Isotopic Radiolabeling of the Antiretroviral Drug [18F]Dolutegravir for Pharmacokinetic PET Imaging

Deciphering the drug/virus/host interactions at infected cell reservoirs is a key leading to HIV-1 remission for which positron emission tomography (PET) imaging using radiolabeled antiretroviral (ARV) drugs is a powerful asset. Dolutegravir (DTG) is one of the preferred therapeutic options to treat HIV and can be isotopically labeled with fluorine-18. [18F]DTG was synthesized via a three-step approach of radiofluorination/nitrile reduction/peptide coupling with optimization for each step. Radiofluorination was performed on 2-fluoro-4-nitrobenzonitrile in 90% conversion followed by nitrile reduction using sodium borohydride and aqueous nickel(II) chloride with 72% conversion. Final peptide coupling reaction followed by HPLC purification and formulation afforded ready-to-inject [18F]DTG in 5.1 ± 0.8% (n = 10) decay-corrected radiochemical yield within 95 min. The whole process was automatized using a TRACERlab® FX NPro module, and quality control performed by analytical HPLC showed that [18F]DTG was suitable for in vivo injection with >99% chemical and radiochemical purity and a molar activity of 83 ± 18 GBq/µmol (n = 10). Whole-body distribution of [18F]DTG was performed by PET imaging on a healthy macaque and highlighted the elimination routes of the tracer. This study demonstrated the feasibility of in vivo [18F]DTG PET imaging and paved the way to explore drug/virus/tissues interactions in animals and humans.


Introduction
Increasing evidence makes remission an achievable goal for new human immunodeficiency virus (HIV-1) therapies [1][2][3]. Combined antiretroviral therapies (cART) suppress viral replication and drastically reduce morbidity and mortality [4]. However, despite the existence of a few patients capable of controlling HIV replication to undetectable levels (i.e., HIV controllers, <1% of the HIV-infected population) [2], cART do not eradicate infected cells, and viral reservoirs still remain [5]. Early control of these reservoirs together with the limitation of pharmacological sanctuaries where the exposure to cART is suboptimal [6] are among the key scientific challenges to make remission a reality for a majority of patients.
The whole-body distribution of the virus is a key parameter to predict the virus/host/ drugs interactions. Indeed, localization of infected cell reservoirs and their interactions with both drugs and the tissue micro-environment remain poorly understood [7]. In this context, in vivo tomographic imaging technologies may be of considerable help to tackle anti-HIV therapy challenges [8][9][10]. In particular, positron emission tomography (PET) imaging is

Late-Stage Radiolabeling Approach
The most efficient strategy to radiolabel DTG with fluorine-18 would be a late-stage approach consisting of the incorporation of fluorine-18 at the final step of the synthesis on the aromatic core of DTG bearing two fluorine atoms (Scheme 1). However, this aromatic core is electron-rich, hampering radiofluorination by standard aromatic nucleophilic substitution (Sn Ar ). Alternative approaches have been described using iodonium salts [21][22][23] or ylides [24][25][26] as precursors or by copper-catalyzed Cham-Lam-like fluorination using boronic acid [27,28], boronic ester [29][30][31][32], or trialkyltin precursors [33,34]. Late-stage isotopic radiolabeling of DTG was considered at the para position of the aromatic core (compared to the methyl amino group), as better radiochemical yields were depicted on this position compared to ortho [27,29,31,33]. Limitations of these fluorination methods were also described in the presence of unprotected alcohol or amine [29,31]. Therefore, the hydroxyl function of DTG was protected with a methyl moiety, the cleavage of such function being well described for DTG [35]. Labeling precursors for late-stage radiofluorination of DTG, i.e., iodonium ylide, boronic acid, and ester and tributyltin derivatives were synthesized in two steps (Scheme S1 in Supplementary Materials). Unfortunately, even if the boronic ester precursor was observed by liquid chromatography coupled with mass spectroscopy (LC-MS), it was too unstable to be isolated. The ylide precursor was synthesized but could not be isolated from its starting material and was unstable at 120 °C, whereas a minimal temperature of 150 °C is required to perform radiofluorination with this type of precursor [24,25]. Several attempts of copper-catalyzed radiofluorination using the tributyltin precursor were realized under the standard conditions described in the literature using the Cu(OTf)4Py4 complex [33], but incorporation of fluorine-18 was never observed (Scheme S2 in Supplementary Materials).

