Synthesis of Diverse ¹¹C-Labelled PET Radiotracers via Direct Incorporation of [¹¹C]CO₂

Abstract

Three new positron emission tomography (PET) radiotracers of interest to our functional neuroimaging and translational oncology programs have been prepared through new developments in [¹¹C]CO₂ fixation chemistry. [¹¹C]QZ (glutaminyl cyclase) was prepared via a tandem trapping of [¹¹C]CO₂/intramolecular cyclization; [¹¹C]tideglusib (glycogen synthase kinase-3) was synthesized through a tandem trapping of [¹¹C]CO₂ followed by an intermolecular cycloaddition between a [¹¹C]isocyanate and an isothiocyanate to form the 1,2,4-thiadiazolidine-3,5-dione core; [¹¹C]ibrutinib (Bruton’s tyrosine kinase) was synthesized through a HATU peptide coupling of an amino precursor with [¹¹C]acrylic acid (generated from [¹¹C]CO₂ fixation with vinylmagnesium bromide). All radiochemical syntheses are fully-automated on commercial radiochemical synthesis modules and provide radiotracers in 1 – 5% radiochemical yield (non-corrected, based upon [¹¹C]CO₂). All three radiotracers have advanced to rodent imaging studies and preliminary PET imaging results are also reported.

TOC image

INTRODUCTION

Positron emission tomography (PET) imaging is a type of functional molecular imaging using probes, known as radiotracers, composed of a bioactive molecule conjugated with a positron-emitting radionuclide.¹ The functional information garnered from a PET scan, when combined with the anatomical information from the co-registered CT or MRI scan, provide unprecedented insight into biochemical pathways, mechanisms of disease pathology, and in vivo behavior of drug molecules. Reflecting this, PET imaging is having far reaching impact on personalized medicine² and drug discovery.³

Commonly used PET radionuclides include carbon-11 (t_1/2 = 20 min), fluorine-18 (t_1/2 = 110 min), and gallium-68 (t_1/2 = 68 min). The choice of radionuclide depends on a number of factors ranging from synthetic considerations about how it will be incorporated into the bioactive molecule of choice, to practical aspects associated with intended application (e.g. the short half-life of ¹¹C allows patients to receive multiple PET scans in a single hospital visit, while the half-life of ¹⁸F facilitates radiotracer distribution from centralized manufacturing facilities).

Carbon-11 is an attractive choice of PET radionuclide because multiple scans can be conducted in series during a single patient visit (e.g. scans with 2 different radiotracers, baseline and intervention studies with 1 tracer). Moreover, it can be frequently incorporated into bioactive or endogenous molecules without any structural modification to the original (non-radioactive) molecule, which may or may not be the case with other radionuclides (e.g. use of radioactive metal ions such as ⁶⁸Ga require decoration of the bioactive molecule with a suitable metal-chelating group prior to radiolabeling). Carbon-11 is produced by a cyclotron, reacting with oxygen added to the cyclotron target gas to generate [¹¹C]CO₂, which is delivered to the radiochemistry laboratory and used to synthesize radiotracers. The short half-life of carbon-11 is advantageous for the reasons outlined above, but it presents challenges. Most notably, the short half-life necessitates that all reactions used to synthesize ¹¹C-radiotracers are reasonably high yielding over a very short time course (e.g. 2–10 min) so that they provide usable amounts of radiotracer, thereby limiting the number of reactions that are practical. Typically, [¹¹C]CO₂ is converted into a secondary synthon such as [¹¹C]CH₃I, [¹¹C]CH₃OTf or [¹¹C]KCN, which is then reacted with a suitable precursor to yield the ¹¹C-labeled compound. Such radiochemical reactions have been used to great effect to synthesize ¹¹C-radiotracers (for recent reviews of carbon-11 radiochemistry, see:^4,5,6). However, there are limitations in the types of radiotracers that can be accessed from such synthons. For example, there must be a place to introduce a methyl group if [¹¹C]CH₃I or [¹¹C]CH₃OTf are to be used for labeling. Given the prevalence of carbonyl groups in bioactive molecules (e.g. many of the best-selling drugs contain one or more C=O bonds⁷), there is significant interest in developing methods that enable incorporation of a ¹¹C-carbonyl unit into bioactive molecules to increase the number and diversity of available PET radiotracers. One such approach involves synthesis of PET radiotracers directly from [¹¹C]CO₂.

