New methodologies for the preparation of carbon-11 labeled radiopharmaceuticals

Purpose This short review aims to cover the more recent and promising developments of carbon-11 (11C) labeling radiochemistry and its utility in the production of novel radiopharmaceuticals, with special emphasis on methods that have the greatest potential to be translated for clinical positron emission tomography (PET) imaging. Methods A survey of the literature was undertaken to identify articles focusing on methodological development in 11C chemistry and their use within novel radiopharmaceutical preparation. However, since 11C-labeling chemistry is such a narrow field of research, no systematic literature search was therefore feasible. The survey was further restricted to a specific timeframe (2000–2016) and articles in English. Results From the literature, it is clear that the majority of 11C-labeled radiopharmaceuticals prepared for clinical PET studies have been radiolabeled using the standard heteroatom methylation reaction. However, a number of methodologies have been developed in recent years, both from a technical and chemical point of view. Amongst these, two protocols may have the greatest potential to be widely adapted for the preparation of 11C-radiopharmaceuticals in a clinical setting. First, a novel method for the direct formation of 11C-labeled carbonyl groups, where organic bases are utilized as [11C]carbon dioxide-fixation agents. The second method of clinical importance is a low-pressure 11C-carbonylation technique that utilizes solvable xenon gas to effectively transfer and react [11C]carbon monoxide in a sealed reaction vessel. Both methods appear to be general and provide simple paths to 11C-labeled products. Conclusion Radiochemistry is the foundation of PET imaging which relies on the administration of a radiopharmaceutical. The demand for new radiopharmaceuticals for clinical PET imaging is increasing, and 11C-radiopharmaceuticals are especially important within clinical research and drug development. This review gives a comprehensive overview of the most noteworthy 11C-labeling methods with clinical relevance to the field of PET radiochemistry.


Introduction
Positron emission tomography (PET) is a highly sensitive imaging modality that can provide in vivo quantitative information of biological processes at a biochemical level [1]. PET relies upon the administration of a chemical probe, often called a radiopharmaceutical, that is labeled with a short-lived positron-emitting radionuclide [e.g. 11 C (t 1/2 = 20.4 min) and 18 F (t 1/2 = 109.7 min)]. Several PET radiopharmaceuticals have been developed for imaging applications, predominantly within oncology [2] and neuroscience [3,4]. The development of novel radiopharmaceuticals requires multiple considerations, where aspects like radionuclide selection, labeling position, metabolic stability, precursor synthesis, radiolabeling procedure, automation, quality control and regulatory affairs all have to be considered [5,6].
Carbon-11 is one of the most useful radionuclides for PET chemistry, since its introduction into a biologically active molecule has minimal effects on the (bio)chemical properties of the compound [7,8]. In addition, there is a vast literature on carbon-based chemistry that can be consulted in the development of radiosynthetic procedures with carbon-11. Moreover, the short half-life of 11 C allows for longitudinal in vivo studies with repeated injections in the same subject (patient or animal) and on the same experimental day. Although the advances in 11 C chemistry have enabled the preparation of a great number of radiolabeled molecules, there are still relatively few that have been applied for the direct preparation of novel radiopharmaceuticals for PET. The present review will provide an overview of the most recent and promising developments within carbon-11 chemistry since year 2000.

General considerations in radiopharmaceutical chemistry
A few general comments are required to provide a context for a discussion of PET radiopharmaceutical production [2,5]. First of all, the radionuclides used in PET emit highenergy radiation and, therefore, the traditional hands-on manipulations used in synthetic chemistry are not feasible. Thus, in order to avoid unnecessary radiation exposure, radiolabeling is performed in fully automated and preprogrammed synthesis modules housed inside lead-shielded fume hoods (hot-cells). One could say that radiochemistry, in particular that with 11 C, is a hybrid science between organic chemistry and engineering. Time is another factor of major importance in PET chemistry. A radiopharmaceutical used in PET is typically synthesized, purified, formulated and analyzed within a timeframe of roughly 2-3 physical half-lives of the employed radionuclide. For example, to obtain 11 C-labeled radiopharmaceutical in optimal radiochemical yield (RCY), a compromise has to be made between the chemical yield and the radioactive decay. The chemical yield of a reaction is thus not as important as the obtained radioactivity of the target compound at end of synthesis. Furthermore, since only trace amount of the radiolabeling synthon is used in PET, the amount of the non-radioactive reagents is in large excess, which implies that the reaction follows pseudo firstorder kinetics. By consequence, small impurities in reagents or solvents may have a significant influence on the reaction outcome. The radiochemist has to further consider the specific activity (SA), which is a measure of the radioactivity per unit mass of the final radiolabeled compound. Since high SA is often required in neuroreceptor imaging studies to avoid saturation of the receptor system, the methods that are highlighted in this review are all non-carrier-added nature.