Three-Step Radiolabeling Approach
Considering the difficulties encountered for the late-stage labeling of DTG, the strategy described for raltegravir radiosynthesis was explored [36,37]. Mimicking the recent synthesis of dolutegravir described by Dietz et al. [38], [ 18 F]DTG can be synthesized by peptide coupling between carboxylic acid 3 and radiolabeled difluorobenzylamine [ 18 F]2 (Scheme 2). Although the radiosynthesis of [ 18 F]2 has never been reported, p-[ 18 F]fluorobenzylamine has been extensively described in the literature [39]. Following a similar approach, [ 18 F]2 could be synthesized by reduction of a nitrile moiety from [ 18 F]1 which could be obtained by radiofluorination of 2-fluoro-4-nitrobenzonitrile. This approach will benefit from the poor electron density of 2-fluoro-4-nitrobenzonitrile due to the nitrile moiety at the para position of the nitro group, which allows for the introduction of the fluorine-18 ion via a standard SnAr reaction.  [39]. A kinetic monitoring of the reaction using thin layer chromatography (TLC) analysis on silica gel (see Figure S12 in Supplementary Materials) revealed that performing the radiofluorination in acetonitrile at 95 °C for 5 min was the best compromise to reach a high conversion into [ 18 F]1 (96%) in a short time (Table 1, entry 1). However, acetonitrile, even at trace quantities, could be deleterious for the following step of nitrile reduction, and DMSO is usually preferred [40]. Radiofluorination in DMSO afforded [ 18 F]1 in 90% conversion (Table 1, entry 4), comparable with acetonitrile under the same conditions. Labeling precursors for late-stage radiofluorination of DTG, i.e., iodonium ylide, boronic acid, and ester and tributyltin derivatives were synthesized in two steps (Scheme S1 in Supplementary Materials). Unfortunately, even if the boronic ester precursor was observed by liquid chromatography coupled with mass spectroscopy (LC-MS), it was too unstable to be isolated. The ylide precursor was synthesized but could not be isolated from its starting material and was unstable at 120 • C, whereas a minimal temperature of 150 • C is required to perform radiofluorination with this type of precursor [24,25]. Several attempts of copper-catalyzed radiofluorination using the tributyltin precursor were realized under the standard conditions described in the literature using the Cu(OTf) 4 Py 4 complex [33], but incorporation of fluorine-18 was never observed (Scheme S2 in Supplementary Materials).

Three-Step Radiolabeling Approach
Considering the difficulties encountered for the late-stage labeling of DTG, the strategy described for raltegravir radiosynthesis was explored [36,37]. Mimicking the recent synthesis of dolutegravir described by Dietz et al. [38], [ 18 F]DTG can be synthesized by peptide coupling between carboxylic acid 3 and radiolabeled difluorobenzylamine [ 18 F]2 (Scheme 2). Although the radiosynthesis of [ 18 F]2 has never been reported, p-[ 18 F]fluorobenzylamine has been extensively described in the literature [39]. Following a similar approach, [ 18 F]2 could be synthesized by reduction of a nitrile moiety from [ 18 F]1 which could be obtained by radiofluorination of 2-fluoro-4-nitrobenzonitrile. This approach will benefit from the poor electron density of 2-fluoro-4-nitrobenzonitrile due to the nitrile moiety at the para position of the nitro group, which allows for the introduction of the fluorine-18 ion via a standard Sn Ar reaction. Labeling precursors for late-stage radiofluorination of DTG, i.e., iodonium ylide, boronic acid, and ester and tributyltin derivatives were synthesized in two steps (Scheme S1 in Supplementary Materials). Unfortunately, even if the boronic ester precursor was observed by liquid chromatography coupled with mass spectroscopy (LC-MS), it was too unstable to be isolated. The ylide precursor was synthesized but could not be isolated from its starting material and was unstable at 120 °C, whereas a minimal temperature of 150 °C is required to perform radiofluorination with this type of precursor [24,25]. Several attempts of copper-catalyzed radiofluorination using the tributyltin precursor were realized under the standard conditions described in the literature using the Cu(OTf)4Py4 complex [33], but incorporation of fluorine-18 was never observed (Scheme S2 in Supplementary Materials).

