A Potential Dubin-Johnson Syndrome Imaging Agent: Synthesis,Biodistribution,and MicroPET Imaging

Abstract

Dubin-Johnson syndrome (DJS) is caused by a deficiency of the human canalicular multispecific organic anion transporter (cMOAT). A new lipophilic copper-64 complex of 1,4,7-tris(carboxymethyl)-10-(tetradecyl)-1,4,7,10-tetraazadodecane (5) was prepared and evaluated for potential as a diagnostic tool for DJS. The prepared ligand was labeled with ⁶⁴Cu citrate in high radiochemical purity. In vivo uptake and clearance of the complex was determined through biodistribution studies using normal Sprague-Dawley rats and mutant cMOAT-deficient (TR⁻) rats. In normal rats, the radioactive copper complex was cleared quickly from the body exclusively through the hepatic pathway. The ⁶⁴Cu complex was taken up rapidly by the liver and quickly excreted into the small intestine and then the upper large intestine, whereas < 1% ID/organ was found in the kidney at all time points post injection. Whereas activity was accumulated continuously in the liver of TR⁻ rats, it was not excreted into the small intestine. MicroPET studies of normal and TR rats were consistent with biodistribution data and showed dramatically different images. This study strongly suggests that cMOAT is involved in excretion of ⁶⁴Cu-5. The significant difference between the biodistribution data and microPET images of the normal and TR⁻ rats demonstrates that this new ⁶⁴Cu complex may allow noninvasive diagnosis of DJS in humans.

Introduction

Dubin-Johnson syndrome (DJS) is a rare autosomal recessive disorder characterized by a defect in the transfer of endogenous and exogenous anionic conjugates from hepatocytes into the bile, which causes conjugated hyperbilirubinemia and deposition of black pigments in the liver [1]. A mutation in the human canalicular multispecific organic anion transporter gene (cMOAT) was found to cause DJS [2–5]. cMOAT, also known as multidrug resistance protein 2 (MRP2) or ATP-binding cassette (ABC) transporter ABCC2, is expressed predominantly in the liver and only weakly in the kidneys, and mediates the transport of various amphipathic anions [6]. Previously, DJS was diagnosed from increased levels of coproporphyrine isomer in the urine, a delayed rise in the sulfobromophthalein sodium test, and liver biopsy to see pigments [7]. Because of the complexity and invasiveness of these diagnostic methods, hepatobiliary scintigraphy [8,9] and CT imaging [7] were recently tried as a new diagnostic approach to DJS. In hepatobiliary scintigraphy, ^99mTc-labeled iminodiacetic acid derivatives such as ^99mTc-HIDA [10], ^99mTc-disofenin [9], and ^99mTc-mebrofenin [8] showed swift blood clearance and liver uptake followed by retardation of excretion into the bile and gallbladder.

An animal model of human DJS is the transporter-deficient (TR⁻) Wistar rat [11]. The TR⁻ rat has a defective cMOAT and shows the typical phenotypic and diagnostic features of DJS such as hyperbilirubinemia and delayed hepatic clearance of intravenously injected sulfobromophthalein [12,13].

The hepatobiliary excretion of a compound depends on several physicochemical characteristics [14,15]: an ideal molecular weight of 300-1000; the presence of a strong anionic polar group ionized at plasma pH; the presence of a nonpolar group to decrease renal excretion [16]; lipophilic character; and binding to plasma proteins, which prevents filtering by the glomeruli of the kidney and promotes transfer into the hepatocyte.

We prepared a Cu(II) complex of a DO3A derivative that fulfills the above physicochemical properties and can be labeled with copper-64 for imaging by positron emission tomography (PET). Copper-64 (t_1/2 = 12.7 hr; β⁺, 19%; β⁻, 39%) has favorable properties as a radionuclide for use in both PET imaging and targeted radiotherapy [17–19]. DO3A derivatives are known to form stable metal complexes with many metal ions and are widely used as bifunctional chelators [20] and chelating agents for gadolinium magnetic resonance imaging contrast agents [21,22].

Herein, we report the synthesis of a DO3A derivative that has a tetradecyl group; its radiolabeling with ⁶⁴Cu and log p determination, and the in vivo behavior of this ⁶⁴Cu complex as determined by biodistribution and microPET imaging.

Materials and Methods

General

1,4,7,10-Tetraazacyclododecane (cyclen) was purchased from Strem Chemicals (Newburyport, MA) and all other solvents and reagents were obtained from Aldrich (St. Louis, MO) and used as received without further purification. Water was distilled and then deionized (18 MΩ/cm²) by passing through a Milli-Q water filtration system (Millipore Corp., Bedford, MA). ¹H and ¹³C NMR spectra were measured using a Varian Gemini 300 instrument, and chemical shifts are reported in parts per million on the δ scale relative to TMS or solvent peak. Proton chemical shifts are annotated as follows: ppm [multiplicity, coupling constant (Hz), integral]. Elemental microanalyses were performed by Galbraith Laboratories (Knoxville, TN). Mass spectra were obtained from Washington University Mass Spectrometry Resource.

Copper-64 was prepared on the Washington University School of Medicine CS-15 cyclotron by a ⁶⁴Ni(p,n)⁶⁴Cu nuclear reaction at a specific activity range of 50–200 mCi/μg as previously described [18]. EM Science Silica Gel 60 F254 thin-layer chromatography (TLC) plates were purchased from Fisher Scientific (Pittsburgh, PA). Radio-TLC was accomplished using a Bioscan 200 imaging scanner (Bioscan, Inc., Washington, DC). Radioactivity was counted with a Beckman Gamma 8000 counter containing a NaI crystal (Beckman Instruments, Inc., Irvine, CA).

