The feasibility of 18F-AlF-NOTA-PRGD2 PET/CT for monitoring early response of Endostar antiangiogenic therapy in human nasopharyngeal carcinoma xenograft model compared with 18F-FDG

Purpose Radiolabeled arginine-glycine-aspartic acid (RGD) peptides have been developed for PET imaging of integrin avβ3 in the tumor vasculature, leading to great potential for noninvasively evaluating tumor angiogenesis and monitoring antiangiogenic treatment. The aim of this study was to investigate a novel one-step labeled integrin-targeted tracer, 18F-AlF-NOTA-PRGD2, for PET/CT for detecting tumor angiogenesis and monitoring the early therapeutic efficacy of antiangiogenic agent Endostar in human nasopharyngeal carcinoma (NPC) xenograft model. Experimental design and results Mice bearing NPC underwent 18F-AlF-NOTA-PRGD2 PET/CT at baseline and after 2, 4, 7, and 14 days of consecutive treatment with Endostar or PBS, compared with 18F-FDG PET/CT. Tumors were harvested at all imaging time points for histopathological analysis with H & E and microvessel density (MVD) and integrin avβ3 immunostaining. The maximum percent injected dose per gram of body weight (%ID/gmax) tumor uptake of 18F-AlF-NOTA-PRGD2 PET/CT was significantly lower than that in the control group starting from day 2 (p < 0.01), much earlier and more accurately than that of 18F-FDG PET/CT. Moreover, a moderate linear correlation was observed between tumor MVD and the corresponding tumor uptake of 18F-AlF-NOTA-PRGD2 PET/CT (r = 0.853, p < 0.01). Conclusions 18F-AlF-NOTA-PRGD2 PET/CT can be used for in vivo angiogenesis imaging and monitoring early response to Endostar antiangiogenic treatment in NPC xenograft model, favoring its potential clinical translation.


INTRODUCTION
Angiogenesis, the formation of new blood vessels, has long been recognized to play a critical role in tumor growth, invasion, and metastasis [1,2]. Targeting of tumor angiogenesis pathway has been exploited as an effective therapeutic approach for cancer treatment [3][4][5]. Nasopharyngeal carcinoma (NPC), one of the most common malignancies of the head and neck in China and Southeast Asia, is highly vascularized. Moreover, some studies have investigated the potential of antiangiogenic agents alone or in combination with chemoradiotherapy in advanced stages of NPC with encouraging results [6][7][8][9]. However, not all patients would benefit from a specific antiangiogenic therapy and some patients may not respond [10,11]. It is therefore in great demand to develop an alternative Oncotarget 27244 www.impactjournals.com/oncotarget approach to identify patients who most likely benefit from antiangiogenic treatment, to detect emerging resistance, and to monitor early therapeutic efficacy [12].
Histopathologic evaluation of microvessel density (MVD), a prognostic indicator of progression, is not practical for routinely evaluating tumor angiogenesis due to its invasive nature of the procedure [13]. Noninvasive imaging technologies such as dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) or computed tomography (CT), used to provide evidence on tumor blood flow, permeability, and volume, are technically challenging and cannot directly and effectively quantify the changes of post-treatment tumoral vascularity [14][15][16]. Positron emission tomography (PET) using 18 F-FDG (2-deoxy-2-18F-fluoro-D-glucose) also was used to monitor antiangiogenic therapy by determining glucose metabolism changes, but it may not be an ideal modality because it is not a tumor-specific radiotracer. Therefore, molecular imaging targeting specific pathways involved in angiogenesis is warranted for specific monitoring of some molecular changes as early therapeutic effects via antiangiogenesis, with the benefit that it allows repetitive noninvasive followups during the course of therapy.
Integrin avβ3, an adhesion molecule responsible for the regulation of tumor angiogenesis and invasion, is highly expressed on activated and proliferating endothelial cells during tumor angiogenesis and several types of cancer cells, but not in quiescent blood vessels [17][18][19]. Moreover, peptides containing arginine-glycine-aspartic acid (RGD) can specifically and strongly bind to integrin avβ3, therefore, a series of radiolabeled RGD peptide probes have been developed in a bid to visualize and quantify integrin αvβ3 expression in vivo and some of them are undergoing clinical trials [20][21][22]. However, only a few of studies have been performed on the use of RGD tracers to evaluate the therapeutic efficacy of antiangiogenic agents, especially in the territory of human NPC [23][24][25][26].
Endostar, a novel modified recombinant human endostatin, approved by the China Food and Drug Administration (CFDA) for the treatment of non-small cell lung cancer (NSCLC) in 2005, has showed strong antiangiogenic effect on a variety of xenotrans planted tumors and cancer patients, including NPC [27][28][29][30][31][32]. In the present study, we evaluated the potential of 18 F-AlF-NOTA-PRGD2, a novel one-step labeled integrin-targeted tracer, for PET/CT imaging to detect tumor angiogenesis in human NPC xenograft model. In addition, we investigated its feasibility to monitor the early response to Endostar antiangiogenic therapy, with the comparison of 18 F-FDG.