Carbon-11 chemistry
Carbon-11 is commonly generated via the 14 N(p, a) 11 C nuclear reaction. The reaction is performed by high-energy proton bombardment of a cyclotron target containing nitrogen gas with small amounts a second gas. [ 11 C]Carbon dioxide ( 11 CO 2 ) and [ 11 C]methane ( 11 CH 4 ), are formed, when either small amounts oxygen or hydrogen are present in the cyclotron target. Sometimes, these simple primary precursors are used directly as labeling agents (e.g. 11 CO 2 ), but more often they are converted via on-line synthetic pathways into more reactive species before being used in 11 C-labeling reactions. However, several reactive 11 C-labeled precursors have been developed over the years [8], but the 11 C-precursors that will be discussed in this review are displayed and highlighted in Fig. 1.

C-methylation reaction
By far, the most common method in modern carbon-11 chemistry is heteroatom methylation using the methylating agents [ 11 C]methyl iodide ( 11 CH 3 I) [9,10] or [ 11 C]methyl triflate ( 11 CH 3 OTf) [11,12]. This reaction can either be performed using a traditional vial-based approach or alternatively using solid support (''on-cartridge'' [13] or ''in-loop'' [14] methods), which is very convenient from an automation prospective. A majority of the 11 C-labeled radiopharmaceuticals that are used on a regular basis, with a few exceptions, are thus produced by these two methylating agents. However, these methylating agents are sluggishly reactive towards arylamines. Especially difficult are substrates where the aryl group in a primary arylamine electron density has been further reduced by an electronwithdrawing group. In such situations, the more reactive methylating agent, 11 CH 3 OTf, may even fail to react. However, Pike and co-workers presented a method that utilized inorganic bases (e.g. Li 2 O) paired with DMF to permit methylation of a wide range of arylamines using 11 CH 3 I at room temperature [15]. Moreover, in a recent study, the research group of Billard described the application of 11 CO 2 as a C 1 building block for the catalytic methylation of amines [16]. Importantly, this one-pot approach eliminates the time-consuming preparation of the active methylating agent. The proposed mechanism is outlined in Table 1. In brief, an appropriate amine precursor, initially traps 11 CO 2 to form complex 1, which is reduced in two-steps with ZnCl 2 /iPr and PhSiH 3 , to furnish the expected methylamine. It was realized that the 11 CO 2 trapping was dependent on the basicity of the amine in use, varying between 65 and 80%. A large number of substrates, including the well-established radioligand, [ 11 C]PIB [17], was obtained in acceptable yields (Table 1).
In recent years, the application of 11 CH 3 I in transitionmetal-mediated reactions has become more widespread for 11 C-labeling of radiopharmaceuticals [18,19]. Figure 2 shows a brief overview of compounds labeled via Pd-mediated 11 C-methylation. Two radioligands for the serotonin transporter, [p-methyl- 11 C]MADAM [20] and [ 11 C]5methyl-5-nitroquipazine [21], as well as a novel radioligand for the nicotinic acetylcholine receptor [22] (nAChRs) was methylated using the transition-metal-mediated reaction. [ 11 C]A-85,380 displayed favorable in vivo properties for quantification of the nAChRs in living brain [23]. The nAChRs represents major neurotransmitter receptor responsible for various brain functions, and changes in the density of nAChRs have been reported in various neurodegenerative diseases, including Alzheimer's disease and Parkinson's disease [24]. Another example is the radiosynthesis of the 15R-[ 11 C]TIC methyl ester, a prostacyclin receptor radioligand, which was the first radioligand approved for investigation in humans [25,26]. The same radioligand was later used to image variations in organic anion-transporting polypeptide function in the human hepatobiliary transport system [27].
