Clinical Translational Evaluation of Al 18 F-NOTA-FAPI for Fibroblast Activation Protein Targeted Tumour Imaging

Purpose In this study, a novel Al 18 F-NOTA-FAPI probe was developed for fibroblast activation protein (FAP) targeted tumour imaging, which was available to achieve curie level radioactivity by automatic synthesizer. The tumour detection efficacy of Al 18 F-NOTA-FAPI was further validated both in preclinical and clinical translational studies. Methods The radiolabeling procedure of Al 18 F-NOTA-FAPI was optimized. Cell uptake and competitive binding assay were completed with U87MG and A549 cell lines, to evaluate the affinity and specificity of Al 18 F-NOTA-FAPI probe. The biodistribution, pharmacokinetics, radiation dosimetry and tumour imaging efficacy of Al 18 F-NOTA-FAPI probe were researched with healthy Kunming (KM) and/or U87MG model mice. After the approval of ethical committee, Al 18 F-NOTA-FAPI probe was translated into clinical for the PET/CT imaging of first 10 cancer patients. Al 18 F-NOTA-FAPI ± through manually (n with the purity specific activity of 9.3-55.5 MBq/nmol. Whole body effective dose of Al 18 F-NOTA-FAPI was estimated to be 1.24E-02 mSv/MBq, lower than several other FAPI probes ( 68 Ga-FAPI-04, 68 Ga-FAPI-46 and 68 Ga-FAPI-74). In U87MG tumour bearing mice, Al 18 F-NOTA-FAPI showed good tumor detection efficacy from the results of micro PET/CT imaging and biodistribution studies. In organ biodistribution study of human patients, Al 18 F-NOTA-FAPI showed lower SUVmean than 2-[ 18 F]FDG in most organs, especially in liver (1.1 ± 0.2 vs. 2.0 ± 0.9), brain (0.1 ± 0.0 vs. 5.9 ± 1.3), and bone marrow (0.9 ± 0.1 vs. 1.7 ± 0.4). Meanwhile, Al 18 F-NOTA-FAPI do not show extensive bone uptakes, and was able to find out more tumour lesions than 2-[ 18 F]FDG in the PET/CT imaging of several patients. Conclusion Al 18 F-NOTA-FAPI probe was successfully fabricated and applied in fibroblast activation protein targeted tumour PET/CT imaging, which showed excellent imaging quality and tumour detection efficacy in U87MG tumour bearing mice as well as in human cancer patients.


