RESEARCH ARTICLE
Breast Cancer Imaging Using the Near-Infrared Fluorescent Agent, CLR1502 Melissa L. Korb, Jason M. Warram, Joseph Grudzinski, Jamey Weichert, Justin Jeffery, and Eben L. Rosenthal
Abstract Positive margins after breast conservation surgery represent a significant problem in the treatment of breast cancer. The nearinfrared fluorescence agent CLR1502 (Cellectar Biosciences, Madison, WI) was studied in a preclinical breast cancer model to determine imaging properties and ability to detect small islands of malignancy. Nude mice bearing human breast cancer flank xenografts were given a systemic injection of CLR1502, and imaging was performed using LUNA (Novadaq Technologies Inc., Richmond, BC) and Pearl Impulse (LI-COR Biosciences, Lincoln, NE) devices. Normal tissues were examined for fluorescence signal, and conventional and fluorescence histology was performed using the Odyssey scanner. Peak tumor to background ratio occurred 2 days after injection with CLR1502. The smallest amount of tumor that was imaged and detected using these devices was 1.9 mg, equivalent to 1.9 3 106 cells. The highest fluorescence signal was seen in tumor and normal lymph node tissue, and the lowest fluorescence signal was seen in muscle and plasma. Human breast cancer tumors can be imaged in vivo with multiple optical imaging platforms using CLR1502. This pilot study supports further investigations of this fluorescent agent for improving surgical resection of malignancies, with the goal of eventual clinical translation.
NCOLOGIC RESECTION remains the primary treatment modality for solid tumors, but effectiveness is limited by the inability to obtain clear margins. In addition to visual inspection and palpation by the surgeon, current strategies for identifying tumor margins intraoperatively include conventional frozen section or permanent section histology, both of which require removal of several centimeters of adjacent normal tissue to obtain margins that are pathologically free of disease. Removal of this additional tissue adjacent to the tumor has the potential to negatively impact functional and cosmetic outcomes.1,2 Therefore, there is an opportunity to improve detection of nonpalpable, subclinical disease through precise visualization of tumor boundaries without excessive removal of adjacent normal tissue. The most widely proposed technique for real-time surgical imaging of cancer margins remains the use of optical near-infrared (NIR) imaging and intraoperative magnetic resonance
O
From the Department of Surgery, University of Alabama at Birmingham, Birmingham, AL; Cellectar Biosciences, Madison, WI; and Department of Radiology, University of Wisconsin, Madison, WI. Address reprint requests to: Melissa L. Korb, MD, Department of Surgery, University of Alabama at Birmingham, 1922 7th Avenue South, KB 217, Birmingham, AL 35294; e-mail: [email protected].
DOI 10.2310/7290.2014.00040 #
2014 Decker Intellectual Properties
imaging (MRI).3,4 With respect to NIR imaging, the depth of penetration (5–10 mm), wide field, and nontoxic nature of fluorescence-based imaging make it ideal for the surgical setting.5–9 Several clinical trials in this area have defined potential benefit10 or are currently ongoing (clinicaltrials. gov, identifiers NCT0197273, NCT01508572, and NCT 01987375). Breast conservation surgery, by definition, attempts to minimize the ablative defect and provides an especially relevant platform for NIR-based applications. Positive margins in breast cancer are defined as tumor cells present within 2 mm of the surgical margin.11 Unfortunately, breast conservation surgery is associated with a positive margin rate of 20 to 60%,1,11 requiring 15 to 60% of these cases to undergo reexcision.12,13 Thus, there is significant opportunity for improvement over current techniques such as frozen section and wire-guided localization.14 The clinical introduction of CLR1404 in positron emission tomographic (PET) imaging of solid tumors provides motivation for developing a structurally similar compound, CLR1502 (Cellectar Biosciences, Madison, WI), for NIR fluorescence imaging. CLR1502 is an alkylphosphocholine compound that targets cellular and intracellular membranes. Through disruption of cholesterol homeostasis, the phospholipid carrier of CLR1502 is incorporated into cancer cells by cholesterol-rich plasma membrane lipid rafts, which are present in increased amounts in cancer cells over
Molecular Imaging, Vol 13, 2014: pp 1–9
1
2
Korb et al
normal cells.15 Furthermore, retention of the agent is prolonged secondary to deficiency of phospholipid catabolizing agents in tumor cells compared to normal cells in various cancer types, including ovary, colon, melanoma, lung, and breast.16–18 Considerable research has been done with analogous agents in PET imaging19–21 (clinicaltrials.gov, identifiers NCT01516905, NCT01540513, NCT00582283, NCT01662284, NCT01898273) and has shown them to be well tolerated, without evidence of acute toxicity at high doses.21 However, optical fluorescence imaging represents a relatively new area of application. Implementation of these innovative imaging agents in the operating room would be facilitated by the use of established hardware. To this end, we assessed the potential of a currently available intraoperative device (LUNA/SPY, Novadaq Technologies Inc., Richmond, BC) optimized for indocyanine green (ICG) (excitation and emission 780 and 830 nm, respectively) to detect CLR1502 (excitation and emission 767 and 796 nm, respectively). LUNA is available at most tertiary care facilities for fluorescence angiography in plastic surgery procedures. Additionally, the Pearl Impulse device (LI-COR Biosciences, Lincoln, NE) is a small animal imaging system designed to image the IRDye products, but it also has spectral overlap with CLR1502. Pearl has the advantage of an enclosed system that can be used to assess tissue fluorescence with greater flexibility and sensitivity.6,22 In this study, we used the LUNA and Pearl instruments to characterize CLR1502 localization within tumors in a preclinical model.