Two of the most applied cross-coupling reactions in radiochemical synthesis today are the Stille and Suzuki reactions, where organotin and organoborane compounds function as starting materials, respectively, and 11 CH 3 I as a coupling partner. A wide variety of functional groups such as amino, hydroxyl, or carboxylate are tolerated in these reactions and protective groups are usually not required. One unfortunate drawback with Stille coupling is, however, the inherent toxicity of the organotin reagent. Because of the regulatory aspects associated with radiopharmaceuticals that are to be used in human subjects, the less toxic organoborate substrates are usually preferred. As an alternative route to 11 C-methylated arenes, Kealey and coworkers describe a convenient two-step Pd-mediated crosscoupling of 11 CH 3 I with organozinc reagents (Scheme 1) [28]. The Nagishi-type reaction was used to synthesize a series of simple arenes in excellent yields. The same protocol was finally applied in the radiosynthesis of an mGluR5 radioligand, [ 11 C]MPEP [29]. Even though organozinc reagents are known to be moisture sensitive, it is much likely, that in the near future, Nagishi cross-coupling reaction will be considered a good alternative to the established protocols.
Enolates are a class of carbon centered nucleophiles that shortly may have a major importance in the radiopharmaceutical community. To generate an active enolate, a strong base, such as alkyl lithium or lithium diisopropylamide is typically needed. Using lithium bases to remove a a-proton is not always adequate because of their moisture sensitivity. However, in 2010, two methods for the synthesis of 11 C-labeled arylpropionic acid derivatives have been presented [30,31]. The rapid sp 3 -sp 3 11 C-methylation reaction relied on the formation of benzylic enolates, using either sodium hydride or tetrabutylammonium fluoride as base (Scheme 2). The reaction proceeds smoothly under mild conditions. However, until recently, the 11 C-methylated product formed under these conditions was obtained in low enantiomeric purity. The use of chiral phase-transfer catalyst has enabled enantioselective synthesis of the amino acid, [ 11 C]L-alanine, in high enantioselective purity [enantiomeric excess (ee) of 90%] [32]. Moreover, this year, our group presented a novel (carbonyl)cobalt-mediated, and microwave-assisted, carbonylative protocol for the direct preparation of 11 Clabeled aryl methyl ketones using 11 CH 3 I as the labeling agent [33]. The method uses CO 2 (CO) 8 as a combined aryl halide activator and carbon monoxide source for the carbonylation reaction. The method was used to label a set of functionalized (hetero)arenes with yields ranging from 22 to 63% (Scheme 3).

