Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/radiol.2423052065
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Montet, X.
Right arrow Articles by Weissleder, R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Montet, X.
Right arrow Articles by Weissleder, R.
(Radiology 2007;242:751-758.)
© RSNA, 2007

Experimental Studies

Tomographic Fluorescence Imaging of Tumor Vascular Volume in Mice1

Xavier Montet, MD, Jose-Luiz Figueiredo, MD, Herlen Alencar, MD, Vasilis Ntziachristos, PhD, Umar Mahmood, MD, PhD and Ralph Weissleder, MD, PhD

1 From the Center for Molecular Imaging Research, Massachusetts General Hospital and Harvard Medical School, Building 149, 13th St, Room 5403, Charlestown, MA 02129. Received December 19, 2005; revision requested February 10, 2006; revision received March 22; accepted April 19; final version accepted July 6. R.W. supported in part by P50 CA86355, P01 CA69246, R24 CA92782, and R33 CA91807 grant. X.M. supported by the Swiss National Science Foundation. Address correspondence to R.W. (e-mail: weissleder{at}helix.mgh.harvard.edu).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 References
 
Purpose: To prospectively determine the feasibility of imaging vascular volume fraction (VVF) and its therapeutic inhibition in mouse models of cancer with three-dimensional fluorescence molecular tomography (FMT).

Materials and Methods: All studies were approved by the institutional animal review committee and were in accordance with National Institutes of Health guidelines. CT26 colon tumor–bearing mice were imaged with FMT after intravenous administration of long-circulating near-infrared fluorescent blood-pool agents optimized for two nonoverlapping excitation wavelengths (680 and 750 nm). A total of 58 mice were used for imaging VVF to evaluate the following: (a) differences in ectopically and orthotopically implanted tumors (n = 10), (b) cohorts of mice (n = 24) treated with anti–vascular endothelial growth factor (VEGF) antibody, (c) serial imaging in same animal to determine natural course of angiogenesis (n = 4), and (d) dose response to anti-VEGF therapy (n = 20). To compare groups receiving antiangiogenic chemotherapy, analysis of variance was used.

Results: Fluorochrome concentrations derived from FMT measurements were reconstructed with an accuracy of ±10% at 680 nm and ±7% at 750 nm and in a depth-independent manner, unlike at reflectance imaging. FMT measurements of vascular fluorescent probes were linear, with concentration over several orders of magnitude (r > 0.98). VVFs of colonic tumors, which varied considerably among animals (3.5% ± 1.5 [standard deviation]), could be depicted with in vivo imaging in three dimensions with less than 5 minutes of imaging and less than 3 minutes of analysis. The natural course of angiogenesis and its inhibition could be reliably imaged and depicted serially in different experimental setups.

Conclusion: FMT is a tomographic optical imaging technique that, in conjunction with appropriate fluorescent probes, allows quantitative visualization of biologic processes.

© RSNA, 2007


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 References
 
Angiogenesis is a key biologic process underlying many clinically relevant diseases, such as cancer (1,2), atherosclerosis (3,4), and autoimmune diseases (5). Given its key role in supplying nutrients and immune cells to these disease areas, angiogenesis has received widespread attention as a "drugable" process. A number of successful angiogenic inhibitors are in clinical use (eg, bevacizumab, AE-941, thalidomide), while a myriad of other agents are in developmental stages. Imaging continues to play a role in testing the effectiveness of antiangiogenic agents in preclinical trials and in clinical practice (6,7). One reason for using imaging as an objective end point has been the high cost associated with recent cytostatic treatments (8).

Different imaging approaches and imaging agents have been used to measure angiogenic processes. The diverse techniques have relied on either kinetic measurements of conventional contrast agents (9), steady-state measurements of long-circulating imaging agents (10,11), or endothelial-targeted agents (12–14). More recently, genetically engineered mice that express green fluorescent protein (gfp) or luciferase in endothelial cells (or tumor cells) have become available and thus allow optical measurements (15,16). Developed clinical techniques have variable degrees of accuracy and are often not robust enough to measure subtle drug effects. Recent experimental techniques used in drug development are often time-consuming or computationally intensive, which limits throughput; and, in some cases, recent experimental techniques are not translatable into clinical practice (gfp or luciferase imaging).

We have previously used quantitative near-infrared optical tomography to measure fluorochrome concentrations in deep tissues (17–19). More recently, we have also validated the use of fluorescence molecular tomography (FMT) in the measurement of angiogenesis by using magnetofluorescent nanoparticles and compared its performance with the more time-consuming magnetic resonance (MR) imaging and nuclear medicine techniques (20). Because optical measurements can be obtained at multiple wavelengths in the near-infrared spectrum, FMT is particularly useful for serial and temporal imaging in the steady state. Thus, the goal of our study was to prospectively determine the feasibility of imaging vascular volume fraction (VVF) and its therapeutic inhibition in mouse models of cancer by using three-dimensional FMT (21).