The electrophilic carbon in [¹¹C]CO₂ means it can be used as a carbonyl source, and can be trapped by an appropriate nucleophilic component. For example, this approach can be used to synthesize radiolabeled carboxylic acids, such as [¹¹C]acetate and [¹¹C]palmitate, by reacting [¹¹C]CO₂ with an appropriate Grignard reagent.⁸ New advances in the synthesis of [¹¹C]carboxylic acids involve treating organoboron precursors with [¹¹C]CO₂ in the presence of a copper catalyst.^9,10 More recently, there has also been a surge in [¹¹C]CO₂ fixation chemistry (for a review of current developments, see:¹¹). For example, [¹¹C]CO₂ fixation chemistry has recently been employed in the synthesis of [¹¹C]ureas (both symmetrical¹² and unsymmetrical^{13,14,15,16,17}) and [¹¹C]carbamates.^{14,17,18,19,20,21} In an interesting variant of the latter, Miller also demonstrated that analogous reactions with [¹¹C]CS₂ can be employed to generate [¹¹C]dithiocarbamates.²²

These impressive new developments in [¹¹C]CO₂ fixation chemistry were of particular interest to us because they have greatly opened up the synthetic transformations possible with carbon-11, and we believed that we could now employ [¹¹C]CO₂ fixation to synthesize three radiotracers of interest to our neuroimaging and translational oncology programs that would be extremely challenging to prepare by other means (Figure 1). From a neuroimaging perspective, we were interested in accessing [¹¹C]3-(3-(1H-imidazol-1-yl)propyl)quinazoline-2,4(1H,3H)-dione ([¹¹C]QZ, 1) and [¹¹C]tideglusib (2) as potential radiotracers for glutaminyl cyclase (QC) and glycogen synthase kinase-3 (GSK-3), respectively.^23,24 In our growing translational oncology program, we were also interested in a method for synthesizing [¹¹C]ibrutinib (3), a radiolabeled version of the anticancer drug targeting B-cell malignancies.²⁵ We reasoned that all three radiotracers could be synthesized from labeled building blocks accessible using [¹¹C]CO₂ fixation. Herein we report three new variants of [¹¹C]CO₂ fixation chemistry to generate intermediates that can be used in intramolecular cyclization-, intermolecular cycloaddition- and peptide coupling-reactions to generate [¹¹C]QZ, [¹¹C]tideglusib and [¹¹C]ibrutinib, respectively.

Figure 1.

New radiotracers labeled by [¹¹C]CO₂ fixation (* = ¹¹C)

RESULTS AND DISCUSSION

Automation and Specific Activity Considerations

We initially wished to have a means of automating [¹¹C]CO₂ fixation chemistry using our General Electric (GE) TRACERLab FX_C-Pro systems and made modifications to the synthesis modules accordingly (see ESI for detailed description). Briefly, because of the harsh, often sparingly soluble, reagents used in [¹¹C]CO₂ fixation chemistry, we initially replaced the reagent reservoirs of the synthesis module with disposable v-vials to enable handling of air sensitive reagents and to prevent cross contamination between syntheses. Moreover, valves were modified or installed to allow delivery of [¹¹C]CO₂ directly to the TRACERLab reaction vessel. [¹¹C]CO₂ was produced with a GE PETTrace cyclotron via the ¹⁴N(p,α)¹¹C reaction. Carbon-11 produced in this manner has very high specific activity (for example, previously reported specific activities for ¹¹C-labeled tracers produced in our laboratory range from 1 – 40 Ci/μmol²⁶). While high specific activities are required for potent radiotracers such as [¹¹C]carfentanil (low specific activity doses could elicit unwanted pharmacological responses), in our hands very high specific activity ¹¹C has proven problematic for [¹¹C]CO₂ fixation chemistry. For example, we have previously reported the need to include carrier [¹²C]CO₂ when producing [¹¹C]acetate to avoid contaminating doses with [¹¹C]acetone and [¹¹C]tert-butanol resulting from over alkylation of [¹¹C]CO₂ with the Grignard reagent.⁸ No-carrier-added (NCA) reactions proved to be an issue in the present work too (vide infra). For example, while [¹¹C]ibrutinib could be prepared carrier-free, [¹¹C]tideglusib could not. The synthesis of [¹¹C]ibrutinib was also notably cleaner and higher yielding when carrier [¹²C]CO₂ was added. Our standard protocol is to introduce 0.5 – 1.0 mL (~22.5 – 45 μmoles) of [¹²C]CO₂ to the molecular sieves prior to delivery of [¹¹C]CO₂ from the cyclotron.

