Molecular Imaging-Derived Biomarker of Cardiac Nerve Integrity - Introducing High NET Affinity PET Probe 18F-AF78.

Background: Radiolabeled agents that are substrates for the norepinephrine transporter (NET) can be used to quantify cardiac sympathetic nervous conditions and have been demonstrated to identify high-risk congestive heart failure (HF) patients prone to arrhythmic events. We aimed to fully characterize the kinetic profile of the novel 18F-labeled NET probe AF78 for PET imaging of the cardiac sympathetic nervous system (SNS) among various species.
Methods: 18F-AF78 was compared to norepinephrine (NE) and established SNS radiotracers by employing in vitro cell assays, followed by an in vivo PET imaging approach with healthy rats, rabbits and nonhuman primates (NHPs). Additionally, chase protocols were performed in NHPs with NET inhibitor desipramine (DMI) and the NE releasing stimulator tyramine (TYR) to investigate retention kinetics in cardiac SNS.
Results: Relative to other SNS radiotracers, 18F-AF78 showed higher transport affinity via NET in a cell-based competitive uptake assay (IC50 0.42 ± 0.14 µM), almost identical to that of NE (IC50, 0.50 ± 0.16 µM, n.s.). In rabbits and NHPs, initial cardiac uptake was significantly reduced by NET inhibition. Furthermore, cardiac tracer retention was not affected by a DMI chase protocol but was markedly reduced by intermittent TYR chase, thereby suggesting that 18F-AF78 is stored and can be released via the synaptic vesicular turnover process. Computational modeling hypothesized the formation of a T-shaped π-π stacking at the binding site, suggesting a rationale for the high affinity of 18F-AF78.
Conclusion: 18F-AF78 demonstrated high in vitro NET affinity and advantageous in vivo radiotracer kinetics across various species, indicating that 18F-AF78 is an SNS imaging agent with strong potential to guide specific interventions in cardiovascular medicine. © The author(s).

Introduction

Various clinical conditions, such as heart failure (HF) 1, 2, diabetes-associated cardiac autonomic neuropathy (CAN) 3, dilated or hypertrophic cardiomyopathy 4, 5, aortic stenosis 6 or infarcted myocardium, are associated with alterations of the sympathetic nervous system (SNS) 7. Serving as the backbone of neurohumoral radionuclide imaging, cardiac SNS radiotracers using either single-photon emission computed tomography (SPECT) or positron emission tomography (PET) are mostly structurally related to physiological norepinephrine (NE) 8, 9. Neurohumoral cardiac radiotracers share similar pathways with NE and are transported into neuronal cells via the NE transporter (NET), and some of these radiotracers are subsequently stored in presynaptic vesicles.

Such radiotracers have been suggested to provide a noninvasive read-out indicating the level of denervated myocardium 10. The radiotracer 123I-meta-iodobenzylguanidine (123I-MIBG) was the first and only approved molecular imaging agent by the U.S. Food and Drug Administration for the scintigraphic assessment of myocardial sympathetic innervation. In contrast to such a global myocardial assessment, the increased spatiotemporal resolution of PET technology allows for a more precise read-out of regional neuronal denervation, e.g., the amount of denervated myocardium in the border zone after myocardial infarction relative to remote myocardium 11.

