Copper-Mediated Radiofluorination of Arylstannanes with [18F]KF.

A copper-mediated nucleophilic radiofluorination of aryl- and vinylstannanes with [18F]KF is described. This method is fast, uses commercially available reagents, and is compatible with both electron-rich and electron-deficient arene substrates. This method has been applied to the manual synthesis of a variety of clinically relevant radiotracers including protected [18F]F-phenylalanine and [18F]F-DOPA. In addition, an automated synthesis of [18F]MPPF is demonstrated that delivers a clinically validated dose of 200 ± 20 mCi with a high specific activity of 2400 ± 900 Ci/mmol.

Positron emission tomography (PET) imaging is a noninvasive diagnostic imaging technique that provides in vivo physiochemical information.[1−4] Fluorine-18 (18F) is the most commonly used radionuclide for PET, largely due to its attractive half-life of 110 min. This half-life enables the sophisticated synthesis of target molecules, allows distribution of 18F-labeled tracers to facilities that lack cyclotrons, and facilitates in vivo pharmacokinetic studies.[2] With the increasing number of PET scans, there is a need for new, practical methods for the late-stage introduction of 18F into bioactive molecules. Methods that enable the nucleophilic radiofluorination of readily accessible and stable precursors are in particularly high demand due to the availability of [18F]fluoride from small medical cyclotrons.[5,6]

Arylstannanes are appealing precursors for PET radiolabeling for a number of reasons. First, they can be easily prepared from inexpensive starting materials,[7,8] and important examples are already known intermediates and/or commercially available (i.e., TriBoc-l-DOPA methyl ester).[9] Second, the Sn–C bond is stable to most functional group manipulations,[10,11] which enables flexible and modular syntheses of complex precursors that can then undergo selective late-stage radiofluorination. Third, arylstannanes have already been validated as precursors to radiopharmaceuticals for human clinical trials,[12−15] thereby mitigating concerns about toxicity and the feasibility of Sn removal. However, to date, the major limitation in this field is that existing methods for the fluorination[16,17] and radiofluorination[18,19] of arylstannanes require electrophilic fluorine sources (e.g., F2, N-fluoropyridinium salts, N-fluorobenzenesulfonimide, Selectfluor) (Scheme a).[20] These reagents are much more expensive and less available than fluoride (particularly in the 18F form). Furthermore, electrophilic radiofluorination methods result in products with dramatically lower specific activity (ratio of 18F/19F), which greatly reduces imaging sensitivity.[3]

Scheme 1

Radiofluorination of Arylstannanes

We sought to address this limitation through the development of a method for the Cu-mediated nucleophilic radiofluorination of arylstannanes (Scheme b). This work builds on recent reports from our group[6e,21] and others[6c,6f] demonstrating a related Cu-mediated (radio)fluorination of arylboron precursors. In addition to the generally attractive features of arylstannane precursors discussed above, we anticipated that they would be particularly well suited for Cu-mediated radiofluorination. Specifically, a key fundamental step of the Cu-mediated process, the transmetalation of the aryl group to Cu, is expected to proceed significantly faster from tin versus boron, which should result in faster reaction rates (crucial for radiolabeling applications) as well as fewer side reactions and higher yields.[22,23] We report herein the development, optimization, and scope of the Cu-mediated radiofluorination of arylstannanes with [18F]KF. This transformation exhibits high functional group tolerance and has been applied to the late-stage radiofluorination of a number of complex molecules. Furthermore, it has been scaled to an automated synthesis module and used to prepare a clinically validated high specific activity dose of the radiotracer 2′-methoxyphenyl-(N-2′-pyridinyl)-p-18F-fluorobenzamidoethylpiperazine ([18F]MPPF).

Our initial investigations focused on the Cu(OTf)2-mediated fluorination of 1-SnBu with KF under conditions analogous to those demonstrated for aryltrifluoroborate substrates.[21] After 18 h at 60 °C with 4 equiv of KF in CH3CN, the fluorinated product 1 was obtained in 42% yield as determined by 19F NMR spectroscopy (Table , entry 1). While this yield is lower than that reported with 1-BFK under closely analogous conditions (70%, 20 h), the addition of 18-crown-6 as a phase-transfer catalyst could be used to boost the yield with 1-SnBu to 55% (entry 2). Furthermore, as predicted, the fluorination of 1-SnBu is significantly faster than that of 1-BFK. For example, 1-SnBu affords 51% yield after just 15 min under these conditions. In contrast, the analogous reaction of 1-BFK requires more than 2 h to afford 50% yield. The faster rate with 1-SnBu is highly desirable for PET applications. We also examined the sensitivity of this reaction to the stoichiometry of fluoride. This is another key consideration for translation because [18F]fluoride is typically the limiting reagent during radiofluorination. We were encouraged to see that this reaction still proceeds (albeit in reduced yield) with KF as the limiting reagent (Table , entry 6).