[¹¹C]QZ (1)

Increased levels of glutaminyl cyclase (QC) in AD are thought to result in the formation of pyroglutamate (pGlu)-modified Aβ peptides. These derivatives of Aβ are suspected to be more neurotoxic and resistant to clearance than regular amyloid-β, and they are prion-like, functioning as a template for misfolding and as a nucleus for aggregation with unmodified amyloid-β. In a post-mortem study of human neocortical brain samples it was found AD patients had upregulated levels of QC mRNA compared to patients without dementia, while immunohistochemistry suggests 20–30% higher expression of QC in all cortical layers of AD patients compared to controls.^27,28 We postulate that upregulated QC could therefore be a prodromal marker for AD and were extremely interested to determine if this upregulation could be detected using PET. Our initial efforts to develop a PET radiotracer for QC concentrated on [¹¹C]PBD150.²⁹ However, despite strong literature precedent for the compound (PBD150 is the only QC inhibitor to date that has been shown to reduce the formation of Aβ aggregates in vitro and in vivo^27,30), we were disappointed to discover that [¹¹C]PBD150 does not cross the blood-brain barrier in rodents. Given the importance of QC as both a therapeutic and diagnostic target in AD, our efforts to develop a radiotracer for QC continue. It is unclear why [¹¹C]PBD150 did not enter the CNS. P-glycoprotein transporter involvement was ruled out,²⁹ but one potential explanation is the compound is too hydrophilic to cross the BBB (CLogP = 0.9). Reflecting this, we were interested in next evaluating QC radiotracers that were more lipophilic. To this end, we selected [¹¹C]QZ (1) as a 2^nd generation QC radiotracer for further evaluation (CLogP = 1.5). QZ is a closely related desmethyl analog of a recently reported QC inhibitor (IC₅₀ = 123 nM), Figure 2.²³

Figure 2.

QC Inhibitors

Our strategy for accessing [¹¹C]QZ (1) was through an intramolecular [¹¹C]CO₂ fixation reaction with precursor 5 (Scheme 1). Precursor 5 was synthesized in 39% yield from isatoic anhydride (4) by coupling with 1-(3-aminopropyl)imidazole. Treatment with [¹¹C]CO₂/[¹²C]CO₂ and BEMP, followed by POCl₃, generated carrier-added [¹¹C]QZ (CA-[¹¹C]QZ, 1). Purification by semi-preparative HPLC provided 2 mCi of the isolated and formulated radiotracer (non-decay corrected radiochemical yield (RCY) based upon [¹¹C]CO₂ = 1%, radiochemical purity (RCP) >99% and specific activity = 3 Ci/mmol).

Scheme 1.

Synthesis of [¹¹C]QZ (1)

A preliminary rodent microPET imaging study was conducted with CA-[¹¹C]QZ (Figure 3). CA-[¹¹C]QZ was administered to a female Sprague Dawley rat and dynamic PET scans were acquired for 60 min post-injection of the radiotracer (n = 1). Unfortunately, like [¹¹C]PBD150 previously,²⁹ analysis of summed frame images showed no brain uptake and early frames showed no first pass uptake in the brain. Given these findings, at this time we do not recommend the imidazolyl propyl urea or thiourea scaffolds for further development of a radiotracer for quantifying brain QC activity without further work to address brain permeability of the compounds.

Figure 3.

