EDB-FN targeted probes for the surgical navigation, radionuclide imaging, and therapy of thyroid cancer

Extradomain B of fibronectin (EDB-FN) is a promising diagnostic and therapeutic biomarker for thyroid cancer (TC). Here, we identified a high-affinity EDB-FN targeted peptide named EDBp (AVRTSAD) and developed three EDBp-based probes, Cy5-PEG4-EDBp(Cy5-EDBp), [18F]-NOTA-PEG4-EDBp([18F]-EDBp), and [177Lu]-DOTA-PEG4-EDBp ([177Lu]-EDBp), for the surgical navigation, radionuclide imaging, and therapy of TC. Based on the previously identified EDB-FN targeted peptide ZD2, the optimized EDB-FN targeted peptide EDBp was identified by using the alanine scan strategy. Three EDBp-based probes, Cy5-EDBp, [18F]-EDBp, and [177Lu]-EDBp, were developed for fluorescence imaging, positron emission tomography (PET) imaging, and radiotherapy in TC tumor-bearing mice, respectively. Additionally, [18F]-EDBp was evaluated in two TC patients. The binding affinity of EDBp to the EDB fragment protein (Kd = 14.4 ± 1.4 nM, n = 3) was approximately 336-fold greater than that of the ZD2 (Kd = 4839.7 ± 361.7 nM, n = 3). Fluorescence imaging with Cy5-EDBp facilitated the complete removal of TC tumors. [18F]-EDBp PET imaging clearly delineated TC tumors, with high tumor uptake (16.43 ± 1.008%ID/g, n = 6, at 1-h postinjection). Radiotherapy with [177Lu]-EDBp inhibited tumor growth and prolonged survival in TC tumor-bearing mice (survival time of different treatment groups: saline vs. EDBp vs. ABRAXANE vs. [177Lu]-EDBp = 8.00 d vs. 8.00 d vs. 11.67 d vs. 22.33 d, ***p < 0.001). Importantly, the first-in-human evaluation of [18F]-EDBp demonstrated that it had specific targeting properties (SUVmax value of 3.6) and safety. Cy5-EDBp, [18F]-EDBp, and [177Lu]-EDBp are promising candidates for the surgical navigation, radionuclide imaging, and radionuclide therapy of TC, respectively.


Introduction
Thyroid cancer (TC) is the most common endocrine carcinoma, accounting for 1-2% of human tumors [1]. The incidence of TC is increasing year by year, and the mortality rate of advanced TC is also increasing [2,3]. The common pathological types of TC mainly include papillary thyroid cancer (PTC), follicular thyroid cancer (FTC), medullary thyroid cancer (MTC), and anaplastic thyroid carcinoma (ATC). Except for ATC, TC is mainly treated with surgery [1,4,5]. And up to 30% of patients with DTC experience disease recurrence, and 30% of the cancers eventually become refractory to radioactive iodine therapy [1,6,7]. Patients with radioactive iodine-refractory DTC (RAIR-DTC) have a poor prognosis, with a 10-year survival rate of 10%. The 5-year survival rates of MTC and ATC respectively only were 68.75% and 16.81% [8,9]. Patients with radioactive iodine-refractory differentiated thyroid cancer (RAIR-DTC), anaplastic thyroid carcinoma (ATC), and progressive medullary thyroid cancer (MTC) have no chance of cure; conventional therapy plays only a palliative role. More novel effective therapeutic approaches are still urgently needed for these hard-to-treated TC.
Fibronectin (FN) is an extracellular matrix protein with pivotal physiological effects that are due to the alternative mRNA splicing of its EDA, EDB, and IIICS regions due to posttranslational modifications [10]. Generally, the extradomain B of fibronectin (EDB-FN) is expressed in fetal tissues and in cancerous tissues of adults, but not in many noncancerous tissues. Due to this feature, EDB-FN has been extensively used for the targeted delivery of radioisotopes to exert therapeutic effects on primary cancers and metastatic lesions [11][12][13][14][15][16][17]. Recently, Phei Er Saw et al. reported that there are fifteen kinds of cancers with EDB-FN overexpression including TC. The cancer-to-normal ratio of EDB-FN expression in TC was the fourth highest among these cancers [12]. These results suggest that EDB-FN may be a promising biomarker for the diagnostic and therapeutics of TC.
An EDB-FN targeted peptide named ZD2 (CTVRT-SADC) has been identified by Professor Zheng-Rong Lu's team using phage display technology. Based on ZD2 peptide, they developed EDB-FN targeted fluorescent and magnetic resonance imaging (MRI) probes for prostate cancer and pancreatic cancer imaging [18]. Although the current EDB-FN targeted peptide ZD2 has been successfully employed for molecular imaging of several kinds of cancers in mouse models, the binding affinity of ZD2 to the EDB fragment protein (micromolar) is still needs to be improved. In this study, we identified an optimized EDB-FN targeted peptide named EDBp (AVRTSAD, with nanomolar affinity) by using an alanine scan strategy based on ZD2 (with micromolar affinity) and developed three EDBp-based probes, Cy5-PEG4-EDBp, [ 18 F]-NOTA-PEG4-EDBp, and [ 177 Lu]-DOTA-PEG4-EDBp, for fluorescence imaging, positron emission tomography (PET) imaging, and radionuclide therapy in TC tumor-bearing mice, respectively. Additionally, we carried out the first-in-human evaluation of [ 18 F]-NOTA-PEG4-EDBp in two TC patients.

