PRODUCTION OF ¹¹C-LABELED AMIDES VIA “IN-LOOP” ¹¹C-CARBONYLATION

Abstract

The carbon-11 (¹¹C) labeling of benzamides holds great promise in the development of PET radiotracers given the prevalence of benzamides in bioactive molecules. This work establishes an improved automated route to carbon-11 labeled primary benzamides through carbonylative cross-coupling of aryl halides with carbon-11 carbon monoxide using electrophilic aroyl dimethylaminopyridinium salts as an intermediate. The methodology uses GE TRACERLab FX automated radiochemistry synthesis modules and allows for the synthesis of a variety of benzamide containing compounds, including known PET imaging agent [¹¹C]LY2795050, a KOR antagonist, in moderate to excellent radiochemical yields. This not only provides a practical means of synthesizing primary [¹¹C]benzamides, but it sets the stage for potential future applications in PET tracer synthesis. Overall, this development offers a versatile tool for generating ¹¹C-labeled compounds for PET imaging applications, thus opening new avenues in the field of molecular imaging and advancing our understanding of various biological processes.

Introduction:

Positron emission tomography (PET) imaging is a powerful technique that utilizes radiopharmaceuticals, bioactive molecules tagged with positron-emitting radionuclides.[1,2] Positrons are emitted from the radionuclide and annihilate with an electron to generate two 511 keV gamma photons that are detected by the PET scanner. PET is used for imaging and quantification of biochemical processes, and finds application in diagnosis and monitoring of diseases such as cancer and dementia, as well as to support drug discovery. Of the various radionuclides used for PET imaging, carbon-11 (¹¹C) is particularly well-suited due to its favorable imaging characteristics and the ubiquity of carbon in bioactive molecules.[1] With a relatively short half-life of 20.3 minutes, ¹¹C decays quickly, minimizing radiation exposure to the subject while allowing for multiple scans during a single day visit. The high percentage (99.8%) of decay by positron emission and low energy of the emitted positrons (0.96 MeV) contribute to high-resolution PET imaging, enabling precise visualization of molecular processes in vivo. Furthermore, since carbon is a major constituent of organic molecules (as carbon-12, the most abundant stable isotope), introducing ¹¹C into biologically relevant compounds does not significantly alter their physicochemical or pharmacological properties. Additionally, carbon-13 and carbon-14 labeled isotopologs can be used for hyperpolarized MRI applications and autoradiography/binding affinity studies, respectively. This makes carbon-11 an attractive choice for labeling molecules of interest for PET imaging.[3] By using carbon-11 labeled compounds, researchers can non-invasively study various molecular processes and pathways in living subjects, including humans and animal models, provided there is a means to incorporate carbon-11. This capability has broad applications in biomedical research, drug development, and clinical diagnostics, allowing for a deeper understanding of disease mechanisms and treatment responses at the molecular level.

In the presence of oxygen (approximately 0.5–1%) when irradiating nitrogen-14, a cyclotron produces [11C]CO₂. This compound, while useful as a synthon in a limited number of reactions, is more often further derivatized to create a range of carbon-11-labeled synthons for an even greater variety of ¹¹C-labelelling methods.[4–6] For instance, [11C]CO can be obtained by partially reducing [11C]CO₂ over a surface of molybdenum, zinc, or activated charcoal.[7,8] [11C]CO has proven to be highly useful in medical imaging. Notably, [11C]CO was one of the first tracers used in human blood volume measurements as early as 1968 [9], and was utilized even earlier in 1945 to confirm CO is not metabolized in vivo. [10]