Substantial evidence has been obtained on neurohumoral image-guided strategies in various clinical scenarios, including implantable cardiac device (ICD) insertion or prognostication 8, 12, 13. For instance, the “AdreView Myocardial Imaging for Risk Evaluation in Heart Failure (ADMIRE-HF)” trial demonstrated the superior prognostic value of 123I-MIBG scintigraphy for major cardiovascular events in ischemic and nonischemic HF patients compared to that of other established parameters 14. Based on these encouraging results, the PAREPET trial linked denervated myocardium quantified by 11C-meta-hydroxyephedrine (11C-HED) PET to an increased risk for sudden cardiac death in ischemic cardiomyopathy patients, independent of left ventricular ejection fraction and infarct volume 15. The use of 18F-labeled SNS PET radiotracers has also increased recently; these tracers have the advantage of a significantly longer half-life (110 min vs. 20 min for carbon-11), allowing the dispatch from central cyclotron facilities for improved cost-effectiveness or implementation of additional delayed image protocols. Thirty years ago, 18F-6-fluorodopamine (18F-6F-DA) was introduced as the first 18F-labeled NET radiotracer 16, demonstrating its usefulness in evaluating various scenarios, such as neuroendocrine tumors 17, cardiac SNS changes 18 and sympathetic denervation in Parkinson's disease (PD) 19. Recently, the novel 18F-labeled radiotracer AF78 was introduced as a NET-targeting probe with a phenethylguanidine core structure. AF78 is characterized by an easy radiolabeling procedure and high radiolabeling yield, allowing for high throughput even in a busy PET practice 20. However, prior to routine human use, a precise understanding of the catecholamine radiotracer handling of 18F-AF78 at the nerve terminal is indispensable. Such accurate characterization of 18F-AF78 at the nerve terminal may pave the way for image-guided molecular strategies in various clinical scenarios, including prediction of arrhythmias, guiding resynchronization therapies, or monitoring complex cardiac interventions, such as valve replacement. In the present study, we aimed to fully decipher the kinetic profile of 18F-AF78 of the cardiac SNS by employing in vitro cell assays and high-resolution PET imaging among various species.

Methods

Experimental protocols were approved by the Animal Ethics Committee of the National Cerebral and Cardiovascular Center, Research Institute, Osaka, Japan (Approval number 18019) and conducted in strict accordance with the Guide for the Care and Use of Laboratory Animals, published by the U.S. National Institute of Health 21, and the ARRIVE guidelines.

Radiolabeling

18F-AF78 was synthesized as described previously 20. In short, a solution of the precursor (4.8 mg, 0.066 mmol) and Kryptofix222 (22.5 mg, 0.060 mmol) in dry acetonitrile (3 mL) was added to the isolated and dried 18F-KF. The mixture was heated at 110 °C for 10 min, followed by the addition of hydrochloric acid (6 N, 0.3 mL) and further heating at 120 °C for 20 min (Figure ). After quenching the reaction, the mixture was purified via semipreparative high-performance liquid chromatography. The collected fraction was diluted with water (10 mL) and trapped on a preconditioned Sep-Pak C18 cartridge, which was then washed with water (5 mL). 18F-AF78 was eluted with ethanol (1 mL). After removing the solvent with nitrogen flow, the target tracer was diluted with saline to the required concentration for further application. The radiochemical purity of 18F-AF78 was consistently > 97%, with molar activity > 56 GBq/µmol.

Cell culture and competitive uptake assay

SK-N-SH (human neuroblastoma cell line) cells naturally expressing NET were cultivated according to the instructions from the supplier (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany). The cells were seeded one day before the competitive assay into a 24-well plate at 1 × 105 cells/well and incubated overnight at 37 °C. In the presence of 100 µM pargyline (monoamine oxidase inhibitor) and 20 µM pyrogallol (catechol-O-methyltransferase inhibitor), 3H-NE (PerkinElmer, Germany) in 0.1% BSA/DMEM (12 kBq/mL) was added to each well together with various concentrations of either NET selective inhibitor desipramine (DMI) or testing compounds, including NE as reference, ranging from a final concentration of 10-10 to 10-3 M. The plate was incubated at 37 °C for 60 min. The supernatant was removed, and the cells were washed with ice-cold phosphate-buffered saline buffer (2 × 1 ml) followed by the addition of NaOH solution (0.1 N, 500 μl). Cell lysates were collected and measured in a liquid scintillation analyzer 10. Nonspecific uptake was determined by incubating at 4 °C, which was negligible in all cases. The cold references of 1-(3-bromo-4-(18F-3-fluoropropoxy)benzyl)guanidine/flubrobenguane (FBBG), 4-fluoro-3-hydroxyphenethylguanidine (4F-MHPG) and 3-fluoro-4-hydroxyphenethylguanidine (3F-PHPG) were synthesized as described with final purity ≥ 95% on HPLC, and were structures confirmed by NMR in accordance with the literature 22, 23. The other reference compounds are commercially available with purity ≥ 95% and were as follows: norepinephrine bitartrate and phenoxybenzamine (PhB) hydrochloride (TCI Europe, Zwijndrecht, Belgium), desipramine (DMI) hydrochloride (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany), 6-fluorodopamine and meta-hydroxylephedrine (HED) hydrochloride (ABX advanced biochemical compounds GmbH, Radeberg, Germany), nonradioactive meta-iodobenzylguanidine (MIBG) hydrochloride (Activate Scientific, Prien, Germany), and 123I-MIBG (FUJIFILM Toyama Chemical Co., Ltd., Chiba, Japan).