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 References
 
All studies were approved by the institutional animal review committee and were in accordance with National Institutes of Health guidelines.

One author (R.W.) is a founder of and one author (U.M.) is a consultant to VisEn Medical, Woburn, Mass. VisEn Medical did not provide any support for this study and had no influence over study design or outcome. One author (X.M.) had control of inclusion of any data and information that might present a conflict of interest.

Imaging System and Probes
FMT experiments were performed by using a commercially available imaging system (FMT-Solaris; VisEn Medical). The system allows data acquisition at different sets of wavelengths (680 nm for excitation and 700 nm for emission; 750 nm for excitation and 780 nm for emission). Images were acquired at both sets of wavelengths in phantoms and anesthetized mice. Briefly, objects were positioned in the imaging chamber and surrounded by fluid composed of 1% fat emulsion solution (Intralipid; Fresenius, Bad Homburg, Germany) and 0.5% India ink; this fluid closely matched the optical properties of tissues. Image data sets were reconstructed by using a normalized Born forward model adapted to small mouse models (18,19). Details of the algorithm have been published (19). Image acquisition time per animal was 3–5 minutes, and reconstruction time was 1–3 minutes. Reconstructions were performed on a personal computer (3.6-GHz Pentium 4 processor; Dell, Round Rock, Tex) by using FMT-Solaris 3.0 software. Images were displayed as raw data sets (excitation, emission, masks) and as reconstructed three-dimensional data sets in transverse, sagittal, and coronal planes. Fluorochrome concentration in the target was automatically calculated from reconstructed images and expressed as femtomols of fluorochrome per defined target volume. With a known blood concentration of the tracers, tumoral VVFs were calculated as VVF = (Ct/Cb) · 100, where Ct is the fluorochrome concentration in the tumor and Cb is the fluorochrome concentration in blood.

The vascular imaging probes used in this study were AngioSense 680 and AngioSense 750 (VisEn Medical). Probes were injected intravenously at 2 nmol fluorochrome per mouse, which corresponded to 150 µL per mouse. Both probes are high-molecular-weight (250 kDa) pegylated graft copolymers with an indocyanine-type fluorophore optimized for nonquenching. The probes are nonimmunogenic because of pegylation and isotonia (290 mOsm/L). To demonstrate the in vivo behavior of both probes, intravital microscopy (X.M., H.A.) was performed with a prototype laser scanning microscope (0V100; Olympus, Tokyo, Japan). The system was equipped with four lasers (488 nm for gfp and/or fluorescein isothiocyanate, 561 nm for rhodamine, 633 nm for indocyanine 5.5, and 748 nm for Alexa Fluor 750 imaging) and custom-built heated stages for intravital imaging (22). The 633- and 748-nm lasers were used to excite AngioSense 680 and AngioSense 750, respectively. Tumor-bearing mice, coinjected with both imaging probes, underwent intravital microscopy of tumor neovasculature. The use of a thin objective with a millimeter footprint allowed imaging of the neovessels of the tumor through a keyhole incision (22).

Cell Culture and Tumor Models
Tumor cell lines used in this study were obtained from the American Tissue Culture Collection (Manassas, Va). CT26 is a mouse colon carcinoma cell line. It was cultured in Roswell Park Memorial Institute medium with 2 mmol/L L-glutamine, 1.5 g/L sodium bicarbonate, 4.5 g/L glucose, 10 mmol/L HEPES, 1.0 mmol/L sodium pyruvate, and 10% heat-inactivated fetal bovine serum at 37°C in a humidified 5% CO2 atmosphere. Two million cells were implanted (X.M., J.L.F.) into the lateral part of the lower limb in athymic female nude mice (Charles River Laboratories, Wilmington, Mass) that weighed 25–30 g. Alternatively, tumor cells were implanted orthotopically in the colon by using a previously described model (23). All in vivo experiments were performed in mice anesthesized with 80 mg of ketamine (Ketalar; Pfizer, New York, NY) per kilogram of body weight or 12 mg/kg xylazine (Rompun; Bayer, Leverkusen, Germany) intraperitoneally or with gas (isoflurane 2% [Isofluran; Baxter, Deerfield, Ill]).

In Vitro Imaging
To compare the relationship between fluorescence signal and imaging depth, we filled a capillary tube (1 mm in diameter) with 5 fmol of AngioSense 680 and/or AngioSense 750 and sequentially imaged it with reflectance fluorescence and FMT at different depths (0–8 mm) (X.M.). To study the dual-wavelength capabilities of the FMT system, similar phantom experiments (n = 3) were performed at variable AngioSense concentrations and mixtures, all of which were suspended in opaque tissue-simulating matching fluid.