Beyond its direct applications, [11C]CO serves as a precursor for synthesizing a variety of carbonyl-containing compounds through transition metal-mediated 11C-carbonylation processes. [7],[11] The carbonyl group is a prevalent functional group in many drugs and bioactive molecules, making [11C]CO an attractive synthon for preparing carbon-11 labeled radiopharmaceuticals.[12] Despite its versatility and importance, the widespread application of [11C]CO has faced several challenges. Compared to other carbon-11-labeled reagents, such as [11C]CH₃I and [11C]CO₂, [11C]CO has seen less frequent use due to its limited availability and the specialized equipment historically required for its preparation. Additionally, the low reactivity and limited solubility of [11C]CO in organic solvents present significant synthetic challenges. As a result, traditional labeling procedures have often involved complex setups with high-pressure reactors [13] and/or relatively high reaction temperatures.[14]

To address these historical challenges, there have been significant advancements in the field of [11C]CO radiochemistry with the goal of making it more accessible for routine use. The increased availability of modern cyclotrons and the development of automated radiosyntheses are facilitating the production of patient doses with more complicated methodologies. One notable development is the one-step “in-loop” method for carbon-11 carbonylation,[15] which allows for the synthesis of labeled amides at ambient pressure and temperature, addressing previous reactivity challenges.[16] This fully automated carbon-11 carbonyl labeling process has been successfully adapted for use with synthesis modules, enabling cGMP manufacture of radiopharmaceuticals suitable for clinical applications.[15][16][17][18]

To date, one-step approaches have been reported for the synthesis of acids, esters, and primary amides, and while they provide reasonable radiochemical yields, they have seen limited adoption by the wider community, perhaps because they remain manual in nature or use bespoke synthesis modules [17]. In an attempt to make [¹¹C]CO chemistry more accessible and simpify labeling diverse sets of scaffolds, we have investigated conducting [¹¹C]CO chemistry on a commercially available GE TRACERLab synthesis module, and have had reasonable success forming the intermediate carbonylation complex before reacting it with the appropriate nucleophiles. For example, we recently reported a two-step approach involving the initial carbon-11 carbonylation of a (hetero)aryl halide precursor followed by trapping with a nucleophile (HO- or RO-), and expanded the range of radiopharmaceuticals accessible from [11C]CO to include carbon-11 labeled carboxylic acids and esters.[19] A second focus of our carbonylation work has been applying it to bioactive molecules relevant to PET imaging and automating the chemistry using GE TRACERLab synthesis modules. We feel the latter is particularly important as these modules have a large install base (hundreds in the US), thus making [11C]CO accessible to many more sites. The developments from our lab and the other groups working in this space represent promising progress in the automated synthesis and production of carboxylic acid- and ester-based radiopharmaceuticals for clinical use from [11C]CO, taking a step towards routine use and potentially opening doors to new diagnostic (and therapeutic) applications. In marked contrast to acids and esters, labeling of primary amides using [11C]CO has remained somewhat elusive and adapting this reaction so that it can benefit from the latest loop chemistry and automation developments around [11C]CO is the focus of this work.

Primary benzamides play crucial roles in various biologically relevant compounds, including drugs used in cancer and central nervous system (CNS) treatments (Figure 1). We were interested in developing a more convenient method for labeling primary benzamides with carbon-11 for PET imaging. Existing methods for labeling substituted benzamides with carbon-11 have taken precedence over the exploration of labeling primary benzamides as they can be accessed with the use of amine reactants as opposed to ammonia (or an ammonia surrogate). Previous methods for labeling primary benzamides with ¹¹C have relied on transition metal-mediated protocols, often starting from [11C]CN or [11C]CO (Figure 2).[20–27] While these methods have shown success in yielding the target benzamides with high radiochemical yields (RCYs), they come with certain drawbacks. For instance, strategies involving the use of toxic ammonia gas and/or requiring high-pressure micro-autoclave equipment inherently makes such approaches less convenient and potentially hazardous, particularly for non-experts conducting routine radiopharmaceutical production day-to-day. [21–22]

Figure 1.

Benzamide core containing drug like molecules.

Figure 2.

Previous approaches for radiosynthesis for primary benzamides. (A) Cu-mediated radiocyanation [23][24] (B) Pd-mediated radiocyanation [25] (C) Pd-mediated radiocarbonylation with ammonia [26] (D) Pd-mediated radiocarbonylation with formamide [27].