Introduction
Fibroblast activation protein (FAP) is a type II transmembrane glycoprotein consisting of 760 amino acids, which belongs to the serine protease family and is selectively expressed in the stroma fibroblasts associated with epithelial cancers [1]. Reactive tumour stroma or fibrosis generally presents with increased number of activated fibroblasts that usually express FAP, whereas normal stroma in most adult organs only contains a small number of quiescent or resting fibroblasts with low or undetectable FAP expression, making FAP a novel metabolic target in cancer theranostics [2,3]. Till now, plenty of therapies targeting FAP have been explored, including FAP inhibitors [4], peptide drug complexes [5,6], antibodies [7], CAR-T cell therapy [8], vaccines [9] and tumour immunotherapy [10].
Especially in recent several years, (4-Quinolinoyl)glycyl-2-cyanopyrrolidine based organic small molecules that exhibited excellent affinity with FAP [11], known as FAP inhibitor (FAPI), have been radiolabeled with different radionuclides including 68 Ga,90 Y, 99m Tc, 64 Cu and 225 Ac, and have been translated into clinical for the nuclear imaging and radionuclide therapy of various types of cancer [12][13][14][15]. Earlier study indicated that several highly prevalent cancers showed remarkably high uptake and image contrast on 68 Ga-FAPI PET/CT imaging [13]. More importantly, in contrast to 2-[ 18 F]FDG, no diet or fasting is required in preparation for the 68 Ga-FAPI examination, and the image acquisition in 68 Ga-FAPI imaging can be potentially started much earlier than 2- [16].
However, 68 Ga-FAPI PET/CT imaging suffered the disadvantage of radionuclide supply, since 68 Ga was usually eluted from 68 Ge- 68 Ga generator, one single synthesis would only achieve a small amount of radiopharmaceutics avaliable to 2-4 patients.
Meanwhile, the supply of 68 Ga-FAPI to distant centres that requiring this probe would be restricted because of the short half-life of 68 Ga (t1/2 = 68 min). Remolding of 68 Ga-FAPI into 18 F radiolabeled probes would fundamentally resolve its intrinsic disadvantages mentioned above, in consideration of the longer half-life of 18 F (t1/2 = 109.8 min) and the availability of 18 F from cyclotron that owned by many medical centres. Additionally, the lower positron energy of 18  MeV of 68 Ga) rendered it a shorter positron range than 68 Ga (0.6 mm vs. 3.5 mm), which would result in a higher spatial resolution on PET/CT imaging [17]. Most recently, two 18 F radiolabeled FAPI probes has been reported, namely 18 F-FAPI-74 and [ 18 F]FGlc-FAPI, respectively [18,19]. Both of this two probes showed excellent tumour imaging efficacy, however, the structure of this novel Al 18 F-NOTA-FAPI probe was different from them.
Herein, we report the rapid and efficient radiolabeling strategy of a novel Al 18 F-NOTA-FAPI probe and carried out its preclinical investigations. Furthermore, Al 18 F-NOTA-FAPI is translated into clinical application for the PET/CT imaging of patients with different types of cancer, and its imaging efficacy is compared with 2-[ 18 F]FDG.

Chemicals and reagents
Metal basis chemicals used in this study include the following: potassium hydrogen phthalate (KHP) was purchased from Acros Organics (USA), anhydrous aluminum chloride (AlCl3) and sodium acetate were purchased from Alfa Aesar (China) Chemicals Co. Ltd. Reagents including ethanol and acetonitrile were purchased from Honeywell International Inc (USA), trifluoroacetic acid (TFA) was purchased from Shanghai Aladdin Biochemical Technology Co., Ltd (China). All chemicals and reagents were applied directly without further purification. NOTA-FAPI precursor was purchased from HUAYI Co. Ltd (China). Reagents including phosphate buffer saline (PBS), cell culture medium, fetal bovine serum (FBS), penicillin-streptomycin solution (PS), Glutamax, non-essential amino acid (NEAA) and trypsin for cell culturing and following experiments were obtained from Gibco (Thermo Fisher Scientific, China). were purchased from Beijing Vital River Laboratory Animal Technology Co. Ltd (China). Mice of specific pathogen free (SPF) grade were breeding in individual ventilated cages (IVC) with free available of food and water. Tumour on the right flank of mice were allowed to reach an volume of approximately 0.5-0.8 cm in diameter, and animals were further allpied for following imaging and in vivo biodistribution experiments. All animal experiments were completed in accordance with relavant guides and regulations of Beijing Cancer Hospital.

Stability and partition coefficient
The in vitro stability of Al 18 F-NOTA-FAPI was assessed in saline and 5% human serum albumin (HSA) for 1 h, 2 h and 4 h with radio-HPLC. For in vivo stability assessment, KM mice (female, 18-20 g, n = 3) was administered with approximately 37 MBq of Al 18 F-NOTA-FAPI intravenously, urine and blood was collected at 10 min, 0.5 h and 1 h post-injection, radiochemical purity of Al 18 F-NOTA-FAPI in the urine and blood of mice was analyzed with radio-HPLC.
The octanol-water partition coefficient of Al 18 F-NOTA-FAPI was detected by mixting 10 μL of purified Al 18 F-NOTA-FAPI with 590 μL of PBS (0.1 M, pH 7.4) and 600 μL of octanol in a 1.5 mL tube (n = 5). The mixture was vortexed for 3 min and further centrifugated at 3000 × rpm for 5 min, a small portion (10 μL) was removed from each phase and further counted with a γ counter (Wizard II, Perkin Elmer Inc., Germany).
The partition coefficient was calculated as the counts in octanol divided by the counts in phosphate buffered saline, the value was expressed as logD7.4 (mean ± SD).

