Assessment of an 18F-Labeled Phosphoramidate
`Peptidomimetic as a New Prostate-Specific
`Membrane Antigen–Targeted Imaging Agent
`for Prostate Cancer
`
`Suzanne E. Lapi1, Hilla Wahnishe1, David Pham1, Lisa Y. Wu2, Jessie R. Nedrow-Byers2, Tiancheng Liu2,
`Kaveh Vejdani1, Henry F. VanBrocklin*1, Clifford E. Berkman*2,3, and Ella F. Jones*1
`
`1Department of Radiology and Biomedical Imaging, Center for Molecular and Functional Imaging, University of California, San
`Francisco, California; 2Department of Chemistry, Washington State University, Pullman, Washington; and 3Cancer Targeted
`Technology, LLC, Woodinville, WA
`
`Prostate-specific membrane antigen (PSMA) is a transmembrane
`protein commonly found on the surface of late-stage and meta-
`static prostate cancer and a well-known imaging biomarker for
`staging and monitoring therapy. Although 111In-labeled caprop-
`mab pendetide is the only approved agent available for PSMA
`imaging, its clinical use is limited because of its slow distribution
`and clearance that leads to challenging image interpretation.
`A small-molecule approach using radiolabeled urea-based
`PSMA inhibitors as imaging agents has shown promise for pros-
`tate cancer imaging. The motivation of this work is to explore
`phosphoramidates as a new class of potent PSMA inhibitors
`to develop more effective prostate cancer imaging agents
`with improved specificity and clearance properties. Methods:
`N-succinimidyl-4-18F-fluorobenzoate (18F-SFB) was conjugated
`to S-2-((2-(S-4-amino-4-carboxybutanamido)-S-2-carboxyethoxy)-
`hydroxyphosphorylamino)-pentanedioic acid (Phosphoramidate
`(1)), yielding S-2-((2-(S-4-(4-18F-fluorobenzamido)-4-carboxy-
`butanamido)-S-2-carboxyethoxy)hydroxyphosphorylamino)-
`pentanedioic acid (3). In vivo studies were conducted in mice
`bearing either LNCaP (PSMA-positive) or PC-3 (PSMA-negative)
`tumors. PET images were acquired at 1 and 2 h with or without
`a preinjection of a nonradioactive version of the fluorophosphor-
`amidate. Tissue distribution studies were performed at the end
`of the 2 h imaging sessions. Results: Phosphoramidate (1)
`and its fluorobenzamido conjugate (2) were potent inhibitors of
`PSMA (inhibitory concentration of 50% [IC50], 14 and 0.68 nM,
`respectively). PSMA-mediated tumor accumulation was noted
`in the LNCaP versus the PC-3 tumor xenografts. The LNCaP tu-
`mor uptake was also blocked by the administration of nonradio-
`active (2) prior to imaging studies. With the exception of the
`kidneys, tumor-to-tissue and tumor-to-blood ratios were greater
`than 5:1 at 2 h. The strong kidney uptake may be due to the
`known PSMA expression in the mouse kidney, because signifi-
`cant reduction (.6-fold) in kidney activity was seen in mice
`
`Received May 26, 2009; revision accepted Aug. 28, 2009.
`For correspondence or reprints contact: Ella F. Jones, Department of
`Radiology and Biomedical Imaging, Center for Molecular and Functional
`Imaging, University of California, San Francisco, 185 Berry St., Suite 350,
`Box 0946, San Francisco, CA 94107.
`E-mail: ella.jones@radiology.ucsf.edu
`*Contributed equally to this work.
`COPYRIGHT ª 2009 by the Society of Nuclear Medicine, Inc.
`
`injected with (2). Conclusion: 18F-labeled phosphoramidate (3)
`is a representative of a new class of PSMA targeting peptidomi-
`metic molecules that shows great promise as imaging agents for
`detecting PSMA1 prostate tumors.
`
`Key Words: molecular imaging; PET; radiopharmaceuticals; 18F;
`PSMA; phosphoramidate; prostate cancer
`
`J Nucl Med 2009; 50:2042–2048
`DOI: 10.2967/jnumed.109.066589
`
`Prostate cancer is the second leading cancer found in
`
`men. Each year in the United States, more than 186,000
`new cases are diagnosed, with approximately 29,000
`prostate cancer–related deaths (1). Prostate cancer imaging
`not only is an important component of diagnosis and
`staging, it has also become an integral part of treatment
`planning, especially in radiation therapy (2).
`Prostate-specific membrane antigen (PSMA) is a 100 to
`120 kDa transmembrane protein upregulated on the tumor
`cell surface of late-stage, androgen-independent, and me-
`tastatic prostate cancer (3). Because of its restricted over-
`expression on prostate cancer cells, PSMA is an important
`biomarker for prostate cancer prognosis and an attractive
`target for therapy (4). 111In-labeled capropmab pendetide
`(ProstaScint; Cytogen) is an antibody-based agent approved
`by the Food and Drug Administration for PSMA imaging.