Chemistry and radiochemistry
The labeling efficiency for 18 F-AlF-NOTA-PPRGD2 varied from 5% to 25% depending on the reaction volumes. The total synthesis time was about 30 min without HPLC purification. The radiochemical purity was over 97%.

Integrin αvβ3 expression validation
As shown in Figure 1A, the FACS analysis demonstrated that positive rate of CNE-2 cells stained with anti-human integrin αvβ3 monoclonal antibody was only 0.7%, with 84.4% for HUVECs as compared. The Western blot study confirmed the FACS findings ( Figure 1B), indicating that the integrin β3 expression on the CNE-2 cell was negligible. The representative micro PET/CT imaging of CNE-2 bearing mice 1 h after injection of 18 F-AIF-NOTA-PRGD2 was shown in Figure 1C. The tumor could be clearly visualized with excellent contrast to contralateral background. Moreover, high radioactivity accumulation of the tumor was detected.
For the validation of intergrin αvβ3 expression on CNE-2 tumor tissues, we should stained tumor sections with both anti-human and anti-murine integrin αvβ3 antibodies, because the CNE-2 tumor cells are of human origin and the tumor vasculature endothelial cells are of mouse origin [33]. On one hand, we have validated that there is no integrin αvβ3 expression on CNE-2 cell in vitro as above. On the other hand, most of β3 was found to bind to αv (αvβ3) or αIIb (αIIbβ3), but in tumors the latter was hardly expressed. Hence, we only performed immunofluorescence against murine β3 to confirm the integrin αvβ3 expression on the tumor vasculature endothelial cells in CNE-2 tumor tissue. As shown in Figure 1D, the CNE-2 tumor tissues were highly vascularized indicated by strong CD31 staining. Moreover, the co-localization of integrin β3 and CD31 indicated that the integrin β3 expression in the CNE-2 tumor tissues was mainly derived from the tumor blood vessels.

Effects of Endostar on CNE-2 tumor growth
The effective antiangiogenic therapy with Endostar was carried out in CNE-2 tumor-bearing mice. Daily administration of Endostar consecutively for 14 days tended to slow the tumor growth. As displayed in Figure 2, a much less time-dependent increase in tumor growth was observed in the treated group than the control. No early resistance or tumor progression was found in our study. In addition, the tumor volumes in the treated group were significantly reduced compared with those in the control group starting from day 8 (P < 0.05). At the end of the study, the tumor volumes reached 937.97 ± 91.32 mm 3 for the treatment group versus 588.32 ± 70.89 mm 3 for the control group (P < 0.01). No observable body weight loss or any other side effects were observed during the treatment period, indicating that the dosage was safe.

