Background
Since its discovery [
1], sigma (σ) receptor nomenclature had undergone considerable changes and has now been generally accepted as a distinct biochemical entity whose exact functions still remain elusive. There are at least two subtypes of σ receptors, σ
1 and σ
2, which are distinguished from each other by their ligand selectivity, molecular size and postulated pharmacological profile [
2‐
4].
The σ receptors have been characterised in several species, including human, and have been found unevenly distributed in the brain [
5,
6], the endocrine system including the liver, spleen, and pancreas [
4,
7] and in other peripheral organs such as the testis, ovary, heart and kidney [
8‐
10]. Sigma receptor subtypes were also found to be both region- and species-specific. For instance, σ
1 receptors were found in higher concentration compared to σ
2 receptors in the brain [
5,
6] and in the liver of guinea pig compared to other species [
4].
The σ
1 gene has been cloned from a variety of mammalian tissue sources [
11‐
15] and shares no sequence homology to any other cloned mammalian protein. The σ
1 receptor was recently identified as a receptor chaperone in the endoplasmic reticulum (ER) and is involved in ER-mitochondrial Ca
2+ signalling and cell survival [
16,
17]. To date, the structure of σ
2 receptor is not known, whose gene has also not been purified, sequenced or cloned. As a result, the exact role that this receptor has in tumour and normal cell proliferation is currently unknown. However, recent data have implicated that σ
2 binding sites is likely to be the progesterone receptor membrane component 1 (PMRMC1) or within PMRMC1 complex which shares homology with cytochrome
b
5, a heme-binding protein that activates cytochrome P450 proteins [
18]. This finding suggested an involvement of σ
2 receptors in progesterone signalling and lipid and drug metabolism [
19], and is in fact in agreement with the hypothesis previously postulated by a number of original reports [
12,
20‐
25]. However, the finding remains to be clarified due to the discrepancies in molecular weights, opposite responses to PMRCM1 antibody or σ
2 receptor antagonists and binding to P450 enzymes of PMRCM1 and σ
2 receptors [
26].
Overexpression of σ receptors in a variety of human tumour cells and tissues such as pancreatic, breast, bladder, melanoma, colon and brain [
27‐
32] has been reported. Interestingly, σ
2 expression was found to be higher than σ
1 expression in certain cell lines [
27,
33,
34]. Several studies have shown that σ
2 receptor expression is a biomarker for the proliferative status of tumour cells and solid tumours [
32,
35,
36], and are 8 to 10 times overexpressed in proliferative cells compared to quiescent cells [
32,
35]. The PGRMC1 had been reported to be more abundant in several tumours and/or tumour cell lines than in healthy control tissues including the lung [
37], breast, colon and thyroid [
38] cancer cell lines. Taken collectively, these findings make the σ
2 receptor an attractive molecular target for the development of new radiotracers not only for tumour detection and assessment of proliferative status, but also for the treatment of cancer using radiolabelled high-affinity/selectivity σ
2 ligands.
Substantial evidence also demonstrates anti-proliferative and pro-apoptotic activity of σ
2 agonists in tumour cells [
32,
35]. Sigma ligands with high affinity, but no selectivity for σ
2 receptors, have been reported since the 1990s. Over the years, the search for σ
2 selective ligands has led to the identification of a number compounds having modest to high selectivity for σ
2 receptors. To date, the majority of σ radioligands used for positron emission tomography (PET) and single-photon emission computed tomography (SPECT) imaging in cancer have either been non-subtype selective or σ
1 selective [
39]. Ligands with some selectivity for σ
2 receptors have also been reported; however, they are presenting a variety of drawbacks (as summarised in Table
1) including non-specificity (i.e. cross reactivity for other neuroreceptors), low tumour-to-background contrast, short biological half-lives or radiolabelling shortfalls.
18F-labelled radiotracers have been implicated to be superior compared to other PET radionuclide-labelled tracers in providing higher quality PET images (higher energy), fewer time constraints (adequate half-life) and in permitting longer scan sessions of higher tumour/normal tissue ratios [
40]. Therefore, the need still remains for the development of
18F-labelled σ
2 selective radioligands with optimal tumour uptake and rapid washout from non-target tissues that are suitable for PET or SPECT imaging.