Animal preparation

Male Wistar rats (n = 4, weighing 250-400 g) and New Zealand white rabbits (n = 5, weighing 3.2-3.6 kg) were used for cardiac uptake studies. Anesthesia was induced and maintained during the experiment by 2% isoflurane 24. Large animal PET studies were performed on cynomolgus monkeys (n = 4, weighing 3.9-4.2 kg). Induction of anesthesia in NHPs was conducted by intramuscular injection of ketamine (1.5 mg/kg) and xylazine (0.6 mg/kg) to allow the preparation and handling of the animals. After a tracheal cannula was inserted, 1.5% sevoflurane (SEVOFLURANE Inhalation Solution, Pfizer Japan Inc., Tokyo, Japan) vaporized with 100% oxygen was inhaled, and the tidal volume and respiratory rate of the ventilator were monitored and kept in the normal range throughout the imaging sessions with an anesthesia workstation (Apollo®, Drägerwerk AG & Co. KGaA, Lübeck, Germany) 25.

PET imaging, biodistribution and kinetic studies

A small-animal PET system (microPET FOCUS 120, Siemens, Erlangen, Germany) was used for the rat studies. A 30-min dynamic scan was initiated immediately after the radiotracer injection (10-20 MBq) via the tail vein (in a bolus). For the blocking study in rats, PhB (50 mg/kg, n = 4) was injected 10 min before radiotracer administration. The list-mode data were sorted into three-dimensional reconstructions, which were then rebinned with full three-dimensional binning to reconstruct dynamic images. Data were collected in the following frames: 10 s × 12, 30 s × 6, and 300 s × 5. For the biodistribution studies, 10 min after radiotracer injection (1-2 MBq) under anesthesia, the animals were euthanized for the control or PhB blocking group (50 mg/kg, n = 4-6). Another group of animals was euthanized 60 min after radiotracer administration (1-2 MBq, n = 4-6) to investigate late-phase uptake. Organs of interest were harvested for tissue counting with a γ-counter. The organ-to-blood ratios were calculated following weight and decay correction of tissue counts.

18F-AF78 (3-4 MBq) and 123I-MIBG (approximately 2.5 MBq) were injected into rabbits maintained under anesthesia via the marginal ear vein in a single dose. The NET blocker DMI (1.5 mg/kg) was injected 10 min before radiotracer administration in the blocking study. A 10-min dynamic scan was initiated immediately after the radiotracer injection. The animals were euthanized immediately after the scan, and the organs of interest were harvested and measured in a gamma counter to calculate biodistribution.

NHPs were maintained under anesthesia during imaging and were studied with a PET scanner (PCA-2000A, Toshiba Medical Systems Corporation, Tochigi, Japan). After a 5-min transmission scan and a bolus intravenous (i.v.) injection of the radiotracer 18F-AF78 (approximately 20 MBq) via the saphenous vein, a 60-min dynamic PET scan (10 s × 12, 30 s × 6, and 300 s × 11) was initiated. For the blocking study, DMI was injected 10 min before radiotracer administration, after which the PET scan commenced immediately (n = 4). For the DMI chase study, DMI (1 mg/kg, n = 4) was injected 10 min after radiotracer administration, while the PET scan was initiated simultaneously with radiotracer administration. The catecholamine releasing stimulator tyramine (TYR) was also used in a chase protocol, with 50 µg/kg each injection (n = 4) injected intermittently (5 times, 10-min intervals) after radiotracer administration. The radiotracers 18F-AF78 and 123I-MIBG were formulated in saline for the animal studies. Control animals received normal saline. The inhibitors PhB, DMI and TYR were first dissolved in a minimum amount of DMSO and diluted with saline for administration with a final concentration of DMSO lower than 10%. The experimental timeline is provided in the supplementary material.

