Radiolabeling and initial biological evaluation of [¹⁸F]KBM-1 for imaging RAR-α receptors in neuroblastoma

Abstract

Retinoic acid receptor alpha (RAR-α) plays a significant role in a number of diseases, including neuroblastoma. Children diagnosed with high-risk neuroblastoma are treated 13-cis-retinoic acid, which reduces risk of cancer recurrence. Neuroblastoma cell death is mediated via RAR-α, and expression of RAR-α is upregulated after treatment. A molecular imaging probe that binds RAR-α will help clinicians to diagnose and stratify risk for patients with neuroblastoma, who could benefit from retinoid-based therapy. In this study, we report the radiolabeling, and initial in vivo evaluation of [¹⁸F]KBM-1, a novel RAR-α agonist. The radiochemical synthesis of [¹⁸F]KBM-1 was carried out through KHF₂ assisted substitution of [⁻¹⁸F] from aryl-substituted pinacolatoesters-based retinoid precursor. In vitro cell uptake assay in human neuroblastoma cell line showed that the uptake of [¹⁸F]KBM-1 was significantly inhibited by all three blocking agents (KBM-1, ATRA, BD4) at all the selected incubation times. Standard biodistribution in mice bearing neuroblastoma tumors demonstrated increased tumor uptake from 5 min to 60 min post radiotracer injection and the uptake ratios for target to non-target (tumor: muscle) increased 2.2-fold to 3.7-fold from 30 min to 60 min post injection. Tumor uptake in subset of 30 min blocking group was 1.7-fold lower than unblocked. These results demonstrate the potential utility of [¹⁸F]KBM-1 as a RAR-α imaging agent.

Retinoic acids (RA) exhibit various important biological functions in the control of normal growth, differentiation and fetal development.^1,2 Their activity is mediated through retinoic acid receptors (RAR), which are ligand activated nuclear receptors.^3,4 This receptor family consists of α, β, and γ subtypes; each subtype has multiple isoforms. Recent studies have demonstrated that both all-trans retinoic acid (ATRA) and 13-cis-retinoic acid (isotretinoin) promote morphological differentiation and inhibit growth of neuroblastoma cells, mainly through activation of RAR-α subtypes.^5,6 Human neuroblastoma cell lines treated with ATRA differentiate in culture, forming neurites and exhibiting growth arrest. Haber et al. demonstrated that RAR-α expression was specifically induced in neuroblastoma cell lines by ATRA and other differentiating agents tested.⁷ RAR-α expression appears to be necessary for ATRA to have therapeutic effects on neuroblastoma cells in vitro.⁷ Furthermore, reduced expression of RAR-α in vivo may contribute to the process of neuroblastoma tumorigenesis and resistance to ATRA therapy.⁸ Although attempts have been made to measure RAR-α activity using neuroblastoma cell lines in vitro, no reliable in vivo clinically relevant probe exists to measure and quantify RAR-α expression.⁷ A RAR-α binding PET radiotracer could be a key tool to measure real-time RAR-α expression and elucidate effects of RAR-α on the pathogenesis of neuroblastoma.

Das et al. developed a novel boron-based RAR-α agonist (BD4) specifically for treatment of kidney diseases.⁹ They analyzed structure-activity relationships (SAR) derived from in vitro binding assays and demonstrated that 10-fold concentration of ATRA was required to induce a 50% competitive inhibition of BD4 binding to a GST-tagged RAR-α. This confirms that BD4 binds to RARα with a higher affinity than ATRA.^9,10 Using BD4 as a lead structure, we designed an F-18 based PET tracer by replacing the boronic ester group with a borontrifluoride group.¹¹ Here we report for the first time a novel radiolabeling procedure, and initial in vitro and in vivo biological evaluation of [¹⁸F]KBM-1 as a potential imaging probe for RAR-α in a murine model of neuroblastoma.

The precursor BD4 for the radiochemical reaction was synthesized following the previously reported method from aryl aldehyde derivative using Michael-Aldol and Witting reactions.^9,12–14 The non-radioactive standard KBM-1 was obtained from the precursor aryl bornic ester moiety BD4 using KF/KHF2 chemistry as previously reported.^9,13,15 Radiochemical synthesis of [¹⁸F]KBM-1 was achieved by adding the precursor BD4 (200 mM) to the azeotropically dried [¹⁸F⁻] residue followed by addition of 2 M HCl (6 μL) and 0.4 M KHF₂ (1.0 μL) at a pH of ~3, using published methods (Scheme 1).^15–17 The reaction mixture was heated in a closed vial setup at 40 °C for 130 min. The final product was purified on silica gel chromatography using EtOH:NH₄OH (95:5) as the eluent. The radiotracer [¹⁸F]KBM-1 was formulated with 0.9% NaCl solution and filtered through a 0.22 μm pyrogen-free Millipore filter. [¹⁸F] KBM-1 was produced in 6–8% decay corrected radiochemical yield in a total synthesis and formulation of 2.5 h (n = 6). The overall radiochemical purity was ~98% and the specific activity of the [¹⁸F]KBM-1 was ~1800–2000 mCi/μmol (decay corrected to end of synthesis).

Scheme 1.

Radiochemical synthesis of [¹⁸F]KBM-1.