Female Sprague-Dawley rats were purchased from Charles River Laboratories (Wilmington, MA) and TR⁻ Wistar rats were obtained from a colony maintained at Washington University. Blood was obtained by cardiac puncture from healthy Sprague-Dawley rats. The rat serum was obtained by spinning down blood using 13 mL Vacutainer® (SST gel and clot activation, Becton Dickinson, Franklin Lakes, NJ). Serum was stored on ice before use. All animal experiments were performed in compliance with the Washington University Animal Care Committee guidelines.

Ligand Synthesis

1-(Benzyloxycarbonyl)-1,4,7,10-tetraazacyclododecane·3HCl (1·3HCl) was prepared as previously described [23].

Synthesis of 1,4,7-tris(tert-Butoxycarbonylmethyl)-10-(Benzyloxycarbonyl)-1,4,7,10-Tetraazacyclododecane (2)

To a solution of 1-(benzyloxycarbonyl)-1,4,7,10-tetraazacyclododecane·3HCl (865 mg, 2.09 mmol) in 35 mL of acetonitrile were added N,N-diisopropylethylamine (2670 μL, 15.33 mmol) and tert-butyl bromoacetate (935 μL, 6.33 mmol). The reaction mixture was slowly heated to 80°C and allowed to stir for 2 days. After evaporation of solvents in vacuo, the residue was dissolved in 20 mL of methylene chloride and washed with 30 mL of H₂O. The aqueous layer was further extracted by 2 × 20 mL of methylene chloride. The combined organic layer was dried over MgSO₄ and rotary evaporated. The residue was purified via flash column chromatography on alumina, eluting with ethyl acetate/hexane (1:2) to afford a clear oil (999 mg, 74%): ¹H NMR (CDCl₃) δ 1.37 (s, 9H), 1.38 (s, 9H), 1.39 (s, 9H), 2.68 (br s, 4H), 2.72 (br s, 4H), 2.89-2.93 (m, 4H), 3.18 (s, 2H), 3.26 (s, 4H), 3.45-3.47 (m, 4H), 5.04 (s, 2H), 7.20-7.30 (m, 5H); ¹³C NMR (CDCl₃) δ 28.19 & 28.24 (CCH₃), 47.0, 47.5, 51.8, 52.0, 52.5, 52.7, 53.6 & 53.7 (NCH₂CO₂), 54.9 (NCH₂CO₂), 57.2, 57.7, 66.7 (OCH₂Ph), 80.75 & 80.81 (OCCH₃), 127.6, 127.8, 128.4, 137.1 (Ph ring C), 156.3 (NCO₂), 170.8 & 170.9 (NCH₂CO₂). Anal. Calcd for C₂₆H₅₀N₄O₆·2H₂O: C, 59.63; H, 8.83; N, 8.18. Found: C, 59.45; H, 8.47; N, 8.27. HRMS (ESI): m/z 649.4147 ([M + H]⁺, C₃₄H₅₇N₄O₈, calcd 649.4176).

Synthesis of 1,4,7-tris(tert-Butoxycarbonylmethyl)-1,4,7,10-Tetraazacyclododecane (3) To a solution of 1,4,7-tris(tert-butoxycarbonylmethyl)-10-(benzyloxycarbonyl)-1,4,7,10-tetraazacyclododecane (999 mg, 1.54 mmol) in 35 mL of methanol was added 10% Pd on activated carbon (193 mg). The reaction was stirred under hydrogen gas at room temperature for 18 hr. The Pd catalyst was filtered off using a Celite pad and the filtrate was concentrated under reduced pressure. The residue was redissolved in 30 mL of methylene chloride and then reduced to complete dryness under vacuum (739 mg, 93%): ¹H NMR (CDCl₃) δ 1.398 (s, 9H), 1.404 (s, 18H), 2.57 (br s, 4H), 2.72 (s, 8H), 2.79 (br m, 4H), 3.27 (s, 2H), 3.28 (s, 4H); ¹³C NMR (CDCl₃) δ 28.3 and 28.4 (OCCH₃), 47.7 (CH₂CH₂NH), 51.0, 52.0, 52.4, 57.4 (NCH₂CO₂), 81.0 (OCCH₃), 171.1 and 171.4 (CO₂). HRMS (ESI): m/z 515.3798 ([M + H]⁺, C₂6H₅₁N₄O₆, calcd 515.3809).

Synthesis of 1,4,7-tris(tert-Butoxycarbonylmethyl)-10-(Tetradecyl)-1,4,7,10-Tetraazacyclododecane (4)

To a solution of 1,4,7-tris(tert-butoxycarbonylmethyl)-1,4,7,10-tetraazacyclododecane (152 mg, 0.30 mmol) in 35 mL of acetonitrile were added N,N-diisopropylethylamine (127 μL, 0.59 mmol) and 1-bromotetradecane (79 μL, 0.32 mmol). The reaction mixture was slowly heated to 70°C and allowed to stir for 64 hr. After evaporation of solvents in vacuo, the residue was dissolved in 30 mL of methylene chloride and the layer was separated between methylene chloride and 30 mL of water. The aqueous layer was further extracted with 2 × 20 mL of methylene chloride. The combined organic layer was dried over MgSO₄ and concentrated to dryness. The crude product was chromatographed using alumina to wash out impurities first by methylene chloride, then to elute the product by MeOH:methylene chloride (1:20) (101 mg, 48%): ¹H NMR (CDCl₃) δ 0.88 (t, J = 6.6 Hz, 3H), 1.25 (s, 24H), 1.45 (s, 27H), 2.36 (m, 2H), 2.59 (t, J = 5.6 Hz, 4H), 2.82 (s, 12H), 3.28 (s, 4H), 3.31 (s, 2H); ¹³C NMR (CDCl₃) δ 14.3 (CH₂CH₂CH₃), 22.9 (CH₂CH₂CH₃), 27.4, 27.9, 28.2, 28.5 (OCCH₃), 29.6, 29.9, 32.1, 52.08 and 52.28 and 52.34 and 52.58 (cyclen CH₂), 56.50, 56.57, 56.9 (NCH₂CO₂), 80.8 (OCCH₃), 171.4 (CO₂). HRMS (ESI): m/z 711.5960 ([M + H]⁺, C₄₀H₇₉N₄O₆, calcd 711.6000).

