| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Research Articles: Therapeutics, Targets, and Development
Use of fluorescent labeled anti–epidermal growth factor receptor antibody to image head and neck squamous cell carcinoma xenografts
1 Division of Otolaryngology-Head and Neck Surgery, Department of Surgery and 2 Department of Medicine, University of Alabama at Birmingham, Birmingham, Alabama
Requests for reprints: Eben L. Rosenthal, Division of Otolaryngology, Department of Surgery, University of Alabama at Birmingham, BDB Suite 563, 1808 7th Avenue South, Birmingham, AL 35294-0012. Phone: 205-934-9767; Fax: 205-934-3993. E-mail: oto{at}uab.edu
Abstract
Physicians and surgeons rely on subtle tissue changes to detect the extent of tumors and the presence of residual disease in the clinical setting. The development of a cancer-specific fluorescent contrast agent has the potential to provide real-time tumor imaging in the clinic or operating room. Because epidermal growth factor receptor (EGFR) is highly overexpressed on the surface of head and neck squamous cell carcinoma (HNSCC), we sought to determine if fluorescently labeled anti-EGFR antibody could be used to image HNSCC xenografts in vivo. Cetuximab or control isotype-matched IgG1 was conjugated with the Cy5.5 fluorochrome and systemically injected into mice bearing human split thickness skin grafts, tumor cell line xenografts, transplanted human tumor xenografts, or mouse mesothelioma tumors. Xenografts were imaged by time-domain fluorescence imaging or fluorescence stereomicroscopy. Both imaging modalities detected specific uptake of cetuximab-Cy5.5 in HNSCC xenografts with significantly higher fluorescence levels relative to control IgG1-Cy5.5. Tumor xenograft fluorescence was higher compared with background (before injection), human split thickness skin grafts, or mouse mesothelioma tumors at 24, 48, and 72 h. Fluorescence was detected in multiple HNSCC tumor cell lines with variable EGFR expression levels. Mock resections of flank tumors using fluorescence stereomicroscopy showed that small (2 mm) specimens could be detected in the surgical wound bed. These results show the feasibility of using fluorescently labeled anti-EGFR antibody to detect human tumors in the surgical setting. [Mol Cancer Ther 2007;6(4):1230–8]
Introduction
Surgeons determine tumor margins intraoperatively by gross palpation, a technique that has not changed in the past 60 years. Unfortunately, this technique results in positive or close surgical margins in 
50% of cases (1, 2). Because positive margins predict poor survival, accurate identification of intraoperative margins may be of significant value in guiding the surgeon during ablative procedures. The high rate of positive margins in head and neck cancers is often due to the deeply infiltrative nature of the disease, and tumor cannot be well visualized during clinical assessment. Identification of an optical contrast agent that could provide the surgeon with real-time information about the tumor margins or the presence of residual disease may improve outcomes and reduce removal of uninvolved tissues. Although of insufficient tissue penetration for whole-body imaging, in the surgical setting, optical fluorescence may provide real-time diagnostic imaging that would allow clinicians to visualize the local extent of tumor spread or the presence of regional disease in the clinic or operating room. Fluorescent imaging has been widely applied in small animal models of cancer to accurately detect or monitor tumor growth (3).
Whereas magnetic resonance imaging and computed tomography can effectively image tumors preoperatively, optical imaging has the potential to provide real-time information about the extent of tumor, regional disease, or the presence of residual disease in the operating room or clinic. Previous optical imaging work has focused on the use of antibody-based fluorescent imaging to assess receptor or protein expression (4, 5), rather than guide therapy. Although the focus of medical imaging is often on the detection of early lesions or very late lesions (metastatic disease), there are only a limited number of studies to show the potential use of tumor-specific contrast agents being used for fluorescent imaging to guide oncological resections. Neurosurgeons have developed nonspecific fluorescent probes to assess tumor margins of glioblastomas intraoperatively (6, 7). Cancer-specific probes have been used for detection of gastrointestinal cancer (8, 9) and intrathoracic cancer (10, 11). These highly investigational techniques have not yet been applied to tumors of the upper digestive tract.
