|Home | About | Journals | Submit | Contact Us | Français|
In vivo imaging of HER2 expression may allow direct assessment of HER2 status in tumor tissue and provide means to quantify changes in receptor expression following HER2-targeted therapies. This work describes in vivo characterization of HER2-specific 18F-FBEM-ZHER2:342–Affibody molecule and its application to study the effect of 17-DMAG on HER2 expression by PET imaging.
To assess the correlation of signal observed by PET with receptor expression, the tracer was administered to athymic nude mice bearing subcutaneous human breast cancer xenografts with different levels of HER2 expression. To study the down-regulation of HER2, mice were treated with four doses (40 mg/kg) of 17-DMAG, an inhibitor of Hsp90, known to decrease the HER2 expression. Animals were scanned before and after the treatment. After the last scan, mice were euthanized and tumors were frozen for receptor analysis.
The tracer was eliminated quickly from the blood and normal tissues, providing high tumor/blood and tumor/muscle ratios as early as 20 min post injection. The high contrast images between normal and tumor tissue were recorded for BT474 and MCF7/clone18 tumors. Very low but still detectable uptake was observed for MCF7 tumors and none for MDA-MB-468. The signal correlated with the receptor expression assessed by immunohistochemistry as well as western blot and ELISA. The levels of HER2 expression, estimated by post-treatment PET imaging, decreased 71% (p < 4 × 10−6) and 33% (p < 0.002), respectively, for BT474- and MCF7/clone18-tumor bearing mice. These changes were confirmed by the biodistribution studies, ELISA and western blot.
Our results suggest that the described 18F-FBEM-ZHER2:342–Affibody molecule can be used to assess HER2 expression in vivo by PET imaging and monitor possible changes of receptor expression in response to therapeutic interventions.
Human epidermal growth factor receptor type 2 (HER2, ErbB-2, neu,) is a well-established tumor biomarker that is overexpressed in a wide variety of carcinomas including breast, ovarian, prostate, and lung cancer [1, 2]. Since HER2 overexpression plays an important role in aggressive tumor behavior and poor clinical outcome , the early-stage detection and quantification of HER2 is clinically relevant and could be used for selection of optimal therapy for individual patients.
Hsp90 (heat shock protein 90) is a molecular chaperone highly expressed in most of the tumor cells. It is required for the stability and function of many client proteins that promote cancer cell growth and survival . The Hsp90 inhibitors, by interfering with the chaperone activity, result in targeting of client proteins to the proteasome for degradation. Since HER2 is one of the Hsp90 client proteins, it can be indirectly down-regulated by Hsp90 inhibitors such as the naturally occurring ansamycin antibiotic geldanamycin (GA) known for its tumoricidal potential  or the recently developed 17(dimethylaminoethylamino)-17-demethoxygeldanamycin (17-DMAG) which is a hydrophilic GA derivative that can be administered orally with good bioavailability and better activity in vitro and in vivo than its predecessor, 17- allylamino-17-demethoxygeldanamycin (17-AAG) . In vivo downregulation of HER2 in human tumor xenografts following treatment with 17-AAG was reported by Smith-Jones et al.  and, later, by Orlova et al. . These GA analogues are currently being tested in the Phase I clinical trials .
Thus far, HER2 expression pattern has been routinely determined by ex vivo analysis of tissue samples using fluorescence in situ hybridization (FISH) and/or immunohistochemistry (IHC). These methods, although commonly used in clinical practice, have several limitations. Most notably, they require tissue removal from the body, which restricts their analysis only to the sampled parts and may not properly represent the overall tumor characteristics. The variability in scoring between these techniques, whether as a result of true heterogeneity or artifacts in preparation, has led to decreased reliability of the final HER2 status determination [10, 11].
Introducing a new methodology for in vivo quantification of HER2 receptors using PET imaging would present a complementary, noninvasive option to obtain real-time information that could not only facilitate selection of patients for HER2-targeted therapy [3, 12], but also could provide information regarding the immediate response to therapeutic interventions, both in primary lesions and in distant metastases. This feedback would allow adjusting the dose and treatment schedule for individual patients, based on the actual status of HER2 receptors. PET imaging with its high sensitivity, high special resolution, and proven quantification abilities, could also reduce the number of false-negative or false-positive results of the currently utilized methods: FISH and IHC.