Cell lines
TC cells were obtained from the cell bank of the State Key Laboratory of Oncology in South China. Specific culture methods and related experiments are provided in the supplementary materials.

Mice
A total of 1 × 10 6 BHT-101 cells were injected into the cervical thyroid position of each BALB/c nude mouse to construct TC tumor-bearing mouse models. The mice were divided into a surgical excision group (n = 7), imaging group (n = 16), and therapy group (n = 24). When the tumors reached a diameter of 0.5-1.0 cm, in vivo imaging experiments and treatment were performed. All animal experiments were approved by the Institutional Animal Care and Use Committee at Sun Yat-sen Cancer Center (Ethical code: L102012021080H).

Acquired of EDBp
All non-alanine residues in EDB-FN were mutated using the alanine scan strategy. We obtained a new set of peptides by the alanine scan mutation technique (in which alanine is substituted by another amino acid in the peptide). Micro-Scale Thermophoresis (MST) was used to evaluate the affinity between alanine scanned peptides and EDB-FN proteins, and the peptide with strongest affinity was selected, namely EDBp. Details are shown in the supplementary materials.

Cy5-PEG4-EDBp fluorescent imaging and directed resection of mass
Cy5-PEG4-EDBp was synthesized by the Chinese Peptide Company (lot no.: CU-09-00,019, Hangzhou, China). Mice in the blocking group (2000 µM EDBp was injected via the tail vein and circle for 1 h, n = 4) and the unblocking group (n = 3) were injected with 2 µM Cy5 at the same time, and NIRF imaging was performed after a 1-h cycle. The fluorescence intensity of the orthotopic tumors was monitored by an IVIS 200 imaging system. Then, the tumors of the mice were removed as cleanly as possible and imaged to observe whether there was any tumor tissue remaining after removal. After the mice were sacrificed, the main organs, such as the kidney and heart, were collected for imaging, and the corresponding ROIs were recorded. The imaging data were analyzed using Living Image 4.7.3 (IVIS 200 Imaging Systems) software.

Biodistribution
ATC tumor-bearing mice (n = 6) were blocked with 3000 µM EDBp for 1 h. Mice in the unblocking group (n = 10) and blocking group (n = 6) were injected with 3.7 MBq (100 μCi)/150 μl of [ 18 F]-NOTA-PEG4-EDBp through the tail vein. One hour later, both the unblocking group (n = 6) and blocking group (n = 6) were immediately sacrificed and dissected for the biological distribution study. Two hours after the treatment, the remaining mice from the unblocking group (n = 4) were sacrificed and examined for biological distribution. Organs and tissues, such as the liver, heart, lung, and bone, were measured by a γ counter (Automatic Gamma Counter, 2480 WIZ-ARD2), and the distribution of [ 18 F]-NOTA-PEG4-EDBp in different organs was analyzed.