Table 1
Overall review of reported selective σ2 ligands for PET and SPECT
|
125I | 554a
| 1.0a
| Tumour-to-muscle ratio = 6 (4 h, EMT-6 tumoured mice) | |
|
76Br | 5484a
| 12a
| Tumour-to-muscle ratio = 2 (2 h, EMT-6 tumoured mice) | |
|
76Br | 12900a
| 8.2a
| Tumour-to-muscle ratio = 8 (2 h, EMT-6 tumoured mice) | |
|
11C | 3078a
| 10a
| Tumour-to-muscle ratio = 3 (1 h, EMT-6 tumoured mice) | |
|
18F | 330a
| 6.9a
| Tumour-to-muscle ratio = 3 (2 h, EMT-6 tumoured mice) | |
|
18F | 2150a,c
| 0.26a,c
| Tumour-to-muscle ratio = 8 (2 h, EMT-6 tumoured mice) | |
125I, 18F | 2.8a,d
| Tumour-to-muscle ratio = 8 (2 h, Line 66 breast tumoured mice) | |
|
18F | 1711a
| 0.8a
| Tumour-to-muscle ratio = 2 (2 h, EMT-6 tumoured mice) | |
|
18F | 6b
| 2b
| Brain-to-muscle ratio = 3 (2 h, naïve mice) | |
Recently, to address the shortfalls of previously reported σ2 ligands and improve the imaging properties of new σ2 ligands, our laboratory has developed a new class of phthalimido σ2 ligands with high affinity and selectivity for σ2 receptors. The in vitro characterisation and structure-affinity analysis of these phthalimido compounds indicated that σ2 affinity is significantly enhanced by the phthalimido ring. Functionalising with halogens, such as fluorine, bromine or iodine on the phthalimido ring, would even further increase this σ2 affinity. Using this approach has identified two lead PET compounds, 18F-SIG343 and 18F-SIG353. In the current study, our aim was to explore the tumour imaging potential of the two ligands through further in vitro and in vivo investigation in mice bearing the A375 human amelanotic melanoma, a cell line that had been reported to express approximately 100 times of σ2 receptors higher than its σ1 receptor counterpart.
Methods
General
All reagents and solvents used were obtained from commercially available sources and used with no further purification. 4-Fluorophthalic anhydride was purchased from Alfa Aesar (Ward Hill, MA, USA) while 4-nitrophthalic anhydride was from Frinton Laboratories Inc. (Hainesport, NJ, USA). Nuclear magnetic resonance (NMR) spectra were performed on a Bruker Avance DPX 400 (Bruker Corporation, Billerica, MA, USA) operating at 400 MHz for 1H NMR spectra and 100 MHz for 13C NMR spectra. 18F-HF was produced on a GE PET trace via the 18O(p, n)18F nuclear reaction (RPA Hospital, Camperdown, NSW, Australia). Haloperidol was obtained from Tocris Bioscience (Bristol, UK), (+)-pentazocine from Research Biochemicals Incorporated (Natick, MA, USA) and σ ligands synthesised in house at LifeSciences, ANSTO (Lucas Heights, NSW, Australia). For cell uptake studies, drugs were dissolved in PBS buffer or in saline for injection for the animal studies, with the help of a few drops of 0.5% acetic acid. Semi-preparative high-performance liquid chromatography (HPLC) purification was performed with a Waters 600 HPLC controller (Waters Company, Milford, MA, USA) and pump, an in-line UV detector (Waters 486, 254 nm) and a single sodium diode crystal flow radioactivity detector (Carrol & Ramsey Associates, Berkeley, CA, USA) using a Pheonomenex Bondclone (Lane Cove, New South Wales, Australia) (C18, 10 μm, 300 × 7.8 mm) eluting at 3 mL/min with 30% MeCN/70% water containing 0.1% TFA.
Purity analysis and specific activity
Purity analysis and specific activity of 18F-SIG343 and 18F-SIG353 were performed on a Varian 9002 pump (Varian Medical Systems, Palo Alto, CA, USA), a linear UV-VIS detector (λ = 221 nm) in series with an Ortec ACE Mate Scint 925 γ-detector (Ortec, South Illinois Ave., Oak Ridge, TN, USA) on a Phenomenex Synergi Max-RP (C12, 4 μm, 250 × 4.6 mm) eluting at 1 mL/min with 40% MeCN/60% ammonium acetate (0.1 M) as the mobile phase. The identity of the labelled compounds was confirmed by co-injection with the authentic compounds on HPLC. For specific activity calculations, the radioactivity of the injected product for the radiochemical analysis was measured with a Capintec R15C dose calibrator (Capintec, Inc., Ramsey, NJ, USA), while the mass of SIG343 and SIG353 was determined by comparing the area of the UV absorbance peak corresponding to the carrier product with a calibrated standard curve relating its mass to UV absorbance.
Lipophilicity
The lipophilicity of SIG343 and SIG353 were assessed using RP-HPLC by determining the log
P
7.5 value using literature procedures [
46]. Samples, dissolved in methanol, were analysed using a C18 column (RP C18, Xterra, 5 μm, 4.6 × 150 mm) and a mobile phase consisting of MeOH and phosphate buffer (0.1 M, pH 7.5) compounds were estimated by a comparison of its retention time to that of standards of known log
P values.
Chemistry
Synthesis of SIG343 and SIG353
2-(4-(6,7-Dimethoxy-3,4-dihydroisoquinolin-2(1H)-yl)butyl)-5-fluoroisoindoline-1,3-dione, SIG343
A mixture of 4-(6,7-dimethoxy-3,4-dihydroisoquinolin-2(1H)-yl) butan-1-amine (132 mg, 0.5 mmol) [
47], 4-fluorophthalic anhydride (83 mg, 0.5 mmol) and p-xylene (3 mL) was stirred and gently boiled under a stream of nitrogen. As the xylene evaporated, more was added to maintain the volume. Within a few minutes, a viscous pale yellow gum had formed, but this redissolved slowly, disappearing completely after 1.25 h to form a pale yellow solution. Heating was continued for a total of 2 h (approximately 3 mL of additional xylene required), then the hot solution was treated with charcoal, filtered through celite and evaporated. The crystalline residue was recrystallised from 95% ethanol to give 177 mg (85.9%) of large colourless plates.
