Toxicology in Vitro Measuring inhibition of monoamine reuptake transporters by new psychoactive substances (NPS) in real-time using a high-throughput, ﬂ uorescence-based assay

The prevalence and use of new psychoactive substances (NPS) is increasing and currently over 600 NPS exist. Many illicit drugs and NPS increase brain monoamine levels by inhibition and/or reversal of monoamine reuptake transporters (DAT, NET and SERT). This is often investigated using labor-intensive, radiometric endpoint measurements. We investigated the applicability of a novel and innovative assay that is based on a ﬂ uorescent monoamine mimicking substrate. DAT, NET or SERT-expressing human embryonic kidney (HEK293) cells were exposed to common drugs (cocaine, DL -amphetamine or MDMA), NPS (4- ﬂ uoroamphetamine, PMMA, α -PVP, 5-APB, 2C-B, 25B-NBOMe, 25I-NBOMe or methoxetamine) or the antidepressant ﬂ uoxetine. We demonstrate that this ﬂ uorescent microplate reader-based assay detects inhibition of di ﬀ erent transpor- ters by various drugs and discriminates between drugs. Most IC 50 values were in line with previous results from radiometric assays and within estimated human brain concentrations. However, phenethylamines showed higher IC 50 values on hSERT, possibly due to experimental di ﬀ erences. Compared to radiometric assays, this high-throughput ﬂ uorescent assay is uncomplicated, can measure at physiological conditions, requires no speci ﬁ c facilities and allows for kinetic measurements, enabling detection of transient e ﬀ ects. This assay is therefore a good alternative for radiometric assays to investigate e ﬀ ects of illicit drugs and NPS on monoamine reuptake transporters.

