Late-Stage Carbon-14 Labeling and Isotope Exchange: Emerging Opportunities and Future Challenges.

Carbon-14 (14C) is a gold standard technology routinely utilized in pharmaceutical and agrochemical industries for tracking synthetic organic molecules and providing their metabolic and safety profiles. While the state of the art has been dominated for decades by traditional multistep synthetic approaches, the recent emergence of late-stage carbon isotope labeling has provided new avenues to rapidly access carbon-14-labeled biologically relevant compounds. In particular, the development of carbon isotope exchange has represented a fundamental paradigm change, opening the way to unexplored synthetic transformations. In this Perspective, we discuss the recent developments in the field with a critical assessment of the literature. We subsequently discuss research directions and future challenges within this rapidly evolving field.

Introduction

Carbon-14 (14C) is a naturally abundant radioactive isotope of carbon-12 (12C), with a half-life of 5730 years. This long-lived radioisotope emits low-energy β-particle radiation (i.e., electrons with a mean energy of radiation of 56 keV) that is commonly utilized as a traceless tag for organic molecules to study their fate. This unique tool, in association with β-counting and β-imaging technologies, provides vital knowledge on the fate of synthetic organic molecules (Figure ).[1] This information is critical to establish potential issues affecting human health and is required worldwide by regulatory agencies such as OECD, FDA, and EMA for the following:

Figure 1

Carbon-14 radiolabeling: a traceless tool for tracking organic molecules, supporting drug development and the agrochemical industry.

pharmaceutical development: to unveil drug metabolism, disposition, and pharmacokinetics of novel pharmaceuticals[2,3]

animal health drug development: to determine the metabolism, disposition, and pharmacokinetics of veterinary drugs[4]

crop science: to understand the plant metabolism of agrochemicals and pesticides[5,6]

human food safety evaluation: to ensure that food derived from animals that have been treated by veterinary drugs is safe for human consumption[7]

environmental fate studies: to perform soil dissipation studies and assess potential environmental impacts associated with human and animal health product excretions that might enter the aquatic and terrestrial environment[8,9]

Rapid and straightforward access to 14C-radiolabeled organic molecules is a strict requirement for accelerating research in these fields with high societal impacts. Surprisingly, carbon radiolabeling still represents a bottleneck and a largely unsolved fundamental problem.

These limitations are related to challenges associated with the fundamental source of the radioisotope, as carbon-14 is generated in nuclear reactors as Ba[14C]CO3.[10] Consequently, all 14C-labeled compounds are derived from this common carbon source. The radioactivity is incorporated into the chemical scaffold at an early stage of the synthesis and transformed, by a series of consecutive steps, into the desired molecule. This classical approach is marred by several major drawbacks, namely, the generation of massive amounts of radioactive waste (extremely difficult to dispose of) and the multistep, time-consuming nature of these synthetic approaches, which mandates the development of a specific route in line with radio-synthetic requirements and safety regulations. Last but not least, 14C radiosynthesis is highly resource-demanding, as the price for 37 GBq of Ba[14C]CO3 (ca. 3.3 g, corresponding to ca. 800 mg of [14C]CO2) is 25 k€, and any additional step will skyrocket the overall cost.

While 14C-labeling is constantly utilized by pharmaceutical and agrochemical companies, methodological innovation toward this isotope has been scarce until recently. In the past five years, we witnessed a sudden revival in interest toward this isotope, and new technologies have appeared, providing effective alternatives to multistep procedures. The advent of late-stage 14C-radiolabeling and the conceptualization of the first practical examples of carbon isotope exchange (CIE) have provided much enthusiasm in this field with a resurgence of interest in these long-lasting challenges. In this Perspective, we aim to paint a portrait of the vibrant state of the art, focusing exclusively on the recent developments in late-stage labeling that have offered tangible evidence of its applicability to carbon-14 labeling. Consequently, all methodologies that apply uniquely to stable 13C, but do not provide concrete use of the radioisotope, will not be included in this Perspective.

Brief History

Since its discovery, made on February 27, 1940, by Martin Kamen and Sam Ruben at the Lawrence Berkeley National Laboratory,[11]14C has been instrumental in a wide range of applications, and it particularly helped revolutionize many fields of life science (Figure ).[12,13] It is well-known that the utilization of [14C]CO2 turned out to be key in the determination of how plants utilize carbon dioxide in the process of photosynthesis.[14,15] This was immediately recognized by Melvin Calvin who stated, in his notorious Nobel Prize Lecture in 1961, that “Ruben and Kamen provided the ideal tool for the tracing of the route along which carbon dioxide travels on its way to carbohydrate.”[16,17]

Figure 2

Historical timeline of carbon-14 uses and applications.

Another early recognition of this “supremely important tracer”[18] was made by Willard Libby, who conceptualized in 1946[19] his groundbreaking idea on the most notorious technology related to this radionuclide: radiocarbon dating.[20] Libby realized that 14C, formed in the atmosphere by a reaction between neutrons and nitrogen-14,[21,22] would find its way into living organisms, which would thus be tagged with the radioisotope.[23] The “radiocarbon revolution” had tremendous effects on the fields of archeology, geology, geophysics, and other branches of science,[24] and Libby was recognized with the Nobel Prize in Chemistry in 1960.[25]

14C has also been utilized in organic chemistry as a radioactive tracer element to elucidate and provide evidence on reaction mechanisms.[26] This work was prevalently performed in the 1950s, before NMR and its value in chemistry had been elucidated and commercial instruments became available.[27] A prominent example is the elucidation of the Claisen rearrangement by Ryan and O’Connor in 1952.[28] The authors managed to provide undisputable evidence of the [3,3]-sigmatropic rearrangement by inserting a 14C-tag on the terminal carbon of the allyl group. After the rearrangement, the precise location of the 14C-tag was unequivocally established by a series of degradative processes onto the resulting phenol, in association with beta-counting on the derived fractions.[29] Other examples of the use of 14C in mechanistic investigations are the confirmation of the existence of a benzyne intermediate[30,31] and the rearrangements of carbon atoms in t-butyl and t-amyl derivatives by J. D. Roberts;[32−34] the series of studies on the Wagner rearrangement by Collins and co-workers;[35−39] the reaction of diazo compounds with nitroolefins by O’Connor;[40] and the investigation on benzidine rearrangement by Smith.[41]