Synthesis of 1,4,7-tris(Carboxymethyl)-10-(Tetradecyl)-1,4,7,10-Tetraazacyclododecane Hydrochloride Salt (5·4HCl)

A 3 M HCl solution (20 mL) was added to 22 mg (0.031 mmol) of 1,4,7-tris(tert-butoxycarbonylmethyl)-10-(tetradecyl)-1,4,7,10-tetraazacyclododecane. The flask was slowly heated to 110°C and allowed to stir for 14 hr. The clear solution was concentrated to dryness overnight using Savant SpeedVac to give white solid 5 as a hydrochloric acid salt (16 mg, 75%): ¹H NMR (D₂O) δ 0.86 (br s, 3H), 1.26 (br s, 22H), 1.75 (br s, 2H), 3.22 (br m, 10H), 3.47 (br s, 8H), 3.60 (br s, 4H), 3.94 (br s, 2H); ¹³C NMR (D₂O) δ 16.9 (CH₂CH₂CH₃), 25.6 (CH₂CH₂CH₃), 26.0, 29.6, 32.0, 32.3, 32.48, 32.53, 32.62, 32.69, 34.8 (CH₂CH₂CH₃), 51.3 and 51.7 and 52.9 and 54.5 (cyclen CH₂), 56.9 (CH₂CO₂), 57.5 (NCH₂CH₂CH₂), 59.2 (CH₂CO₂), 172.7 and 177.0 (CH₂CO₂). HRMS (FAB): m/z 543.4119 ([M - 4HCl + H]⁺, C₂₈H₅₅N₄O₆, calcd 543.4122).

Synthesis of Cu(II) Complex of 1,4,7-tris(Carboxymethyl)-10-(Tetradecyl)-1,4,7,10-Tetraazacyclododecane (Cu-5)

To an aqueous solution (15 mL) of 1,4,7-tris(carboxymethyl)-10-(tetradecyl)-1,4,7,10-tetraazacyclododecane hydrochloride salt (5·4HCl, 17 mg, 0.025 mmol) was added 5 mL of aqueous solution of Cu(ClO₄)₂·6H₂O (11 mg, 0.030 mmol), and the mixture was stirred overnight at room temperature. The light blue powder precipitated was filtered by glass frit and washed by water, small amounts of ethanol and ether, then dried in air (3.2 mg, 21%). R_f = 0.69 (silica gel, H₂O:MeOH = 1:2): HRMS (ESI): m/z 604.3267 ([M + H]⁺, C₂8H₅₃CuN₄O₆, calcd 604.3261).

Radiochemistry

Copper-64 chloride was converted to ⁶⁴Cu citrate by stirring with 100 μL of 0.1 M ammonium citrate, pH 6.5. Copper-64 citrate (1.2-12.5 mCi) was added to a ligand solution containing less than 0.3 mg of ligand in 100 μL of the same buffer. The reaction mixture was incubated at 40°C for 1 hr. The radiochemical purity was determined by radio-TLC using silica gel plate (R_f = 0.70, H₂O:MeOH = 1:2).

Log p Measurement

The partition coefficients were determined by adding 15 μL of ⁶⁴Cu-labeled complex (21 μCi) to a solution containing 1.5 mL of octanol and 1.5 mL of buffer solution (0.01 M ammonium acetate, pH 7.4), which were obtained from saturated octanol buffer solutions. The resulting solutions were then vortexed for 1 min and centrifuged for 5 min at 5000 rpm. A 1-mL aliquot of the octanol layer was removed, back extracted with 1 mL of buffer, vortexed, and centrifuged for 2 min at 5000 rpm. Aliquots (150 μL) of octanol and buffer were removed, weighed, and counted (first partition coefficient). A 700-μL aliquot of the previous octanol layer was removed and back extracted again with 700 μL of buffer, vortexed, and centrifuged as before. Aliquots (150 μL) of octanol and buffer were removed, weighed, and counted (second partition coefficient). A 400-μL aliquot of octanol layer of second back extraction was removed and back extracted one more time with 400 μL of buffer, vortexed, and centrifuged as before. Aliquots (150 μL) of octanol and buffer were removed, weighed, and counted (third partition coefficient). The partition coefficient was calculated as a ratio of counts in the octanol fraction of unit mass to counts in the buffer fraction of unit mass per back extraction. The average log p value of the three back extractions is reported.

