Synthesis, Radiolabeling, and Biological Evaluation of the trans-Stereoisomers of 1-Amino-3-(fluoro-¹⁸F)-4-fluorocyclopentane-1-carboxylic Acid as PET Imaging Agents

Abstract

The enantiomeric non-natural cyclic amino acids (3R,4R)-1-amino-3-fluoro-4-(fluoro-¹⁸F)cyclopentane-1-carboxylic acid and (3S,4S)-1-amino-3-fluoro-4-(fluoro-¹⁸F)cyclopentane-1-carboxylic acid ([¹⁸F]5) have been prepared as a racemic mixture in 1.3% decay corrected radiochemical yield and in greater than 99% radiochemical purity. [¹⁸F]5 is transported primarily via system L with some transport occurring via system ASC, as assessed in rat 9L gliosarcoma, human U87 ΔEGFR glioblastoma, and human DU145 androgen-independent prostate carcinoma tumor cells. In rats bearing intracranial 9L gliosarcoma, [¹⁸F]5 gave tumor to contralateral brain tissue ratios of up to 2.8. Biodistribution studies in healthy rats demonstrated that bladder accumulation is delayed until 10 min postinjection.

Positron emission tomography (PET), in combination with computed tomography (CT), is a vital imaging modality in clinical oncology. PET-CT is routinely applied for the visualization of numerous types of neoplasia and is FDA approved for a variety of indications, including cancer staging, assessing tumor response to therapy, guiding tumor biopsy, and evaluating potential recurrences, among others.¹⁻³ However, certain types of tumors, such as those of the brain and prostate, are challenging to image via PET-CT. This limitation is not inherent to PET-CT but arises from the use of 2-[¹⁸F]fluoro-2-deoxy-d-glucose (¹⁸[F]FDG), which has high avidity for normal brain tissue and rapidly concentrates in the bladder, resulting in large background signals that obscure neoplasia in these regions.⁴ To date, ¹⁸[F]FDG is one of only four radiotracers with FDA approval for an oncologic indication, and it is the only radiotracer indicated for broad oncologic application. Accordingly, the development of new PET radiotracers capable of imaging brain and prostate cancers is an important task.

Relative to normal tissue, tumors upregulate several metabolic processes to maintain heightened rates of cell proliferation. For example, neoplasia display enhanced avidity for glucose owing to an overexpression of glucose transporters (GLUT). As an analogue of glucose, ¹⁸[F]FDG is recognized by GLUT and concentrates in cells with heightened GLUT expression, which is the basis for the differentiation of tumors via ¹⁸[F]FDG-PET.⁵ Like glucose, amino acids (AAs) are central to many cellular processes, and the membrane bound proteins responsible for their transport are often overexpressed in tumors.^6,7 Consequently, the use of radiolabeled compounds that are substrates for amino acid transporters (AATs) is a viable strategy for imaging neoplasia.^8,9 Substrates for each of the three ubiquitous AATs, which include those from systems L (leucine preferring, sodium independent), A (alanine preferring, sodium dependent), and ASC (alanine, serine, cysteine, and threonine preferring, sodium dependent) have proven to be effective radiotracers in certain preclinical models,¹⁰ although only system L mediates AA transport into the brain.¹¹ Therefore, only AAs capable of undergoing transport by system L offer promise as general brain PET tracers.

To date, many radiolabeled endogenous and non-natural AAs with affinity for system L transport have been evaluated. One notable example of the former is [¹¹C]-methionine (MET), which has been employed in multiple PET imaging centers worldwide for more than 30 years. MET is an effective agent for the visualization of some prostate cancers and is more effective than ¹⁸[F]FDG in imaging glioma.¹²⁻¹⁴ Nonetheless, the clinical utility of MET, and natural AAs in general, is limited by the short half-life of the ¹¹C radionuclide and their propensity to act as substrates for protein synthesis, which can complicate the interpretation of PET data.¹⁵

In contrast, non-natural AAs offer a platform for the incorporation of ¹⁸F, a radionuclide with a more than fivefold longer half-life and a shorter positron range than ¹¹C. Moreover, certain non-natural AA scaffolds have been demonstrated to confer metabolic stability while maintaining avidity for transport by AATs. Several decades ago, Washburn and coworkers reported that ¹⁴C labeled 1-aminocyclobutane carboxylic acid (ACBC) and 1-aminocyclopentane carboxylic acid (ACPC) have high tumor avidity and good in vivo stability.¹⁶⁻¹⁹ Following these reports, Goodman and coworkers prepared and biologically evaluated several ¹⁸F radiolabeled analogues of ACBC.²⁰ Their biological transport was assessed in 9L gliosarcoma cells and, in each case, was mediated primarily by system ASC, though some transport by system L was observed. In Fischer rats bearing 9L gliosarcoma, [¹⁸F]FACBC derivatives demonstrated good brain uptake and much higher tumor to contralateral brain tissue ratios than [¹⁸F]FDG. Ultimately, these studies lead to the development and FDA approval of one fluorinated ACBC derivative, 3-anti-FACBC, as a PET imaging agent for the detection of recurrent prostate cancer.²¹ 3-anti-FACBC was also evaluated in humans as a PET imaging agent for the detection of primary glioma and metastatic brain cancer.²²⁻²⁴ Goodman and coworkers later reported the synthesis of racemic anti-1-amino-2-fluoro-cyclopentane-1-carboxylic acid ((R,S)-anti-2-FACPC), the first fluorinated ACPC derivative.²⁵ In contrast to fluorinated ACBC compounds, (R,S)-anti-2-FACPC demonstrated a preference for system L transport with some contribution by ASC, and in the 9L gliosarcoma model, (R,S)-anti-2-FACPC provided a nearly twofold increase in tumor to contralateral brain tissue compared to 3-anti-FACBC. While these preclinical data appeared promising, a study in humans demonstrated that (1R,2S)-anti-2-FACPC, isolated by HPLC purification of the racemic anti-2-FACPC mixture, rapidly accumulates in the bladder, precluding the detection of primary prostate tumors and limiting its utility.²⁶

