Simple ammonium salts acting on sigma-1 receptors yield potential treatments for cancer and depression

Sigma-1 and sigma-2 receptors are emerging therapeutic targets. We have identified that simple ammonium salts bind to these receptors and are effective in vivo. Radioligand binding assays were used to obtain structure-activity relationships of these salts. MTS assays were performed to determine their effect on growth in MCF7 and MDA-MB-486 cells. Anticancer properties were tested in NMRI mice transplanted with a fragment of mouse adenocarcinoma (MAC13). Antidepressant activity was tested using the forced-swim test and tail suspension tests. Dipentylammonium (Ki 43 nM), tripentylammonium (Ki 15 nM) and trihexylammonium (Ki 9 nM) showed high affinity for the sigma-1 receptor. Dioctanoylammonium had the highest affinity (K50 0.05 nM); this also showed the highest affinity for sigma-2 receptors (Ki 13 nM). Dipentylammonium was found to have antidepressant activity in vivo. Branched-chain ammonium salts showed lower affinity. Bis(2-ethylhexyl)ammonium (K50 29 µM), triisopentylammonium (K50 196 µM) and dioctanoylammonium showed a low Hill slope, and fitted a 2-site binding model for the sigma-1 receptor. We propose this two-site binding can be used to biochemically define a sigma-1 receptor antagonist. Bis(2-ethylhexyl)ammonium and triisopentylammonium were able to inhibit the growth of tumours in vivo. Cheap, simple ammonium salts act as sigma-1 receptor agonists and antagonists in vivo and require further investigation.


Results
Binding to the sigma-1 and sigma-2 receptors. Previous data have shown that MDA-MB-468 cells contain high concentrations of sigma-1 receptors (pK d ± SEM of [ 3 H] (+) pentazocine 7.8 ± 0.6, K d 17 nM, B max 2300 ± 200 fmol.mg −1 ) 19 . In contrast, no [ 3 H] (+) pentazocine binding was observed in MCF7 cells, identifying these cells to be a good model to study sigma-2 receptor binding. This is in agreement with previous studies 9 . The binding of [ 3 H] DTG to MCF7 cells showed a pK d ± SEM of [ 3 H] DTG 7.92 ± 0.04, K d = 12 nM, B max 1700 ± 400 fmol.mg −1 . We have recently shown that [ 3 H] DTG binding is reduced by (+) pentazocine with low affinity, further confirming the absence of sigma-1 receptors from MCF7 cells 20 .
This systematic study compared affinities of 34 ammonium salts for the sigma-1 and sigma-2 receptors. Those of interest were further assessed for cross-reactivity with other systems and in vivo biological activity.
Primary ammonium salts showed increasing affinity associated with increasing molecular weight. The highest affinity primary ammonium salt we tested was decylammonium with a K i of 310 nM (Table 1, Supplementary  Table ST1, Supplementary Fig. SF1). Affinity for the sigma-2 receptor showed a similar graded but less dramatic Ammonium salt σ-1 R affinity pK i ± SEM (n) σ-1 R affinity (µM) σ-2 R affinity pK i ± SEM (n) σ-2 R affinity (µM) σ-1 R K i /σ-2 R K i  Table 1.
Affinities of key ammonium salts tested. Affinities for the sigma-1 and sigma-2 receptors (σ-1 R and σ-2 R, respectively) are shown as pK i and K i (µM) values from competition binding assays. Hill slopes were determined as unity for all salts tested, except those labelled *. σ-1 R K i /σ-2 R K i shows the ratio between K i values. Sigma-1 receptor affinity has been previously published 17 . These data are reproduced from the PhD theses by J.M. Brimson 44 and H. Abbas 45 .