Today, while 13C has replaced 14C for routine mechanistic investigations, the role of 14C is still eminent in the elucidation of biosynthetic pathways.[42] The use of simple radiolabeled precursors, often amino acids, allows the unequivocal tracking of their evolution into final products in highly sophisticated biological environments.[43] A recent example is the remarkable work from Evanno, Poupon, and Thomas on the biosynthesis of the cyclic guanidine alkaloids from Crambeidae marine sponges.[44]

The use of 14C-labeled biologically active molecules to determine their metabolism and disposition started in the late 50s. While we cannot state with certitude the first reported study, early representative examples were the labeling of N-[14C] methyl-erythromycin by Flynn, Murphy, and McMahon in 1955[45] and the determination of its elimination and metabolism in rats in 1956 at Eli Lilly and Company.[46,47] The use of 14C has since become common practice and a gold standard for a variety of applications in drug development such as the determination of drug absorption, distribution, metabolism, and excretion (ADME); mass balance studies;[48] and quantifying drug concentrations in target organs using whole body autoradioluminography (WBAL).[49] In this context, the capacity to introduce the radioisotope into molecules of interest in a cost-effective, timely, and efficient manner is of utmost importance.

Production of 14C Reagents and Isotope Specificities

Carbon-14 is artificially generated by neutron bombardment of solid beryllium or aluminum nitride, a process that might range from one to three years. After treatment, this nuclear reaction leads to the isolation of barium carbonate Ba[14C]CO3 as the first carbon-14 building block in organic radiosynthesis. This starting material is transformed into four major derivatives from which all 14C building blocks derive, namely, 14C-carbon dioxide ([14C]CO2), metal 14C-cyanides ([14C]MCN), 14C-acetylene, and 14C-cyanamide ([14C]BaN-CN).[50] These primary building blocks are further converted into more complex secondary intermediates such as [14C]CH3OH or H[14C]COOK. Of course, even more sophisticated precursors can be prepared such as mono- or multilabeled aromatic rings (Scheme , top).[51,52]

Scheme 1

(Top) Arborescence of 14C-Labeled Stating Materials, (Bottom Left) Monomers That Undergo Polymerization and/or Radiolysis, and (Bottom Right) Example of a Reaction Providing Different Product Outcomes between Cold and Hot Synthesis

In most cases, the presence of radiocarbon in the structure of a given drug has no notable effect on its biological properties, as the size increase of the isotope is negligible compared to hydrogen–deuterium or hydrogen–tritium, and the kinetic isotope effect of carbon is minor, as well.

Nonetheless, due to the inherent electron emission of the isotope, covalent bond cleavage is possible. Indeed, molecules labeled with 14C might not behave as stably labeled 12C- and 13C-isotopologues and might undergo decomposition, a process known as radiolysis.[53] For example, benzene labeled with more than three 14C atoms tends to polymerize spontaneously (Scheme , bottom left).[10] The same process has been reported for labeled monomers such as methyl methacrylate, methyl acrylate, vinyl acetate, and vinyl chloride, even at low molar activity (Am).[54,55]Am is defined as the amount of radioactivity per unit mol of the element. From the theoretical point, the maximum molar activity of carbon-14 is 2.31 GBq mmol–1.[10]

Besides radiolysis, differences in reactivity between stable 12C- and 14C-labeled compounds have been sporadically reported. Parker described three examples including the trifluoromethylation of 1-chloro-4-iodobenzene (Am 1.8 GBq mmol–1), where product distributions differed quite remarkably between the nonlabeled and the radioactive syntheses (Scheme , bottom right).[56,57] Additional examples concerned aniline bromination and a Johnson Claisen rearrangement, where differences in reactivity were noticed.[57] In 1986, Parker made the hypothesis that “if a chemical reaction can proceed by two or more pathways, whereof one involving the participation of free radicals, the latter might be favored, when a high molar activity radiolabeled species is involved.”

For the above-mentioned reasons, in the context of synthetic methodology development, we strongly encourage not considering the transfer from a stable isotope (i.e., 13C) to radiocarbon as a negligible and automatic step. Additionally, as technical and stoichiometric differences are often present when using 14C reagents, the case of the stoichiometry of [14C]CO2 being a crystal clear example, we advise against comparing results obtained with an excess of cold [13C]CO2 to 14C radioactive molar activities.

Analysis of the Chemical Reactions and Linear Sequences in 14C Radiosynthesis from the Most Common Building Blocks

In order to gain precise and quantitative information on the current state of the art in radiocarbon synthesis, we compiled an inventory of 129 publications reported on the radioisotope-specialized Journal of Labeled Compounds and Radiopharmaceuticals, between 2010 and 2020 (see the Supporting Information for a detailed list of publications). These articles described 154 syntheses aiming to insert 14C into the molecules of interest. Of note, pharmaceutical companies published more than 75% of these articles, thus underpinning the strategic importance of this isotope in the pharmaceutical industry. An analysis of the starting material revealed that cyanide salts are by far the most used precursors, accounting for 32% of all processes (Figure A). The simple access to these primary sources, their high reactivity, and their simpler handling compared to gas sources might explain this observation. In comparison, [14C]CO2, the second most frequent source, is found in 12% of the syntheses (generated either in situ from solid Ba[14C]CO3 or directly in its gaseous form). Building blocks such as carbonyl derivatives and alkyl halides represent 12% and 7% of the isotopic sources, respectively. Technically advanced arenes account for 11% of the distribution. Finally, in a similar proportion, formate salts (6%), urea derivatives (6%), and cyanate salts (5%) are also significant precursors for radiolabeling. Chemical reactions utilized to incorporate the radionuclide strictly depend on the selected labeled building blocks (i.e., M[14C]CN, [14C]CO2, alkyl halide, etc.). For the six 14C precursors selected in Figure B, generally one procedure prevailed over the others. For carbon dioxide, carboxylation reactions in the presence of organomagnesium or organolithium species prevail over direct reductions, while condensation reactions are most common for urea derivatives. Concerning cyanide salts, transition-metal catalysis and nucleophilic substitutions are the most representative pathways to highlight. However, for other sources such as formate salts and alkyl halides, various sequences have been reported (see the Supporting Information for a detailed list of publications).