Several molecular probes based on antibodies (Abs) have been recently tested in experimental animal tumor models, but a PET tracer for routine clinical use has not yet been developed. The application of antibodies (~150 kDa) for molecular imaging is limited because of their large size, resulting in low tumor penetration and slow clearance, both of which hamper their clinical imaging applications. Often several days are needed to obtain reasonable tumor/blood ratios, making most of the short-lived PET-imaging radionuclides inapplicable. To circumvent these problems, several different ligands have been developed and extensively studied over the last few years. Among them are: Abs fragments and engineered variants like F(ab’)2, F(ab’), single-chain Fv (scFv), diabodies and minibodies .
Recently, a new class of relatively small (~6.5 kDa) proteins has become available for studies, and several groups are now using Affibody molecules to image HER2 [14, 15] or EGFR-positive tumors . The small size, resulting in rapid blood clearance, good tumor penetration, and high binding affinity to selected targets make Affibody molecules ideal candidates for imaging purposes. HER2-specific Affibody molecules, which bind with picomolar affinity to HER2 epitope distinct from those involved in binding of trastuzumab or pertuzumab, have been radiolabeled with several radioisotopes, including 99mTc, 111In, 68Ga, 90Y, and 125I [8, 17-19]. A methodology for labeling Affibody molecules with 18F has recently been developed in our laboratory [14, 20]. We have reported a great potential for 18F-labeled FBEM-ZHER2:342–Affibody molecules to be used for in vivo monitoring of HER2 expression by PET imaging. Cheng et al.  used a similar approach to test various clones of anti-HER2 Affibody molecules, obtaining comparable results.
In the current study, we tested the hypothesis that the use of 18F-FBEM-ZHER2:342–Affibody molecule would enable quantitative assessment of HER2 down-regulation after anti-HER2 therapy. We are the first to show that semiquantitative analysis of PET imaging data can be used to assess different levels of HER2 expression, and to monitor changes in their expression in response to 17-DMAG treatment in mice bearing xenografts tumors.
Unless otherwise specified, all reagents were analytical grade and were obtained from commercial sources. The ZHER2:342–Cys Affibody molecules were kindly provided by our Cooperative Research and Development Agreement (CRADA) partner in Sweden (Affibody AB; http://www.affibody.com). 18F radionuclide was produced via the 18O(p,n)18F nuclear reaction in the Clinical Center of the National Institutes of Health (CC/NIH) cyclotron facility by irradiating 18O-enriched water. 18F-FBEM-ZHER2:342–Affibody was prepared as previously described . 17-DMAG was purchased from InvivoGen, San Diego, CA, USA, as lyophilized purple powder and reconstituted with 0.9% sodium chloride for injection.
Human breast (BT474, MDA-MB-361, MCF7, and MDA-MB-468) cancer cell lines that express different levels of HER2 were purchased from the American Type Culture Collection (Rockville, MD, USA). The breast cancer cell line, stably transfected with HER2 (MCF7/clone18), was kindly provided by Drs. John W. Park and Byron Hann, University of California, San Francisco, CA, USA. The cells were cultured in RPMI1640 (BT474, MDA-MB-468), DMEM (MCF7, MDA-MB-361, MCF7/clone18) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (GIBCO, Grand Island, NY, USA) and penicillin/streptomycin (100 U/ml of each). Cells were grown as a monolayer at 37°C in a humidified atmosphere containing 5% CO2. In the case of MCF7/clone18 transfectants, the culture medium also contained 400 μg/ml Geneticin (GIBCO, Grand Island, NY, USA).
All animal studies were conducted in accordance with the principles and procedures outlined in the NIH Guide for the Care and Use of Animals on approved studies from the NIH Institutional Animal Care and Use Committee. Five- to 7-week-old female athymic nude mice were implanted subcutaneously in the shoulder region with 4.5 × 106 of MDA-MB-361, MCF7, MDA-MB-468, MCF7/clone18 or with 5.5 × 106 of BT474 cells, suspended in Matrigel (BD Bioscience, San Jose, CA, USA). Estrogen pellets 1.72 mg (Innovative Research of America, FL, USA) were implanted 48 h before tumor cell inoculation and remained in place until the end of the study. Tumors (100–250 mg) developed after 3–5 weeks. To monitor the down-regulation of HER2 receptors, the mice were treated with four doses (i.v., 40 mg/kg) of 17-DMAG, an inhibitor of Hsp90. Following PET imaging, the animals were dosed 12, 36, 48, 60 hours and then imaged prior to 72 h after the first scan.