In an effort to improve upon the biodistribution profile of (1R,2S)-anti-2-FACPC, we recently disclosed the synthesis of the two cis diastereomers of 3,4-DFACPC, (1S,3R,4S)-1-amino-3,4-difluorocyclopentane-1-carboxylic acid (anti-cis-3,4-DFACPC) and (1S,3S,4R)-1-amino-3,4-difluorocyclopentane-1-carboxylic acid (syn-cis-DFACPC).²⁷ Similar to (1R,2S)-anti-2-FACPC, system L was established as the primary means of transport for both cis diastereomers of 3,4-DFACPC, and both compounds provided high tumor to contralateral brain tissue ratios in the 9L gliosarcoma bearing Fischer rat model. However, in normal Fischer rats, bladder uptake of both 3,4-DFACPC stereoisomers was delayed until at least 5 min postinjection, providing a narrow window of opportunity to image the pelvic region that was not available with (1R,2S)-anti-2-FACPC. Interestingly, the magnitude of uptake of anti-cis-DFACPC was noticeably larger than that of syn-cis-DFACPC in most normal tissues as well as in the 9L tumor, suggesting that the orientation of the fluorine atoms plays a role in the transport of these compounds. Given the promising preclinical data obtained with the cis stereoisomers of 3,4-DFAPC, we reasoned that the trans stereoisomers should also be evaluated as PET imaging agents. Accordingly, described herein is the synthesis of (3R,4R)-1-amino-3,4-difluorocyclopentane-1-carboxylic acid and (3S,4S)-1-amino-3,4-difluorocyclopentane-1-carboxylic (trans-3,4-DFACPC) as a racemic mixture and its in vitro and in vivo evaluation as a PET imaging agent.

Results and Discussion

Chemistry

The target compound, racemic trans-3,4-DFACPC (5), was prepared in four steps starting from racemic triflate 1 (Scheme 1), which was obtained as previously described.²⁷ Briefly, compound 1 underwent a clean S_N2 inversion hydrolysis with aqueous NaHCO₃ to give racemic fluorohydrin 2 in nearly quantitative yield. Treatment of 2 with triflic anhydride in dichloromethane afforded highly unstable triflate 3, which was used directly for the next step without further purification and is the precursor for the radiosynthesis of trans-DFACPC 5. Many attempts to fluorinate triflate 3 under a variety of conditions were unsuccessful due to the formation of olefinic byproducts via elimination of triflate rather than substitution. However, treatment of 3 with CsF in tBuOH proved to be an efficient method for the formation of racemic trans difluoride 4(28) in 38% yield over two steps from 2. Subsequent deprotection of the N-Boc and ester groups in 4 was carried out with concentrated HCl to obtain the target compound 5 in 93% yield. The relative stereochemistry of compound 5 was determined by X-ray crystallography (Figure 1).

Scheme 1. Synthesis of Racemic trans-3,4-DFACPC (5).

Figure 1.

X-ray crystal structure of racemic trans-DFACPC (5). Atom labels are as follows: white = hydrogen, black = carbon, red = oxygen, blue = nitrogen, light green = fluorine, dark green = chlorine.

Radiochemistry

Racemic trans-3,4-[¹⁸F]-DFACPC ([¹⁸F]5) was prepared as previously described for the synthesis of anti-cis-3,4-[¹⁸F]-DFACPC,²⁷ except that in the fluorination reaction, [¹⁸F]CsF was employed instead of [¹⁸F]KF to overcome the problem of elimination, as described in the cold chemistry. The method involves a one-pot, two-step reaction sequence consisting of nucleophilic fluorination of the corresponding triflate precursor 3 with [¹⁸F]CsF in 1:1 ^tbutanol/acetonitrile with Kryptofix 222 for 10 min, followed by cleavage of the protecting groups using 6 M HCl (Scheme 2). After labeling and deprotection, the product [¹⁸F]5 was isolated by passing through an ion-retardation resin packed column in sequence with an alumina cartridge and an Oasis HLB reverse phase cartridge followed by elution with saline into fractions. The eluted fractions were used directly for biological studies. The radiosynthesis was completed in 83 min from the end of bombardment with 1.3% (12 mCi) decay-corrected yield, and the radiochemical purity was >99% (determined by radiometric HPLC). The specific activity of [¹⁸F]5 in the dose vial was not measured, although given that 20 mg of the triflate precursor was employed in the radiosynthesis, the maximum amount of unlabeled material in the final dose would not exceed 3.15 mg/mCi.

Scheme 2. Synthesis of [¹⁸F]5 (Racemic trans-3,4-[¹⁸F]-DFACPC).