Figure 3

Analysis of 129 publications reporting 14C radiosynthesis over the past decade. (A) Pie chart of building blocks utilized as a 14C-source. (B) Types of reactions utilized to introduce the isotope. (C) 14C-labeled positions on the final molecule of interest. (D) Number of linear steps required to complete the radiosynthesis.

The analysis of the labeled position on the final molecule clearly shows that carbon C(sp2) hybridization is preferred, as observed in three-quarters of the labeled molecules (74%, Figure C). Carbonyl (29%), imine derivatives (24%), and labeled arenes (15%) are the most recurrent labeled positions encountered. With regard to the other hybridization states of carbon, the C(sp3) distribution nearly reaches one-quarter of the substrate inventory (24%), while C(sp1) labeling is rarely targeted (3%).

Lastly, we looked to the number of linear steps required for the incorporation of the radionuclide onto the final labeled molecule (Figure D). Most commonly, two, three, or five steps are necessary to obtain the final substrate with a median value that was determined to be 4 steps. However, in 30% of reports, 6 steps or more are required to provide the labeled compounds, including 6% of processes with a nightmarish 10 step radiosynthesis. In this context, it is somehow shocking to notice that late-stage procedures (i.e., the use of one single radioactive reaction) account only for 6% of the total number of radiosyntheses analyzed.

Radiocarbon Labeling

Compared to long multistep radiosynthesis (Scheme A), a tardive insertion of the isotope has the advantage to reduce the overall cost of the approach, the time required to obtain the molecule, and the amount of radioactive waste generated, with a beneficial impact on the overall efficiency of the process.

Scheme 2

Comparison between (A) Classical Multi-Step Radiosynthesis, (B) a Late-Stage Approach, and (C) Carbon Isotope Exchange

In this area, we can distinguish two possible approaches. On one hand, the late-stage incorporation of 14C, ideally at the last step of the process (Scheme B), has the benefit of inserting the 14C without isotopic dilution (i.e., high Am). The main drawback is the necessity to prepare a suitably decorated precursor, elaborated specifically to adapt to the methodology. Though this process does not require handling of radioactive material, it might be time-demanding and require the design of an alternative retrosynthesis compared to the molecule of interest.

Recently, an alternative strategy has been reported: carbon isotopic exchange (CIE, Scheme C). This approach aims to replace a 12C functional group (−COOH, −CN) with its radiolabeled counterpart in one single operation, directly on the final compound of interest. CIE is appealing as it avoids the requirement for the elaboration of a precursor and is the most straightforward way to access 14C-labeled molecules. As these transformations are based on equilibria, 14C is inserted with some degree of isotope dilution, and the Am will be lower compared to the other strategies.

Late-Stage 14C-Labeling with High Molar Activity

Late-Stage Labeling from [14C]CO2

Carbon dioxide is the primary radiocarbon source and one of the most attractive building block to incorporate the radionuclide. Nevertheless, its high thermodynamic stability represents a severe challenge and a limitation for effective utilization.[58] In addition, compared to cold unlabeled chemistry, the use of a large excess and high pressures of [14C]CO2 is simply inconceivable from both financial and safety standpoints.

In 2012, radiochemists at AstraZeneca published two strategies based on the partial reduction of CO2 and its subsequent utilization for ring closure reactions (Scheme ). Using an excess of [14C]CO2 in the presence of the Schwartz reagent,[59]14C-formaldehyde was generated in situ and immediately reacted with the sulfonamide precursor 1 to obtain the [14C]hydrochlorothiazide 2, a diuretic drug (Scheme , top).[60] Though the radiochemical yield (RCY) from labeled CO2 was modest (18%), the advantage of this one-pot strategy is not negligible.

Scheme 3

One-Pot CO2 Reduction Strategies: (Top) Reduction to Formaldehyde to Label Hydrochlorothiazide and (Bottom) Reduction to Formate and Subsequent Cyclization

Xanthine derivative 4, a myeloperoxidase inhibitor, was labeled on the imidazole backbone via a 2-step process from CO2 (Scheme , bottom). Lithium 14C-formate was prepared in situ after a reduction of [14C]CO2 with superhydride LiEt3BH. After the solution was concentrated, the precursor 3 was added in the presence of EDCI to form the corresponding formylated intermediate, which was further converted into the desired product 4 under basic conditions.[61] The overall radiochemical yield from CO2 was low (8%), but sufficient compound was delivered to be administered to rats for quantitative whole body autoradiography (QWBA) experiments.

Urea derivatives are common functional groups in medicinal chemistry and are often found in pharmaceuticals.[62,63] In 2018, Hesk and co-workers at Merck applied a strategy, published a few years earlier,[64] for the labeling of 7, a potent inhibitor of the NPY5 receptor.[65] A trapping of radioactive carbon dioxide by the aniline 5 under basic conditions, followed by an addition of POCl3 as a dehydrating agent, led to the generation of an isocyanate intermediate. The subsequent addition of aliphatic amine 6 afforded the expected labeled urea 7 with a radiochemical yield of 32% and high Am (Scheme ). While effective, the disadvantage of this reaction, which is commonly used in 11C-labeling, is the requirement for a dehydrating agent that is poorly tolerant with elaborated functional groups.

Scheme 4

Labeling of Urea from [14C]CO2 Using a Dehydration Protocol in the Presence of POCl3

The same year, our group reported a last-step Staudinger aza-Wittig procedure (SAW) for the labeling of cyclic ureas.[66] Inspired by the biorthogonal Staudinger ligation,[67] we aimed to develop a broadly tolerant methodology suitable to a large variety of substructures (Scheme ). A treatment of an azido-amine derivative with dimethylphenylphosphine resulted in the instantaneous formation of diatomic nitrogen and the corresponding iminophosphorane. To this frozen solution, a stoichiometric amount of [14C]CO2 was precisely delivered to the reaction using an RC Tritec carboxylation manifold, generating an intermediate isocyanate, and subsequent intramolecular nucleophilic addition delivered the cyclized urea products (Scheme , left).