Mice bearing BT474 and MCF7/clone18 tumors were injected with 3.7–4.4 MBq (3.4–4.0 μg, 100 μL) of 18F-FBEM-ZHER2:342–Affibody into the tail vein. Groups of 6–7 mice were sacrificed, and their major organs dissected and weighed 2 h post-injection. The radioactivity uptake in the tissues was measured, along with a standard of the injected dose using a γ-counter. The results were expressed as a percentage of injected dose per gram of tissue (% ID/g).
Mice were anesthetized using isoflurane/O2 (1.5–5% v/v), placed in a prone position in the center of the field of view of the scanner, and injected with approximately 3.7–4.4 MBq (3.4–4.0 μg, 100 μL) of the 18F-FBEM-ZHER2:342–Affibody via the tail vein. PET scans were performed using the Advanced Technology Laboratory Animal Scanner (ATLAS) PET scanner . Whole-body (four bed positions, each 15 min) or dynamic data acquisition (1 frame, 10 min; 6 or 12 frames) were started about 2 min after radiotracer injection and recorded with a 100–700 keV energy window. In some cases, additional static scans (15 min) were performed 1 h post tracer injection. The time points were chosen based on the previous studies in which maximal uptake was observed at 1 h post-injection and then maintains a plateau until 2 h . The images were reconstructed by a two-dimensional ordered-subsets expectation maximum (2DOSEM) algorithm, and no correction was applied for attenuation or scatter. For each scan, regions of interest (ROI) were drawn over the tumor, normal tissue, and major organs. The maximum counts per pixel within the tumor or organs were obtained from the multiple ROI (counts per second per cubic centimeter). The results were calculated as a percentage of injected dose per gram (% ID/g) by means of a calibration constant obtained from scanning the 18F source of known activity, assuming the tissue density of 1 g/mL, and dividing by the injected dose, decay corrected to the time of scanning.
Sub-confluent cell cultures were incubated with 50 μg/μL Affibody molecules or just with growth medium before fixation for 1 h at 37 °C. Then cells were washed twice with phosphate-buffered saline (PBS−) and detached with Cellstripper™, (Cellgro, Herndon, VA, USA). For every sample, 1 × 106 cells were collected, fixed with 4% formaldehyde and incubated 1 h on ice. Then cells were washed three times with PBS− and resuspended in stain buffer (PBS− plus 5% FBS and 0.5% Triton X-100). Subsequently, the cells were covered and incubated on ice 30 min, and after that appropriate dilution of goat-anti-Affibody antibody was added for 45 min to cells treated before with Affibody molecules. After that, cells were washed once and resuspended in 200 μL of stain buffer, followed by 30-min incubation with secondary rabbit-anti-goat-AlexaFluor488 conjugated antibody (Molecular Probe, Eugene, OR, USA). For HER2 detection, cells, after permeabilization, were incubated 45 min with anti mouse Neu(24D2)-FITC monoclonal antibody or with IgG1-FITC Mouse Isotypic Control (Santa Cruz Biotechnology, INC, Santa Cruz, CA, USA).
After antibody staining, all cells were washed with stain buffer two times and resuspended in 300 μL of PBS− for data acquisition. All solutions were kept on ice.
Flow cytometry was done using a FACS Calibur instrument (BD Biosciences, San Jose, CA, USA). CellQuest Pro software was used for data acquisition and FlowJo software for analysis (Tree Star Inc., Ashland, OR, USA). 10,000 events were recorded for each sample and the population corresponding to single cells was gated and analyzed as a histogram plot.
To determine HER2 protein level in cell lines, cell cultures were harvested by Cellstripper™ and washed three times in PBS. Then, cell pellets were resuspended in a proper amount of resuspension buffer provided with the ELISA Kit (Calbiochem, La Jolla, CA, USA). For each 100 μL of cell suspension, 20 μL of Antigen Extraction Agent (also provided with the kit) was added and cells were incubated at 4 °C for 20 min with a gentle agitation, afterwards centrifuged at 13,000 rpm for 15 min. The supernatant was collected and stored at −80 °C.