PET/CT imaging
Mice were injected with [ 18 F]-NOTA-PEG4-EDBp (3.7 MBq/150 μl) through the tail vein. PET/CT scans were performed after 1 h of drug circulation. The mice were divided into an unblocking group and a blocking group. In the blocking group, each mouse was injected with 4 mg/l EDBp in the tail vein before [ 18 F]-NOTA-PEG4-EDBp was injected. PET scanning parameter setting is as follows: 600-s one-bed acquisition, MLEM algorithm, iteration number 12, scattering, decay, random correction, and no attenuation correction. CT parameter settings are as follows: 0.4 mA and 45 kV. Reconstruction algorithm FBP (small animal PET/CT, Albira II, Bruker). The survival status of mice from treatment to death or termination of the experiment was recorded, and tumor tissues were collected for IHC analysis using Ki-67 antibodies.

Ex vivo autoradiography
Paraffin sections of TC tissues of different pathological types were obtained from the Pathology Department of SYSUCC. The dewaxed and hydrated sections were incubated in 20 μCi/mL [ 18 F]-NOTA-PEG4-EDBp Tris-HCl buffer (pH = 8.2) at room temperature for 30 min. Then, the sections were washed with ultrapure water 6 times and Tris-HCl buffer 2 times. The slices were dried in an oven at 60 °C for 5 min until they became white and posted on an imaging film for 2 h. The imaging film was scanned with a phosphor screen imaging system (Typhoon FLA7000IP, GE) at a pixel size of 100 μm, and the PMT was 1000. The data were analyzed using ImageQuant TL8.1 software.

PET/CT imaging of patients with [ 18 F]-NOTA-PEG4-EDBp
The study protocol was approved by the Ethics Committee of the Affiliated Hospital of Jiangnan University and conducted in accordance with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards (ethical code: LS2023011). Each patient gave written and informed consent before the PET/CT study. No patient preparation was required for the [ 18 F]-NOTA-PEG4-EDBp PET scan. At 5 min and 45 min after intravenous injection (dose of 4.07 ± 0.1 MBq/kg, a total of 296.0 MBq) of patient no. 1 and at 1 h after intravenous injection (dose of 3.7 ± 0.1 MBq/kg, a total of 281.2 MBq) of patient no. 2, the two patients were scanned on a Biograph 64 PET/CT scanner (Siemens). Whole-body (vertex to thigh) PET/CT images were obtained in 3D mode (2 min per bed position). A continuous low-dose CT scan for attenuation correction was acquired in spiral mode with the settings 120 kV, 170 mAs, slice thickness 2 mm, and pitch 0.8. For image analysis, ROIs were measured with a multimodal workstation (Syngo.via; Siemens Medical Solutions).

Statistical analysis
Statistical analysis and mapping were performed using SPSS 22.0 and GraphPad Prism 8. Data are presented as mean ± SD. Data were compared using two-way analysis of variance (ANOVA) with Bonferroni correction and Student's t-test. Overall survival was determined by the Kaplan-Meier method, and the p-value was calculated with the Wilcoxon test. The significance is given as *p < 0.05, **p < 0.01, and ***p < 0.001.