improvement. Decylammonium was found to have the highest affinity with a K i of 3.9 µM (Table 1, Supplementary  Table ST1, Supplementary Fig. SF1). Secondary ammonium salts bound with much higher affinity to the sigma-1 receptor. Affinity for the sigma receptors increased with chain length, with dioctylammonium showing the highest affinity (K 50 0.05 nM and 13 nM, for sigma-1 and sigma-2 receptors respectively) ( Table 1, Supplementary Table ST1, Supplementary  Fig. SF2). Of the compounds tested, dioctylammonium showed the greatest selectivity (245-fold) towards the sigma-1 receptor over the sigma-2 receptor. The binding of dioctylammonium to sigma-1 receptors showed a particularly low Hill slope (0.43 ± 0.02, mean ± SEM, n = 4). A two-site binding model fits the data better ( Supplementary Fig. SF3), giving pK i high 11.9 ± 0.3 (50 ± 7% of total binding) and pKi low 9.4 ± 0.2.
Tertiary ammonium salts also showed a graded increase in affinity. Tripentylammonium and trihexylammonium were found to have the highest affinity with K i values for the sigma-1 receptor of 15 nM and 9 nM, respectively. The affinity for sigma-2 receptors showed a very similar pattern; trihexylammonium was found to have highest affinity with K i of 230 nM. Increasing the chain length above six carbons was detrimental to binding to both sigma-1 and sigma-2 receptors (Table 1, Supplementary Table ST1, Supplementary Fig. SF4).
All quaternary ammonium salts tested had low affinity for the sigma-1 receptor. Tetrabutylammonium bound with highest affinity with a K i of 1.6 mM, and there was little increase in affinity as the carbon chain length increased, therefore no larger quaternary ammonium salts were tested for sigma-1 receptor affinity. Similarly, quaternary ammonium salts showed low affinity for the sigma-2 receptor (Table 1, Supplementary Table ST1, Supplementary Fig. SF5).
In addition to these straight-chained ammonium salts, 10 branched-chain salts were tested. These tended to have lower affinity for the sigma-1 receptor than comparable straight-chained analogues ( Table 1, Supplementary  Table ST1, Supplementary Fig. SF6). Their affinity for the sigma-2 receptor was not as greatly reduced. Triisopentylammonium showed the greatest selectivity towards the sigma-2 receptor, with a 200-fold higher affinity for sigma-2 receptors over sigma-1 receptors. The binding of most straight-chained ligands to the sigma-1 receptors showed a simple displacement with a Hill slope of unity. In contrast, the branched-chain ammonium salts appeared to have shallower Hill slopes. A one-sample t test (GraphPad Prism) showed three had a Hill slope significantly less than unity at the sigma-1 receptor: bis(2-ethylhexyl)ammonium (0.24 ± 0.14, P = 0.001, n = 5), di-sec-butylammonium (0.41 ± 0.12, P = 0.040, n = 3) and triisopentylammonium (0.42 ± 0.12, P = 0.008, n = 6). Dicyclohexylammonium showed a low, but not statistically significant Hill slope (0.4 ± 0.2, P = 0.052, n = 5). The binding of bis(2-ethylhexyl)ammonium and triisopentylammonium was found to fit a 2-site model in preference to a one-site fit. Bis(2-ethylhexyl)ammonium was found to have pK i high 6.4 ± 0.5 (18 ± 6% of total binding) and pK i low 4.4 ± 0.3. Similarly, triisopentylammonium was found to have pK i high 5.6 ± 0.4 (70 ± 13% of total binding) and pKi low 1.8 ± 0.2. (Supplementary Fig. SF3). Di-sec-butylammonium was found to best fit a one-site model.
The Hill slopes of binding to the sigma-2 receptor showed less of a pattern: eleven of the 34 tested showed a Hill slope significantly below unity (Supplementary Table ST1). All agents were found to best fit a one-site binding model at the sigma-2 receptor. The high affinity of some of these simple ligands led us to further examine their profile is a series of tests. Cellular proliferation. The MTS assay was used to determine whether our sigma receptor ligands affected cellular metabolism in MDA-MB-468 cells, which express both sigma-1 and sigma-2 receptors and MCF7 cells, which appear to express only sigma-2 receptors. All but one of the straight-chain ammonium salts were unable to decrease cell proliferation at concentrations 10 times above the K i for the relevant receptor (data not shown). Due to its high affinity, dioctylammonium was tested at concentrations well above its K i and it was found able to reduce cellular proliferation, albeit with pIC 50 value of 5.0 ± 0.2, well above our reported affinity (pK i for the low affinity state 9.4).