Monitoring the antiangiogenic effect of Endostar by microPET/CT
The static 18    Oncotarget 27246 www.impactjournals.com/oncotarget post-treatment were shown in Figure 3 and Figure 4, respectively. Visually, the radioactivity was obviously lower in the tumors of treatment group than those in the control mice, especially for the group with 18 F-AIF-NOTA-PRGD2. Furthermore, the radioactivity distribution in the treatment group become more and more heterogeneous due to tumor necrosis, while the radioactivity accumulated in the control tumors was relatively homogeneous.

Ex vivo tumor tissue analyses for assessing response to antiangiogenic therapy
The immunofluorescence staining of CD31 and CD61 in ex vivo CNE-2 tumor tissues in the treated group at baseline and 2, 4, 7, and 14 days after Endostar treatment were evaluated. As displayed in Figure 6, the tumor angiogenesis was found significantly inhibited as early as day 2 post-treatment, compared with that of pre-treatment (day 0), and the inhibition lasted to the end of the 2-week treatment. Interestingly, the expression of integrin αvβ3 on the CNE-2 tumor cell was negligible, and almost all the murine β3 was co-localized with the CD31, all these results demonstrated that the integrin αvβ3 was solely expressed on the tumor vasculature in the CNE-2 tumor tissue.
For the macrophage-specific marker F4/80 (Figure 7), a slightly increased macrophage infiltration indicating inflammatory reaction was observed at the early stage of Endostar treatment, whereas it began to decrease from day 7 after treatment, compared with control tumors, contrary to the trend of tumor cell proliferation Ki-67.
Furthermore, the correlations between 18 F-AIF-NOTA-PRGD2 tumor uptake at the end of therapy and the corresponding MVD are shown in Figure 8. A positive and significant linear relationships with r 2 = 0.728 was observed between them. The MVD was significantly lower in the Endostar-treated tumors than that in the control tumors (3.21 ± 0.72 % vs. 6.94 ± 0.68 %; P < 0.001), consistent with the tendency of the corresponding tumor uptake. Oncotarget 27248 www.impactjournals.com/oncotarget DISCUSSION In our present study, we demonstrated that 18 F-AIF-NOTA-PRGD2 enable the visualization and quantification of tumor angiogenesis in the CNE-2 NPC xenograft model, for which intergrin αvβ3 was solely expressed on the tumor vasculature, similar to most kinds of HNSCC cell line and cancer patients [22,34]. Moreover, 18 F-AIF-NOTA-PRGD2 could effectively monitor the early response to the antiangiogenic therapy of Endostar in the CNE-2 NPC xenograft model.
Endostar, as an antiangiogenic agent, on one hand, can cause cytostatic effects rather than cytotoxic effects, that is, antiangiogenic treatment could lead to a delay or stop of tumor progression rather than tumor shrinkage. With a therapy regimen of consecutive daily injection for 14 days in our study, Endostar impeded CNE-2 tumor growth without evident tumor regression and a significant tumor growth inhibition was not observed until day 8. On the other hand, Endostar can inhibit the formation of tumor newborn vascular endothelial cells, resulting in the decrease of MVD, which led to the decrease of integrin expression levels and corresponding targeted imaging probe uptake in tumors. In our study, changes in the tumor uptake of 18 F-AIF-NOTA-PRGD2 were observed as early as day 2 after the initiation of Endostar treatment, before any significant volumetric changes of tumors. Moreover, the reduction in tumor accumulation of 18 F-AIF-NOTA-PRGD2 was much earlier than 18 F-FDG detectable tumor metabolic changes on day 7 post-treatment. 18 F-FDG is the most commonly used PET imaging agent and has been widely applied in the clinical context for tumor diagnosis, cancer staging, and therapeutic effect evaluation. However, 18 F-FDG is not a tumor-specific radiotracer and may shows complex changes when used to monitor antiangiogenic therapy. It is interesting to note  Oncotarget 27249 www.impactjournals.com/oncotarget that, 18 F-FDG uptake was somewhat increased on day 2 and 4 compared with that in the control group, possibly because of the combined effect of slightly increased macrophage infiltration (F4/80 staining) and reduced tumor cell proliferation (Ki-67 index) at the early stage of Endostar-treated tumors. It has been well documented that 18 F-FDG can lead to a false-positive in forms of inflammation, infection, and granulomatous disease [35,36], which may be the limitation of 18 F-FDG in the determination of therapeutic response.