Protein Binding Study

The fraction of the labeled compound to proteins was determined by ultrafiltration (Amicon Centrifugal Filter Devices, Centricon YM-10, 10,000 MWCO, Beverley, MA). To 7 mL of the rat serum, 5 μL of the labeled compound ⁶⁴Cu-5 (29 μCi, radiochemical purity of 97%) was mixed and the contents were incubated at 37°C for 15 min. A control study was done in saline to ensure that the activity did not bind to the filter. Copper-64-5 (1 μL, 6.8 μCi) was diluted with 2 mL of saline. Three samples (1.5 mL each) from the serum and one aliquot (1.5 mL) from the saline solution were taken and placed in the loading cell of the Centricon. The ultrafiltration devices were centrifuged in a Sorvall Biofuge Primo R with a fixed angle rotor at 1500 × g for 30 min at 23°C to collect approximately 27% (400 μL) of the original volume of serum as ultrafiltrates. Copper-64 complex concentrations (cpm/mL) in the unfiltered protein or saline solutions and their ultrafiltrates were determined by counting measured aliquots (100 μL) in a Beckman Gamma 8000 gamma counter. The activity left on the empty Centricon membrane filters was also counted in the gamma counter to measure nonspecific binding. Serum stability of the labeled complex was analyzed by radio-TLC right after 15-min incubation at 37°C, and the serum left at room temperature after the incubation was analyzed further 5 hr later. The stability of ⁶⁴Cu-5 in the unfiltered and ultrafiltrate saline solution was also determined by the same method when the activity counting was finished. The percentage of protein-bound ⁶⁴Cu-5 was calculated as [24]:

$\%\text{Binding}=\left[1-\frac{\left(\frac{{}^{64}\text{Cu}-5\text{concentration in protein ultrafilter}}{{}^{64}\text{Cu}-5\text{concentration in unfiltered serum}}\right)}{\left(\frac{64\text{Cu}-5\text{concentration in saline ultrafilter}}{64\text{Cu}-5\text{concentration in unfiltered saline}}\right)}\right]\times 100$

Biodistribution Study

The radioactive complex was diluted with saline for injection. Mature female Sprague-Dawley rats (n = 4 per time point) weighing 180–200 g were anesthetized with 1-2% isofluorane and injected with 15 μCi of activity in a volume of 140 μL via the tail vein. At selected time points post injection (5, 15, 30 min, 1, 2, and 4 hr), rats were sacrificed. For comparison, the labeled compound was injected into mature TR⁻ Wistar rats (n = 4 per time point) weighing 300–400 g after dilution with saline (15 μCi in 150 μL each). At 30 min, 2 hr, and 4 hr post injection, rats were sacrificed. Organs of interest were removed, weighed, and counted. The percent injected dose per gram (% ID/g) and percent injected dose per organ (% ID/organ) were calculated by comparison to weighed, counted standard solutions.

MicroPET Imaging Study

The microPET imaging studies were carried out using the microPET-R4 rodent scanner (Concorde Microsystems, Knoxville, TN) [25]. Images were reconstructed using Fourier rebinning followed by 2-D filtered back projection [26,27]. A normal mature Sprague-Dawley rat weighing 139 g and a mature TR⁻ Wistar rat weighing 199 g were anesthetized with 1-2% vaporized isofluorane and injected with 1.4 mCi of activity in 140–145 μL saline via the tail vein. At specific time points, the rats were reanesthetized and then immobilized in a supine position on custom-built support beds with attached anesthetic gas nose cones for data collection. The regions of interest (ROIs) were quantified by viewing these areas in the selected tissues and averaging the activity concentration over the contained voxels. ROI analysis software consisted of two programs: Analyze AVW 3.0 (Biomedical Imaging Resource, Mayo Foundation, Rochester, MN) and a viewing application program developed in-house by R. Laforest (Washington University School of Medicine, St. Louis, MO) in International Data Language (Research Scientific, Boulder, CO).

Dynamic Imaging. The rats were first anesthetized with 1-2% isofluorane and placed near the center of the field of view (CFOV) of the microPET scanner where the highest image resolution and sensitivity are obtained. Dynamic imaging was performed for 30 min beginning at the injection of 1.4 mCi of activity in 140–145 μL saline into a normal mature Sprague-Dawley rat weighing 139 g and a mature TR⁻ Wistar rat weighing 199 g via a microcatheter (Harvard Apparatus, Holliston, MA) placed in the external jugular vein. The image frame duration was 300 sec per frame × 6 frames.

Static Imaging. For the baseline PET scans, rats were first anesthetized with 1-2% isofluorane and placed in the supine position in a custom-designed holder. The entire holder was placed near the CFOV of the microPET scanner. Ten-minute static scans were obtained 1, 2, 3, and 4 hr post injection. The methods of data collection and the reconstruction for microPET imaging have been described previously [28,29]. Briefly, all raw data were first sorted into 3-D sinograms, followed by Fourier rebinning and 2-D filtered backprojection reconstruction using ramp filter cutoff at the Nyquist frequency. An ROI was placed on organs in the transaxial microPET images that include the entire organ volume. The average radioactivity concentration within an organ was obtained from the average pixel values within the multiple ROI volume. From the dynamic study, time-activity curves were also calculated from the standard uptake value (SUV) of an organ, which was defined as the average radioactivity concentration divided by the total injected activity that is divided by body weight with the following formula:

$\text{SUV}=\frac{\text{Average radioactivity concentration in ROI}(\text{nCi}/\text{mL})}{\text{Injected dose}(\text{nCi})/\text{body weight}(\mathrm{g})}$

In these microPET scans, the attenuation corrections were not applied because the accuracy of the measured attenuation correction was poor with this scanner and the amount of attenuation from a rat body was relatively small; instead, the attenuation correction factors were incorporated into the system calibration. With the image reconstruction algorithm and filter described above, the resolution of this microPET system was between 2.0 and 3.0 mm full width at half maximum in each of the three dimensions within a 5-cm imaging field of view around the central axis of the tomograph.