In the Netherlands, the Drugs Information and Monitoring System (DIMS) offers a drug testing service to drug users. Data showed that, although NPS are also sold as common illicit drugs, the use of NPS as a drug of choice is increasing. The most frequently detected NPS in drug http://dx.doi.org/10.1016/j.tiv.2017.05.010 Received 6 January 2017; Received in revised form 3 April 2017; Accepted 11 May 2017 samples were 2,5-dimethoxyphenethylamine (2C-B), 4-fluoroamphetamine (4-FA) and methoxetamine (MXE). In addition, these NPS were also reported most frequently to the Dutch Poisons Information Center by health care professionals (Hondebrink et al., 2015a).
Exposure to NPS results in many desired effects, such as euphoria, mental stimulation, intensification of sensory perception, increased sociability, increased energy, increased empathy, openness, less inhibitions, and sexual arousal (Miliano et al., 2016). However, most users are unaware of possible adverse effects, which depend on the specific NPS used. For many different NPS severe health effects have been reported, including confusion, psychosis, suicidal thoughts and extreme aggression. In addition, life-threatening neurological and cardiovascular effects have been reported, such as arrhythmias, reverse Takotsubo cardiomyopathy, myocardial infarction, brain hemorrhage, convulsion and coma (Scottish Government Social Research, 2014;Wijers et al., 2017;Hohmann et al., 2014;Madias, 2015;Al-Abri et al., 2014;Butterfield et al., 2015). As a result, 9% of all drug-related emergency department visits involved the use of NPS (EMCDDA, 2015a(EMCDDA, , 2015b. The actual number is likely higher, due to difficulties in detecting NPS in blood or urine samples of users. In addition, patients visiting the emergency department for a drug intoxication have high admission rates, reported up to 70% (Duineveld et al., 2012).
Commonly used illicit drugs are well known to increase extracellular brain levels of monoamines, including dopamine, norepinephrine and serotonin. Monoamine levels can be increased via vesicular release of monoamines, decreased breakdown of these neurotransmitters and via inhibition and/or reversal of monoamine reuptake transporters including the dopamine transporter (DAT), norepinephrine transporter (NET) and serotonin transporter (SERT) (Korpi et al., 2015). Such increased monoamines levels can be related to clinical outcomes. For example, increased dopaminergic activity is related to reinforcing and behavioral-stimulating effects of drugs (Kimmel et al., 2001;Volkow et al., 2009). Substances with a primary site of action at DAT are also known to have a high abuse liability and they can induce strong adverse effects (Howell and Kimmel, 2008;Koob and Volkow, 2010). On the other hand, increased adrenergic activity can induce a wide range of cardiovascular effects such as tachycardia, hypertension and hyperthermia (Greene et al., 2008). Finally, increased serotonergic activity can induce entactogenic effects, but can also result in adverse effects including the potentially life-threatening serotonin syndrome (Mugele et al., 2012).
Cocaine, amphetamine and MDMA are known inhibitors of monoamine transporters. In addition, amphetamine and MDMA can also induce reversal of membrane transporters, thereby further increasing extracellular brain levels of monoamines (Torres et al., 2003;Fleckenstein et al., 2007;Verrico et al., 2007;Rietjens et al., 2012). Since many NPS have molecular structures comparable to illicit drugs and also induce comparable intended effects, their mechanisms of action likely overlap. In support of this, inhibition and reversal of monoamine transporters has been reported for several NPS (Eshleman et al., 2013;Nagai et al., 2007;Rickli et al., 2015aRickli et al., , 2015bSimmler et al., 2013Simmler et al., , 2014. Since both the use and the number of available NPS (currently over 600, UNODC, 2016) are increasing and severe adverse health effects have been reported, there is an urgent need to rapidly assess the hazard and risk for human health.
Several assays can be used to determine the (neurotoxic) effects of NPS. Preferably, applied assays allow for rapid screening of a large number of substances. For example, effects on neuronal activity can be determined with considerable throughput using multi-well micro-electrode arrays. Recently, it was shown that common illicit drugs and NPS reduce neuronal activity at concentrations relevant for human exposure (Hondebrink et al., 2016). This integrated endpoint provides valuable information, but provides limited insight in the mechanisms of action. Targeted assays allow for investigation of specific mechanisms, including drug-induced effects on GABA receptors (Hondebrink et al., 2011a;Hondebrink et al., 2013;Hondebrink et al., 2015b), voltagegated calcium channels (Hondebrink et al., 2011b;Hondebrink et al., 2012) or acetylcholine receptors . Moreover, the function of monoamine reuptake transporters is often investigated as a mechanism of action for psychoactive drugs, including NPS. Assays measuring transporter function often rely on measurement of the uptake of radio-labelled transporter ligands by e.g. human embryonic kidney (HEK) cells transfected with the transporter of interest. To perform such assays, specific laboratory requirements are needed for handling radio-labelled material. In addition, this method only allows for examining effects at the end of a particular exposure, precluding real-time kinetic measurements during drug exposure.
In addition to radiometric assays, neurotransmitter transporter uptake activity can be measured using fluorescent substrates such as 4-(4-(dimethylamino)styryl)-N-methylpyridinium (ASP + ), 4-(4-dimethylamino)-phenyl-1-methylpyridinium (APP + ), 1-methyl-4-phenylpyridium (MPP + ) and fluorescent false neurotransmitters (FFN) (Oz et al., 2010;Karpowicz et al., 2013;Schwartz et al., 2003;Fowler et al., 2006). Recently, a commercially available method was described using a fluorescent transporter substrate combined with a masking dye. Innovative aspects of this assay are that it does not require specific laboratory facilities or techniques, which makes it easy to use with a lower labor intensity. Also, high-throughput and real-time kinetic measurements can be performed using a plate reader (Jørgensen et al., 2008;Bernstein et al., 2012). The possibility to measure over time, for example, allows to investigate the reversibility of a drug-induced effect by adding potential antidotes, which is not possible using radiometric assays. Despite its benefits compared to radiometric assays, this method has rarely been used to measure effects of illicit drugs or NPS on the activity of neurotransmitter transporters. If this assay is applicable, it could aid in classifying NPS and quickly provide information on their mechanism of action.
The current research therefore investigates the applicability of this novel fluorescent assay to determine the potency of drugs, including NPS ( Fig. 1), to inhibit monoamine reuptake transporters in comparison to radiometric assays.