Epidermal growth factor receptor (EGFR) provides an excellent target for cancer-specific imaging in head and neck squamous cell carcinoma (HNSCC) because of its location on the cell surface and it is highly overexpressed, even compared with other cancers, and is up-regulated early in tumor progression (12, 13). Cetuximab is a monoclonal antibody directed against EGFR (14). Cetuximab selectively binds to the external domain for the EGFR with high affinity (Kd = 1 nmol/L; ref. 15) and has been effective in treating HNSCC when used alone or in combination with radiotherapy (16, 17). The transition to the clinic would certainly be facilitated by selecting cetuximab as an imaging agent because it is Food and Drug Administration approved as a therapeutic agent in HNSCC.
Accurate identification of cancer using cancer-specific fluorescent contrast agents has the potential to improve surgical outcomes. We hypothesize that fluorescently labeled anti-EGFR antibody can be used as a cancer-specific contrast agent to guide surgical therapy in HNSCC. To this end, we show that fluorescently labeled cetuximab can be used in vivo to detect HNSCC tumor xenografts, with minimal fluorescence in normal human skin xenografts.
Materials and Methods
Reagents
We used cetuximab (ImClone Systems, Inc., New York, NY), a recombinant, human/mouse chimeric monoclonal antibody that binds specifically to the extracellular domain of the human EGFR. Cetuximab is composed of the Fv regions of a murine anti-EGFR antibody with human IgG1 heavy and 
light chain constant regions and has an approximate molecular weight of 152 kDa. Human IgG1 antibody (Alexis Biochemicals, San Diego, CA) was used as a control antibody (molecular weight of 146 kDa). The second control antibody was human IgG1() (Southern Biotechnology Associates, Inc., Birmingham, AL) antibody (molecular weight of 144 kDa). Cy5.5 (CyDye deoxynucleotides, Amersham Biosciences, Piscataway, NJ) was used as the far-red fluorescent marker. Cy5.5 has a broad absorption peak with its maximum at 683 nm. Its emission maximum when coupled to IgG is at 707 nm, with a relative quantum yield of 0.28. Cy5.5 has a degree of labeling of 4.2 mol of dye per mole of protein using a 
of 250,000 mol/L–1cm–1 at the absorbance maximum.
Cell Lines and Tissue Culture
Four human cancer cell lines were used: UM-SCC-1 (Thomas Carey, University of Michigan, Ann Arbor, MI), FaDu, CAL 27, and AB12. The HNSCC lines (UM-SCC-1 and CAL 27) were maintained in DMEM containing 10% fetal bovine serum. All cell lines were supplemented with L-glutamine, penicillin, and streptomycin and incubated at 37°C in 5% CO2. CAL 27, a human tongue squamous cell carcinoma, was obtained from the American Type Culture Collection (Manassas, VA). CAL 27 cells were cultured and maintained in 90% DMEM + 10% fetal bovine serum. The mouse mesothelioma cell line AB12 was provided by Dr. Katri Selander (University of Alabama at Birmingham, Birmingham, Alabama). AB12 cells were cultured and maintained in complete medium consisting of high-glucose DMEM (Mediatech, Washington, DC) supplemented with 10% heat-inactivated FCS, 100 units/mL penicillin, 100 g/mL streptomycin, and 2 mmol/L glutamine. The green fluorescent protein (GFP) vector (pcDNA 3.1 vector, Invitrogen, Carlsbad, CA) was used to transfect the SCC-1 cell line using the calcium phosphate precipitation method (18). Under G418 selective conditions (concentrations, 0.1–0.8 mg/mL), resistant clones were screened by fluorescent appearance. Stably resistant clones were then selected from each cell population, propagated, and analyzed for GFP protein expression by fluorescence microscopy.
