| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Molecular Imaging |
1 From the Center for Molecular Imaging Research, Massachusetts General Hospital, Harvard Medical School, Bldg 149, 13th St, Room 5408, Charlestown, MA 02129. From the 2004 RSNA Annual Meeting. Received December 23, 2005; revision requested February 21, 2006; revision received June 19; accepted July 11; final version accepted November 1. Supported in part by National Institutes of Health grants RO1-EB001872, R24-CA92782, and by a grant from the Dana-Farber/Harvard Cancer Center Technology Innovation Fund. Address correspondence to U.M. (e-mail: mahmood{at}helix.mgh.harvard.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
Materials and Methods: The institutional animal care committee approved all animal experiments. CT26 tumor cells were orthotopically implanted into the descending colon of C57BL6/J mice (n = 10). Thirteen days later, mice intravenously received either 2 nmol of a protease-sensing probe that had cathepsin B as a major activator (n = 5) or saline (control animals [n = 5]). One day later, animals were noninvasively examined to the point of the splenic flexure by using microcatheter imaging. Excised colons were subsequently evaluated with epifluorescence imaging, histologic examination, and cathepsin B immunohistochemistry. Student t test was used for statistical analysis, with P < .05 considered to indicate a significant difference.
Results: Results with fiberoptic imaging demonstrated that all tumors were visible with the protease-activatable probe, even when they were not readily apparent at white light imaging. A target-to-background ratio (TBR) of 8.86 for tumor to adjacent normal mucosa was achieved in the NIR channel after probe administration (P = .001), whereas white light images resulted in a TBR of 1.14 (P > .5) based on luminosity. The tumoral NIR fluorescence intensity was more than 30-fold greater in probe-injected animals than in control animals, indicating that essentially all of the signal recorded in lesions was from activatable probe administration. Results of immunohistochemistry confirmed cathepsin B overexpression in the tumor compared with adjacent mucosa.
Conclusion: The use of NIR imaging microcatheters combined with protease-activatable smart probes results in a beacon effect that highlights tumors with high TBRs; this technique thus may be a potentially useful adjunct to white light colonoscopy in the future.
© RSNA, 2007
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
Proteases have been shown to be overexpressed in a number of cancers, including colon cancer (12,13). In particular, cathepsins, including cathepsin B, are often overexpressed in human colonic neoplasms (14,15). Results of a previous study (16) demonstrated the ability of a protease-sensitive probe to improve detection of adenomatous polyps in the small bowel after resection and flushing in an animal model. To our knowledge, there has been no demonstration of the feasibility of using minimally invasive imaging catheters that can simultaneously and independently depict full-spectrum white light and near-infrared (NIR) fluorescence (17) in combination with an NIR optical "smart" probe (18) that is activated by proteases, such as cathepsin B, for noninvasive early disease detection of colonic neoplastic lesions. Thus, the purpose of our study was to prospectively evaluate the ability of micro-fiberoptic catheters, which simultaneously record white light and NIR images, to reveal colonic neoplasms after the intravenous administration of activatable smart probes that increase their NIR fluorescence subsequent to protease activation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
Cell Line
CT26 murine colon cancer cells were purchased from a supplier (American Type Culture Collection, Manassas, Va) and cultured according to standard protocols in RPMI 1640 medium with 2 mmol/L L-glutamine (Invitrogen, Carlsbad, Calif) adjusted to contain 1.5 g/L sodium bicarbonate, 4.5 g/L glucose, 10 mmol/L HEPES, 1.0 mmol/L sodium pyruvate, and 10% fetal bovine serum. Once the cells reached 90% confluence, they were trypsinized and suspended in Hanks balanced salt solution (Mediatech, Herndon, Va) immediately before tumor implantation.
Orthotopic Colon Cancer Model
The institutional animal care committee approved all animal experiments. Ten C57BL6/J mice (Jackson Laboratory, Bar Harbor, Me) underwent surgery, with gas anesthesia (2% isoflurane [Baxter, Deerfield, Ill] in O2 2 L/min) delivered by face mask. Surgery was performed by using a previously described technique for orthotopic tumor implantation (19). Two authors (H.A., J.F.) performed the implantation. Briefly, a midline incision was created, a nontraumatic clamp (Fine Science Tools, Foster City, Calif) was positioned on the descending colon, and a polyethylene catheter was inserted rectally. The isolated portion of the colon (distal to the clamp) was washed twice with 500 µL of phosphate-buffered saline (PBS) to clean out the bowel contents.