`Although it is the only commercially available agent, its
`clinical use is limited largely because of its slow distribu-
`tion and clearance, and the images produced are often
`difficult
`to interpret. More recently, a small-molecule
`approach has generated a class of promising urea-based
`PSMA-targeted agents (Fig. 1) for PET and SPECT (5–7).
`Although optimization of
`these tracers is under way,
`exploration of more potent PSMA-targeting groups and
`their use for prostate cancer imaging will produce new
`imaging agents with improved sensitivity and specificity.
`
`2042
`
`THE JOURNAL OF NUCLEAR MEDICINE • Vol. 50 • No. 12 • December 2009
`
`

`

`FIGURE 1. Structural elements of known PSMA substrate and inhibitors, compared with phosphoramidate (1). Highlighted
`portions indicate structural features similar to phosphoramidate design.
`
`Also known as folate hydrolase I and glutamate car-
`boxypeptidase II, PSMA is reported to possess proteolytic
`activities toward g-glutamyl folic acid derivatives (Fig. 1)
`and the neuropeptide N-acetylaspartylglutamate (8,9). Further
`studies on the PSMA substrate specificity have indicated that
`acidic residues at the P1 and P19 positions are more pref-
`erable, and several folate-like and N-acetylaspartylglutamate–
`like dipeptides with modest hydrolytic efficiency have been
`identified (10). In our own work, while adapting the di-
`peptide motif for PSMA targeting, we have developed
`a library of tetrahedral phosphoramidates for PSMA inhibition
`(11). Through molecular pruning, we have systematically
`identified several potent
`inhibitors
`(12)
`that pseudo-
`irreversibly bind to PSMA (13). In vitro studies of the
`fluorescently labeled phosphoramidates further reveal their
`ability to localize and internalize in PSMA-positive
`(PSMA1) cells (14), making these compounds ideal
`candidates for PSMA-targeted delivery for prostate cancer
`imaging and therapeutic applications.
`Herein we present our effort in using the phosphorami-
`date scaffold as a targeting element for prostate cancer
`imaging. The synthesis and characterization of an 18F-
`labeled analog of phosphoramidate (1) and its in vivo PET
`and biodistribution data in murine xenografts are reported.
`
`MATERIALS AND METHODS
`
`Cell Lines, Reagents, and General Methods
`LNCaP and PC-3 cells were obtained from the American Type
`Culture Collection. 1H, 13C, and 31P nuclear magnetic resonance
`(NMR) spectra were recorded on a 300- and 500-MHz (DRX;
`Bruker) or a 400-MHz (Varian) NMR spectrometer. 1H NMR
`chemical shifts are relative to tetramethylsilane (d 5 0.00 parts
`
`per million [ppm]), CDCl3 (d 5 7.26 ppm), or D2O (d 5 4.87
`ppm). 13C NMR chemical shifts are relative to CDCl3 (d 5 77.23
`ppm). 31P NMR chemical shifts in CDCl3 or D2O were externally
`referenced to 85% H3PO4 (d 5 0.00 ppm) in CDCl3 and D2O,
`respectively. Aqueous solutions were prepared with deionized
`distilled water (Milli-Q water system; Millipore). All liquid flash
`chromatography (silica or C18) was performed using a Biotage
`SP4 flash chromatography system. The high-performance liquid
`chromatography (HPLC) analysis and purification system for
`radioactive compounds consisted of a Rheodyne injector with
`a 2-mL loop, a Waters model 590 pump, a Shimudzu model SPD-
`10A ultraviolet detector, and an in-line radioactivity detector
`(model 105 s; Carroll and Ramsey Associates) coupled to a data
`collection system (PeakSimple, model 304; SRI). The phosphor-
`amidate precursor 2-benzyloxycarbonylamino-4-(1-benzyloxycar-
`bonyl-2-hydroxy-ethylcarbamoyl)-butyric acid benzyl ester was
`synthesized as previously described (14).
`t-Butyl 4-N,N,N,tri-
`methyl ammonium benzoate triflate salt was synthesized accord-
`ing to the literature method (15). N-succinimidyl-4-fluorobenzoate
`(SFB) was prepared as previously described (16). All other
`reagents and solvents were purchased from Sigma-Aldrich and
`Fisher Scientific and were used without further purification unless
`indicated.
`
`Synthesis of S-2-((2-(S-4-Amino-4-
`Carboxybutanamido)-S-2-Carboxyethoxy)-
`Hydroxyphosphorylamino)-Pentanedioic Acid (1)
`A tetrahydrofuran solution (15 mL) of 2-benzyloxycarbonyl-
`amino-4-(1-benzyloxycarbonyl-2-hydroxy-ethylcarbamoyl)-butyric
`acid benzyl ester (71 mg, 0.069 mmol) was added with 10% Pd/C
`(12 mg), K2CO3 (23 mg, 2 equivalent), and distilled H2O (1 mL).
