Improved Radiosynthesis and Automation of [¹¹C]2-(2,6-difluoro-4-((2-(N-methylphenylsulfonamido)ethyl)thio)phenoxy)acetamide ([¹¹C]K2) for Positron Emission Tomography of the Glutamate α-Amino-3-hydroxy-5-methyl-4-isoxazole Propionic Acid (AMPA) Receptor

Abstract

A new automated radiosynthesis of [¹¹C]2-(2,6-difluoro-4-((2-(N-methylphenylsulfonamido)ethyl)thio)phenoxy)acetamide ([¹¹C]K2), a radiopharmaceutical for the glutamate α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptor is reported. While manual syntheses have been described, these are unsuitable for routine production of larger batches of [¹¹C]K2 for (pre)clinical PET imaging applications. To meet demands for the imaging agent from our functional neuroimaging collaborators, herein we report a current good manufacturing practice (cGMP)-compliant synthesis of [¹¹C]K2 using a commercial synthesis module. The new synthesis is fully automated and has been validated for clinical use. The total synthesis time is 33 min from end-of-bombardment and the production method provides 2.66±0.3 GBq (71.9±8.6 mCi) of [¹¹C]K2 in 97.7±0.5% radiochemical purity and 754.1±231.5 TBq/mmol (20,382.7±6,256.1 Ci/mmol) molar activity (n=3). Batches passed all requisite quality control testing confirming suitability for clinical use.

1. Introduction

The glutamate α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptor is believed to play an important role in many neurophysiological disorders. It is thought that differences in the function and abundance of AMPA receptors in the brain are reflected in disorders such as epilepsy, depression, schizophrenia and Alzheimer’s disease.^{1a, b, c} Having a means of visualizing and quantifying AMPA in vivo in humans would greatly aid in the understanding of its role in the various neurological disorders in which its dysfunction is implicated. In 1996–1997, Segkiguchi and coworkers reported the synthesis of 4-[2-(phenylsulfonylamino)ethylthio]-2,6-difluoro-phenoxyacetamide (PEPA, 1), an allosteric potentiator of AMPA receptors.^{2a, b} Subsequently, PEPA was modified through additional N-methylation to yield 2-(2,6-difluoro-4-((2-(N-methylphenylsulfonamido)ethyl)thio)phenoxy)acetamide (K2). The ¹¹C-labeled version of K2 ([¹¹C]K2, 2) was successfully synthesized (Figure 1), and the first example of quantifying AMPA receptors in living human brain using PET imaging was disclosed.³ In an autoradiographic saturation binding study, Miyazaki and colleagues found that [¹¹C]K2 exhibited high binding affinity (K_d = 48.0 nM) and appropriate binding capacity (B_max = 364.6 fmol per mg tissue, or 364.6 nM assuming 1 kg of tissue corresponds to 1L).^3a The binding potential (BP), estimated from B_max/K_d, is a metric used by imaging scientists when considering the suitability of a radiotracer for imaging a target binding site in human subjects, particular in the brain. Although the ideal binding potential is unknown, ≥5 has been recommended for brain tracers by the National Institutes of Health. Given these reported values, the BP for K2 = 364.6/48 = 7.6, confirming suitability for imaging AMPA receptors.

Figure 1.

AMPA Radioligand [¹¹C]K2

We set out to adapt the published radiosynthesis of [¹¹C]K2 for use with an automated synthesis module to enable large scale production of the radiotracer for use in pre-clinical and clinical AMPA PET studies at the University of Michigan. During the initial brainstorming and planning stages, we briefly considered designing a ¹⁸F-labeled derivative of K2. However, an inspection of the literature revealed that ¹⁸F-labeled K2 analogs have had many difficulties related to their radiosyntheses.⁴ Therefore, we decided to focus on adapting the previously reported manual radiosynthesis of [¹¹C]K2 for validated and automated cGMP production as well as establishing quality control (QC) release criteria, in order to enable (pre)clinical studies. In this note, we report a fully automated radiosynthesis of [¹¹C]K2 using a General Electric (GE) Tracerlab FX_C-Pro that employs only class 3 solvents, along with complete QC testing results for three validation batches.

2. Experimental

2.1. General Considerations

Full experimental details for syntheses and radiosyntheses, including copies of nuclear magnetic resonance (NMR) spectra and radio-high-performance liquid chromatography (radio-HPLC) traces, are provided in the supporting information.