The development of a more convenient method for labeling primary benzamides with carbon-11 is thus of great interest to researchers in the field of PET imaging. Such a method would not only streamline the labeling process but also eliminate the need for hazardous reagents and specialized equipment beyond those commonly found in a radiochemistry facility, thus making the production of primary benzamide PET tracers more accessible and safer for researchers and non-expert staff. Motivated by recent progress in carbonylative formation of aroyl dimethylaminopyridinium (aroyl-DMAP) salts, [27] [28] [29] which simplify acylation reactions, we were inspired to develop a user-friendly approach for synthesizing carbon-11 labeled primary benzamides. By leveraging recent advancements from “in-loop” carbon-11 carbonylation protocols and drawing inspiration from the carbonylative formation of aroyl-DMAP salts, our goal was to devise a method for the synthesis of carbon-11 labeled primary benzamides using a commercially available TRACERLab module “out-of-the-box”, improving accessibility and efficiency of production for PET imaging applications.

Results and Discussion

Based on our prior investigations, we initiated our exploration by employing iodobenzene as a representative substrate for a one-step “in-loop” carbon-11 carbonylation process.[16] However, despite reports of one-step processes for labeling primary benzamides in the literature [27], in our hands such an approach yielded 1-¹¹CONH₂ (Nu = NH₂) in less than 1% radiochemical conversion (RCC, see Supplementary Information for detailed information on RCC determination) (Scheme 1). Reasons for this difference are unclear, but we hypothesize that poisoning of the catalyst by a nitrogen containing nucleophile in our set up may be the cause. As Xantphos ligated aryl palladium complexes are stable, we also evaluated a two-step approach wherein the intermediate [11C]CO complex was prepared in-loop first, and then transferred to the synthesis module reactor pre-charged with the nucleophile, avoiding possible poisoning of our catalyst by prolonged exposure to a nitrogen containing nucleophile. The second reaction was carried out at 100 °C for 10 min. Under this modified protocol, increased radiochemical conversions were observed, providing [¹¹C]-1-CONH₂ in 13% RCC (Scheme 1).

Scheme 1.

One-step and two-step [¹¹C]-carbonylation methods for synthesizing primary amides.

Observing the promising outcome of the two-step process over the one-step process, we pivoted to investigating different conditions to optimize radiochemical conversions of the two-step process. We decided to first investigate different palladium complexes. To begin with, iodobenzene (1-I), xantphos, and Pd₂(π-cinnamyl)Cl₂ in THF was reacted at room temperature over 20 minutes and the resultant mixture was loaded onto an HPLC loop.[16] [11C]CO was prepared using a molybdenum catalyst at 850 °C as previously described,[16] and trapped using a SiO₂ trap (100 mg silica gel in a 1/8 inch tube) immersed in liquid nitrogen. The [11C]CO was released and pushed by helium at 5 ml/min into the HPLC loop containing the precursor catalyst complex for reaction. The loop was sealed for 5 min at room temperature and the reaction mixture was then analyzed using radio-HPLC to determine RCC to 1-I. Optimization of the reaction conditions was initiated by gradually increasing the amount of ammonia solution (0.5 M in THF) ranging from 7.0 to 200 μL (Table 1, entries 1–4), but unfortunately, none of the conditions yielded significant amounts of [¹¹C]-1-CONH₂.

Table 1.

Optimization of two-step conditions for the radiosynthesis of primary amides.