Mice were scarified by cervical dislocation and tumors were isolated, frozen on dry ice, and stored at −80°C. After that, the specimens were weighed and sliced into small pieces. Then tumor tissue was homogenized on ice, using Polytron homogenizer, and lysed in resuspension buffer, provided with the ELISA kit. To remove cell debris, the suspensions were centrifuged at 13,000 rpm (4 °C) for 15 min. The supernatants were collected and stored at −80 °C. Before use, the protein concentration was estimated by the BCA assay kit (Pierce, Rockford, IL, USA), according to the manufacturers protocol. HER2 protein level was measured by an ELISA kit, according to the manufacturer's recommended procedure. Results are expressed in nanograms of HER2 per milligram of protein. The data are presented as the mean of two replicates, with each experiment repeated three times.
All reagents for NuPAGE and Western blots were from Invitrogen (Carlsbad, CA, USA). Total cellular protein (30 μg for all analyses) was mixed with sample buffer and incubated at 70 °C for 10 minutes. 30 μl aliquots were resolved on 4–12% NuPAGE Novex tris-acetate gel by electrophoresis. Thereafter, proteins were transferred to a nitrocellulose membrane, which was subsequently blocked for 1 h at room temperature with 5% non-fat milk blocking buffer, and then incubated with rabbit polyclonal antibody (anti-HER2), or a loading control, the mouse monoclonal antibody anti-α–Tubulin (clone DM1A) (Cell Signaling Technology, Beverly, MA, USA) at 4 °C overnight. After extensive washing with TBST, membranes were incubated with secondary antibodies HRP-conjugated goat anti-rabbit and anti-mouse IgG1 for 1 h at room temperature. After final washing, proteins were visualized with a chemiluminescence detection system (Pierce, Rockford, IL, USA) and subsequent exposure to the Kodak Molecular Imaging System (Kodak, Rochester, NY, USA). Protein band densities were analyzed using the spot density analysis software provided with the image station.
HER2 was analyzed immunohistochemically in paraffin sections using the HercepTest™, an FDA-approved assay for identification of tissue overexpressing p185 HER2 (K5205, Dako, Denmark), in accordance with the manufacturer's protocol and scoring guidelines. The percentage of positive tumor cells was determined by experienced pathologists from Pathology/Histotechnology Laboratory, SAIC Frederick, Inc./NCI-Frederick (http://web.ncifcrf.gov/rtp/lasp/phl/).
Data are presented as mean ± SD. Statistical comparisons were made using the Student t Test. A p value of less than 0.05 was considered significant.
To assess the correlation between 18F-FBEM-ZHER2:342–Affibody uptake and HER2 receptor density in vivo, the tracer was administered into mice bearing subcutaneous tumors with five levels of HER2 expression: 1. BT474 (very high); 2. MCF7/clone18 (high); 3. MDA-MB-361 (medium); 4. MCF7(very low); and 5. MDA-MB-468 (negative). For all animals, 15-min static images were recorded 1 h post tracer injection. Afterward, tumor uptake was evaluated by semiquantitative analysis of PET images and confirmed by the biodistribution studies (Figure 1). Very good agreement was found between the results obtained by these two methods. These results also correlated very well with HER2 density corresponding to each particular tumor model, except for MCF7/clone18, where the specific uptake of 18F-FBEM-ZHER2:342–Affibody was much lower than we expected, based on ex-vivo analysis of HER2 expression as measured in the cells and tissue lysates by ELISA assay (Table 1). Since this discrepancy could result from different affinity of tracer binding to MCF7/clone18 cells, we checked 18F-FBEM-ZHER2:342–Affibody binding to MDA-MB-361 and MCF7/clone18 cells in vitro by FACS analysis (Figure 2). The results showed higher binding of Affibody molecules to MCF7/clone18 than to MDA-MB-361 cells. These data corresponded to the HER2 expression measured in in vitro and ex vivo samples by ELISA and FACS, suggesting that the observed lower-than-expected accumulation of the tracer in MCF7/clone18 tumors is not caused by the receptor binding characteristics but most likely by the differences in tumor microenvironment.
To check the feasibility of PET analysis as complementary to IHC in the routine clinical application, the same tissue samples were also analyzed by standard immunostaining assay, HercepTest, according to score guidelines. Images of representative tumor cryosections stained for HER2 expression are shown in Figure 3. The receptor expression was found to be relatively homogenous throughout the field of the sections. The IHC scores for BT474, MCF7/clone18, MDA-MB-361 cell lines were +3 and for MCF7 +1; MDA-MB-468 cells were assessed as HER2-negative.