1H NMR (CDCl
3) δ 1.62 (m, 2H), 1.76 (m, 2H), 2.52 (t,
J = 7.5 Hz, 2H), 2.68 (t,
J = 5.8 Hz, 2H), 2.79 (t,
J = 5.7 Hz, 2H), 3.52 (s, 2H), 3.72 (t,
J = 7.0 Hz, 2H), 3.82 (s, 3H), 3.83 (s, 3H), 6.50 (s, 1H), 6.57 (s, 1H), 7.36 (ddd,
J = 8.8, 8.2, 2.4 Hz, 1H), 7.50 (dd,
J = 7.0, 2.1 Hz, 1H), 7.83 (dd,
J = 8.2, 4.5 Hz, 1H).
13C NMR (CDCl
3) δ 24.5, 26.5, 28.6, 38.1, 51.0, 55.7, 55.8, 55.9, 57.6, 109.5, 111.2 (d,
J = 24.5 Hz), 111.5, 121.0 (d,
J = 23.0 Hz), 125.6 (d,
J = 9.6 Hz), 126.1, 126.6, 127.8 (d,
J = 2.3 Hz), 135.1 (d,
J = 9.2 Hz), 147.2, 147.5, 166.4 (d,
J = 230.8 Hz), 167.1 (d,
J = 3.0 Hz), 167.8. Anal. C H N (C
23H
25FN
2O
4); theoretical C, 66.98; H, 6.11; N, 6.79; found C, 67.07; H, 6.16; N, 6.73.
2-(5-(6,7-Dimethoxy-3,4-dihydroisoquinolin-2(1H)-yl)pentyl)-5-fluoroisoindoline-1,3-dione, SIG353
4-Fluorophthalic anhydride and 5-(6,7-dimethoxy-3,4-dihydroisoquinolin-2(1H)-yl)pentan-1-amine were treated under the same reaction conditions for the synthesis of SIG343 to give the title compound as colourless crystals. 1H NMR (CDCl3) δ 1.41 (m, 2H) 1.71 (m, 4H), 2.60 (m, 2H), 2.84 (m, 4H), 3.66 (s, 2H), 3.68 (t, J = 7.1 Hz, 2H), 3.82 (s, 3H), 3.83 (s, 3H), 6.51 (s, 1H), 6.57 (s, 1H), 7.36 (ddd, J = 8.8, 8.2, 2.4 Hz, 1H), 7.50 (dd, J = 7.0, 2.4 Hz, 1H), 7.83 (dd, J = 8.2, 4.7 Hz, 1H).
Synthesis of the radiolabelling precursors
2-(4-(6,7-Dimethoxy-3,4-dihydroisoquinolin-2(1H)-yl)butyl)-5-nitroisoindoline-1,3-dione (1)
4-Nitrophthalic anhydride and 4-(6,7-dimethoxy-3,4-dihydroisoquinolin-2(1H)-yl)butan-1-amine were treated under the same reaction conditions for the synthesis of SIG343 to give the title compound as bright orange crystals. 1H NMR (CDCl3) δ 1.64 (m, 2H), 1.81 (m, 2H), 2.53 (t, J = 7.5 Hz, 2H), 2.68 (t, J = 5.8 Hz, 2H), 2.79 (t, J = 5.8 Hz, 2H), 3.51 (s, 2H), 3.79 (t, J = 7.2 Hz, 2H), 3.82 (s, 3H), 3.83 (s, 3H), 6.48 (s, 1H), 6.56 (s, 1H), 8.01 (d, J = 7.9 Hz, 1H), 8.58 (dd, J = 7.9, 2.4 Hz, 1H), 8.64 (d, J = 2.4 Hz, 1H). Anal. C H N (C23H25N3O6); theoretical C, 62.86; H, 5.73; N, 9.56; found C, 63.03; H, 5.64; N, 9.44.
2-(5-(6,7-Dimethoxy-3,4-dihydroisoquinolin-2(1H)-yl)pentyl)-5-nitroisoindoline-1,3-dione (2)
4-Nitrophthalic anhydride and 5-(6,7-dimethoxy-3,4-dihydroisoquinolin-2(1H)-yl) pentan-1-amine were treated under the same reaction conditions for the synthesis of SIG343 to give the title compound as bright orange crystals. 1H NMR (d6-DMSO) δ 1.95 (m, 2H), 2.14 (m, 2H), 2.27 (m, 2H), 2.98 (t, J = 7.0 Hz, 2H), 3.13 (t, J = 5.7 Hz, 2H), 3.22 (t, J = 5.7 Hz, 2H), 3.55 (s, 2H), 3.69 (t, J = 7.0 Hz, 2H), 4.29 (s, 6H), 7.11 (s, 1H), 7.14 (s, 1H), 8.53 (d, J = 7.9 Hz, 1H), 9.04 (d, J = 1.8 Hz, 1H), 9.09 (dd, J = 7.9, 1.8 Hz, 1H).