Entry	Pd Catalyst/Ligand	Ammonia source	Observation/RCC
1.	Pd2(π-cinnamyl)Cl2/xantphos/THF/precursor	0.5 M NH3 in THF (7.0 μL)	n.d.a
2.	Pd2(π-cinnamyl)Cl2/xantphos/THF/precursor	0.5 M NH3 in THF (14.3 μL)	n.d.a
3.	Pd2(π-cinnamyl)Cl2/xantphos/THF/precursor	0.5 M NH3 in THF (20 μL)	n.d.a
4.	Pd2(π-cinnamyl)Cl2/xantphos/THF/precursor	0.5 M NH3 in THF (200 μL)	n.d.a
5.	Pd(dba)2/N-xantphos/THF/precursor	0.5 M NH3 in THF (10 μL)	n.d.b
6.	Pd(dba)2/N-xantphos/THF/precursor	LiNH2 (2.0 mg)	n.d.b
7.	Pd(dba)2/N-xantphos/THF/precursor	LiNH2 (2.0 mg)	1% (n = 2)c
8.	Pd(dba)2/N-xantphos/THF/precursor	LiNH2 (2.0 mg)	13%d
9.	Pd(dba)2/N-xantphos/THF/precursor	DABCO/KI/NH2COONH4	10 (n = 2)%e
10.	Pd(dba)2/N-xantphos/THF/precursor	DMAP/formamide	74 (n = 2)%f

^aConditions: Pd₂(π-cinnamyl)Cl₂ (2.5 mg, 4.8 μmol), xantphos (5.0 mg, 8.65 μmol), precursor 1-I (0.8 μL, 7.14 μmol), THF (200 μL);

^bConditions: Pd(dba)₂ (2.0 mg, 3.82 μmol), N-xantphos (2.2 mg, 3.98 μmol), precursor 1-I (0.91 μL, 8.16 μmol), THF (200 μL).

^cHeated at 50 °C for 5 minutes,

^dHeated at 100 °C for 5 minutes,

^eDABCO (6.0 mg, 53.6 μmol), KI (1.0 mg, 6.0 μmol), ammonium carbamate (3.0 mg, 38.5 μmol) and heated at 100 °C for 10 minutes.

^fDMAP (25 mg, 205 μmol) and formamide (100 μL, 2.5 mmol) heated at 100 °C for 5 minutes. n.d. = not detected.

We observed poor solubility of Pd₂(π-cinnamyl)Cl₂ in THF, prompting us to transition to the use of Pd(dba)₂ and N-xantphos for subsequent reactions for improved solubility, again with an eye on simplifying automation. However, we still obtained none of the desired product when using 0.5 M NH₃ in THF as the ammonia source (Table 1, entry 5). This prompted us to consider alternative ammonia sources for the synthesis of primary benzamides, and we initially switched to lithium amide as a readily available alternative. Unfortunately, under room temperature conditions, the desired product was not obtained (Table 1, entry 6). However, when the reaction was conducted under heating at 50 °C for 5 minutes, the desired product was generated with a RCC of 1% (Table 1, entry 7), which increased to 13% under heating at 100 °C for 5 minutes (Table 1, entry 8). Given the ammonia source had an effect on reaction yield, but that lithium amide was not giving usable amounts of product, we considered other alternatives. In 2015, Bhanage et al. used solid ammonium carbamate as an effective ammonia surrogate replacing ammonia for aminocarbonylation of aryl iodides.[30] We decided to see if these conditions would translate for the in-loop two-step carbonylation of aryl halides. Initial complex formation was conducted in loop, and then transferred to the reactor that had been pre-charged with DABCO/KI and NH₂COONH₄ in dry acetonitrile.[30] The reactor was heated at 100 °C for 10 minutes. The crude reaction mixture was analyzed using HPLC and confirmed product in a RCC of 10% (Table 1, entry 9), similar to the results obtained using lithium amide.