Ammonium salt specificity for the sigma receptors. Having determined that these simple ammonium salts are non-toxic in vitro and some have high affinity for the sigma receptors, we sought to determine whether these agents also bound other targets. We chose to study dipentylammonium and tripentylammonium salts from the straight-chained ammonium salts and bis(2-ethylhexyl)ammonium and triisopentylammonium from the (+) pentazocine (c) are also shown. SiRNA targeting the sigma-1 receptor caused a shift in the pIC 50 for bis(2ethylhexyl)ammonium from 3.37 to 2.7 (P = 0.044, unpaired t-test), and for triisopentylammonium from 2.9 to 2.3 (P = 0.0095, unpaired t-test). Error bars show SEM, n = 3. SiRNA also caused a reduction in the sigma-1 receptor number, as monitored using a fixed concentration (30 nM) of [³H] (+) pentazocine (P < 0.0001, unpaired t-test, n = 8).This figure is reproduced from the PhD thesis by J.M. Brimson 44 . Panel (c) has been previously presented 19  www.nature.com/scientificreports www.nature.com/scientificreports/ branched-chain ammonium salts. In terms of "druglikeness", the free amine dipentylamine fits well into the limitations proposed by Lipinski 21 , including a partition coefficient (logP) value of 3.82. In contrast, tripentylamine, bis(2-ethylhexyl)amine and triisopentylamine, have logP values above 5 (6.44, 6.64 and 5.89, respectively) (ACD prediction, (www.chemspider.com)), outside the Lipinski guidelines. The partition coefficients of the ammonium salts at pH 7.4 (logD (7.4)) was as follows: dipentylammonium 0.65; tripentylammonium 3.73; bis(2-ethylhexyl) ammonium 3.38 and; triisopentylammonium 2.68 (ACD prediction, (www.chemspider.com)). According to Clark's 5 "rules of thumb" a logD (7.4) of 1-3 is favoured. In addition, he suggests that a charge of +1 aids access to the brain 22 .

Straight-chained ammonium salts as sigma-1 receptor agonists and antidepressants.
We have determined that our simple ammonium salts show moderate to high affinity for the sigma receptors. Based on MTS data with MCF7 cells, they appear to show no toxicity. We have, therefore assessed these compounds for their behavioural effects using the forced-swim test 23 and the tail suspension test 24 , the most used animal models for screening antidepressants.
Five groups of animals (5 animals per group) were submitted to the forced-swim test. Control mice (10 ml. kg −1 saline) exhibited despair behaviour, remaining still (only making movements to keep their head above the water) for 233 ± 10 seconds out of a total of six minutes in the water. The selective serotonin reuptake inhibitor, fluoxetine (10 mg.kg −1 ), significantly reduced the time spent immobile to 153 ± 19 seconds (P = 0.0022). Similarly, dipentylammonium (5 mg.kg −1 ) significantly reduced the time spent immobile to 80 ± 13 seconds (P < 0.0001) (Fig. 3). The sigma-1 receptor antagonist N'-[2-(3,4-dichlorophenyl)ethyl]-N,N,N'-trimethylethane-1,2-diamine (BD1047) alone at 4 mg.kg −1 had no significant effect, immobility time being 217 ± 17 seconds (P > 0.99). However, a 4 mg.kg −1 dose of the sigma-1 receptor antagonist BD1047 prevented a 5 mg.kg −1 dose of dipentylammonium from having any effect, immobility time was 185 ± 5 seconds (P = 0.0001, compared to dipentylammonium alone), indicating that the effects of dipentylammonium were reversed (Fig. 3). In contrast, tripentylammonium did not have any significant effect on the despair like behaviour of the mice (at doses up to 10 mg.kg −1 ), as measured by the immobility time in the forced-swim test, and was therefore not tested further.