CO 2 -fixation reaction
[ 11 C]Carbon dioxide is in itself a highly attractive starting material for radiolabeling, since it is produced directly in the cyclotron. However, due to low chemical reactivity, the direct incorporation of CO 2 into organic molecules poses a significant challenge. High pressures, high temperatures or catalysts are commonly required to activate the molecule. The traditional method for 11 CO 2 ''fixation'' is the Grignard reaction, which involves the conversion of alkyl or aryl magnesium halides to [ 11 C]carboxylic acids. However, Grignard reagents require great care and the rigorous exclusion of atmospheric moisture and CO 2 during storage and manipulation. To overcome these limitations, two independent research groups presented what arguably can be viewed as the most ground-breaking advance in the field of carbon-11 chemistry since 11 CH 3 I was introduced in the early 1970s. The innovative method, that was inspired by the recent advances in ''green chemistry'' and reported in 2009, uses sub-milligram amounts of precursor compound, reacts at low temperature (typically room temperature), for 1-3 min reaction time and does not require advanced technical equipment [34,35]. To overcome the low reactivity of CO 2 , organic amines such as DBU or BEMP act as organomediators by activating CO 2 prior to the covalent bond formation [36,37]. The first report on 11 CO 2 fixation was on the synthesis of 11 Fig. 3). These includes, the carbonyl analogue radioligand of [ 11 C-methyl]AR-A014418, a compound developed for imaging of synthase kinase 3b (GSK-3b) [39]. However, unfortunately, the in vivo evaluations of AR-A014418 revealed an undesirably low brain uptake [40]. Moreover, two potent and irreversible fatty acid amide hydrolase (FAAH) inhibitors, [ 11 C]PF-04457845 [41] and [ 11 C]CURB [42], have also been prepared. The latter, [ 11 C]CURB, have recently been translated to a clinical setting for reginal quantification of FAAH activity in human brain [43]. Furthermore, the reversible monoamine oxidase B (MAO-B) radioligand, [ 11 C]SL25.1188, previously prepared using the technical demanding [ 11 C]phosgene approach, was radiolabeled in high yield via 11 CO 2 -fixation [44,45]. This radioligand was recently translated for human PET imaging [46]. Later, on a related subject, Dheere and co-workers presented a further refinement to the methodology to obtain [ 11 C]ureas from less reactive amines, such as anilines [47,48]. Once again, DBU was used to trap 11 CO 2 in solution, but in this case, it was shown that treatment of the carbamate anion intermediate (5) with Mitsunobu reagents, DBAD and PBu 3 , provided the corresponding asymmetric ureas in high radiochemical conversion (Scheme 5).
In the interest of expanding the scope of 11 CO 2 as a feedstock in radiochemical synthesis, copper-mediated approaches to carboxylic acids and their derivatives have been described [49,50]. In the most recent example, the combination of Cu(I) with boronic esters enabled CO 2 activation in the presence of a soluble fluoride additive and an organic base. In this reaction, the use of TMEDA was found to be crucial for obtaining high radiochemical yields, an observation likely explained by its dual action as both a trapping agent for 11 CO 2 and a ligand for the copper catalyst. A variety of functional groups were tolerated under optimized conditions, and the generated 11 C-carboxylic acids could be further converted into either amines or esters, as exemplified in the one-pot two-step preparation of a candidate radioligand for the oxytocin receptor. Scheme 4 Proposed pathways of 11 C-labeled urea and carbamates via 11 CO 2 -fixation chemistry [37] Carbonylation reactions using 11 CO [ 11 C]Carbon monoxide ( 11 CO) has many attractive features as a synthon for PET radiochemistry, including its facile production [51,52] and high versatility in transition-metal-mediated carbonylation reactions [53][54][55][56]. The widespread use of 11 CO in radiosynthetic chemistry was until recently hampered by its poor reactivity. Several solutions have been introduced to overcome the above shortcomings, both from a technical and chemical point of view [57][58][59][60]. A breakthrough was reported in 1999, where Kihlberg and co-workers introduced a method wherein 11 CO was allowed to react in a small autoclave under high solvent pressure ([350 Bar) [61]. The high-pressure reactor methodology exhibited nearly quantitative 11 CO trapping efficiency and high radiochemical yield. Even though this method has exemplified the importance of 11 CO as a labeling precursor it has not gained broad adoption in the PET radiochemistry community. This can partly be attributed to the overall complexity of the autoclave system and the relatively high level of service needed to maintain the system operational. Moreover, the repeated use of an integrated stainless steel reactor may infer issues related to transition metal build up over time, which is problematic in reaction development and system validation.