`The mixture was stirred vigorously under Ar(g), followed by the
`charge of H2(g) from a balloon for 7 h at room temperature. After
`the reaction was completed,
`the solvent was removed under
`vacuum. The remaining residue was dissolved in 1:1 ratio of
`
`18F-LABELED PHOSPHORAMIDATE • Lapi et al.
`
`2043
`
`

`

`methanol/water, filtered through a 0.2-mm PTFE micropore
`filtration disk (Whatman), and dried under vacuum. The yield
`was 89.9%. 1H NMR (300 MHz, D2O): d 1.75–1.94 (m, 2H),
`1.98–2.10 (m, 2H), 2.12–2.20 (m, 2H), 2.43–2.51 (m, 2H), 3.47–
`3.54 (q, 1H, J 5 6Hz, 15Hz), 3.63–3.67 (q, 1H, J 5 5 Hz, 8 Hz),
`3.98–4.10 (m, 2H), and 4.32–4.35 (t, 1H, J 5 4 Hz). 13C NMR (75
`MHz, D2O): d 26.8, 31.4, 31.7, 31.8, 33.7, 48.7, 54.2, 55.6, 55.8,
`56.6, 64.6, 64.7, 174.3, 175.21, 176.0, 181.5, 181.6, and 182.9.
`31P NMR (300 MHz, D2O): d 7.63.
`
`stream. O-(N-succinimidyl)-N,N,N9,N9-tetramethyluroniumtetra-
`fluoroborate (10 mg, 33 mmol) in 0.5 mL of acetonitrile was
`added to the resulting residue and allowed to react at 60°C for
`5 min. After cooling, the solution mixture was diluted with 0.5%
`acetic acid to 10 mL and loaded onto a second Chromafix HR-P
`cartridge. The cartridge was washed with 5 mL of 20% acetoni-
`trile and eluted with 1.5 mL of acetonitrile. Radiochemical
`yields were typically approximately 20% decay-corrected from
`fluoride.
`
`Synthesis of S-2-((2-(S-4-(4-Fluorobenzamido)-
`4-Carboxybutanamido)-S-2-Carboxyethoxy)
`Hydroxyphosphorylamino)-Pentanedioic Acid (2)
`A solution of N-succinimidyl-4-fluorobenzoate (SFB) (6 mmol)
`in 100 mL of DMSO was added to a stirred solution of (1)
`(2 mmol, 100 mL of 20 mM in H2O), 160 mL of H2O, and 40 mL
`of 1 M NaHCO3. The reaction mixture was stirred for 6 h in the
`dark at room temperature. The pH of the resulting solution was
`then adjusted to 9.3 by the addition of 8 mL of 1 M Na2CO3. The
`unreacted (1) was scavenged by stirring with 25 mg of the Si-
`Isocyanate resin (SiliCycle, Inc.) overnight at room temperature.
`The solution was subsequently centrifuged (7,800 relative centrif-
`ugal force, 10 min), and the supernatant was lyophilized in a 2-mL
`microcentrifuge tube. The unreacted or hydrolyzed SFB was
`removed by successively triturating the lyophilized solid with 1-mL
`portions of DMSO and centrifuging the mixture (16,200 relative
`centrifugal force, 1 min) after each wash;
`this process was
`repeated 10 times. The resulting solid was dried in vacuo
`providing the desired 4-fluorobenzamido-phosphoramidate (2) in
`quantitative yield. 1H NMR (D2O): d 1.56–1.68 (m, 2H), 1.86–
`2.14 (m, 4H), 2.27–2.35 (m, 2H), 3.27–3.34 (m, 1H), 3.73–3.87
`(m, 2H), 4.05–4.08 (t, 1H, 3.9 Hz), 4.15–4.19 (dd, 1H, 4.4 Hz, 9.15
`Hz), 7.04–7.10 (t, 2H, 9.0 Hz, 9.0 Hz), and 7.66–7.70 (dd, 2H, 9.0
`Hz, 5.3 Hz). 13C NMR (D2O): d 27.51, 32.04, 32.08, 32.69,33.95,
`55.96, 56.05, 56.83, 64.90, 115.63, 115.80, 139.90, 129.97,
`160.81, 163.88, 165.87, 169.66, 175.37, 176.13, 178.75, 181.66,
`and 183.19. 31P NMR (D2O): d 8.33. Matrix-assisted laser
`desorption ionization high-resolution mass spectrometry (M-K)2
`calculated 715.9266, found 715.9359 for C20H20FK4N3O10P.