In Vivo Imaging
All in vivo imaging was performed by using nude mice bearing CT26 tumors (colon adenocarcinoma). Tumors had been implanted 5–7 days prior to imaging (in a total of 58 animals). After such time, the tumors were small (3–5 mm in diameter) and had not developed central necrosis. Immediately prior to imaging, gas-anesthetized animals received an intravenous injection of AngioSense 680 and/or 750 (2 nmol fluorochrome per mouse) via the tail vein. A number of different FMT imaging experiments were performed (X.M., J.L.F.) after intravital microscopic validation studies (X.M., H.A.).

Orthotopic versus ectopic tumors.—This experiment was performed to determine the feasibility of FMT imaging of orthotopically implanted colonic tumors and ectopically implanted flank tumors in live mice and to compare VVFs of tumors. Ten animals (five with ectopic tumors, five with orthotopic tumors) were used for this study. Presence of orthotopic tumors was confirmed with MR imaging (4.7-T Bruker; Pharmascan, Karlsruhe, Germany) by using a T2-weighted sequence (repetition time msec/echo time msec, 2500/38.2; section thickness, 1 mm), as well as a T1-weighted sequence (700/14.2; section thickness, 1 mm), after injection of gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany).

Treatment cohorts.—To determine whether treatment effects could be determined in cohorts of treated mice, three experimental groups (of eight mice each) with subcutaneous tumors were investigated (for a total of 24 mice). The first group received intraperitoneal injections of anti–vascular endothelial growth factor (VEGF) antibody (bevacizumab, Avastin; Genentech, San Francisco, Calif) (5 mg/kg) twice weekly and was imaged after 1 week of treatment. The second group received intraperitoneal injection of phosphate-buffered saline. The third group of mice did not receive any treatment and was imaged at identical time points.

Serial imaging in same animal.—To determine the feasibility of repeated imaging in the same animal (n = 4), serial imaging studies were performed by alternating AngioSense 680 and AngioSense 750 between imaging sessions. Imaging sessions were performed 5 days after tumor implantation (baseline), after 5 more days to allow the tumor to grow, and after 1 week of treatment with an anti-VEGF antibody (bevacizumab) (5 mg/kg) (inhibition).

Therapeutic dose response.—To study the angiogenic response to different doses of anti-VEGF therapy (bevacizumab), we examined 20 mice. Mice with 5-day-old tumors were imaged after intravenous injection of AngioSense 680 (2 nmol) to obtain baseline VVF and then were given different doses of anti-VEGF twice weekly with intraperitoneal injection (0, 1, 3, 5, and 7 mg/kg; each dose was given to four mice). After 1 week of treatment, animals were reimaged after intravenous injection of AngioSense 750 (2 nmol).

Histologic Examination
For the treatment cohort group, after in vivo imaging, colonic tumors were removed, fast frozen, and cut into 5-µm slices (X.M.). To identify microvessels, we stained slices with a primary anti-CD31 antibody (Santa Cruz Biotechnology, Santa Cruz, Calif). Microvessels were revealed with a biotinylated secondary antibody (Abcam, Cambridge, Mass) and developed for 12 minutes by using EAC substrate (Dako, Carpinteria, Calif). Microvessel density was determined automatically by using software (Image J, version 1.32j; National Institutes of Health, Bethesda, Md) as previously described (24). For this procedure, a cutoff level for positive immunohistochemical staining was defined. The counting fields were selected by finding neovascular hot spots. In these areas of highest vascularity, a computerized analysis of the CD31-positive endothelial area was performed under the microscope with a magnification of x200. The structures selected by Image J as microvessels were verified by the generation of a microvessel map. The analysis was repeated in at least five regions of the tumor. The number of microvessels counted corresponded to the mean vessel density. Control sections were processed identically, with the exception of omitting the incubation with the primary antibody. Hematoxylin-eosin staining was performed routinely for all groups.

Statistical Analysis
Statistical analysis was performed with software (Prism, version 4a, 2003; GraphPad Software, San Diego, Calif). Data are reported as means ± standard deviations. A significant difference was determined by using the Student t test. P values less than .05 were considered to indicate a significant difference.

Linear regression between known concentrations and FMT-reconstructed values was used to demonstrate the adequacy of FMT at measuring absolute concentration of fluorophores.