Results

Synthesis, Radiolabeling, and Lipophilicity

Mono-protected cyclen compound, mono-Cbz-cyclen (1) was reacted with tert-butyl bromoacetate to give trialkylation on available nitrogen atoms on cyclen (Scheme 1). Deprotection of Cbz by hydrogenation afforded 1,4,7-tris(tert-butoxycarbonylmethyl)-1,4,7,10-tetraazacyclododecane (3) in high yield. Further alkylation of 1-bromotetradecane on one remaining nitrogen atom of cyclen and removal of the tert-butyl group by acidic hydrolysis gave the final compound in the hydrochloride salt form. Due to the intrinsic amphiphilic nature of compound 5 and its tendency to form micelles in aqueous solution, the aqueous solvent was evaporated after hydrolysis using vacuum centrifugation (Savant SpeedVac).

The prepared ligand 5 was labeled with ⁶⁴Cu in ammonium citrate buffer, pH 6.5. The radiochemical purity was determined by radio-TLC to be ≥91% when labeled at 40°C. The same reaction at room temperature gave lower labeling yield, 76%. Copper-64-labeled compound was used for the following studies without further purification.

The log p (octanol/water partition coefficient) of the radiolabeled compound was determined. Average log p values of three measurements of three consecutive back extractions was 1.18 ± 0.16.

Protein Binding Study

The bound and unbound fractions of the labeled compound were separated from serum samples by ultrafiltration. All values have been corrected for nonspecific binding of the lipophilic tracer to the filter material using binding data independently measured for ⁶⁴Cu-5 in saline solution. The nonspecific binding was less than 11% in all three measurements. The percentage of protein binding for ⁶⁴Cu-5 was calculated to be 99.2 ± 0.05%. There was no appreciable degradation of the complex in both serum and saline at room temperature even at 5 hr after 15-min incubation at 37°C.

Biodistribution Studies

In vivo uptake and clearance of the complex was determined through biodistribution studies using normal Sprague-Dawley rats and cMOAT-deficient (TR⁻) rats. The TR⁻ rat model is a rodent model of DJS, which is defective in cMOAT. Biodistribution results are summarized in Tables 1 and 2. Due to limited availability of TR⁻ rats in our facility, there were fewer biodistribution time points for TR⁻ rats than for normal rats. More direct comparisons in clearance organs of ⁶⁴Cu-5 in normal Sprague-Dawley and TR⁻ rat are reported as % ID/organ in Figures 1 and 2.

Scheme 1.

Synthesis of ligand 5.

Table 1.

Biodistribution of ⁶⁴Cu-5 in Mature Female Sprague-Dawley Rats

Tissue	5 min	15 min	30 min	1 hr	2 hr	4 hr
Blood	1.57 ± 0.275	0.81 ± 0.091	0.44 ± 0.139	0.26 ± 0.065	0.09 ± 0.014	0.08 ± 0.023
Lung	1.28 ± 0.229	0.72 ± 0.073	0.47 ± 0.115	0.29 ± 0.086	0.14 ± 0.019	0.13 ± 0.052
Liver	6.39 ± 1.157	4.98 ± 0.142	4.71 ± 1.423	2.49 ± 0.074	1.46 ± 0.173	1.03 ± 0.238
Spleen	0.74 ± 0.111	0.39 ± 0.026	0.27 ± 0.049	0.22 ± 0.051	0.12 ± 0.019	0.10 ± 0.030
Kidney	1.07 ± 0.144	0.70 ± 0.056	0.75 ± 0.157	0.82 ± 0.142	0.87 ± 0.094	0.74 ± 0.260
Muscle	0.21 ± 0.041	0.18 ± 0.028	0.12 ± 0.038	0.07 ± 0.023	0.02 ± 0.006	0.02 ± 0.005
Fat	0.17 ± 0.041	0.15 ± 0.025	0.10 ± 0.039	0.05 ± 0.019	0.02 ± 0.003	0.02 ± 0.011
Heart	0.66 ± 0.143	0.33 ± 0.019	0.18 ± 0.054	0.12 ± 0.030	0.05 ± 0.006	0.06 ± 0.019
Small intestine			10.77 ± 1.160		2.20 ± 1.556
Upper large intestine			1.78 ± 0.904		22.93 ± 3.248
Lower large intestine			0.07 ± 0.024		0.04 ± 0.004

Data are expressed as the %ID/g ± SD with four animals per data point.

The ⁶⁴Cu-complex was taken up rapidly by the liver (49% ID/organ at 5 min), and quickly excreted through the small intestine (34% ID/organ at 30 min) and then the upper large intestine (91% ID/organ at 2 hr), whereas < 1% ID/organ of activity was found in the kidney at all time points post injection. The ratio of the percentage of injected dose per gram (% ID/g) of liver versus kidney is 5.9, 7.1, and 6.3 at 5, 15, and 30 min, respectively. Blood retention was gradually decreased from 1.6% ID/g at 5 min to 0.09% ID/g by 2 hr. Because most of the activity was already excreted into the upper large intestine by 2 hr, there was very little change in the % ID/g in organs except liver between the 2- and 4-hr time points. The liver activity at 2 hr (29% ID/g) was further excreted by 4 hr (1.0% ID/g). The uptake (% ID/g) in the kidney was actually lower than that of blood by 15 min, and from 30 min to 4 hr post injection, very constant uptakes were measured (0.74-0.87% ID/g).

Table 2.