2.2. Radiosynthesis of [¹¹C]K2

[¹¹C]K2 was produced using a GE Tracerlab FX_C-Pro automated radiochemistry synthesis module. [¹¹C]CO₂ (111 GBq, 3 Ci) was generated via ¹⁴N(p, 〈)¹¹C nuclear reaction using a GE PETtrace cyclotron (60 μA beam for 30 min) and was then allowed to react with H₂ gas in the presence of Ni to generate [¹¹C]CH₄. The resulting [¹¹C]CH₄ was exposed to iodine vapors at 720 °C to afford [¹¹C]MeI which was passed through a silver triflate-graphpac column to furnish the desired electrophilic reagent, [¹¹C]MeOTf (~ 37 GBq, 1 Ci). The precursor PEPA (1, 1.0 mg) was dissolved in anhydrous EtOH (200 μL). To this solution was added 0.5 M NaOH (5 μL), and the reaction mixture was then allowed to react with [¹¹C]MeOTf at room temperature, for 3 mins. The crude reaction mixture was purified by semi-preparative HPLC (for a representative trace, see Figure 2) using a Phenomenex 10μ C18(2) 100, 150 × 10 mm column at a flow rate of 4 mL/min (mobile phase: 40% EtOH, 10 mM NH₄OAc, pH 5.0; retention time (t_R) ~9–10.5 min) and reformulated using a C18 sep-pak cartridge. The purified product was eluted from the C18 cartridge using 0.5 mL of ethanol (USP) and diluted with 9.5 mL USP saline (containing 25 μL of ascorbic acid (500 mg/mL) USP for injection), resulting in a final formulation of 5% ethanol in saline, before passing through a 0.22 μm sterile filter into a sterile 10 mL dose vial. A sample of the formulated dose was submitted for quality control testing as described below.

Figure 2:

Representative semi-preparative HPLC trace for purification of [¹¹C]K2 (t_R ~9–10.5 min)

2.3. Quality Control Testing

Quality control of [¹¹C]K2 was conducted according to the guidelines outlined in Chapter <823> of the US Pharmacopeia and standard procedures that have been previously reported.^{5a, b} In Section 2.3.1, HPLC analysis is described, while Section 3 summarizes three process verification batches using the developed method. All acceptance criteria were met or exceeded by the three validation batches confirming their suitability for clinical use and validating the radiosynthesis method for future production use.

2.3.1. HPLC Analysis

The chemical and radiochemical purity of [¹¹C]K2 was analyzed using a Shimadzu LC2010 HPLC equipped with a Bioscan/Eckert and Ziegler radioactivity detector and an ultraviolet (UV) detector (Column: Luna-5u-C18–150×4.6, flow rate: 2 mL/min, mobile phase: 35% MeCN, 10 mM NH₄OAc, pH 5, t_R ~ 5.7 min). A representative HPLC trace is shown in Figure 3, and additional details, including zoomed in traces for determining RCP and quantifying chemical impurities can be found in the Supporting Information.

Figure 3:

Representative analytical UV (A) and Radiochemical (B) HPLC traces for formulated [¹¹C]K2 (t_R ~5–6min)

2.3.2. pH Test

The pH of formulated [¹¹C]K2 was checked by placing a small sample of the dose onto a pH indicator strip. Visual comparison to the scale provided with the strips was used to determine the pH. The acceptable pH range is 4.5 to 7.5.

2.3.3. Filter Integrity Test

The sterile filter that was used to collect the dose (with the needle still attached) was connected to a supply of nitrogen via a regulator. The needle was then submerged in water and the nitrogen pressure was slowly increased. If the pressure reaches above the filter acceptance pressure of 44 psi without a visible stream of bubbles, the filter’s integrity is verified.

2.3.4. Endotoxin Test

A Charles River Laboratories EndoSafe Portable Testing System was used to measure the endotoxin levels in the radiopharmaceutical doses in accordance with the United States Pharmacopeia’s standards. Acceptable does must contain less than 175 Endotoxin Units (EU)/batch (<175 EU/mL) to meet quality specifications.