`
`Synthesis of N-Succinimidyl-4-18F-Fluorobenzoate
`(18F-SFB)
`18F-SFB was synthesized based on the literature method with
`modifications (17). Briefly, 18F-fluoride ion was produced by
`proton bombardment of 18O-H2O using the University of Cal-
`ifornia, San Francisco (UCSF), PETtrace (GE Healthcare) 17-
`MeV cyclotron. The 18F-fluoride ion was concentrated on an anion
`exchange column (Oak Ridge Technology Group Inc.) and eluted
`with 0.5 mL (95% acetonitrile/water) in a solution of Kryptofix
`222
`(4,7,13,16,21,24-hexaoxa-l,10-diazabicyclo[8.8.8]hexaco-
`sane-K222; 30 mg/mL) and K2CO3 (5.5 mg/mL). The water was
`removed azeotropically by the addition and evaporation of
`acetonitrile. A DMSO solution (0.5 mL) of t-butyl 4-N,N,N,tri-
`methyl ammonium benzoate triflate salt (2 mg, 11 mmol) was
`added and allowed to react at 160°C for 10 min. After cooling, 0.6
`mL of 1 M HCl was added, and the solution was heated to 120°C
`for 15 min. The solution mixture was diluted to 10 mL with water
`and passed through a Chromafix HR-P cartridge (Macherey-
`Nagel). 18F-fluorobenzoic acid was eluted from the cartridge with
`acetonitrile, treated with 20 mL of methanolic (CH3)4NOH in
`0.4 mL of acetonitrile, and brought to dryness under a nitrogen
`
`Synthesis of 18F-S-2-((2-(S-4-(4-Fluorobenzamido)-
`4-Carboxybutanamido)-S-2-Carboxyethoxy)
`Hydroxyphosphoryl-Amino)-Pentanedioic Acid (3)
`The synthesized 18F-SFB was brought to dryness under reduced
`pressure with a nitrogen stream. It was dissolved in 50 mL of
`acetonitrile and allowed to react with 3–4 mg of phosphoramidate
`(1) (dissolved in 100 mL of H2O) and 100 mL of sodium carbonate
`buffer (1 M), pH 10. The reaction was allowed to proceed for 10
`min at room temperature, and the radioconjugate was purified by
`HPLC (Synergi Fusion-RP 80A column; Phenomenex) (250 mm ·
`4.6 mm, 4 mm, 100 mM sodium phosphate buffer, pH 6.8, as
`eluant; flow rate, 1 mL/min; elution time, 5 min). Radiochemical
`yields (from 18F-SFB) were typically greater than 90%.
`
`General Method of Determining Inhibitory Concentration
`of 50% (IC50)
`Inhibition studies were performed as previously described, with
`minor modifications (12,14). Working solutions of the substrate
`(N-[4-(phenylazo)-benzoyl]-glutamyl-g-glutamic acid [PABGgG])
`and inhibitor were made in Tris buffer (50 mM, pH 7.4). Working
`solutions of purified PSMA were diluted in Tris buffer (50 mM,
`pH 7.4, containing 1% Triton X-100 [Sigma-Aldrich]) to provide
`15%220% conversion of substrate to product in the absence of
`inhibitor. A typical incubation mixture (final volume, 250 mL) was
`prepared by the addition of either 25 mL of an inhibitor solution or
`25 mL of Tris buffer (50 mM, pH 7.4) to 175 mL of Tris buffer
`(50 mM, pH 7.4) in a test tube. PABGgG (25 mL, 10 mM) was
`added to the above solution. The enzymatic reaction was initiated
`by the addition of 25 mL of the PSMA working solution. In all
`cases, the final concentration of PABGgG was 1 mM; the enzyme
`was incubated with 5 serially diluted inhibitor concentrations,
`providing a range of inhibition from 10% to 90%. The reaction
`was allowed to proceed for 15 min with constant shaking at 37°C
`and was terminated by the addition of 25 mL of methanolic
`trifluoroacetic acid (2% trifluoroacetic acid by volume in metha-
`nol), followed by vortexing. The quenched incubation mixture was
`quickly buffered by the addition of 25 mL of K2HPO4 (0.1 M),
`vortexed, and centrifuged (10 min at 7,000g). An 85-mL aliquot of
`the resulting supernatant was subsequently quantified by HPLC as
`previously described (18,19). IC50 values were calculated using
`KaleidaGraph 3.6 (Synergy Software).
`
`In Vivo PET Studies
`All animal experiments were conducted in accordance with the
`UCSF Institutional Animal Care and Use Committee guidelines.
`The 18F-fluorobenzamido-phosphoramidate (3) was evaluated in
`vivo using murine xenografts. Prostate cancer cell lines LNCaP
`(PSMA1, in 50/50 serum-free medium and Matrigel (BD Bio-
`sciences) matrix) and PC-3 (PSMA-negative, in 100% serum-free
`medium) were used to inoculate the animals. Approximately 107
`LNCaP (n 5 6) or PC-3 (n 5 3) cells were injected subcutaneously
`on the right shoulder of male NCr nude mice. Animals with tumors
`
`2044
`
`THE JOURNAL OF NUCLEAR MEDICINE • Vol. 50 • No. 12 • December 2009
`
`

`

`FIGURE 2. General scheme of modular synthetic approach to phosphoramidates.