Materials and Methods Cell Lines and Tissue Culture 2LMP (23 lung metastatic pooled) triple-negative human breast cancer cells, derived from MDA-MB-231 (MD Anderson metastatic breast), were obtained from the laboratory of Dr. Donald Buchsbaum at the University of Alabama in Birmingham. Cells were maintained in Dulbecco’s Modified Eagle’s Medium, 10% fetal bovine serum, and 1% L-glutamine. Cells were incubated at 37uC in 5% CO2 and cultured to 80% confluence. Cell number was determined by hemocytometer and trypan dye exclusion, and 2 3 106 cells in phosphate-buffered saline (PBS) were injected subcutaneously on the flank. Reagents The imaging agent used was CLR1502, which was synthesized as previously described.15 CLR1502 was reconstituted
to 0.5 mg/mL and 0.05 mg/mL in 0.4% (v/v) polyoxyethylene sorbitan monolaurate (Polysorbate 20, JT Baker, Avantor, Center Valley, PA), 7% (v/v) ethanol (EMD Chemicals Inc., Gibbstown, NJ), and 0.9% (w/v) sodium chloride (APP Pharmaceuticals, Schaumburg, IL). The final solution was filtered through sterile 0.22 mm syringe filters. Prior to injection, the weight-based dose of CLR1502 was diluted to a total volume of 200 mL in PBS. Animal Models Nude (nu/nu) female mice (Charles River Laboratories, Hartford, CT), age 4 to 6 weeks, were purchased and allowed to acclimate for 1 week before experiments were started. Mice were housed in accordance with the Institutional Animal Care and Use Committee guidelines, and all experiments were conducted and mice were euthanized according to approved protocols. All mice received 200 mL subcutaneous injection of 2LMP cells (2 3 106) suspended in PBS on either the right or left flank. Tumors were allowed to grow for 2 weeks. To determine the time of the peak tumor to background ratio (TBR), mice bearing flank tumors were intravenously injected with either a 3 mg/kg (n 5 2) or a 0.3 mg/kg (n 5 2) dose of CLR1502. The tumors were imaged daily after injection with the clinical imaging system LUNA and the preclinical Pearl Impulse system. LUNA images were analyzed using SPY-Q Image Analysis software (Novadaq Technologies Inc.), and Pearl images were analyzed using Image Studio software (LI-COR Biosciences). Tumor volume was measured by standard caliper assessment twice weekly. One mouse from the 0.5 mg/mL group required early euthanasia due to tumor ulceration on postinjection day 11. The remaining mice were euthanized on postinjection day 21, and tumors and ipsilateral lymph nodes were collected for further analysis. To evaluate the threshold of detection for CLR1502, mice were intravenously injected with a 3 mg/kg dose of CLR1502 (n 5 3). On postinjection day 2, retro-orbital bleeding was done to collect blood from each mouse. Blood was spun at 1,500 revolutions per minute for 10 minutes to collect the plasma fraction. Mice were then euthanized, and serial sectioning of resected tumors was performed. The skin overlying the tumor was incised to create a flap, and the whole tumor was imaged using LUNA and Pearl. In this model, the fluorescence of the wound bed represented the background fluorescence that would be present in a surgical field. The whole tumor was weighed and then dissected in half. One half was placed aside in a tissue cassette. The other was weighed and placed
CLR1502 for Near-Infrared Imaging of Breast Cancer
back into the tumor bed for imaging as described above. The serial sectioning of each tumor proceeded until the tumor could no longer be grossly divided. At the conclusion of serial tumor imaging, multiple other organs were collected for imaging, including lymph nodes, fat, muscle, liver, spleen, kidneys, and skin. Tissues were weighed and placed in individual cassettes for imaging, and 25 mL of plasma was imaged in an Eppendorf tube. Fluorescent Imaging and Measurement Images were captured on the LUNA for 34 seconds with a frame rate of 15 frames/s. The tumor was in the center of the field of view for all animal images. The SPY-Q Image Analysis software used for analysis is specific to the LUNA device. LUNA captures were stacked in SPY-Q to create a single image with total fluorescence values. For daily imaging, one region of interest (ROI) was constructed over the center of the tumor and four additional background ROI were placed in the normal tissue over the ipsilateral side of the mouse at the posterior flank, midbody, shoulder, and neck. For serial sectioning, one ROI was constructed over the center of the tumor and one background ROI was placed next to the tumor in the tissue of the wound bed. Images were exported as bitmap, and total point pixel intensity was recorded. TBR was calculated for each animal at each time point. The daily imaging group TBRs represent transcutaneous (TC) values for the tumor, whereas TBRs for the serial sectioning group represent a direct value with the skin removed. For each tissue examined ex vivo, an ROI was constructed over the tissue of interest. Images were exported as bitmap, and total point pixel intensity was recorded. To ensure consistency, one author performed all ROI measurements and calculations. Pearl images were collected on the 800 nm channel and white filters with the settings resolution 85 mm and focus 0. The tumor was in the center of the field of view for all animal images. The Image Studio software used for analysis is specific to the Pearl device. For daily imaging, one ROI was constructed over the center of the tumor and four additional background ROI were placed in the normal tissue over the ipsilateral side of the mouse at the posterior flank, midbody, shoulder, and neck. For serial sectioning, one ROI was constructed over the center of the tumor and one background ROI was placed next to the tumor in the tissue of the wound bed. Total counts were recorded for each ROI. All ROI were the same area. Background ROI were averaged, and TBR was calculated for each animal at each time point. The daily imaging group TBRs represent
3
TC values for the tumor, whereas TBRs for the serial sectioning group represent a direct value with the skin removed. For each tissue examined ex vivo, size-matched ROI were constructed over the tissue of interest. Total counts were recorded. To ensure consistency, one author performed all ROI measurements and calculations. Histology Evaluation Conventional hematoxylin-eosin (H&E)-stained slides were examined under 103 magnification with a Fisher Micromaster Scientific Microscope (120V, 12-561B, Fisher Scientific, Fair Lawn, NJ). Unstained slides corresponding to the H&E slides were placed in the Odyssey scanner (LICOR Biosciences) to generate fluorescent histology images of the ipsilateral lymph nodes, contralateral lymph nodes, tumor, skin, and muscle. The intensity was standardized for all images. Images were analyzed using Image Studio software. Five size-matched ROI were constructed and evenly distributed within the tissue of interest. Total counts were recorded for each ROI and averaged for each tissue. To ensure consistency, one author performed all ROI measurements and calculations. Statistical Analysis To determine differences in tumor volume at baseline between the daily imaging and serial sectioning groups, volumes were compared using a two-tailed t-test (Excel 2010, Microsoft Corporation, Redmond, WA). Tumor fluorescence was compared to background fluorescence for each time point in the daily imaging and serial sectioning studies also using two-tailed t-tests. Statistical significance was considered p , .05.
Results Tumor Properties The average starting volume of the tumors was 184 mm3 6 80 mm3 for the high-dose cohort and 108 mm3 6 83 mm3 for the low-dose cohort. The average final volume of the tumors was 971 mm3 6 923 mm3 for the high-dose group and 1,263 mm3 6 993 mm3 for the low-dose group. Average tumor volume prior to serial sectioning was 193 mm3 6 28 mm3, and average weight was 77.4 mg 6 23.7 mg. The starting tumor volumes for the daily imaging and serial sectioning groups were not significantly different (p 5 .11).
4
Korb et al
NIR Fluorescence Imaging of Tumors Following a single systemic administration of CLR1502, tumor fluorescence in the high-dose group (0.5 mg/mL) was significantly greater than that for the tumors in the low-dose cohort (0.05 mg/mL) for all time points (p , .05) by both LUNA and Pearl. When the tumor fluorescence was divided by the background to obtain the TC TBR (Figure 1), the time points 1 day, 30 hours, and 2 days showed the high-dose group to have a statistically significantly higher TC TBR than the low-dose group by LUNA (p , .05). By Pearl, day 2 and days 12 to 16 and 19 to 21 showed the high-dose group to have a statistically significantly higher TC TBR than the low-dose group (p , .05). The peak TC TBR value for the high-dose cohort was 1.4 by LUNA and 1.5 by Pearl, which was similar to the peak TC TBR value for the low-dose group of 1.3 by both LUNA and Pearl. Peak TC TBRs were achieved near postinjection day 2. Timing and TC TBR values were very consistent between LUNA and Pearl imaging modalities for the high-dose group; however, the low-dose group demonstrated more variability with respect to peak values. Peak TC TBR for the low-dose group was achieved at 72 hours after injection on LUNA but was observed to peak and remain relatively constant between 24 and 72 hours on Pearl. Whereas TC TBR values fell below detection (TBR , 1.0) on day 10 by LUNA for both dosing cohorts, Pearl TC TBRs did not fall below detection until day 12 for the low-dose group and day 16 for the high-dose group. Once TC TBR values fell below 1.0, they remained below detection throughout the remainder of the imaging study. Pathologic sectioning of tumor was used to confirm that fluorescence correlated with the presence of tumor. Determination of Detection Threshold for NIR Fluorescence Imaging To determine the smallest amount of tumor that can be detected in a surgical wound bed using CLR1502 and the Food and Drug Administration–approved LUNA device, the tumor was excised and then serially divided. Pearl was used as a validation tool for LUNA imaging. Prior to serial sectioning, the TC TBR was 1.8 for LUNA and Pearl. After removal of the overlying skin, the direct TBR values (using the wound bed as background) for LUNA and Pearl were 1.8 and 2.1, respectively. Serial sectioning of tumor tissue resulted in a decline in direct TBR as tumor size decreased (Figure 2); however, tumor fluorescence remained statistically significantly higher than the background tumor bed
fluorescence for each division by both LUNA and Pearl (p , .05). The smallest fragment of tumor that was imaged weighed an average of 1.9 mg and was detectable by both LUNA and Pearl, with TBR values of 1.4 and 1.6, respectively. The value 1.9 mg can be extrapolated to equal 1.9 3 106 cells23 that were visible. Fluorescence Uptake in Nontarget Tissues In the daily imaging, detection threshold, and fluorescent histology studies, lymph nodes were highly fluorescent, although malignant disease could not be identified. The daily imaging study demonstrated signal-to-background ratio (SBR) values for the lymph nodes in both the highand low-dose groups that remained 1.3 to 1.8 from 7 hours after injection until day 21. SBR values for lymph nodes were 1.5 for the detection threshold study. These values are close to and, in the case of the daily imaging experiment, greater than the peak TBR values for the tumors in each study. When tissues were examined ex vivo, the lymph nodes ipsilateral to the flank tumors had the highest fluorescence values, followed (in descending order) by normal skin, contralateral lymph nodes, kidneys, spleen, fat, liver, muscle, and plasma (Figure 3). Microscopic evaluation of all nontumor tissues revealed normal histology without evidence of metastasis or tumor invasion. As demonstrated in Figure 4, fluorescent histology images obtained with the Odyssey scanner show increased fluorescence in the normal lymph nodes relative to the tumor, skin, and muscle.