Radiation dosimetry estimation
The internal radiation dosimetry of Al 18   were drawn on the CT images and further mapped on PET, SUVmax of tumour and muscle were calculated and referred to as tumour to muscle ratio. In the block group, mouse were co-injected with an extra amount (20 nmol) of non-radiolabeled precursor.

Patient enrollment
The clinical translational study of Al 18 F-NOTA-FAPI was approved by the Ethics

Radiopharmaceutical preparation
The synthetic procedure of Al 18 F-NOTA-FAPI was presented in Fig. 1, and the molecular structures of this probe as well as earlier reported [ 18 F]FGlc-FAPI were compared directly. This study evaluated several parameters which might influence the radiolabeling yield, including the amount of AlCl3 (2 mM), KHP (0.5 M) and precursor (4 mM), as well as the temperature and raction volume. As shown in Fig. S1, all parameters exhibited significant impact on radiolabeling yield. Meanwhile, the optimal radiolabeling parameters are as follows, both the volumes of AlCl3 and KHP were 6 μL, the amount of precursor was 5 μL, the volume of 18 F  ions in saline was 100 μL. The reaction mixture was heated at 110 o C for 15 min, and further purified with Sep-Pak Light C18 cartridge. The highest radiolabeling yield under the optimal reaction conditions, appoximately 36.5% can be achieved under manual operation (33.8 ± 3.2% on average, n = 10). The specific activity of Al 18 F-NOTA-FAPI was around 9.3-55.5 MBq/nmol, according to the activity of 18 F  ions added. Radiochemical purity of the final product was higher than 99%. Furthermore, Al 18 F-NOTA-FAPI with much higher activity (> 100 mCi) can be achieved under automatic synthesis with this optimal reaction condition.  (Fig. S3). Moreover, the IC50 of NOTA-FAPI was calculated to be 1.73 ± 0.93 nM (Fig. 2b), which was lower than several other FAPI probes (Table   S1), indicating the high affinity between NOTA-FAPI and FAP.

Small animal biodistribution and pharmacokinetics study
In KM mice, the tracer showed rapid clearence through the kidney, with extremely low uptake in most normal organs including liver, lung, spleen and brain (Fig. 3a). At 1 h post injection, the highest uptake was observed in the gallbladder (17.83 ± 2.92%ID/g), followed by the bone (6.34 ± 0.98%ID/g) and muscle (1.25 ± 0.47%ID/g). The tracer uptake in most organs decrased rapidly along with time, except in the bone and gallbladder. A comparison of the biodistribution at 1 h post injection in KM mice between 68 Ga-FAPI and Al 18 F-NOTA-FAPI was presented in Tabel 1. In all tested organs, the uptake of Al 18 F-NOTA-FAPI was lower than 68 Ga-DOTA-FAPI-04, especially in the spleen, stomach wall and muscle (Fig. S4).
Furthermore, pharmacokinetics study of Al 18 F-NOTA-FAPI in mice provided conclusive evidence for its rapid clearance from body. As shown in Fig. 3b and equation In U87MG tumour bearing xenografts, Al 18 F-NOTA-FAPI showed extremely high accumulation in the tumour region (35.29 ± 1.00 %ID/g ) and low uptake in most organs at 1 h post-injection, as shown in Fig. 4a. The uptake of Al 18 F-NOTA-FAPI in the tumour was slightly higher than 68 Ga-DOTA-FAPI-04 both at 1 h and 4 h post-injection ( Fig. 4b), and the uptake can be almost completely blocked (0.50 ± 0.11 %ID/g, p < 0.0001) by the adding of non-radiolabeled NOTA-FAPI. In A549 xenograft mice, the uptake of Al 18 F-NOTA-FAPI in the tumour tissue was much lower, only reached 3.28 ± 0.66 %ID/g (Fig. S5). Al 18 F-NOTA-FAPI showed relatively high uptake in the bone structure, both in U87MG (5.25 ± 0.68 %ID/g) and A549 (6.77 ± 0.93 %ID/g) tumour bearing mice. However, the uptake in bone can be lowered (3.27 ± 0.28 %ID/g) by coinjection with non-radiolabeled NOTA-FAPI in U87MG tumour bearing mice (p < 0.01, Figure 4a). Alike 68 Ga-DOTA-FAPI-04 (Fig. S6), the retention of Al 18 F-NOTA-FAPI in the tumour was not satisfied, with a decreasement of 33.72% from 1 h to 4 h postinjection.