Summed sagittal rat PET image with [¹¹C]QZ (1) (0 – 60 min after tail-vein injection of the radiotracer)

[¹¹C]Tideglusib (2)

Glycogen synthase kinase-3β (GSK-3β) dysregulation is hypothesized to play an important role in a number of disease states, and inhibitors of GSK-3 are being explored as potential therapeutics.³¹ Notably, increased levels of the active form of the GSK-3β isoform are linked to hyperphosphorylation and subsequent toxic aggregation of tau protein in Alzheimer’s disease and related tauopathies.^32,33 In connection with our program developing radiotracers for PET imaging of GSK-3β,^34,35 we were interested in radiolabeling tideglusib, an irreversible inhibitor of GSK-3β that was shown to reduce tau hyperphosphorylation in murine models and has recently been used in phase 2 clinical trials for treatment of progressive supranuclear palsy (PSP)³⁶ and AD.³⁷ Radiolabeled tideglusib could also be used as a companion diagnostic during development of the therapeutic allowing, for example, study of brain uptake, target engagement, receptor occupancy and biodistribution, as well as potentially serving as a high affinity GSK-3β probe for future research studies.

We investigated the synthesis of [¹¹C]tideglusib (2) from naphthalen-1-amine (6), Scheme 2. We hypothesized that carbon-11 fixation chemistry could be used to generate [¹¹C]naphthylisocyanate 7, and that the 1,2,4-thiadiazolidine-3,5-dione core of [¹¹C]tideglusib could then be accessed through an intermolecular cycloaddition between [¹¹C]isocyanate 7 and (isothiocyanatomethyl)benzene (8), analogous to reported cold syntheses of such compounds.³⁸ Initially treating 6 with carrier-added [¹¹C]CO₂/[¹²C]CO₂ and BEMP followed by POCl₃ generated CA-[¹¹C]naphthylisocyanate 7. Gratifyingly, exposing 7 to (isothiocyanatomethyl)benzene (8) in the presence of additional POCl₃, followed by hydrolysis with water/atmospheric oxygen, promoted cyclization to CA-[¹¹C]tideglusib (2) as expected. However, yields using this method were poor (0.2% non-decay corrected RCY based upon [¹¹C]CO₂), likely due to inefficient chlorination of 8 with POCl₃ and, consequently, a poor cyclization yield. Considering alternative chlorinating agents identified in the literature,³⁹ we investigated whether N-chlorosuccinimide would be more effective at promoting the cyclization. This proved to be the case, and CA-[¹¹C]tideglusib (2) was synthesized in 2% non-decay corrected RCY based upon [¹¹C]CO₂. Attempts to synthesize non-carrier-added [¹¹C]tideglusib (NCA-[¹¹C]tideglusib) from [¹¹C]CO₂ were unsuccessful, possibly because of competing side reactions of NCA-[¹¹C]7, and studies to address this will be conducted in due course. We therefore focused upon automating the procedure using CA-[¹¹C]CO₂ on a GE TRACERLab FX_C-Pro synthesis module (see SI for further details), which included purification by semi-preparative HPLC. However, initial attempts at automation employing a 60% acetonitrile buffered eluent for semi-preparative HPLC purification led to rapid decomposition of 2 and a RCP of only 25% at 10 min post-end-of-synthesis (EOS). Additional investigation of this issue revealed that decomposition was likely occurring due to radiolysis and not chemical instability of 2, as dilute samples (0.0001mg/mL) of non-radioactive reference standard did not show decomposition in this eluent system after 48 hours. The addition of ascorbic acid did not retard the decomposition of CA-[¹¹C]tideglusib, so the eluent composition was switched from acetonitrile to ethanol. A 70% ethanol buffered eluent led to stabilization of CA-[¹¹C]tideglusib, and 1 – 2 mCi were obtained following semi-preparative purification. The isolated product, formulated in 70% ethanolic HPLC mobile phase, exhibited stability over time with no decrease in RCP after ~30 min. Reformulation of a 550 μCi sample provided 150 μCi of CA-[¹¹C]tideglusib in 10% ethanol in USP saline suitable for preclinical use (>99% RCP, specific activity: 37 Ci/mmol). No change in RCP of the reformulated sample was observed over 60 min, indicating radiolytic stability of CA-[¹¹C]tideglusib in the injectable formulation at these levels of radioactivity.

Scheme 2.