Biodistribution of ⁶⁴Cu-5 in Mature TR⁻ Wistar Rats

Tissue	30 min	2 hr	4 hr
Blood	1.00 ± 0.053	0.77 ± 0.156	0.61 ± 0.421
Lung	0.95 ± 0.043	0.70 ± 0.148	0.47 ± 0.036
Liver	1.59 ± 0.292	3.71 ± 1.365	3.36 ± 0.312
Spleen	0.57 ± 0.043	0.46 ± 0.074	0.36 ± 0.096
Kidney	0.91 ± 0.066	1.97 ± 0.439	2.82 ± 0.890
Muscle	0.15 ± 0.010	0.13 ± 0.038	0.07 ± 0.014
Fat	0.09 ± 0.015	0.06 ± 0.024	0.05 ± 0.019
Heart	0.40 ± 0.034	0.33 ± 0.075	0.21 ± 0.013
Stomach	0.22 ± 0.074	0.15 ± 0.035	0.12 ± 0.017
Small intestine	0.51 ± 0.046	0.70 ± 0.131	0.61 ± 0.118
Upper large intestine	0.30 ± 0.044	1.01 ± 0.322	1.00 ± 0.263
Lower large intestine	0.11 ± 0.002	0.21 ± 0.156	0.57 ± 0.067

Data are expressed as the %ID/g ± SD with four animals per data point.

The TR⁻ rats showed a completely different biodistribution pattern. The activity was accumulated continuously in the liver (22% ID/organ at 30 min and 42% ID/organ at 2 hr), but was not excreted into the small intestine (1.8% ID/organ at 30 min and 2.4% ID/organ at 2 hr). Although renal excretion is known to serve as the alternative route of clearance for all ^99mTc-IDA complexes in cases of reduced hepatic function [30], the uptake of activity in the kidney was very slowly increased and was only marginally higher than that of normal rats (1.3% ID/organ at 30 min and 3.8% ID/organ at 4 hr). Liver uptake nearly doubled between 30 min and 2 hr (21.6% ID/organ at 30 min and 41.7% ID/organ at 2 hr); however, only 8% of the activity of 2 hr was further accumulated in liver by 4 hr post injection, whereas only 3.8% more of the activity was excreted to the gastrointestinal track. Compared to 91% ID/organ of ⁶⁴Cu-5 in the upper large intestine of normal rats at 2 hr, less than 8% ID/organ was found in the entire gut of TR⁻ rats at the same time point. It is not clear whether blood retention and liver uptake at 4 hr were higher compared to those at 2 hr, respectively, due to the large variability of the 4-hr blood retention and 2-hr liver data.

MicroPET Imaging Studies

MicroPET was also used to follow in vivo uptake and clearance of activity. Dynamic imaging was performed for the first 30 min after ⁶⁴Cu-5 injection into a normal Sprague-Dawley rat and a TR⁻ rat, and then 10-min static images were obtained at 1, 2, 3, and 4 hr post injection. An ROI was placed on the liver and kidney in the transaxial microPET images. A time-activity curve (TAC) was generated from each ROI (Figure 3).

MicroPET images of normal and TR⁻ rats, shown in Figure 4, showed great differences in clearance between the two rats at all time points post injection. 3-D rotating images were captured as a projection image at the angles that best represented the activity uptake in organs. Sectional images could not show all liver, kidney, and intestine in the same plane.

Figure 1.

Biodistribution of ⁶⁴Cu-5 in the clearance organs of mature female Sprague-Dawley rats. The data are reported as %ID/organ ± SD (n = 4).

Discussion

We synthesized a DO3A-based cyclen compound that has a long alkyl chain. The ligand 1,4,7-tris(carboxymethyl)-10-(tetradecyl)-1,4,7,10-tetraazacyclododecane (5) was prepared from the mono-protected cyclen compound, 1-(benzyloxycarbonyl)-1,4,7,10-tetraazacyclododecane·3HCl, without difficulty. Compound 3 and its hydrobromide salt were previously prepared by direct alkylation of cyclen at 0°C [31,32]. However, compared to our current method, those synthetic procedures required intensive column purification of product from other alkylated side products to give a relatively low yield of 41-42%. The prepared ligand 5 was labeled with ⁶⁴Cu in high radiochemical purity.

The log p measurement using octanol/water partition coefficient showed that ⁶⁴Cu-labeled complex was lipophilic. Previous studies [33] in our group have shown that the lipophilicity of Cu-cyclen complexes are influenced by the overall charge of the copper complex, the symmetry within the complex, and the lipophilicity of substituents on N atoms of cyclen ring. The radiolabeled compound, ⁶⁴Cu-5, has three acetate groups, and therefore the overall charge of the Cu(II) complex can be assigned −1. Two acetate groups on the cyclen ring can coordinate to the copper to give a six-coordinated metal complex and to compensate the 2+ charge of Cu(II) ion [34], and the single remaining acetate group adds 1- to the overall charge under physiological conditions. Interestingly, another negatively charged Cu(II) complex of 1,4,7-tri(carboxymethyl)-10-(4-(tert-butyl)benzyl)-1,4,7,10-tetraazacyclododecane has very similar log p value, 1.27 [33].

Figure 2.

Biodistribution of ⁶⁴Cu-5 in the clearance organs of mature TR Wistar rats. The data are reported as %ID/organ ± SD (n = 4).

Figure 3.

Time - activity curves for liver and kidney of one Sprague-Dawley rat and one TR⁻ Wistar rat (n = 1 each).