Radiation dosimetry estimation
The internal radiation dosimetry in adult female patients was estimated based on the biodistribution results in KM mice. As shown in Tabel

Small animal PET/CT imaging
In A549 tumour bearing mice, no significant uptake of Al 18 F-NOTA-FAPI was observed at the tumour region (SUVmax = 0.87), the tracer showed rapid clearence from body, with only high uptake in the gallbladder was noticed. However, in U87MG tumour xenografts, the tumour showed extremely high uptake of Al 18 F-NOTA-FAPI (SUVmax = 3.53), meanwhile, high uptake in the gallbladder was also noticed.
Interestingly, high uptake of Al 18 F-NOTA-FAPI in the tumour region of U87MG xenografts can be almost totally blocked by co-injection of non-radiolabeled FAPI precursor, while most of the injected radiopharmaceutics were excreted through the biliary and intestine tract, indicating the specific uptake of Al 18 F-NOTA-FAPI in the tumour area. (Fig. 5a-c). Futhermore, immunohistochemical staining results indicating high expression of FAP in U87MG xenograft but not in A549 tumour xenograft ( Fig.   5d and 5e).
Interestingly, unlike the high uptake of Al 18 F-NOTA-FAPI in the gallbladder in animal micro-PET/CT imaging, such phenomenon was not validated in some human patients, and the tracer uptake in the bone of human patients was negligible.