Discussion Alkyl phospholipids are a class of compounds that are known to be selectively taken up and retained by tumor cells. These agents have been extensively studied in PET/ computed tomographic (CT) imaging19–21 and are currently being investigated in multiple clinical trials but have only recently been adapted for NIR optical fluorescent imaging. Adaptation of the core molecule to multiple imaging modalities leverages similar toxicology and manufacturing information and facilitates clinical translation. In this pilot study, we examined the innovative NIR fluorescent agent CLR1502 and its ability to detect breast cancer using the wide-field imaging device LUNA, the enclosed imaging device Pearl, and the scanning device Odyssey. After systemic injection, CLR1502 signal within the tumor using the Pearl and LUNA systems was found to peak at 2 days, with maximum TBR values of 1.8. Tumor signal fell below background after 1.5 to 2 weeks. As few as
CLR1502 for Near-Infrared Imaging of Breast Cancer
5
Figure 1. Daily imaging of CLR1502 tumors. Transcutaneous tumor to background ratio (TC TBR) for the high-dose (0.5 mg/mL) and low-dose (0.05 mg/mL) groups using the (A) LUNA and (B) Pearl Impulse imaging systems. Error bars represent standard deviation. C, Representative LUNA and Pearl images from the high-dose group are shown for several points over the course of the study. Arrows indicate the location of the tumor.
6
Korb et al
1.9 3 106 tumor cells were able to be fluorescently detected using this imaging compound with these devices. Strong uptake of this agent was also noted in normal lymph nodes as early as 7 hours postinjection and remaining for several weeks. The timing of peak uptake by the tumor (2 days) and peak uptake by the normal tissues (7 hours) is consistent with studies of radiolabeled alkyl phospholipid agents.21 Fluorescence signal was lowest in muscle tissue and plasma. The rapid expansion of cancer imaging into the NIR field in recent years has led to the development of multiple other agents, including antibody-based, folate-based, and cathepsin-activatable probes. Each of these techniques, in addition to the alkylphosphocholine CLR1502, relies on differential, cancer-specific targeting of the tumor. The application of such different NIR probes to a diverse set of cancers (breast, colon, gastric, head and neck, lung, melanoma, ovarian, prostate) has produced varied results.
For antibody-based imaging, peak TBRs of 1.5 to 9 have been reported,5,6,22,24 folate-based agents cite TBRs of 3 to 12 and a significant uptake of folate agent within cancer cells,25–27 and cathepsin-activatable probes have been shown to produce a 1.8-fold increase in fluorescence signal intensity over normal tissue and TBRs of 8 to 9.28–30 In this study, the observed TBRs of 1.3 to 1.8 were lower than what is reported for these other modalities; however, high variability in TBR for each of these techniques is a byproduct of differences in cancer type, tumor size, imaging modality, and ROI analysis. No form of standardization currently exists to compare results across studies, and the development of industry wide standards is essential prior to large-scale clinical trials.31 Nonspecific uptake in normal tissue such as skin and in normal lymph nodes was a noted limitation of this technique. At certain time points, the normal lymph nodes were observed to have higher fluorescence than the tumor;
Figure 2. Serial sectioning of CLR1502 tumors. Average fluorescence and tumor to background ratio for each serial section of the tumor using the (A) LUNA and (B) Pearl Impulse imaging systems. Error bars represent standard deviation. C, Representative images of serial tumor sections imaged within the wound bed to account for background fluorescence. Weights represent the average across all mice for each serial section.
CLR1502 for Near-Infrared Imaging of Breast Cancer
7
Figure 3. CLR1502 fluorescence in tissues. A, LUNA and Pearl fluorescence for tissues examined ex vivo. Error bars represent standard deviation. B, Representative gross, LUNA, and Pearl images of the tissues examined ex vivo. A 5 skin overlying the tumor; B 5 skin remote to tumor; C 5 muscle; D 5 fat; E 5 tumor fragments from serial sectioning; F 5 ipsilateral lymph nodes; G 5 contralateral lymph nodes; H 5 liver; I 5 spleen; J 5 kidneys; K 5 plasma.
however, it is unknown if the presence of metastatic cancer in the lymph node would further increase fluorescent signal. Additionally, the lower fluorescent values for the tumor may suggest an unexpected quenching effect of CLR1502 as it concentrates within the tumor. Further support for this conclusion is reflected in the relatively early peak in fluorescence at day 2 followed by a steady decline in TBR values to less than 1 as the agent was allowed to circulate and accumulate within the tumor. Certainly, tumor size must also be considered a possible confounder in this process. The potential depth of penetration for NIR is approximately 1 cm32; however, this may be influenced by quantum yield, instrument
parameters, dye quenching, and tissue type. In this case, as the tumors enlarged, attenuation of fluorescent signal may have contributed to the decrease in TBR values. Low absolute TBR values could also be a reflection of using instrumentation not specifically designed around the excitation and emission wavelengths of CLR1502 or from calculating TBR for daily imaging using background from skin, which had high fluorescence. In the operative setting, tumor exposure via skin retraction would limit this artifact, as was seen in the serial sectioning model. Further studies with this agent are warranted to determine the sensibility of this technique in orthotopic models of breast and other cancers, to identify if quenching is
8
Korb et al
Figure 4. Fluorescence histology. A, Odyssey fluorescence for tissues examined on histology slides. Error bars represent standard deviation. B, Representative images by traditional hematoxylin-eosin (H&E) and Odyssey fluorescence microscopy of lymph nodes, tumor, skin, and muscle (310 original magnification).
occurring, and to examine modifications that can be made to improve TBR.
Equipment was donated by Novadaq and LI-COR Biosciences. CLR1502 was supplied through a material transfer agreement with Cellectar Biosciences. Financial disclosure of reviewers: None reported.
Conclusion CLR1502 can be used with current NIR fluorescence imaging devices to visualize breast cancer in the early period after injection. Contrast between fluorescence of the tumor and normal tissue is greatest for tumors located in or adjacent to muscle and fat tissues, such as breast cancers at the level of the pectoralis. Additionally, the dual capacity of alkyl phospholipid agents for PET and NIR imaging allows for both tumor detection and margin assessment, thereby enhancing clinical impact and facilitating clinical translation.
Acknowledgments Financial disclosure of authors: This work was supported by grants from the National Institutes of Health (5T32CA091078-12).