Radiochemistry
An aqueous 18F-fluoride solution (18F-HF, 6 to 8 GBq) was added to a 10-mL vial containing anhydrous acetonitrile (1 mL), Kryptofix 2.2.2 (2.0 mg; Sigma-Aldrich Corporation, St. Louis, MO, USA) and K2CO3 (0.7 mg). The solvent was evaporated under a stream of nitrogen at 100°C under vacuum. This azeotropic drying was repeated twice by further addition of anhydrous acetonitrile (2 × 1 mL). The nitro precursors (1) or (2) (2 mg) was dissolved in DMF (0.5 mL) and added to the dried K222.K2CO3.K18F complex. The reaction was stirred and heated at 150°C for 5 min before the reaction mixture was diluted with mobile phase (500 μL) and purified by semi-preparative reverse-phase chromatography. The peak with the retention time corresponding to SIG343 (16 min) or SIG353 (21 min) was collected and diluted with water (10 mL) and then trapped on a Waters C18 Light Sep Pak®. The trapped radiotracer was eluted off the cartridge with ethanol (1 mL). The ethanol was concentrated in vacuo and diluted with saline for in vivo studies while it was diluted with PBS (pH 7.2) for cell studies.
In vitro studies
Radioreceptor binding assays
The test compounds were solubilised in DMSO and diluted in 50 mmol/L Tris-HCl (pH 8.0). Membrane homogenates were prepared from male Sprague-Dawley rat brains as previously described [
48,
49]. The binding of σ ligands to σ
1 and σ
2 receptors was determined according to literature methods with minor modification [
50]. Briefly, the percentage of inhibition was determined by incubating, in triplicate, aliquots of diluted rat brain membrane (300 μg of protein) with 10
-11 to 10
-5 mol/L concentrations of the studied drugs in 50 mmol/L Tris-HCl (pH 8.0) with
3H-(+)-pentazocine (3 nmol/L) at 37°C for 2.5 h for σ
1 receptors or with
3H-DTG (10 nmol/L) and (+)-pentazocine (1 μmol/L) at 25°C for 1.5 h for σ
2 receptors. In both assays, non-specific binding was determined in the presence of haloperidol (10 μmol/L). After incubation, the reaction was terminated by rapid filtration using a Brandel 48-well cell harvester (Brandel, Gaithersburg, MD, USA) over Whatman GF/B glass fibre filters that were soaked in a solution of 0.5% polyethyleneimine at 4°C for at least 2 h before use. Filters were washed three times with 5 mL of ice-cold wash buffer (50 mmol/L Tris HCl, pH 7.4). The filters were collected and the amount of bound radioactivity was measured using a β-scintillation counter (Tri-Carb 2100TR, Packard Instrument Co., Downers Grove, IL, USA). The percentage inhibition of the studied drugs, at concentrations of 10
-5 and 10
-7 mol/L, was also determined for the cross activity at a number of other neuroreceptors (Caliper LifeSciences, MA, USA). The
IC
50 values were then converted to apparent
K
i values using the Cheng-Prusoff equation and radioligand
K
d values [
51].
Cell culture
A375 (human amelanotic melanoma) cells were purchased from American Type Culture Collection (Manassas, VA, USA) and cultured in RPMI-1640 medium (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% foetal calf serum (Invitrogen, CA, USA) and 2 mM l-glutamine (Sigma). Cells were maintained in 175 cm2 flasks in at 37°C in 5% CO2:95% atmosphere humidified incubator and grown to sub-confluent monolayers before being detached using trypsin for use in animal models.
Cell uptake and inhibition studies
In 24-well culture plates, 2.5 × 105 cells were seeded in complete medium and left to attach overnight. The following day, the number of cells per well was counted in triplicate. Before incubation with the radiotracer, the growth media was removed and the cells were washed once with warm PBS to remove all traces of growth media. 18F-SIG343 or 18F-SIG353 was formulated in warm PBS containing 0.1% Tween-80. Freshly prepared radiotracer (0.37 MBq in 500 μL) was added and the cells were incubated at 37°C for 2, 15, 30, 60 and 120 min. Uptake was terminated by removing the tracer solution and washing cells with ice-cold PBS. Subsequently, the cells were lysed with 500 μL of 0.2 N NaOH. The radioactivity in the cell lysate was measured with a Wallac 1480 γ-counter (PerkinElmer, MA, USA). The results were expressed as percentage of applied dose per 1 × 105 cells. All activities were corrected for decay. Optimal uptake time will be selected for inhibition cell uptake studies.