Two-hour dynamic studies were done for mice bearing BT474 tumors in order to assess the kinetics of 18F-FBEM-ZHER2:342–Affibody accumulation in the tumor. The results showed that high-contrast PET images data can be collected 1 h post-tracer injection. At this time point, tumors were clearly distinguishable and tumor accumulation was the dominant feature (data not presented). The biodistribution of the radioactivity 2 h post tracer injection was studied in mice bearing BT474, MCF7/clone18, and MCF7 tumors through monitoring by PET imaging. The results agreed with previously published data . The biodistribution of the tracer as determined by gamma counter (Figure 4) showed a high activity accumulation in BT474 tumors (21.3 ± 3.7% ID/g), and MCF7/clone18 (14.8 ± 2.0% ID/g), and was clearly lower (p < 0.004) in MCF7 (4.2 ± 0.7% ID/g) tumors. The highest normal tissue concentration of radioactivity was found in the kidneys. However, the renal uptake of 18F-FBEM-ZHER2:342–Affibody was significantly lower than in the tumors of BT474 and MCF7/clone18 tumor-bearing mice and corresponded to 11.1 ± 2.7% ID/g (p < 0.02) and 5.8 ± 2.1% ID/g (p < 0.0002), respectively. Even for MCF7 tumor-bearing mice, the signal detected in the tumor and in the kidneys, although less pronounced due to very low expression of HER2 receptors on those tumors (Table 1), was significantly different (p < 0.05). Representative images of dynamic study and whole body coronal images of mice bearing xenografts with different HER2 expression are shown on Figure 4. Very high tumor to background contrast images were recorded for BT474 tumors and it was still possible to detect MCF7 tumors with very low HER2 expression.
The effect of therapeutic regimens on the HER2 expression after 17-DMAG treatment was compared in mice bearing BT474 and MCF7/clone18 tumors. Dynamic scans were carried out by acquiring a continuous series of 6 frames post-tracer injection before and after 17-DMAG treatment. Next, 15-min static images were taken. In each case ROI was placed around the tumor and muscle tissue and a time-activity curve or radioactivity concentration were calculated (Figure 5). We decided to use maximal tumor uptake instead of mean tumor uptake, because maximal tumor uptake measurement is free of user-dependent variation in defining ROI, and is less susceptible to the effects of the necrosis that appears when tumors grow or respond to the treatment during the monitoring time. Interestingly, no particular differences in the tumor uptake were observed, based on image contrast, comparing data before and after the treatment (even though the animals received a high dose of the drug). However, the subsequent numerical analysis of recorded images clearly showed different 18F-FBEM-ZHER2:342–Affibody accumulations, confirming that the HER2 receptor expression declined, compared to the pretreatment level. In the case of animals bearing BT474 tumors, after four doses of 17-DMAG (40 mg/kg), the tracer uptake was 71% lower (p < 4 × 10−6), compared to non-treated animals. The tumor response was less effective in the case of mice bearing MCF7/clone18 tumors (only 33% decrease, p < 0.002). It is noteworthy that the final uptake of the tracer, determined by analysis of the images, was the same for both tumor models. To confirm these data, the HER2 expression in the tissues extracted from the same tumors was also measured ex vivo by an ELISA, an immunoblot analysis, as well as by an immunostaining assay using HercepTest. The numerical analysis is presented in Table 2 and the representative Western blots for BT474 and MCF7/clone18 tumor models are shown in Figure 5. In this case we also obtained some unexpected results. According to the IHC, BT474 tumors expressing a high level of HER2 protein showed the same strong membrane and cytoplasmic staining corresponding to score 3+ (data not presented) before and after the treatments. Whereas ex vivo analyses of BT474 tumor tissue lysates by Western blot and ELISA assay indicated that the treatment with 17-DMAG resulted in significant downregulation of HER2 protein.