Thus far, an increasing number of RGD-based pepetides carrying a range of radionyclides (e.g., 18 F, 99m Tc, 89 Zr , 68 Ga, 111 In, and 64 Cu) have been developed using various strategies for radiolabeling techniques and peptide modification for integrin αvβ3 expression. With the wellsuited physical half-life of 109.7 min, 18 F is the most popular radioisotope for routine clinical use. Consequently, several 18 F-labeled RGD peptide tracers, including 18 F-galacto-RGD [34], 18 F-AH11585 [21], have been tested in oncologic patients. Futhermore, 18 F-FPPRGD2 [37] has been approved by the Food and Drug Administration (FDA). Many attempts, including the cyclization of the peptides, the synthesis of multimeric RGD peptides and PEGlaytion, have been made to pursue more ideal RGD probes with high tumor targeting capability [38]. However, the monomeric cyclic RGD and time-consuming multiple-step synthetic procedure and relatively low yield had hindered their widespread use [39]. Therefore, a novel one-step labeled tracer 18 F-AlF-NOTA-PRGD2 has been prepared successfully by direct labeling 18 F-aluminum-fluoride complex with a pre-attached chelator on the dimeric RGD peptides with the application of chelation chemistry, without the need of HPLC [40,41].
In our study, we proved that 18 F-AlF-NOTA-PRGD2 enabled visualization of tumor angiogenesis by targeting integrin avβ3 in CNE-2 xenograft model with good imaging quality.
Additionally, we investigated the potential of 18 F-AlF-NOTA-PRGD2 to accurately assess the early antiangiogenic efficacy of Endostar. Our longitudinal imaging results indicated that even though much higher tumor uptake was found in 18 F-FDG PET/CT imaging, therapeutic effect was more clearly and much earlier reflected by 18 F-AlF-NOTA-PRGD2 PET/CT imaging. Furthermore, the tumor uptake showed significantly changes in the Endostar-treated tumors after two-week treatment than those in the control tumors, consistent with the tendency of the corresponding MVD. More importantly, a positive and significant correlation was found between the tumor MVD and the corresponding 18 F-AlF-NOTA-PRGD2 tumor uptake. The co-localization of CD31 and CD61 support these results, which strongly demonstrate that the antiangiogenic effects of Endostar can be monitored by quantitative 18 F-AlF-NOTA-PRGD2 PET/ CT imaging.
Our results are in agreement with those of recently published studies about the use of radiolabeled RGD peptides to monitor the antiangiogenic treatment efficacy in animal models. For example, Battle et al. [26]. demonstrated that the tumor uptake of 18 F-AH111585 decreased significantly from day 2 onward during the 14-days treatment of sunitinib, a small-molecular tyrosine kinase inhibitor (TKI), in integrin αvβ3-positive U87MG tumors models, in line with the corresponding MVD reduction. Similarly, ZD4190, another TKI, appeared to exert a significant decrease of 18 F-AH111585 uptake in Oncotarget 27250 www.impactjournals.com/oncotarget Calu-6 tumors [42] as well as 18 F-FPPRGD2 uptake in MDA-MB-435 breast tumors [25], although the integrin αvβ3 expression level on the latter cell surface was much higher than the former [42]. Collectively, all these results demonstrated that PET/CT with RGD-based probe trace was promising for the assessment of antiangiogenic therapy response.
However, we have to bear in mind that different antiangiogenic agents may have distinct effects on tumor integrin expression, since integrin αvβ3 is not only expressed on the tumor vascular endothelial cells, but also on some kinds of tumor cells [19]. For example, it has been reported that only murine integrin expression on tumor vasculature is down-regulated after treatment with low-dosage Abraxane, but that on tumor cells was almost unaffected [33], whereas the ZD4190 may suppress the integrin expression both on the tumor vasculature endothelial cells and tumor cells by primarily targeting vascular endothelial growth factor receptor (VEGFR) and epidermal growth factor receptor (EGFR) tyrosine kinase activity [25]. In our study, the intergrin αvβ3 was solely expressed on the tumor vasculature in the CNE-2 tumor model, making it possible that 18 F-AlF-NOTA-PRGD2 can directly correlate the tumor uptake of the tumor with corresponding MVD as well as can effectively monitor the early response to antiangiogenic effect of Endostar in the NPC xenograft model. However, we have to bear in mind that tumor uptake of a given radiotracer are not solely dependent on receptor expression. Several factors, such as vascular permeability, vascular density and volume, may affect the specific uptake. Trace kinetic modeling has been validated valuable for separating specific and nonspecific binding and may allow more sensitive and detailed quantification than simple SUV analysis [43,44]. Therefore, further investigations are warranted to eliminate the influence of tumor microenvironment on the pharmacokinetics of 18 F-AlF-NOTA-PRGD2.