References 1. Ruidiaz ME, Blair SL, Kummel AC, Wang-Rodriguez J. Computerized decision support system for intraoperative analysis of margin status in breast conservation therapy. ISRN Surg 2012; 2012:546721. 2. Weinstock YE, Alava I 3rd, Dierks EJ. Pitfalls in determining head and neck surgical margins. Oral Maxillofac Surg Clin North Am 2014;26:151–62, doi:10.1016/j.coms.2014.01.003. 3. Vahrmeijer AL, Hutteman M, van der Vorst JR, et al. Imageguided cancer surgery using near-infrared fluorescence. Nat Rev Clin Oncol 2013;10:507–18, doi:10.1038/nrclinonc.2013.123. 4. Rosenthal EL, Warram JM, Bland KI, Zinn KR. The status of contemporary image-guided modalities in oncologic surgery. Ann Surg 2014. [In press] 5. Terwisscha van Scheltinga AG, van Dam GM, Nagengast WB, et al. Intraoperative near-infrared fluorescence tumor imaging with
CLR1502 for Near-Infrared Imaging of Breast Cancer
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
vascular endothelial growth factor and human epidermal growth factor receptor 2 targeting antibodies. J Nucl Med 2011;52:1778– 85, doi:10.2967/jnumed.111.092833. Day KE, Beck LN, Heath CH, et al. Identification of the optimal therapeutic antibody for fluorescent imaging of cutaneous squamous cell carcinoma. Cancer Biol Ther 2013;14:271–7, doi: 10.4161/cbt.23300. Korb ML, Hartman YE, Kovar J, et al. Use of monoclonal antibody-IRDye800CW bioconjugates in the resection of breast cancer. J Surg Res 2014;188:119–28, doi:10.1016/j.jss.2013.11.1089. Adams KE, Ke S, Kwon S, et al. Comparison of visible and nearinfrared wavelength-excitable fluorescent dyes for molecular imaging of cancer. J Biomed Opt 2007;12:024017, doi:10.1117/ 1.2717137. Marshall MV, Draney D, Sevick-Muraca EM, Olive DM. Singledose intravenous toxicity study of IRDye 800CW in SpragueDawley rats. Mol Imaging Biol 2010;12:583–94, doi:10.1007/ s11307-010-0317-x. Stummer W, Pichlmeier U, Meinel T, et al. Fluorescence-guided surgery with 5-aminolevulinic acid for resection of malignant glioma: a randomised controlled multicentre phase III trial. Lancet Oncol 2006;7:392–401, doi:10.1016/S1470-2045(06)70665-9. Au GH, Shih WY, Shih WH, et al. Assessing breast cancer margins ex vivo using aqueous quantum-dot-molecular probes. Int J Surg Oncol 2012;2012:861257. Jung W, Kang E, Kim SM, et al. Factors associated with re-excision after breast-conserving surgery for early-stage breast cancer. J Breast Cancer 2012;15:412–9, doi:10.4048/jbc.2012.15.4.412. Yu CC, Chiang KC, Kuo WL, et al. Low re-excision rate for positive margins in patients treated with ultrasound-guided breastconserving surgery. Breast 2013;22:698–702, doi:10.1016/j.breast. 2012.12.019. Pleijhuis RG, Graafland M, de Vries J, et al. Obtaining adequate surgical margins in breast-conserving therapy for patients with early-stage breast cancer: current modalities and future directions. Ann Surg Oncol 2009;16:2717–30, doi:10.1245/s10434-009-0609-z. Weichert JP, Clark PA, Kandela IK, et al. Alkylphosphocholine analogs for broad spectrum cancer imaging and therapy. Sci Transl Med 2014;6:240ra75, doi:10.1126/scitranslmed.3007646. Plotzke KP, Fisher SJ, Wahl RL, et al. Selective localization of a radioiodinated phospholipid ether analog in human tumor xenografts. J Nucl Med 1993;34:787–92. Plotzke KP, Rampy MA, Meyer K, et al. Biodistribution, metabolism, and excretion of radioiodinated phospholipid ether analogs in tumor-bearing rats. J Nucl Biol Med 1993;37:264–72. Rampy MA, Brown RS, Pinchuk AN, et al. Biological disposition and imaging of a radioiodinated alkylphosphocholine in two rodent models of breast cancer. J Nucl Med 1996;37:1540–5. Grudzinski JJ, Yoriyaz H, Deluca PM Jr, Weichert JP. Patient specific treatment planning for systemically administered
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
9
radiopharmaceuticals using PET/CT and Monte Carlo simulation. Appl Radiat Isot 2010;68:59–65, doi:10.1016/j.apradiso.2009. 09.070. Grudzinski JJ, Floberg JM, Mudd SR, et al. Application of a wholebody pharmacokinetic model for targeted radionuclide therapy to NM404 and FLT. Phys Med Biol 2012;57:1641–57, doi:10.1088/ 0031-9155/57/6/1641. Pinchuk AN, Rampy MA, Longino MA, et al. Synthesis and structure-activity relationship effects on the tumor avidity of radioiodinated phospholipid ether analogues. J Med Chem 2006; 49:2155–65, doi:10.1021/jm050252g. Korb ML, Hartman YE, Kovar J, et al. Use of monoclonal antibody-IRDye800CW bioconjugates in the resection of breast cancer. J Surg Res 2014;188:119–28, doi:10.1016/j.jss.2013.11.1089. Epub 2013 Nov 22. Kim H, Chaudhuri TR, Buchsbaum DJ, et al. High-resolution single-photon emission computed tomography and x-ray computed tomography imaging of Tc-99m-labeled anti-DR5 antibody in breast tumor xenografts. Mol Cancer Ther 2007;6:866–75, doi:10.1158/1535-7163.MCT-06-0230. Day KE, Beck LN, Deep NL, et al. Fluorescently labeled therapeutic antibodies for detection of microscopic melanoma. Laryngoscope 2013;123:2681–9, doi:10.1002/lary.24102. Epub 2013 Apr 24. van Dam GM, Themelis G, Crane LM, et al. Intraoperative tumorspecific fluorescence imaging in ovarian cancer by folate receptoralpha targeting: first in-human results. Nat Med 2011;17:1315–9, doi:10.1038/nm.2472. Sun JY, Shen J, Thibodeaux J, et al. In vivo optical imaging of folate receptor-beta in head and neck squamous cell carcinoma. Laryngoscope 2014;124:E312–9, doi:10.1002/lary.24606. Epub 2014 Mar 4. Heidari Majd M, Barar J, Asgari D, et al. Targeted fluoromagnetic nanoparticles for imaging of breast cancer mcf-7 cells. Adv Pharm Bull 2013;3:189–95. Ding S, Eric Blue R, Chen Y, et al. Molecular imaging of gastric neoplasia with near-infrared fluorescent activatable probes. Mol Imaging 2012 l11:507–15. Alencar H, Funovics MA, Figueiredo J, et al. Colonic adenocarcinomas: near-infrared microcatheter imaging of smart probes for early detection—study in mice. Radiology 2007;244:232–8, doi: 10.1148/radiol.2441052114. Figueiredo JL, Alencar H, Weissleder R, Mahmood U. Near infrared thoracoscopy of tumoral protease activity for improved detection of peripheral lung cancer. Int J Cancer 2006;118:2672–7. Rosenthal EL, Zinn KR. Putting numbers to fluorescent guided surgery. Mol Imaging Biol 2013;15:647–8, doi:10.1007/s11307-0130662-7. Keereweer S, Van Driel PB, Snoeks TJ, et al. Optical image-guided cancer surgery: challenges and limitations. Clin Cancer Res 2013; 19:3745–54, doi:10.1158/1078-0432.CCR-12-3598.
