Detection of Dysplastic Intestinal Adenomas Using a ... - SAGE Journals

RESEARCH ARTICLE

Molecular Imaging . Vol. 4, No. 1, January 2005, pp. 67 – 74

67

Detection of Dysplastic Intestinal Adenomas Using a Fluorescent Folate Imaging Probe Wei-Tsung Chen1,*,y, Khashayarsha Khazaie1,2,*, Guoying Zhang 2, Ralph Weissleder 1, and Ching-Hsuan Tung 1 1

Harvard Medical School, Charlestown, and 2Harvard Medical School, Boston

Abstract Macrophages have long been recognized as a prominent component of tumors. Activated macrophages overexpress folate receptors and we used this phenomenon to image inflammatory reactions in colon dysplasia using a fluorescent folate probe (FFP). APC#468 mice injected with FFP showed fluorescent adenomas (target-to-background ratio, adenoma vs. adjacent normal mucosa, of 2.46 ± 0.41), significantly higher ( p < .001) than adenomas in animals injected with a non-folate-containing control probe. Fluorescence-activated cell-sorting analysis revealed a 3-fold higher content of Mac1positive cells in colonic adenomas compared with normal adjacent mucosa (6.8% vs. 2.2%), and confirmed the source of FFP-positive cells to be primarily an F4/80-positive macrophage subpopulation. Taken together, these results indicate that FFP potentially can be used to image dysplastic intestinal adenomas in vivo. Mol Imaging (2005) 4, 67 – 74. Keywords: Adenoma, dysplasia, folate receptor, macrophage, colon cancer, imaging.

Introduction Adenomatous polyps represent the earliest stage of colon cancer, characteristically arising from the deregulated growth of aberrant crypt epithelial cells, and can progress to invasive and metastatic cancer [1]. Cross talk between tumor-infiltrating and neoplastic cells is increasingly recognized to influence various stages of carcinogenesis. Early during tumor formation, stromal cells and tumor-infiltrating leukocytes might provide signals that regulate cancer cell growth and differentiation. It would therefore be useful to develop methods to noninvasively measure the degree of infiltration of early dysplasia by inflammatory cells in situ [2,3]. Dysplasia is characterized histopathologically by the disorganized growth of undifferentiated crypt epithelial cells, exhibiting a high mitotic index and harboring enlarged nuclei with multiple prominent nucleoli. We, as well as others, have observed that macrophages make up a prominent cellular component of dysplasia, and that several receptors and enzymes could be used to image activated macrophages. Among these is the folate receptor-b, a nonepithelial isoform of the folate receptor, which

expressed in activated macrophages [4]. Similar to the epithelial form, it is a 38-kDa glycosyl phosphatidylinositol-anchored protein that binds folic acid and folate conjugates with high affinity (< 1 nM) [5,6]. With the exception of the kidney and placenta, normal tissues express low to undetectable levels of folate receptor [5]. Folate-receptor-targeted radionuclide agents have been suggested for in vivo imaging application [7]. Recently, we have synthesized a fluorescent folate probe (FFP) [8] and investigated its utility for the detection of cancer in vivo [9]. Here we reasoned that the developed FFP could be useful to image macrophages, which are associated with dysplastic intestinal adenomas, and thus enhance the detection of adenomas. Our results show that the probe is specifically enriched within the dysplastic lesions and recognizes a population of tumorinfiltrating cells also expressing the Mac1 and F4/80 antigens, characteristic markers of macrophages. These observations indicate that the probe is suitable for in situ imaging of local inflammatory reactions associated with early dysplasia in colon cancer.

Materials and Methods Probe and Fluorochrome The FFP was prepared as previously described [8,9]. Both FFP and its nonfolate fluorochrome, NIR2 [10], have an excitation wavelength maximum at 662 nm and an emission wavelength maximum at 686 nm. Mouse Model APCD468 mice (n = 9, age 12 –25 weeks) were obtained from the animal core laboratory of the Dana Farber Harvard Cancer Center. The animal protocol was approved by the Institutional Review Board. Mice were Corresponding author: Ching H. Tung, PhD, Center for Molecular Imaging Research, Massachusetts General Hospital, 149 13th Street, Room 5406, Charlestown, MA 02129; e-mail: [email protected] *These authors contributed equally to this work. y Current address: Radiology Department, Taipei Municipal Jen-Ai Hospital, Taipei, Taiwan. Received 22 December 2004; Accepted 14 January 2005. D 2005 Neoplasia Press, Inc.