Molecular imaging of specific markers may provide the information on both the expression of targeted molecules and the early assessment of response to a particular therapeutic intervention. Particularly PET, as a functional imaging technology, can provide rapid, reproducible, and noninvasive in vivo assessment of the receptor expression. Radioligands for in vivo imaging of HER2 by PET have been developed by several groups. Robinson et al. labeled a divalent antibody (C6.5) fragment with 124I and showed that it could be used for the imaging of HER2-positive tumors using a clinical PET/CT (computed tomography) scanner . While Garmestani et al. labeled intact trastuzumab with 86Y or 111In , Olafsen et al. used trastuzumab to create a minibody and labeled it with 64Cu . The resulting radioconjugate was successfully applied to image HER2 in tumor xenografts by micoPET. Smith-Jones et al. used an F(ab′)2 fragment of trastuzumab labeled with 68Ga for PET . Their report presents the first attempt to monitor in vivo changes of HER2 expression following therapeutic intervention.
Many reports describing the application of Affibody molecules for imaging and therapy of HER2-posistive tumors have been published by the Swedish Group [8, 25-27]. We were the first to independently confirm the uniquely advantageous characteristics of these molecules as an HER2 targeting agent, especially for diagnostic application. By labeling the Affibody molecule containing a C-terminal cysteine (ZHER2:342-Cys) with 18F, we have created a new tracer that will allow the application of a well-established PET methodology to quantify HER2 expression in vivo. As described in our recent publication , the elimination half-life of about 36 min and high accumulation of the radioactivity in HER2-positive tumors as compared with the background, facilitated the imaging studies. Our findings were then confirmed by Cheng et al., who also labeled Affibody molecules with 18F .
In this work, we have tested the sensitivity of our approach combining 18F-labeled Affibody molecules and the PET imaging technique to detect differences in HER2 expression levels and to monitor the changes in HER2 expression following therapeutic intervention. To assess the correlation of signal observed by PET with receptor expression, we administered the new tracer to athymic nude mice bearing subcutaneous tumors with five levels of HER2 expression. The results showed that 18F-FBEM-ZHER2:342–Affibody rapidly accumulated in HER2-positive tumors and was eliminated quickly from the blood and normal tissues, providing high tumor/blood and tumor/muscle ratios as soon as 20 min post-injection. Notably, it was also possible to detect MCF7 tumors expressing very low level of HER2. The only other organs with higher accumulation than the tumors were kidneys, indicating that the clearance occurred exclusively via the renal tract. However, for monitoring breast tumors and even distant metastasis (except for adrenal and liver), the reconstruction artifacts resulting in difficulties visualizing target will be negligible.
Semi-quantitative PET analysis of tumors showed that the signal recorded by PET depended on the number of receptors expressed in those particular cell lines, as assessed by ex vivo methods. We found that PET data analysis corresponds very well with the biodistribution results for each studied tumor model and depended on the HER2 expression, except for MCF7/clone18. The level of the tracer binding in vivo in these tumors was comparable to the binding obtained in MDA-MB-361 xenografts (Figure 1) in spite of the fact that the receptor expression in the later was three-fold lower (Table 1). In vitro FACS analysis of cell lysates (Figure 2) showed that this divergence could not be explained by lower affinity of Affibody molecules to MCF7/clone18 cells. Similar discrepancy was observed by McLarty et al . In these studies, the radioactivity uptake in MCF7/clone18 xenografts following injection of 111In-DTPA-trastuzumab was two-fold lower than that observed in MDA-MB-361 suggesting non-linear relation of the imaging data to the actual concentration of the receptors. The authors showed also that the correlation can be disturbed by non-specific binding. In our recent work, we investigated the nonspecific binding of 18FFBEM-ZHER2:342–Affibody in both HER2-positive tumors after pre-saturation of the receptors with non-labeled Affibody and HER2-negative tumor xenografts . Since our results showed the same level of binding in both cases, we assumed that non-specific binding should be similar for other subcutaneous tumor xenografts and we decided to forego these experiments in the present work. It is likely that the lower than expected accumulation of radioactivity in MCF7/clone18 tumors could be due to the vascular and microenvironmental abnormalities leading to inefficient penetration of the tracer. More rigorous PET studies allowing multi-compartment kinetic analysis of the data might circumvent this problem.