To compare the group receiving antiangiogenic chemotherapy (bevacizumab) with the control group receiving phosphate-buffered saline, analysis of variance was used. When P < .05 was found, a Bonferroni correction was performed.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 ADVANCES IN KNOWLEDGE
 References
 
In Vitro Imaging
Initial phantom experiments were performed to determine the effect of target depth on fluorochrome quantization and to obtain correlation curves for the different imaging probes. FMT-derived fluorochrome quantization was depth independent (0–8 mm), with a variability of ±10% for the 680 channel and ±7% for the 750 channel. In contradistinction, reflectance imaging–derived signals decreased exponentially with depth (Fig 1). Results of phantom experiments (Fig 2a–2c) showed linear correlation between FMT signal and probe concentration over several orders of magnitude when the probes were measured alone or in mixtures (r = 0.98, P < .003 for AngioSense 680; r = 0.99, P < .001 for AngioSense 750; r = 0.99, P < .02 for a mixture of AngioSense 680 and AngioSense 750).


Figure 1A
View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a: Validation of FMT imaging data. (a) Graphs of comparison of FMT and reflectance (FRI) data sets as function of imaging depth. Note that FMT-calculated fluorochrome concentrations are linear with increasing imaging depth, whereas there is exponential decrease with planar reflectance fluorescence imaging. Data in 750-nm channel are tighter (±7% variation) than those in 680-nm channel (±10% variation). (b) Photomicrographs of validation of two near-infrared vascular probes with confocal in vivo microscopy in live animals bearing CT26 tumors. Tumor microvasculature is chaotic—a typical feature. Note that imaging probes are truly intravascular (at least for 20 minutes after intravenous administration) and that there is concordance between the two probes. Note absence of bleed-through when only one channel is turned on. Scale bar = 100 µm, recon = reconstruction.

 

Figure 1B
View larger version (23K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b: Validation of FMT imaging data. (a) Graphs of comparison of FMT and reflectance (FRI) data sets as function of imaging depth. Note that FMT-calculated fluorochrome concentrations are linear with increasing imaging depth, whereas there is exponential decrease with planar reflectance fluorescence imaging. Data in 750-nm channel are tighter (±7% variation) than those in 680-nm channel (±10% variation). (b) Photomicrographs of validation of two near-infrared vascular probes with confocal in vivo microscopy in live animals bearing CT26 tumors. Tumor microvasculature is chaotic—a typical feature. Note that imaging probes are truly intravascular (at least for 20 minutes after intravenous administration) and that there is concordance between the two probes. Note absence of bleed-through when only one channel is turned on. Scale bar = 100 µm, recon = reconstruction.

 

Figure 2A
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a: Graphs of FMT reconstruction values of fluorochromes. A 1-mm capillary tube was filled with known amounts of (a) AngioSense 680, (b) AngioSense 750, or (c) both and was immersed in opaque medium to simulate tissue at FMT data acquisition. There was excellent correlation between reconstructed and true probe amounts (r = 0.98, P < .003 for AngioSense 680; r = 0.99, P < .001 for AngioSense 750; r = 0.99, P < .02 for mixture). Nevertheless, slope is not equal to 1 due to error in calibration factor. (d) In vivo correlation of both imaging probes in animals bearing CT26 tumors. Animals received 2 nmol AngioSense 680 and 2 nmol AngioSense 750 before imaging. Note good correlation between the two near-infrared channels (r = 0.91, P < .002).

 

Figure 2B
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b: Graphs of FMT reconstruction values of fluorochromes. A 1-mm capillary tube was filled with known amounts of (a) AngioSense 680, (b) AngioSense 750, or (c) both and was immersed in opaque medium to simulate tissue at FMT data acquisition. There was excellent correlation between reconstructed and true probe amounts (r = 0.98, P < .003 for AngioSense 680; r = 0.99, P < .001 for AngioSense 750; r = 0.99, P < .02 for mixture). Nevertheless, slope is not equal to 1 due to error in calibration factor. (d) In vivo correlation of both imaging probes in animals bearing CT26 tumors. Animals received 2 nmol AngioSense 680 and 2 nmol AngioSense 750 before imaging. Note good correlation between the two near-infrared channels (r = 0.91, P < .002).

 

Figure 2C
View larger version (13K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2c: Graphs of FMT reconstruction values of fluorochromes. A 1-mm capillary tube was filled with known amounts of (a) AngioSense 680, (b) AngioSense 750, or (c) both and was immersed in opaque medium to simulate tissue at FMT data acquisition. There was excellent correlation between reconstructed and true probe amounts (r = 0.98, P < .003 for AngioSense 680; r = 0.99, P < .001 for AngioSense 750; r = 0.99, P < .02 for mixture). Nevertheless, slope is not equal to 1 due to error in calibration factor. (d) In vivo correlation of both imaging probes in animals bearing CT26 tumors. Animals received 2 nmol AngioSense 680 and 2 nmol AngioSense 750 before imaging. Note good correlation between the two near-infrared channels (r = 0.91, P < .002).