The specific uptake of the radiotracers into the cancer cells was examined in the presence of σ ligands as competitors (final concentration 5 μM): (+)-pentazocine (σ1), haloperidol (non-selective σ1/σ2) and unlabelled SIG343 or SIG353 (σ2). Cells were prepared as described for uptake studies. Prior to the addition of the radiotracer, 400 μL of blocking drug (6.25 μmol/L in PBS) or PBS (for controls) was added to the wells. Freshly prepared radiotracer (0.37 MBq in 100 μL) was added, and the cells were incubated at 37°C for 15 min as the optimal time previously determined by the cell uptake studies. Uptake was terminated as described for uptake studies. The percentage of uptake in the treatment groups relative to the control group was determined.
Ethics
Animal experiments were performed according to the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes and were approved by the ANSTO Animal Care and Ethics Committee.
In vivo studies
A375 tumour-bearing mice model
Female Balb/C nude mice aged 5 to 6 weeks old were obtained from the Animal Resource Centre (Perth, Australia). The animals were kept at a constant temperature of 22°C ± 2°C on a 12/12 h light/dark cycle with lights on at 09:00 am. Food and water were freely available. After a week of acclimatisation, mice were injected subcutaneously in the left flank with 1 × 106 A375 human amelanotic melanoma cells, in 100 μL of Ca2+/Mg2+-free phosphate buffered saline. The procedure was performed without anaesthetic. For in vivo studies, A375 tumour-bearing mice were used 26 days after tumour inoculation.
Biodistribution studies
Tumour-bearing mice were used to examine the tissue distribution of radioactivity after intravenous injection into the tail vein of 1 MBq of 18F-SIG343 or 18F-SIG353 in 100 μL of saline. At 15, 30, 60, and 120 min post injection of the radiotracer, groups of mice (n = 5) were sacrificed by CO2 administration followed by cervical fracture and dissection. Selected organs were weighed, and the radioactivity was measured using a γ-counter. The percentage of injected dose (%ID) was calculated by comparison with suitable dilutions of the injected dose. Radioactivity concentrations were expressed as percentage of injected dose per gram of wet tissue (%ID/g), assuming a uniform density of 1 g/cm3. Data were corrected for the radioactivity decay and tail injected dose and normalised for the standard mouse body weight (20 g). The remaining activity in the carcass was also determined in order to obtain the total activity in the mouse (background activity) at each time point.
Blocking studies
Blocking studies were performed to investigate the specific uptake of the tracer via σ2 receptor mechanism. Haloperidol and the unlabelled compound SIG343 or SIG353 (1 mg/kg) were administered by intravenous injection 5 min prior to radiotracer administration. Control mice received saline only. Five minutes after injection of the blocking drug, 18F-SIG343 or 18F-SIG353 (1 MBq/100 μL of saline) was injected, and groups of mice (n = 5) were sacrificed 30 min post injection of the radiotracer. Organs were processed as described for the biodistribution study. Radioactivity concentrations in the organs in the treatment groups were compared to that measured in control mice.
The amount of intact 18F-SIG343 and 18F-SIG353 in the plasma, urine, tumour and brain cortex was quantified by thin layer chromatography (TLC) and radio-HPLC analysis. Mice were injected with 20 MBq of the radiotracer in 100 μL of saline. Mice were sacrificed 15, 60 and 120 min post injection of radiotracer. Whole organ samples of brain cortex and tumour samples (10 to 60 mg, minced) were added to unlabelled SIG343 or SIG353 (5 μL; 1 mg/mL), KF (5 μL; 1 mg/mL), MeCN (0.3 mL) and water (0.2 mL). Samples were exposed to an ultrasonic probe (Ultrasonic processor, Misonix Inc., Farmingdale, NY, USA) for 2 min before being centrifuged (5,000 rpm, 5 min). Plasma (50 μL) or urine (10 μL) samples were added to unlabelled SIG343 or SIG353, KF and 0.5 mL MeCN. Plasma samples were centrifuged (Heraeus Biofuge PrimoR, Thermo Fisher Scientific, Hudson, NH, USA; 5,000 rpm, 5 min), and the supernatant was removed and the radioactivity of the precipitated pellets was measured using a gamma counter to determine the extraction efficiency. If necessary, multiple extractions were performed to ensure maximum recovery of the radioactivity. An appropriate amount of the supernatant, based on the activity level (cpm), was collected (approximately 100 μL), diluted with water (up to 1.5 mL) for HPLC analysis or evaporated to dryness under vacuum for TLC analysis.
The TLC sample was reconstituted in methanol (25 to 50 μL) and mixed before being applied to the concentrating zone of the silica TLC plate. In a separate lane, the corresponding 18F-labelled and 19F-standard was also spotted. The TLC solvent systems of EtOAc/MeOH (70:30) were utilised for 18F-SIG343 (rf 0.50) and 18F-SIG353 (rf 0.55). The UV of the standard was identified using a UV lamp, while the movement and integration of the radioactive spots were visualised and measured using a phosphorimager (BAS 2500 Phosphorimager, Fujifilm, Tokyo, Japan) with Fujifilm Multigage 3.0 software. The intact radiotracer was identified as the radioactive spot containing the identical rf value to the corresponding 19F-standard seen under the UV lamp. The integration of the active spot in relation to all the activity in the TLC lane gave the percentage of intact radiotracer.