PET imaging for rats, rabbits and NHPs was performed from the chest to the upper abdomen, including the heart, lungs, liver and part of the intestines. The reconstructed images were analyzed using AMIDE imaging software (version 1.0.1). For kinetic studies in rats and NHPs, the tissue uptake was determined as standardized uptake values (SUVs), from which time-activity curves of different organs, e.g., heart, blood, lung and liver, were generated with transmission computed tomography images for topographic localization.

Statistics

All results are displayed as the mean ± SD. A two-tailed paired Student's t test was used to compare differences between two dependent groups, and a two-tailed independent Student's t test was used to compare differences between independent groups. Multiple group comparisons were performed using analysis of variance (two-way ANOVA). A P value of less than 0.05 was assumed to be statistically significant. Statistical analysis was performed with GraphPad Prism (version 8.4.3 GraphPad Software, San Diego, USA).

Results

The affinity of AF78 to NET is similar to that of NE and is higher to that of established SNS radiotracers. AF78 showed almost identical transporting affinity by NET when compared to the physiological reference NE (IC50 values 0.42 ± 0.14 µM vs. 0.50 ± 0.16 µM, respectively, n.s.). When compared to other established 18F-labeled SNS radiotracers, the cell uptake affinity of AF78 by NET was significantly higher than that of 4F-MHPG (IC50, 2.78 ± 1.22 µM, P ≤ 0.01) and FBBG (IC50, 2.28 ± 1.05 µM, P ≤ 0.05) but not 3F-PHPG (IC50, 0.78 ± 0.060 µM, n.s.). In addition, a 4.2-fold higher affinity of AF78 relative to MIBG was noted (IC50, 1.86 ± 0.30 µM, P ≤ 0.05). HED revealed an IC50 of 0.39 ± 0.11 µM, (n.s.), which is comparable to that of AF78. DMI, a selective and specific inhibitor of NET, was used as a positive control (Table ).

Biodistribution of Homogeneous 18F-AF78 distribution throughout the left ventricle (LV) could be visualized in healthy Wistar rats. Administration of the radiotracer after PhB pretreatment demonstrated a focal decrease in radiotracer accumulation that remained stable throughout the imaging protocol (Figure ). Time-activity curves showed sustained intense radiotracer uptake in the heart but continuous low values for both the blood pool and liver in control rats (Figure ). Upon PhB pretreatment, cardiac uptake decreased rapidly after the initial perfusion, reaching comparable values to that of the blood pool. Stable cardiac uptake was observed in the biodistribution study showing unchanged the organ-to-blood ratios in the early (10 min) and late (60 min) phases. In the blocking study using the nonselective NET blocker PhB, only uptake in the heart, not in any other organ, was reduced, demonstrating NET-mediated specific radiotracer uptake (Figure ).

Specific cardiac uptake of Distinct cardiac imaging showed a homogeneous distribution of 18F-AF78 throughout the LV in healthy rabbits, while this uptake diminished after pretreatment with the NET blocker DMI (Figure ). By comparing 18F-AF78 and 123I-MIBG in the same animals, although the uptake of the former was lower than that of the latter in the controls, the mean SUV after DMI blockade — which can be considered background — was lower, demonstrating comparable contrast with and a better signal-to-background ratio than obtained with 123I-MIBG (Figure ).