 

Figure 2D
View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2d: Graphs of FMT reconstruction values of fluorochromes. A 1-mm capillary tube was filled with known amounts of (a) AngioSense 680, (b) AngioSense 750, or (c) both and was immersed in opaque medium to simulate tissue at FMT data acquisition. There was excellent correlation between reconstructed and true probe amounts (r = 0.98, P < .003 for AngioSense 680; r = 0.99, P < .001 for AngioSense 750; r = 0.99, P < .02 for mixture). Nevertheless, slope is not equal to 1 due to error in calibration factor. (d) In vivo correlation of both imaging probes in animals bearing CT26 tumors. Animals received 2 nmol AngioSense 680 and 2 nmol AngioSense 750 before imaging. Note good correlation between the two near-infrared channels (r = 0.91, P < .002).

 
In Vivo Imaging
Dual-wavelength imaging.—Subsequent in vivo experiments were performed to validate that the imaging probes exhibited intravascular residence during imaging. By using intravital laser scanning microscopy, we determined that the coinjected imaging probes were exclusively located within chaotic tumor blood vessels without any apparent leakage during imaging time (30 minutes after injection). The vascular half-life of the imaging probes was more than 5 hours (25). In time series, blood flow could be seen in either channel, with circulating cells being enhanced with fluorescent plasma. Furthermore, there was close correlation between the probes observed in the 680-nm channels and those in the 750-nm channels, indicating identical behavior of the two imaging probes. When respective excitation sources (680 or 750 nm) were turned off, no imaging signal was observed in that channel, confirming that the signals in both channels were not due to bleed-through of signal. In another set of experiments, we coinjected both imaging probes into animals bearing CT26 subcutaneous tumors and performed imaging with FMT in both channels to correlate signal concordance at the macroscopic scale (Fig 2d). Imaging signals were linear as expected, and the correlation coefficient (r) was 0.91, which was determined with f(x) = 0.85(±0.16)x + 15.7(±20.8).

Orthotopic versus ectopic tumors.—Subsequent experiments were conducted to determine whether implanted tumors could be depicted tomographically. On average, ectopic tumors (Fig 3) contained 30 pmol fluorochrome, which is consistent with a VVF of 3.5% ± 1.5. Orthotopic tumors contained 28 pmol fluorochrome, which is consistent with a VVF of 3.3% ± 0.8 (P = .85). Orthotopic tumors and ectopic tumors thus exhibited similar levels of VVF, with considerable variation among tumors and animals, as has been reported previously for other experimental cancers (11).


Figure 3A
View larger version (115K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3a: Serial coronal reconstructions (reconstruction voxel, 200 x 200 x 400 µm) of tumors. Imaging was performed in (a) ectopically and (b) orthotopically implanted colonic tumors. FMT enabled imaging of near-infrared fluorescent probes in tumor neovasculature in (a) superficial and (b) deep-seated tumors. Interestingly, tumors had similar mean VVF in the two locations, although standard deviation was large due to tumor heterogeneity. Arrows = tumor, * = bladder.

 

Figure 3B
View larger version (98K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3b: Serial coronal reconstructions (reconstruction voxel, 200 x 200 x 400 µm) of tumors. Imaging was performed in (a) ectopically and (b) orthotopically implanted colonic tumors. FMT enabled imaging of near-infrared fluorescent probes in tumor neovasculature in (a) superficial and (b) deep-seated tumors. Interestingly, tumors had similar mean VVF in the two locations, although standard deviation was large due to tumor heterogeneity. Arrows = tumor, * = bladder.

 
Antiangiogenic treatment.—To explore the effect of a VEGF blockade on angiogenesis, we imaged cohorts of mice that had received 1 week of treatment for established tumors (Fig 4). Animals treated with the anti-VEGF antibody (5 mg/kg) showed lower angiogenesis (VVF, 1.33% ± 0.96) than did the phosphate-buffered saline (control) group (VVF, 5% ± 1.3). FMT imaging findings were corroborated with mean vessel density measurements (baseline, mean vessel density = 197 ± 25; control group, mean vessel density = 244 ± 21; anti-VEGF antibody group, mean vessel density = 155 ± 35), and CD31 immunohistochemistry results are shown in Figure 4. The difference between the group treated with phosphate-buffered saline and the group treated with anti-VEGF antibody reached significance (P < .01). There was no statistically significant difference between the baseline and the phosphate-buffered saline–treated group (Fig 4).


Figure 4
View larger version (69K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4: In vivo angiogenic data. To determine whether chemotherapeutic drug-induced effects on tumor neovasculature could be detected, we performed FMT imaging in three cohorts of mice: animals receiving no treatment (normal) (n = 8), animals receiving saline (PBS) (n = 8), and animals receiving treatment (bevacizumab [Avastin]) (n = 8). Note lower VVF for treatment group compared with control and baseline groups. Imaging results were validated with immunohistochemistry. (CD31 microvascular stain and hematoxylin-eosin (HE) stain; original magnification, x200.) Scale bar = 100 µm. NS = not significant. Error bars = standard deviation.