`
`measuring between 5 and 10 mm (1–2 wk after injection) were
`anesthetized by isoflurane inhalation. The synthesized 18F-fluoro-
`benzamido-phosphoramidate (3)
`(3,700–7,400 kBq [100–200
`mCi]) was administered through tail vein injection. The animals
`were imaged by a microPET/CT system (Inveon; Siemens) at 0, 1,
`and 2 h for 10-min acquisition times. For blocking studies, animals
`were injected with 1 mg of nonradioactive fluorobenzamido-
`phosphoramidate (2) in 200 mL of Tris buffer 1 h before injection
`of the radioactive tracer.
`The PET data were acquired in list mode and reconstructed with
`the iterative ordered-subset expectation maximization 2-dimensional
`reconstruction algorithm provided with the Siemens Inveon System.
`
`Biodistribution Studies
`After imaging at 2 h, animals were euthanized for biodistribu-
`tion analysis. Blood was collected by cardiac puncture. Major
`organs—heart, lung, liver, pancreas, spleen, kidney, brain, and
`testes—and tumor xenografts were harvested, weighed, and
`counted in an automated g-counter (Wizard 2; PerkinElmer).
`The percentage injected dose per gram (%ID/g) of tissue was
`calculated by comparison with standards of known radioactivity.
`Statistical analysis was performed using a t test (Microsoft
`Excel software). All analyses were 1 tailed and considered a type 3
`(2-sample unequal variance). A P value of less than 0.05 was
`considered statistically significant.
`
`RESULTS
`
`Synthesis and Characterization of Phosphoramidate (1)
`and Its Conjugates
`Using bis-(diisopropylamino) chlorophosphine (Cl-P-
`[N(iPr)2]2) and protected glutamate, we have been able to
`generate the inhibitor core routinely. On the basis of this
`modular synthetic approach, a variety of primary alcohols
`have been incorporated with ease (Fig. 2). As previously
`described, systematic molecular pruning and computational
`chemistry have suggested that serine and glutamate at P1
`and P19 positions are responsible for interacting with the
`PSMA Arg536 and Arg210/Tyr700/Tyr227 binding pockets,
`
`respectively (12). Further cell studies have led to the
`discovery of the leading phosphoramidate (1) with potent
`PSMA inhibition (IC50, 14 nM) (14). Following Figure 2,
`phosphoramidate (1) was synthesized at close to 90% yield.
`By incorporating the glutamate–serine dipeptide as a pri-
`mary alcohol, this leading inhibitor also possesses an amine
`functional linker for incorporation of 18F-labeled prosthetic
`groups for in vivo PET (Fig. 3).
`To investigate the effect of the prosthetic group on the
`inhibitory potency of phosphoramidate (1), we opted to
`synthesize the nonradioactive analog of the fluorobenza-
`mido-phosphoramidate (2) conjugate. Using standard bio-
`conjugation techniques (20), we reacted a 3-fold excess of
`SFB with phosphoramidate (1) in buffered conditions. For
`the ease of separation, resin-bound isothiocyanate was used
`as a scavenger for unreacted phosphoramidate (1). Un-
`reacted or hydrolyzed SFB was removed by successively
`triturating (2) with DMSO before subsequent inhibition
`studies.
`Using the same conjugation chemistry through an
`N-succinimidyl ester, we efficiently coupled 18F-SFB to
`phosphoramidate (1) in 10 min, producing the radiolabeled
`PSMA inhibitor (3) at greater than 90% radiochemical yield.
`With the nonradioactive fluorobenzamido-phosphoramidate
`(2) as a standard, we have optimized HPLC conditions
`using a Fusion RP column (Phenomenex, Inc.) to purify the
`18F-labeled product (3). Six different batches of the labeled
`product have been generated with consistent reproducibil-
`ity.
`
`PSMA Inhibition
`The effect of SFB conjugation on PSMA inhibition by
`the phosphoramidate was investigated using an HPLC-
`based IC50 assay (12,18,19). The nonradioactive fluoro-
`benzamido conjugate (2) exhibited a higher inhibitory potency
`than did the parent phosphoramidate with the IC50 value of
`
`FIGURE 3. Synthetic scheme for 4-
`fluorobenzamido-phosphoramidate
`conjugate (2) and 4-18F-fluorobenzami-
`do-phosphoramidate conjugate (3).
`
`18F-LABELED PHOSPHORAMIDATE • Lapi et al.