2.3.5. Sterility test

The sterility of the [¹¹C]K2 radiopharmaceutical doses were evaluated using culture tubes containing fluid thioglycolate media (FTM) and tryptic soy broth (TSB), which were inoculated with dose samples. These cultures were then incubated along with positive and negative controls for 14 days. The FTM was used to detect anaerobes, aerobes, and microaerophiles and the TSB was used to test for non-fastidious and fastidious microorganisms. The culture tubes underwent visual inspections on the 3^rd, 7^th and 14^th days. The sterility was assessed by comparing the cultures to both the positive and negative controls; no observed microbial growth in the dose and negative control samples indicates sterility, while the positive control must show growth (turbidity) after 14 days.

3. Results and Discussion

Precursor PEPA (1) and K2 (2) standard were synthesized using reported procedures, with only slight variations (see Supporting Information for details).^6a–c PEPA precursor (1) was stable for ~1 year when divided into 1.0 mg aliquots and stored at −20 °C.⁷ Before developing an automated radiosynthesis of [¹¹C]K2, an initial screening of reaction conditions was explored using several solvent and electrophilic ¹¹C-labeled synthons (Table 1, entries 1–5). It was found that starting from PEPA precursor (1), [¹¹C]K2 (2) was generated in good radiochemical conversion (RCC) using [¹¹C]CH₃I in DMF with 0.5 M NaOH (1.01 equiv.) at 80 °C for 3 min 30 secs (83% RCC, n = 1, Table 1, entry 1) and [¹¹C]CH₃OTf in EtOH with 0.5 M NaOH (1.01 equiv.) at rt for 3 mins (81±4% RCC, n = 2, Table 1, entry 4). The in-loop method using [¹¹C]CH₃OTf and EtOH with 0.5 M NaOH (1.01 equiv.) was also explored but resulted in variable RCCs (Table 1, entry 5). Interestingly, [¹¹C]CH₃I in EtOH (Table 1, entry 2) or DMSO (Table 1, entry 3) with 0.5 M NaOH (1.01 equiv.) at 80 °C for 3 min 30 secs resulted in low RCCs. Considering the similar RCCs between entries 1 and 4 we decided to move forward with automated radiosynthesis development using the conditions in entry 4. Using ethanol as reaction solvent instead of DMF eliminates the detection of residual solvents during quality control tests as ethanol is a class 3 solvent,⁸ whereas DMF is a more toxic class 2 solvent. Eliminating GC testing decreases total production time, which given the short half-life of ¹¹C (20 min), reduces losses due to radioactive decay and improves final activity yields.

Table 1.

Optimization of Radiochemical Reaction Conditions

entry	solvent (200 μL)	base (equiv.)	alkylating agent	temperature/time	RCC (%)
1	DMF	0.5 M NaOH (1.01)	11CH3I	80 °C/3 min 30 s	83 (n = 1)
2	EtOH	0.5 M NaOH (1.01)	11CH3I	80 °C/3 min 30 s	11 (n = 1)
3	DMSO	0.5 M NaOH (1.01)	11CH3I	80 °C/3 min 30 s	36 ± 4 (n = 2)
4	EtOH	0.5 M NaOH (1.01)	11CH3OTf	rt/3 mins	81 ± 4 (n = 2)
5	EtOH (in-loop)	0.5 M NaOH (1.01)	11CH3OTf	rt/5 mins	32 ± 43 (n = 2)