Effects of branched-chain ammonium salts of tumour growth. MAC13 assay; in vitro assay. We have shown above that bis(2-ethylhexyl)ammonium and triisopentylammonium both cause a reduction of metabolic activity of MDA-MB-468 cells in the MTS cell survival assay. We subsequently wanted to test their effects  www.nature.com/scientificreports www.nature.com/scientificreports/ on tumours in vivo using MAC13 cell transplant 28 . Initially MAC13 cells were tested for sigma 1 receptor expression, using radioligand saturation binding. [ 3 H] (+) pentazocine saturation resulted in mean ± SEM B max of 1310 ± 140 fmol.mg −1 , and the mean ± SEM pK d was 7.4 ± 0.1 (n = 3). Both bis(2-ethylhexyl)ammonium and triisopentylammonium caused a reduction in MAC13 cell metabolism in vitro (Fig. 6), with mean ± SEM pIC 50 (n = 6) values of 5.3 ± 0.2 and 4.4 ± 0.2 respectively. In this respect, MAC13 cells are more sensitive to our ligands than MDA-MB-468 cells (pIC 50 3.3 and 2.96, respectively, see above). www.nature.com/scientificreports www.nature.com/scientificreports/ MAC13 assay; in vivo assay. MAC13 tumours were established in the mice 12 days before treatment with the ammonium salts began. Daily treatment with 10 mg.kg −1 of bis(2-ethylhexyl)ammonium and triisopentylammonium resulted in a statistically significant reduction in tumour growth compared to vehicle control treated animals (P < 0.001) (Fig. 7), suggesting that the two ammonium salts exhibit anti-tumour properties similar to that seen with the sigma-1 receptor antagonists IPAG and rimcazole 11,12 . In addition, the animals showed no negative signs to the use of these drugs: grooming activity was maintained during the study (data not shown).

Discussion
We have performed structure-activity relationship study using 34 simple ammonium salts for sigma-1 and sigma-2 receptor affinity. Our molecules are far more simple than the pharmacophore model previously proposed for the sigma-1 receptor 29 . The major findings of this study were that several salts show high affinity for these receptors. Selectivity between the ammonium salts ranged from 245-fold preference for the sigma-1 receptor (dioctylammonium) to a 200-fold preference for the sigma-2 receptor (triisopentylammonium). We also identified branched-chained ammonium salts have lower affinities than the straight-chained ammonium salts for the sigma-1 receptor. However, branched-chain ammonium salts, along with the outlier dioctylammonium, showed a binding profile more in line with what has been previously seen with antagonists such as IPAG and rimcazole 19 .
The binding profile of all-but-one (dioctylammonium) of the straight-chained ammonium salts showed a simple profile with Hill slope of unity, whereas branched-chain ammonium salts bound with a low Hill slope. This appears to be a novel means to distinguish antagonists from agonists using in vitro binding assays and identifies most straight-chained ammonium salts as agonists and the branched-chain ammonium salts as antagonists at this receptor. Further evidence is provided by the observation that, while binding the receptor, the straight-chained ammonium salts did not effect calcium influx or reduce metabolic activity (excepting dioctylammonium, which reduced metabolic activity; we were unable to obtain calcium data for this compound, as precipitation was observed in this assay), whereas some of the branched-chain ammonium salts did, as is the case for antagonists 11 .  www.nature.com/scientificreports www.nature.com/scientificreports/ Quaternary ammonium salts showed very low affinity for both sigma receptors. However, tetrapropylammonium and tetrabutylammonium are frequently used at mM concentrations for blocking potassium channels 30 and NMDA channels 31 . The effects of these quaternary ammonium salts on sigma receptors could be relevant when they are being used to block the channels.