Three-Step Radiolabeling Approach
Considering the difficulties encountered for the late-stage labeling of DTG, the strategy described for raltegravir radiosynthesis was explored [36,37]. Mimicking the recent synthesis of dolutegravir described by Dietz et al. [38], [ 18 F]DTG can be synthesized by peptide coupling between carboxylic acid 3 and radiolabeled difluorobenzylamine [ 18 F]2 (Scheme 2). Although the radiosynthesis of [ 18 F]2 has never been reported, p-[ 18 F]fluorobenzylamine has been extensively described in the literature [39]. Following a similar approach, [ 18 F]2 could be synthesized by reduction of a nitrile moiety from [ 18 F]1 which could be obtained by radiofluorination of 2-fluoro-4-nitrobenzonitrile. This approach will benefit from the poor electron density of 2-fluoro-4-nitrobenzonitrile due to the nitrile moiety at the para position of the nitro group, which allows for the introduction of the fluorine-18 ion via a standard SnAr reaction.

Radiofluorination Optimization
Radiofluorination of 2-fluoro-4-nitrobenzonitrile was realized by SnAr under standard conditions using the K [ 18 F]F/K222 complex in acetonitrile (Scheme 3) [39]. A kinetic monitoring of the reaction using thin layer chromatography (TLC) analysis on silica gel (see Figure S12 in Supplementary Materials) revealed that performing the radiofluorination in acetonitrile at 95 °C for 5 min was the best compromise to reach a high conversion into [ 18 F]1 (96%) in a short time ( Table 1, entry 1). However, acetonitrile, even at trace quantities, could be deleterious for the following step of nitrile reduction, and DMSO is usually preferred [40]. Radiofluorination in DMSO afforded [ 18