Scheme 5

Last-Step Radiolabeling of Ureas and Carbamates Using a Staudinger Aza-Wittig Approach with [14C]CO2

This procedure was shown to be extremely mild and effective: [14C]CO2 was converted into the desired cyclic ureas within 5 min at room temperature. This protocol was applied with success to the labeling of four pharmaceutically relevant cyclic urea derivatives, including flibanserin (8) and oxatomide (9) in suitable radiochemical yields. Functional group compatibility was showcased by the labeling of an unprotected heptapeptide bearing multiple functional groups, including alcohol, carboxylic acid, amine, and indole. Finally, it was shown that the methodology could be successfully applied to the short-lived carbon-11 isotope.

In 2020, this procedure was applied to the labeling of 5- and 6-membered cyclic carbamates (Scheme , left).[68] Using the same approach with more challenging hydroxyl nucleophiles (both alcohols and phenols), nine carbamates were radiolabeled, including five pharmaceuticals in 40–59% radiochemical isolated yields. Once more, the applicability of the procedure to 11C radiolabeling was identified on 24 substrates, including a linear carbamate. To avoid precursor elaboration, we validated a 3-step disconnection/reconnection strategy, involving ring opening/isotopic closure, which allowed the labeling of 14C-zolmitriptan starting from the commercially available unlabeled pharmaceutical.

In 2021, we reported a linear version of the SAW procedure for the labeling of urea substructures (Scheme , right).[69] By using all combinations of aliphatic/aromatic azides and amines, access to dissymmetric linear ureas was granted, including derivatives such as hydroxyureas, semicarbazides, or sulfonylureas. This procedure was applied on 43 derivatives using [13C]CO2 under identical stoichiometric conditions compared to 14C. It was shown that, in the presence of less nucleophilic anilines, the use of N-methylimidazole additive (2 equiv) was required to obtain satisfactory yields. This sequence enabled the 14C-labeling of four drugs from 26% to 82% isolated yields.

Late-Stage Labeling from [14C]CO

Carbonylation reactions are very frequently encountered in late-stage labeling, in particular for stable 13C and short-lived 11C.[70] On the contrary, this is not the case for [14C]CO, which has only rarely been reported as a building block. Besides the inherent toxicity of [14C]CO that requires obvious safety precautions, its use has been underexploited due, in particular, to its poor stability. Hargraves and co-workers investigated the cryogenic isotopic distillation of [14C]CO as a technology for isotope concentration and found that [14C]CO undergoes significant radiolysis even when stored at low temperatures.[71] Similar observations were reported by Hardy and co-workers.[72] Consequently, this gas is generated in situ and utilized directly for further functionalization.

[14C]Formate salts have been successfully reported to generate [14C]CO in the presence of concentrated sulfuric acid.[73] In 2011, Elmore and co-workers at AstraZeneca published the synthesis of a delta opioid agonist (15) based on a palladium-catalyzed amino-carbonylation as the final step (Scheme , top). In this procedure, [14C]CO was produced by [14C] sodium formate dehydration after the addition of sulfuric acid at 110 °C.[74] The same strategy was used for a series of other carbonylations from aryl halides, such as hydroxycarbonylations,[75,76] alkoxycarbonylations,[77] and intramolecular aminocarbonylations.[78]

Scheme 6

Carbonylative Cross-Coupling Utilizing [14C]CO Generated from [14C]Formate Sources

In 2013, Whitehead et al. developed an alternative procedure that enables [14C]CO generation under milder conditions and applied it to the radiosynthesis of celivarone 17 (Scheme , bottom).[79] Carbon monoxide was liberated after an addition of lithium [14C]formate to a chlorophosphonium cation, previously generated from triphenylphosphine and perchloroacetone.[80] Under these conditions, the formate salt is immediately transformed into [14C]formyl chloride that readily dissociate at temperatures exceeding −60 °C to give [14C]CO. [14C]Celivarone was labeled in 53% RCY by a subsequent palladium-catalyzed alkoxycarbonylation reaction.

Skrydstrup’s group pioneered the development of solid carbon monoxide sources, which allows the release of the poisonous gas under safer and simple conditions.[81] In collaboration with Elmore at AstraZeneca, they reported in 2012 a method to generate [14C]CO based on a solid and bench stable [14C]CO precursor, COgen (20).[82] Utilizing a two-chamber glassware reactor, the release of [14C]CO takes place in the first chamber through a palladium-catalyzed decarbonylation of 20, and gas diffusion in the second chamber allows the desired carbonylation reaction (Scheme B). 14COgen is synthesized in two steps from [14C]CO2 by lithiation of 9-methyl-9H-fluorene 19 and subsequent carboxylation, which is followed by activation of the carboxylic acid with oxalyl chloride with an overall radiochemical yield of 91% (Scheme A). Using this method, a variety of combinations of aryl halides and amines or boronic acids led to the labeling of relevant amides or ketones, respectively. In the presence of isotopically diluted [14C]COgen, three drugs were labeled with success (Scheme ). Olaparib 21 and thalidomide 23 were obtained by aminocarbonylation in 37% and 70% RCY, respectively. Fenofibrate 22 was labeled in 72% RCY starting from the corresponding boronic acid.

Scheme 7

Formation of [14C]CO from [14C]COgen and Subsequent Use for Pd-Catalyzed Carbonylation Reactions

Prof. T. Skrydstrup is kindly acknowledged for providing the image of the two-chamber glassware reactor.

In 2019, Sardana et al. developed a visible-light-mediated late-stage aminocarbonylation methodology, starting from alkyl iodides and secondary amines in the presence of stoichiometric amounts of labeled CO (Scheme ).[83] After extensive optimization with 13C, palladium catalysis afforded the labeled amide at ambient temperature using COgen as a limiting reagent. Using isotopically diluted 14COgen with unlabeled material (at 5% 12C/14C, Am 0.11 GBq mmol–1), this procedure was successfully applied to the radiocarbonylation of two amides 24 and 25 with the expected molar activities (Scheme ).