Animal Models
Severe combined immunodeficient (SCID) male mice, age 4 to 6 weeks (Charles River Laboratories, Wilmington, MA), were obtained and housed in accordance with the institution's Institutional Animal Care and Use Committee guidelines. Furthermore, all experiments were conducted and the animals were euthanized according to our institution's Institutional Animal Care and Use Committee guidelines. Because mouse dermis does not express human EGFR, waste split thickness skin graft was obtained from split thickness skin grafting procedures where excess skin graft tissue was available. Split thickness skin grafts were secured onto the left dorsal surface of SCID mice with absorbable suture after removal of a 1- to 2-cm2 section of dermis. Hair was removed before the procedure. Although viable graft could be determined on gross exam, histology was obtained to confirm viable skin. There was no evidence of hair-bearing skin to suggest murine dermal regrowth.
SCC-1, FaDu, or CAL 27 was used as a stringent model because SCC-1 has low EGF expression compared with other HNSCC cell lines. Cells (1 x 106) were injected into the left flank of SCID mice. In all experiments, the tumor size was <1.5 cm in greatest diameter during imaging. After 2 to 3 weeks of subdermal infiltration, mice were systemically injected with 50 µg human IgG1 antibody (Alexis Biochemicals) conjugated with Cy5.5 (IgG1a-Cy5.5) or human IgG1 antibody (Southern Biotechnology Associates) conjugated with Cy5.5 (IgG1b-Cy5.5) or with 50 µg cetuximab conjugated with Cy5.5 (cetuximab-Cy5.5). It is possible that extravasation of protein secondary to permeable vasculature or elevated blood flow in tumors (blood-borne fluorescence) may result in nonspecific sequestering of Cy5.5-cetuximab in the tumor. Human tumor explant model was developed using metastatic moderately differentiated cervical metastasis from a HNSCC. The tumor was passaged in SCID mice and tumors between passages 2 and 3 were used in this study. Histologic examination by a pathologist did not reveal any histologic change in the tumor through passage 5. Skin grafts are waste human skin graft tissue engrafted on the exposed muscle of SCID mice. For the orthotopic floor of mouth model, we injected mice under sterile technique. In a volume of 0.015 mL, 7.5 x 105 cells were injected transcervically into the floor of mouth. Use of this model in our laboratory has revealed that these tumors grow in the tongue musculature and floor mouth in an invasive manner based on serial histologic sections (19, 20).
A total of two animals for each group (IgG1k-Cy5.5 control or cetuximab-Cy5.5 conjugate) for each cell line were assessed. Mice underwent systemic tail vein administration of the conjugated antibody (50 µg) before optical imaging. Transcervical surgical resections were done after the mice were euthanized. Stereomicroscopic imaging was done before and after each partial and near-total surgical resection. After surgical resections, the mandible and associated structures were removed and paraffin embedded. Serial (bread loaf) histologic sections and routine H&E staining were then done to assess the presence and location of residual tumor.
Imaging
Cetuximab or human IgG1 isotype control was incubated with Cy5.5 Reactive Dye in 0.15 mol/L phosphate buffer (pH 7.8) for 1.5 h. The nonconjugated Cy5.5 was removed by Centricon Centrifugal Filter Unit, YM-30 (Millipore, Billerica, MA). All procedures were conducted under aseptic technique. To control for nonspecific uptake, IgG1 antibody that was isotype matched with cetuximab was conjugated with Cy5.5, and parallel imaging was done in animals after systemic injection of the IgG1-Cy5.5 conjugate.
The purpose of this study is to determine if fluorescent imaging can be used clinically to guide surgical therapy. To determine if cetuximab-Cy5.5 fluorescence can be used to accurately assess tumor dimensions in the clinical setting, the Cy5.5 fluorescent image size (based on observer size measurements on the captured images) was correlated with caliper measurements. SCC-1 cells were transfected with a vector carrying GFP and xenografted into the flank of SCID mice. GFP fluorescence was used to confirm the presence of the injected SCC-1 tumor before and after systemic injection of cetuximab (50 µg) labeled with Cy5.5.