68

Dysplastic Intestinal Adenoma Imaging Chen et al.

randomly divided into three experimental groups: (1) animals receiving an intravenous injection of 2 nmol FFP (n = 6, 162 adenomas,); (2) animals receiving an intravenous injection with 2 nmol free fluorochrome (n = 2, 78 adenomas); and (3) noninjected control mice (n = 1, 28 adenomas). In addition, we also included strain-matched non-APC mice (C57BL6, Jackson Laboratories, Bar Harbor, ME) and injected them with 2 nmol FFP (n = 2, no true adenomas). An additional three APC mice were used for in vitro fluorescence-activated cellsorting (FACS) analysis. Adenomas were identified by using a stereomicroscope (Leica DC 500, Bannockburn, IL); the lesions showed typical pedunculated/tubular or sessile/villous morphology under the stereomicroscope. Flow Cytometry Analysis Six hours after FFP injection, adenomas were excised and pooled. Normal intestinal mucosa from the vicinity of the adenomas was collected in a similar fashion. Tissues were minced in small fragments and reacted with collagenase (1800 U/ml) and hyaluronidase (1200 U/ ml) in Hank’s solution containing DNAse I (20 mg/ml) and MgCl2 (2 mM) for 20 min at 37C. Isolated cells were obtained by passing through a 20-mm filtration membrane (Becton Dickinson, San Jose, CA). After blocking with 1% bovine serum albumin in phosphate-buffered saline (PBS), cells were incubated with Fc block (2.4G2, Pharmingen, San Diego, CA), fluorescein (FITC)-labeled anti-C-Kit antibody, phycoerythrin-labeled anti-Mac1 (CD11b) antibody (Pharmingen), and FITC-labeled F4/ 80 (eBioscience, San Diego, CA) for identifying macrophage population [11]. Primary goat antifolate receptor and FITC-labeled secondary antibody was used to detect folate receptor expression. Flow cytometry was performed on FACSCalibur (Becton Dickinson) using samples with or without the primary antibody and analyzed by CellQuest software (Becton Dickinson). Alternatively, the entire gut was minced, digested as above, and then fractionated on a two-step Percoll gradient (44% and 67%). The interface cells were collected and stained as above before FACS analysis. Imaging and Lesion Assessment To each APC mouse, 2 nmol FFP was injected via tail vein 6 hr before imaging. Normal saline flushing was first performed from the proximal jejunum to distal ileum by laparotomy and a 19-gauge feeding needle was indwelled into the lumen of the proximal jejunum to remove food residuals that contained autofluorescence signals. After flushing, the animal was placed in a decubitus position in a home-built fluorescence imaging system, which has

been described in detail previously [12], and the small bowels were pulled and placed on the imaging table for imaging. For fluorescence acquisition, a 615- to 645-nm excitation filter and a 680- to 720-nm emission filter (Omega Optical, Brattleboro, VT) were used. After imaging, animals were euthanatized with CO2 inhalation, then the entire bowels were removed, and ex vivo fluorescence imaging was performed for correlation. The same procedure was also performed in a strainmatched non-APC C57BL6 mouse. Images were acquired for 2 min and analyzed using commercially available software (Kodak Digital Science 1D software, Rochester, NY). Regions of interests were obtained from the entirety of each adenoma and from adjacent size matched intestinal mucosa. Mean signal intensities (SI) were recorded. The target (adenoma) to background (adjacent normal mucosa) contrast ratio (TBR) was calculated as follows: TBR (%) = (SIadenoma SIbackground)/ (SImucosa SIbackground) where SIbackground represents background optical noise (offset). All results are presented as mean ± standard deviation. Statistical analysis of the two groups was conducted using a two-tailed Student t test for unpaired samples by using SPSS 9.0 (SPSS Inc., Chicago, IL). A P value < .05 was considered to be significant.