Scheme 8

Visible-Light-Mediated Aminocarbonylation of Unactivated Alkyl Iodides with Stoichiometric [14C]CO

Ketones are common functional groups in organic chemistry and are found in fine chemicals and particularly in pharmaceuticals. While (hetero)aryl ketones can be prepared and labeled by carbonylative cross-couplings using palladium catalysis (see fenofibrate 22, Scheme ), the propensity for palladium-alkyl intermediates to undergo competing beta-hydride elimination has limited the development of effective carbonylative solutions to aliphatic ketones.[84] In 2020, Skrydstrup’s group reported a stoichiometric nickel-mediated carbonylative coupling of alkyl zinc reagents with alkyl iodide to access labeled alkyl ketones (Scheme ).[85] Starting from NN2 pincer nickel(II) complex (A), an addition of alkyl zinc reagent produced a nickel(II) alkyl complex (B). Ex situ generation of CO in a double-chamber COware reactor was followed by insertion into the NiII-alkyl complex (B), thus generating the corresponding NiII-acyl species (C). A subsequent addition of alkyl iodide to a second nickel(I) complex, reduced by an addition of pincer nickel species (A) and manganese Mn, led to the generation of an alkyl radical that provided the NiIII-(alkyl)acyl complex (D). Finally, reductive elimination afforded the expected labeled alkyl ketone (E). Based on this procedure and using a stoichiometric amount of nickel(II) and nickel(I) complexes, 14C-labeled nabumetone 27, a nonsteroidal anti-inflammatory drug (NSAID), was isolated with a radiochemical yield of 36% and an Am of 0.5 GBq mmol–1.

Scheme 9

Nickel Complexes Mediated Access to 14C-Labeled Aliphatic Ketones

Late-Stage Labeling from [14C]CH3I

14C-labeled methyl iodide represents a tempting secondary building block because it provides a wide range of opportunities to introduce the radionuclide: in particular, access to labeled O- and N-[14C]methylated substrates granted by electrophilic methylation. [14C]CH3I is prepared by LiAlH4 reduction of [14C]CO2 to [14C]CH3OH, followed by thermal treatment with HI, and is subsequently employed for alkylation.[10]

In 2012, Tanga and co-workers published the radiosynthesis of the pimaradiene diterpene acanthoic acid [C]28, using a 3-step deconstruction/reconstruction strategy (Scheme , top). Acanthoic acid was dihydroxylated with osmium tetroxide to give vicinal diol 29; subsequent oxidative cleavage with sodium periodate provided the aldehyde precursor 30. A late-stage Wittig reaction was utilized for the introduction of the carbon-14 radioisotope.[86] The labeled phosphonium ylide was prepared from [14C]CH3I and triphenylphosphine under basic conditions. An excess of the 14C-methylene Wittig reagent was added to the solution containing aldehyde 30, affording the labeled compound with a modest radiochemical yield of 9% and high molar activity (1.9 GBq mmol–1).

Scheme 10

Radiosynthesis of Acanthoic Acid and Methylsulfones by Deconstruction/Reconstruction Strategies

In 2016, Gauthier Jr. and co-workers at Merck presented a deconstruction/reconstruction strategy based on the demethylation/[14C]methylation of methylsulfones (Scheme , bottom).[87] Under optimal conditions, methyl phenyl sulfones 31 were treated with benzyl bromide in the presence of excess tBuOK followed by in situ elimination of the resulting styrene, thus leading to the formation of sulfinate salt 32. A final alkylation with [14C]CH3I promoted the formation of the labeled methylsulfones [C]31. This strategy was applied to the synthesis of two pharmaceutically active molecules, etoricoxib 33 and odanacatib 34. Using high-molar-activity [14C]CH3I (Am = 2.1 GBq mmol–1), 33 was isolated in 21% RCY. According to the authors, the reaction with stably labeled CH3I proceeded in >90% yield, and the low RCY is due to the incomplete transfer of the radioactive gas alkylating agent. This is another example where technical issues due to the manipulation of the radioisotope might deeply affect the transposition of the cold synthesis. 14C-labeling of odanacatib 34 (30% RCY), a selective inhibitor of cathepsin K, was required for environmental risk assessment studies.

Late-Stage Labeling from [14C]CH2O

In 2015, Baran and co-workers reported a methodology for the hydromethylation of unactivated alkenes promoted by Fe(acac)3 (Scheme ).[88] In the first step, a sulfonylhydrazone was generated by condensation of formaldehyde on the corresponding sulfonylhydrazide 35. An excess of this reagent was added to a solution of alkene, in the presence of phenylsilane and an iron complex carrying out the formation of alkylhydrazine 37. Finally, a solvent switch to methanol and gentle heating of the reaction at 60 °C yielded the hydromethylated adduct 38, by reductive C–N bond cleavage, and a release of sulfinic acid.

Scheme 11

Hydromethylation of Unactivated Alkenes with Formaldehyde

This procedure was successfully applied to a variety of substrates using unlabeled formaldehyde (6 equiv) and even on rotenone, an olefin bearing insecticide, in 50% yield on a gram scale. Using [14C]HCOH (Am = 2.1 GBq mmol–1), iron-mediated hydromethylation of rotenone was successfully realized in collaboration with the Maxwell group at Bristol Myers Squibb. The 14C-labeled compound 39 was isolated in 18% yield from rotenone (6% RCY from [14C]HCOH). We might speculate that, to reduce costs associated with the labeled reagent for the 14C reaction, the number of equivalents of 14C-formaldehyde was limited to 3.2, which can explain the lower yield compared to the unlabeled experiment.

Based on their previous findings on the metallaphotoredox cross-electrophile coupling strategy,[89,90] in 2021, MacMillan and co-workers reported a radiomethylation procedure suitable to multiple isotopes, including tritium, 11C, and 14C (Scheme ).[91]

Scheme 12

Metallaphotoredox Cross-Electrophile Coupling Strategy for the Radiomethylation of Aryl and Alkyl Bromides

Starting from the alkyl or aryl bromide derivatives and a corresponding radiolabeled methyl source ([14C]CH3ONp), the desired molecules were synthesized based on a dual photoredox and metal catalysis mechanism, mediated by silyl radical activation of alkyl halides. Aiming for practical applicability, the authors selected isotopically labeled methyl naphthalene-2-sulfonate as a limiting reagent, owing to its nonvolatility compared to MeI. In the presence of the additive LiBr, [14C]CH3ONp undergoes an in situ Finkelstein-like reaction to form [14C]CH3Br.