High positive rate of EDB-FN overexpression in TC
The expression of EDB-FN in different pathological types of TC tumor tissues were analyzed by immunohistochemistry. Two experienced pathologists evaluated the results according to this classification. EDB-FN expression was 87.1% (27/31) Table. S1). There is no significant correlations between the EDB-FN expression and clinical features including gender, age, TNM stage, and prognosis (Table S2). The lower the expression of EDB-FN, the better the prognosis of ATC, and there is no significant correlation with the prognosis of other types of TC (Fig. S2). The extreme high positive rate of EDB-FN overexpression makes it a potential diagnostic and therapeutic biomarker for TC.
Based on the EDBp peptide, three probes were synthesized for diagnosis and treatment. According to previous studies, DOTA and NOTA-modified peptide can be easily radiolabeled with radionuclide to form molecular probes; meanwhile, the introduction of glyceryl stearate (PEG4) can increase the water solubility and better interact with the target. EDBp was connected to NIRF dye Cy5, chelating agents NOTA, and DOTA by PEG4-Lys as linker to form the target molecular probe Cy5-PEG4-EDBp, NOTA-PGE4-EDBp, and DOTA-PEG4-EDBp ( Fig. 2A, Fig. S3). Two radiolabeled ligands were generated in high purity and specific activity:  Fig. 2C). [ 18 F]-EDBp incubation with FBS at 37 °C for 180 min was still stable, and the purity is > 99% (Fig. S7), and the synthetic yield is about 60-70%. [ 177 Lu]-EDBp incubation with FBS at 37 °C for 7 days was still stable, and the purity is > 99% (Fig. S7), and the synthetic yield is about 90-95%.

NIRF imaging with Cy5-EDBp for surgical resection of TC
Cy5-EDBp bound specifically to TC cell lines of different pathological types (Fig.s S4, S5A). Whole-body nearinfrared fluorescence (NIRF) imaging showed that under the same imaging parameters, fluorescence aggregation was observed in the neck tumor of the unblocking group. The blocking group, without any tumor tissue accumulation, was used as a negative control (Fig. 3A). NIRF imaging of the collected organs and tumors further demonstrated 1 3 Fig. 1 The binding affinities of seven alanine mutated peptides and the ZD2 peptides to the EDB fragment protein. Seven alanine mutated peptides (A1-A7) and the ZD2 peptides were synthesized, and their binding affinities to the EDB fragment protein were measured by using the microscale thermophoresis (MST) technology specific accumulation of the Cy5-EDBp peptide in tumors but not in other organs except the kidney (Fig. S5C). There was severe adhesion between the neck tumor and the surrounding muscle and vessel tissue (Fig. S5B), and no residue was found on the cutting edge by fluorescence imaging after blunt tumor separation (Fig. 3B). NIRF imaging of the tumor tissues showed that the Cy5-EDBp peptide but not the blocking group specifically accumulated in tumors (Fig. 3C). The radiation efficiency of tumor tissue in the unblocking group was 44.95 ± 20.89 (n = 4) and in the blocking group was 9.07 ± 3.92 (n = 3, p = 0.035). The radiation efficiency of other organs except the kidney was lower than that of  (Fig. 3D). These findings indicate that the EDBp peptide specifically homes to EDB-FN-positive TC and fluorescent imaging with Cy5-EDBp can facilitate the complete removal of tumor in TC-bearing mice.  (Fig. S6). Biodistribution in normal mice of the liver, spleen, and kidney had more concentration than other organs, which may be related to blood flow and metabolic pathways (Fig. S8A). Biodistribution in TC tumor-bearing mice was significantly higher than that of other organs except for the kidney in 1 h and 2 h (Fig.  S8B). The tumor-to-organ ratios at 1 h and 2 h indicated that 1 h had obvious advantages, so 1 h was selected for subsequent imaging experiments (Fig. S8C) (Fig. 4A). The maximum intensity projection (MIP) diagram showed that the neck nuclide enrichment in the unblocking group was not observed in other organs (except the liver, kidney, and bladder) and the blocking group (Fig. 4A). The distribution of radionuclides in tumor tissues of the unblocking group was significantly higher than that in the blocking group (Fig. 4B). The tumor tissue uptake peaks were 16.43 ± 1.00%ID/g (n = 6, p < 0.001) in the unblocking group and 2.07 ± 0.98%ID/g (n = 6, p = 0.0035) in the blocking group at 1 h and 6.52 ± 0.54%ID/g (n = 4, p < 0.001) in the unblocking group at 2 h. The unblocking group was eight times higher than that of the blocking group (p < 0.001), and EDBp blocked the binding of [ 18 F]-EDBp to EDB-FN-related targets in tumor tissue (Fig. 4B)