Inhibition of uptake by monoamine transporters
Uptake activity of hNET, hDAT and hSERT was measured using the Neurotransmitter Transporter Uptake Assay Kit from MDS Analytical Technologies (Sunnyvale, CA). The kit contained a mix consisting of a fluorescent substrate, which resembles the biogenic amine neurotransmitters, and a masking dye that extinguishes extracellular fluorescence. This product is patented by the manufacturer and the exact identity of the fluorescent substrate and masking dye therefore remains unknown. Uptake of the fluorescent substrate increases intracellular fluorescence, while extracellular fluorescence is blocked by the masking dye (Jørgensen et al., 2008). The fluorescent substrate solution was prepared by dissolving the mix in HBSS according to the protocol provided by the supplier and stored at − 18°C for a maximum of 4 days.

Drug-induced monoamine transporter uptake inhibition (preincubation with the fluorescent substrate)
On day 0, HEK 293 cells were seeded at a density of approximately 60.000 cells/well in clear-bottom, black-walled, 96-well plates (Greiner Bio-one, Solingen Germany) coated with PLL buffer (50 mg/L). Cells were allowed to proliferate overnight in a humidified 5% CO 2 /95% air atmosphere at 37°C. Experiments were performed the next day (day 1). Cells were pre-incubated with the fluorescent substrate for 12 min prior to a 30 min drug exposure (t = − 12 to t = 0). Culture medium was replaced by 100 μL/well fluorescent substrate solution, and uptake measurements were started. At t = 0, 100 μL/well HBSS without (control) or with drug was added to each well and uptake was measured continuously for 30 min. Background wells were pre-incubated with 100 μL/well HBSS without fluorescent substrate solution and exposed at t = 0 min to 100 μL/well HBSS without drugs. Non-transfected HEK 293 cells pre-incubated with 100 μL/well fluorescent substrate solution and exposed at t = 0 to 100 μL/well HBSS without drugs served as negative controls. Drugs ( Fig. 1) were prepared daily in HBSS from 2 or 100 mM stock solutions. Cocaine, DL-amphetamine, MDMA, 4-FA, MXE, PMMA, 2C-B, 25B-NBOMe, 25I-NBOMe, α-PVP and 5-APB were measured at final concentrations of 0.01-1000 μM. For 25B-NBOMe and A. Zwartsen et al. Toxicology in Vitro 45 (2017) [60][61][62][63][64][65][66][67][68][69][70][71] 25I-NBOMe the maximum concentration tested was 100 μM, as higher concentrations were cytotoxic. While continuous measurements (with temporal resolution determined by the speed of the plate reader) are possible, we measured fluorescence every 3 min, starting directly after addition of the fluorescent substrate solution (t = − 12). Fluorescence was measured with a microplate reader (Tecan Infinite M200 microplate; Tecan Trading Männedorf, Switzerland) at 37°C at 430/515 nm excitation/emission wavelength in bottom-reading mode using optimal gain values for each cell type (number of cycles: 21, time interval: 3 min, number of flashes: 19, integration time: 20 μs, no lid). Cell attachment was visually examined following experiments.

Drug-induced monoamine transporter uptake inhibition (preincubation with drugs)
In many radiometric assays, cells are pre-incubated with drugs prior to incubation with the radio-labelled substrate. We therefore also tested this experimental condition using MDMA and cocaine as reference chemicals. On day 1, medium was removed and 100 μL/well HBSS without (control) or with MDMA or cocaine was added to each well for 10 min prior to addition of 100 μL fluorescent substrate solution/well. Following addition of the fluorescent substrate solution (t = 0), fluorescence was measured every 3 min for 30 min as described above.

Possible drug-induced reversal of monoamine transporters
Single-cell imaging was performed to investigate if the fluorescent substrate can be released via reverse transport. Changes in fluorescence of hSERT-transfected cells were measured at room temperature with the Neurotransmitter Transporter Uptake Assay Kit from MDS Analytical Technologies (Sunnyvale, CA). On day 0, cells were seeded on PLLcoated glass-bottom dishes (MatTek, Ashland, Massachusetts) at a density of 18.000 cells/dish. On day 1, medium was replaced with 300 μL buffer consisting of 50% fluorescent substrate solution and 50% HBSS comparable to the plate reader experiments. The dish was placed on the stage of an Axiovert 35 M inverted microscope (40 × oil-immersion objective, NA 1.0; Zeiss, Göttingen, Germany), equipped with a TILL Photonics Polychrome IV (Xenon Short Arc lamp, 150 W; TILL Photonics, GmBH, Gräfelfing, Germany). Fluorescence was measured every 3 min at 430/515 nm excitation/emission wavelength using an Image SensiCam digital camera (TILL Photonics GmBH). Cells were preincubated with the fluorescent substrate solution for 21 min (t = −21 to t = 0), after which the fluorescent substrate was removed and cells were washed with 1 mL HBSS (t = 0). Subsequently, HBSS was replaced by 500 μL HBSS without (control experiments) or with MDMA (1 mM). Fluorescence was continuously measured in single cells and in areas without cells for 30 min following exposure.