Synthesis of [¹¹C]Tideglusib (2)

CA-[¹¹C]Tideglusib was advanced to a microPET imaging study in a female Sprague Dawley rat. CNS permeability was observed with peak brain uptake of 200 nCi/cc, corresponding to a peak SUV of 0.7, after 3 minutes, followed by washout (Figure 4). In the whole brain time-radioactivity curve (Figure 4B), there is an apparent increase in SUV during the last two time points. We see this frequently in our data analysis, independent of radiotracer, and attribute it to increased noise resulting from the low residual counts that remain in the brain during the late frames of the PET scans. The imaging results are somewhat inconsistent with reported pharmacokinetic data, which suggests that tideglusib is an irreversible inhibitor of GSK-3β (K_i= 60nM in vitro, [³⁵S]tideglusib/GSK-3β binding study), albeit with a time-dependent inhibition profile (IC₅₀ = 5 nM with 60 min pre-incubation, 105 nM without pre-incubation in a double ATP and inhibitor study).⁴⁰ Thus, it was expected that brain uptake should increase, or at least not decrease, over time. It is uncertain whether washout was due to A) the low specific activity of the radiotracer resulting in GSK-3β saturation, B) rapid metabolism to a non-GSK-3β active agent, or C) that tideglusib PK in vivo differ from PK in vitro for another reason. As such, studies are currently underway to improve RCY and specific activity of [¹¹C]tideglusib, so that a better understanding of the in vivo characteristics of this clinically-relevant pharmaceutical agent can be realized.

Figure 4.

A: Summed sagittal rat PET image with CA-[¹¹C]tideglusib (2) (0 – 60 min after tail-vein injection of the radiotracer); B: whole brain time-radioactivity curve.

[¹¹C]Ibrutinib (3)

Bruton’s tyrosine kinase (BTK) is a mediator of the B-cell-receptor signaling pathway implicated in the pathogenesis of B-cell cancers. Early-stage clinical trials found ibrutinib, a BTK inhibitor, showed antitumor activity in several types of B cell lymphoma (for a review of the clinical use of ibrutinib, see:²⁵), and is approved by the U.S. Food and Drug Administration (FDA) for the treatment of mantle cell lymphoma and chronic lymphocytic leukemia. Radiolabeled ibrutinib has the potential to function as a companion diagnostic for predicting response to ibrutinib therapy (using, for example, comparable approaches to those recently employed for [¹¹C]docetaxel⁴¹), as well as to support future drug discovery efforts around the BTK pathway.

Synthesis of [¹¹C]ibrutinib (Scheme 3) commenced from vinylmagnesium bromide (8). The Grignard reagent was treated with [¹¹C]CO₂ directly. In our first approach, we then generated [¹¹C]acryloyl chloride 10 by treatment with phthaloyl dichloride, per a literature procedure.⁴² We next attempted to distill 10 into a solution of precursor 11 and 2,6-di-tert-butylpyridine (DTBP), and were encouraged to generate [¹¹C]ibrutinib (3), albeit in low yields (15 mCi, ~0.5% non-decay corrected RCY based upon [¹¹C]CO₂). Analysis of reaction mixtures suggested that ~90% of the [¹¹C]acryloyl chloride polymerized during distillation. To avoid this polymerization issue at elevated temperatures, we next used oxalyl chloride to generate 10, and then transferred the entire [¹¹C]acryloyl chloride reaction mixture into a vial containing amino precursor 11 and DTBP at room temperature. We manually produced injectable doses of 30mCi (~1% non-decay corrected RCY, based upon [¹¹C]CO₂) with >90% RCP, but had a difficult time adapting this method to the automated radiochemistry synthesis modules. Both the Grignard reagent and oxalyl chloride are very reactive and are sensitive to air, or can react with each other. In addition, any excess of either reagent would negatively impact the next step of the reaction, so quantitative quenching was required for each step. Although the synthetic method works, we anticipated that these issues would make the method practically unreliable and complicate routine automated production.

Scheme 3.