 
Serial imaging.—We determined whether it would be possible to image the natural course of angiogenesis and its inhibition serially in the same animal. Results of serial imaging in the same animal (Fig 5a) showed that dual-wavelength FMT was able to depict angiogenesis over time (in individual animals). VVF varied from 3.00% ± 1.85 to 3.57% ± 0.75 during development and decreased on average to 1.87% ± 1.02 during anti-VEGF (5 mg/kg) treatment.


Figure 5A
View larger version (18K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5a: Graphs of angiogenic progression and regression data. (a) Serial imaging experiments were conducted at different wavelengths to measure angiogenesis during tumor progression ("pre") and regression ("post") after antiangiogenic therapy (Tx, "pre 1 wk"). (b) Imaging of bevacizmab (Avastin)-induced antiangiogenic effects with different doses. Fifty percent effective dose was 2.5 mg/kg; beyond 5 mg/kg, no further antiangiogenic effects were observed.

 

Figure 5B
View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5b: Graphs of angiogenic progression and regression data. (a) Serial imaging experiments were conducted at different wavelengths to measure angiogenesis during tumor progression ("pre") and regression ("post") after antiangiogenic therapy (Tx, "pre 1 wk"). (b) Imaging of bevacizmab (Avastin)-induced antiangiogenic effects with different doses. Fifty percent effective dose was 2.5 mg/kg; beyond 5 mg/kg, no further antiangiogenic effects were observed.

 
Therapeutic dose response.—Finally, we performed dose-response experiments to determine optimal antiangiogenic drug concentrations in the CT26 model. We report the effective dose where 50% of the effect can be seen (hereafter, 50% effective dose). As shown in Figure 5b, the 50% effective dose of bevacizumab was 2.5 mg/kg, which is in a similar range as described previously for other cancer models (26).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
ADVANCES IN KNOWLEDGE
 References
 
Our results show that FMT-derived fluorochrome measurements are less depth dependent than those obtained with conventional reflectance imaging. That fact, together with a linear dose-response curve over several orders of magnitude, a true intravascular residence time of the fluorescent tracer, and the ability to perform calculations without pharmacokinetic assumptions of rapid extravasation, lend to the robustness of the described technique. Indeed, our results show that serial imaging and dose-response imaging are feasible in experimental mouse models in short time periods. Animal preparation and injection and image acquisition, reconstruction, and analysis took about 10 minutes per mouse, and throughputs of 50 mice per day have been achieved.

The simplest technique for detecting optical reporter molecules in vivo is generally a form of planar imaging (fluorescence reflectance imaging) with the use of simple charge-coupled device cameras. Planar imaging is technically easy to implement but has limitations, as shown by our and other experiments (17). The strong photon scattering and weighting of signals toward surface activity, limited penetration depth, single-projection viewing, and several nonlinear effects of photon propagation can lead to erroneous interpretation of planar data. Although there is a nonlinear relation between depth and imaging signal, FMT has the potential to improve on these inherent limitations and offer robust and accurate quantitation in three dimensions. In a typical FMT experiment, tissue is illuminated at different angles, and diffuse light photons are detected by using a charge-coupled device camera. Each source-detector pair effectively implements a different projection through tissue, albeit through tortuous pathways. These measurements are then combined in a tomographic reconstruction scheme that in its simplest form can be written as a matrix equation, y = Wx, where y is a vector of measurements, x is the vector of unknown optical properties for each of the voxels in the image, and W is the weight matrix that is calculated on the basis of a theoretic model of photon propagation in tissues (forward model). As in positron emission tomographic imaging, data can be reconstructed in different planes, and three-dimensional regions of interest yield quantitative fluorochrome data, which simplifies analysis.

FMT has been validated for a number of fluorescent imaging agents, such as magnetofluorescent nanoparticles (20), fluorescent trastuzumab (20), annexin V (18), and activatable imaging probes (27), among others. Our study used long-circulating highly fluorescent graft copolymers similar in design to the ones previously tested in clinical trials (21). This agent contains methoxy polyethylene glycol attached to a polylysine backbone and is optimized for vascular imaging (25), as well as for high fluorescence. Our intravital microscopic data indicate a blood half-life of the agent of more than 5 hours, with essentially no extravasation into tumors during the first 30 minutes after intravenous administration. At later time points, extravasation does occur, so tumors become brighter over time. This effect has previously been investigated in depth for different-sized markers (28) and exploited for tumoral drug delivery (electron paramagnetic resonance effect) (29).