Estimated drug concentration in the brain
The estimated brain concentrations were calculated using human recreational serum/blood levels obtained from literature (voluntary intake, driving under the influence or non-fatal intoxications, except for 5-APB which was derived from human overdose cases). Next, a brain partitioning factor (BPF) was determined for each drug by dividing the brain concentration by the serum/blood concentration found in human post mortem or animal studies. These human (recreational) serum/ blood levels were multiplied with the corresponding BPF to estimate human brain levels resulting from recreational drug use.
2.6. Data analysis 2.6.1. Calculating inhibition of uptake by monoamine transporters (in control experiments) The fluorescence of each well was background corrected (time-and plate-matched). Linearity of the uptake curves (raw data, fluorescence units (FU)) was assessed for hDAT, hNET, hSERT and non-transfected HEK cells using linear regressions (GraphPad Prism, version 6.05). Linearity was assessed in cells pre-incubated only with the fluorescent substrate (t = − 12 to t = 0) as well as when HBSS without drugs was added after 12 min of pre-incubation with the fluorescent substrate (t = 0 to t = 30), comparable to the actual experimental conditions. 2.6.2. Calculating drug-induced monoamine transporter uptake inhibition (pre-incubation with the fluorescent substrate) The fluorescence of each well was background corrected (time-and plate-matched). Uptake of the fluorescent substrate was first determined per well by calculating the change in fluorescence (ΔFU) at 12 min after drug exposure (t = 12) compared to the fluorescence prior or 10 min pre-incubation with the fluorescent substrate (A) or a drug/HBSS (B) respectively), as a percentage of the fluorescence prior to exposure (t = 0). Secondly, uptake in drugexposed wells was expressed as a percentage of control wells (not shown in figure). The fictional effects of the addition of HBSS (Control; A), a drug inhibiting transporter function (Drug 1; A) and a drug reducing fluorescence lower than prior the fluorescence to exposure (for details see discussion)(Drug 2; A) to cells pre-incubated with the fluorescent substrate are shown.
The fictional effects on monoamine uptake of cells pre-incubated with HBSS (Control) or drug (Drug 1) before the addition of the fluorescent substrate are also depicted (B).
to exposure (i.e. the fluorescence following 12 min pre-incubation with the fluorescent substrate at t = 0), as a percentage of the fluorescence prior to exposure ( Fig. 2A). Notably, as cells were pre-incubated with fluorescent substrate solution, fluorescence following drug exposure can be below the fluorescence prior to drug exposure and changes in fluorescence can therefore be negative (Fig. 2). Secondly, the percentage uptake in control wells of all plates was averaged and wells that showed values 2 × SD above or below average were considered as outliers and were excluded from further analysis (2%). Uptake in drug-exposed wells was expressed as a percentage of control wells. Outliers in exposed groups (effects 2 × SD above or below average) were removed (2%) and data was expressed as the mean ± SEM of n wells obtained from at least 3 independent experiments (N plates) (cell were seeded from different passages or different thawings), with at least 3 wells (n) per plate. Concentration-response curves were made for each transporter and each exposure. IC 50 values for multiple time points after exposure were based on full concentration-response curves (GraphPad Prism, version 6.05).

Calculating drug-induced monoamine transporter uptake inhibition (pre-incubation with drugs)
In a separate set of experiments, cells were pre-incubated with MDMA or cocaine for 10 min prior to addition of the fluorescent substrate (t = − 10 to t = 0). The following 12 min (t = 0 to t = 12), cells were exposed to both the drug and the fluorescent substrate. The fluorescence of each well was background corrected (time-and platematched). Uptake of the fluorescent substrate was first determined per well by calculating the change in fluorescence (ΔFU) at 12 min after drug and substrate exposure (t = 12) compared to the fluorescence prior to the drug and substrate exposure (i.e. the fluorescence following 10 min pre-incubation with the drug (t = 0), as a percentage of the fluorescence just after addition of the fluorescent substrate (t = 0) (Fig. 2B). Secondly, these values were analyzed as mentioned above, starting from averaging uptake in control wells of all plates.