Radiofluorination Optimization
Radiofluorination of 2-fluoro-4-nitrobenzonitrile was realized by Sn Ar under standard conditions using the K [ 18 F]F/K 222 complex in acetonitrile (Scheme 3) [39]. A kinetic monitoring of the reaction using thin layer chromatography (TLC) analysis on silica gel (see Figure S12 in Supplementary Materials) revealed that performing the radiofluorination in acetonitrile at 95 • C for 5 min was the best compromise to reach a high conversion into [ 18 F]1 (96%) in a short time ( Table 1, entry 1). However, acetonitrile, even at trace quantities, could be deleterious for the following step of nitrile reduction, and DMSO is usually preferred [40]. Radiofluorination in DMSO afforded [ 18  Borane complexes are also standard reagents for nitrile reduction [43] which have been used for the reduction of [ 18 F]fluorobenzonitrile [39]. In our case, reduction of [ 18 F]1 at either room temperature or 65 °C resulted in only 5% of [ 18 F]2 and mostly degradation of the starting material ( Table 2, entries 3 and 4, respectively). Reduction of nitrile moieties by hydrogenation in the presence of palladium on charcoal has also been extensively described in the literature [44], although it has never been used for radiochemical reactions, due to the heterogeneity of the medium. In our case, hydrogenation of compound [ 18 F]1 in methanol at room temperature only resulted in the formation of compound [ 18 F]2′ with 30% conversion (Table 2, entry 5). Indeed, palladium-catalyzed hydrodefluorination has been reported in the case of poly-fluoro arenes [45]. Considering the sensitivity of [ 18 F]1 to defluorination and the significant proportion of side products formed using standard reducing agents, milder conditions were explored. Werkmeister et al. have described an original method to reduce nitrile by hydrogenation under mild conditions using [RuCl2(pCymene)2]2 and 3-bis(diphenylphosphino)propane (DPPP) as the catalytic system and isopropanol as the hydrogen donor [46]. Under those conditions, no side reactions were observed, but [ 18 F]2 was obtained with only 10% conversion ( Table 2, entry 6). The in situ generation of metal boride using NaBH4 in the presence of transition metal (Co, Ni) and water is an efficient reducing agent to promote nitrile reduction [47]. This catalytic system has already been described for the reduction of [ 18 F]fluorobenzonitrile to [ 18 F]fluorobenzylamine [40,48,49]. Using excess amounts of NaBH4 and hexahydrated   Figure S13 in Supplementary Materials for radio HPLC examples). Moreover, [ 18 F]2 will be able to react in the next step of peptide coupling in the same way as [ 18 F]2. Therefore, a compromise has to be found between maximizing the conversion into [ 18 F]2 and minimizing [ 18 F]2 formation. In addition, results with LAH were not reproducible, and ratios observed between [ 18 F]2 and [ 18 F]2 were not constant. Therefore, other reducing agents were explored. Borane complexes are also standard reagents for nitrile reduction [43] which have been used for the reduction of [ 18 F]fluorobenzonitrile [39]. In our case, reduction of [ 18 F]1 at either room temperature or 65 • C resulted in only 5% of [ 18 F]2 and mostly degradation of the starting material ( Table 2, entries 3 and 4, respectively). Reduction of nitrile moieties by hydrogenation in the presence of palladium on charcoal has also been extensively described in the literature [44], although it has never been used for radiochemical reactions, due to the heterogeneity of the medium. In our case, hydrogenation of compound [ 18 F]1 in methanol at room temperature only resulted in the formation of compound [ 18 F]2 with 30% conversion (Table 2, entry 5). Indeed, palladium-catalyzed hydrodefluorination has been reported in the case of poly-fluoro arenes [45]. Considering the sensitivity of [ 18 F]1 to defluorination and the significant proportion of side products formed using standard reducing agents, milder conditions were explored. Werkmeister et al. have described an original method to reduce nitrile by hydrogenation under mild conditions using [RuCl 2 (pCymene) 2 ] 2 and 3-bis(diphenylphosphino)propane (DPPP) as the catalytic system and isopropanol as the hydrogen donor [46]. Under those conditions, no side reactions were observed, but [ 18 F]2 was obtained with only 10% conversion ( Table 2, entry 6). The in situ generation of metal boride using NaBH 4 in the presence of transition metal (Co, Ni) and water is an efficient reducing agent to promote nitrile reduction [47]. This catalytic system has already been described for the reduction of [ 18 [48]. In our hands, almost no reaction was observed ( Table 2, entry 8).  [48]. In our hands, almost no reaction was observed ( Table 2, entry 8). Clearly the homogeneous nitrile reduction using NaBH4 and hexahydrated NiCl2 was the best compromise to achieve a moderate conversion into the desired benzylamine [ 18 F]2 while minimizing the formation of side products. In order to further improve the conversion into [ 18 F]2, different parameters (quantity of reagents, solvent, time, and temperature) were studied (Table 3). THF was also explored to realize this reduction, but conversion into [ 18 F]2 was always lower than in methanol (data not shown). Decreasing the quantity of NaBH4 resulted in a total conversion of the starting material [ 18 F]1 but significantly increased the proportion of side products compared to the desired compound [ 18 F]2 ( Table 3, entry 2). Increasing the time of reaction to 10 min (Table 3, entry 3) or the temperature to 50 °C (Table 3, entry 4) did not significantly increase the conversion into [ 18 F]2. Additional water (100 μL) was introduced to the reaction to promote the formation of the Ni2B reducing agent and improve the homogeneity of the reaction ( Clearly the homogeneous nitrile reduction using NaBH 4 and hexahydrated NiCl 2 was the best compromise to achieve a moderate conversion into the desired benzylamine [ 18 F]2 while minimizing the formation of side products. In order to further improve the conversion into [ 18 F]2, different parameters (quantity of reagents, solvent, time, and temperature) were studied (Table 3). THF was also explored to realize this reduction, but conversion into [ 18 F]2 was always lower than in methanol (data not shown). Decreasing the quantity of NaBH 4 resulted in a total conversion of the starting material [ 18 F]1 but significantly increased the proportion of side products compared to the desired compound [ 18 F]2 (Table 3, entry 2). Increasing the time of reaction to 10 min (  (Table 3, entry 11). Overall, the best compromise to optimize the conversion into [ 18 F]2 (72%) while minimizing the formation of deleterious side products (28%) for the next reaction step was found when using a rather low quantity of NaBH 4 (40 µmol) and a large excess of hexahydrated nickel chloride with additional water ( Table 3, entry 8). 1 Experiments were realized in duplicate. 2 The conversion rate was measured by radio-HPLC as the ratio of formed [ 18 F]2 over the total radioactivity. The conversion is the mean value of two experiments. 3 The side products represent the combination of [ 18 F]2 and other unidentified compounds formed during the reduction process. Best conditions were highlighted in red.