`
`2045
`
`

`

`control. At 2 h post-injection, the PSMA1 tumor accumu-
`lation was 1.24 6 0.17 %ID/g, with a tumor-to-blood ratio
`of 9:1 (Fig. 7). With the exception of the kidneys, there was
`minimal nonspecific uptake in all other organs (,0.25
`%ID/g). Tumor uptake in the PSMA2 control was com-
`parable to the uptake in the normal organs (0.32 6 0.14
`%ID/g), with a tumor-to-blood ratio of 1:1. The differences
`of tumor uptake collected in the PSMA1 and PSMA2
`animal models were statistically significant, as confirmed
`by a 1-tailed Student t test with the P value less than 0.002.
`When animals were treated in advance with the nonradio-
`active fluorobenzamido-phosphoramidate (2), the uptake
`by the LNCaP PSMA1 tumor was decreased by 8-fold
`(0.13% 6 0.14), with a tumor-to-blood ratio of 0.8:1.
`The kidney uptake in both PSMA1 and PSMA2 models
`was relatively high at 2.24 6 0.6 %ID/g and 2.83 6 0.9
`%ID/g, respectively. However, a significant decrease in
`kidney uptake (.6-fold) was observed in mice pretreated
`with the nonradioactive blocking agent.
`
`DISCUSSION
`
`Design of Phosphoramidates as PSMA Inhibitors
`Phosphoramidates, first described by Maung et al. (19),
`are potent PSMA inhibitors. The design strategy of this
`class of compounds is largely based on the binding features
`of PSMA endogenous substrates and potent inhibitors. As
`shown in Figure 1, the phosphoramidate scaffold is in-
`corporated with L-glutamate at the P19 position, possessing
`a binding feature closely resembling L-glutamate in the
`folyl-g-glu substrate. Compared with 2-PMPA, a known
`PSMA potent inhibitor, and the urea-based PSMA target
`agents (Fig. 1), phosphoramidates not only have similar
`structural features but also are well suited for carrying
`amine-reactive payloads while retaining excellent inhibi-
`tory potency.
`In the past, a variety of phosphoramidates have been
`synthesized using a modular approach (12,14) from Cl-P-
`[N(iPr)2]2, protected glutamate, and primary alcohols. To
`closely mimic a PSMA substrate, we introduced the
`glutamate–serine dipeptide as a primary alcohol building
`block to complete the synthesis of phosphoramidate (1).
`In this particular compound, whereas the serine residue
`occupies the P1 position to provide an additional binding
`feature to the Arg536 pocket (12), the glutamate residue
`serves as a linker with amine functionality for convenient
`coupling of reporter molecules. Taken together, the overall
`design of phosphoramidate (1) possesses key functionalities
`
`FIGURE 4. PET coronal images of male nude mice bearing
`subcutaneous LNCaP (A) and PC-3 (B) tumor xenografts 2 h
`after injection of 18F-fluorobenzamido-phosphoramidate (3).
`Arrows indicate tumor placement.
`
`0.68 nM. A similar trend was also observed for the 5FAMX–
`phosphoramidate conjugate (14), indicating that subsequent
`tagging of an imaging reporter does not adversely affect the
`inhibitory property of the leading phosphoramidate (1).
`
`In Vivo Imaging and Biodistribution Study
`18F-fluorobenzamido-phosphoramidate (3) was injected
`via a tail vein into male NCr nude mice bearing LNCaP
`PSMA1 or PC-3 PSMA2 tumor xenografts. As shown in
`Figures 4 and 5, the in vivo uptake of (3) can clearly be
`observed in the LNCaP PSMA1 model at 2 h post-
`injection. In contrast, there was no detectable tumor signal
`in the PC-3 PSMA2 xenograft. In both models, there is
`a significant uptake in kidneys but a modest degree of
`signal found in all other organs. The specificity of 18F-
`fluorobenzamido-phosphoramidate (3) to PSMA was dem-
`onstrated by the competition study with the blocking agent.
`The LNCaP PSMA1 model was injected with 1 mg of the
`nonradioactive version of fluorobenzamido-phosphorami-
`date (2) an hour before the administration of 18F-fluoro-
`benzamido-phosphoramidate (3). The resulting PET images
`showed a substantial decrease in tumor uptake at 120 min
`after injection of the imaging probe (Fig. 5). A reduction in
`kidney uptake was also observed in the animals in the
`blocking experiment.
`The ex vivo biodistribution data confirm the imaging
`findings (Fig. 6). The averaged tumor uptake in the
`PSMA1 model is 4 times higher than that of the PSMA2
`
`images of
`FIGURE 5. PET transaxial
`male nude mice bearing subcutaneous
`LNCaP (A and C) and PC-3 (B) tumor
`xenografts at 2 h after injection of 18F-
`fluorobenzamido-phosphoramidate (3).
`Arrows indicate tumor placement.
`Blocked LNCaP is shown in C.