Approaches to [¹¹C]Ibrutinib (3)

To address these issues, we next explored the possibility of generating [¹¹C]acrylic acid 12 and then using it in a peptide coupling with precursor 11 (Scheme 3). Quenching the carboxylate, resulting from treatment of vinylmagnesium bromide with [¹¹C]CO₂ (diluted with 0.5 mL of ¹²CO₂), using HCl efficiently generated CA-[¹¹C]acrylic acid (12). We selected HATU as the peptide-coupling reagent, and efficient coupling of precursor 11 with CA-[¹¹C]acrylic acid 12 generated CA-[¹¹C]ibrutinib (3). Purification by semi-preparative HPLC was followed by reconstitution into ethanolic saline using a C18 Sep-Pak cartridge and gave 150 mCi of isolated and formulated product (5% non-decay corrected RCY based upon [¹¹C]CO₂, RCP >99% and specific activity ~55 mCi/μmol). Total synthesis time was 70 minutes from end of beam and the procedure was fully automated on a GE TRACERLab FX_C-Pro synthesis module. We also tried this method no-carrier added, obtaining similar amounts (60 – 150 mCi) of isolated and formulated NCA-[¹¹C]ibrutinib (2 – 5% non-decay corrected RCY based upon [¹¹C]CO₂) but with higher specific activities (600 – 2100 mCi/μmol). Notably however, RCP (>90%) and chemical purity were lower for the NCA product (see SI for further details of both CA and NCA syntheses).

In an initial baseline PET study in a female CD-1 mouse, NCA-[¹¹C]ibrutinib showed significant uptake and retention in spleen, liver and intestines (Figure 5), consistent with the known distribution and excretion of ibrutinib.⁴³ Further investigations of the biodistribution and metabolism of [¹¹C]ibrutinib using normal and tumor bearing rodents, as well as nonhuman primates, are underway and will be reported in due course.

Figure 5.

Summed coronal mouse PET image with [¹¹C]ibrutinib (3) (0 – 60 min after tail-vein injection of the radiotracer)

In summary we have demonstrated the synthesis of complex PET radiotracers using different building blocks all accessed by [¹¹C]CO₂ fixation chemistry. This work demonstrates that the ability to rapidly trap [¹¹C]CO₂ and incorporate it into building blocks using a number of different approaches allows quite sophisticated downstream chemistry to take place in spite of the 20 min half-life of carbon-11. For example, we demonstrate that intramolecular cyclizations, intermolecular cycloaddition reactions and peptide coupling reactions are all possible. Each process is fully automated through new modifications to a standard radiochemical synthesis module, and the radiochemical syntheses provide isolated and formulated doses of radiotracers, which are appropriate for pre-clinical in vivo imaging in rodents and non-human primates. To date, radiochemical yields are somewhat low but reflect the challenging chemistry being conducted with short-lived carbon-11. Notably, yields could be improved by adding carrier [¹²C]CO₂ to the reactions rather than using NCA-[¹¹C]CO₂, which we attribute to competing side reactions that become significant when working with NCA carbon-11. This is a current limitation of using this new chemistry with very high specific activity carbon-11, and should be taken into consideration when considering [¹¹C]CO₂ fixation approaches for radiolabeling. Exploring the scope of [¹¹C]CO₂ fixation reactions is ongoing in our laboratory, including efforts to determine the minimal amount of [¹²C]CO₂ that can be added to minimize side reactions, and maximize specific activities of the resulting radiotracers. Nevertheless, the new reactions reported herein have enabled access to 3 new radiotracers inaccessible by other carbon-11 radiochemistry in amounts suitable for initial pre-clinical evaluation. [¹¹C]QZ does not cross the BBB and so is unsuitable for imaging glutaminyl cyclase in the CNS. Contrastingly, initial results with [¹¹C]tideglusib and [¹¹C]ibrutinib are promising and suggest that additional research with both radiotracers is warranted.

EXPERIMENTAL PROCEDURES

Chemistry

General considerations

All solvents and reagents were commercially available and used without further purification unless otherwise stated. NMR spectra were recorded with a Varian 400 MHz instrument at room temperature with tetramethylsilane (TMS) as an internal standard. Mass spectra were performed on an Agilent Q-TOF HPLC-MS or VG (Micromass) 70-250-S Magnetic sector mass spectrometer employing the electrospray ionization (ESI) method or electron ionization (EI) method. High performance liquid chromatography (HPLC) was performed using a Shimadzu LC-2010A HT system equipped with a Bioscan B-FC-1000 radiation detector. All procedures including anhydrous solvents were performed with rigorously dried glassware under inert atmosphere.