Breast Cancer Imaging Using the Near-Infrared Fluorescent Agent, CLR1502 Melissa L. Korb, Jason M. Warram, Joseph Grudzinski, Jamey Weichert, Justin Jeffery, and Eben L. Rosenthal
Abstract Positive margins after breast conservation surgery represent a significant problem in the treatment of breast cancer. The nearinfrared fluorescence agent CLR1502 (Cellectar Biosciences, Madison, WI) was studied in a preclinical breast cancer model to determine imaging properties and ability to detect small islands of malignancy. Nude mice bearing human breast cancer flank xenografts were given a systemic injection of CLR1502, and imaging was performed using LUNA (Novadaq Technologies Inc., Richmond, BC) and Pearl Impulse (LI-COR Biosciences, Lincoln, NE) devices. Normal tissues were examined for fluorescence signal, and conventional and fluorescence histology was performed using the Odyssey scanner. Peak tumor to background ratio occurred 2 days after injection with CLR1502. The smallest amount of tumor that was imaged and detected using these devices was 1.9 mg, equivalent to 1.9 3 106 cells. The highest fluorescence signal was seen in tumor and normal lymph node tissue, and the lowest fluorescence signal was seen in muscle and plasma. Human breast cancer tumors can be imaged in vivo with multiple optical imaging platforms using CLR1502. This pilot study supports further investigations of this fluorescent agent for improving surgical resection of malignancies, with the goal of eventual clinical translation.
NCOLOGIC RESECTION remains the primary treatment modality for solid tumors, but effectiveness is limited by the inability to obtain clear margins. In addition to visual inspection and palpation by the surgeon, current strategies for identifying tumor margins intraoperatively include conventional frozen section or permanent section histology, both of which require removal of several centimeters of adjacent normal tissue to obtain margins that are pathologically free of disease. Removal of this additional tissue adjacent to the tumor has the potential to negatively impact functional and cosmetic outcomes.1,2 Therefore, there is an opportunity to improve detection of nonpalpable, subclinical disease through precise visualization of tumor boundaries without excessive removal of adjacent normal tissue. The most widely proposed technique for real-time surgical imaging of cancer margins remains the use of optical near-infrared (NIR) imaging and intraoperative magnetic resonance
O
From the Department of Surgery, University of Alabama at Birmingham, Birmingham, AL; Cellectar Biosciences, Madison, WI; and Department of Radiology, University of Wisconsin, Madison, WI. Address reprint requests to: Melissa L. Korb, MD, Department of Surgery, University of Alabama at Birmingham, 1922 7th Avenue South, KB 217, Birmingham, AL 35294; e-mail: [email protected].
DOI 10.2310/7290.2014.00040 #
2014 Decker Intellectual Properties
imaging (MRI).3,4 With respect to NIR imaging, the depth of penetration (5–10 mm), wide field, and nontoxic nature of fluorescence-based imaging make it ideal for the surgical setting.5–9 Several clinical trials in this area have defined potential benefit10 or are currently ongoing (clinicaltrials. gov, identifiers NCT0197273, NCT01508572, and NCT 01987375). Breast conservation surgery, by definition, attempts to minimize the ablative defect and provides an especially relevant platform for NIR-based applications. Positive margins in breast cancer are defined as tumor cells present within 2 mm of the surgical margin.11 Unfortunately, breast conservation surgery is associated with a positive margin rate of 20 to 60%,1,11 requiring 15 to 60% of these cases to undergo reexcision.12,13 Thus, there is significant opportunity for improvement over current techniques such as frozen section and wire-guided localization.14 The clinical introduction of CLR1404 in positron emission tomographic (PET) imaging of solid tumors provides motivation for developing a structurally similar compound, CLR1502 (Cellectar Biosciences, Madison, WI), for NIR fluorescence imaging. CLR1502 is an alkylphosphocholine compound that targets cellular and intracellular membranes. Through disruption of cholesterol homeostasis, the phospholipid carrier of CLR1502 is incorporated into cancer cells by cholesterol-rich plasma membrane lipid rafts, which are present in increased amounts in cancer cells over
Molecular Imaging, Vol 13, 2014: pp 1–9
1
2
Korb et al
normal cells.15 Furthermore, retention of the agent is prolonged secondary to deficiency of phospholipid catabolizing agents in tumor cells compared to normal cells in various cancer types, including ovary, colon, melanoma, lung, and breast.16–18 Considerable research has been done with analogous agents in PET imaging19–21 (clinicaltrials.gov, identifiers NCT01516905, NCT01540513, NCT00582283, NCT01662284, NCT01898273) and has shown them to be well tolerated, without evidence of acute toxicity at high doses.21 However, optical fluorescence imaging represents a relatively new area of application. Implementation of these innovative imaging agents in the operating room would be facilitated by the use of established hardware. To this end, we assessed the potential of a currently available intraoperative device (LUNA/SPY, Novadaq Technologies Inc., Richmond, BC) optimized for indocyanine green (ICG) (excitation and emission 780 and 830 nm, respectively) to detect CLR1502 (excitation and emission 767 and 796 nm, respectively). LUNA is available at most tertiary care facilities for fluorescence angiography in plastic surgery procedures. Additionally, the Pearl Impulse device (LI-COR Biosciences, Lincoln, NE) is a small animal imaging system designed to image the IRDye products, but it also has spectral overlap with CLR1502. Pearl has the advantage of an enclosed system that can be used to assess tissue fluorescence with greater flexibility and sensitivity.6,22 In this study, we used the LUNA and Pearl instruments to characterize CLR1502 localization within tumors in a preclinical model.