In a further effort to improve the method we also investigated conditions from Schou et al. who generated [¹¹C]aroyl dimethylaminopyridinium salts using DMAP and aryl halides to improve intermediate formation en route to labeled benzamides.[27] In their study, the in situ generated salts were reacted with formamide as the ammonia source to generate intermediate formamides which were heated at 100 °C for 5 minutes in a reactor, leading to improved yields of carbon-11 labeled primary benzamides. Inspired by these encouraging results, we next explored these conditions for the synthesis of terminal amides by “in-loop” carbon-11 carbonylation (Table 1, entry 10). The reactive complex was formed from iodobenzene (1-I), N-xantphos, and Pd(dba)₂ in THF in an external vial at room temperature over 20 minutes. The resultant mixture was loaded onto an HPLC loop, treated with incoming [11C]CO and then the HPLC loop was sealed for 5 min. After this time, the intermediate complex was transferred into the reactor containing DMAP and formamide. The second reaction was conducted at 100 °C for 5 minutes. Water/acetonitrile (1:1) was added to the reactor and again heated at 100 °C for 5 minutes. Gratifyingly, this process resulted in 74% RCC of the desired product, [¹¹C]-1-CONH₂, perhaps due to the use of larger quantities of reagents sued in Schou et al.’s report [27] compared to other methods (e.g. 205 μmol DMAP; 2.5 mmol formamide).

With suitable labeling conditions in hand, we next examined the substrate scope for producing carbon-11 amide derivatives using this transformation (Scheme 2). A variety of aryl iodides with both electron-withdrawing (2-I – 5-I) and electron-donating (6-I – 11-I) substituents successfully reacted to produce a series of carbon-11 labeled primary benzamides. First, we tested para-substituted aryl iodides, including those substituted with electron-withdrawing groups (p-NO₂ 2-I, p-CF₃ 3-I, p-CN 4-I,), which produced [¹¹C]-2-CONH₂, [¹¹C]-3-CONH₂ and [¹¹C]-4-CONH₂ in 51%, 52% and 54% RCC, respectively. Aryl iodides substituted with two meta electron-withdrawing substituents (5-I) also worked well; bis-trifluoromethyl substituted aryl iodide (5-I) produced [¹¹C]-5-CONH₂ in 47% RCC.

Scheme 2.

Substrate scope for two-step carbon-11 carbonylation.

^aRadiochemical conversion (RCC) determined via radio-HPLC as the average of two separate reactions conducted at least 2 replicates. See supplementary information for additional details.

Next, we evaluated effect of electron-rich groups for this conversion. For example, 1-iodo-4-methoxybenzene (6-I) afforded [¹¹C]-6-CONH₂ in 66% RCC. We also tested the impact of an unmasked phenol group on RCC and found that, in the case of unprotected phenol analog 4-iodophenol (7-I), the reaction still proceeded but the product was obtained in lower RCC (22%) than the methoxy derivative (6-I). As such, a protecting group for phenol analogues is recommended for best results.

We next investigated the effect of the p- and o-substituted methyl group for the synthesis of carbon-11 labeled primary benzamides. Both 4- and 2- methyl substituted aryl iodides were suitable substrates, generating [¹¹C]-8-CONH₂ and [¹¹C]-9-CONH₂ in 71% and 47% RCC respectively. Steric effects play a significant role in complex formation with palladium catalysts, leading to a marked difference in the RCC values of the products [¹¹C]-8-CONH₂ and [¹¹C]-9-CONH₂. The sterically hindered ortho-substituted aryl iodide substrate (9-I) yielded only 47% RCC, compared to the para-substituted aryl iodide substrate (8-I), which achieved a higher RCC of 71%. Next para-chloro-substituted aryl iodide 10-I gave [¹¹C]-10-CONH₂ in 71% RCC. Substrates flanked with electron-withdrawing groups performed better than electron-donating groups. Furthermore, carbon-11 aminocarbonylation of biaryl ether 11-I gave [¹¹C]-11-CONH₂ in 66% RCC.

Of note, HPLC analysis of the crude reaction mixtures for the majority of substrates investigated revealed not only the peak corresponding to the desired benzamide, but also a second peak eluting approx. 1 – 2 min later (see Supplementary Information). From prior work preparing carbon-11 labeled carboxylic acids [19], we know this is the corresponding acid that is formed in varying amounts alongside the benzamide due to adventitious water in the process. To date we have not been able to suppress concomitant formation of both the labeled acid with the benzamide.