Human tumor xenografts were imaged using a custom-built Leica fluorescent stereomicroscope (Leica MZFL3 stereo research microscope, Leica Microsystems, Bannockburn, IL) that was fitted with a GFP and Cy5.5 filter and an ORCA ER charge-coupled device camera (Hamamatsu, Bridgewater, NJ) to allow for real-time imaging of Cy5.5 fluorescence. GFP excitation was done between 450 and 490 nm, and emission was measured using a GFP filter (filter set 41023, GFP3, Leica Microsystems) between 500 and 550 nm. Cy5.5 filter (Chroma filter set 41023; Chroma Technology, Rockingham, VT) provided excitation between 630 and 670 nm, and emission was measured at 685 to 735 nm. Mice were measured with calipers before imaging. Measurements were taken of the length and width of the tumor at all time points. The purpose of these experiments was to determine if fluorescent tumor measurements could predict actual size (caliper measurements); therefore, size rather than intensity was measured. Dimensions of fluorescent images were obtained by observer measurements of captured images. The tumor area was determined by multiplying length and width and was recorded in square centimeters.
The mice tumor regions were also imaged with an eXplore Optix time-domain fluorescence imaging system (ART/GE Healthcare, Princeton, NJ) to detect Cy5.5 antibody in tumor. The system featured a 635-nm pulse laser for excitation, with a 670-nm bandpass filter on the emission side. The tumor regions were scanned with a 1-mm grid and integration time of 0.3 s per point, using a laser setting of 27.4 to 89 microwatts. Region of interest analyses of the tumors determined the average photon counts per millimeter of tumor, as well as fluorescence lifetime. Lifetime measurements were based on the entire tumor region of interest, using a single exponential model for data collected between 2 to 10 ns after the laser pulse. In a typical experiment, up to four mice were anesthetized and scanned at the same time.
Tumors and skin grafts were evaluated at 2, 6, 24, 48, and 72 h after administration of Cy5.5-cetuximab or Cy5.5-IgG1. Using the stereomicroscope, each tumor and skin graft was imaged at exposure times of 0.05, 0.1, 0.2, 0.4, and 0.8 s. Fluorescence captured by the stereomicroscope was quantified after digital capture using ImageJ.3 For the images, the mean fluorescence intensity of the tumor or skin graft was determined by outlining the tissue as a region of interest in ImageJ and subtracting the background fluorescence for the same region.
Statistical Analyses
The linear relationship between the GFP versus Cy5.5 measurements and caliper versus Cy5.5 measurements was computed as Spearman's correlation coefficient. The bias present between the measurements was expressed in relationship to the caliper measurements for both GFP and Cy5.5 and expressed as the mean difference and 95% confidence interval. Continuous variable means (e.g., fluorescence intensity; counts per square millimeter) were compared by two-tailed t test.
Confocal Microscopy
After the 72-h time point, images were collected, the mice were sacrificed, and the tumor and skin grafts were resected for confocal microscopy. A Leica DMIRBE Inverted Microscope was used in conjunction with a Leica SP1 Confocal Scanning System to image the tissue. A helium neon laser was used to excite Cy5.5-labeled tissue at 633 nm. Emission from the Cy5.5 was captured at a range of 647 to 800 nm. Leica confocal software was used to acquire the images and a two-dimensional maximum projection was created from optical slices of the tissue to produce a final image. The computer workstation used to take pictures is a Windows NT 4.0 Workstation.
Results
Cetuximab-Cy5.5 Detects SCC-1 Xenografts In vivo
For antibody-directed fluorescence to be clinically useful, measurement of antibody-mediated tumor fluorescence should be very similar to actual tumor size. To assess this, actual tumor size was determined by both caliper measurements and GFP fluorescence. Using GFP-expressing SCC-1 cells, SCC-1 xenografts showed Cy5.5 fluorescence measurements that corresponded to the gross measurement (Fig. 1
) and GFP fluorescent measurement. The Spearman's correlation coefficient was 0.93 (P = 0.0009) for the Cy5.5 versus the GFP measurements and 0.90 (P = 0.002) for the caliper versus the Cy5.5 measurements. Compared with the caliper measurements, both other methods gave nonstatistically significant lower mean measurements of –0.01 (95% confidence interval, –0.08–0.06) for GFP versus caliper and –0.02 (95% confidence interval, –0.12–0.07) for Cy5.5 versus caliper. The slightly lower means may be a result of the caliper measurements, including the skin thickness on either side of the tumor as part of the measurement.