Precursors

Precursors 6 and 8 were commercially available (Aldrich); precursor 11 was obtained from Bristol-Myers Squibb and used as received; precursor 5 was synthesized as described below.

N-(3-(1H-imidazol-1-yl)propyl)-2-aminobenzamide (5)

Isatoic anhydride 4 (0.300 g, 2.4 mmol) was dissolved in THF (14.4 mL, 0.17 M) and 1-(3-aminopropyl)imidazole (0.286 mL, 2.4 mmol) was added. The solution was stirred overnight at room temperature, at which point solvent was removed under vacuum. The resulting product was extracted with DCM and the organic layer was washed with 5 mL saturated sodium bicarbonate solution, 2× with water, dried (MgSO₄), and concentrated in vacuo. The product was purified by silica gel chromatography (DCM/MeOH gradient). After purification, product 5 was collected as an oil (0.23 g, 39% yield); R_f 0.17 (9:1 DCM:MeOH); ¹H NMR (400 MHz d4-MeOD): 7.68 (1H, s), 7.40 (1H, d, J=7.8), 7.17-7.14 (2H, m), 6.95 (1H, br), 6.72 (1H, d, J=8.1), 6.60 (1H, t, J=7.5), 4.08 (2H, t, J=7.0), 3.31 (2H, t, J=6.8), 2.05 (2H, quin, J=6.9); HRMS: calculated for C₁₃H₁₆N₄O: 244.1397, found: 245.1394 ([M+H]⁺).

Reference Standards

The identity of all ¹¹C-labeled radiotracers was confirmed by co-injection onto HPLC with the corresponding unlabeled reference standard. Details on commercial suppliers and/or synthesis of reference standards is described in the supporting information (SI).

Radiochemistry

General Considerations

Standard reagents and solvents were commercially available and used without further purification, unless otherwise noted: sodium chloride, 0.9% USP and sterile water for Injection, USP were purchased from Hospira; Dehydrated Alcohol for Injection, USP was obtained from Akorn Inc.; Ascorbic Acid for Injection, USP was acquired from Bioniche Pharma; molecular sieves were purchased from Alltech; and HPLC columns were acquired from Phenomenex. Other synthesis components were obtained as follows: sterile filters were acquired from Millipore; C18 Sep-Paks were purchased from Waters Corporation; 10 cc sterile vials were obtained from HollisterStier. Sep-Paks were flushed with 10 mL of ethanol followed by 10 mL of sterile water prior to use. [¹¹C]CO₂ was produced with a General Electric (GE) PETTrace cyclotron. High purity nitrogen containing 0.5% O₂ was irradiated with 16.5 MeV to generate [¹¹C]CO₂, which was delivered to a GE TRACERlab FX_C-Pro synthesis module via stainless steel lines. Radiochemical syntheses were carried out using GE TRACERLab FX_C-Pro synthesis modules housed in Comecer hot-cells. Doses produced were assessed via standard quality control techniques (see below) and were appropriate for rodent studies.

[¹¹C]QZ (1)

A solution of precursor 5 (4 mg) and BEMP (5 μL) dissolved in 0.1 mL acetonitrile was sparged for 4 min with [¹¹C]CO₂ (300 mCi) that had been diluted with 1 mL of ¹²CO₂ at 7 mL/min (Ar used for transfer). A solution of POCl₃ (0.4 μL in 0.1 mL of acetonitrile) was added and the resulting mixture was stirred for 4 minutes. Then 1 mL of HPLC mobile phase was added and used to transfer the reaction mixture to the HPLC loop. The product was purified by reverse phase HPLC (column: Luna C18 semi-prep, 150 × 10 mm, 10 μ; mobile phase: 25% EtOH, 10 mM NaH₂PO₄, pH 8.0, 0.50 g/L of L-ascorbic acid; flow rate: 4 mL/min) and the product peak (t_R ~ 9 – 10 min) was collected and diluted with USP saline to achieve a 10% ethanol solution suitable for injection. Yield was 2 mCi (1% non-corrected isolated and formulated RCY based upon [¹¹C]CO₂; RCP = 100%; pH = 6.5).