CONCLUSIONS
In our present study, we proved that 18 F-AlF-NOTA-PRGD2 PET/CT enabled the clear visualization of tumor angiogenesis and could monitor the therapeutic efficacy of antiangiogenesis in the CNE-2 human NPC xenograft model, which was much earlier and more accurately than 18 F-FDG metabolic imaging and the anatomical structure changes. Therefore, it may be a promising radiotracer used for monitoring the early response from antiangiogenic therapy on tumor and could potentially be translated into clinical practice.

General materials
Unless otherwise specified, all reagents obtained commercially were of analytical grade and used without further purification. No-carrier-added 18 F-fluoride was obtained from an in-house PET trace cyclotron (HM-10, Sumitomo Heavy Industries Ltd, Japan). 18 F-FDG was purchased from the Nuclear Pharmacy of Cardinal Health and reconstituted with sterile saline. The peptides pegylated dimeric RGD peptide PEG 3 -E[c(RGDyK)] 2 (denoted as PRGD2) and NOTA-PRGD2 were synthesized by Chinese Peptide Company (Hangzhou, China). The female BALB/c nude mice (6-8 weeks old, 18-22 g body weight) were purchased from Shanghai Experimental Animal Center (Shanghai, China). The Endostar was provided by Simcere Pharmaceutical Research Co., Ltd.

Synthesis and quality control of 18 F-AlF-NOTA-PRGD2
18 F-AlF-NOTA-PRGD2 was synthesized following previously reported procedure [45]. Briefly, aluminum chloride in 0.2 M sodium acetate buffer at pH = 4 (3 uL, 2 mM) was added into 0.1 mL aqueous [18F] fluoride (0.37 GBq) in a 1.0 mL V-vial. The solution was then heated at 100°C for 10 min to form the aluminum-fluoride complex. After the vial was cooled, NOTA-PRGD2 in 0.2 M sodium acetate buffer at pH = 4 (6 uL, 2 mM) was added, and then the vial was heated at 100°C for another 10 min. At the end of the reaction, the mixture was directly loaded on the C18 cartridge without High Performance Liquid Chromatography (HPLC) purification and then eluted with 0.3 mL of 80 % ethanol/water with 2% acetic acid. The ethanol solution was evaporated with an argon flow, and the final product was then formulated in PBS solution for in vivo study.

Cell culture
Human umbilical vein endothelial cells (HUVECs), obtained from American Type Culture Collection (ATCC, Manassas, VA), were cultured in DMEM culture medium. A poorly differentiated NPC cell lines CNE-2 (FDCC, Fu Dan IBS Cell Center) were grown in RPMI 1640 culture medium. Both culture media were supplemented with 10% fetal bovine serum (Gibco, Paisley, UK) and 1% penicillin / streptomycin at 37°C in a humidified atmosphere containing 5% CO 2 . The expression of integrin αvβ3 on the surface of CNE-2 cells lines was confirmed using flow cytometry and Western blot analysis. HUVECs, supposed to express high amounts of integrin αvβ3, were simultaneously analyzed as the positive control. Antimouse IgG was used as control. This protocol is also applied to HUVECs cells.