|
100-fold over background. Note that the skin graft appears darker compared with both the previous measurement and to the surrounding tissue, this is likely a result of variable background noise and some surrounding wound response, respectively.
|
|
|
To determine if cetuximab-Cy5.5 was specific for human EGFR-expressing tumors, SCID mice bearing either mouse mesothelioma tumors or SCC-1 xenografts were assessed. There was no significant uptake in the AB12 tumor compared with the SCC-1 tumor cell line by fluorescent stereomicroscopy (Fig. 5 ).
|
50% of head and neck cancer tumor resections, providing evidence to support what most surgeons know that the tumor margins and residual disease can be hard to detect during cancer resections. To simulate the clinical applicability of using immunofluorescent tumor resections, a SCC-1 flank tumor underwent partial and full resection (Fig. 6
). The tumor showed increased fluorescence after removal of the cutaneous covering. After surgical exposure, tumors have a distinct edge in the xenograft model and a subtotal resection was done leaving a 2-mm tumor fragment of tumor. Subsequent imaging could be used to detect this residual tumor when reimaged by Cy5.5 fluorescent stereomicroscopy. After complete removal of the xenograft, fluorescent imaging did not reveal residual disease. In an orthotopic model of HNSCC, a near-total resection left behind residual fluorescence, which was confirmed to be tumor on histologic analysis of the left floor of mouth (Fig. 6).
|
The results of this study suggest that systemically administered fluorescently labeled anti-EGFR antibody can be used to image HNSCC xenografts in vivo. Tumor fluorescence as measured by time-domain fluorescence and fluorescence stereomicroscopy was significantly higher compared with human epithelial background (xenografted human split thickness skin grafts) and higher than nonhuman EGFR-expressing tumors (AB12 mesothelioma cell line). In a mock surgical resection, fluorescently labeled cetuximab could be used as a cancer-specific contrast agent to detect small fragments of after partial resection of xenografted tumor.
Because size evaluation of head and neck tumors is complicated by their location and variable appearance, we sought to determine if tumor size could accurately determine size in a preclinical model as a preliminary step to determining the potential to translate this technique into head and neck cancer patients. Our results suggest that fluorescence size in this preclinical model of tumor growth does correlate with tumor size as measured by calipers. Although the effectiveness of systemic administration of fluorescently labeled antibody has not been shown previously as a cancer-specific contrast agent in HNSCC, studies have explored the use of fluorescent imaging agents in the laboratory, and clinical setting optical imaging has been applied to detection of tumors using fluorescently labeled antibodies with success in Karposi's sarcoma and breast cancer (4, 21). In addition to the attachment of fluorescent probes, anti-EGFR antibodies have been conjugated to quantum dots and gold nanoparticles to identify premalignant lesions of the cervix using topical application (22, 23). Soukos et al. (24) have used EGFR-targeted agents to detect premalignant lesions in the hamster cheek model after i.v. administration. More recently, Gruss et al. (25) used topical application of far-red fluorescent dye conjugated to cetuximab to accurately demarcate abnormal from normal tissue in 500 µm thick multilayer tissue constructs. This study provides evidence that anti-EGFR–targeted imaging can be used to identify malignant tissue on histologic sections after in vitro treatment. Unlike these studies, we focus on the potential clinical application of this technology: (a) we show that fluorescence from engrafted human skin is relatively low compared with tumor fluorescence, (b) we use human tumor explants and tumor cell lines with variable EGFR expression to show wide application, and (c) we show the feasibility of detecting small fragments of tumor within a wound bed using a stereomicroscope similar to those used clinically (Fig. 6).