In recent years, the development of low-pressure techniques has been in focus. In 2012, an efficient protocol was reported by Eriksson and co-workers, in which 11 C-carbonylation reactions were achieved without the need for high-pressure equipment [62]. The high solubility of xenon gas in organic solvents was exploited as an effective way of transferring 11 CO into a sealed standard disposable reaction vial (1 ml) without significant pressure increase. The utility of the method was exemplified by 11 C-labeling of amides, ureas, and esters. The use of disposable glass reaction vessels eliminates carry over issues associated with the high-pressure autoclave system, thus simplifying the transition to clinical applications. Recently, three reports were published using the same 11 CO transfer protocol (Fig. 4). Windhorst et al. produced three 11 C-labeled acryl amide radioligands for in vivo PET imaging of the tissue transglutaminase (TG2) enzyme [63]. Moreover, with the ''xenon-method'' as the synthesis platform, the Uppsalagroup presented two novel approaches to 11 C-carbonyl labeled compounds. Firstly, a new multicomponent reaction for 11 C-labeling of sulfonyl carbamates was described [64]. The method was further applied as a synthetic tool for the in vivo evaluation of an angiotensin II receptor subtype 2 (AT 2 R) agonist. Secondly, a method to access 11 C-labeled alkyl amides via a thermally-initiated radical reductive dehalogenative approach [65]. One of the restrictions of transition-metal-mediated reactions is the competing bhydride elimination of the resulting metal-substrate intermediate, which precludes the use of alkyl electrophiles containing b-hydrogen. This unfortunate competing reaction is non-excitant for radical pathways. A series of unactivated alkyl iodides was successfully converted into the corresponding alkyl amide in good RCY, including the radiosynthesis of an 11b-HSD1 inhibitor. Specific Pd-complexes have also been shown to trap 11 CO at ambient pressure and without the need for any high-pressure equipment [66,67]. XantPhos, a hindered bidentate phosphine ligand, in combination with palladium (l-cinnamyl) chloride dimer were found to be excellent for promoting 11 C-carbonylation reactions. Notably, in this study it was discovered that, depending on the palladiumligand complex in use, different 11 CO trapping efficiency  Table 2). The reaction proceeds smoothly at close to atmospheric pressure with aryl halides or triflates as substrates using simple disposable glass vials. This method was recently also applied in the preparation of well-known D 2 radioligand, [ 11 C]raclopride, but with the carbon-11 labeled in the more metabolically stable carbonyl group (Fig. 4) [68]. Interestingly, in a direct comparison between ([ 11 C]methyl)raclopride (produced using the standard 11 C-methylation approach) and ([ 11 C]carbonyl)raclopride, both radioligands showed similar in vivo properties with regards to quantitative outcome measurements, radiometabolite formation and protein binding.
The protocol was further improved by Andersen and coworkers, where pre-generated (Aryl)Pd(I)L n oxidative addition complexes were utilized as precursors for the following 11 C-carbonylation reaction [69]. This is exemplified in the preparation of [ 11 Ccarbonyl]raclopride in Scheme 6. The isolated complexes, (Aryl)Pd(I)L n , have already undergone the potentially challenging oxidative addition step before their employment in carbonylative 11 C-labeling. In this case, Pd-XantPhos complexes appeared to be among the most reactive precursors, although, electron-deficient aryl precursors demanded Pd-P(t-Bu) 3 to prevent aryl scrambling with phosphine ligand. The simplicity of these low-pressure techniques, and especially the ''xenon-method'' delivery protocol, may offer a potential for being widely adopted in radiopharmaceutical research and development.
As mentioned previously, one restriction with transitionmetal-mediated reactions is the competing b-hydride elimination. However, in contrast to Pd or Rh catalyst, nickel has been known to suppress the b-hydride elimination reaction. In the light of this, Rahman and co-workers recently reported the first successful use of nickel-mediated carbonylative cross-coupling of non-activated alkyl iodides using 11 CO at ambient pressure [70]. The best conditions identified in this study made use of a nickel(0) precatalyst, Ni(COD) 2 , in the presence of bathophenantroline as ligand (Table 3). Six model compounds were successfully radiolabeled in acceptable to good yields. However, more data is required to establish if the method is suitable of preparing more complex molecules.