Under blue light irradiation, the photoredox catalyst (A) is excited to generate, after single-electron oxidation of bromide (a dissociable ligand of nickel B), a bromine radical and the reduced photocomplex. Bromine-radical-promoted hydrogen-atom abstraction from tris(trimethylsilyl)silane (TMS)3SiH generates the stabilized silyl radical intermediate (C), which in turns affords the radiolabeled methyl radical (D) group through halogen-atom abstraction. On the other hand, zerovalent nickel catalyst undergoes oxidative addition with the corresponding aryl bromide to furnish a NiII intermediate (F). Oxidative capture of a labeled methyl radical leads to the formation of NiIII (G), and subsequent reductive elimination provides the expected product (I) and the corresponding NiI complex (H). Lastly, single-electron transfer from the reduced photocatalyst to the NiI (H) simultaneously regenerates both catalysts. This procedure was applied to the labeling of several 3H and 11C radiotracers. On the other hand, only a single example using 14C was reported: glipizide 41 was labeled with 50% isolated radiochemical yield and high molar activity (Scheme ).

Late-Stage Labeling from [14C]CN

Cyanide salts are the most used reagents for 14C incorporation due to their suitable handling, facile accessibility, and versatility of the corresponding nitrile functional group (Figure A). While the synthesis and reactivity of electrophilic cyanating sources have been largely investigated,[92−94] most are unfortunately unsuited to carbon radioisotope labeling. In 2016, our group developed an access to 2-aminobenzothiazole and 2-aminobenzoxazole derivatives via a KCN polarity inversion strategy using 1-chlorobenzotriazole (BtCl, Scheme ).[95] By mixing BtCl and [14C]KCN, electrophilic cyanating agent [14C]BtCN 42 was generated in situ. Subsequent addition of the aminophenol or aminothiophenol led to the formation of the labeled heterocycles. This method allowed the labeling of an advanced herbicide synthetic intermediate 44 in a one-pot approach. Starting from isotopically diluted [14C]KCN (Am = 0.3 GBq mmol–1), compound 44 was labeled with 45% RCY based on this procedure coupled with sequential peptide bond formation with the corresponding carboxylic acid.

Scheme 13

14C-Labeled Electrophilic Cyanating Reagents

In 2017, Song’s group at Johnson & Johnson described the synthesis of the electrophilic cyanating reagent [14C]NCTS 43.[96] A treatment of low-molar-activity [14C]KCN (Am = 0.1 GBq mmol–1) with bromine formed in situ cyanogen bromide, which was reacted directly with N-phenyl-p-toluenesulfonamide to provide the desired reagent in 68% RCY. In the paper, while the authors reported a method for decarboxylative cyanation catalyzed by palladium using stable [13C]NCTS, no application with 14C was provided.

Labeling Using Biocatalytic Enzymatic Reactions

Biocatalytic cascade reactions are a cost-effective alternative to multistep synthesis for carbon-14 incorporation. Using well-known biological mechanisms, this approach might provide access to complex labeled molecules in a one-pot procedure based on successive enzymatic catalysis. In 2000, Bacher’s group reported the biosynthetic labeling of an isoprenoid biosynthesis intermediate, 4-diphosphocytidyl-2C-methyl-d-erythritol, through a 4-step enzyme cascade reaction system that provided the expected 14C-labeled compound in 47% isolated yield from [2-14C]pyruvate.[97] In 2007, Roy’s group reported the use of nitrile hydrolyzing enzymes (nitrilases) as an attractive alternative to selectively hydrolyze nitrile derivatives under mild conditions.[98]

In 2021, Ren and co-workers from Merck published a biocatalytic synthesis of [14C]islatravir 46 (Scheme , top). This labeled adenosine-based nucleoside reverse transcriptase translocation inhibitor, under investigation for the treatment of HIV infection, was required to support drug metabolism and pharmacokinetics studies.[99]

Scheme 14

(Top) Comparison of Synthetic versus Biocatalytic Approaches to [14C]Islatravir Radiosynthesis and (Bottom) Biocatalytic Route to [14C]50 from [14C]Tyrosine

Their chemical radiosynthesis started from [14C]t-butyl acetate 45, providing 46 in an overall 6% RCY, in 9 linear steps. The low yield and the high cost, time, and waste generated by this approach necessitated the development of a “greener” alternative. Based on previous developments at Merck on the manufacturing of unlabeled islatravir by means of biocatalytic cascades,[100] the authors aimed to adapt it to 14C. Using [14C]acetaldehyde 47, a delicate starting material that undergoes slow radiolysis (10% per month), the biocatalytic approach provided 46 in a much improved 37% RCY and high Am in just one single reaction step. This work is a beautiful example of how biocatalysis might support the development of 14C tracers.

11-De-O-methyltomaymycin 50 is a highly toxic antitumor antibiotic utilized as a payload for antibody drug conjugate (ADC) development. To support in vivo pharmacokinetic and metabolism studies at Sanofi, Aubert and co-workers required a 14C-labeled version of the molecule.[101] While a multistep approach appeared to be too ambitious, the authors turned their attention to a fermentation process starting from labeled precursors.

After precise optimization, the authors were able to achieve their goal using the Streptomyces sp. FH6421 strain (Scheme , bottom). In the presence of 14C-labeled anthranilic acid 48 (2.1 GBq mmol–1) and tyrosine 49 (6.3 GBq mmol–1) as radiolabeled precursors, 14C-labeled 11-de-O-methyltomaymycin 50 could be isolated in 21% RCY and high Am (4.4 GBq mmol–1) and up to 97% radiochemical and chemical purity.

While biocatalytic cascade reactions have not been extensively utilized for 14C-labeling, the stunning achievements presented herein show that when a biocatalytic approach is possible, efforts should be directed toward its implementation.