Fluorescent imaging and gamma camera imaging technologies have not been applied to the surgical setting to guide surgical resections of malignancies because of the nonspecificity of the probe for disease detection. Imaging of radioisotopes has been assessed previously and found to have poor resolution, nonspecific uptake, and toxic to iodine avid tissues (26, 27). Furthermore, the rapid half-life of radioisotopes adds significant complexities in manufacture and administration. Fluorescence has been applied recently to sentinel lymph node detection in gastrointestinal cancer (8) and intrathoracic cancer (9, 11). Neurosurgeons have developed fluorescent probes that are preferentially incorporated into tumor tissue but not normal brain. These probes have been successful in assessing tumor margins of glioblastomas intraoperatively (6). Given the recent success of these probes in neurosurgery, we sought to develop a cancer-specific detection in the surgical setting because it has not been addressed previously for other disease sites.
To assess the clinical relevance of tumor detection on epithelial surfaces, we used engrafted human skin as a control to measure expected background. Cetuximab is targeted against human EGFR; therefore, background levels in mice may not reflect that expected in humans. Although engrafted human mucosa would serve as a more appropriate control for head and neck cancer, technical difficulties make these experiments impractical. EGFR expression in the present experiments did not correlate with fluorescent intensity (counts per pixel). The failure of EGFR expression to positively correlate with imaging intensity may relate to different rates of EGFR turnover and incorporation of the fluorescently labeled antibody over the 48 to 72 h of peak fluorescence. Higher fluorescence in tumors with low EGRF expression might be due to increased internalization of the receptor that could result in higher levels of fluorescent probe within the tumor cells, despite lower overall expression. Alternatively, differences in vascular ingrowth between tumor cell lines may not allow equal permeability of the antibody within the tumor resulting in lower rates of binding.
To assess the specificity of Cy5.5-labeled cetuximab, we used fluorescently labeled IgG1 control antibodies. Nonspecific uptake of fluorescently labeled antibody, as a result of vascular permeability, changes in hydrostatic pressure or other vascular changes was not detected using IgG1 antibodies of similar molecular weight and class as cetuximab. Fluorescence detected after injection of IgG1-Cy5.5 peaked at 6 h and then rapidly dissipated. It is possible to speculate that this early fluorescence is a result of fluorescent antibody circulating within the vascular spaces of the tumor and is subsequently cleared from the circulation or more widely distributed in other compartments leading to less tumor-specific fluorescence.
Conclusion
Currently, surgeons rely on subtle tissue changes and frozen section for intraoperative assessment of tumor margins. Although the limited tissue penetration of fluorescent probes will limit their utility in whole-body detection of malignancy, fluorescent imaging could provide real-time information to direct surgical therapy. In the present study, xenografted HNSCC can be accurately detected after systemic injection of cetuximab-Cy5.5 using fluorescent stereomicroscopy.
Footnotes
Grant support: American Cancer Society grant RSG-06-1006-01-CCE (E.L. Rosenthal) and the University of Alabama at Birmingham Comprehensive Cancer Center core grant NIH, P30 CA13148.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 11/30/06; accepted 2/ 9/07.
References
vß3 integrin-targeted optical contrast agents with dynamic optical imaging. J Biomed Opt 2005;10:034019.[CrossRef][Medline] Asgari S, Rohrborn HJ, Engelhorn T, Stolke D. Intra-operative characterization of gliomas by near-infrared spectroscopy: possible association with prognosis. Acta Neurochir (Wien) 2003;145:453–9; discussion 459–60.[Medline] Kircher MF, Mahmood U, King RS, Weissleder R, Josephson L. A multimodal nanoparticle for preoperative magnetic resonance imaging and intraoperative optical brain tumor delineation. Cancer Res 2003;63:8122–5.This article has been cited by other articles:
|
|
 |
|
  D. A. Mankoff, J. M. Link, H. M. Linden, L. Sundararajan, and K. A. Krohn Tumor Receptor Imaging J. Nucl. Med., June 1, 2008; 49(Suppl_2): 149S - 163S. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