Other recent advances in carbon-11 chemistry
Hydrogen cyanide is well established as a versatile precursor in PET radiopharmaceutical chemistry [71][72][73], and its involvement in metal-mediated cyanation of aryl Scheme 6 Labeling based on pre-isolated (Aryl)Pd(I)L n complexes with 11 CO Table 3 Nickel-mediated 11 C-aminocarbonylation of iodoalkyl compounds (pseudo)halides is well documented [74,75]. A limitation of such reactions is that they require rather harsh conditions, such as high temperature, long reaction times, and inorganic bases (e.g. KOH), which reduces the substrate scope. Recently, a novel method was reported describing near instantaneous, room temperature Pd-mediated coupling of [ 11 C]HCN to aryl halides or triflates [76]. The method is based on sterically hindered biaryl phosphine ligands ( Table 4) that facilitate rapid transmetalation with [ 11 C]HCN and reductive elimination of aryl nitriles at ambient temperature. A wide variety of (hetero)arenes and drug-like molecules were radiolabeled in high yields, including the j-opioid receptor radioligand [ 11 C]LY2795050 (Fig. 5). Moreover, two known antidepressants were also 11 C-labeled in this study. This further illustrates the usefulness of the current method in the preparation of radiopharmaceuticals.
Carbon disulfide, the sulfur analogue of carbon dioxide, has recently been synthesized on-line from 11 CH 3 I using either P 2 S 5 and elemental sulfur (S 8 ) at elevated temperatures in excellent yields [77,78]. Due to the weaker C=S bond CS 2 is considered more reactive than CO 2 . Moreover, CS 2 reacts very rapidly with many primary amines at room temperature to form the dithiocarbamate salts, which in turn, upon treatment with a suitable alkylating reagent will give the corresponding thiocarbamates. Heating, on the other hand, induces rearrangement to form the symmetrical thiourea (Scheme 7). Some model compounds were radiolabeled using this protocol in near quantitative yields. Finally, a progesterone receptor agonist, Tanaproget, was also produced in high RCY (Fig. 5).
Lastly, an improved, mild synthesis of [ 11 C]formaldehyde have opened up new to carbon-11 labeled radiopharmaceuticals [79]. The treatment of trimethylamine N-oxide  [80,81]. In addition, since [ 11 C]formaldehyde was reported already in 1972 [82], there are other molecules previously reported in the literature that may now be synthesized in a simplified fashion using this protocol.

Final remarks
The increasing importance of PET in drug development and clinical research has motivated researchers to initiate programs directly dedicated to development of new radiolabeling methods. This review summarizes some of the most recent and promising strategies to obtain carbon-11 labeled products. In the past two decades, and well before this, efforts have brought to bear an impressive range of methods for 11 C-radiochemistry. However, there are still issues to be addressed. Take for example, the heteroatom 11 C-methylation reaction, which is now considered as an established method by the broader radiochemical community. Why is this? The main reason is the access to dedicated commercially available radiochemical equipment for this radiochemistry. Consequently, to streamline new methodologies, and make them widely available, new radiochemical equipment is needed. A possible approach to attack the problem could be to develop radiosynthesis equipment with a higher flexibility. A fundamental question is if microscale technology (microfluidic or microreactor) can provide a breakthrough in radiochemistry? Its compact design, flexible attributes, and its suitability for automation make microscale technology an ideal platform for performing the rapid radiolabeling reactions required for PET. So far, efforts made to adapt microscale technology for PET radiolabeling purposes have focused on proof-of-principle studies and to illustrate the advantages associated with the technology and significant further development is needed for the technology to reach its full potential. Although the authors recognize the importance of microreactor technologies, other technical approaches towards the development of more flexible radiochemical synthesis equipment are equally attractive at this point. Regardless of which direction is taken in the future, we firmly believe that a stronger collaboration between radiochemists and technical engineers is vital for succeeding in the development of the next generation of PET radiochemistry equipment. Finally, radiochemistry is the foundation for PET imaging. By broadening the spectrum of radiochemical reactions within clinical PET radiochemistry, radiochemists will not only be able to increase the number of compounds that can be labeled with carbon-11 but also provide an increased opportunity to label a given compound in different positions.
Acknowledgements The authors thank members of the PET group at Karolinska Institutet.

Compliance with ethical standards
Conflict of interest KD declares that he has no conflict of interest. CH declares that he has no conflict of interest. MS declares the he is an employee and shareholder at AstraZeneca Pharmaceuticals.
Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://crea tivecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.