`
`2046
`
`THE JOURNAL OF NUCLEAR MEDICINE • Vol. 50 • No. 12 • December 2009
`
`

`

`for PSMA targeting and inhibition. With the established
`modular synthetic approach, phosphoramidate (1) can be
`routinely produced in high yield.
`Phosphoramidate (1) itself is a potent PSMA inhibitor
`(IC50, 14 nM), exhibiting pseudo-irreversible inhibition that
`is common to this structural framework (13). Interestingly,
`the fluorobenzamido conjugate exhibits an enhanced in-
`hibitory potency by 20-fold. This trend is consistent with
`other conjugates such as 5FAMX (14) and presumably
`arises through the neutralization of the N-terminal amine
`through conjugation. Unlike some of the urea-based in-
`hibitors (21), the glutamate residue also acts as a spacer for
`incorporation of reporter molecules, without causing any
`adverse effects on the intrinsic PSMA inhibition. This gives
`phosphoramidate (1) the flexibility to be used as a targeting
`ligand for different imaging strategies.
`
`In Vivo Imaging and Biodistribution
`Encouraged by the in vitro inhibition data (14), we
`examined the in vivo uptake of 18F-fluorobenzamido-
`phosphoramidate (3) in PSMA1 and PSMA2 tumor models
`by PET and ex vivo biodistribution study. Our initial assess-
`ment shows that the phosphoramidate probe (3) rapidly
`localized at the LNCaP PSMA1 tumor, whereas there was
`significantly less activity found in the PC-3 PSMA2 tumor.
`The specificity of the probe was tested using fluorobenza-
`mido-phosphoramidate (2) in a competition study. When
`PSMA1 animals were pretreated with 1 mg of the non-
`radioactive analog of phosphoramidate (3) to block the
`binding pockets of PSMA, the resulting tumor uptake was
`significantly decreased. Such in vivo findings clearly
`demonstrate that the leading phosphoramidate is a PSMA-
`specific targeting ligand.
`In addition to high PSMA1 tumor uptake, the phosphor-
`amidate probe also exhibits low background and negligible
`nonspecific binding to most organs. In particular, there is
`minimal uptake in bone, suggesting that no metabolic
`defluorination or hydrolysis of phosphoramidate to phos-
`phate causing deposits in the bone hydroxyapatite matrix.
`The strong kidney uptake may in part be due to the
`hydrophilic nature of the probe, which favors renal clear-
`ance. More important, this may due to the strong PSMA
`expression found in mice that is also noted by others in the
`
`FIGURE 6. Biodistribution of 18F-fluo-
`robenzamido-phosphoramidate (3) as
`determined by radioactivity assays in
`tumor-bearing mice (n 5 3 in each
`group). Tissues were harvested at 2 h
`after injection of 18F-fluorobenzamido-
`phosphoramidate (3). Uptake values are
`expressed as %ID/g of tissue.
`
`literature (22–24). This is further supported by the re-
`duction of kidney signal by 6-fold when PSMA1 animals
`were treated in advance with the nonradioactive fluoro-
`benzamido-phosphoramidate in the blocking study. Collec-
`tively, the strong uptake in the kidneys may be confined to
`the mouse model. Nonetheless, the mouse xenografts still
`serve as a good model for the initial evaluation for our
`probe development.
`Biodistribution data confirm our findings from PET. At
`120 min,
`the phosphoramidate probe achieved tumor
`uptake of 1.24 %ID/g in the LNCaP PMSA1 model, with
`a tumor-to-blood ratio of 9:1, which is significantly higher
`than the tumor-to-background criteria set by the Food and
`Drug Administration–approved agent ProstaScint
`(3:1)
`(25). Equally important, our data show that the phosphor-
`amidate probe is specific to PSMA1 tumors, with averaged
`PSMA1 tumor uptake 4 times higher than the PC-3
`PSMA2 model (0.32 %ID/g). The PSMA specificity is
`further supported by the blocking study using a pretreatment
`of the nonradioactive probe (1 mg), with the resulting
`tumor uptake reduced by 8-fold. Compared with a pre-
`viously published 18F-labeled urea-based agent
`(6),
`
`FIGURE 7. Tumor-to-blood ratios of male nude mice
`bearing subcutaneous LNCaP and PC-3 tumor xenografts
`2 h after injection of 18F-fluorobenzamido-phosphoramidate
`(3). LNCaP blocked indicates injection in advance of 1 mg of
`nonradioactive fluorobenzamido-phosphoramidate (2).
`
`18F-LABELED PHOSPHORAMIDATE • Lapi et al.
`
`2047
`
`

`

`form may
`our phosphoramidate probe in the current
`have lower tumor uptake (1.24 %ID/g vs. 3.7 %ID/g) at
`2 h after injection; however, our probe exhibits significantly
`lower kidney and liver uptake (0.21–0.23 %ID/g vs.