Radio-HPLC analysis was performed following the method of Hilton et al. [
52]. A pre-column (Waters Oasis HLB, 25 μm, 20 × 3.9 mm) and a reversed phase HPLC column (Phenomenex Bondclone C18, 10 μm, 250 × 4.6 mm or Phenomenex Synergi Max-RP 80A C18, 4 μm, 250 × 4.6 mm) in series, with a switching valve between columns was utilised. The pre-column was washed with 1% acetonitrile in water for 3 min at 1.5 mL/min and then the solvent direction was switched to include the HPLC column. Both columns in series were then eluted over 10 min. The radioactivity peak corresponding to the authentic radiotracer was compared to the total activity registered in the radiochromatogram to give the fraction of unchanged radiotracer in the sample.
PET/CT imaging studies
The mouse was anaesthetised via inhalant isoflurane (Forthane, Abbott Laboratories, IL, USA) (5% induction, 1% to 3% maintenance in 200 mL/min oxygen). Respiration and heart rates of the animal were monitored (BioVet; m2m Imaging Corp, Cleveland, OH, USA) during the entire scanning period. The core body temperature of the animals was maintained via a temperature-controlled heating pad. The mouse was injected intravenously with 4.6 to 16.7 MBq of
18F-SIG343 or
18F-SIG353, with a constant mass of the unlabelled compounds of 0.06 nmol in 100 μL of saline (
n = 3). A 120-min PET scan was performed on an Inveon multimodality positron emission tomography-computed tomography (PET/CT) system (Siemens Medical Solutions, Knoxville, TN, USA) followed by a 10-min CT scan on each subject for anatomical information. Image acquisition commenced simultaneously with radiotracer injection. The data was histogrammed into 25 consecutive frames, and activity volumes were reconstructed with an iterative reconstruction (OSEM/MAP) including attenuation and scatter correction, achieving a reconstructed spatial resolution of 1.5 mm [
53]. Briefly, each individual PET scan was co-registered to its respective CT (automatic and visual quality control) (Anatomist/Brainvisa, version 3.1.4). Regional activity data (Bq/cc) were extracted from nine selected volumes/regions of interest (VOIs/ROIs) where σ receptor distribution was reported and the tail and expressed as percentage of injected dose per cubic centimetres (%ID/cc).
Data and statistical analyses
Significant outliers were identified by Grubb tests and removed from the raw data set. Subsequently, data were analysed using the GraphPad Prism 5.04 (GraphPad Software Inc., San Diego, CA, USA) statistical package software. Significance was set at P ≤ 0.05 for all statistical analyses.
In the cell uptake blocking studies, separate one-way ANOVAs [drug treatment as a factor (control, (+)-pentazocine, haloperidol and unlabelled SIG343 or SIG353)], followed by Bonferroni's post hoc tests were performed to statistically determine the significant changes in the uptake percentage of the radiotracers into the cells compared to controls.
Selected tumour-to-organ (tumour-to-blood and tumour-to-muscle) uptake ratios (TORs), were calculated by dividing the tumour radioactivity concentration (%ID/g) by the radioactivity concentration in the respective organ at time t. Comparisons of organ radioactivity concentrations (%ID/g) and TORs in the biodistribution studies or %ID/cc in PET imaging studies between the two radiotracers were performed using separate unpaired, two-tailed Student's t tests, matched organs and time points.
Meanwhile, for the blocking studies, separate within-group two-way ANOVAs [drug treatment (control, haloperidol and unlabelled SIG343 or SIG353) × regions] followed by Bonferroni's post hoc tests were performed for each radiotracer to determine any statistically significant changes in the uptake (%ID/g) amongst selected ROIs compared to the control.
Discussion
In the current study, radioreceptor binding assays identified two lead σ
2 ligands, SIG343 and SIG353, from a series of phthalimide derivatives that have high affinity and are highly selective for σ
2 receptors. From literature, it is interesting to note that the structure designed for the σ receptor also shows affinity for the related dopamine receptors [
56,
57]. However, small structural differences can greatly influence a compounds affinity and selectivity for any of the receptors. As a result, our
in vitro data indicated that the compounds have no affinity for dopamine receptors.