To circumvent uncertainties associated with an extravasated imaging agent during serial imaging, we took advantage of imaging at a second wavelength for clearance to occur with time. While we have used only two excitation wavelengths (680 and 750 nm), we envision that similar probes with different fluorescence peaks (eg, 630 nm, 800 nm, 850 nm) could further extend the number of measurements made in the same animal. Furthermore, we have also shown that repeated measurements in the same channel are possible if pre- and postimage series are acquired during a subsequent imaging session.

Future studies on the described techniques may focus on additional animal models with spontaneous tumors. One limitation of our study was the use of orthotopic and heterotopic tumors. In these models, antiangiogenic therapeutic response may be enhanced due to more rapid tumor growth, which results in more rapid microvessel growth and more abnormal neovasculature than that typically seen in spontaneous tumors. However, the blood-pool agents used in this study remained within the vasculature beyond the imaging time, even for the models used, and would be even less likely to leak from the vasculature in spontaneous tumors. Therefore, no methodologic changes would be required in such evaluation of therapeutic drug dosing.

Practical applications: Our results match with reported literature values (26), as well as with correlative histologic findings, and indicate the potential for quantitative optical imaging in angiogenesis inhibition. The described methods may prove useful for investigating the biology of angiogenesis, for comparing microscopic (intravital microscopy) and macroscopic imaging data, and for testing the effectiveness of angiogenesis inhibitors.


   ADVANCES IN KNOWLEDGE
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
References
 


   ACKNOWLEDGMENTS
 
The authors acknowledge the assistance of Betty Zheng, BS, and Peter Waterman, BS, with image acquisition and Elena Aikawa, MD, PhD, for helpful comments regarding immunohistochemistry.


   FOOTNOTES
 

Abbreviations: FMT = fluorescence molecular tomography • VEGF = vascular endothelial growth factor • VVF = vascular volume fraction

See Materials and Methods for pertinent disclosures.

Author contributions: Guarantors of integrity of entire study, V.N., U.M., R.W.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, X.M., J.L.F., H.A.; experimental studies, X.M., J.L.F., H.A., V.N.; statistical analysis, X.M., J.L.F., H.A.; and manuscript editing, V.N., U.M., R.W.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 References
 