Treatment with [ 177 Lu]-EDBp in TC tumor-bearing mouse models
Albumin paclitaxel (ABRAXANE) is the best choice of single-agent chemotherapy for ATC because of its good efficacy and small side effects [19]. Therefore, the efficacy of ABRAXANE treatment group and [ 177 Lu]-EDBp treatment group on ATC tumor-bearing mice was compared. To avoid the influence of the characteristics of peptides themselves, we observed the efficacy of EDBp as a separate treatment group, at the same time, the saline group as the blank control group. The experiment was terminated when the mice died or the tumor diameter reached 1.5 cm. The growth rate of tumor volume in the [ 177 Lu]-EDBp group was the slowest, and in the albumin paclitaxel (ABRAXANE) group, it was slower than that in the saline and EDBp groups. There was no significant difference in the tumor growth rate between the saline and EDBp treatment groups (Fig. 5A, Fig. S12 A-E  (Fig. 5B). IHC of Ki67 was performed in tumor tissues of different treatment groups (Fig. 5C) (Fig. 5D). The main organs of mice in the [ 177 Lu]-EDBp treatment group were embedded in paraffin for HE staining, no apparent abnormal damage was found (Fig. S13), and the [ 177 Lu]-EDBp therapy was well tolerated.   SUVmax value of 3.6 ( Fig. 6 B and C). There were no adverse or clinically detectable pharmacologic effects in the patients. These first-in-human results demonstrate that [ 18 F]-EDBp has specific targeting properties and safety. In addition to the high uptake of [ 18 F]-EDBp in the urinary system, there was no obvious specific uptake in other organs and tissues; the uptake of tumor lesions is superior to that of normal organs or tissues (  (Fig. 7).

