24-hour time point. Localization of the signal intensities in the tumor area became increasingly evident and intense after 24 hours and peaked around 48 hours. After that, the signals from the tumor area lasted up to 2 weeks, while slowly growing weaker. On the contrary, mice injected with the Affibody ABD-(Z_HER2:342)₂-Alexa Fluor 750 showed high, intense signals around the tumor and kidney area until the 8-hour time point. Accumulation of the signals in the tumor area became evident and remained intense around 24 hours and peaked before 48 hours. After peaking, the signals from the tumor area decreased gradually and almost disappeared after 72 hours. However, mice injected with the Affibody monomer Z_HER2:342-Alexa Fluor 750 showed high fluorescence intensities only in the kidney area until 4 hours, and tumor signals were washed out quickly, although weak, localized signals were detected around 6 to 8 hours. Most signal intensities were cleared after 24 hours. Signal-to-background ratios of each probe from peak postinjection times were calculated by normalizing signals from tumor area to those from contralateral normal tissue area. The signal-to-background ratios were highest in Affibody ABD-(Z_HER2:342)₂-Alexa Fluor 750 followed by trastuzumab-Alexa Fluor 750 and Affibody monomer Z_HER2:342-Alexa Fluor 750 (Table 2). The clearance rates of each probe in the tumor area were calculated based on the derived equations after fitting each curve to single exponential decay. The derived half-lives were also shown in Table 2. As expected, the half-life was longest with the trastuzumab-Alexa Fluor 750 conjugate (108 hours) followed by Affibody ABD-(Z_HER2:342)₂-Alexa Fluor 750 (45 hours) and Affibody monomer Z_HER2:342-Alexa Fluor 750 (11 hours).

The fluorescence intensities from each isolated tissue, 24 hours after the injection of Affibody ABD-(Z_HER2:342)₂-Alexa Fluor 750 conjugate, were also measured using a fiber optic device and compared to see whether our in vivo measurements reflect the signal intensities from the tumor and other tissues (Fig. 4C). The tumor signal intensity was the highest, with those from the kidney and liver showing the second and third highest levels. All other tissues showed minimal intensities. To validate the specificity of the Affibody ABD-(Z_HER2:342)₂-Alexa Fluor 750, we did the blocking experiment as we did with the cultured cells. One group of mice was preinjected with an excess amount (100-fold) of unlabeled Affibody ABD-(Z_HER2:342)₂ and another group of mice was preinjected with the same excess of unlabeled trastuzumab 30 minutes before injection with the Affibody ABD-(Z_HER2:342)₂-Alexa Fluor 750 conjugate. The kinetics of the fluorescence intensities in the tumor and contralateral normal tissue area were then measured over time using a fiber optic device (Fig. 4D). Preinjecting the mice with an excess amount of unlabeled Affibody ABD-(Z_HER2:342)₂ blocked most of the fluorescent signals from the Affibody ABD-(Z_HER2:342)₂-Alexa Fluor 750 conjugate, whereas preinjecting with an excess amount of unlabeled trastuzumab failed to block the signals. Specificity of our HER2 imaging analyses were further strengthened by the experiments with HER2-nonexpressing U251 tumors as shown in Fig. 4E.

Molecular imaging has its roots in nuclear medicine, which focuses on the management of patients through the use of radiotracers in conjunction with imaging technologies, such as positron emission tomography and single photon emission tomography. The underlying principle can now be tailored to other imaging modalities, such as optical imaging. Recent, fast-moving trends in this field will allow us to establish noninvasive, specific, sensitive, quantitative, and cost-effective NIR optical imaging techniques. One goal is to provide an ideal probe for HER2 imaging in living animals and, ultimately, in humans. HER2-specific Affibody molecules have been shown to be good candidates for in vivo molecular imaging using radionuclide modalities (13–17). However, there is a lack of information regarding the application of these HER2-specific Affibody molecules to in vivo NIR optical imaging. In this study, we assessed the feasibility of using of HER2-specific Affibody molecules to create a NIR probe for in vivo optical imaging of HER2.

Because part of the goal is to provide probes that will minimally interfere with the studied system, we have examined what effects HER2-specific Affibody molecules might have on the target cells and whether their binding to HER2 interferes with trastuzumab binding and its therapeutic effects. As a first step toward this goal, we assessed the cellular toxicity of HER2-specific Affibody molecules, because, to be an ideal candidate, the probe should minimally affect the target cells. Our data show that the HER2-specific Affibody is not toxic to HER2-expressing cells, whereas treatment with trastuzumab significantly decreased the viability of the same cells. Furthermore, the presence of the Affibody molecules did not affect the trastuzumab toxicity (Fig. 1). Our results agree with previous findings on cellular toxicity and intracellular signaling pathways (29, 30), supporting our claim that our probe does not affect the system to be studied.