Materials and Methods Cell Lines and Tissue Culture 2LMP (23 lung metastatic pooled) triple-negative human breast cancer cells, derived from MDA-MB-231 (MD Anderson metastatic breast), were obtained from the laboratory of Dr. Donald Buchsbaum at the University of Alabama in Birmingham. Cells were maintained in Dulbecco’s Modified Eagle’s Medium, 10% fetal bovine serum, and 1% L-glutamine. Cells were incubated at 37uC in 5% CO2 and cultured to 80% confluence. Cell number was determined by hemocytometer and trypan dye exclusion, and 2 3 106 cells in phosphate-buffered saline (PBS) were injected subcutaneously on the flank. Reagents The imaging agent used was CLR1502, which was synthesized as previously described.15 CLR1502 was reconstituted
to 0.5 mg/mL and 0.05 mg/mL in 0.4% (v/v) polyoxyethylene sorbitan monolaurate (Polysorbate 20, JT Baker, Avantor, Center Valley, PA), 7% (v/v) ethanol (EMD Chemicals Inc., Gibbstown, NJ), and 0.9% (w/v) sodium chloride (APP Pharmaceuticals, Schaumburg, IL). The final solution was filtered through sterile 0.22 mm syringe filters. Prior to injection, the weight-based dose of CLR1502 was diluted to a total volume of 200 mL in PBS. Animal Models Nude (nu/nu) female mice (Charles River Laboratories, Hartford, CT), age 4 to 6 weeks, were purchased and allowed to acclimate for 1 week before experiments were started. Mice were housed in accordance with the Institutional Animal Care and Use Committee guidelines, and all experiments were conducted and mice were euthanized according to approved protocols. All mice received 200 mL subcutaneous injection of 2LMP cells (2 3 106) suspended in PBS on either the right or left flank. Tumors were allowed to grow for 2 weeks. To determine the time of the peak tumor to background ratio (TBR), mice bearing flank tumors were intravenously injected with either a 3 mg/kg (n 5 2) or a 0.3 mg/kg (n 5 2) dose of CLR1502. The tumors were imaged daily after injection with the clinical imaging system LUNA and the preclinical Pearl Impulse system. LUNA images were analyzed using SPY-Q Image Analysis software (Novadaq Technologies Inc.), and Pearl images were analyzed using Image Studio software (LI-COR Biosciences). Tumor volume was measured by standard caliper assessment twice weekly. One mouse from the 0.5 mg/mL group required early euthanasia due to tumor ulceration on postinjection day 11. The remaining mice were euthanized on postinjection day 21, and tumors and ipsilateral lymph nodes were collected for further analysis. To evaluate the threshold of detection for CLR1502, mice were intravenously injected with a 3 mg/kg dose of CLR1502 (n 5 3). On postinjection day 2, retro-orbital bleeding was done to collect blood from each mouse. Blood was spun at 1,500 revolutions per minute for 10 minutes to collect the plasma fraction. Mice were then euthanized, and serial sectioning of resected tumors was performed. The skin overlying the tumor was incised to create a flap, and the whole tumor was imaged using LUNA and Pearl. In this model, the fluorescence of the wound bed represented the background fluorescence that would be present in a surgical field. The whole tumor was weighed and then dissected in half. One half was placed aside in a tissue cassette. The other was weighed and placed
CLR1502 for Near-Infrared Imaging of Breast Cancer
back into the tumor bed for imaging as described above. The serial sectioning of each tumor proceeded until the tumor could no longer be grossly divided. At the conclusion of serial tumor imaging, multiple other organs were collected for imaging, including lymph nodes, fat, muscle, liver, spleen, kidneys, and skin. Tissues were weighed and placed in individual cassettes for imaging, and 25 mL of plasma was imaged in an Eppendorf tube. Fluorescent Imaging and Measurement Images were captured on the LUNA for 34 seconds with a frame rate of 15 frames/s. The tumor was in the center of the field of view for all animal images. The SPY-Q Image Analysis software used for analysis is specific to the LUNA device. LUNA captures were stacked in SPY-Q to create a single image with total fluorescence values. For daily imaging, one region of interest (ROI) was constructed over the center of the tumor and four additional background ROI were placed in the normal tissue over the ipsilateral side of the mouse at the posterior flank, midbody, shoulder, and neck. For serial sectioning, one ROI was constructed over the center of the tumor and one background ROI was placed next to the tumor in the tissue of the wound bed. Images were exported as bitmap, and total point pixel intensity was recorded. TBR was calculated for each animal at each time point. The daily imaging group TBRs represent transcutaneous (TC) values for the tumor, whereas TBRs for the serial sectioning group represent a direct value with the skin removed. For each tissue examined ex vivo, an ROI was constructed over the tissue of interest. Images were exported as bitmap, and total point pixel intensity was recorded. To ensure consistency, one author performed all ROI measurements and calculations. Pearl images were collected on the 800 nm channel and white filters with the settings resolution 85 mm and focus 0. The tumor was in the center of the field of view for all animal images. The Image Studio software used for analysis is specific to the Pearl device. For daily imaging, one ROI was constructed over the center of the tumor and four additional background ROI were placed in the normal tissue over the ipsilateral side of the mouse at the posterior flank, midbody, shoulder, and neck. For serial sectioning, one ROI was constructed over the center of the tumor and one background ROI was placed next to the tumor in the tissue of the wound bed. Total counts were recorded for each ROI. All ROI were the same area. Background ROI were averaged, and TBR was calculated for each animal at each time point. The daily imaging group TBRs represent
3
TC values for the tumor, whereas TBRs for the serial sectioning group represent a direct value with the skin removed. For each tissue examined ex vivo, size-matched ROI were constructed over the tissue of interest. Total counts were recorded. To ensure consistency, one author performed all ROI measurements and calculations. Histology Evaluation Conventional hematoxylin-eosin (H&E)-stained slides were examined under 103 magnification with a Fisher Micromaster Scientific Microscope (120V, 12-561B, Fisher Scientific, Fair Lawn, NJ). Unstained slides corresponding to the H&E slides were placed in the Odyssey scanner (LICOR Biosciences) to generate fluorescent histology images of the ipsilateral lymph nodes, contralateral lymph nodes, tumor, skin, and muscle. The intensity was standardized for all images. Images were analyzed using Image Studio software. Five size-matched ROI were constructed and evenly distributed within the tissue of interest. Total counts were recorded for each ROI and averaged for each tissue. To ensure consistency, one author performed all ROI measurements and calculations. Statistical Analysis To determine differences in tumor volume at baseline between the daily imaging and serial sectioning groups, volumes were compared using a two-tailed t-test (Excel 2010, Microsoft Corporation, Redmond, WA). Tumor fluorescence was compared to background fluorescence for each time point in the daily imaging and serial sectioning studies also using two-tailed t-tests. Statistical significance was considered p , .05.
Results Tumor Properties The average starting volume of the tumors was 184 mm3 6 80 mm3 for the high-dose cohort and 108 mm3 6 83 mm3 for the low-dose cohort. The average final volume of the tumors was 971 mm3 6 923 mm3 for the high-dose group and 1,263 mm3 6 993 mm3 for the low-dose group. Average tumor volume prior to serial sectioning was 193 mm3 6 28 mm3, and average weight was 77.4 mg 6 23.7 mg. The starting tumor volumes for the daily imaging and serial sectioning groups were not significantly different (p 5 .11).
4
Korb et al
NIR Fluorescence Imaging of Tumors Following a single systemic administration of CLR1502, tumor fluorescence in the high-dose group (0.5 mg/mL) was significantly greater than that for the tumors in the low-dose cohort (0.05 mg/mL) for all time points (p , .05) by both LUNA and Pearl. When the tumor fluorescence was divided by the background to obtain the TC TBR (Figure 1), the time points 1 day, 30 hours, and 2 days showed the high-dose group to have a statistically significantly higher TC TBR than the low-dose group by LUNA (p , .05). By Pearl, day 2 and days 12 to 16 and 19 to 21 showed the high-dose group to have a statistically significantly higher TC TBR than the low-dose group (p , .05). The peak TC TBR value for the high-dose cohort was 1.4 by LUNA and 1.5 by Pearl, which was similar to the peak TC TBR value for the low-dose group of 1.3 by both LUNA and Pearl. Peak TC TBRs were achieved near postinjection day 2. Timing and TC TBR values were very consistent between LUNA and Pearl imaging modalities for the high-dose group; however, the low-dose group demonstrated more variability with respect to peak values. Peak TC TBR for the low-dose group was achieved at 72 hours after injection on LUNA but was observed to peak and remain relatively constant between 24 and 72 hours on Pearl. Whereas TC TBR values fell below detection (TBR , 1.0) on day 10 by LUNA for both dosing cohorts, Pearl TC TBRs did not fall below detection until day 12 for the low-dose group and day 16 for the high-dose group. Once TC TBR values fell below 1.0, they remained below detection throughout the remainder of the imaging study. Pathologic sectioning of tumor was used to confirm that fluorescence correlated with the presence of tumor. Determination of Detection Threshold for NIR Fluorescence Imaging To determine the smallest amount of tumor that can be detected in a surgical wound bed using CLR1502 and the Food and Drug Administration–approved LUNA device, the tumor was excised and then serially divided. Pearl was used as a validation tool for LUNA imaging. Prior to serial sectioning, the TC TBR was 1.8 for LUNA and Pearl. After removal of the overlying skin, the direct TBR values (using the wound bed as background) for LUNA and Pearl were 1.8 and 2.1, respectively. Serial sectioning of tumor tissue resulted in a decline in direct TBR as tumor size decreased (Figure 2); however, tumor fluorescence remained statistically significantly higher than the background tumor bed
fluorescence for each division by both LUNA and Pearl (p , .05). The smallest fragment of tumor that was imaged weighed an average of 1.9 mg and was detectable by both LUNA and Pearl, with TBR values of 1.4 and 1.6, respectively. The value 1.9 mg can be extrapolated to equal 1.9 3 106 cells23 that were visible. Fluorescence Uptake in Nontarget Tissues In the daily imaging, detection threshold, and fluorescent histology studies, lymph nodes were highly fluorescent, although malignant disease could not be identified. The daily imaging study demonstrated signal-to-background ratio (SBR) values for the lymph nodes in both the highand low-dose groups that remained 1.3 to 1.8 from 7 hours after injection until day 21. SBR values for lymph nodes were 1.5 for the detection threshold study. These values are close to and, in the case of the daily imaging experiment, greater than the peak TBR values for the tumors in each study. When tissues were examined ex vivo, the lymph nodes ipsilateral to the flank tumors had the highest fluorescence values, followed (in descending order) by normal skin, contralateral lymph nodes, kidneys, spleen, fat, liver, muscle, and plasma (Figure 3). Microscopic evaluation of all nontumor tissues revealed normal histology without evidence of metastasis or tumor invasion. As demonstrated in Figure 4, fluorescent histology images obtained with the Odyssey scanner show increased fluorescence in the normal lymph nodes relative to the tumor, skin, and muscle.