Peptide Coupling Optimization
Inspired by the recent work of Dietz and colleagues [38], [ 18 F]DTG was finally obtained in a last step of peptide coupling between [ 18 F]2 and carboxylic acid 3 using 2-(1Hbenzotriazole-1-yl)-1,1,3,3-tetramethylaminium tetrafluoroborate (TBTU) as coupling agent and diisopropylethylamine (DIPEA) in DMF (Scheme 4). Compound 3 was synthesized in one-step from compound GSK3210346A [50] (see Scheme S3 in Supplementary Materials). In those conditions, [ 18 F]DTG was obtained in 30% conversion after 10 min of reaction at room temperature. Increasing the temperature or the reaction time did not result in any significant improvement of the conversion.
found when using a rather low quantity of NaBH4 (40 μmol) and a large excess of hexahydrated nickel chloride with additional water (  1 Experiments were realized in duplicate. 2 The conversion rate was measured by radio-HPLC as the ratio of formed [ 18 F]2 over the total radioactivity. The conversion is the mean value of two experiments. 3 The side products represent the combination of [ 18 F]2′ and other unidentified compounds formed during the reduction process. Best conditions were highlighted in red.

Peptide Coupling Optimization
Inspired by the recent work of Dietz and colleagues [38], [ 18 F]DTG was finally obtained in a last step of peptide coupling between [ 18 F]2 and carboxylic acid 3 using 2-(1Hbenzotriazole-1-yl)-1,1,3,3-tetramethylaminium tetrafluoroborate (TBTU) as coupling agent and diisopropylethylamine (DIPEA) in DMF (Scheme 4). Compound 3 was synthesized in one-step from compound GSK3210346A [50] (see scheme S3 in Supplementary Materials). In those conditions, [ 18 F]DTG was obtained in 30% conversion after 10 min of reaction at room temperature. Increasing the temperature or the reaction time did not result in any significant improvement of the conversion.

Automated Radiosynthesis of [ 18 F]DTG
For in vivo PET imaging applications, the radiosynthesis of [ 18 F]DTG has to be automated for reproducibility and radioprotection reasons. With optimized conditions for each step in hand, the automation was carried out using a TRACERlab ® FX N Pro (GE Healthcare, Chicago, IL, USA) module ( Figure 1). The presence of two separated reactors on the module was mandatory, as the reaction conditions, the solvents in particular, between the fluorination and the reduction steps are not compatible. Moreover, unreacted fluoride ions could promote the decomposition of sodium borohydride [51]. In contrast, the reduction and the peptide coupling reaction could be realized in one pot in the second reactor. After using a standard procedure for the preparation of the K [ 18 F]F/K222 complex

Automated Radiosynthesis of [ 18 F]DTG
For in vivo PET imaging applications, the radiosynthesis of [ 18 F]DTG has to be automated for reproducibility and radioprotection reasons. With optimized conditions for each step in hand, the automation was carried out using a TRACERlab ® FX N Pro (GE Healthcare, Chicago, IL, USA) module ( Figure 1). The presence of two separated reactors on the module was mandatory, as the reaction conditions, the solvents in particular, between the fluorination and the reduction steps are not compatible. Moreover, unreacted fluoride ions could promote the decomposition of sodium borohydride [51]. In contrast, the reduction and the peptide coupling reaction could be realized in one pot in the second reactor. After using a standard procedure for the preparation of the K [ 18 F]F/K 222 complex with azeotropic drying, the radiofluorination step was realized in reactor 1 following the optimized conditions depicted above. In order to remove unreacted fluoride ions before the reduction step, the crude mixture from radiofluorination was pre-purified by solid-phase extraction (SPE) on a C18 cartridge. This step also achieved a solvent exchange from DMSO to methanol for the reduction step. The crude [ 18 F]1 was eluted with methanol from the C18 cartridge to the reactor 2 to avoid any contamination for the reduction step from possible residual DMSO or fluoride ions in reactor 1. To realize the reduction step, the sodium borohydride solution in methanol and the nickel chloride hexahydrate solution in water were added from separate vials into the reactor 2. Mixing the two solutions beforehand resulted in the degradation of the formed Ni 2 B reducing agent and a poor conversion into [ 18 F]2. The coupling reaction was then performed in one pot by adding a mixture of compound 3, TBTU, and DIPEA in DMF directly into reactor 2 at the end of the reduction step. At the end of the radiosynthesis, purification of the crude reaction mixture was realized by semi-preparative HPLC on a Symmetry C18 column (Waters) using a mixture of acetonitrile, water, and trifluoroacetic acid (TFA) as eluent. [ 18 F]DTG was isolated with a retention time of about 17 min (see Figure S17 in Supplementary Materials). Final formulation by SPE on a C18 cartridge afforded ready-to-inject [ 18 F]DTG in 95 min and 5.1 ± 0.8% (n = 10) decay-corrected radiochemical yield. This yield is comparable to the similar radiosynthesis of [ 18 F]Raltegravir (4.61 ± 0.3% within 135 min) described by Blecha et al. [37]. The quantity of [ 18 F]DTG produced was in the range 500-700 MBq, consistent with administration to humans (typically in the range 200-400 MBq).
DMSO to methanol for the reduction step. The crude [ 18 F]1 was eluted with methanol from the C18 cartridge to the reactor 2 to avoid any contamination for the reduction step from possible residual DMSO or fluoride ions in reactor 1. To realize the reduction step, the sodium borohydride solution in methanol and the nickel chloride hexahydrate solution in water were added from separate vials into the reactor 2. Mixing the two solutions beforehand resulted in the degradation of the formed Ni2B reducing agent and a poor conversion into [ 18 F]2. The coupling reaction was then performed in one pot by adding a mixture of compound 3, TBTU, and DIPEA in DMF directly into reactor 2 at the end of the reduction step. At the end of the radiosynthesis, purification of the crude reaction mixture was realized by semi-preparative HPLC on a Symmetry C18 column (Waters) using a mixture of acetonitrile, water, and trifluoroacetic acid (TFA) as eluent. [ 18 F]DTG was isolated with a retention time of about 17 min (see Figure S17 in Supplementary Materials). Final formulation by SPE on a C18 cartridge afforded ready-to-inject [ 18 F]DTG in 95 min and 5.1 ± 0.8% (n = 10) decay-corrected radiochemical yield. This yield is comparable to the similar radiosynthesis of [ 18 F]Raltegravir (4.61 ± 0.3% within 135 min) described by Blecha et al. [37]. The quantity of [ 18 F]DTG produced was in the range 500-700 MBq, consistent with administration to humans (typically in the range 200-400 MBq).