The potential of these tracers as imaging agents for identification of σ
2 receptor localisation
in vivo, particularly in tumours, was explored. A number of cancer cell lines that express various concentrations of σ receptors have been described; however, the level of expression of the two σ subtypes was often not reported [
39]. Amongst these cell lines, MDA-MB-231 and PC-3 cell lines were reported to express 16% and 33% of σ
2 receptors, respectively [
27]. The A375 human amelanotic cell line was reported to express high concentration of σ
2 receptors, approximately 100 times higher than that of the σ
1 subtype [
27]. The relationship of σ
2 receptors and cellular proliferation was explored and demonstrated that proliferative cells have ten times more σ
2 receptors than quiescent cells, although full down-regulation of the σ
2 receptor was reached only several (>3) days after the cells had become quiescent [
32]. Thus, the half-life of the σ
2 receptor appeared to be rather long (>12 h) [
39]. In our study we seeded the cell cultures overnight after the cells had reached 80% confluence (i.e. proliferative phase) to ensure the cells were expressing the maximum amount of σ
2 receptors in the cells, also to optimise signal-to-noise ratio in cell uptake studies. The results from the cell uptake inhibition studies are in agreement with the
in vitro receptor binding data confirming that SIG343 has a higher selectivity for σ
2 receptors, compared to SIG353.
Biodistribution studies in tumoured mice indicated that statistically significant uptake of 18F-SIG343 and 18F-SIG353 were retained in the organs known to express high densities of σ receptors such as the brain, liver, kidneys, lungs, pancreas and ovaries. As the uptake of the radiotracers in the liver and pancreas was still increasing, it is possible that the terminal half-lives of these ligands are longer than 120 min which is beyond the last time point of the study. Lack of clearance of the tracers observed at 120 min may be attributable to either continuing accumulation of the radiotracer and/or its radioactive metabolite in the liver and pancreas.
Despite some similarities in distribution profiles and clearance in the organs, the two radiotracers exhibited some differences in the extent and the rate of uptake. For instance, a significantly higher uptake of 18F-SIG343 compared to that of 18F-SIG353 was observed in the brain and the lung. However, the brain clearance was more rapid for 18F-SIG353, and the lung clearance was faster for 18F-SIG343. High to moderately high brain uptake of the radiotracers together with their appropriate lipophilicity indicated that the radiotracers are also suitable as imaging agent for receptors in the brain.
Moderately high and prolonged tumour uptake, 3% to 4% and 4% to 5% for
18F-SIG343 and
18F-SIG353, respectively, throughout the study and up to 120 min, showed promising tumour imaging potential for these radiotracers. Both radiotracers showed similar trends of increasing tumour-to-blood and tumour-to-muscle uptake ratios over time. Selective σ
2
18F-radioligands for PET imaging from the literature were reported to have tumour-to-muscle ratios ranging from 2 to 8 at 120 min in various tumoured mouse models (Table
1). Tumour-to-muscle uptake ratios of
18F-SIG343 and
18F-SIG353 are 4 and 7 at 120 min respectively, which are comparable to those reported. Similarly, the tumour-to-blood uptake ratios were high at 8 to 9 for both radiotracers as a result of the increasing radiotracer uptake in tumour over time, and the radioactivity clearance from the systemic circulation (increasing tumour-to-blood concentration ratios) or from the muscle (increasing tumour-to-muscle uptake ratios) over time. Bone uptake of both radiotracers was low and has steadily declined over time which indicates no significant
in vivo de-fluorination and subsequent fluoride bone uptake.
In the blocking study, we used only selective σ2 ligands as competitors as the σ2 specificity of these radiotracers had been previously established in naïve animals (manuscript in preparation). The competing drugs (1 mg/kg in 100 μL) including haloperidol (Ki(σ2) = 20 nM) and unlabelled SIG343 and SIG353 (Ki(σ2) = 8 and 2.7 nM, respectively) were administered 5 min prior to radiotracer injection. The concentration used deems to be in excess for all compounds used (i.e. 532, 485 and 469 nM for haloperidol, SIG343 and SIG353, respectively) which are at least 50 times greater than the K
i values of the radiotracers for σ2 receptors to ensure a complete saturation of receptor occupancy.
It is interesting to note that the uptake in the pancreas were found to increase (↑55% to 89%, P s < 0.001) in both treatment groups (haloperidol-treated and unlabelled SIG343-treated) compared to controls for 18F-SIG343. However, this increased uptake phenomenon in the pancreas was only observed in the unlabelled SIG353-treated (↑37%, P < 0.001), but not in the haloperidol-treated group, for 18F-SIG353 (P > 0.05). It is possible that pancreas is a site of radioactive metabolite accumulation, and the metabolism is a dynamic process which would make it difficult to measure specific binding in this organ. This observation was also evident in, and supported by, the biodistribution data which showed that the uptake in the liver and pancreas was constantly increasing over time. Although there was no significant increase in radioactivity of the tracer in the blood or bone as a result of the blocking drug treatment (data not shown), based on the increased radiotracer uptake values over time in many organs expressing σ receptors, we hypothesise that the continuing increase in pancreas and liver uptake is likely due to an increase in radioactive metabolites, rather than specific σ uptake.