Discussion Alkyl phospholipids are a class of compounds that are known to be selectively taken up and retained by tumor cells. These agents have been extensively studied in PET/ computed tomographic (CT) imaging19–21 and are currently being investigated in multiple clinical trials but have only recently been adapted for NIR optical fluorescent imaging. Adaptation of the core molecule to multiple imaging modalities leverages similar toxicology and manufacturing information and facilitates clinical translation. In this pilot study, we examined the innovative NIR fluorescent agent CLR1502 and its ability to detect breast cancer using the wide-field imaging device LUNA, the enclosed imaging device Pearl, and the scanning device Odyssey. After systemic injection, CLR1502 signal within the tumor using the Pearl and LUNA systems was found to peak at 2 days, with maximum TBR values of 1.8. Tumor signal fell below background after 1.5 to 2 weeks. As few as
CLR1502 for Near-Infrared Imaging of Breast Cancer
5
Figure 1. Daily imaging of CLR1502 tumors. Transcutaneous tumor to background ratio (TC TBR) for the high-dose (0.5 mg/mL) and low-dose (0.05 mg/mL) groups using the (A) LUNA and (B) Pearl Impulse imaging systems. Error bars represent standard deviation. C, Representative LUNA and Pearl images from the high-dose group are shown for several points over the course of the study. Arrows indicate the location of the tumor.
6
Korb et al
1.9 3 106 tumor cells were able to be fluorescently detected using this imaging compound with these devices. Strong uptake of this agent was also noted in normal lymph nodes as early as 7 hours postinjection and remaining for several weeks. The timing of peak uptake by the tumor (2 days) and peak uptake by the normal tissues (7 hours) is consistent with studies of radiolabeled alkyl phospholipid agents.21 Fluorescence signal was lowest in muscle tissue and plasma. The rapid expansion of cancer imaging into the NIR field in recent years has led to the development of multiple other agents, including antibody-based, folate-based, and cathepsin-activatable probes. Each of these techniques, in addition to the alkylphosphocholine CLR1502, relies on differential, cancer-specific targeting of the tumor. The application of such different NIR probes to a diverse set of cancers (breast, colon, gastric, head and neck, lung, melanoma, ovarian, prostate) has produced varied results.
For antibody-based imaging, peak TBRs of 1.5 to 9 have been reported,5,6,22,24 folate-based agents cite TBRs of 3 to 12 and a significant uptake of folate agent within cancer cells,25–27 and cathepsin-activatable probes have been shown to produce a 1.8-fold increase in fluorescence signal intensity over normal tissue and TBRs of 8 to 9.28–30 In this study, the observed TBRs of 1.3 to 1.8 were lower than what is reported for these other modalities; however, high variability in TBR for each of these techniques is a byproduct of differences in cancer type, tumor size, imaging modality, and ROI analysis. No form of standardization currently exists to compare results across studies, and the development of industry wide standards is essential prior to large-scale clinical trials.31 Nonspecific uptake in normal tissue such as skin and in normal lymph nodes was a noted limitation of this technique. At certain time points, the normal lymph nodes were observed to have higher fluorescence than the tumor;
Figure 2. Serial sectioning of CLR1502 tumors. Average fluorescence and tumor to background ratio for each serial section of the tumor using the (A) LUNA and (B) Pearl Impulse imaging systems. Error bars represent standard deviation. C, Representative images of serial tumor sections imaged within the wound bed to account for background fluorescence. Weights represent the average across all mice for each serial section.
CLR1502 for Near-Infrared Imaging of Breast Cancer
7
Figure 3. CLR1502 fluorescence in tissues. A, LUNA and Pearl fluorescence for tissues examined ex vivo. Error bars represent standard deviation. B, Representative gross, LUNA, and Pearl images of the tissues examined ex vivo. A 5 skin overlying the tumor; B 5 skin remote to tumor; C 5 muscle; D 5 fat; E 5 tumor fragments from serial sectioning; F 5 ipsilateral lymph nodes; G 5 contralateral lymph nodes; H 5 liver; I 5 spleen; J 5 kidneys; K 5 plasma.
however, it is unknown if the presence of metastatic cancer in the lymph node would further increase fluorescent signal. Additionally, the lower fluorescent values for the tumor may suggest an unexpected quenching effect of CLR1502 as it concentrates within the tumor. Further support for this conclusion is reflected in the relatively early peak in fluorescence at day 2 followed by a steady decline in TBR values to less than 1 as the agent was allowed to circulate and accumulate within the tumor. Certainly, tumor size must also be considered a possible confounder in this process. The potential depth of penetration for NIR is approximately 1 cm32; however, this may be influenced by quantum yield, instrument
parameters, dye quenching, and tissue type. In this case, as the tumors enlarged, attenuation of fluorescent signal may have contributed to the decrease in TBR values. Low absolute TBR values could also be a reflection of using instrumentation not specifically designed around the excitation and emission wavelengths of CLR1502 or from calculating TBR for daily imaging using background from skin, which had high fluorescence. In the operative setting, tumor exposure via skin retraction would limit this artifact, as was seen in the serial sectioning model. Further studies with this agent are warranted to determine the sensibility of this technique in orthotopic models of breast and other cancers, to identify if quenching is
8
Korb et al
Figure 4. Fluorescence histology. A, Odyssey fluorescence for tissues examined on histology slides. Error bars represent standard deviation. B, Representative images by traditional hematoxylin-eosin (H&E) and Odyssey fluorescence microscopy of lymph nodes, tumor, skin, and muscle (310 original magnification).
occurring, and to examine modifications that can be made to improve TBR.
Equipment was donated by Novadaq and LI-COR Biosciences. CLR1502 was supplied through a material transfer agreement with Cellectar Biosciences. Financial disclosure of reviewers: None reported.
Conclusion CLR1502 can be used with current NIR fluorescence imaging devices to visualize breast cancer in the early period after injection. Contrast between fluorescence of the tumor and normal tissue is greatest for tumors located in or adjacent to muscle and fat tissues, such as breast cancers at the level of the pectoralis. Additionally, the dual capacity of alkyl phospholipid agents for PET and NIR imaging allows for both tumor detection and margin assessment, thereby enhancing clinical impact and facilitating clinical translation.
Acknowledgments Financial disclosure of authors: This work was supported by grants from the National Institutes of Health (5T32CA091078-12).