Next, a fully automated test run was carried out using the optimal conditions developed in the initial optimization screen (vida supra). Briefly, ~111GBq (~3 Ci) of ¹¹C was produced via ¹⁴N(p, α)¹¹C nuclear reaction and converted by standard procedures to the desired electrophilic reagent, [¹¹C]MeOTf.^5a Approximately 15 mins before end of bombardment, PEPA (1, 1.0 mg) was dissolved in anhydrous EtOH (200 μL) and 0.5 M NaOH was added (5 μL). The mixture was vortexed for 1 minute and then added to the TRACERLab reactor. [¹¹C]MeOTf (~37 GBq, 1 Ci) was bubbled through the reaction mixture at room temperature for 3 mins. The crude reaction mixture was diluted with HPLC mobile phase and purified by semi-preparative HPLC using a Phenomenex C18 column at 4 mL/min. The product peak at ~9–10.5 min was collected for 90 s (see Figure 2 and Supporting Information for representative HPLC traces), diluted and reformulated using a C18 sep-pak cartridge. [¹¹C]K2 was trapped on the cartridge and washed with water to remove residual HPLC mobile phase. In our initial efforts, the labeled product was then eluted using ethanol and diluted with USP saline. The formulated product was passed through a 0.22 μm Millipore filter to afford a final injectable dose of [¹¹C]K2 (2) (2.79 GBq (75.5 mCi), 2.5% non-corrected radiochemical yield based upon 111 GBq (3 Ci) of [¹¹C]CO₂;) in 33 mins from end-of-bombardment, with a radiochemical purity of 96% (n = 1) (Scheme 1). However, during our initial exploration of this automated radiolabeling sequence, it was found that after purification and reformulation, slow radiolysis was observed within 1 hour. Therefore, in the finalized automated radiosynthesis, a small amount (25 μL) of USP ascorbic acid was added to the final dose to inhibit radiolysis. With an optimized process in hand, three consecutive batches were produced to give 2.66±0.3 GBq (71.9±8.6 mCi) of [¹¹C]K2 in 97.7±0.5% RCP and 754.1±231.5 TBq/mmol (20,382.7±6,256.1 Ci/mmol) molar activity (n=3). Quality control test results are summarized in Table 2, and all doses met or exceeded release criteria,⁹ validating the production method of [¹¹C]K2 for (pre)clinical applications. Inclusion of ascorbic acid is known to reduce radiolysis¹⁰ and, reflecting this, RCP of the formulated [¹¹C]K2 remained >96% 1 h post-end of synthesis. The average mass of K2 in a batch was 0.16±0.06 μg/mL (1.6 μg/10 mL batch) which, given K2 has been administered up to 0.349±0.069 μg/kg (i.e. 20.94 μg to a 60 kg subject) without adverse events,^3c further confirms suitability of this method for routine production of [¹¹C]K2 for clinical use. Using these numbers, the K2 concentration specification was set at ≤2.1 μg/mL, allowing administration of an entire 10 mL batch to a single research subject. Total impurity limits (including precursor) were set to ≤1.0 μg/mL, or ≤ 10 μg/batch, consistent with batch impurity limits in our recent regulatory submissions to FDA.

Scheme 1.

Automated Test Radiosynthesis of [¹¹C]K2 (2)

Table 2.

QC Data for the Process Verification Batches of [¹¹C]K2 (2)

QC Test	Release Criteria	Batch 1	Batch 2	Batch 3
Radiochemical Purity	≥ 90%	97.6 %	98.2 %	97.2 %
K2 Concentration	≤ 2.1 μg/ml	0.16 μg/mL	0.096 μg/mL	0.22 μg/mL
Total impurities (including precursor)	≤ 1.0 μg/mL	0.06 μg/mL	0.008 μg/mL	0.012 μg/mL
Molar Activity	55.5 TBq/mmol (≥ 1500 Ci/mmol)	690 TBq/mmol (18,641 Ci/mmol)	1011 TBq/mmol (27,325 Ci/mmol)	562 TBq/mmol (15,182 Ci/mmol)
pH	4.5 – 7.5	5.5	5.5	5.5
Visual Inspection	Clear, colorless, no precipitate	Pass	Pass	Pass
Radiochemical Identity	RRT*: 0.9–1.1	1.02	1.02	1.02
Radionuclide Identity	t1/2: 18.4–22.4 min	20.5 min	19.8 min	20.3 min
Filter Membrane Integrity Test	≥ 44 psi	50 psi	47 psi	48 psi
Bacterial Endotoxin Test (initiated prior to release)	< 17.5 EU†/mL	< 2.00 EU/mL	< 2.00 EU/mL	< 2.00 EU/mL
Sterility (post release)	Sterile	Sterile	Sterile	Sterile

^*RRT = relative retention time (retention time of [¹¹C]K2]/[retention time of [¹²C]K2 reference standard); † EU = endotoxin units.

4. Conclusions

A fully automated cGMP method for the radiosynthesis of [¹¹C]K2 (2) has been developed. The synthesis utilizes class 3 solvent EtOH in both the radiosynthesis and purification steps. This reduces the time needed for QC testing since there is no need to check for residual solvents using GC on a daily basis which, in turn, reduces losses in radiochemical yield due to radioactive decay. We believe this validated production process will facilitate more widespread use of the AMPA radioligand [¹¹C]K2 (2) at PET Centers around the world.