Calculating possible drug-induced reversal of monoamine transporters
Single cells were incubated with fluorescent substrate for 18 min after which the fluorescent substrate was removed (t = − 18 to t = 0). Subsequently, MDMA or HBSS (control) was added for 30 min. The fluorescence over time was analyzed in single cells and in areas without cells (background fluorescence). ΔFU in single cells at 6, 12 and 30 min after drug exposure (t = 6, 12 and 30) was corrected for background fluorescence and normalized to the fluorescence at t = 0. A possible change in background fluorescence was also analyzed. Data is expressed as mean ± SEM from n cells, obtained from N dishes.
Non-transfected HEK cells did not show an increase in fluorescence ( Fig. 3; n = 12 wells, N = 1 plate), indicating that the increase in fluorescence in cells transfected with transporters is due to transporter function and not to passive diffusion (See also Jørgensen et al. (2008) for additional uptake characteristics of the fluorescent substrate).

Drug-induced monoamine transporter uptake inhibition (pre-incubation with the fluorescent substrate)
Exposure to NPS and commonly used illicit drugs concentrationdependently inhibited uptake of monoamine transporters following 12 min of exposure ( Fig. 4; Table 1). Cocaine potently inhibited all three transporters with IC 50 values of 1.3-1.9 μM. α-PVP was over ten times more potent than cocaine in inhibiting uptake of hDAT and hNET (IC 50 0.1 μM), although α-PVP only weakly inhibited hSERT. DL-Amphetamine also potently inhibited hNET and to a lesser extent hDAT, but only weakly inhibited hSERT.

Drug-induced monoamine transporter uptake inhibition (pre-incubation with drugs)
When transporter inhibition is investigated using radiometric assays, cells are often pre-incubated with the drug of interest prior to the addition of the radio-labelled substrate. Therefore, we also investigated this experimental condition using the fluorescent substrate for hSERT cells using cocaine and MDMA as reference chemicals. Cells transfected with hSERT were pre-incubated with cocaine or MDMA for 10 min prior to addition of the fluorescent substrate for 30 min to determine a possible difference in potency. Concentration-response curves of MDMA and cocaine when cells were pre-incubated with drugs before fluorescent substrate addition were then compared to concentrationresponse curves where cells were pre-incubated with fluorescent substrate before drug exposure (Fig. 4). Data showed comparable concentration response curves (Fig. 5). The IC 50 values at 6, 12 and 30 min after addition of the fluorescent substrate (t = 6, 12 and 30) were also comparable, although pre-incubation with drugs resulted in slightly higher IC 50 values for MDMA (Supplemental Table 2).

Possible drug-induced reversal of monoamine transporters
To determine whether the fluorescence-based assay could detect reversal of monoamine transporters, single-cell imaging was used. This method is the preferred choice to study reversal as the sensitivity of the photodetectors of the imaging system exceeds that of the plate reader (at the expense of throughput). In addition, the single cell fluorescence imaging experiments allowed for measurement of the fluorescent intensity in the individual cells as well as in the surrounding medium Fig. 4. Concentration-response curves of different drugs for the inhibition of uptake of fluorescent substrate via monoamine transporters (hDAT, hNET and hSERT) following 12 min of drug exposure. Prior to drug exposure, cells were pre-incubated with fluorescent substrate. Curves were fitted using nonlinear regression and data points are expressed as the mean ± SEM (n = 9-29 wells, N = 3-6 plates). The corresponding IC 50 values can be found in Table 1. (areas without cells). Since MDMA is known to reverse hSERT (Verrico et al., 2007;Rudnick and Wall, 1992;Mlinar and Corradetti, 2003), we investigated the effect of MDMA on single cells expressing hSERT.
Following pre-incubation with fluorescent substrate, the substrate was removed and cells were exposed to HBSS or MDMA. In control experiments, a limited increase in fluorescence was observed at 6, 12 and 30 min after the replacement of the substrate with HBSS (t = 6, 12 and 30) of respectively 4% ± 2, 6% ± 2 and 13% ± 1, which was likely due to incomplete removal of the fluorescent substrate (Fig. 6). Following MDMA exposure (1 mM), a comparable, limited increase in fluorescence was observed at 6, 12 and 30 min of 0% ± 1, 2% ± 2 and 3% ± 2 respectively, which suggests that the fluorescent substrate is not subject to reverse transport. In addition, no increase in fluorescence was observed in the extracellular medium.