`12–29 %ID/g) and a comparable tumor-to-blood ratio of 9:1.
`
`CONCLUSION
`
`Phosphoramidate (1) is a potent PSMA inhibitor with
`structural elements that closely resemble PSMA substrates.
`With L-glutamate as a spacer, the phosphoramidate scaffold
`possesses the flexibility to accommodate a variety of
`reporter molecules without compromising the PSMA in-
`hibition properties, making it a unique platform for PSMA
`targeting. As an imaging agent, the 18F-labeled phosphor-
`amidate has high uptake in the PSMA-expressing tumor
`and low background. The in vivo data also show that the
`tumor uptake is specific to PSMA. These collectively
`positive findings provide confidence that further develop-
`ment of this class of compounds will yield a highly
`sensitive and specific imaging agent for prostate cancer.
`
`ACKNOWLEDGMENTS
`
`We thank Joseph K. Choi for expert assistance with IC50
`determinations, Dr. Youngho Seo for his helpful discussions
`and advice, and William Mannone for reliable operation of
`the UCSF cyclotron. This work was supported by the
`National Cancer Institute (R21CA122126).
`
`REFERENCES
`
`1. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2008. CA Cancer J Clin.
`2008;58:71–96.
`2. Speight JL, Roach M, III. Advances in the treatment of localized prostate cancer:
`the role of anatomic and functional imaging in men managed with radiotherapy.
`J Clin Oncol. 2007;25:987–995.
`target prostate specific membrane anti-
`3. Ghosh A, Heston WD. Tumor
`gen (PSMA) and its regulation in prostate cancer. J Cell Biochem. 2004;91:
`528–539.
`4. Tricoli JV, Schoenfeldt M, Conley BA. Detection of prostate cancer and
`predicting progression: current and future diagnostic markers. Clin Cancer Res.
`2004;10:3943–3953.
`5. Foss CA, Mease RC, Fan H, et al. Radiolabeled small-molecule ligands for
`prostate-specific membrane antigen: in vivo imaging in experimental models of
`prostate cancer. Clin Cancer Res. 2005;11:4022–4028.
`6. Chen Y, Foss CA, Byun Y, et al. Radiohalogenated prostate-specific membrane
`antigen (PSMA)-based ureas as imaging agents for prostate cancer. J Med Chem.
`2008;51:7933–7943.
`
`7. Mease RC, Dusich CL, Foss CA, et al. N-[N-[(S)-1,3-dicarboxypropyl]carba-
`moyl]-4-[18F]fluorobenzyl-L-cysteine, [18F]DCFBC: a new imaging probe for
`prostate cancer. Clin Cancer Res. 2008;14:3036–3043.
`8. Pinto JT, Suffoletto BP, Berzin TM, et al. Prostate-specific membrane antigen:
`
`a novel folate hydrolase in human prostatic carcinoma cells. Clin Cancer Res.
`
`1996;2:1445–1451.
`9. Heston WD. Characterization and glutamyl preferring carboxypeptidase
`
`function of prostate specific membrane antigen: a novel folate hydrolase.
`
`Urology. 1997;49:104–112.
`10. Barinka C, Rinnova M, Sacha P, et al. Substrate specificity, inhibition and
`
`enzymological analysis of recombinant human glutamate carboxypeptidase II.
`
`J Neurochem. 2002;80:477–487.
`11. Mallari JP, Choy CJ, Hu Y, et al. Stereoselective inhibition of glutamate
`
`carboxypeptidase by organophosphorus derivatives of glutamic acid. Bioorg Med
`
`Chem. 2004;12:6011–6020.
`12. Wu LY, Anderson MO, Toriyabe Y, et al. The molecular pruning of a
`
`phosphoramidate peptidomimetic inhibitor of prostate-specific membrane anti-
`
`gen. Bioorg Med Chem. 2007;15:7434–7443.
`13. Liu T, Toriyabe Y, Kazak M, Berkman CE. Pseudoirreversible inhibition of
`
`prostate-specific membrane antigen by phosphoramidate peptidomimetics.
`
`Biochemistry. 2008.
`14. Liu T, Wu LY, Kazak M, Berkman CE. Cell-surface labeling and internalization
`
`by a fluorescent
`
`inhibitor of prostate-specific membrane antigen. Prostate.
`
`2008;68:955–964.
`15. Wust F, Hultsch C, Bergmann R, Johannsen B, Henle T. Radiolabelling of
`
`isopeptide N epsilon-(g-glutamyl)-L-lysine by conjugation with N-succinimidyl-
`4-[18F]fluorobenzoate. Appl Radiat Isot. 2003;59:43–48.
`16. Lapatsanis L, Milas G, Froussios K, Kolovos M. N-2,2,2-(trichloroethoxycar-
`
`bonyl)-L-amino acids and synthesis of N-(9-fluoroenymethoxycarbonyl)-L-amino
`
`acids in