Carbon Isotope Exchange

Carbon isotope exchange (CIE, Schemes C and 15) aims to selectively replace a 12C functional group (i.e., −COOH, −CN) with its radiolabeled counterpart directly on the final compound of interest, in one single operation.

Scheme 15

Carbon Isotope Exchange

Isotope exchange technologies have been extensively developed over the past decades for hydrogen, providing a vast arsenal of opportunities to introduce deuterium and tritium on pharmaceutical compounds.[102] On the other hand, this concept lay largely unexplored for the carbon isotope until recently.[103] Despite early examples on decarboxylative carboxylation of aliphatic carboxylates,[104−108] the synthetic applicability of such processes, requiring harsh thermal conditions (T > 300 °C), often pyrolytic,[109,110] did not find practical applications. One report on a biochemical isotope exchange strictly limited to 4-hydroxybenzoate, and catalyzed by phenol carboxylase, was reported by groups Aresta and Fuchs.[111]

CIE Using [14C]CO2

In early 2019, three independent reports based on transition-metal-assisted decarboxylative carboxylation sequences appeared. First, carboxylic acids are activated allowing the release of CO2 molecules and the creation of a transient C-metal bond. Then, in the presence of labeled CO2, competition between [12C]CO2 and [14C]CO2 for carboxylation occurred, providing the desired labeled carboxylic acid products as an inseparable mixture of 12C and 14C isotopomers.

As CIE is oftentimes based on reversible transformations and equilibria, 14C is incorporated in the desired products with some degree of isotope dilution with natural 12C. In order to evaluate the efficiency of isotope exchange and compare different procedures, a major parameter of interest is the isotope enrichment (% IE, indicating the ratio of 12C over 14C), which is directly proportional to the Am of the molecule (Figure ).

Figure 4

Correlation between isotopic enrichment (% IE) and molar activity (Am) for carbon-14. Color code assignment is arbitrarily selected by the authors, with the aim to present in a scholarly manner the CIE section of the review. Note: correlations between Am and their applications provide general indications. They might vary according to national laws, the internal company rules, and the inherent molecule specifications and study.

For applications in drug development, the Am is a parameter of utmost importance. For example, low-Am compounds (i.e., low IE < 10%) will only be useful for accelerator mass spectrometry (AMS) studies.[112] On the contrary, CIE methods allowing for an IE higher than 40% might be of interest for broader practical applications (human-ADME, mass balance).[113,114] On the other hand, for animal mass balance[115] and QWBA[116] studies as well as for agrochemical studies, a high IE (>1.9 GBq mmol–1) is normally used. From the theoretical point of view, the maximum molar activity (i.e., 100% IE) of carbon-14 is 2.31 GBq mmol–1 (or 62.4 mCi mmol–1).

Our group in collaboration with Thibault Cantat patented in 2018[117] a copper-catalyzed CIE starting from cesium carboxylate salts.[118] Using the RC Tritec carboxylation manifold to precisely deliver 3 equiv of labeled carbon dioxide, the mixture was heated at 150 °C leading to [12C]CO2 extrusion of aromatic carboxylates in the presence of copper(I) catalyst. The organometallic intermediate subsequently reacted with 14C-labeled CO2 affording the expected carboxylic acids (Scheme ). This procedure was applied with success to various aryl and heteroaryl carboxylates allowing carbon isotope incorporation in 2 h. Finally, 14C-labeling on several biological acids with good molar activities was performed. Interestingly, only 3 equiv of labeled CO2 was required (with a maximal theoretical IE of 75%) to obtain an IE in the range 40–54% (Scheme ). Limitations of the current method are the sensitivity to labile protons that favors the formation of the proto-decarboxylation byproduct, and the scope which tolerates aromatic carboxylated and is preferable with electron-withdrawing substituents. On the other hand, the 1-step radiolabeling of flumequine 51 showcases the effectiveness of this procedure over multistep approaches. Previously, [C]51 was labeled in 10 linear steps from [14C]CO2 for rat and dog disposition studies.[119]

Scheme 16

Copper-Catalyzed Carbon Isotope Exchange

In 2019, Baran’s group reported a procedure based on a stoichiometric nickel(II) transformation utilizing activated N-hydroxyphthalimide (NHP) redox-active esters from alkyl carboxylate derivatives (Scheme ).[120] The introduction of labeled CO2 allows the CIE at room temperature, and the reaction exhibited a wide tolerance toward primary and secondary alkyl carboxylates. In collaboration with the department of radiochemistry at Bristol Myers Squibb, the sequence was implemented to biologically relevant compounds using a large excess of [14C]CO2 (5.5–32 equiv). In particular, chlorambucil 54 and mycophenolic acid 56 were isolated with 17% and 20% IE, respectively.

Scheme 17

Stoichiometric Nickel-Assisted CIE

Concomitantly to this work, Martin’s group independently published an analogous CIE procedure suitable for aliphatic NHP redox active esters.[121] The direct exchange was performed using a substituted 2,2′-bipyridine ligand, which allowed a reduction in the catalyst loading to 10 mol %. The procedure was only performed with [13C]CO2, and no data on 14C were provided. In addition, the authors reported a 2-step full isotope replacement of alkyl and aryl carboxylic acids by merging decarboxylative halogenations with catalytic carboxylation of organic halides (Scheme ). Silver-catalyzed decarboxylation halogenation of aliphatic carboxylic acids[122] followed by nickel-catalyzed carboxylation enables isolating the corresponding labeled acids with high carbon isotope incorporation for a broad scope of substrates using 13C, including phenyl acetic acid [C]57. This method is also suitable for aryl acids. Compound 59 underwent Hunsdiecker-type decarboxylative bromination, according to the procedure reported by Larrosa.[123] Further Ni-catalyzed carboxylation delivered [C]59 with excellent molar activity.[124,125]