References 1. Ruidiaz ME, Blair SL, Kummel AC, Wang-Rodriguez J. Computerized decision support system for intraoperative analysis of margin status in breast conservation therapy. ISRN Surg 2012; 2012:546721. 2. Weinstock YE, Alava I 3rd, Dierks EJ. Pitfalls in determining head and neck surgical margins. Oral Maxillofac Surg Clin North Am 2014;26:151–62, doi:10.1016/j.coms.2014.01.003. 3. Vahrmeijer AL, Hutteman M, van der Vorst JR, et al. Imageguided cancer surgery using near-infrared fluorescence. Nat Rev Clin Oncol 2013;10:507–18, doi:10.1038/nrclinonc.2013.123. 4. Rosenthal EL, Warram JM, Bland KI, Zinn KR. The status of contemporary image-guided modalities in oncologic surgery. Ann Surg 2014. [In press] 5. Terwisscha van Scheltinga AG, van Dam GM, Nagengast WB, et al. Intraoperative near-infrared fluorescence tumor imaging with
CLR1502 for Near-Infrared Imaging of Breast Cancer
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
vascular endothelial growth factor and human epidermal growth factor receptor 2 targeting antibodies. J Nucl Med 2011;52:1778– 85, doi:10.2967/jnumed.111.092833. Day KE, Beck LN, Heath CH, et al. Identification of the optimal therapeutic antibody for fluorescent imaging of cutaneous squamous cell carcinoma. Cancer Biol Ther 2013;14:271–7, doi: 10.4161/cbt.23300. Korb ML, Hartman YE, Kovar J, et al. Use of monoclonal antibody-IRDye800CW bioconjugates in the resection of breast cancer. J Surg Res 2014;188:119–28, doi:10.1016/j.jss.2013.11.1089. Adams KE, Ke S, Kwon S, et al. Comparison of visible and nearinfrared wavelength-excitable fluorescent dyes for molecular imaging of cancer. J Biomed Opt 2007;12:024017, doi:10.1117/ 1.2717137. Marshall MV, Draney D, Sevick-Muraca EM, Olive DM. Singledose intravenous toxicity study of IRDye 800CW in SpragueDawley rats. Mol Imaging Biol 2010;12:583–94, doi:10.1007/ s11307-010-0317-x. Stummer W, Pichlmeier U, Meinel T, et al. Fluorescence-guided surgery with 5-aminolevulinic acid for resection of malignant glioma: a randomised controlled multicentre phase III trial. Lancet Oncol 2006;7:392–401, doi:10.1016/S1470-2045(06)70665-9. Au GH, Shih WY, Shih WH, et al. Assessing breast cancer margins ex vivo using aqueous quantum-dot-molecular probes. Int J Surg Oncol 2012;2012:861257. Jung W, Kang E, Kim SM, et al. Factors associated with re-excision after breast-conserving surgery for early-stage breast cancer. J Breast Cancer 2012;15:412–9, doi:10.4048/jbc.2012.15.4.412. Yu CC, Chiang KC, Kuo WL, et al. Low re-excision rate for positive margins in patients treated with ultrasound-guided breastconserving surgery. Breast 2013;22:698–702, doi:10.1016/j.breast. 2012.12.019. Pleijhuis RG, Graafland M, de Vries J, et al. Obtaining adequate surgical margins in breast-conserving therapy for patients with early-stage breast cancer: current modalities and future directions. Ann Surg Oncol 2009;16:2717–30, doi:10.1245/s10434-009-0609-z. Weichert JP, Clark PA, Kandela IK, et al. Alkylphosphocholine analogs for broad spectrum cancer imaging and therapy. Sci Transl Med 2014;6:240ra75, doi:10.1126/scitranslmed.3007646. Plotzke KP, Fisher SJ, Wahl RL, et al. Selective localization of a radioiodinated phospholipid ether analog in human tumor xenografts. J Nucl Med 1993;34:787–92. Plotzke KP, Rampy MA, Meyer K, et al. Biodistribution, metabolism, and excretion of radioiodinated phospholipid ether analogs in tumor-bearing rats. J Nucl Biol Med 1993;37:264–72. Rampy MA, Brown RS, Pinchuk AN, et al. Biological disposition and imaging of a radioiodinated alkylphosphocholine in two rodent models of breast cancer. J Nucl Med 1996;37:1540–5. Grudzinski JJ, Yoriyaz H, Deluca PM Jr, Weichert JP. Patient specific treatment planning for systemically administered
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
9
radiopharmaceuticals using PET/CT and Monte Carlo simulation. Appl Radiat Isot 2010;68:59–65, doi:10.1016/j.apradiso.2009. 09.070. Grudzinski JJ, Floberg JM, Mudd SR, et al. Application of a wholebody pharmacokinetic model for targeted radionuclide therapy to NM404 and FLT. Phys Med Biol 2012;57:1641–57, doi:10.1088/ 0031-9155/57/6/1641. Pinchuk AN, Rampy MA, Longino MA, et al. Synthesis and structure-activity relationship effects on the tumor avidity of radioiodinated phospholipid ether analogues. J Med Chem 2006; 49:2155–65, doi:10.1021/jm050252g. Korb ML, Hartman YE, Kovar J, et al. Use of monoclonal antibody-IRDye800CW bioconjugates in the resection of breast cancer. J Surg Res 2014;188:119–28, doi:10.1016/j.jss.2013.11.1089. Epub 2013 Nov 22. Kim H, Chaudhuri TR, Buchsbaum DJ, et al. High-resolution single-photon emission computed tomography and x-ray computed tomography imaging of Tc-99m-labeled anti-DR5 antibody in breast tumor xenografts. Mol Cancer Ther 2007;6:866–75, doi:10.1158/1535-7163.MCT-06-0230. Day KE, Beck LN, Deep NL, et al. Fluorescently labeled therapeutic antibodies for detection of microscopic melanoma. Laryngoscope 2013;123:2681–9, doi:10.1002/lary.24102. Epub 2013 Apr 24. van Dam GM, Themelis G, Crane LM, et al. Intraoperative tumorspecific fluorescence imaging in ovarian cancer by folate receptoralpha targeting: first in-human results. Nat Med 2011;17:1315–9, doi:10.1038/nm.2472. Sun JY, Shen J, Thibodeaux J, et al. In vivo optical imaging of folate receptor-beta in head and neck squamous cell carcinoma. Laryngoscope 2014;124:E312–9, doi:10.1002/lary.24606. Epub 2014 Mar 4. Heidari Majd M, Barar J, Asgari D, et al. Targeted fluoromagnetic nanoparticles for imaging of breast cancer mcf-7 cells. Adv Pharm Bull 2013;3:189–95. Ding S, Eric Blue R, Chen Y, et al. Molecular imaging of gastric neoplasia with near-infrared fluorescent activatable probes. Mol Imaging 2012 l11:507–15. Alencar H, Funovics MA, Figueiredo J, et al. Colonic adenocarcinomas: near-infrared microcatheter imaging of smart probes for early detection—study in mice. Radiology 2007;244:232–8, doi: 10.1148/radiol.2441052114. Figueiredo JL, Alencar H, Weissleder R, Mahmood U. Near infrared thoracoscopy of tumoral protease activity for improved detection of peripheral lung cancer. Int J Cancer 2006;118:2672–7. Rosenthal EL, Zinn KR. Putting numbers to fluorescent guided surgery. Mol Imaging Biol 2013;15:647–8, doi:10.1007/s11307-0130662-7. Keereweer S, Van Driel PB, Snoeks TJ, et al. Optical image-guided cancer surgery: challenges and limitations. Clin Cancer Res 2013; 19:3745–54, doi:10.1158/1078-0432.CCR-12-3598.