With the catheter still in place, a second clamp was applied to the colon 1 cm distal to the first one. This second clamp encompassed both the colon wall and the catheter, creating a 1-cm-long closed bowel loop. The clamp had a calibrated pressure of 10 g/mm2, which is sufficient for creating the closed loop but which does not damage the bowel wall or cause collapse of the catheter. One hundred microliters of 0.05% trypsin in edetic acid (Mediatech) was injected into the bowel loop, the catheter was removed, and the trypsin was left in place for 30 minutes to disrupt the mucosal layer only. The distal clamp was then removed, the catheter was reinserted, and the colon was flushed twice with 500 µL PBS. Subsequently, the distal clamp was repositioned, and 1 x 106 CT26 cells suspended in 100 µL Hanks balanced salt solution were injected into the colon lumen. The catheter was removed, and cells were allowed 30 minutes for implantation. Both clamps were then removed, and the abdominal wall was closed in a two-layer fashion. A heated surgery table was used in all cases, and the exposed bowel loops were kept moist by irrigation with warm saline solution.
On the basis of results of previous tumor growth studies in this model, all animals were imaged 14 days after CT26 tumor cell implantation. The total surgery time for the tumor implantation was approximately 80 minutes per mouse, and there were no deaths related to the surgical procedure. All mice appeared healthy 14 days after the surgery.
NIR Activatable Probe
The smart probe used in our study represents a class of optical imaging agents that have their fluorescent emission effectively inhibited in the native state by fluorescence resonance energy transfer caused by the proximity of the fluorochromes to one another. The quenched protease-activatable NIR fluorescent probe we used (Prosense 680; VisEn Medical) is similar to a previously described probe (18) in that it is a synthetic graft copolymer consisting of poly-L-lysine that is sterically protected by multiple methoxypolyethylene glycol side chains and to which multiple fluorochromes are attached. The probes are dequenched by a number of biologically relevant proteases that cleave lysine-lysine bonds and result in signal intensity increases of the imaging probe of 15–30 fold. In particular, cathepsin B has been demonstrated in vivo to be a major contributor to cleavage and activation (18). The imaging probe has a peak absorption of approximately 680 nm and a peak emission of 700 nm. Thirteen days after implantation and 24 hours before imaging, five mice (group 1) were injected intravenously with 150 µL of the protease-activatable probe (2 nmol of fluorochrome). The remaining five mice (group 2, the control group) were injected with equivalent volume of saline solution. Each mouse was randomly assigned to each group.
Fluorescent Endoscopy
The imaging catheter we used has been described elsewhere (17). Briefly, the catheter has an outer diameter of 0.8 mm and consists of 10 000 ordered fibers surrounded by a group of 14 bundles of illumination fibers, resulting in an imaging matrix of 100 x 100. Visible light is separated from NIR light through a 670-nm dichroic mirror. The NIR fluorescent light is also passed through an emission filter with a 690–800-nm six-cavity band-pass design (Omega Optical, Brattleboro, Vt) to completely eliminate any remaining excitation light in the fluorescent image. The two components are recorded separately but simultaneously with white light or NIR video cameras (StellaCam EX; Adirondack Video, Hudson Falls, NY), allowing display and video capture of full color spectrum white light images adjacent to NIR images, which, on the basis of probe fluorescence, represent protease tissue activity. Excitation light was provided by a mercury vapor lamp (SUV-DC; Lumatec, Munich, Germany).
Two authors (H.A., M.A.F.) performed the endoscopy procedures. During imaging, the catheter was lubricated with water and introduced rectally into mice anesthetized with 2% isoflurane in O2 2 L/min. The colon was gently insufflated with air while keeping the mean pressure to less than 10 mm Hg to avoid overinsufflation of the entire bowel, which could lead to perforation or regurgitation of fluid through the esophagus. As the catheter was gently advanced into the colon, the abdomen was observed, both to localize the tip of the catheter with transillumination and to monitor the colon for proper insufflation. Images from both cameras were recorded simultaneously and in real time on a personal computer (Dell, Round Rock, Tex) by using custom software created in house. The catheter-based imaging session was easily performed in both animal groups. The average length of catheter insertion was 4 cm, and each examination required 10 to 15 minutes to perform. There were no deaths or complications. In both animal groups, white light and NIR images were acquired simultaneously and in real time.