Discussion
The application of EDB-FN targeted probes in other tumors laid the theoretical foundation for our study. Previous studies of EDB-FN targeted probes have been based on antibodies and peptides. Antibody-based approaches can be specific and sensitive. Unfortunately, their large molecular weights result in slow delivery and diffusion into tumor tissues. Compared to antibodies or proteins, low-molecular-weight peptides have a higher cell/tissue permeation ability. Their pharmacokinetics can be improved by chemical modification, whereas their targeting ability is rarely affected. In this study, we identified an optimized EDB-FN-targeting peptide named EDBp (AVRTSAD, with nanomolar affinity) by using an alanine scan strategy based on ZD2 (CTVRTSADC, with micromolar affinity). MST was used to detect the binding force between alanine-scanned peptides and the EDB-FN protein.  The sketch map of the study may provide more detail information about the high binding affinity of EDBp peptide to the EDB fragment protein.
The current surgical methods of TC usually rarely lead to the complete removal of all tumor tissue, and they might cause parathyroid injury [20,21]. Although MTC and ATC resection can improve the prognosis, some patients miss the opportunity for surgery because complete surgical resection is not achieved [22]. The value of fluorescence imaging in tumor imaging has been widely reported [23,24]. Zhenhua Hu team reported PET/NIR-II fluorescence image-guided surgery of glioblastoma using a folate receptor α-targeted dual-modal nanoprobe and first-in-human liver-tumor surgery guided by multispectral fluorescence imaging in the visible and near-infrared-I/II windows [25,26] [28]. On the basis of these studies, we used the fluorescent tracer Cy5-EDBp to detect TC tumor-bearing mice and guided the complete resection of tumors by NIRF imaging. The radiation efficiency of tumor tissue in the unblocking group was 5 times that in the blocking group. Cy5-EDBp was used for surgical navigation, which was beneficial for completely removing the thyroid gland and protecting the parathyroid gland during DTC operations. This technique may also help change the status quo of inoperable tumors of MTC and ATC that cannot be completely removed. We demonstrated the feasibility of the EDB-FN targeted PET tracer [ 18 F]-EDBp for detection in TC cells (Fig. S5) and ATC tumor-bearing mice with high sensitivity and specificity. We herein demonstrated the feasibility of the EDB-FN-targeted PET tracer [ 18 F]-EDBp for detection in TC tumor-bearing mice with high sensitivity and specificity. Biodistribution in the tumor tissues of TC tumor-bearing mice was significantly higher than that of other organs except for the kidney at 1 h and 2 h (Fig. S7). Drug toxicity studies showed that EDBp and [ 18 F]-EDBp exerted no obvious bone marrow suppression and no obvious damage to the major organs of normal mice (Figs. S8/9).
Peptide receptor radionuclide therapy (PRRT) has emerged as one of the most pivotal modalities of treatment for neuroendocrine tumors (NETs), and one of the actively investigated topics is the utility of PRRT in the treatment of TC [29]. On the basis of positive imaging, the target peptide EDBp was used for radionuclide therapy. Using a β-emitter, radionuclide such as [ 177 Lu] (path length of ~ 1.5 mm) should enable the killing of the tumor stromal cells and the surrounding tumor cells [30,31]. [ 177 Lu]-based probes targeting to different tumor-associated target markers have been reported for the treatment of malignant tumors [32,33]. Here, we radiolabeled the peptide EDBp with the therapeutic nuclide [ 177 Lu]. We observed that the tumor volume in the [ 177 Lu]-EDBp treatment group was reduced, and the prognosis was improved more so than that in the ABRAXANE group. There was no significant difference in the patients' mental state between the [ 177 Lu]-EDBp treatment group and other treatment groups, and no apparent major organ damage of [ 177 Lu]-EDBp was observed (Fig. S13). We observed that EDB-FN was overexpressed in different pathological types of TC, so [ 177 Lu]-EDBp is applicable to different pathological types of TC and lays a foundation for the treatment of RAIR-DTC.
We found that EDB-FN was overexpressed in different pathological types of TC (Table S1). EDB-FN expression was less affected by various conditions, such as sex, age, and stage, thus reducing the rate of missed diagnosis in clinical application (Table. S2). The overexpression of EDB-FN in all TCs of different pathological types laid a theoretical foundation for the application of FN in all pathological types. We demonstrated that the probe [ 18 F]-EDBp is sensitive to human TC tumor tissues by autoradiography. PET imaging of [ 18 F]-EDBp in two TC patients showed specificity and sensitivity, and after [ 18 F]-EDBp injection, the vital signs of the patients were stable, and no distinct discomfort was observed. Patient no. 2 underwent surgery, chemotherapy, radiotherapy, and multiple nuclide treatments, which clinically indicated that the focal glucose metabolic rate was reduced, and the sensitivity of [ 18 F]-FDG imaging was poor. [ 18 F]-EDBp PET imaging could detect multiple bone metastases, and some lesions had not obvious bone destruction. The PET imaging value of [ 18 F]-EDBp in TC is worthy of affirmation and expectation. In addition to the high uptake of [ 18 F]-EDBp in the urinary system, there was no obvious specific uptake in other organs and tissues, the uptake of tumor lesions is superior to that of normal organs or tissues (Table. S3), which provides a favorable basis for the application of EDBp in radionuclide therapy. EDB-FN has been reported to be overexpressed in many malignancies. Therefore, in addition to TC, EDBp-based probes have potential applications for these cancers.
This study has several limitations. First, the rapid blood clearance of [ 177 Lu]-EDBp led to a shorter time of tumor concentration, which may reduce therapeutic efficacy. This may be due to the fast blood clearance of EDBp. At the same time, it can reduce the residence time of nuclides in normal organs more and mitigate the damage to normal organs. Second, to follow the 3Rs (Reduction, Replacement and Refinement) principle for laboratory animals, only a limited number of mice have been used to preliminarily evaluate the treatment of [ 177 Lu]-EDBp. Third, only two TC patients were included for the evaluation of [ 18 F]-EDBp. To further evaluate its specific targeting properties and safety, more patients should be included in future studies.

Conclusion
In conclusion, EDBp is an optimized EDB-FN targeted peptide with nanomolar affinity. Cy5-EDBp facilitates surgical navigation of TC. [ 18 F]-EDBp is sensitive and safe for the detection of TC. [ 177 Lu]-EDBp is effective for the treatment of ATC.