After successfully testing different carbocyclic iodoarenes, we next investigated ¹¹C- aminocarbonylation of heteroaromatic halides. ¹¹C-Aminocarbonylation of 2-bromofuran 12-Br resulted in [¹¹C]-12-CONH₂ in 2.9% RCC. Thiophenes were also compatible with this new methodology, and ¹¹C-amino carbonylation of 2-iodothiophene (13-I) generated [¹¹C]-13-CONH₂ in 61% RCC. Similarly, 4-iodopyridine analogue 14-I generated [¹C-isonicotinamide [¹¹C]-14-CONH₂ in 69% RCC. Aryl iodides were found to work better than aryl bromides, and we believe this is the reason for the low RCC of [¹¹C]-12-CONH₂.

Next, we investigated the labeling of bioactive molecules of interst using this new method. The synthesis of PARP inhibitor carbon-11 veliparib ([¹¹C]-15-CONH₂) was investigated using the optimized carbon-11 carbonylation conditions, and was obtained from 15-I in 27% RCC (n = 5). Expanding the scope further, radiolabeling of amino acids is of enormous interest to our imaging program. Thus, to determine the applicability of the current method for radiolabeling unprotected amino acids, we evaluated unprotected haloaromatic amino acid 16-I, and we gratified to obtain [¹¹C]-16-CONH₂ in 31% RCC.

Lastly, the labeling of kappa opioid receptor (KOR) antagonist LY2795050 was also investigated under the optimized carbon-11 carbonylation methodology (Scheme 3). Originally developed at the Yale PET Center [25], this imaging agent is now also in routine use at our PET Center and we have spent considerable time evaluating the most effective method for its routine production. [24][20] We were thus interested in whether it could be readily accessed using this new ¹¹C-carbonylation method. [¹¹C]LY2795050 ([¹¹C]-17-CONH₂) was obtained from 17-I using the optimized conditions in 8.4 ± 3.2% RCY. To adapt the method for routine production requires automation, and so we also used LY2795050 as a proof-of-concept case with which to demonstrate automated labeling using a TRACERLab synthesis module. The fully automated synthesis of [¹¹C]-17-CONH₂ gave the product in a decay-corrected yield of 209 ± 78 mCi (7748 ± 2898 MBq) and RCY of 8.4 ± 3.2% (25 ± 9 mCi (942 ± 352 MBq) at end-of-syntheses, corresponding to 1.0 ± 0.41% non-decay-corrected RCY based upon ~92.5 GBq (~2.5 Ci) of starting [11C]CO, 62 min synthesis, n = 3), enough for clinical research use. Furthermore, the molar activity was 79 ± 25 TBq/μmol (2137 ± 679 Ci/mmol), and the product was obtained in 99.5 ± 0.7% radiochemical purity, further confirming suitability of the new method for routine production. Given these encouraging results, future work will validate the synthesis for clinical production and compare it with the other methods in use at our facility to determine which is best for routine production based on factors such as yield, reliability and reproducibility.

Scheme 3.

Pd-mediated carbonylation for the synthesis of [¹¹C]LY2795050.

Conclusion:

In summary, a simple and efficient method was developed for the ¹¹C-labeling of primary benzamides. This approach utilized carbon-11 labeled aroyl dimethylaminopyridinium salts in a two-step in-loop carbonylative cross-coupling reaction of aryl halides with [11C]CO, yielding a diverse range of carbon-11 labeled primary benzamides with good to excellent yields. The automated synthesis of the kappa antagonist compound [¹¹C]LY2795050 demonstrated good radiochemical yields can be achieved for clinically relevant PET imaging agents. The highlight of this method includes its straightforward chemistry, ease of execution, moderate to high radiochemical yields and purity, and compatibility with commercial radiochemistry modules for the automated production of the carbon-11 labeled compounds. The method enables ready access to carbon-11 labeled primary benzamides using standard automated synthesis modules, and routine chemistry suitable for use in most radiochemistry production facilities.