It has been shown previously that sigma-1 receptor ligands such as (+) pentazocine and 1 S,2R-cis-N-[2-(3,4-dichlorophenyl)ethyl]-N-methyl-2-(1-pyrrolidinyl)-cyclohexylamine (BD737) are capable of affecting muscarine-induced intracellular calcium changes in SH-SY5Y neuroblastoma cells which have been shown to express the M 1 and M 3 subtypes of the muscarinic acetylcholine receptor 32 . These changes could not be reversed by sigma-1 receptor antagonist haloperidol. Therefore, it was deduced that (+) pentazocine (pK i 6.29 ± 0.06 32 ) and BD737 were antagonists at the muscarinic receptor. Our binding data suggest that these simple ammonium salts have lower affinity for the muscarinic receptor than rimcazole and (+) pentazocine, suggesting they are more selective than this widely used ligand. Dipentylammonium has a pK i of 7.4 for the sigma-1 receptor 17 , which is significantly higher than that of rimcazole (pK i 6.5). It also has a significantly lower affinity for the muscarinic receptor with a pK i 3.4 compared to rimcazole with a pK i of 4.5. Therefore, dipentylammonium shows greater selectivity than (+) pentazocine, haloperidol and rimcazole.
We have previously shown that IPAG regulates intracellular calcium in a manner regulated by cholera toxin 19 . The concentrations required to effect calcium signalling (pEC 50 = 3.9, EC 50 123 µM) and reduce cellular metabolic activity (pEC 50 = 4.6, EC 50 24 µM) were significantly higher than the equilibrium binding constants (pK i high 12.8, K i high 175 fM and pK i low 6.8, K i low 168 nM) 19 . We describe similar observations here: Bis(2-ethylhexyl) ammonium, dioctylammonium and triisopentylammonium all required concentrations higher than their binding affinities to effect a decrease in cellular metabolism; only dicyclohexylammonium required similar concentrations. This suggests that the equilibrium binding affinities of certain compounds may not represent the binding affinity to the active form of a receptor. The induction of a high-affinity state accompanied by desensitisation following prolonged incubation with ligands has previously been observed with a limited number of other receptors, including the nicotinic acetylcholine receptor 33 , µ-opiate receptor 34 and GABA receptor 35 . Whilst outside the remit of this publication, we suggest that these observations are further investigated.
There is much evidence to suggest that the sigma-1 receptor is involved in depressive-like behaviours. Sigma-1 knock-out mice show a depressive-like phenotype 36 and previous studies have identified sigma-1 receptor agonists, such as igmesine, SKF-10,047, and dehydroepiandrosterone sulphate, have antidepressant-like properties in the forced-swim test assay; furthermore, sigma-1 receptor antagonists such as progesterone or BD1047 reverse these effects 37 . It has previously been shown that dipentylammonium has sigma-1 receptor dependant neuroactive properties 17 . Our in vivo results suggest that one of these simple ammonium salts, dipentylammonium, is a potential antidepressant acting through the sigma-1 receptor, since the antidepressant-like effects were reversed by BD1047.
Dipentylammonium required very high doses to have any effects on monoamine transporters, excluding them as targets for the effects observed. Neither dipentylammonium per se nor in combination with BD1047 had an effect on locomotion suggesting that neither have psychostimulant or psychoinhibitory effects, providing strong evidence that dipentylammonium causes its antidepressant effects through a sigma-1 receptor-mediated mechanism.
Despite having an affinity similar for the sigma-1 receptor to that of dipentylammonium, tripentylammonium failed to reduce despair behaviour in mice subjected to the forced-swim test. This can be explained by the physicochemical properties of the two salts: dipentylammonium met the guidelines proposed by Lipinski 21 , whereas tripentylammonium did not.