Histology, Immunohistochemistry, and Immunofluorescence Microscopy Bowel tissue of animals was excised and fixed in 4% paraformaldehyde for 24 hr, paraffin embedded, cut into 6-mm sections, and stained with hematoxylin and eosin. Immunohistochemistry for folate receptor expression in adenomas was performed on frozen sections using a diluted 1:100 primary goat antifolate receptor polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Binding of the primary antibody was revealed with a diluted 1:250 biotinylated secondary donkey anti-goat antibody (Santa Cruz Biotechnology). The staining procedure was performed with a modified avidin– biotin –peroxidase complex technique. The slides were visualized with a chromogen of diaminobenzidine (DAB, Vectastain; Vector Laboratories, Burlingame, CA). Sections were counterstained with hematoxylin ( Vector Laboratories). Frozen sections from APC mouse intestine were fixed in acetone for 10 min, blocked with 10% normal donkey serum for 1 hr at room temperature, then incubated with a 1:10 dilution of primary rat anti-Mac1 monoclonal antibody (BD Biosciences) or a 1:100 dilution of primary goat antifolate receptor polyclonal antibody (Santa Cruz Biotechnology) overnight at room temperature. After Molecular Imaging . Vol. 4, No. 1, January 2005


washing in PBS, they were reacted for 1 hr at room temperature with a dilution of 1:250 secondary antibody (FITC-conjugated donkey anti-rat IgG for Mac1 antibody and Cy3-conjugated donkey anti-goat F(ab0)2 fragment for folate receptor antibody, respectively). Immunofluorescence caused by binding of the primary antibody to Mac1 and folate receptor was observed on a Zeiss LSM 5 PASCAL confocal microscope (Zeiss, Thornwood, NY). FFP uptake cells were observed in fluorescence channels for colocalization with Mac1- and folate-receptorpositive cells.

Results Adenomatous Lesions A mouse model with a targeted truncating mutation of the adenomatous polyposis coli (APC) gene at exon 468 was used in this study (Khazaie, in preparation). From 2 months of age, adenomas occurred in the small and large intestine (n = 265) and had typically a tubovillous appearance (Figure 1A), ranging in size

69

from 0.9 to 3.7 mm in diameter (mean, 2.1 mm). Analysis of histologic sections showed the characteristic architecture of dysplastic crypts with hyperchromatic nuclei, expanding at the luminal surface of the mucosa, neighboring well spaced, nondysplastic crypts and villi (Figure 1B). The adenomas were largely devoid of differentiated cells including mucin producing goblet cells (Figure 1C). Besides the size of the lesions, there is very limited heterogeneity among the adenomas that arise in APC-deficient mouse models of polyposis. Immunohistochemical staining for folate receptor revealed an abundance of positive cells dispersed in the stroma of dysplastic intestinal adenomas, whereas no staining of the aberrant epithelium was detected (Figure 1D and E).

Fluorescence Imaging Enhanced fluorescent signal emanating from adenomas was readily detected in vivo by fluorescence imaging 6 hr after probe injection (Figure 2A). The

Figure 1. Macroscopic image, histology, and immunohistochemistry of dysplastic adenoma. (A) Macroscopic view of several closely arranged adenomas in the duodenum of transgenic mice. Arrow points to an adenoma. (B) Hematoxylin and eosin stain of a typical tubovillous adenoma characteristic of the phenotype of the lesions that appear in the small intestine. Dysplasia was characterized by tall, hyperchromatic disorderly cells with cigar-shaped nuclei and concomitant crypt budding (arrow), which contrasts with the appearance of discrete villi lined with a monolayer of secretory epithelial cells (arrowhead). (C) PAS staining of the same lesion as in B, showing sulfomucin-expressing goblet cells in healthy neighboring tissue (arrowhead) and their absence within the adenoma (arrow) due to predominance of undifferentiated cells in the lesion. (D and E) Immunoperoxidase staining on cryosections of folate receptor in dysplastic intestinal adenomas. Note that folate-receptor-positive cells (red arrows) are located in the stroma, whereas the aberrant epithelium is unstained (black arrows). Magnification 400.

Molecular Imaging . Vol. 4, No. 1, January 2005

70


Figure 2. In vivo and ex vivo fluorescence imaging of dysplastic intestinal adenomas 6 hr after FFP injection. (A and B) Mice were anesthetized, and a small incision was introduced in the abdomen through which a length of the small intestine of approximately 10 cm was gently pooled out, flushed with PBS and imaged. (A) APC mice showed multiple fluorescing nodules. (B) No focal fluorescence-enhanced lesions were detectable in the control healthy mouse. (C) The section of the bowel imaged in the live APC mouse was excised from the sacrificed mouse and imaged ex vivo, and fluorescence imaging correlated with the in vivo image. (D). Target-tobackground ratio measurements showing increased fluorescence intensity of signal from adenomas of folate-probe-injected (n = 162) as compared with freefluorochrome-injected (n = 78) or control APC mice not injected with probe (n = 28).