Quality Control
Quality control was performed on [ 18 F]DTG by analytical HPLC with both UV (258 nm) and gamma detection to confirm the identity of the radiotracer and its chemical and radiochemical purity and to assess the molar activity (MA) of the radiotracer (Figure 2). The identity of the radiotracer was confirmed by the retention time of [ 18 F]DTG (t R = 3.74 min, Figure 2C), which was within the range t R ref ± 10% compared to the retention time of the DTG reference (t R ref = 3.60 min., Figure 2A). [ 18 F]DTG was obtained with both a radiochemical and chemical purity above 99% ( Figure 2B and C respectively). [ 18 F]DTG was obtained with a MA of 83 ± 18 GBq/µmol (n = 10) calculated with the help of a calibration curve (see Figure S18 in Supplementary Materials). In summary, the quality control of [ 18 F]DTG demonstrated that the radiotracer was suitable for in vivo injection to perform PET imaging. Pending additional tests, [ 18 F]DTG would be suitable for human injection according to guidance from the European Pharmacopeia [52]. the DTG reference (tR ref = 3.60 min., Figure 2A). [ 18 F]DTG was obtained with both a radiochemical and chemical purity above 99% ( Figure 2B and C respectively). [ 18 F]DTG was obtained with a MA of 83 ± 18 GBq/μmol (n = 10) calculated with the help of a calibration curve (see Figure S18 in Supplementary Materials). In summary, the quality control of [ 18 F]DTG demonstrated that the radiotracer was suitable for in vivo injection to perform PET imaging. Pending additional tests, [ 18 F]DTG would be suitable for human injection according to guidance from the European Pharmacopeia [52].

PET Imaging
Feasibility of in vivo PET imaging using [ 18 F]DTG was first assessed in a non-human primate. PET images unveiled the whole-body distribution of [ 18 F]DTG (Figure 3). Highest PET signal was observed in the liver (SUV = 6.06 ± 0.14), kidneys (SUV = 6.45 ± 0.31), gallbladder (SUV = 12.16 ± 1.26), and urinary bladder (SUV = 13.73 ± 0.97), consistent with the known elimination routes of DTG [53]. PET signal in the brain was strikingly low (SUV = 0.04 ± 0.001) compared with peripheral organs, suggesting limited brain penetration of [ 18 F]DTG. Minimally invasive [ 18 F]DTG PET, therefore, sheds light on the elimination of DTG and its distribution to tissues. Interestingly, it suggests a minimal distribution to the brain which could correspond to a pharmacological sanctuary. Previous studies suggest a low brain distribution of [ 18 F]DTG in mice [54]. Further experiments with dynamic brain PET imaging and kinetic modelling will however be necessary to correctly estimate the brain distribution of [ 18 F]DTG in macaques.