Discussion
Fibroblast activation protein targeted tumour imaging strategy developed rapidly in recent several years, especially utilizing FAP specific small molecule inhibitors.
Although plenty of work needs to be further finished, FAPI has been deemed the most competitive radiotracer with 18 F-FDG in clinical tumour PET/CT imaging, because of its fast clearance from body, universality in the diagnosis of various types of cancer, and high tumour to background ratio. As we have successfully fabricated the Al 18 F labeled PSMA ligand for tumour detection [20], it is also convenient to translate the 68 Ga-DOTA-FAPI into Al 18 F-NOTA-FAPI probe. In this study, we investigated both the preclinical and clinical utilization of Al 18 F-NOTA-FAPI probe, which showed great significance in the PET/CT imaging of cancer patients.
Compared with 68 Ga-FAPI, Al 18 F-NOTA-FAPI showed faster clearance from body in mice biodistribution, rendered lower uptake in most normal organs. Meanwhile, the lower effective dose of Al 18 F-NOTA-FAPI (1.24E-02 mSv/MBq) calculated in this study might also ascribe to its faster clearance. The effective dose of Al 18 F-NOTA-FAPI was almost the same with 18 F-FAPI-74 (0.014 mSV/MBq) reported earlier, which contains a NOTA chelator and can also be radiolabeled with 18 F-AlF [18] . Although the structure of FAPI-74 was not public yet, its molecular weight was reported to be 735.8 g/mol, different from the probe (mw. 771.82) utilized in this study.
The high uptake in the gallbladder of mice was a characteristic of Al 18 F-NOTA-FAPI, since such phenomenon was not noticed in 68 Ga-FAPI biodistribution. This can be ascribed to the higher lipophilicity of Al 18 F-NOTA-FAPI, since the LogD7.4 of Al 18 F-NOTA-FAPI was calculated to be -1.88 ± 0.01, bigger than the LogD7.4 of 68 Ga-DOTA-FAPI-04 (-2.87 ± 0.02). High accumulation of radiopharmaceutics in the bone of mice was appreciable, which was validated by small animal PET/CT imaging, especially in the shoulder joints, knee joints, skull and the spine of mice, which might be the explanation of the high absorbed dose and effective dose in osteogenic cells and red marrow, respectively. Such phenomenon was also noticed in previous study with the [ 18 F]FGlc-FAPI probe, which might because of the relatively high expression of FAP in the osteoblast and bone marrow stem cells (BMSC) of mice, hence the high accumulation of FAPI probe in the skeleton of mice was ascribed to physiological uptake [19].
FAP was expressed primarily on the cancer associated fibroblasts but not on tumour cells, only several cancer cell lines showed high expression of FAP, including several glioblastoma cell line [21]. Indeed, much higher tumour cell uptake of Al 18 F-NOTA-FAPI in U87MG cell line was noticed, compared with FAP low expressed A549 cell line in vitro. Since the uptake of Al 18 F-NOTA-FAPI in U87MG cells can be blocked by extra adding of non-radiolabeled precursor, the specific binding between the radiopharmaceutical and FAP can be thus confirmed. Interestingly, the highest uptake of Al 18 F-NOTA-FAPI in U87MG cells was observed at 10 min, and decreased in the following 2 hours, which was not reported by previous studies. We ascribe this to the rapid binding of Al 18 F-NOTA-FAPI with FAP and internalization in U87MG cells, since the IC50 of NOTA-FAPI was lower than any previously reported FAPI probes (Table S2) [12,22].
Also, from the biodistribution analysis results of tumour bearing mice, tracer uptake in U87MG xenografts was approximate 10 folds higher than that in A549 xenografts, while no significant differences were noticed in most normal organs. Meanwhile, in small animal PET/CT imaging, extremely high uptake of Al 18 F-NOTA-FAPI in the tumour region of U87MG xenograft can be observed, the tumour-to-muscle SUVmax ratio was 23.5, much higher than A549 mice xenograft (tumour/muscle SUVmax ratio = 6.2). While in the mice co-injected with non-radiolabeled NOTA-FAPI, most of the radioactivity was accumulated in the gallbladder and intestine, further indicated the specific binding of Al 18 F-NOTA-FAPI to FAP. Moreover, in this study, the immunohistological chemistry staining of both A549 and U87MG tumours rescted from mice also demonstrated the high expression of FAP in U87MG xenografts.
In the PET/CT imaging of human cancer patients, there existed several differences compared with micro PET/CT imaging with mice. Firstly, the high uptake in the gallbladder of mice was not appreciable in several human patients, which might because the fasting of mice will promote the secretion of bile, while eating in some human patients might accelerate the removal of gallbladder contents, hence the difference of SUVmean in human gallbladder was not significant between Al 18 F-NOTA-FAPI and 2-  [23]. However, when comparing the SUVmean values in primary tumours and normal organs, Al 18 F-NOTA-FAPI showed higher tumor-to-background ratios than 2-[ 18 F]FDG thanks to its low uptake in normal organs, especially in the spleen and brain. Therefore, Al 18 F-NOTA-FAPI was a promising candidate and alternative to 68

Conclusion
In the present study, we developed the Al 18 F-NOTA-FAPI probe for fibroblast activation protein targted imaging. Al 18 F-NOTA-FAPI can be achieved under convenient manual operation with high radiolabeling yield and specific activity. NOTA-FAPI in this study showed higher affinity with FAP compared with several other FAPI probes. Al 18 F-NOTA-FAPI showed excellent imaging efficacy in U87MG tumour bearing mice and was successfully translated into the clinical PET/CT imaging of cancer patients. Furthermore, Al 18 F-NOTA-FAPI can be achieved by batch of production with high radioactivity using automatic synthesis, render it a promising candidate and alternative to 68 Ga-DOTA-FAPI-04.

Compliance with ethical standards
Conflicts of interest: The authors declare that they have no conflict of interest.