Our in vitro data for some ligands suggests that they behave as antagonists at the sigma-1 receptor. The binding profiles of bis(2-ethylhexyl)ammonium and triisopentylammonium mimic other sigma-1 receptor antagonists such as IPAG and rimcazole, with shallow Hill slopes and possible multiple binding states 19 , they are able to induce a calcium response, as are other antagonists such as IPAG 11 . Furthermore, treatment of MDA-MB-468 cells with these branched-chain ammonium salts resulted in a reduction of cellular metabolism, which could be reversed by sigma-1 receptor knock-down using siRNA selective for the sigma-1 receptor. Sigma-1 receptor antagonists IPAG and rimcazole have previously been shown to reduce tumour growth in vivo 11 . We were able to carry out in vivo studies of tumour growth using MAC13 tumours, implanted into NMRI mice. Both bis(2-ethylhexyl)ammonium and triisopentylammonium were able to prevent tumour growth in vivo in much the same was as with other studies using sigma-1 receptor antagonists.

conclusion
We have identified a non-toxic, high affinity sigma-1 receptor ligand, dipentylammonium, which possesses potent antidepressant activity in animals in vivo. Dipentylammonium shows low affinity for classic monoamine transporters yet high affinity for the sigma-1 receptor. Further development of this molecule should be carried out to determine whether this compound can proceed into clinical trials as an antidepressant. We have also identified 2 potential anticancer agents, which act as sigma-1 receptor antagonists: bis(2-ethylhexyl)ammonium and triisopentylammonium. These too, should be further investigated as to whether they may have clinical use.

Materials and methods
General materials. Tissue culture media, antibiotics, trypsin and serum were purchased from Invitrogen www.nature.com/scientificreports www.nature.com/scientificreports/ (Fisher Scientific, Loughborourgh, UK) as either hydrochloride salts (HCl•) or as a free base amines. The free base amines were made soluble in water by making the HCl salt: 5 ml of free base amine was dissolved in 15 ml of ether, and hydrochloric acid was added slowly until the mixture turned acidic, measured using universal indicator paper (Sigma-Aldrich, Dorset, UK). The resulting solid was filtered and any remaining solvent was removed under vacuum. If the product was a liquid at room temperature, the solvents were removed under vacuum. Other reagents were purchased from Sigma-Aldrich (Poole, UK). Before use, drugs were dissolved in an appropriate vehicle and diluted into assay buffer, and the pH of each solution was determined and adjusted to 7.4.
Sigma-2 binding. Assays (100 μl) were performed at room temperature for two hours with [ 3 H] DTG in the presence of DMSO (10% final concentration). Nonspecific binding was determined using 100 μM reduced haloperidol. No sigma-1 receptor "masking agent" (e.g., (+) pentazocine) was used, as this adversely affects the determination of K d and K i values at the sigma-2 receptor 20 . Assays were terminated by addition of ice-cold sigma-2 wash buffer (150 mM NaCl, 10 mM Tris-HCl; pH 7.2) and filtration, using a cell harvester, through glass fibre filters (GF/C, Sigma-Aldrich, Poole, UK). Tubes and filter discs were washed (2 ×3 ml) with ice-cold wash buffer, and the filter discs dried under vacuum. Scintillation counting was carried out after overnight incubation of the discs with ProSafe FC + cocktail. Muscarinic receptor binding assays. Rat brain membranes were prepared from freshly killed rats. Rats were exposed to an increasing concentration of carbon dioxide until dead in accordance with Schedule 1 of Animals (Scientific Procedures) Act 1986 and decapitated. This method was approved by Bath University Ethics Committee. The brain was immediately removed and placed on ice with TBS. The brain was homogenised in 10 ml of ice cold TBS by a Teflon-in-glass homogeniser for two 30 s periods, the homogenates were centrifuged at 40,000 g for 30 minutes at 4 °C. The resultant supernatant was discarded and the pellets resuspended by homogenisation in fresh TBS. 1.35 nM [ 3 H] L-quinuclidinyl [phenyl-4-3 H] benzilate (specific activity 48.0 Ci.mmol −1 , GE Healthcare, Little Chalfont, UK) (pK d 9.5 ± 0.1, n = 4 data not shown), was incubated with rat brain membranes (0.5 mg.ml −1 ) to which varying concentrations of ammonium salt compounds were added as with the sigma-1 receptor competition assays. The membranes were incubated and harvested as with the [ 3 H] (+) pentazocine binding assays.