Table 1

Cu-Mediated Nucleophilic Fluorination of Arylstannanes in Acetonitrile with Excess Fluoridea

entry	R	R1	time (h)	additive	KF (equiv)	yield (%)
1	Bu	F	18	none	4	42
2	Bu	F	18	18-crown-6	4	55
3	Bu	F	2	none	4	42
4	Bu	F	2	18-crown-6	4	53
5	Bu	F	0.25	18-crown-6	4	51
6	Bu	F	0.25	18-crown-6	0.5	15b
7	Bu	MeO	0.25	18-crown-6	4	23
8	Me	MeO	0.25	18-crown-6	4	24
9	Bu	Ph	0.25	18-crown-6	4	34
10	Me	Ph	0.25	18-crown-6	4	64
11	Bu	Ac	0.25	18-crown-6	4	57
12	Me	Ac	0.25	18-crown-6	4	65

General conditions: Arylstannane (0.025 mmol, 1 equiv), Cu(OTf)2 (4 equiv), KF, 18-crown-6 (4 equiv), CH3CN (0.083 M), 60 °C. Yield determined by 19F NMR spectroscopic analysis of the crude reaction mixture using 1,2-difluorobenzene as an internal standard.

Yield based on KF.

We next applied this fluorination to a small set of arylstannanes (Table , entries 7–12) to establish the effectiveness of our 15 min nucleophilic fluorination protocol for electronically diverse arylstannanes and examine the impact of the alkyl substituents on tin (R) on the reaction. As shown in Table (entries 7–12), stannanes bearing electron-donating (p-MeO), electron-neutral (p-Ph), and electron-withdrawing (p-Ac) substituents react to form fluorinated products in moderate to good yields within just 15 min at 60 °C. The p-MeO derivative is particularly noteworthy, as the corresponding aryltrifluoroborate is poorly reactive, affording <5% yield under any of the fluorination conditions examined.[21] Substitution of the alkyl group on the stannane had a significant impact on yield over this short reaction time, with the Me-substituted stannanes affording comparable or higher yield than the Bu derivatives in all cases. This likely reflects a faster rate of transmetalation from the less hindered tin center.[24]

We next focused on translating this nucleophilic fluorination to achieve the 18F-fluorination of 3-SnBu. As is common in the radiofluorination field,[3,25] significant reoptimization was required, as the best cold fluorination conditions afforded no detectable 18F-labeled product (Table , entry 1). This is likely a consequence of the dramatic change in fluoride stoichiometry (from 0.3 M with KF to approximately 1 nM with [18F]KF).[26] Three modifications were found to be critical for achieving radiofluorination in this system: (1) changing from CH3CN to an amide-based solvent (DMF or DMA); (2) the addition of pyridine (which was also an important additive in related radiofluorinations of arylborons);[6e] and (3) increasing the reaction temperature to between 110 and 140 °C. The optimal radiofluorination conditions for 3-SnBu were as follows: 2 equiv of Cu(OTf)2, 15 equiv of pyridine in 0.1 M DMA at 140 °C. This afforded [F]3 in 65 ± 2% radiochemical conversion (RCC) within just 5 min (entry 8).

Table 2

Cu-Mediated Nucleophilic Fluorination of 3-SnBu3 with [18F]KFa

entry	solvent	temp (°C)	pyridine (equiv)	RCCb (%)
1	CH3CN	60	0	nd
2	CH3CN	60	50	nd
3	DMF	60	50	nd
4	DMF	110	50	22 ± 2
5	DMA	110	50	51 ± 1
6	DMA	110	15	53 ± 1
7	DMA	140	15	55 ± 10
8c	DMA	140	15	65 ± 2
9d	DMA	140	15	44 ± 1

General conditions: 3-SnBu (0.01 mmol, 1 equiv), Cu(OTf)2 (2 equiv), pyridine (15 equiv), [18F]KF, DMA (0.01 M), 140 °C, 30 min. RCC determined by radio-TLC (n ≥ 2).

nd = no product detected by radio-TLC or HPLC.

Reaction time = 5 min.

3-SnMe used as substrate.