PET Imaging
Feasibility of in vivo PET imaging using [ 18 F]DTG was first assessed in a non-human primate. PET images unveiled the whole-body distribution of [ 18 F]DTG (Figure 3). Highest PET signal was observed in the liver (SUV = 6.06 ± 0.14), kidneys (SUV = 6.45 ± 0.31), gallbladder (SUV = 12.16 ± 1.26), and urinary bladder (SUV = 13.73 ± 0.97), consistent with the known elimination routes of DTG [53]. PET signal in the brain was strikingly low (SUV = 0.04 ± 0.001) compared with peripheral organs, suggesting limited brain penetration of [ 18 F]DTG. Minimally invasive [ 18 F]DTG PET, therefore, sheds light on the elimination of DTG and its distribution to tissues. Interestingly, it suggests a minimal distribution to the brain which could correspond to a pharmacological sanctuary. Previous studies suggest a low brain distribution of [ 18 F]DTG in mice [54]. Further experiments with dynamic brain PET imaging and kinetic modelling will however be necessary to correctly estimate the brain distribution of [ 18 F]DTG in macaques.

Chemistry
Chemicals were purchased from Aldrich (Saint-Quentin-Fallavier, France) and used as received. 2-Fluoro-4-nitrobenzonitrile was purchased from Fluorochem. Dolutegravir and compound GSK3210346A were kindly provided by GSK. Reactions were monitored by thin layer chromatography (TLC) on aluminum pre-coated plates of silica gel 60F254 (VWR, Rosny-sous-bois, France). The compounds were localized at 254 nm using a UV

Chemistry
Chemicals were purchased from Aldrich (Saint-Quentin-Fallavier, France) and used as received. 2-Fluoro-4-nitrobenzonitrile was purchased from Fluorochem. Dolutegravir and compound GSK3210346A were kindly provided by GSK. Reactions were monitored by thin layer chromatography (TLC) on aluminum pre-coated plates of silica gel 60F 254 (VWR, Rosny-sous-bois, France). The compounds were localized at 254 nm using a UV lamp. 1 H and 13 C NMR spectra were recorded on a Bruker Advance 400 MHz apparatus using DMF-d 4 as solvent. The chemical shifts (δ) are reported in ppm (s, d, t, q, and b for singlet, doublet, triplet, quadruplet, and broad signal, respectively) and referenced with the solvent residual chemical shift. Ultra-performance liquid chromatography-mass spectroscopy (UPLC-MS) was realized on an Ultimate 3000 (Thermo Scientific, Waltham, MA, USA) device equipped with an Acquity BEH 2.1 × 50 mm, 1.7 µm column (Waters, Milford, MA, USA). A gradient of water with 0.1% of formic acid and acetonitrile with 0.1% of formic acid (3% of CH 3 CN/HCHO for 2 min, then rising to 100% during 7 min, then decreasing to 3% during 1 min, then keeping 3% for 2 min) at a flow rate of 0.  Figure S14, S15 and S16 respectively).

Radiochemistry
All reactions were carried out using a TRACERlab ® FX N Pro module (GE Healthcare). No carrier-added [ 18 F]fluoride ion was produced via the 18 O(p, n) 18 F nuclear reaction by irradiation of a 2 mL [ 18 O]water (>97% enriched, Rotem) target with an IBA Cyclone-18/9 (IBA) cyclotron.
Analytical HPLC was performed using a 717 plus Autosampler system, a 1525 binary pump, a 2996 photodiode array detector (Waters, Milford, MA, USA), and a Flowstar LB 513 (Berthold, Thoiry, France) gamma detector. The system was operated with the Empower 3 (Waters, Milford, MA, USA) software. HPLC were realized on a reverse phase analytical Symmetry ® C18 (150 × 3.9 mm, 5 µm, Waters) column using a mixture of H 2 O/CH 3 CN/PicB7 ® as eluent. The chemical identification of the peak was assessed by comparing the retention time of the radiotracer with the retention time of the non-radioactive reference (t R ref ). For acceptance, the retention time must be within the t R ref ± 10% range. Radiochemical and chemical purities were calculated as the ratio of the AUC of the radiotracer peak over the sum of the AUCs of all other peaks on gamma and UV chromatograms, respectively. The conversion of the reaction was calculated as the ratio of the decay-corrected activity of the radiotracer at the end of the synthesis over the activity of the starting material, both measured in an ionization chamber (Capintec ® , Berthold, Thoiry, France).