We conducted in vitro cell uptake assays using SH-SY5Y cell line, a human neuroblastoma cell line previously reported to express RAR-α with and without the blocking agents to evaluate the specificity and binding efficiency of [¹⁸F]KBM-1.^7,18–21 The precursor molecule, BD4 has been reported to demonstrate high and specific binding towards RAR-α over ATRA. The fluorometric binding assay showed that BD4 binds specifically to RARα with an affinity of 14 ± 5 nM (K_D) over ATRA.⁹ We used nonradioactive standard KBM1, ATRA, and precursor BD4 as blocking agents in our in vitro cell uptake assay. The uptake of [¹⁸F]KBM-1 was ~62% and 60% inhibited by non-radioactive KBM-1 as blocking agent, ~37% and 32% inhibited by ATRA as blocking agent, and ~58% and 51% inhibited by precursor BD4 as blocking agent after 15 min and 60 min incubation times respectively (Fig. 1). These findings indicate that [¹⁸F]KBM-1 uptake was significantly decreased in neuroblastoma cell lines with all three blocking agents (1) non-radioactive standard KBM-1, (2) ATRA, the widely used RAR-α targeting agent and (3) BD4, the previously published highly selective RAR-α ligand; thus demonstrating binding of [¹⁸F]KBM-1 to RAR-α.

Fig. 1.

In vitro cell uptake of [¹⁸F]KBM-1 by at baseline and blockade conditions at 15 min and 60 min incubation time points. The data were expressed as % injected dose (ID)/mg of protein present in each well with p values ≤0.05 considered statistically significant (n = 6). All the blocking sections received 60-fold excess concentration of the blocking agent including, KBM-1, ATRA, and BD4 five min prior to the radiotracer addition.

We then conducted a biodistribution study with [¹⁸F]KBM-1 in male BALB/c mice with neuroblastoma tumors. Mice were grouped into three groups (n = 4) based on 5 min, 30 min and 60 min post injection (p.i). [¹⁸F]KBM-1 displayed rapid clearance from most of the major organs from 5 min to 60 min post injection (Fig. 2). The initial uptake at 5 min p.i in blood, kidney and liver was 4.37 ± 1.131, 26.22 ± 2.351 and 2.92 ± 0.209 (%ID/g) respectively. The uptake had mostly cleared by 60 min, with blood levels at 0.93 ± 0.512, kidney at 10.50 ± 1.509 and liver at 1.95 ± 0.233 (% ID/g) respectively. However, from 5 min to 60 min, bone uptake was constant or slightly increased-%ID/g of 1.65 ± 0.781 (5 min), 2.65 ± 0.112 (30 min) and 2.48 ± 0.131 (60 min), suggesting minor metabolic in vivo defluorination.

Fig. 2.

Biodistribution of [¹⁸F]KBM-1 in tumor-bearing mice (n = 4) after 5 min, 30 min and 60 min post injection. Results are expressed in % injected dose (ID)/g ± standard deviation, with p values ≤0.05 considered statistically significant.

Tumor uptake remained stable, while the uptake ratios for target to non-target (tumor: muscle) increased from 2.2-fold to 3.7-fold at 5 min to 60 min p.i respectively (Fig. 3). To demonstrate specific binding, we performed a blocking experiment in a subset of mice from the 30 min group (n = 2). The blocking agent was the nonradioactive KBM-1 (15 mg/kg) and was administered 15 min before the radiotracer injection. Tumor uptake in the 30 min blocking group was 1.7-fold lower than in the unblocked ones i.e., 0.80 ± 0.23 (%ID/g) versus 1.46 ± 0.81 in the unblocked group, demonstrating specificity for RAR-α receptors in the tumor (Table 1).

Fig. 3.

Ratio of target (tumor): nontarget (muscle) at 5 min, 30 min and 60 min post injection of [¹⁸F]KBM-1 in tumor bearing mice (n = 4).

Table 1.

Biodistribution of [¹⁸F]KBM-1 in tumor bearing mice (n = 3) at 30 min post injection, with and without blocking agent. Results are expressed in % injected dose (ID)/g ± standard deviation.

Organs	30 min-nonblockade	30 min-blockade
Brain	0.23 ± 0.09	0.15 ± 0.07
Heart	1.91 ± 1.12	1.70 ± 0.75
Blood	4.01 ± 2.31	2.91 ± 1.19
Lung	2.78 ± 0.77	2.88 ± 0.11
Liver	2.32 ± 0.22	2.44 ± 0.84
Pancreas	2.03 ± 1.23	2.13 ± 0.65
Kidneys	24.33 ± 2.36	20.43 ± 2.78
Tumor	1.46 ± 0.81	0.81 ± 0.23
Muscle	0.49 ± 0.11	0.40 ± 0.07
Bone	2.65 ± 1.12	3.18 ± 1.65

In conclusion, [¹⁸F]KBM-1 was synthesized for the first time using ${\text{ArBF}}_3^{-}/{\text{KHF}}_2$ chemistry with high radiochemical purity and specific activity. The in vitro cell uptake assay indicated that [¹⁸F] KBM-1 has the potential to serve as a RAR-α imaging agent. Initial in vivo biodistribution studies in BALB/c tumor-bearing mice demonstrate the potential of [¹⁸F]KBM-1 as a radiotracer for imaging neuroblastoma. Although we demonstrate novel radiochemistry and initial evaluation of [¹⁸F]KBM-1 binding in vitro and in vivo towards RAR-α, we are currently characterizing the complete receptor specificity, in vivo metabolic profile, kinetics and its performance in microPET imaging studies.