Cell Uptake Studies

Competitive uptake inhibition experiments of [¹⁸F]5 were accomplished under the conditions previously reported for cis-difluoro stereoisomers.²⁷ Cell uptake of [¹⁸F]5 was evaluated in the presence and absence of inhibitors in the following cell lines: rat 9L gliosarcoma, human U87 ΔEGFR glioblastoma, and human DU145 androgen-independent prostate carcinoma. 2-Aminobicyclo[2.2.1]heptane-2-carboxylic acid (BCH) and N-methyl-R-aminoisobutyric acid (MeAIB) were used as system L and system A transport inhibitors, respectively, and the combination of alanine, cysteine, and serine (equal molar amounts) was used to inhibit system ASC. The uptake data were normalized and expressed as mean percent uptake relative to the control condition. The results of the amino acid uptake and inhibition studies are depicted in Table 1.

Table 1. 9L, U87 ΔEGFR, and DU145 Cell Uptake of [¹⁸F]5 with and without Inhibitors after 30 min of Incubation^a.

tumor cell line		[18F]5	% inhibition
9L	control	23.5 ± 1.5
	BCH	3.91 ± 0.72*	83
	MeAIB	24.8 ± 2.3**	no inhibition
	ASC	11.1 ± 0.54*	53
U87 ΔEGFR	control	6.45 ± 0.46
	BCH	2.16 ± 0.11*	67
	MeAIB	5.47 ± 0.16**	no inhibition
	ASC	2.34 ± 0.18*	64
DU145	control	34.0 ± 4.1
	BCH	2.03 ± 0.19*	94
	MeAIB	31.9 ± 1.4**	7
	ASC	5.68 ± 0.60*	83

^aData are presented as percent ligand uptake of the initial dose per 0.5 million cells (%ID/5 × 10⁵ cells) ± standard deviation (n = 3–4) and normalized for the dose and number of cells. p-Values represent comparisons of uptake in the presence of inhibitor to control uptake using two-tailed paired t-tests. p < 0.05 is considered statistically significant. *p < 0.05. **p ≥ 0.05.

Cellular uptake of [¹⁸F]5 was inhibited 83, 67, and 94% by BCH in 9L, U87 ΔEGFR and DU145 cells, respectively. Similarly, ASC inhibition resulted in reduced [¹⁸F]5 uptake, but to a lesser extent compared with BCH inhibitor (53, 64, and 83% by alanine, serine, and cysteine in 9L, U87 ΔEGFR, and DU145 cells, respectively). In the case of MeAlB, very little inhibition (7% in DU145 cells) or no reduction (in 9L, U87 ΔEGFR cells) of uptake was observed. These studies suggest that [¹⁸F]5, like its cis-difluoro stereoisomers (anti-cis-3,4-[¹⁸F]-DFACPC and the syn-cis-3,4-[¹⁸F]-DFACPC), undergoes cellular transport predominantly by system L with some transport occurring through system ASC. Moreover, the absolute uptake of radioactivity of [¹⁸F]5 was greater than that of syn-cis-3,4-[¹⁸F]-DFACPC in all of the cell lines studied. These uptake concentrations were comparable to anti-cis-3,4-DFACPC and anti-3-[¹⁸F]-FACBC uptake in rat 9L and U87 ΔEGFR cells. Notably, the absolute uptake of [¹⁸F]5 was nearly 2-fold higher than that of anti-cis-3,4-[¹⁸F]-DFACPC (18%ID/5 × 10⁵ cells) and 1.7-fold higher than the uptake of anti-3-[¹⁸F]-FACBC (20%ID/5 × 10⁵ cells) in DU145 cells (human androgen-independent prostate carcinoma cells), indicating cell specific superiority of [¹⁸F]5 over the cis-3,4-DFACPC stereoisomers and anti-3-[¹⁸F]-FACBC for imaging prostate cancers (Table 2). Furthermore, [¹⁸F]5 was evaluated as a racemic mixture in this study; the separation and biological evaluation of these enantiomers is ongoing and will be disclosed in a future paper.

Table 2. Comparison of 9L, U87 ΔEGFR, and DU145 Cell Uptake of [¹⁸F]5, anti-cis-3,4-[¹⁸F]-DFACPC, syn-cis-3,4-[¹⁸F]-DFACPC, and anti-3-[¹⁸F]-FACBC without Inhibitors after 30 min of Incubation^a.

tracer	9L	U87 ΔEGFR	DU145
[18F]5	23.5 ± 1.5	6.45 ± 0.46	34.0 ± 4.1
anti-cis-3,4-[18F]-DFACPC	23.4 ± 3.0	7.53 ± 1.2	17.7 ± 3.0
syn-cis-3,4-[18F]-DFACPC	7.82 ± 0.37	4.29 ± 0.29	12.2 ± 0.46
anti-3-[18F]-FACBC	23.2 ± 2.6	6.19 ± 1.22	20.2 ± 3.4

^aData are presented as percent ligand uptake of the initial dose per 0.5 million cells (%ID/5 × 10⁵ cells) ± standard deviation (n = 3–4) and normalized for the dose and number of cells.

Biodistribution Studies in Rats with Intracranial 9L Tumors

In vivo biodistribution studies were performed in Fischer rats with intracranial 9L gliosarcoma tumors as described in our previous study.²⁷ The uptake of radiotracer at 4.5, 12.5, and 52.5 min after intravenous administration of [¹⁸F]5 is summarized in Table 2. The highest tumor uptake of [¹⁸F]5 was observed at 12.5 min postinjection (p.i.) with 1.66 ± 0.35% ID/g and remained relatively steady throughout the experimental paradigm. The brain radiotracer levels were much lower with 0.39 ± 0.03% ID/g at 4.5 min, 0.43 ± 0.04*% ID/g at 12.5 min, and 0.46 ± 0.05% ID/g at 52.5 min p.i., resulting in relatively high tumor-brain ratios (T/N) (2.6, 2.8, 2.4 at 4.5, 12.5, and 52.5 min p.i., respectively) (Figure 2).

Figure 2.

Representative PET image of 9L tumor-bearing Fischer rat acquired with [¹⁸F]5, summed from 30–60 min postinjection. Activity is expressed in standard uptake values (SUV).

These ratios are comparable to those obtained with anti-cis-3,4-[¹⁸F]-DFACPC (T/N = 3.5 to 2.1 over 60 min p.i.) and syn-cis-3,4-[¹⁸F]-DFACPC (T/N = 4.3 to 2.2 over 60 min p.i.) tested previously in the same animal model, except at the 4.5 min time point with a lower T/N ratio obtained with [¹⁸F]5. However, T/N ratios fell in each following time point with the cis-difluoro stereoisomers owing to the progressive loss of activity in the 9L tumors or by the accumulation of activity in the normal regions of brain. In contrast, the candidate compound [¹⁸F]5 concentration remained steady in both the regions over the course of the study (T/N of 2.4 at 52.5 min), suggesting a superior window of [¹⁸F]5 in PET imaging (Figure 3).

Figure 3.

Biodistribution as percent of injected dose per gram (%ID/g) of radioactivity in tissues of 9L tumor-bearing Fischer rats following intravenous administration of [¹⁸F]5.

Biodistribution Studies in Normal Rats

The results of biodistribution data obtained with [¹⁸F]5 in normal rats at 10.4, 35.6, and 53.4 min are summarized in Table 3. [¹⁸F]5 displayed relatively similar uptake and clearance patterns in the brains of normal rats (0.45 ± 0.13 to 0.54 ± 0.16%ID/g) and tumor bearing rats (0.39 ± 0.03 to 0.46 ± 0.05%ID/g), suggesting that the radiotracer can cross the blood–brain barrier through biological transport and that its presence in the brain does not simply reflect heterogeneous permeability of the BBB. On the other hand, the highest initial uptake of radioactivity of [¹⁸F]5 was observed in the kidneys (3.99 ± 0.16%ID/g), indicating a normal renal excretion for the radiotracer. The remainder of the measured tissues, including liver, heart, lungs, bowel, testes, and spleen, showed an uptake of the radiotracer during the initial time point that decreased quickly with the passage of time (Table 3). These findings are comparable to our previously reported cis-difluoro isomers, but [¹⁸F]5 was found to be less concentrated in all the tissues than our previous compounds at all time points p.i. In addition, radioactivity levels of [¹⁸F]5 in the bladder were increased with time, similar to the cis stereoisomers, but were comparably lower than the previous isomers at all time points evaluated. At an early time point, [¹⁸F]5 (0.65 ± 0.27%ID/g at 10.4 min) showed comparable bladder uptake with anti-3-[¹⁸F]-FACBC (0.51 ± 0.13%ID/g at 4.5 min). The lesser radioactive concentration of [¹⁸F]5 in nontargeted tissues and lower initial radioactivity levels in the bladder during the time course of the PET study make this radiotracer more favorable for imaging tumors in the pelvic region (Table 4 and Figure 4).

Table 3. Biodistribution as Percent of Injected Dose Per Gram (%ID/g) of Radioactivity in Tissues of 9L Tumor-Bearing Fischer Rats Following Intravenous Administration of [¹⁸F]5^a.

tissue	4.5 min	12.5 min	52.5 min
liver	1.36 ± 0.14	0.97 ± 0.09	0.62 ± 0.05
heart	1.42 ± 0.10	0.98 ± 0.12	0.67 ± 0.08
lung	0.84 ± 0.08	0.64 ± 0.06	0.59 ± 0.11
muscle	0.56 ± 0.15	0.61 ± 0.12	0.68 ± 0.12
brain	0.39 ± 0.03*	0.43 ± 0.04*	0.46 ± 0.05*
tumor	1.00 ± 0.22*	1.20 ± 0.43*	1.11 ± 0.40*
bone	0.68 ± 0.23	0.75 ± 0.16	0.80 ± 0.09
spine	0.80 ± 0.26	0.75 ± 0.11	0.75 ± 0.15
T/N	2.6	2.8	2.4

^aData are reported as mean percent dose per gram ± standard deviation (n = 4) at each time point. p-Values represent comparisons of uptake in the 9L tumor and normal brain using two-tailed paired t-tests. p < 0.05 is considered statistically significant. *p < 0.05.

^bT/N denotes tumor to brain ratio.

Table 4. Biodistribution as Percent of Injected Dose Per Gram (%ID/g) of Radioactivity in Tissues in Normal Fischer Rats Following Intravenous Administration of [¹⁸F]5^a.

tissue	10.4 min	35.6 min	53.4 min
liver	1.20 ± 0.25	0.64 ± 0.22	0.71 ± 0.13
heart	0.97 ± 0.25	0.68 ± 0.14	0.66 ± 0.18
lung	0.73 ± 0.17	0.66 ± 0.32	0.55 ± 0.10
kidney	3.99 ± 0.16	2.14 ± 0.79	2.54 ± 0.64
bladder	0.65 ± 0.27	4.94 ± 2.96	9.89 ± 3.05
muscle	0.42 ± 0.11	0.59 ± 0.17	0.61 ± 0.18
brain	0.45 ± 0.13	0.49 ± 0.10	0.54 ± 0.16
bowel	0.59 ± 0.02	0.37 ± 0.02	0.37 ± 0.06
testes	0.50 ± 0.16	0.54 ± 0.21	0.50 ± 0.16
bone	0.76 ± 0.22	0.62 ± 0.12	0.65 ± 0.17
spine	0.72 ± 0.17	0.65 ± 0.08	0.66 ± 0.13

^aData are reported as mean percent dose per gram ± standard deviation (n = 3) at each time point.

Figure 4.

Biodistribution as percent of injected dose per gram (%ID/g) of radioactivity in tissues in normal Fischer rats following intravenous administration of [¹⁸F]5. Data are reported as mean percent dose per gram.

In summary, the synthesis of cold and ¹⁸F radiolabeled racemic trans-3,4-DFACPC is reported. [¹⁸F]5 was subjected to in vitro competitive inhibition experiments in rat 9L gliosarcoma, human U87 ΔEGFR glioblastoma, and human DU145 androgen-independent prostate carcinoma cells and, in each case, was demonstrated to undergo active transport mediated primarily by system L with some contribution from system ASC. Relative to [¹⁸F]-FACBC and the cis diastereomers of 3,4-DFACPC, [¹⁸F]5 demonstrated similar absolute uptake in 9L and U87 ΔEGFR cells and much greater uptake in DU145 cells. In microPET imaging experiments with Fischer rats bearing intracranial 9L gliosarcoma, tumor accumulation of [¹⁸F]5 reached 1.20 ± 0.43%ID/g, while uptake in contralateral brain remained relatively low, permitting tumor to normal brain tissue ratios of up to 2.8. In normal Fischer rats, the accumulation of [¹⁸F]5 was delayed until at least 10 min postinjection, whereas the cis diastereomers of DFACPC and anti-3-[¹⁸F]-FACBC demonstrated similar levels of bladder activity by the 5 min time point. Given its lag in bladder accumulation and high DU145 cell uptake, [¹⁸F]5 is a promising preclinical candidate for the imaging of primary prostate tumors. Additionally, the high tumor to contralateral brain signal obtained with [¹⁸F]5 in the 9L gliosarcoma model suggests that it may also be useful for imaging intracranial tumors.

Experimental Section

General Information

All solvents were purchased from Fisher Scientific or Sigma-Aldrich and dried over 4 Å mol sieves (8–12 mesh, Sigma-Aldrich). Unless otherwise noted, all commercially available reagents and substrates were used directly as received. Ultra-high purity dry air was purchased from nexAir LLC. Thin layer chromatography was performed on Merck silica gel plates and visualized by UV light and/or potassium permanganate. ¹H, ¹³C, and ¹⁹F NMR spectra were recorded on Bruker 600, Varian INOVA 600, INOVA 500 and INOVA 400 spectrometers. Residual solvent resonances were treated as internal reference signals. ¹⁹F spectra were referenced to either trifluoroacetic acid (−76.55 ppm) or fluorobenzene (−113.15 ppm). IR spectra were recorded on a Nicolet iS10 FT-IR spectrometer, and the absorption peaks were reported in cm^–1. The purification of products was performed via flash chromatography²⁹ unless otherwise noted. A Thomas capillary melting point apparatus was used to determine the melting points (uncorrected). High resolution mass spectra were obtained from the Emory University Mass Spec Facility Inc. X-ray crystal structure data were obtained from Dr. John Bacsa of the Emory University X-ray Crystallography Center. The [¹⁸F]fluoride was produced at Emory University Center for Systems Imaging with an 11 MeV Siemens RDS 111 negative-ion cyclotron (Knoxville, TN) by the ¹⁸O(p, n) ¹⁸F reaction using [¹⁸O]H₂O (95%). Alumina N SepPaks and HLB Oasis cartridges were purchased from Waters, Inc. (Milford, MA). The ion retardation (IR) chromatography columns and the IR resin AG 11A8 (50–100 mesh) were purchased from BioRad Laboratories (Hercules, CA). Trap/release cartridges model DW-TRC were purchased from D&W, Inc. (Oakdale, TN). Radiometric TLC was performed with the same type of silica plates from Whatman and analyzed using a Raytest system (model Rita Star, Germany). Isolated radiochemical yields were determined using a dose-calibrator (Capintec CRC-712M). Analytical HPLC experiments were performed with a Waters Breeze HPLC system equipped with a Bioscan flowcount radioactivity detector and an inline UV detector set to monitor wavelengths 210, 230, and 254 nm (Astec chirobiotic T column, Sigma-Aldrich part number 12021AST; mobile phase: MeOH). All animal experiments were carried out under humane conditions and were approved by the Institutional Animal Use and Care Committee (IUCAC) and Radiation Safety Committees at Emory University.

Chemistry

Compound 1 was prepared as described previously.²⁷

Racemic Ethyl 1-((tert-Butoxycarbonyl)amino)-3-fluoro-4-hydroxycyclopentane-1-carboxylate (2)

To a scintillation vial open to air containing a stir bar and racemic ethyl 1-((tert-butoxycarbonyl)amino)-3-fluoro-4-(((trifluoromethyl)sulfonyl)oxy)cyclopentane-1-carboxylate (1) (131 mg, 0.309 mmol) in THF (2 mL, 0.15 M), saturated aqueous sodium bicarbonate (2 mL) was added. The biphasic mixture was heated to 50 °C and stirred vigorously for 16 h. After cooling to room temperature, the contents of the reaction were poured into a separatory funnel, and the contents were diluted with ethyl acetate (10 mL). The organics were separated, washed with brine (5 mL), dried over Na₂SO₄, filtered, and concentrated in vacuo. The resulting crude residue was purified by silica gel flash chromatography, eluting the desired compound with a 50/50 EtOAc/hexanes gradient (R_f = 0.4). White solid, 88 mg, 0.302 mmol, 98% yield. ¹H NMR (400 MHz, chloroform-d) δ 5.14–4.94 (m, 2H), 4.47–4.34 (m, 1H), 4.22 (qd, J = 7.1, 1.3 Hz, 2H), 2.80 (ddd, J = 27.5, 15.9, 3.1 Hz, 1H), 2.47–2.17 (m, 4H), 1.43 (s, 9H), 1.28 (t, J = 7.1 Hz, 3H). ¹³C NMR (150 MHz, chloroform-d) δ 173.6, 155.2, 93.9 (d, J = 179.5 Hz), 80.6, 73.0 (d, J = 17.8 Hz), 62.2, 61.8, 42.8, 41.4 (d, J = 17.4 Hz), 28.4, 14.2. ¹⁹F NMR (282 MHz, chloroform-d, fluorobenzene reference standard) δ −195.4 – −196.2 (m). IR (neat, cm^–1): 3355, 1709, 1692. HRMS (ESI) Calcd for C₁₃H₂₃O₅NF (M + H)⁺: 292.15548. Found: 292.15546. Melting point: 72–74 °C.

Racemic Ethyl-1-((tert-Butoxycarbonyl)amino)-3-fluoro-4 (((trifluoromethyl)sulfonyl)oxy)cyclopentane-1-carboxylate (3)

A scintillation vial under N₂ containing a stir bar and racemic ethyl 1-((tert-butoxycarbonyl)amino)-3-fluoro-4-hydroxycyclopentane-1-carboxylate (2) (50 mg, 0.17 mmol, 1 equiv) and pyridine (30 μL, 0.38 mmol, 2.2 equiv) in CH₂Cl₂ (0.5 mL) was cooled to 0 °C. A separate vial containing trifluoromethanesulfonic anhydride (50 μL, 0.34 mmol, 2.0 equiv) in CH₂Cl₂ (0.5 mL) was cooled to 0 °C, and this mixture was added dropwise to solution of 2 with vigorous stirring. The mixture was stirred at 0 °C for 15 min, then diluted with hexanes (1 mL). A white powder precipitated and was filtered away, and the supernatant was concentrated at 0 °C to give 54 mg of crude, colorless oil. As the triflate is highly unstable, it was used directly without further purification.

Racemic Ethyl-1-((tert-Butoxycarbonyl)amino)-3,4-difluorocyclopentane-1-carboxylate (4)

To a scintillation vial under N₂ containing a stir bar and crude racemic ethyl-1-((tert-butoxycarbonyl)amino)-3-fluoro-4-(((trifluoromethyl)sulfonyl)oxy)cyclopentane-1-carboxylate (3) (54 mg), tert-butanol (2 mL) was added. Cesium fluoride (78 mg, 0.515 mmol) was then added under a stream of N₂, and the reaction was heated to 50 °C for 12 h. The mixture was diluted with water (5 mL) and CH₂Cl₂ (5 mL), and the phases were separated. The aqueous phase was washed with another portion of CH₂Cl₂ (5 mL), and the organics were collected and dried over Na₂SO₄, filtered, and concentrated. The resulting crude residue was purified by silica gel flash chromatography, eluting the desired compound with a 20/80 EtOAc/hexanes gradient (R_f = 0.4). White solid, 19 mg, 0.065 mmol, 38% yield over two steps. ¹H NMR (500 MHz, chloroform-d) δ 5.40–5.04 (m, 3H), 4.22 (q, J = 7.1 Hz, 2H), 2.83–2.47 (m, 3H), 2.29 (t, J = 19.1 Hz, 1H), 1.43 (s, 9H), 1.27 (t, J = 7.1 Hz, 3H). ¹³C NMR (125 MHz, chloroform-d) δ 172.7, 154.82, 96.6 (dd, J = 179.2, 29.2 Hz), 95.9 (dd, J = 180.5, 30.0 Hz), 80.5, 63.7, 62.2, 41.6, 40.8, 28.4, 14.2. ¹⁹F NMR (376 MHz, chloroform-d, fluorobenzene reference standard) δ −183.2 (broad s, 1F), −184.8 (dddd, J = 60.1, 35.0, 13.0, 9.1 Hz, 1F). IR (neat, cm^–1): 3281, 1735, 1688, 1671. HRMS (ESI) Calcd for C₁₃H₂₂O₄NF₂ (M + H)⁺: 294.15114. Found: 294.15129. Melting point: 64–66 °C.

Racemic 1-Amino-3,4-difluorocyclopentane-1-carboxylic acid hydrochloride (5)

To a scintillation vial containing racemic ethyl-1-((tert-butoxycarbonyl)amino)-3,4-difluorocyclopentane-1-carboxylate (4) (19 mg, 0.065 mmol), concentrated HCl (1 mL) was added. The reaction was heated to 90 °C for 1 h and then allowed to cool to room temperature. On cooling, colorless crystals formed spontaneously. The supernatant was carefully removed with a small gauge needle, and the crystals were freed of further solvent in vacuo. Colorless crystals, 10 mg, 0.061 mmol, 93% yield. ¹H NMR (600 MHz, deuterium oxide) δ 5.40–5.35 (m, 1H), 5.32–5.27 (m, 1H), 2.80 (ddt, J = 40.7, 16.2, 3.6 Hz, 1H), 2.71 (dd, J = 27.5, 16.5 Hz, 1H), 2.57 (ddd, J = 31.2, 16.7, 6.0 Hz, 1H), 2.42 (dd, J = 19.2, 17.3 Hz, 1H). ¹³C NMR (125 MHz, deuterium oxide) δ 173.6, 96.2 (dd, J = 172.7, 31.1 Hz), 94.9 (dd, J = 177.1, 32.0 Hz), 63.4, 40.6 (d, J = 23.5 Hz), 39.7 (d, J = 20.9 Hz). ¹⁹F NMR (376 MHz, deuterium oxide, trifluoroacetic acid reference standard) δ −184.6 – −185.1 (m), −187.6 – −188.1 (m). IR (neat, cm^–1): 2919 (broad), 1746. HRMS (ESI) Calcd for C₆H₁₀O₂NF₂ (M + H)⁺: 166.06741. Found: 166.06783. Melting point (decomposes): 260 °C.

Radiochemistry: Racemic trans-3,4-[¹⁸F]-DFACPC ([¹⁸F]5)

The preparation of racemic [¹⁸F]5 was based on the previously reported automated synthesis of anti-cis-3,4-[¹⁸F]-DFACPC.²⁷ To a glass vessel containing a solution of Cryptand 222 in MeCN (22 mg/mL) (1.0 mL) was added 1460 mCi of no-carrier-added [¹⁸F]HF through a trap/release (T/R) cartridge by using a solution of Cs₂CO₃/H₂O (20 mg mg/mL) (0.6 mL). The solvent was removed at 110 °C with a nitrogen flow, and additional MeCN (3.5 mL) was added followed by evaporation of the solvent with a nitrogen flow to remove residual H₂O. Racemic triflate precursor 4 (20 mg, 0.047 mmol) in dry tBuOH (0.5 mL) and MeCN (0.5 mL) was added to the vial, and the reaction mixture was heated at 110 °C for 10 min. The intermediate product was treated with 6 N HCl (0.5 mL) at 110 °C for 10 min and purified by passing through an IR column assembly consisting of a 7 mm × 120 mm bed of AG 11A8 IR resin, two neutral alumina SepPaks (preconditioned with water), and an HLB Oasis reverse phase cartridge (preconditioned with water). Racemic [¹⁸F]5 eluted in series through the assembly with three successive portions of sterile saline (∼4.0 mL) into dose vials and was ready for in vitro and in vivo studies. Based on analytical chiral HPLC data comparing the dose solution with authentic 5, [¹⁸F]5 was obtained in >99% radiochemical purity (see Figure S9). The pH of the final dose solution was tested with pH paper and found to be 6–7. The isolated radiochemical yield was 12 mCi in 7 mL of saline as determined using a dose calibrator, affording a 1.3% decay corrected radiochemical yield based on a synthesis time of approximately 83 min, which proceeded immediately upon the end of cyclotron bombardment.

Cells and Culture

The cancer cells used in the study include 9L gliosarcoma (rat), DU145 androgen-independent prostate carcinoma (human), and U87 glioblastoma tumor cell lines (human). Cells were cultured as described previously.³⁰ The tumor cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum, 100 U/mL penicillin, and 100 μg/mL streptomycin, maintained in T-150 tissue culture flasks under humidified incubator conditions (37 °C, 5% CO₂/95% air), and were routinely passaged at confluence.³¹ Cells thus prepared were used in cell uptake and inhibition assays and in mice tumor implantations.

Amino Acid Uptake and Inhibition Assays

Amino acid uptake and inhibition experiments were performed as described previously.³⁰ At the time of the experiments, the medium was exchanged to amino acid free Hank’s balanced salt solution (HBSS) and cells were adjusted to a final concentration of 5 × 10⁷ cells/mL. The following standard condition applied to each study (refer to the Supporting Information for optimization information). Approximately 5 × 10⁵ cells were exposed to 5 μCi of [¹⁸F]5, respectively, in 0.1 mL of amino acid/serum-free HBSS in the absence (control condition) or presence of transport inhibitors for 30 min under incubator conditions (37 °C, 5% CO₂/95% air) in 1.5 mL conical tubes. Ten millimolar final concentrations of MeAIB were used to inhibit uptake mediated by system A AATs. Ten millimolar final concentrations of 2-amino-bicyclo[2.2.1]heptane-2-carboxylic acid (BCH) were used to inhibit uptake mediated by system L AATs. The combination of 10 mM alanine-cysteine-serine (ACS, 3.3 mM of each amino acid) was used for uptake inhibition of system ACS AATs. After incubation, cells were twice centrifuged (75 G for 5 min) and rinsed with ice-cold HBSS to remove residual activity in the supernatant. Each assay condition was performed in triplicate. The activity in tubes was counted in a Packard Cobra II Auto-Gamma counter, the raw counts decay corrected, and the activity per cell number determined from the average of the triplicate. The data from these studies were expressed and normalized as percent uptake of the initial dose per 0.5 million cells (% ID/5 × 10⁵ cells) ± standard deviation (SD).