Concluding that the signal from our tracer detected by PET in tumor tissue depends on different levels of HER2 expression, we moved on to evaluate its utility to monitor changes in HER2 expression following treatment. It is known that 17-AAG, or 17-DMAG, causes down-regulation of HER2 by interfering with the protective chaperone activity of Hsp90. Thus, to induce down-regulation of HER2 in human breast cancer xenografts, the animals received four doses of 17-DMAG. This treatment applied to BT474 and MCF7/clone18 tumor-bearing mice resulted in, respectively, 71% and 33% decrease of HER2 expression as observed by PET imaging. Essentially the same reduction was observed by ex vivo assays, indicating that PET imaging represents an attractive, non-invasive alternative to monitor the therapy effects on HER2 expression and to optimize the treatment schedule for individual patients. The results obtained for BT474 were in good agreement with the pharmacodynamic data previously reported by Smith-Jones et al. , who used 17-AAG and the same tumor model. Their study demonstrated an 80% reduction in HER2 expression in BT474 xenografts, which lasted approximately 36 h and, after which point recovery in expression was observed. The same group have shown that downregulation of HER2, as measured by decreased uptake of F(ab’ )2 fragments of trastuzumab, was more predictive than FDG PET of a response to 17-AAG therapy in mice . Orlova et al.  using 111In-DOTA-ZHER2:342-pep2 observed 2.3 to 3.4-fold decrease of signal from tumor volume on gamma camera images in mice bearing SKOV-3 tumors treated with 17-AGG.
We are not aware of any reports comparing the IHC with the more quantitative global ex-vivo analysis of HER2 expression. Our studies indicated significant discrepancy between the results obtained by HercepTest and ELISA or Western Blot. For example, the pathology report based on guidelines scores for HercepTest listed MDA-MB-361 tumors as 3+ HER2-positive, even though our detailed ex vivo analysis showed a medium level of receptor expression, suggesting that the score should be 2+. Similar disagreement between IHC scoring of HER2 expression using ELISA and Western blot was observed in BT474 tumors treated with 17-DMAG. Strong membrane and cytoplasmic staining corresponding to score 3+ (data not presented) was seen on both treated and non-treated tumors, while ex vivo analyses of the same tissues indicated that 17-DMAG caused significant decrease of HER2 expression in the treated tumors. This inconsistency indicated the subjective character of IHC analysis that could be further distorted by the heterogeneity of HER2 expression in individual tumors.
Our approach using PET for in vivo assessment of HER2 expression may assist diagnosis of breast cancer and improve the outcome of HER2-targeted therapies. IHC and FISH analyses are usually restricted to a limited tissue sample obtained at a single time point. After initial diagnosis, follow-up biopsies are routinely not performed, and currently, there are no means to obtain information concerning how long it takes a therapeutic agent, e.g. Herceptin™, to reach the target, how effective it is, and how long its effect lasts. Moreover, the high level of discordance between these two methods has been documented in several studies [30, 31]. Discordance rates may be as high as 20% when HER2 testing is performed in low volume, and at local laboratories . As we have shown in this work, PET imaging using the 18F-labeled Affibody molecules was sufficiently sensitive to detect a 2-3 fold decrease in HER2 expression, while IHC was not only less sensitive but also misclassified some tumors with intermediate levels of HER2 as highly overexpressing. Such inaccurate estimate of HER2 expression by IHC analysis might lead to misdiagnosis and severe clinical consequences. Therefore, PET imaging may provide a considerable advantage and become an attractive alternative providing means for quantitative, objective assessment of HER2 expression in a noninvasive manner, which would allow performing several scans over the course of therapy.
In conclusion, our results indicate that the HER2 expression in vivo and its possible changes in response to therapeutic interventions can be monitored by PET imaging using 18F-FBEM-ZHER2:342–Affibody. This approach represents an attractive, non-invasive alternative to monitor the therapeutic effects on HER2 expression and to optimize the treatment schedule for individual patients.
The authors appreciate the support of experts form Affibody AB and SAIC Frederick, Inc. We appreciate technical assistance of Yesenia Rodriguez and Aliesia Holly as well as constructive discussions with Lucia Martiniova and one of the creators of ATLAS, Jürgen Seidel.
The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organization imply endorsement by the U.S. Government.
This research was supported in part by the Center for Cancer Research, an Intramural Research Program of the National Cancer Institute; National Institute of Biomedical Imaging and Bioengineering;, and by Breast Cancer Research Stamp Fund awarded through competitive peer review, and was funded in part with Federal funds from the National Cancer Institute, National Institutes of Health, under Contracts N01-CO-12400 and N01-CO-12401.