controlled non-APC C57BL6 mice injected with same dose of FFP showed no focal enhanced fluorescence (Figure 2B). Ex vivo fluorescence imaging of distal ileum correlated with in vivo imaging to a high degree (Figure 2C). The TBR of FFP injected group was 2.45 ± 0.41 (n = 162), significantly higher than that of the noninjected group (1.54 ± 0.31, n = 27, p < .001, Figure 2D). It was also higher than that of animals injected with control fluorochrome devoid of folate (2.15 ± 0.37, n = 76, p < .001). Histologic analysis of lesions revealed the tendency for high TBR to correlate with higher cellularity than lesions with relative low TBR (data not shown). Colocalization of Fluorescence Signal with Mac1 and Folate Receptor Positive Cells Tumor Infiltrating Cells Fluorescence confocal microscopy of frozen sections of adenomas showed scattered Mac1-positive cells and FFP-positive cells in the stroma (Figure 3 A – D). The superimposed image showed that FFP-positive cells colocalized to a high degree with Mac1-positive cells; however, not all MAC1-positive cells have fluorescence signal. Furthermore, immunofluorescence staining for folate-receptor-positive cells also revealed a close corre-

lation between receptor expression on the cell surface and FFP signal (Figure 3 E –G). Colocalization of Fluorescence Signal with Mac1 and F4/80, Folate Receptor Positive Macrophages To extend the results obtained by histology, adenomas of the healthy neighboring tissues were microdissected, digested to single cell suspension using collagenase, stained with specific antibodies and subjected to flow cytometry analysis. Mac1-positive cells were significantly increased in the dysplastic lesions as compared with neighboring healthy tissues, both in the small intestine and in the colon of APC defective mice (Figure 4A). Analysis of probe injected APC mice, revealed that Mac1-positive cells had higher fluorescence signal intensity at 700 nm (FFP) as compared with Mac1-negative cells (Figure 4B). Furthermore, expression of folate receptor was higher in adenoma derived from Mac1-positive cells than those derived from the neighboring healthy mucosa (Figure 4C). Similar results were obtained when single cell suspensions from total intestines were compared. As shown in Figure 5, a significant increase in Mac1-positive cells was noted in the preparation from APC as compared with control Molecular Imaging . Vol. 4, No. 1, January 2005


71

Figure 3. Colocalization of Mac1 staining and FFP signal in adenomas. (A) Immunofluorescence staining revealed Mac1-positive cells in the fluorescein channel. (B) FFP-positive cells were revealed in the 680- to 720-nm channel. (C) Bright-field image. (D) Superimposed image showing Mac1-positive cells colocalizing with FFPpositive cells. Original magnification 200. (E – G) Colocalization of folate receptor staining and FFP signal. (E) Immunofluorescence staining for folate receptor revealed in the rhodamine channel. (F) Corresponding FFP-positive cells were revealed in 680 – 720 nm channel. (G) Superimposed image showing folate-receptorpositive cells colocalizing with FFP-positive cells. Original magnification: 400.

intestines. Furthermore, the intensity of FFP signal correlated with the intensity of Mac1 staining by the cells analyzed. These results were in line with the notion that the source of FFP signal was tumor-infiltrating cells associated with local inflammatory reactions. To define the population of tumor-infiltrating cells that expressed folate receptor, single-cell preparations from complete APC intestines harboring adenomas or control healthy intestines were stained with Mac1 as well as antibodies specific to folate receptor and to a marker, F4/80, that is characteristically expressed by inflammatory macrophages [13]. Folate receptor was predominantly expressed by Mac1- and F4/80-positive cells (Figure 5C). All together, our observations are consistent with the FFP signal emanating from this subpopulation of tumorinfiltrating and inflammation-associated macrophages.

Discussion Macrophages are a major component of inflammatory reactions associated with tumor growth. The intimate Molecular Imaging . Vol. 4, No. 1, January 2005

association of macrophages with tumor growth has prompted speculations, sometimes backed by evidence, that the cross talk between aberrant epithelium and tumor-infiltrating macrophages may be a determining factor in tumor growth and progression. Therefore, in situ imaging of tumor-infiltrating macrophages may not only increase the sensitivity of detection of the lesions, but also provide biological information on the status of the tumors and their response to therapy. To image early inflammatory responses in colon cancer development, we developed a novel folate-receptorbinding fluorescence probe and showed that intravenous administration of this probe leads to preferential labeling of intestinal polyps, as compared with healthy neighboring tissue, both in the live mouse as well as in explanted tissue. The TBR of the FFP injection group was higher than that of the noninjection group and significantly higher than that of the nontargeting free fluorochrome injection group ( p < .0001). Thus, the FFP is superior to nonspecific fluorochromes for in vivo imaging, the latter often being used for nontargeted