Epifluorescence Imaging
To enable quantitation of fluorescence intensity in tumors and normal mucosa, epifluorescence images of the explanted colon were acquired immediately after in vivo studies were completed. At the completion of in vivo imaging, animals were euthanized and the descending colon was removed, longitudinally split open, and imaged by using an epifluorescence system (bonSAI; Siemens Medical Solutions, Malvern, Pa) that is based on a custom-built design (20,21). The system consists of a 150-W halogen excitation light source connected to an acquisition box through an optical wave guide. The built-in filter wheel was set to deliver light at 400–745 nm and 25% intensity for white light images and at 660 nm ± 15 and 100% intensity for NIR images. On the detection side, a second filter wheel used a neutral density filter with an optical density of 4 for white light imaging and a 735 HW 30 filter for NIR imaging. A charge-coupled device camera with a matrix size of 1360 x 1024 pixels and a resolution of 0.116 mm/pixel was used for image acquisition. System control and data storage were performed on a personal computer with software (Syngo; Siemens Medical Solutions, Erlangen, Germany). Exposure time was 0.5 second for all images, and the data were stored in Digital Imaging and Communications in Medicine format for later analysis.
Histologic and Immunohistochemical Analysis
The explanted descending colon was embedded in freezing media and frozen, and 10-µm-thick slices were obtained by using a cryostat (Leica, Bannockburn, Ill). Samples were fixed in 4% paraformaldehyde and stained with hematoxylin-eosin for tumor identification. Immunohistochemical analysis for cathepsin B was performed on adjacent slices by one author (H.S.). For immunohistochemical analysis, the tissue slices were fixed with acetone at 4°C and blocked with 10% goat serum (Invitrogen) for 20 minutes. After blocking was completed, the slides were incubated with avidin-biotin blocking solution (Vector Laboratories, Burlingame, Calif) for 30 minutes. Tissue slices were then incubated with 1 µg/mL rabbit polyclonal immunoglobulin G anti–mouse cathepsin B (Upstate, Lake Placid, NY) diluted in 10 times the baseline concentration of PBS diluted to the normal saline concentration (hereafter called 1x PBS) for 1 hour at room temperature.
After two consecutive washes with 1x PBS and 1x PBS 25% washing fluid (NaCl, KCl, Na2HPO4, KH2PO4, sodium azide, and gelatin), tissue slices were incubated with biotin-labeled secondary antibody (goat anti–rabbit immunoglobulin G [Upstate]) at 1:200 dilution in 2% mouse serum (Jackson ImmunoResearch Laboratories, West Grove, Pa) for 40 minutes at room temperature. After several washes with 1x PBS and 1x PBS 25% washing fluid, the endogenous peroxidase was blocked with 0.3% H2O2-PBS for 15 minutes at room temperature, followed by incubation with peroxidase-conjugated avidin-biotin for 40 minutes (Vectastain Elite ABC Kit Standard; Vector Laboratories). The peroxidase reaction was performed by using Vector NovaRED substrate (Vector Laboratories), and counterstaining was performed with hematoxylin for 30 seconds. As a negative control, the primary antibody was omitted for each immunohistochemical staining.
Statistical Analysis
Epifluorescence images were used for signal intensity calculations. Circular fixed region of interest measurements (diameter, 10 pixels) were obtained by the same author (H.A.) in the tumor, in adjacent normal mucosa (at least two region of interest diameters away from the tumor), and in background. Data are presented as means ± standard errors of the mean, and the signal intensity value represents either tumor signal intensity minus background signal intensity (intrinsic camera noise) or normal mucosa signal intensity minus background signal intensity. The significance of differences between signal intensity values was determined by using the Student t test. P < .05 was considered to indicate a significant difference. Analysis was performed using by using software (KaleidaGraph, version 3.6, 2003; Synergy Software, Reading, Pa).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
Group 1.—In the probe-injected group, all tumors were detected by combining the anatomic information provided by the full color spectrum white light channel with the fluorescent signal generated by protease probe activation in the NIR channel (Fig 1). When tumors were present, bright fluorescent signal was detected in all cases. The NIR signal enabled the identification of smaller, flat lesions and confirmed the anatomic findings in the white light channel when larger tumors were present.
|
Epifluorescence
Figure 2 shows white light and NIR fluorescence images in a probe-injected animal and a control animal. The average target-to-background ratio (TBR) between tumor and adjacent normal mucosa in the NIR channel was 8.86 (P = .001) in the probe-injected group and 1.56 (P > .10, not significant) in the control group (Fig 3a). The TBR for the white light images was 1.14 for the probe-injected group (P > .5, not significant) (Fig 3b). The increase in NIR signal intensity due to protease-based probe activation confirmed the location of tumors identified on the white light images and enabled identification of lesions that were not clearly seen in the white light channel. Moreover, the increase in NIR signal intensity from 91 arbitrary units (au) to more than 3000 au in tumors in animals that received intravenous administration of the agent demonstrated that more than 96% of the fluorescence signal derived from exogenously administered probe and that minor differences in tissue autofluorescence essentially did not alter the protease map that was generated. Figure 4 confirms cathepsin B overexpression in the tumor type evaluated in this study.