  1. Hanahan D, Weinberg R. The hallmarks of cancer. Cell 2000;100:57–70.[CrossRef][Medline]
  2. Carmeliet P, Jain R. Angiogenesis in cancer and other diseases. Nature 2000;407:249–257.[CrossRef][Medline]
  3. Libby P, Schonbeck U. Drilling for oxygen: angiogenesis involves proteolysis of the extracellular matrix. Circ Res 2001;89:195–197.[Free Full Text]
  4. Libby P, Zhao D. Allograft arteriosclerosis and immune-driven angiogenesis. Circulation 2003;107:1237–1239.
  5. Sivakumar B, Harry L, Paleolog E. Modulating angiogenesis: more vs less. JAMA 2004;292:972–977.[Abstract/Free Full Text]
  6. McDonald DM, Choyke PL. Imaging of angiogenesis: from microscope to clinic. Nat Med 2003;9:713–725.[CrossRef][Medline]
  7. Miller JC, Pien HH, Sahani D, Sorensen AG, Thrall JH. Imaging angiogenesis: applications and potential for drug development. J Natl Cancer Inst 2005;97:172–187.[Abstract/Free Full Text]
  8. Uyl-de Groot, CA, Giaccone G. Health economics: can we afford an unrestricted use of new biological agents in gastrointestinal oncology? Curr Opin Oncol 2005;17:392–396.[CrossRef][Medline]
  9. Liu G, Rugo H, Wilding G, et al. Dynamic contrast-enhanced magnetic resonance imaging as a pharmacodynamic measure of response after acute dosing of AG-013736, an oral angiogenesis inhibitor, in patients with advanced solid tumors: results from a phase I study. J Clin Oncol 2005;23:5464–5473.[Abstract/Free Full Text]
  10. Bremer C, Mustafa M, Bogdanov AJ, Ntziachristos V, Petrovsky A, Weissleder R. Steady-state blood volume measurements in experimental tumors with different angiogenic burdens: a study in mice. Radiology 2003;226:214–220.[Abstract/Free Full Text]
  11. Lewin M, Bredow S, Sergeyev N, Marecos E, Bogdanov AJ, Weissleder R. In vivo assessment of vascular endothelial growth factor-induced angiogenesis. Int J Cancer 1999;83:798–802.[CrossRef][Medline]
  12. Sipkins D, Cheresh D, Kazemi M, Nevin L, Bednarski M, Li K. Detection of tumor angiogenesis in vivo by alphaVbeta3-targeted magnetic resonance imaging. Nat Med 1998;4:623–626.[CrossRef][Medline]
  13. Kang HW, Josephson L, Petrovsky A, Weissleder R, Bogdanov A Jr. Magnetic resonance imaging of inducible E-selectin expression in human endothelial cell culture. Bioconjug Chem 2002;13:122–127.[CrossRef][Medline]
  14. Bredow S, Lewin M, Hofmann B, Marecos E, Weissleder R. Imaging of tumour neovasculature by targeting the TGF-beta binding receptor endoglin. Eur J Cancer 2000;36:675–681.[CrossRef][Medline]
  15. Huang P, McKee T, Jain R, Fukumura D. Green fluorescent protein (GFP)-expressing tumor model derived from a spontaneous osteosarcoma in a vascular endothelial growth factor (VEGF)-GFP transgenic mouse. Comp Med 2005;55:236–243.[Medline]
  16. Wang Y, Iyer M, Annala A, Wu L, Carey M, Gambhir S. Non-invasive indirect imaging of vascular endothelial growth factor gene expression in a transgenic mouse model using bioluminescence imaging [abstr]. Mol Imaging 2005;4:315.
  17. Ntziachristos V, Ripoll J, Wang L, Weissleder R. Looking and listening to light: the evolution of whole-body photonic imaging. Nat Biotechnol 2005;23:313–320.[CrossRef][Medline]
  18. Ntziachristos V, Schellenberger E, Ripoll J, et al. Visualization of antitumor treatment by means of fluorescence molecular tomography with an annexin V-Cy5.5 conjugate. Proc Natl Acad Sci U S A 2004;101:12294–12299.[Abstract/Free Full Text]
  19. Graves EE, Weissleder R, Ntziachristos V. Fluorescence molecular imaging of small animal tumor models. Curr Mol Med 2004;4:419–430.[CrossRef][Medline]
  20. Montet X, Ntziachristos V, Grimm J, Weissleder R. Tomographic fluorescence mapping of tumor targets. Cancer Res 2005;65:6330–6336.[Abstract/Free Full Text]
  21. Callahan RJ, Bogdanov A Jr, Fischman AJ, Brady TJ, Weissleder R. Preclinical evaluation and phase I clinical trial of a 99mTc-labeled synthetic polymer used in blood pool imaging. AJR Am J Roentgenol 1998;171:137–143.[Abstract/Free Full Text]
  22. Alencar H, Mahmood U, Kawano Y, Hirata T, Weissleder R. Novel multi-wavelength microscopic scanner for mouse imaging. Neoplasia 2005;7:977–983.[CrossRef][Medline]
  23. Alencar H, King R, Funovics M, Stout C, Weissleder R, Mahmood U. A novel mouse model for segmental orthotopic colon cancer. Int J Cancer 2005;117:335–339.[CrossRef][Medline]
  24. Mentzel T, Brown L, Dvorak H, et al. The association between tumour progression and vascularity in myxofibrosarcoma and myxoid/round cell liposarcoma. Virchows Arch 2001;438:13–22.[CrossRef][Medline]
  25. Weissleder R, Bogdanov AJ, Tung C, Weinmann H. Size optimization of synthetic graft copolymers for in vivo angiogenesis imaging. Bioconjug Chem 2001;12:213–219.[CrossRef][Medline]
  26. Mordenti J, Thomsen K, Licko V, Chen H, Meng Y, Ferrara N. Efficacy and concentration-response of murine anti-VEGF monoclonal antibody in tumor-bearing mice and extrapolation to humans. Toxicol Pathol 1999;27:14–21.[Abstract/Free Full Text]
  27. Ntziachristos V, Tung C, Bremer C, Weissleder R. Fluorescence molecular tomography resolves protease activity in vivo. Nat Med 2002;8:757–760.[CrossRef][Medline]
  28. Yuan F, Dellian M, Fukumura D, et al. Vascular permeability in a human tumor xenograft: molecular size dependence and cutoff size. Cancer Res 1995;55:3752–3756.[Abstract/Free Full Text]
  29. Satchi-Fainaro R, Puder M, Davies J, et al. Targeting angiogenesis with a conjugate of HPMA copolymer and TNP-470. Nat Med 2004;10:255–261.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 JCOHome page
L. S. Hart and W. S. El-Deiry
Invincible, but Not Invisible: Imaging Approaches Toward In Vivo Detection of Cancer Stem Cells
J. Clin. Oncol., June 10, 2008; 26(17): 2901 - 2910.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
K. A. Kelly, S. R. Setlur, R. Ross, R. Anbazhagan, P. Waterman, M. A. Rubin, and R. Weissleder
Detection of Early Prostate Cancer Using a Hepsin-Targeted Imaging Agent
Cancer Res., April 1, 2008; 68(7): 2286 - 2291.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Montet, X.
Right arrow Articles by Weissleder, R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Montet, X.
Right arrow Articles by Weissleder, R.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE