Evaluation of D-isomers of 4-borono-2-¹⁸F-fluoro-phenylalanine and O-¹¹C-methyl-tyrosine as brain tumor imaging agents: a comparative PET study with their L-isomers in rat brain glioma

Boron neutron capture therapy (BNCT) is a biologically and physically targeted radiation therapy for tumors with the nuclear capture reaction of boron-10 (¹⁰B) specifically accumulating in tumor cells. The irradiation of low-energy thermal neutrons to ¹⁰B produces high-energy α (⁴He) particles and lithium ions (⁷Li), leading to cytotoxic effects against tumor cells. Due to the shorter track ranges of these particles (9 and 4 μm for α particles and ⁷Li ions, respectively) than the diameter of tumor cells, adjacent normal cells are expected to remain unaffected when these particles are emitted from tumor cells (for a review, see [1]). Two low molecular weight ¹⁰B carrier compounds, sodium sulfhydryl borane (BSH) [2] and L-4-boronophenylalanine conjugated with fructose (L-BPA-Fr) [3], have been clinically applied. BSH enters tumor cells by passive diffusion through the cell membrane [4‐6]. L-BPA and its F-18-labeled PET probe L-¹⁸F-FBPA [7‐12] are transported into tumor cells through L-type amino acid transporter 1 (LAT1) [11‐13] and amino acid transporter B^0,+ (ATB^0,+) [13], both of which are strongly upregulated in tumor cells [13, 14]. Since L-BPA accumulates in tumor cells with a low target tumor to non-target normal brain tissue ratio (TBR) [6], L-¹⁸F-FBPA was developed to predict the exact concentration of ¹⁰B-BPA in tumor cells prior to BNCT [7‐12]. The combined use of PET and inductively coupled plasma optical emission spectrometry confirmed that L-¹⁸F-FBPA-PET has the ability to estimate ¹⁰B concentrations delivered by L-BPA in the tumor and normal tissues of rats [10, 15] and humans [8].

In order to more effectively and safely perform BNCT, selective ¹⁰B delivery to tumor tissues with high TBR and sufficient thermal neutrons delivered to the tumor region are critical issues. One of the major limitations of L-BPA is its low TBR [6], which results in difficulties regulating the delivery of an adequate amount of ¹⁰B to target tumor tissues along with a sufficiently low amount of ¹⁰B to adjacent non-target normal tissues in order to avoid an adverse reaction. In an attempt to enhance L-BPA uptake into tumor cells, transient blood-brain barrier disruption (BBB-D) using hyperosmotic mannitol has been proposed to facilitate the uptake of L-BPA into F98 glioma in the rat brain, thereby enhancing survival times by BNCT over the non-BBB-D control [4, 15]. Another BBB-D by focused ultrasound sonication also increased the uptake of L-¹⁸F-FBPA into F98 glioma in the rat brain [16]. Since these BBB-D manipulations require treatments with drugs or sonication prior to the administration of L-BPA, the delivery of L-BPA into tumor cells may be affected by individual responses to these pre-treatments.

We previously demonstrated that the D-isomers of O-¹¹C-methyl-tyrosine (¹¹C-CMT), O-¹⁸F-fluoromethyl-tyrosine (¹⁸F-FMT), and O-¹⁸F-fluoroethyl-tyrosine (¹⁸F-FET) were better tumor imaging agents than their corresponding L-isomers in HeLa cells inoculated into the mouse hind leg [17‐21]. In contrast to natural amino acid-based tumor imaging probes such as D-¹¹C-methionine (D-¹¹C-MET), the D-isomers of ¹¹C-CMT, ¹⁸F-FMT, and ¹⁸F-FET were artificial amino acid-based probes without metabolic conversion from the D-isomer to L-isomer by D-amino acid oxidase [17, 22]. L-¹¹C-MET converted from D-¹¹C-MET via α-keto-γ-methiolbutyrate was incorporated not only into proteins via the conversion into amino-acyl-tRNA, but also into non-protein materials such as lipids and RNA by a transmethylation process via S-adenosyl-L-MET [23], leading to a worse TBR due to strong uptake into non-target normal tissues. The D-isomers of ¹¹C-CMT, ¹⁸F-FMT, and ¹⁸F-FET exhibited higher TBRs than their corresponding L-isomers in HeLa cells inoculated into the mouse hind leg because of the faster elimination rate of D-isomers than L-isomers from the blood to the urinary tract [17, 18].

In the present study, we conducted the D-isomerization of ¹⁸F-FBPA in order to improve the properties of L-¹⁸F-FBPA as a brain tumor imaging agent with a higher TBR, and the kinetics and distributions of L- and D-¹⁸F-FBPA were compared in the whole body organs of normal F344 rats with a tissue dissection method as well as in C6 glioma inoculated into rat brains using PET. Furthermore, in order to establish whether the D-isomerization of amino acid-based PET probes may be generalized to improve TBR in the brain, the L- and D-isomers of ¹¹C-CMT were also evaluated in rat brain C6 glioma.

The analyzed data in Fig. 1 confirmed that ¹⁸F-FBPA and ¹¹C-CMT preserved their enantiomeric specificities for the corresponding precursors after radiolabeling using F-18 or C-11. These retention times of 6.6 and 4.5 min for L- and D-¹⁸F-FBPA, respectively (Fig. 1c), and 11.1 and 9.2 min for L- and D-¹¹C-CMT, respectively (Fig. 1d) were consistent with those obtained by the corresponding authentic standard compounds of L- and D-FBPA (Fig. 1a) and L- and D-CMT (Fig. 1b), respectively.

The kinetics of the L- and D isomers of ¹⁸F-FBPA in normal rats are summarized in Table 1. The blood level of D-¹⁸F-FBPA was significantly lower than that of L-¹⁸F-FBPA from 1 min after the injection throughout measurements. Metabolic analyses of plasma demonstrated that both isomers of ¹⁸F-FBPA were very stable, showing 90.0 ± 0.6 and 89.0 ± 2.4% of unmetabolized L- and D-¹⁸F-FBPA, respectively, 90 min after the injection. The SUVs of D-¹⁸F-FBPA in the heart, lung, liver, spleen, pancreas, stomach, small intestine, colon, bone, and muscle were lower than those of L-¹⁸F-FBPA throughout the study after the injection. The uptake of D-¹⁸F-FBPA into normal brain tissue was markedly low with 13% of L-¹⁸F-FBPA 90 min after the injection. The SUVs of D-¹⁸F-FBPA in the kidney and bladder were markedly higher than those of L-¹⁸F-FBPA throughout the study after the injection. Differences in whole body distributions between L- and D-¹⁸F-FBPA were generally similar to those between L- and D-¹¹C-CMT (see Table 2 of [17]).

As shown in Fig. 2, PET imaging of rat C6 glioma indicated that the L- and D-isomers of ¹⁸F-FBPA (a and b) and ¹¹C-CMT (c and d) facilitated the detection of the tumor region in the rat brain. Although glioma cells were injected in the striatum as shown in the “Methods” section, the tumor tissues in almost all rat brains were observed in the cortex as shown in Fig. 2. It is assumed that the injected cells might draw back from the striatum to the cortex through the track of injection needle due to the high pressure in the brain, and then, the cells that were retained in the cortex proliferated.

The absolute uptakes, expressed as SUV, of the D-isomers of ¹⁸F-FBPA (Fig. 2b) and ¹¹C-CMT (Fig. 2d) were lower in the tumor and non-target normal tissue regions of the brain than those of the corresponding L-isomers (Fig. 2a, c); however, the image contrast reflecting TBR appeared to be markedly better for D-isomers (Fig. 2b, d) than for the corresponding L-isomers (Fig. 2a, c).

As shown in Fig. 3a, L-¹⁸F-FBPA exhibited slow and gradual uptake over time in tumor and normal regions up to 90 min after the injection. In contrast, although the uptake levels of D-¹⁸F-FBPA were low, TAC remained constant in tumor and normal tissues from 10 min and up to 90 min after the injection (Fig. 3b). The TACs of L- and D-¹¹C-CMT gradually increased and then remained constant in tumor and normal tissues 40 min and more after the injection (Fig. 3c, d).

The uptakes of the D-isomers of ¹⁸F-FBPA and ¹¹C-CMT into brain glioma were 68 and 89% of the corresponding L-isomers, while the uptakes of D-isomers into non-target normal brain tissues were only 15 and 68% of the corresponding L-isomers (Fig. 4a, c). The important results of the present study are shown in Fig. 4b, d, namely, TBRs were improved by D-isomerization from 1.45 to 6.93 for ¹⁸F-FBPA (TBR-D/TBR-L = 4.76) (b) and from 1.33 to 1.75 for ¹¹C-CMT (TBR-D/TBR-L = 1.38) (d).

The rat brains were dissected into slices for an autoradiographic examination just after the PET imaging. As shown in Fig. 5, both L- (a) and D-¹⁸F-FBPA (b) accumulated less in the normal region of the brain, and thus, the tumor regions corresponding to HE-positive area (c) were clearly imaged. As expected from the PET imaging data, although the uptake of D-¹⁸F-FBPA into glioma was lower than that of L-¹⁸F-FBPA, D-isomer uptake into non-target normal tissue was markedly lower than that of the L-isomer, resulting in TBR values of 6.12 and 1.74 for D-¹⁸F-FBPA and L-¹⁸F-FBPA, respectively (Fig. 5d). These results were consistent with those obtained in PET imaging analyses shown in Fig. 4.

The present study demonstrated that the D-isomers of ¹⁸F-FBPA and ¹¹C-CMT had higher TBRs in rat brain C6 glioma than the corresponding L-isomers. In tumor imaging using amino acid-based PET probes, their transport through upregulated LAT1 [13, 19, 25, 26] and ATB^0,+ [13] is important for their high uptake into tumor cells, and the uptake of D-¹⁸F-FBPA into tumor cells may depend on the activity of these transporters, as reported previously for L-¹⁸F-FBPA [11, 12] and L- and D-¹¹C-CMT [19]. Our previous study using mice inoculated with HeLa cells into their hind legs demonstrated that the uptakes of D-¹¹C-CMT into the normal brain region and abdominal organs, except for the kidney and bladder, were significantly lower than those of L-¹¹C-CMT [17], which was consistent with the present results obtained using L- and D-¹⁸F-FBPA. In contrast to the similar uptake levels of L- and D-¹¹C-CMT into HeLa cells inoculated into the mouse hind leg [17], the uptakes of D-¹¹C-CMT and D-¹⁸F-FBPA into glioma cells inoculated into the rat brain were lower than those of the corresponding L-isomers. There were several differences between the two assay conditions: the species used (mouse vs. rat), tumor cells (HeLa vs. C6 glioma), and inoculated regions (hind leg vs. brain). Among these differences, we confirmed that HeLa and C6 glioma cells both strongly expressed LAT1 [19]. In addition to the species difference, the involvement of BBB needs to be taken into account for the different kinetic properties between the L- and D-isomers. The effects of the D-isomerization of amino acid-based PET probes on BBB penetration appear to be controversial; one study showed no selective transport between L- and D-¹¹C-MET in normal and glioma regions [27], whereas another indicated the lower uptake of D-¹¹C-MET than L-¹¹C-MET in normal and astrocytoma regions [28]. Furthermore, the K1 (rate constant of amino acid transport) of D,L-¹⁸F-FBPA in normal and glioma regions was lower than that of L-¹⁸F-FBPA [8]. It is important to note that these three studies found no significant differences in TBRs between the L- and D (D,L) isomers [8, 27, 28]. The present results demonstrated that the uptake levels of D-¹⁸F-FBPA and D-¹¹C-CMT into normal brain regions were 15 and 68%, respectively, of their corresponding L-isomers and also that the uptakes of these D-isomers into brain C6 glioma were 69 and 89%, respectively, of the corresponding L-isomers. These different uptake profiles between the L- and D-isomers in normal and tissue regions resulted in improved TBRs from 1.45 (L-¹⁸F-FBPA) to 6.93 (D-¹⁸F-FBPA) (TBR-D/TBR-L = 4.76) and from 1.33 (L-¹¹C-CMT) to 1.75 (D-¹¹C-CMT) (TBR-D/TBR-L = 1.38). The faster washout of D-isomers than L-isomers from blood resulted in limited tumor uptake due to shorter available time for its uptake into tumor, while the lower background activity was the advantage for higher TBRs. The TBR of D-¹⁸F-FBPA (6.93) was markedly higher than that of L-¹⁸F-FET ranging from 2.0 (U87 glioblastoma) [29] to 3.1 (F98 glioma) [9] in rat brain glioma.

A critical point for more beneficial BNCT is to achieve a high delivery amount of ¹⁰B to tumor cells; however, it is considered to be more important to maximize the TBR of ¹⁰B uptake for proper therapeutic efficacy in tumors with less damage to normal tissues. L-BPA is currently used as a ¹⁰B carrier molecule to tumors, and although some preliminary evidence for the therapeutic effectiveness of BNCT using L-BPA has been reported, clinical outcomes are still considered to be unsatisfactory [30, 31]. The major limitation of BNCT efficacy is the relatively low TBR of ¹⁰B uptake [6]. Tentative BBB-D using hyperosmotic agents or sonication has been proposed to facilitate the brain tumor uptake of L-BPA/¹⁸F-FBPA [4, 15]. When an intracarotid injection of L-BPA was combined with a pretreatment with hyperosmotic mannitol, the TBR of L-BPA was 1.3–1.6-fold higher in F98 glioma than in the non-treated control [4, 15]. However, in addition to BBB-D, the administration of mannitol increases the glomerular filtration rate and renal blood flow [32], which may affect the kinetics and delivery of L-BPA to tissues due to a change in the excretion rate from the blood to the urinary tract. Another pretreatment with focused ultrasound sonication also enhanced the TBR of ¹⁸F-FBPA into the brain tumor by up to 1.75-fold that of the non-treated control [16]. One of the difficulties associated with BBB-D is possible neurotoxicity that may result in the exposure of normal brain tissues to toxic materials. Although L-BPA is not known to be neurotoxic, BBB-D further disseminates ¹⁰B-BPA as well as ¹⁸F-FBPA not only to tumor cells, but also to normal brain tissues that have an intact BBB before manipulations. Furthermore, since these BBB-D manipulations require pre-treatments with drugs or sonication prior to the administration of BPA, the delivery of BPA into tumor cells may be affected by individual responses to these pre-treatments, which are difficult to predict prior to BNCT. In contrast, the present results suggested that D-isomerization improved the TBRs of ¹⁸F-FBPA in C6 glioma without any pre-treatments. Based on the higher TRB of D-¹⁸F-FBPA than L-¹⁸F-FBPA, the D-isomerization of BPA may have potential as a method to improve the TBR of ¹⁰B uptake. Further animal studies are needed not only to investigate the effects of the D-isomerization of BPA on TBR improvement, but also to evaluate treatment efficacy with BNCT.

The present study demonstrated the potential of D-¹⁸F-FBPA as a better brain tumor imaging agent with higher TBR than its original L-isomer and previously reported tyrosine-based PET imaging agents. D-isomerization was more effective for ¹⁸F-FBPA than ¹¹C-CMT to improve TBR for brain tumor imaging agents. This improved TBR of D-¹⁸F-FBPA without any pre-treatments, such as tentative BBB-D using hyperosmotic agents or sonication, suggests that the D-isomerization of BPA results in the more selective accumulation of ¹⁰B in tumor cells that is more effective and less toxic than conventional L-BPA.