2-Fluoro-4-([ 18 F]fluoro)benzylamine ([ 18 F]2).
To a solution of [ 18 F]1 (tracer quantity) in the appropriate solvent (1 mL) were added the reduction agents, and the reaction took place under the conditions depicted in Tables 2 and 3. The reaction was quenched using a mixture of CH 3  The resulting complex was dried upon heating at 60 • C for 7 min under vacuum and a stream of helium followed by heating at 120 • C for 5 min under vacuum only. Upon cooling to 70 • C, a solution of 2-fluoro-4-nitrobenzonitrile (5 mg) in DMSO (700 µL) was added, and the mixture was heated at 95 • C for 5 min. After cooling to room temperature, the mixture was diluted with water (10 mL) and passed through a C18 Sep Pak ® cartridge (Waters). The cartridge was rinsed with water (5 mL) and eluted with methanol (1.5 mL) to reactor 2. A solution of NaBH 4 2M (20 µL) in THF (100µL) was added, followed by the addition of a solution of NiCl 2 ·6H 2 O (25 mg) in water (100 µL). The mixture was stirred at room temperature for 5 min. A solution of compound 3 (12 mg), TBTU (13 mg), and DIPEA (15 µL) in DMF (500 µL) was added, and the mixture was stirred at room temperature for 10 min. The reaction was diluted with a mixture of CH 3 CN/H 2 O/TFA 30/70/0.1 v/v/v (1.5 mL), and the crude was purified by reverse phase semi-preparative HPLC (Waters Symmetry ® C18 7.8 × 300 mm, 7 µm) with a 515 HPLC Pump (Waters, Milford, MA, USA) using a mixture of H 2 O/CH 3 CN/TFA (70/30/0.1 v/v/v, 5 mL/min) as eluent. UV detection (K2501, Knauer, Berlin, Germany) was performed at 254 nm. The purified compound was diluted with water (20 mL) and passed through a Sep-Pak ® C18 cartridge (Waters, Milford, MA, USA). The cartridge was rinsed with water (10 mL) and eluted with ethanol (2 mL), and the final compound was diluted with saline (0.9% w/v, 8 mL). Ready-to-inject [ 18 F]DTG (0.5-0.7 GBq) was obtained in 5.1 ± 0.8% (n = 10) non-decay corrected radiochemical yield within 95 min.
Quality control. Quality control was performed by reverse phase analytical HPLC according to the general procedure described above. A mixture of H 2 O/CH 3 CN/PicB7 ® (70/30/0.2 v/v/v, 2 mL/min) was used as eluent, and UV detection was performed at 258 nm. Molar activity was calculated as the ratio of the activity of the collected peak of [ 18 F]DTG measured in an ionization chamber (Capintec ® , Berthold, Thoiry, France) over the molar quantity of DTG determined using a calibration curve (see Figure S18 in Supplementary Materials). Molar activity is calculated as the mean value of three consecutive runs.

PET Imaging
Whole-body distribution of [ 18 F]DTG was assessed using whole-body (3-bed step) PET acquisition in a healthy male macaque (Macaca fascicularis, 8.35 kg, 10 years). Animal was i.v. injected with 181 MBq of 18 F-DTG followed by PET acquisition performed using a Siemens Biograph PET-CT scanner (Siemens Healthineers, Knoxville, TN, USA) as previously described [55]. Mean PET signal acquired from 150 to 180 min after injection of [ 18 F]DTG was normalized by injected dose of [ 18 F]DTG and animal weight (standardized uptake value, SUV). Representative PET image was co-registered to the anatomical CT scan.

Conclusions
We have developed the first fluorine-18 isotopic radiolabeling of the antiretroviral drug dolutegravir. The three steps of the radiosynthesis, i.e., radiofluorination, reduction, and peptide coupling reactions, have been optimized to maximize the quantity of [ 18 F]DTG produced. The whole process has been automated for reproducibility and radioprotection purposes to ensure the feasibility of preclinical PET imaging studies. For that matter, quality control has been set up, and 500-700 MBq of ready-to-inject [ 18 F]DTG was produced, a quantity compatible with injection into primates or humans. As a proof of concept, the first in vivo injection of [ 18 F]DTG was carried out in a macaque, and PET imaging was performed to observe the biodistribution of the tracer. [ 18 F]DTG is currently being used in PET studies with healthy and simian immunodeficiency virus (SIV)-positive non-human primates to explore the distribution of cART at remaining SIV sanctuaries and to understand their resistance.