The optimal conditions were applied to a series of aryl-, heteroaryl-, and vinylstannane precursors. As summarized in Figure , this method is compatible with aromatic substrates bearing electron-neutral (3), electron-withdrawing (4, 13), and electron-donating substituents (2, 5–10) as well as heteroaromatic (11) and vinylstannane (12) derivatives. Substrates such as 13-SnBu include a functional handle that can be used for further elaboration. Ortho-substitution was well-tolerated, with the o-MeO substrate 6-SnBu affording a yield comparable to that of the p-MeO substrate 2-SnBu (RCC = 57% versus 48%, respectively). This is in contrast to other metal-mediated nucleophilic radiofluorinations, which typically afford much lower yields with ortho-substituted aromatic substrates (see Table S20). The electron-rich substrates (2, 5–10) are noteworthy, as they are challenging to radiofluorinate using traditional SNAr.[1] Overall, the RCCs are comparable, and in many cases significantly higher, than those for the state-of-the-art metal-mediated radiofluorinations of analogous substrates (see Table S20 for a comparison of metal-mediated nucleophilic radiofluorination methods).

Figure 1

Scope of arylstannane substrates. Conditions: aryltributyltin substrate (0.01 mmol, 1 equiv), Cu(OTf)2 (2 equiv), pyridine (15 equiv), [18F]KF, DMA (0.01 M), 140 °C, 5–30 min. RCC determined by radio-TLC; (a) 100 °C; (b) 18-crown-6 (0.5 equiv); (c) 30 equiv of pyridine.

We next applied our method to the preparation of products currently being evaluated as radiotracers in clinical trials or already FDA approved.[27,37] Initial studies were conducted via manual synthesis, which provided useful RCCs (Figure ). For instance, this method is effective for the synthesis of protected phenylalanine derivatives ([F]14 and [F]15) for studying amino acid transport.[28] We also targeted 3-[18F]fluoro-5-[(pyridin-3-yl)ethynyl]benzonitrile ([18F]F-PEB, [F]18), whose previous radiosyntheses suffer from low yields (1–5% RCY, nondecay corrected).[6e,29−31] Our new method, starting from the readily available and stable arylstannane precursor 18-SnBu, delivers the product in 11 ± 2% RCC.

Figure 2

Substrate scope of relevant radiotracers.

This method is also effective for the radiofluorination of the protected l-DOPA stannane 16-SnMe, affording [F]16 in 56 ± 12% yield. This result is noteworthy for several reasons. First, nucleophilic methods for radiolabeling l-DOPA remain highly sought after,[1,28,32,33] and the obtained RCC is among the best reported for this type of transformation (see Figure S10 for a comparison of nucleophilic methods for [18F]F-DOPA derivatives). Second, the precursor 16-SnMe is a single step from a commercial stannane whose derivatives have been used in the clinical production of [18F]F-DOPA via electrophilic radiofluorination.[12,13] As such, the successful radiofluorination of 16-SnMe offers the potential for a direct nucleophilic replacement for this method.

Finally, we targeted [18F]MPPF ([F]17), a serotonin receptor ligand currently synthesized by SNAr radiofluorination of the NO2 precursor (Scheme a). Our manual Cu-mediated procedure from stannane 17-SnMe delivered [F]17 in 33 ± 4% RCC. The [18F]MPPF synthesis was scaled and automated using a TRACERLab FXFN module (Scheme b). The reaction using 1500 mCi of initial activity afforded a formulated and validated 200 ± 20 mCi dose (13% RCY) after radiolabeling and HPLC purification.[34] The dose prepared by this method passed all cGMP quality control testing necessary for clinical use, as outlined in the US Pharmacopeia and 21CFR212, including residual Cu and Sn levels below the allowed limits specific in the ICH Guidelines (see the Supporting Information for complete information).[6g,35] As shown in Scheme , our new method affords nearly double the RCY and reduces the overall time from end of bombardment (EOB) to end of purification by one-third relative to the current commercial synthesis of this tracer.[36]

Scheme 2

Comparison of Current Methods To Synthesize [18F]MPPF ([18F]17)[36,37]

In conclusion, this paper discloses a mild, copper-mediated method for the nucleophilic fluorination and radiofluorination of arylstannanes. This method represents the first practical nucleophilic fluorination of stannanes using 18F, is compatible with aryl, heteroaryl, and vinylstannanes, and fills an important gap in available late-stage fluorination methods. Furthermore, this process can be readily automated and scaled on a commercial radiochemistry synthesis module and applied to clinically relevant radiotracers. Finally, we have shown that the method is tolerant of a reasonable number of functional groups common in drug targets. To expand the utility further, exploration of the compatibility of this method with medicinal chemistry space is ongoing and will be reported in due course.