Western blotting
CNE-2 and HUVECs cells were trypsinized and lysed in RIPA lysis buffer. Lysates were centrifuged at 12000 g for 10 minutes and protein concentration was determined by using an enhanced BCA protein assay kit (Pierce Biotechnology, Inc., Rockford, IL, USA). Fifty ug proteins were separated by using 8% SDS-PAGE and transferred to the polyvinylidene fluoride (PVDF) membrane. After that, the PVDF membrane was blocked with 5% nonfat milk blocking buffer, and then incubated overnight at 4°C with rabbit antihuman CD61 polyclonal primary antibody (1:200; BD Biosciences, San Jose, USA), followed by incubating with the secondary antibodies: donkey anti-rabbit-IRDye 680 (Li-COR Biosciences). β-actin was used as the loading control, and the bands were detected by using enhanced chemiluminescence (ECL) (Millipore, USA).

Animal models
All the experimental protocols were performed in accordance with guidelines of Institutional Animal Care and Use Committee of Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine. After a 1-week adaptation period, mice were subcutaneously injected with 2 × 10 6 CNE-2 cells in 0.2 mL serum-free media into the right hind flank. Tumors were allowed to grow to 100 mm 3 before treatment (approximately one week after implantation).

Experimental design
The detailed experimental designs are shown in Table 1. Mice with CNE-2 human NPC tumors were allocated to two groups: the imaging group (n = 32) and the immunohistochemical (IHC) staining group (n = 20). Each major group was then divided into control group and Endostar treatment group. For the treated group, Endostar dissolved in 0.9% saline was injected intraperitoneally at the dose of 20 mg/kg/d for consecutive 14 days, while mice in the control group received 0.9% saline at the same dose. Mice in the imaging group underwent 18 F-FDG or 18 F-AIF-NOTA-PRGD2 PET/CT at baseline and after 2, 4, 7, 14 days of treatment with Endostar or PBS, while mice in the histopathologic group were sacrificed on the corresponding imaging time points for immunofluorescence analysis with 3 mice on each time point. To evaluate therapeutic response, the tumor volume was assessed with caliper measurement using the formula: V = 1/2 * ab 2 (a, length; b, width). Body weight was measured every other day.

Small-animal PET/CT imaging
Small-animal PET/CT imaging was performed by using an Inveon system (Siemens Preclinical Solutions, Knoxville, Tennessee, USA). Ten minute static PET scans were acquired at the time point of one hour after injection of 3.7 MBq (100 μCi) 18 F-FDG or 18 F-AIF-NOTA-PRGD2 via tail vein, followed by a 10 min micro-CT scan to obtain anatomic information. For the 18 F-FDG scans, mice were fasted for 6 h before tracer injection. Animals were anesthetized with 1.5% isoflurane in oxygen at 2 L/min and kept warm with a temperature-controlled heating system during the entire imaging procedure. The images were reconstructed by a three-dimensional ordered subsets expectation maximum (3D OSEM) algorithm without attenuation or scattering correction and were then processed by using Siemens Inveon Research Workplace 3.0 (IRW 3.0). The 3D regions of interest (ROIs) were manually drawn over the tumor to obtain the maximum standardized uptake value (SUV max ). Given a tissue density of 1 g/cm 3 , these values were then divided by the injected activity to obtain an image ROI-derived mean and maximum percent injected dose per gram of body weight (%ID/g max ).
All quantitative analyses of the immunofluorescence staining, expressed by the percentage of positive area versus the area of the entire section, were assessed by using Image J software. Among them, the percentage of CD31-positive area was introduced to express the relative microvessel density (MVD) (%) of the tumor.