In healthy NHPs, radiotracer accumulation with clear delineation of the LV could already be visualized 2-5 min postinjection. Almost no uptake could be identified in the blood pool, whereas uptake in the liver decreased slowly over time. Therefore, high-contrast images with reduced hepatic uptake were obtained 45-60 min postinjection, supporting the notion of improved image quality at later phases. Little or no radiotracer retention in the lungs was observed over time (Figure ). Time-activity curves derived from 18F-AF78 dynamic images in controls showed high cardiac radiotracer activity after the initial blood pool washout that plateaued throughout the entire scan time with heart-to-blood pool ratios of 2.03 ± 0.53 at 5 min and 1.82 ± 0.71 at 60 min (Figure ). Relative to cardiac uptake, radiotracer accumulation in the liver was elevated at 5 min postinjection (heart-to-liver ratio at 5 min, 0.38 ± 0.29) but decreased over time (heart-to-liver ratios at 30 min, 0.95 ± 0.20 and 60 min, 1.60 ± 0.38, respectively). Further corroborating the improved image quality at later phases, the absolute SUVs of 18F-AF78 in the blood pool (0.72 ± 0.34) and liver (0.91 ± 0.47) were almost identical at 60 min postinjection, whereas cardiac uptake was higher (1.46 ± 0.18). The heart/liver ratio neared the same level as the heart/blood ratio 60 min after tracer administration, resulting in high-contrast cardiac imaging (Figure ).

The kinetic profile of DMI-mediated selective NET inhibition led to markedly reduced radiotracer accumulation in NHP hearts relative to controls, demonstrating high specificity (Figure ). The cardiac uptake of 18F-AF78 was retained at 60 min after radiotracer administration, which could not be diminished by DMI chase but by TYR chase (Figure ). Time-activity curves demonstrated a significant decrease and could not clearly identify the delineation of left ventricular radiotracer uptake after pretreatment with the NET selective blocker DMI (mean SUV at 60 min, DMI blocking 0.45 ± 0.05 vs. controls, 1.46 ± 0.18; P ≤ 0.0001; Figure ). In contrast, cardiac washout was not affected by a single-time chase protocol using the NET inhibitor DMI after the administration of 18F-AF78, suggesting stable radiotracer retention independent of NET activity after initial transport (mean SUV at 60 min, DMI chase and controls, 1.40 ± 0.07 and 1.46 ± 0.18, respectively; n.s.; Figure ). Intermittent chasing with the catecholamine stimulant TYR at a 10-min interval, however, markedly reduced cardiac tracer retention (mean SUV at 60 min, 0.92 ± 0.20, vs. 1.46 ± 0.18; P ≤ 0.005; Figure ), supporting the notion that 18F-AF78 is stored in synaptic vesicles, while its spillover increases in response to TYR stimulation. Hepatic uptake and washout kinetics, however, were not affected in all different conditions of treatment and remained consistent (Figure ), indicating the high specificity of 18F-AF78 for NET.

Discussion

Here, we focused on the NET as the first important step of radiotracer handling at the presynaptic nerve terminal in the heart and systemically evaluated the recently introduced PET agent 18F-AF78 in a cell-based competitive uptake assay. Additionally, we compared its NET affinity to the reference standard NE, MIBG, HED, and other established 18F-labeled SNS radiotracers. In a human neuroblastoma cell line naturally expressing NET, 18F-AF78 demonstrated an IC50 value comparable to that of physiological NE. Among all investigated radiotracers, AF78 shows potential as a fluorine-18-labeled PET ligand for SNS imaging with an affinity comparable to that of HED but significantly more potent than MIBG (Table ). These results serve as an encouraging complement to a previous DMI blocking study, in which NET specificity was similar to that of 131I-MIBG; reverse blocking using either 18F-AF78 or 131I-MIBG has already demonstrated an encouraging in vitro profile 20. Moreover, in vivo studies revealed a favorable biodistribution with good heart-to-blood pool, heart-to-lung and heart-to-liver ratios in various animal species, thereby allowing for clear contrast imaging of the LV, especially at late time points demonstrated in both rats and NHPs. In addition, 18F-AF78 uptake was sensitive to NET-selective inhibition prior to radiotracer injection, indicating a high specificity to neuronal NET in both rabbits and NHPs. Furthermore, unlike HED undergoing continuous cycling (i.e., diffusion and reuptake) at the nerve terminal, cardiac SNS uptake of 18F-AF78 in NHPs was resistant to DMI chase, suggesting stable radiotracer retention independent of NET activity after initial transport. Last, intermittent chase with the NE stimulator TYR led to a marked spillover of 18F-AF78 from the heart, supporting the notion of stable storage and release kinetics similar to its physiological reference, NE. Overall, the kinetic profile can be used to reflect the mechanistic integrity of NE vesicular turnover pathways, particularly at the presynaptic nerve terminal (Figure ) 26, rendering 18F-AF78 a highly promising SNS imaging agent. These encouraging results should trigger further studies of 18F-AF78, including assessment of biodistribution in humans or to test its clinical significance in multiple challenging scenarios, such as ischemic and nonischemic HF, diabetes-related CAN, valvular heart diseases, cardiomyopathies, or cardiac arrhythmias, including the guidance of resynchronization therapies 8.