Hepatocellular specific compounds have several common physicochemical characteristics such as range of molecular weight, the presence of a strong anionic group ionized at plasma pH, the presence of a nonpolar group, lipophilic character, and binding to plasma proteins [14,15]. The Cu(II) complex of 5 fulfills most of these requirements. The molecular weight of Cu-5 is 604, which is in the range of 300-1000. It has one polar acetate group that will be ionized at physiological pH in addition to two acetate groups that will coordinate to the copper ion, and it also has a long alkyl chain that is hydrophobic. The log p value of 1.18 showed that the overall complex Cu-5 is lipophilic.

Biodistribution studies showed that the radioactive copper complex was cleared quickly from body exclusively through the hepatic pathway in normal Sprague-Dawley rats (Figure 1). Kidney uptake of this Cu(II) complex was lower than that of clinically used hepatobiliary specificity reagents such as disofenin (DISIDA) [35], even though blood pool and liver clearance of Cu-5 are much slower than those of the ^99mTc-IDA derivatives [36]. This comparatively slow blood clearance and very low renal excretion could be explained by the plasma protein binding. The protein binding study done using ultrafiltration showed more than 99% of ⁶⁴Cu-5 bound to plasma proteins with 15 min incubation at 37°C. Even though the results of protein binding experiments of radiopharmaceuticals can be highly related to methodology [37] and the in vitro protein binding studies are not fully predictive of the extent of binding under clinical conditions [38], our ultrafiltration data showed substantial binding of the lipophilic ⁶⁴Cu-5 to blood proteins. Nunn's systematic structure-biodistribution study [39] showed that higher protein binding by radiopharmaceuticals reduced their renal excretion. In addition, it is also generally accepted that radiopharmaceuticals with low protein binding have faster blood clearance [40]. For example, ^99mTc-trimethylmonoiodo-IDA, which has lower protein binding, showed faster blood clearance and lower urinary excretion compared to another ^99mTc-IDA derivative, DISIDA [41].

Figure 4.

Comparison of microPET images of a Sprague-Dawley rat (top) and a TR⁻ Wistar rat at 5 min, 25 min, 1 hr, and 3 hr post injection of ⁶⁴Cu-5.

No significant uptake was found outside of the clearance organs. Even though ⁶⁴Cu-5 shares several common structural features with fatty acids such as a long alkyl chain, polar acetic acid end, and amphiphilicity, heart uptake was less than blood at all time points.

The biodistribution data of ⁶⁴Cu-5 in TR⁻ rats strongly suggest that cMOAT is required for excretion of this complex (Figure 2). In TR⁻ rats, liver clearance was severely retarded due to the lack of this transporter, demonstrated by the fact that only a small amount of activity was excreted from the liver to the gastrointestinal tract. It is not clear if there is any negative feedback mechanism that controls delivery of ⁶⁴Cu-5 from the blood to the liver when that compound keeps accumulating without excreting to bile, or if it is just a matter of a less effective carrier-mediated anionic clearance mechanism in TR⁻ rats [30]. whereas 49% ID/organ of ⁶⁴Cu-5 was accumulated in liver within 5 min in normal Sprague-Dawley rats, only 42% ID/organ was found in liver at 2 hr in TR⁻ rats, and even then, there was insignificant liver clearance to the intestine. Even though the kidney seemed to be used as an alternative excretion organ for Cu-5 in TR⁻ rats, kidney uptake was less than 3% ID/g at all time points evaluated. Copper-64-5 seemed to bind to plasma proteins very tightly and the resulting adduct was too big to pass through the glomerulus in Bowman's capsule [15]. Except for excretion organs, the biodistribution of ⁶⁴Cu-5 in other organs in TR⁻ rat was comparable with that of normal rats, which indicates that rat strain difference between Sprague-Dawley and Wistar rats had minimal influences on the biodistribution of the radiolabeled Cu(II) complex (Tables 1 and 2).

MicroPET imaging analysis data of normal and TR⁻ rats were consistent with biodistribution data and showed a more detailed clearance scheme for liver and kidney (Figure 3). Especially at early time points, TAC of dynamic scan images provided very precise uptake and clearance information of liver and kidney because the bed position of the rat remained at the exact same position, and therefore the exact same ROI was applied.

In normal rats, liver uptake of ⁶⁴Cu-5 increased swiftly to reach its maximum at 17.5 min, and most activity was cleared by 1 hr. The liver uptake at 5 min was higher than at 15 min. This discrepancy between microPET and biodistribution data can be explained when we consider the time needed for dissection and organ harvest in the biodistribution study. Otherwise, the general liver clearance data from the microPET study are consistent with that of the biodistribution study. Kidney uptake was highest at the beginning (2.5 min) and gradually decreased over time to 27.5 min. The kidneys could not be identified on the microPET image beyond 1 hr post injection due to the low ratio of kidney uptake to background. The initially high kidney uptake seemed to be more of a reflection of blood retention of activity rather than actual kidney uptake. In the biodistribution study, the %ID/kidney was found to be less than 1% at all time points, and the %ID/g of kidney was lower than that of blood at 5 and 15 min. The relatively high kidney uptake immediately after injection of ⁶⁴Cu-5 and gradual excretion can be explained in another way. The free, intact, radiolabeled complex ⁶⁴Cu-5 is small enough to pass through the glomerulus and accumulate in the kidneys at early time points, but as soon as plasma proteins bind with the labeled complex and deplete the free ⁶⁴Cu-5 compound, the kidney uptake dropped dramatically because the plasma protein-⁶⁴Cu-5 adduct was too big to escape from the glomerulus in Bowman's capsule [42].