72


Figure 4. Flow cytometry analysis of dysplastic intestinal adenoma. Frequency of Mac1-positive cells in healthy bowel and in dysplastic adenoma (A). Averages of four acquisitions with standard deviations are shown. (B) The Mac1-positive cells showed higher fluorescence signal intensity at 700 nm (open histogram) than Mac1negative cells (filled histogram). (C) Histogram of folate receptor expression in normal mucosa and dysplastic adenomas by gating Mac1-positive cells, a larger number of cells in the dysplastic adenomas bound folate receptor probe (open histogram) as compared with cells derived from the adjacent healthy mucosa (filled histogram). The FACS data represent at least three separate experiments.

image enhancement [14,15]. Previously, using different models, we have demonstrated that such folatereceptor-dependent enhancement can be inhibited by adding free folic acid [9,16]. Furthermore, our observations exclude nonspecific phagocytosis by activated macrophages or free fluorochrome pools in the interstitial space at the inflammation site. Expression of folate receptor was expected to be abundant mainly on tumor-infiltrating macrophages. Indeed, immunohistochemistry as well as FACS analysis documented the expression of folate receptor on stromal cell population, and excluded any expression by tumor epithelium. Furthermore, the FFP signal corresponded to Mac1-expressing cells. These observations are consistent with tumor-infiltrating macrophages being responsible for the increase in folate receptor probe signal in the polyps. Different populations of macrophages are expected to infiltrate the gut. Indeed, the intensity of the FFP probe was higher among Mac1positive cells infiltrating tumor as compared with those residing in healthy intestine. To further define the subpopulation of macrophages that expressed folate recep-

tor, direct staining for the receptor was performed in conjunction with stainings for Mac1 and F4/80. F4/80 has been produced by fusing spleen cells from a rat hyperimmunized with cultured thioglycolate-induced mouse peritoneal macrophages with a mouse myeloma, and therefore may be a marker for inflammation-associated activated macrophages [13]. The expression of folate receptor among the Mac1-positive population of cells corresponded accurately with the level of expression of F4/80, identifying the Mac1- and F4/80-positive subpopulation of macrophages as the major source of FFP signal emanating from the adenomas. The current study used a charge-coupled device (CCD) camera to examine the bowel specimen in the reflectance model [12]. This is similar to the technology used for fiber-optic endoscopy. The current detection technology could thus be adapted to real-time endoscopy or multiwavelength channel endoscopy as recently reported [17,18]. The adaptation would essentially require a separate light source, appropriate filters, and a sensitive CCD camera for detection. With advanced technology for single-photon counting or very low noise Molecular Imaging . Vol. 4, No. 1, January 2005


73

Figure 5. Flow cytometry analysis of healthy and of APC intestines. (A and B) Mac1 (CD11b)-positive cells in healthy bowel (B6) and in dysplastic adenomas (APC) were analyzed in total leukocytes from pooled intestines from two mice in each case. Three populations, (a) Mac1-negative, (b) intermediate, and (c) high, were identified. Each subpopulation was then analyzed for binding to the FFP. Note the increased frequency of Mac1 and folate receptor (FR)-positive cells in APC as compared with healthy intestine. (C) Macrophages were identified by double staining for Mac1 and F4/80. The double-positive population of cells was gated and analyzed for expression of folate receptor by staining with specific monoclonal antibody. Expression of FR correlated well with the level of F4/80 expression (compare C, b with c). The FACS data represent at least three separate experiments.

detection systems, it is also potentially feasible to record the molecular signatures from outside the abdomen. Recently, optical tomography with near-infrared light has been described [15] and facilitated by rigorous mathematical modeling of light propagation in tissue and technological advancements in photon sources and detection techniques. These recent developments expand potential applications of the described probe to the imaging of molecular targets in vivo. In conclusion, we have demonstrated that the in vivo detection of macrophage infiltration of dysplastic intestinal adenoma by optical imaging is feasible, using a folate-targeted fluorescence probe. This detection technology can be adapted to fluorescence endoscopy or intraoperative optical imaging methods for screening of suspicious lesions and biological response to therapy.