In light of these encouraging results, the presented findings should trigger further head-to-head comparisons for quantitative assessments of denervated myocardium between 18F-AF78 and other reported radiotracers. These considerations are further fueled by the fact that, in contrast to the relatively constant hepatic uptake of 11C-HED, 18F-AF78 demonstrated fast liver washout in NHPs (Figures ) in a manner similar to other 18F-labeled radiotracers, as demonstrated by the increasing heart-to-liver ratios over time (5 min after tracer i.v., 0.38 ± 0.29 vs. 30 min, 0.95 ± 0.20 vs. 60 min, 1.60 ± 0.38) 27, 28. In comparison, the heart-to-liver ratios of 18F-4F-MHPG and 18F-3F-PHPG measured in rhesus macaque monkeys reached 2.2 ± 0.8 and 2.5 ± 0.3, respectively, not before 90 min after radiotracer injection 27. In contrast, cardiac retention of 18F-AF78 decreased only slightly in the time-activity curves, in which the SUV dropped from 1.74 ± 0.24 at 5 min to 1.46 ± 0.18 at 60 min postinjection (Figure ). A similar pattern of fast liver washout has also been observed with 18F-3F-PHPG and 18F-4F-MHPG or 18F-FBBG in human studies 23, 29. Therefore, clinical protocols should carefully address the issue of high hepatic uptake at early time points, e.g., by employing an appropriate waiting time after tracer injection before scan initiation. In addition, different hepatic time-activity curves seen in rats and NHPs could also be explained by plasma metabolism due to species discrepancy, and future studies are required to investigate such metabolic aspects. Considering the monoamine oxidase-resistant guanidine tracer structure and relatively stable ether bond, the passive diffusion of the tracers from storage vesicles due to increased lipophilicity through the introduction of a 3-fluoropropyl moiety could explain the higher clogP compared to the lead structure 3F-PHPG 1.27 vs. AF78 1.83, and lower value compared to MIBG 2.72 (clogP values calculated with MarvinSketch 20.13.0, ChemAxon). In the heart, secretory vesicles are acidic under physiological conditions and maintain a pH of 5.0-5.7, depending on the conditions 30, while radiotracers with guanidine moieties are more basic than physiological neurotransmitters with higher pKa values (NE 9.74, MIBG 11.75, MHPG 12.3, FBBG 11.68 and AF78 12.08 (pKa values calculated with MarvinSketch 20.13.0, ChemAxon)). Therefore, on the one hand, after active transport into vesicles, they are stably stored in the heart; on the other hand, the liver uptake and washout rate may at least partially relate to their lipophilicity/clogP value. Further cellular studies are necessary, especially to investigate the Vmax and km values of different tracers to establish kinetics models and clarify the structure-activity relationships of NET-targeting radiotracers.