MicroPET imaging of a single TR⁻ rat and subsequent data analysis provided detailed information on in vivo uptake and clearance (Figure 3), whereas a biodistribution study would have required a much larger number of these transgenic rats. The microPET study showed that ⁶⁴Cu-5 continued to accumulate in liver over time up to 2 hr and slowly decreased by 4 hr in TR⁻ rat. Liver clearance from 2 to 4 hr may be due to metabolic degradation of ⁶⁴Cu-5 and re-release back into the blood stream, because the activity in intestine at 4 hr was not greatly increased compared to intestinal activity at 2 hr as shown in the biodistribution study. The intestines of the TR⁻ rat were barely visualized by microPET imaging. Kidney uptake was low during the first 30 min, and then increased gradually by 4 hr, which is consistent with the biodistribution data. As in the normal Sprague-Dawley rat, kidney clearance during the first 30 min seemed to reflect blood retention rather than kidney uptake itself. However, it is not clear if kidney uptake after 30 min is from intact ⁶⁴Cu-5 complex or free ⁶⁴Cu released from the liver by metabolic degradation. Further metabolism studies were not carried out.

In the micro PET images of the normal Sprague-Dawley rat (Figure 4), most of the activity was accumulated in the liver at 7.5 min then excreted rapidly into the intestines. Kidney was not easily identified in microPET images at all time points due to low uptake. However, the bladder was clearly visualized at 7.5 min, which suggests that even though the kidney uptake was low, some activity was still cleared through the renal tubules, which is presumably due to the presence of a small amount of cMOAT in kidney [6].

In the TR⁻ rat, however, liver and kidney were both visualized at all time points (Figure 4). However, the images of liver and kidney at 7.5 and 27.5 min were not clear due to the high retention of activity in blood. The intestines were barely visualized due to severe retardation of liver clearance into the small intestine. At 1 hr, most of the activity remained in the liver and kidney in the TR⁻ rat, which was very different from the microPET image of the normal rat, in which only the intestines were visualized. At 3 hr post injection, the liver and kidney in the TR⁻ rat was more easily visualized than in the earlier images, which was likely due to decreased background activity. Due to the limited availability of the TR⁻ rats, we could not carry out multiple microPET studies.

It is known that the bile flow rate in TR⁻ rat liver is 50% that of normal rats [43]. However, the secretion of a nonsubstrate of cMOAT such as bile salt was normal in TR⁻ rat. Also, some organic anions showed relatively normal excretion patterns in TR⁻ rats. For example, the hepatobiliary secretion of bilirubin ditaurate in TR⁻ rats was entirely normal [43]. Although the bile flow of TR⁻ rats was reduced to 50% of normal, the biliary concentration of bilirubin ditaurate in TR⁻ rats was twice that in normal rats. Therefore, the substrate specificity of cMOAT can be deduced in TR⁻ rats from compounds that are not secreted or poorly secreted into the bile [12,43]. The data from biodistribution and microPET studies presented here show the severe retardation of secretion of ⁶⁴Cu-5 in TR⁻ liver, strongly suggesting that this new copper complex is transported from liver into bile via cMOAT. However, it was not further studied whether ⁶⁴Cu-5 was excreted into bile as an intact compound, or first converted by the liver into glucuronidated, sulfated, or glutathione conjugates then excreted via cMOAT, similar to what is observed with aminacetophen [12].

Copper-64 is one of a limited number of positron-emitting radionuclides that can be produced on both a nuclear reactor and a medical cyclotron. A method was developed at Washington University School of Medicine to produce large quantities of ⁶⁴Cu with low-energy biomedical cyclotrons [18]. Copper-64 offers the advantages of a positron-emitting tracer for imaging and has potential for use as an internally administered therapeutic radionuclide. The potential wide availability of ⁶⁴Cu, both from reactors and biomedical cyclotrons, suggests that ⁶⁴Cu is a radionuclide with great potential for the clinical use. The human studies of ⁶⁴Cu-labeled anticolorectal carcinoma monoclonal antibody (MAb 1A3) [44,45] and ⁶⁴Cu-TETA-OC (where TETA is 1,4,8,11-tetraazacyclotetradecane-N,N′,N¶ime;,N′¶ime;-tetraacetic acid and OC is d-Phe¹-octreotide) [46] showed that these are promising radiopharmaceuticals for PET imaging of patients with smaller colorectal tumors and neuroendocrine tumors, respectively.

Owing to its availability from a generator, Cu-62 (β⁺ 97.6%) might be used for PET imaging in centers relatively remote from a cyclotron unit [47]. However, in order to use ⁶²Cu instead of ⁶⁴Cu in case of Cu-5, the radiochemistry of Cu-5 should be optimized first to decrease the incubation time (1 hr) due to the short half-life of ⁶²Cu (t_1/2, 9.7 min).

Conclusion

In conclusion, we synthesized a novel DO3A derivative having a long alkyl chain and labeled it with ⁶⁴Cu. The log p measurement using octanol/water showed that the labeled Cu(II) complex was lipophilic, whereas the complex was assumed to carry a −1 charge. In normal rats, ⁶⁴Cu-5 cleared quickly through the liver into the intestines, whereas in TR⁻ rats, ⁶⁴Cu-5 accumulated in the liver and kidney with very little excretion into bile and urine. Biodistribution and microPET studies showed that the lipophilic, anionic ⁶⁴Cu-5 was excreted out of the body by aid of cMOAT. The significant difference between the microPET images of the normal and TR⁻ rats further demonstrated that this new ⁶⁴Cu complex might allow noninvasive diagnosis of DJS in humans.