Scheme 18

Full Isotope Replacement of Carboxylic Acids

In 2020, our group described a transition-metal-free decarboxylation carboxylation sequence on phenylacetic acids (PAA, Scheme ).[126] The importance of PAA in medicinal chemistry is unique, as it is a main substructure representative of NSAIDs. Under thermal conditions, phenyl-acetate cesium salts were decarboxylated to generate benzylic anion intermediates (I). From this anionic species, two possible pathways are suggested: On one hand, the benzylic anion can directly trap a labeled molecule of CO2 leading to the expected compound. On the other hand, a dienolate moiety (II) might be generated first. This species can then undergo nucleophilic attack on labeled CO2 leading to the malonate derivative (III). Subsequent decarboxylation followed by protonation affords the expected labeled carboxylic acid. This procedure was applied with success to various drugs such as lonazolac 60, metiazinic acid 61 and fenclofenac 62 with 14C, and the molar activities in line with ADMET studies were obtained. In addition, this CIE was also applied for the first time to short-lived 11C.[127] The same year, Lundgren and co-workers published a very similar methodology starting from potassium salts. The procedure was only applied to 13C.[128] While the effectiveness of the thermal CIE of PAA is evident, the thermal conditions required for exchange are not ideal for radioactive gas handling.

Scheme 19

Transition-Metal-Free Carbon Isotope Exchange of Phenyl Acetic Acids

In order to provide a milder access to CIE of PAA, in 2021, our group reported a photochemical approach with [14C]CO2.[129] Lundgren’s group reported at the same time an analogous transformation for 13C-labeling, but no proof of concept for 14C was reported.[130] Utilizing 4-CzIPN as an organic photocatalyst, carbon isotopes were inserted onto various phenyl acetic acids without any prefunctionalization. Under basic conditions and blue LED light irradiation, the in situ formed carboxylate underwent photocatalytic oxidation followed by a rapid decarboxylation to generate a benzyl radical (Scheme ). The latter was further reduced to a benzyl carbanion through a second single-electron transfer step. Final carboxylation with a suitable source of CO2 (3 equiv) provided the expected labeled material. This procedure was applied to 14C-labeling of model substrate 63, which was isolated in 47% yield and 1.4 GBq mmol–1 activity (62% IE). This mild protocol was utilized in the labeling of ibuprofen, the ultimate NSAID. [C]64 was isolated in 29% yield and 53% IE, which is suitable for a large panel of biological applications.

Scheme 20

Photochemical Strategy for Carbon Isotope Exchange with CO2

CIE Using [14C]CO

In 2018, Gauthier and co-workers from Merck developed a palladium-catalyzed CIE of aliphatic and aromatic carboxylic acids starting from acyl chlorides using [14C]CO as an isotopic source.[131] The authors made the hypothesis that, after an oxidative addition of zerovalent palladium onto the acid chloride, an equilibration between the nonlabeled/labeled CO ligand might take place under optimized conditions. Final reductive elimination and hydrolysis of the solution would thus afford the expected labeled carboxylic acid (Scheme ).

Scheme 21

Palladium-Catalyzed CIE of Acid Chlorides with [14C]CO

Using this protocol, various exchanges on biologically active substrates were performed with carbon-13, and five examples with carbon-14 were reported. Radioactive compounds were isolated in 16–41% yield and with Am ranging from 0.6 to 1.0 GBq mmol–1. Representative examples are compounds 65 and 66. The advantage of this CIE procedure is its adaptability to both aliphatic and aromatic acids. The authors even showed that, under a specific optimization of selected substrates, the chirality in the alpha position to the carbonyl could be maintained. The downside is the necessity to use a secondary more expensive 14C source and the activation of the carboxylic acid to a hydrolysis-sensitive acid chloride intermediate.

CIE Using [14C]CN Salts

Until 2021, CIE procedures could only tackle carboxylic acids. While this functional group is representative of a large variety of biologically active pharmaceuticals and agrochemicals, it is far from ubiquitous. In order to expand the chemical space of CIE, our group and independently Reilly and Strotman simultaneously published in 2021 two similar procedures allowing CIE of aryl nitriles (Scheme ).[132,133] In the presence of nickel(0) catalyst, an η2-CN complex was formed prior to oxidative addition and metal insertion into the C–CN bond. Thus, a nitrile isotope ligand exchange occurred, and subsequent reductive elimination delivers the expected labeled aromatic nitrile. The procedure exhibited a broad scope including heteroaryl nitriles. Using Zn(CN)2 as a nitrile source, three pharmaceuticals were successfully labeled with enough isotope incorporation of carbon-14 radionuclide for potential applications in ADMET studies.

Scheme 22

Late-Stage CIE of Aryl Nitriles through Ni-Catalyzed C–CN Bond Activation

Conclusion

Carbon-14 has a long history of successful use in life science and particularly in the field of human health. Today, its use as a tracer remains a necessary step in the long process of bringing drugs to market. However, despite recent significant improvements, carbon-14 labeling remains difficult for a large number of compounds. Among recent advances, the strategy of isotopic exchange of carbon is certainly the most promising for future developments of carbon radiosynthesis. However, it is still limited to some specific functional groups, such as carboxylic acids and nitriles. Although these functions are common and present in pharmaceuticals, there is a need to extend this strategy to a wider range of chemical functions.

Besides CIE, efforts should be addressed to envision novel late-stage 14C-labeling methodologies for high-molar-activity tracers. For example, direct access to simple building blocks as 14C-labeled carbon monoxide, directly from [14C]CO2, under mild reaction conditions is still a challenge. Though this concept has been largely explored in electro- and photochemistry, the attention has thus far been essentially focused on the efficiency of the catalyst, while the effectiveness of the conversion of CO2 is not investigated, and a large excess of CO2 is most often utilized. Consequently, no methods such as these have been used in 14C radiosynthesis.

Long-lasting 14C-radioactive waste is extremely difficult to deal with and dispose of. We believe that new efforts should be focused on creative ways to recycle radioactive waste. In our opinion, it is past due to adapt the recently coined concept of green radiochemistry to carbon-14. While Scott’s group has conceptualized it for 11C,[134,135]11C waste does not represent a problem due to the short half-life. This is definitely not the case for 14C, and to meet today’s societal and environmental needs, there is a compelling urgency for the development of a green and more sustainable radiochemistry (GMSR).

Although these goals represent major challenges, they might provide a source of inspiration and motivation to develop new fundamental chemistry.