Amongst the competing σ
2 ligands, the pharmacokinetics of unlabelled SIG343 and SIG353 is unknown and limited in rodent studies, particularly in mice for haloperidol, and the only available pharmacokinetic data of haloperidol are those in rats [
58] (terminal half-life of 1.5 h, intravenous bolus). Since haloperidol was reported to have linear pharmacokinetics within the dose range 0.5 to 2.5 mg/kg [
58], we chose the pre-administration time for challenging the ligands for 5 min in the blocking studies. Haloperidol and the unlabelled SIG343/SIG353 were able to block the uptake of the radiotracers in organs of high concentrations of σ receptors in naïve animals (data not shown), especially in the case of
18F-SIG343. This indicates the concentration and the pre-administration time parameters of the competitive ligands used in these experiments seemed to be appropriate. However, the significant blocking of radiotracer uptake by competing ligands was not observed in the tumour for both radiotracers in the current study. The discrepancy in the results of
in vitro cell uptake inhibition studies (
in vitro specificity) and those of blocking studies (
in vivo specificity) indicated that the pre-administration time of 5 min for these challenging ligands used in the current study may not be sufficient for the drugs exerting their pharmacological action. This is highly likely because in an
in vitro system, the drugs have free access to cell membrane receptors and are only limited by their affinity, appropriate concentration used and the intrinsic efficacy. However, in an
in vivo environment, in addition to those parameters mentioned above, the drugs are also limited by their pharmacokinetic properties before they can reach the target for receptor occupancy or saturation. In addition, the non-blocking of the competing drugs observed in the tumour in the present study could also be due to a number of variants such as the non-specific binding of the tracers in the tumour, internalisation or internal localisation of surface receptor-ligand complex in the tumour that was not readily blocked by the competing σ ligands when the radiotracers were administered. Other possibilities are pathophysiological status of the tumour in the animal model used including hypoxic fraction, proliferative or vasculature structure of the tumour, or even the pathophysiology of the diseased animals where the pharmacokinetics of both competing drug and radiotracer could also be compromised. Taken together, these factors could influence the absorption and distribution of competing drugs. Therefore, while competing σ ligands showed high binding affinity in the
in vitro assay system, their pharmacokinetics may not be able to reach a sufficiently high concentration, or at a sufficient time, at the binding sites in solid tumours to exert their competing effects in our
in vivo biodistribution study.
High amount of intact tracer in the tumour is desirable as it represents the stability of the radiotracer in the target. Intact 18F-SIG343 and 18F-SIG353 were 80% and 65% in the tumour, respectively at 120 min, indicating a moderate in vivo stability.
Unlike the biodistribution, PET imaging of 18F-SIG343 and 18F-SIG353 showed lower tumour uptake (0.9%ID/cc and 1.3%ID/cc, respectively), resulting in low tumour-to-background contrast which hampered the potential of the radiotracers as tumour markers. PET images showed high radioactive uptake in the abdominal cavity; hence, high background activity which is consistent with the biodistribution data where high uptake in the liver, spleen, lung, heart, pancreas and kidney was observed. The discrepancy in the results between in vivo PET and ex vivo biodistribution studies may be attributed to several reasons, including methodological differences in the techniques employed or method quantification/analysis of the images.
Firstly, methodological issues of data analysis should be considered. The percentage of injected dose (%ID/g) which is conventionally used in
in vivo studies gives an absolute index in binding which means that it reflects specific and non-specific binding in targets (i.e. tumour in this animal model), radioligand present in the systemic blood circulation and possibly radioactive metabolites in the tumour and blood circulation [
59], as the metabolisms of
18F-SIG343 and
18F-SIG353 have not been fully assessed. However, results of the metabolite studies indicated that the intact radiotracers were 65% to 80% in the tumour; therefore, the fate of the remaining fraction of the radiotracers cannot be fully concluded until the metabolites are identified and quantified. The second factor is the potential effect of inhalation anaesthesia used in
in vivo studies, but not in
ex vivo studies, which may also interfere with the kinetics of the radiotracer, especially in terms of tissue equilibrium and the extent of tissue perfusion [
60]. Finally, another factor is the difference in concentrations of the radioligand used, for instance, 1 MBq in
ex vivo studies compared with approximately 5 to 16 MBq in PET studies. High radioactivity retention in the tail (approximately 20%ID/cc) observed for both radiotracers indicated a misinjection possibility which could result in less radiotracers available for the systemic circulation, hence low radiotracer uptake in the target. Taken together, these factors could contribute to the low uptake of the radiotracers in the tumour and to the discrepancy in results between
ex vivo and
in vivo studies.
Future studies using longer scans (>120 min) and longer half-life radioisotopes compared to 18F and considering that the amount of radiotracers in the tumour remained constant, the issue of low tumour-to-background contrast could be resolved.
Authors’ contributions
VN carried out radioreceptor binding studies, analysis and interpretation of data, performed statistical analysis, and drafted the manuscript. TP synthesised and radiosynthesised the compounds, carried out metabolite studies and revised the manuscript. CF synthesised the compounds. PB, ED, GR and NH participated in the biodistribution and blocking studies. IG developed the HPLC metabolite study's method. AA analysed the PET data. FM wrote the animal ethics protocols and participated in the biodistribution and blocking studies. MG supervised the research group. AK contributed to conception of the study. RS carried out the in vitro and in vivo studies, participated in analysis and interpretation of data, and revised the manuscript. All authors read and approved the final manuscript.