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
The high TBR that was a result of quenched fluorescence signal from the circulating nonactivated probe and the marked differences in protease expression between the colonic tumors and normal mucosa provided a beacon effect in the NIR channel. The TBR after probe injection in the NIR channel was approximately 9:1, compared with a TBR of approximately 1:1 with white light imaging. Thus, large areas of colonic mucosa may be rapidly evaluated for small, anatomically unapparent abnormalities. Areas of concern are brightly highlighted, potentially reducing the risk of missed lesions. A nonsignificant difference was noted between native tumor NIR fluorescence and normal mucosa NIR fluorescence in the control group, with a TBR of 1.56. This small difference, combined with the wide variation seen in the native fluorescence signal intensity of the control group, precludes the use of autofluorescence in this part of the spectrum for reliable detection of lesions.
The current standard of care (22,23), normal magnification white light colonoscopy, leaves room for improvement. Results of one study (3) showed an overall miss rate of 24% with standard white light colonoscopy. Another study in which histologic findings from resected colon segments were used as a reference standard to evaluate the detection rate at endoscopy likewise revealed a sensitivity of 77% (24). A study of 1000 consecutive patients (4) revealed that many flat lesions, which may be more readily missed compared with polypoid lesions when anatomic methods are used, contained areas of dysplasia. In that study, 54% of lesions that had areas of severe dysplasia or Dukes A carcinoma were flat, and 36% of all adenomas found were flat. A molecular imaging adjunct that could highlight adenocarcinomas would be especially useful to call attention to such areas.
Experimental anatomic visualization techniques that rely on magnification, with or without the aid of chromophores such as indigo carmine, suffer from the fact that full evaluation of the entire area of the colonic wall with magnification requires substantial time; moreover, detection is based on a combination of the appearance of subtle anatomic distortions combined with visual differences in topically applied dye localization (25). An area of concern may have a similar luminosity to normal areas. Thus, magnification chromoendoscopy is often limited to more closely examining focal areas during resection (26) or to screening patients at very high risk for cancer, such as those with familial polyposis syndromes (27).
In the current set of experiments, we showed a high target-to-background effect in adenocarcinomas implanted in the descending colon of mice with the use of NIR protease-activatable probes. The sample size for each group in this study, as in many murine evaluations, was small. This would be of concern if we had demonstrated no difference between fluorescence intensity of tumors and normal mucosa after probe injection, because small group sizes confer an increased risk of type II errors. However, small sample sizes do not increase the risk of type I errors compared with larger sample sizes. Human tumors have a more diverse genetic distribution than the single cell line evaluated in this study. Although previous studies (14,15) have revealed an overexpression of proteases in colon cancer, rigorous gene expression profiling to evaluate the prevalence and, especially, the degree of overexpression of the panel of proteases that activate the smart probes used in the current study would provide an understanding of what percentage of human tumors would be enhanced sufficiently to be visualized with such an approach. However, the adjunct role of this method would not be changed by such outcomes. Additionally, further studies would be helpful in evaluating the specificity of tumoral fluorescence enhancement in the setting of other disease, such as in models of ulcerative colitis.
Practical applications: The experimental methods described here are potentially clinically translatable (28), with respect to both smart activatable probes and imaging devices. The microcatheters used in this study can fit through the working channel of many clinical endoscopes. Alternatively, because catheter-based imaging in mice faces more technical constraints than potential human NIR fluorescence endoscopy, given the relative ratios in instrument diameters and rigidity, direct scaling of the entire design is possible. Such imaging would potentially serve as an adjunct to white light colonoscopy, which is not altered but supplemented by the use of the NIR channel and NIR probes. The bright fluorescence from colonic tumors has the potential to both decrease the miss rate of neoplastic colonic lesions and increase the speed of colonic examinations in the future.
|
ADVANCE IN KNOWLEDGE |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviations: NIR = near infrared • PBS = phosphate-buffered saline • TBR = target-to-background ratio
See Materials and Methods for pertinent disclosures.
Author contributions: Guarantors of integrity of entire study, R.W., U.M.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, H.A., M.A.F., J.F., H.S.; experimental studies, H.A., J.F., H.S.; statistical analysis, H.A., M.A.F., J.F., H.S., U.M.; and manuscript editing, R.W., U.M.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Ogawa, C. A.S. Regino, P. L. Choyke, and H. Kobayashi In vivo target-specific activatable near-infrared optical labeling of humanized monoclonal antibodies Mol. Cancer Ther., January 1, 2009; 8(1): 232 - 239. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. V. Frangioni New Technologies for Human Cancer Imaging J. Clin. Oncol., August 20, 2008; 26(24): 4012 - 4021. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