Although all cardiac sympathetic innervation radiotracers are transported to the nerve terminals in the heart, their subcellular kinetics vary in terms of intracellular storage mode, exemplified by 11C-HED 22. Compared to the stable retention of 123I-MIBG or 18F-FBBG in storage vesicles, 11C-HED undergoes continuous diffusion and NET-mediated reuptake to maintain equilibrium, as shown by decreased cardiac radiotracer uptake after the DMI chase 22. On the one hand, 18F-AF78 demonstrated stable radiotracer retention in the nerve terminals, resistant to the DMI chase (Figure ). However, TYR led to the release of 18F-AF78, likely via two different mechanisms 31. First, by releasing the storage vesicles at the nerve terminals, and second, by acting as a substrate to the NET, the initial transport of TYR into the neuron initiates reverse transport of NE and radiolabeled neurotransmitters within the cytoplasm 32. Thus, the releasing effect was initiated after the intermittent administration of TYR (Figure ); thereafter, an equilibrium would be reached over time since the released radiotracer would be transported back into the nerve terminal, likely due to the high specificity of 18F-AF78 to NET, in a pattern similar to that of 18F-FBBG 33.

In addition to the in vitro and in vivo studies, an in silico hypothesis might hint at a potential explanation for the high NET affinity of AF78. Because there is currently no available X-ray structure of NET, PDB6M0Z, a dopamine transporter with NET-like mutations (D121G/S426 M/F471 L) in NE-bound form, was first chosen for the docking study 34. Adopting the conformation of PBD6M0Z, the 3-dimensional structure of NET as a Homo-AF model was established using AlphaFold prediction. In this model, the smaller radiotracers, including HED, MIBG, 4F-MHPG, 3F-PHPG and 6F-DA (Table ), can be docked into the primary binding site in a pattern similar to NE, with amino or guanidine groups that fit into the binding pocket A, while those radiotracers with a hydroxy group may form a hydrogen bond with A145 according to the model established by Pidathala (Figure ) 34. However, the molecules with a “tail” (3-fluoropropyl group, e.g., FBBG and AF78) do not fit well into the proposed binding pockets as the “tail” flips between binding pocket B and C (Figure ). Therefore, induced-fit docking was performed to allow a better and higher range of rotation of the amino acid moieties. After this adjustment, only one AF78 orientation was present, with the 3-fluoropropyl group pointing to a hydrophobic binding pocket between TM3 and TM8, in accordance with pocket B in Pidathala's model. Most importantly, in this binding position, the fluoride points to pocket C, and the benzene ring of AF78 forms a stacked T-shaped π-π interaction with Y152 and F323, i.e., the benzene ring of AF78 is perpendicular to both amino acid moieties (Figure ). Due to the strong electronegativity of the meta-fluorine substituent, this stacked π-π interaction may be enhanced, thereby stabilizing the binding of AF78 to NET, although AF78 lacks an important hydroxy group similar to other radiotracers that are responsible for a hydrogen bond. The T-shaped π-π stacking configuration is positively correlated with the electronegativity of the meta-substitution on the benzene ring, i.e., F

Overall, the resistance to DMI chase together with the susceptibility to intermittent TYR chase indicates the mode of transportation and storage of 18F-AF78 in sympathetic neurons, which is characterized by active transport via NET followed by stable storage in presynaptic vesicles. Of note, such an in vivo kinetic profile represents a similar pattern as 123I-MIBG and 18F-FBBG 22. In addition, it is especially worth mentioning that the pilot imaging studies in patients with heart failure using 18F-3F-PHPG and 18F-4F-MHPG, which share the phenethylguanidine core structure with AF78, have provided reliable quantitative metrics of regional sympathetic nerve density 36. As a result, given its high affinity to the subcellular target and high-contrast images in various animal species, 18F-AF78, as a second-generation SNS radiotracer that also has a phenethylguanidine core structure, may be readily applicable for human use.