It has been well documented that conjugation chemistry using a maleimide group to react with a thiol group is an efficient method of conjugation (23, 31). It can create a specific 1:1 labeling of proteins with fluorescent dyes if the protein contains only one cysteine residue, like the HER2-specific Affibody used in this study. As shown in Table 1, modification of the HER2-specific Affibody molecules with Alexa Fluor-maleimide dyes resulted in a slight reduction of the apparent affinity of the conjugates perhaps due to steric hindrance. However, the reduction was minor and the conjugate still showed a picomolar range of affinity to HER2, so it still has a better affinity than trastuzumab (which exhibits a nanomolar range of affinity to HER2). Through competition studies, we have shown that the epitope recognized by the HER2-specific Affibody is distinct from the epitope recognized by trastuzumab. This would allow the use of the HER2-specific Affibody-based reagents to test for changes in HER2 receptor expression in response trastuzumab therapy. Our data from fluorescence-activated cell sorting and confocal microscopy studies showed clearly that trastuzumab and Affibody molecules do not compete for binding to HER2 (Figs. 2 and 3). Taken together, these data show that the two conjugates are highly HER2 receptor specific and that their HER2 binding epitopes are separate and distinct.

The advantage of an optical method is that it does not involve radioactive materials and is relatively easy to use. Moreover, due to its low absorption in tissue, the NIR optical methods allow us to target even deeply located targets. During the last decade, advanced mathematical modeling (32–35) of propagation of light in tissue and technological improvements of sources and detectors has made possible optical imaging for clinical application. There are several optical methods and systems in use today in clinic where light can penetrate more than 10 cm deep inside the breast tissue using transillumination measurement (36, 37). Additionally, techniques have made possible the application of tomographic principles for imaging of diffuse light (38–41).

In this study, optical imaging system presented for preclinical studies on small animal to study new drug. The current time-domain optical system (28) can be extended for clinical use similar to the Rinneberg group (37) and softScan (Advanced Research Technologies; ref. 42). The unique advantage of our system is that its fixed source-detector separations allow us to obtain signals at different depths in the medium. In conjunction with the time-resolved system, a cooled charge-coupled device camera allows the acquisition of a fluorescence image superimposed on a grayscale photographic image of the small animal using an overlay.

Pharmacokinetics of trastuzumab conjugate was consistent with previous findings by other groups using different fluorophores, thus verifying our imaging system (43, 44). On the contrary, Affibody monomer Z_HER2:342-Alexa Fluor 750 conjugate failed to show clear signs of tumor accumulation and fast clearance. Instead, it showed fast accumulation in the kidney. Previous reports on the use of Affibody monomer Z_HER2:342 in radionuclide imaging also showed fast clearance from the blood and instant uptake into the kidney, but they reported tumor accumulation within several hours after injection (24, 25). Our conflicting data for this aspect could be due to differences in chemical nature between our fluorescent dye and their radiotracers or to differences in the imaging modalities. In our hands, there was no improvement in tumor accumulation of the Affibody monomer Z_HER2:342-Alexa Fluor 750 conjugate, even with different batches and different dosages (data not shown). Furthermore, the dimer form of the Affibody conjugate also failed to show clear tumor accumulation (data not shown). Fusion of the ABD to imaging probes has been successfully adopted to prolong the circulation half-life of the probes, including the HER2-specific Affibody molecule (45, 46). With the use of Affibody ABD-(Z_HER2:342)₂-Alexa Fluor 750 conjugate, we finally observed clear signs of tumor accumulation with an extended half-life. These results showed that the Affibody ABD-(Z_HER2:342)₂ conjugate was the best performer for in vivo NIR optical imaging, in terms of tumor accumulation and clearance, compared with the trastuzumab and Affibody monomer and dimer conjugates.

To our knowledge, this is the first report on the quantitative in vivo NIR optical imaging of HER2-expressing tumors using Affibody molecules. This approach should provide the opportunity for the development of a rapid and cost-effective imaging modality as an improvement over costly radionuclide-based imaging modalities. A noninvasive paradigm for molecular imaging could provide a new tool for optimization of HER2-targeted therapies using preclinical animal models and a unique noninvasive means for in vivo characterization of HER2 expression in patients and possible monitoring of their down-regulation in response to therapeutic interventions. The application of our probe to a combination of conventional and HER2-specific molecular therapies may provide new approaches in the treatment of HER2-positive breast cancer.

No potential conflicts of interest were disclosed.

The authors appreciate the support of Affibody AB. We also thank Michael Kruhlak for confocal microscopy, Barbara Joan Taylor for fluorescence-activated cell sorting, and Lesa Holly for general laboratory assistance.