MTS cellular metabolic activity assay. Cells (10,000 cells per well, 100 μl) were seeded into 96-well plates and allowed to adhere overnight. The following day drugs were added and cells were permitted to proliferate for 24 hours before the addition of 10 μl CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay reagent (Promega, Southampton, UK) as previously described 11 . Absorbance (490 nm) readings were taken at 0, 1, 2 and 3 hours to monitor formazan formation and rates of metabolic activity were compared.
Vector constructs. The human dopamine transporter gene, kindly provided by Professor De Felice of Virginia Commonwealth University, was released from the pGEX 3Z vector using the KpnI and XbaI restriction enzymes in a sequential digestion. The DNA fragment was purified by agarose gel electrophoresis, and inserted into the mammalian expression vector, pcDNA3.1, at the KpnI and XbaI sites. Cloning success was checked using restriction and sequencing analysis.
The sigma-1 siRNA PSilencer plasmid and nontargeting siRNA control were kindly provided by Dr Christopher Palmer (London Metropolitan University, UK). www.nature.com/scientificreports www.nature.com/scientificreports/ Transporter studies. Individual 293 cell line clones stably expressing the human dopamine transporter, human 5-HT transporter or human noradrenaline transporter were created using Lipofectamine 2000 (Invitrogen, Gibthai, Bangkok). The 293 cells were plated in 6-well plates at 2.5×10 5 cells per well, and allowed to adhere overnight. The plasmid constructs were then mixed with Lipofectamine as instructed by the manufacturer and the 293 cells transfected. The following day the cells were harvested, diluted 10-fold and re-plated in 6-well plates before being allowed to adhere overnight. The following day the transfected cells were treated with G418 (Geneticin) (Invitrogen, Gibthai, Bangkok) to begin selection for positive clones. The cells were incubated for 1 to 2 weeks, with the G418-containing media being replaced every 2 days. Once colonies of resistant cells began to appear, they were isolated using sterile cloning rings and transferred to 96-well plates. Each clone was expanded until enough cells were available to verify the expression of the monoamine transporters. The expression was assayed using reverse transcriptase PCR and with the monoamine transporter assay (Molecular Devices, Tokyo Japan).
Monoamine transporter assay. The 293 cells stably expressing human dopamine, 5-HT or noradrenaline transporters respectively were plated in black-walled 96-well plates at approximately 6,000 cells per well. Once adhered, the media was removed, cells washed with phosphate buffered saline (PBS) and 50 μl Hank's buffered saline (HBSS) added to each well. Drugs were diluted to 2× concentration in HBSS-0.1% BSA and 50 μl were pipetted into each well. Control wells (cells without any drugs) contained 100 μl HBSS-0.05% BSA. The cells were incubated with the drugs for 30 minutes before the addition of 100 μl per well of fluorescent monoamine analogue dye, which is recognised and transported by all three transporters, and with cell impermeable quencher (homogeneous Neurotransmitter Transporter Uptake Assay Kit, #R8174, Molecular Devices, Tokyo, Japan). The cells were then incubated for a further 15 minutes at 37 °C in the dark before the fluorescence was measured in a microplate reader (excitation: 440 nm, emission: 520 nm). Uptake of the dye was compared to the control wells to calculate the % dye uptake for each drug concentration. Each data point was obtained from at least 3 independent determinations.
Sigma-1 receptor knock down. Sigma-1 receptor knock down was carried out as previously described 19,40 . Test for antidepressant activity. Animals. Males from an inbred strain of albino mice (originating from the Laca stock) weighing between 22 and 30 g bred in the Central Animal House (CAH) facility of Panjab University, Chandigarh, were used as previously described 41 . The animals were housed under standard laboratory conditions and maintained on natural light and dark cycle, and had free access to food and water. Animals were acclimatised to laboratory conditions before the experiment. Each animal was used only once. All the experiments were carried out between 09:00 and 15:00 h. The experimental protocols were approved by the Institutional Animal Ethics Committee, Panjab University and conducted according to the Indian National Science Academy Guidelines (INSA) for the use and care of experimental animals.