Acknowledgments This research was supported by Department of Defense DAMD17-02-10361, DAMD17-03-1-0210, and by R01-CA104547-01A1 (KK), NIH P50-


CA86355 (RW), RO1-CA99385, R33-CA88365, and NSF BES-0119382 (CT). We thank Dr. Roderick Bronson for histological analysis.

References [1] Winawer S, Fletcher R, Rex D, Bond J, Burt R, Ferrucci J, et al. (2003). Colorectal cancer screening and surveillance: Clinical guidelines and rationale—Update based on new evidence. Gastroenterology. 124:544 – 560. [2] Pasricha PJ, Motamedi M (2002). Optical biopsies, ‘‘bioendoscopy,’’ and why the sky is blue: The coming revolution in gastrointestinal imaging. Gastroenterology. 122:571 – 575. [3] Marten K, Bremer C, Khazaie K, Sameni M, Sloane B, Tung CH, et al. (2002). Detecting of dysplastic intestinal adenomas using enzyme-sensing molecular beacons in mice. Gastroenterology. 122:406 – 414. [4] Nakashima-Matsushita N, Homma T, Yu S, Matsuda T, Sunahara N, Nakamura T, Tsukano M, Ratnam M, Matsuyama T (1999). Selective expression of folate receptor beta and its possible role in methotrexate transport in synovial macrophages from patients with rheumatoid arthritis. Arthritis Rheum. 42:1609 – 1616. [5] Antony AC (1996). Folate receptors. Annu Rev Nutr. 16: 501 – 521. [6] Kamen BA, Wang MT, Streckfuss AJ, Peryea X, Anderson RG (1988). Delivery of folates to the cytoplasm of MA104 cells is mediated by a surface membrane receptor that recycles. J Biol Chem. 263:13602 – 13609.

74


[7] Ke CY, Mathias CJ, Green MA (2004). Folate-receptortargeted radionuclide imaging agents. Adv Drug Deliv Rev. 56: 1143 – 1160. [8] Tung CH, Lin Y, Moon WK, Weissleder R (2002). A receptortargeted near-infrared fluorescence probe for in vivo tumor imaging. ChemBioChem. 3:784 – 786. [9] Moon WK, Lin Y, O’Loughlin T, Tang Y, Kim DE, Weissleder R, Tung CH (2003). Enhanced tumor detection using a folate receptor-targeted near-infrared fluorochrome conjugate. Bioconjug Chem. 14:539 – 545. [10] Lin Y, Weissleder R, Tung CH (2002). Novel near-infrared cyanine fluorochromes: Synthesis, properties, and bioconjugation. Bioconjug Chem. 13:605 – 610. [11] Walker EB, Akporiaye ET, Warner NL, Stewart CC (1985). Characterization of subsets of bone marrow-derived macrophages by flow cytometry analysis. J Leukoc Biol. 37:121 – 136. [12] Mahmood U, Tung CH, Bogdanov A, Jr., Weissleder R (1999). Near-infrared optical imaging of protease activity for tumor detection. Radiology. 213:866 – 870. [13] Austyn JM, Gordon S (1981). F4/80, a monoclonal antibody

[14]

[15]

[16]

[17]

[18]

directed specifically against the mouse macrophage. Eur J Immunol. 11:805 – 815. Licha K, Riefke B, Ntziachristos V, Becker A, Chance B, Semmler W (2000). Hydrophilic cyanine dyes as contrast agents for nearinfrared tumor imaging: Synthesis, photophysical properties and spectroscopic in vivo characterization. Photochem Photobiol. 72:392 – 398. Ntziachristos V, Yodh AG, Schnall M, Chance B (2000). Concurrent MRI and diffuse optical tomography of breast after indocyanine green enhancement. Proc Natl Acad Sci USA. 97: 2767 – 2772. Chen WT, Mahmood U, Weissleder R, Tung CH (2005). Arthritis imaging using a near-infrared fluorescence folate-targeted probe. Arthritis Res Ther. 7:R310 – R317. Funovics MA, Weissleder R, Mahmood U (2004). Catheter-based in vivo imaging of enzyme activity and gene expression: Feasibility study in mice. Radiology. 231:659 – 666. Kelly K, Alencar H, Funovics M, Mahmood U, Weissleder R (2004). Detection of invasive colon cancer using a novel, targeted, library-derived fluorescent peptide. Cancer Res. 64:6247 – 6251.