[¹¹C]Tideglusib (2)

A solution of naphthalen-1-amine (6) (0.72 mg, 5 μmol) and BEMP (5 μL, 17.5 μmol) dissolved in 0.2 mL acetonitrile was sparged for 6 min with [¹¹C]CO₂ (800 mCi) that had been diluted with 0.5 mL of ¹²CO₂ at 5 mL/min (He used for transfer). After this time the reaction mixture was transferred to a vial containing POCl₃ (10 μL, 100 μmol) in 100 μL THF (inhibitor-free) and allowed to react for 2 min to generate [¹¹C]7. The solution was then transferred to a vial containing benzylisothiocyanate (8) (3.7 μL, 25 μmol) and N-chlorosuccinimide (13.4 mg, 100 μmol) in 200 μL THF (inhibitor-free) with 20 mL air and allowed to react for 5 min. The crude reaction mixture was diluted with 600 μL of 1:1 ethanol: water and transferred (using 60 mL air) into an 8 mL sterile vial. Analysis of the reaction mixture revealed 14.6 mCi of [¹¹C]titdeglusib, which equated to 2% non-decay corrected RCY based upon [¹¹C]CO₂. Total synthesis time was 60 minutes from end of beam. Efforts to purify and reformulate [¹¹C]tideglusib were complicated by radioloysis, but were eventually accomplished using ethanolic solvent systems (see SI for detailed information).

[¹¹C]Ibrutinib (3)

[¹¹CO₂] (~3 Ci diluted with 0.5 mL of ¹²CO₂) was bubbled into vinylmagnesium bromide (200 μL of 1.0 M Grignard diluted with 400 μL of THF) at 7 mL/min flow rate for 4 minutes. 250 μL of HCl solution (100 μL of 4.0 M HCl/dioxane diluted with 800 μL of DMF) was then added and the reaction stirred for an additional 1 min. A solution of precursor 11 (3.9 mg) and DIPEA (25 μL) in DMF (150 μL), and a solution of (1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate) (HATU, 6.5 mg) in DMF (150 μL) were then added and the reaction mixture was stirred at room temperature for 10 min. The reaction was then diluted with 1 mL of HPLC mobile phase and purified by semi-preparative HPLC (column: Prodigy ODS 250×10 mm; mobile phase: 30 mM NH₄OAc in 45% MeCN/H₂O; flow rate: 4 mL/min). The product peak (t_R ~11 – 12 min) was collected into 50 mL of water, and loaded onto a C18 Sep-Pak (Waters, 1cc). The final dose was eluted by 0.5 mL of ethanol followed by 4.5 mL of injectable saline. Total synthesis time was 70 minutes from end of beam and gave 150 mCi of isolated and formulated product (5% non-decay corrected RCY based upon [¹¹C]CO_2; RCP >99%; specific activity = 55 mCi/μmol; pH = 6.0). No-carrier-added syntheses were also attempted and further details are provided in the SI).

Quality Control

Radiochemical purity of products was assessed using Shimadzu LC-2010A HT system equipped with UV-Vis and radioactivity detectors as follows:

[¹¹C]QZ

Column: Phenomenex Luna C18, 150×4.6 mm; mobile phase: 10 mM NH₄OAc in 15 % MeCN; flow rate: 2.0 mL/min; wavelength: 254 nm; room temperature; product t_R = 9.9 min.

[¹¹C]Tideglusib

Column: Phenomenex Luna phenyl hexyl, 5 micron, 150×4.6mm; mobile phase: 10 mM KH₂PO₄ in 70% EtOH, pH 4.6; flow rate: 1 mL/min; wavelength: 254 nm; product t_R = 7.2 min.

[¹¹C]Ibrutinib

Column: Kinetex XB-C18 150×4.6 mm; mobile phase: 30 mM NH₄OAc in 45% MeCN; flow rate: 1.0 mL/min; wavelength: 254 nm; product t_R = 6.9 min.