Forced-swim test. Mice were placed inside a rectangular glass jar (25 × 12 × 25 cm) containing water at a depth of 15 cm maintained at 24 ± 1 °C for a period of 6 minutes as previously described 41 . Each mouse was judged to be immobile when it ceased struggling and remained floating motionless in water, making only those movements necessary to keep its head above water. The total immobility period was recorded and compared to vehicle control group.
Tail suspension test. Mice were suspended 58 cm above a table top by using adhesive tape placed approximately 1 cm from the tip of the tail for a period of 6 minutes as previously described 42 . Each mouse was considered immobile when it hung passively and completely motionless. The total immobility period was recorded and compared to vehicle control group.
Locomotive activity. Locomotor activity (ambulations) was measured by using a computerised Actophotometer (IMCORP, India). An array of 16 infrared emitter/detector pairs measured animal activity along a single axis of motion, the digital data being displayed on the front panel as ambulatory movements. Mice were allowed to acclimatise to the observation chamber for a period of 2 minutes. The activity was monitored continuously for a period of 15 minutes. Locomotion was expressed in terms of total photobeam counts per 15 minutes per animal and compared to vehicle control groups as previously described 41 . In vivo cancer mouse model. Animals. Naval Medical Research Institute (NMRI) mice weighing 25 g were used for tumour transplantation, they were kept on a 12 hour day-night cycle, at ambient temperature, with free access to food (standard chow diet, Special Diet Services, Lillico, Wonham Mill, Bletchworth Surrey, UK) and water, the consumption of which was monitored throughout the experiments.
MAC13 cancer model transplant and treatment. The NMRI mice were transplanted with a fragment of mouse adenocarcinoma (MAC13) in the flank, by means of a trocar 43 . Twelve days after the tumour was implanted, (2020) 10:9251 | https://doi.org/10.1038/s41598-020-65849-6 www.nature.com/scientificreports www.nature.com/scientificreports/ animals (6 animals per group) were treated, with either ammonium salt (10 mg.kg −1 i.v., via the tail vein) or control (100 μl PBS i.v., via the tail vein) daily. Animals were terminated before the tumour exceeded 1 cm ³ . The tumour diameters were measured daily from day 12 using a micrometer. Animal studies were conducted under Home Office License according to the U.K. Coordinating Committee on Cancer Research Guidelines for the Care and Use of Laboratory Animals. The experimental protocols were approved by Aston University Ethics Committee. Data analysis. Saturation and competition data were analysed using non-linear regression analysis (GraphPad Prism, version 6.0 h for Macintosh) to fit saturation curves and calculate maximal radioligand binding and dissociation constants. In order to give a more true representation of the drugs' affinities for the receptor, the IC 50 was converted to a K i (where the Hill slope was near unity) value using the Cheng Prusoff equation. Values were converted into pK d or pK i before any mathematical manipulation to ensure normal distribution. Data are presented as mean ± SEM.
Comparisons of Hill slope were performed using a one-sample t test (GraphPad Prism, version 6.0 h for Macintosh), comparing values to unity. P < 0.05 was considered statistically significant.
MTS and protein assays were read using a Versamax plate reader, and the data collected using Softmax Pro software.
Results from behavioural assays and tumour studies are expressed as mean ± SEM. Significance of differences was determined using one-way ANOVA followed by a Dunnett's (when comparing against a control group) or Tukey's (when comparing all means) post hoc test. P < 0.05 was considered statistically significant.
Protein measurements. Protein amounts were determined using Bio-Rad Protein Dye Reagent, based on Coomassie Brilliant Blue G250 (Bio-Rad, Hemel Hempstead, UK). BSA was used to prepare standards.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.