This study has several limitations. Future studies should also focus on the 18F-AF78 interactions in the presynaptic nerve terminal, e.g., by inhibiting the vesicular monoamine transporter 2 (VMAT2) using tetrabenazine. Such studies are important as NE analogs are subject to vesicular packaging by this transporter 37. Another factor affecting uptake and retention is the specific activity of the radiotracer, as has been demonstrated by our research on 11C-HED in rat hearts. 38. The cold mass effect should certainly be addressed after establishing a fully repeatable radiolabeling protocol suitable for clinical application. The interpretation of the tyramine induced catechol release should be carefully considered, because it may interact with several targets , such as VMAT2, monoamine oxidase, and alpha receptors. Therefore, further studies using reserpine may provide more solid support for the submolecular storage and release mechanism of AF78. Moreover, given the low number of investigated animals per group due to ethical issues, statistical analyses are limited, and a weak effect could have been missed. Increased neuronal retention times and lower neuronal uptake rates have been advocated to provide more reliable read-outs of nerve integrity, as the blood flow dependency is minimized, while the NET transport rate would be slower 39. Therefore, future studies should also investigate these parameters of 18F-AF78 relative to established nerve radiotracers, preferably in dedicated disease models at late time points 90-120 min postinjection, e.g., in heart failure following myocardial infarction. Nonetheless, a comprehensive evaluation of the radiotracer kinetics in healthy animals, as presented in the present study, may be valuable as it allows us to adequately interpret findings in preclinical scenarios with deteriorating cardiac performance. As such, the present study mapping the interactions of 18F-AF78 at the NET in various (healthy) species may be interpreted as laying the proper groundwork for important follow-up studies. Further studies will also provide information on tracer uptake related to sympathetic nerve density or NE content 40. Moreover, test-retest studies evaluating the reproducibility of radiotracer kinetics in various species and humans are needed to ensure that the interpreting molecular imaging expert has high certainty regarding the robustness of the derived cardiac nerve integrity information.

Conclusion

Relative to the physiological reference NE, the second-generation NET radiotracer 18F-AF78 demonstrated almost identical in vitro affinity to NET and was comparable to other established 18F-labeled SNS radiotracers and MIBG. Studies across various species demonstrated reassuring biodistribution and a positive kinetic profile, including specific NET-mediated transport, followed by a stable storage and release mode. Computational modeling also provided a rationale for its high NET affinity. These results render 18F-AF78 a promising SNS radiotracer for prognostication and molecular image-guided strategies in future cardiovascular medicine, e.g., to guide interventions such as resynchronization therapies in ischemic and nonischemic HF, diabetes-related CAN or valvular heart diseases. Beyond cardiovascular diseases, further applications include systems-based organ analyses by providing holistic benefit for the target (heart) and remote organ (kidney) in patients with cardiorenal syndrome or in a cardio-oncology setting, e.g., to prevent relevant cardiotoxicity of anticancer therapeutics.

Supplementary figures.

Click here for additional data file.

Table 1

Competitive cell uptake affinity against 3H-norepinephrine (3H-NE) in human neuroblastoma cells.

Testing Compound #	Structure †	IC50 (µM) ‡, §
DMI		0.010 ± 0.0015*
PhB		0.82 ± 0.30 n.s.
NE		0.50 ± 0.16
6F-DA		1.20 ± 0.33*
MIBG		1.75 ± 0.47**
FBBG		2.28 ± 1.05*
4F-MHPG		2.78 ± 1.22**
3F-PHPG		0.78 ± 0.060 n.s.
HED		0.39 ± 0.11 n.s.
AF78		0.42 ± 0.14 n.s.

NE = norepinephrine DMI = desipramine. PhB = phenoxybenzamine. 6F-DA = 6-fluorodopamine. FBBG = flubrobenguane. 4F-MHPG = 4-fluoro-3-hydroxyphenethylguanidine. 3F-PHPG = 3-fluoro-4-hydroxyphenethylguanidine. HED = metahydroxyephedrine. MIBG = metaiodobenzylguanidine. AF78 = 1-(3-fluoro-4-(3-fluoropropoxy)phenethyl)guanidine. In addition to NET blockers DMI and PhB, the radiotracers are colored to show the similarity in their chemical structures. Red represents a radionuclide acting as a radiotracer. Values are presented as the means ± SD for individual assays (n = 4). Compared to NE, where n.s. P > 0.05, * P ≤ 0.05, ** P ≤ 0.01.