Tumor Induction and Animal Preparation

Rat 9L gliosarcoma cells for intracranial implantation experiments were cultured and prepared the same way as the uptake and inhibition assays, then washed with phosphate buffer solution (PBS), and were made a final concentration of 5 × 10⁴/5 μL in PBS. Rat 9L gliosarcoma cells were implanted into the brains of male Fischer 344 rats (160–210 g) as described previously.²³ Briefly, following anesthesia with an intramuscular injection of ketamine (60 mg/mL) and xylazine (7.5 mg/mL) solution, rats were placed in a stereotactic head holder and were injected with 5 μL suspension of rat 9L gliosarcoma cells (5 × 10⁴ cells per rat) in a location 3 mm right of midline and 1 mm anterior to the bregma at 4 mm deep to the outer table. The injection was performed over the course of 2 min, and the needle was withdrawn over the course of 1 min to minimize the backflow of tumor cells. The burr hole and scalp incision were closed, and the animals were returned to their original cages after recovering from the procedure. Intracranial tumors developed that produced weight loss, apathy, and hunched posture in the tumor-bearing rats. Typically, among 25 animals implanted with tumor cells, 20 would develop tumors visible to the naked eye upon dissection in approximately 10–12 days and were used in the study.³²

Anesthesia

Anesthesis was carried out as described previously.³³ Rats were anesthetized using isoflurane gas. Anesthesia was initiated 10 min ahead of imaging experiments by placing the animal in a cage ventilated with oxygen containing 1–2% isoflurane. Body temperature was held at 37 °C using a temperature-controlled warm air convection system.

Injection of the Radiotracer

The tracer was administered as described previously.³³ A catheter placed in the tail vein prior to imaging experiments was filled with isotonic sodium chloride solution. The tracer was diluted with saline to a final volume of 0.4 mL and injected via the catheter.

Rodent Biodistribution Studies

The microPET imaging process was carried out as described previously.³³ After anesthesia and placement of the tail vein catheter, the animal was placed with its body located at the center of the field of view. The rats were injected through the tail vein catheter with 200–250 μCi of [¹⁸F]5 in 0.4 mL of isotonic saline (pH 6–7). PET imaging microPET data were acquired with a Siemens Inveon PET/CT system (Siemens Medical Solutions, Knoxville, TN, United States). Radioactivity in the syringe was measured before and after the tracer was injected into the tail vein catheter using a Capintec CRC 15R (Capintec Inc., 6 Arrow Road Ramsey, NJ) dose calibrator. Data acquisition was performed for 60 min starting immediately following tracer injection. After PET imaging, all animals underwent computed tomography (CT) transmission imaging in the same position as the acquired PET data. The PET emission data were normalized and corrected for decay and dead time. The images were reconstructed with an ordered-subset expectation maximization algorithm, using attenuation correction from CT, into 15 1 min frames followed by 9 5 min frames. The image volume consisted of 128 × 128 × 159 voxels, each of a size of 0.78 × 0.78 × 0.80 mm. PET and CT images were coregistered using ASIPro and processed using software written by the authors. Regions-of-interest (ROIs) were placed on the fused PET/CT data including the tumor, in the right hemisphere of the brain, and contralateral region in the left hemisphere. The time–activity curves represent the mean activity in the ROIs over time.

Glossary

Abbreviations

AA

amino acid

AAT

amino acid transporter

ACBC

1-amino-cyclobutanecarboxylic acid

ACPC

1-aminocyclopentanecarboxylic acid

anti-3-FACBC

anti-1-amino-3-(fluoro-¹⁸F)-cyclobutane carboxylic acid

(R,S)-anti-2-FACPC

anti-1-amino-2-fluoro-cyclopentane-1-carboxylic acid

BBB

blood–brain barrier

BCH

2-aminobicyclo[2.2.1]heptane-2-carboxylic acid

CsF

cesium fluoride

CT

computed tomography

d

doublet

DCM

dichloromethane

DCY

decay-corrected yield

dd

doublet of doublets

ddt

doublet of doublet of triplets

3,4-DFACPC

1-amino-3,4-difluorocyclopentanecarboxylic acid

DMEM

Dulbecco’s modified Eagle’s medium

dq

doublet of quartets

dtd

doublet of triplets of doublets

ESI

electrospray ionization

3-FACPC

1-amino-3-fluorocyclopentane carboxylic acid

¹⁸[F]FDG

2-[¹⁸F]fluoro-2-deoxy-d-glucose

HBSS

Hank’s balanced salt solution

HCl

hydrochloric acid

HPLC

high performance liquid chromatography

HRMS

high resolution mass spectrometry

IR

infrared spectroscopy, KF, potassium fluoride

m

multiplet

MeAIB

2-(methylamino)-2-methylproprionic acid

MgSO₄

magnesium sulfate

NaHCO₃

sodium bicarbonate

NaOH

sodium hydroxide

N-Boc

N-tert-butoxycarbonyl

NMR

nuclear magnetic resonance spectroscopy

PBS

phosphate buffer solution

PET

positron emission tomography

q

quartet

qd

quartet of doublets

ROI

region-of-interest

rt

room temperature

s

singlet

t

triplet

tBuOH

tert-butanol

THF

tetrahydrofuran

TLC

thin layer chromatography.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.1c00062.

• ¹H, ¹³C, and ¹⁹F NMR spectra for all previously unreported compounds; X-ray crystallography data for [¹⁹F]5; time–activity curves for PET data in Tables 2 and 3; radiometric HPLC data for [¹⁸F]5; molecular formula strings; atomic coordinates (PDF)

The authors declare no competing financial interest.