Estimated drug concentration in the brain
To determine if recreational drug use results in brain concentrations of NPS and drugs of abuse comparable to concentrations causing effects on monoamine transporters, human brain concentrations were estimated. Therefore, concentrations of drugs in serum and/or blood and brain partitioning factors (BPF) were gathered from literature (Table 2). Table 2 also shows which transporters are inhibited at the estimated human brain concentration. Almost all compounds inhibit at least one of the transporters at concentrations relevant for humans, except the NBOMe's, Table 1 Inhibition of monoamine transporter uptake by illicit drugs, NPS and fluoxetine.IC 50 values (obtained following 12 min drug exposure) are presented with 95% confidence intervals [CI] (n =9-29 wells, N=3-6 plates). Grey blocks indicate the transporter(s) at which a compound is most potent.    Tim e (m in) U pt ake (% of t =0)

Group
Fig. 6. The kinetic effect of MDMA exposure (1 mM) on hSERT-expressing cells pre-incubated with fluorescent substrate measured using single cell microscopy. Cells were exposed to MDMA (1 mM) or HBSS (control) for 30 min, following 18 min pre-incubation with fluorescent substrate, after which the fluorescent substrate was removed. Data was normalized to the fluorescence following pre-incubation with fluorescent substrate (t = 0) and expressed as the mean ± SEM (Control: n = 8 cells, N = 2 dishes, MDMA: n = 15 cells, N = 4 dishes).
A. Zwartsen et al. Toxicology in Vitro 45 (2017) 60-71 4. Discussion Our results obtained using an innovative, high-throughput, fluorescence-based assay demonstrate that 3 common illicit drugs, 8 NPS and the SSRI fluoxetine concentration-dependently inhibit human monoamine transporter function (hDAT, hNET and hSERT, Table 1, Fig. 4). Drugs were tested at non-cytotoxic concentrations (Rickli et al., 2015a(Rickli et al., , 2015b(Rickli et al., , 2015cSimmler et al., 2013;Hondebrink et al., 2016). Stimulants such as amphetamine, 4-FA and PMMA in general selectively inhibit hNET and to a lesser extent hDAT. On the other hand, cocaine potently inhibits all three transporters, though α-PVP was over 10 times more potent on hNET and hDAT. In contrast, hallucinogenic drugs such as 2C-B, 25B-NBOMe, 25I-NBOMe and MXE were more potent on hSERT compared to hDAT and hNET. Thus, the fluorescencebased assay can effectively be used to investigate inhibition of human monoamine transporters by drugs. Neurotransmitter transporter inhibition profiles may be used to classify known and new drugs including NPS, thereby providing some information on possible adverse effects.
Importantly, inhibition of monoamine transporters was detected at concentrations relevant for human exposure. For almost all substances, the IC 50 value for at least one of the transporters was in range of the estimated brain concentration, while four substances likely affect more than one transporter at estimated brain concentrations. Even the selective serotonin re-uptake inhibitor fluoxetine inhibits both hSERT (IC 50 0.3 μM) and hNET (IC 50 8 μM) at estimated therapeutic brain concentrations, in line with other literature Renshaw et al., 1992;Komoroski et al., 1994;Strauss et al., 2002;Henry et al., 2005;Strauss and Dager, 2001). This suggests that transporter inhibition is a relevant mechanism of action during recreational use of these compounds. 2C-B, 25B-NBOMe and 25I-NBOMe had estimated brain concentrations below transporter inhibition, which may be explained by a potential underestimation of the estimated brain concentration due to the lack of information on brain partitioning. Also, NBOMe derivatives are potent 5-HT 2A receptor agonists (nM range), which is currently considered as their main mechanism of action (Kyriakou et al., 2015).
Particular benefits of this fluorescence-based assay, compared to traditional radiometric assays, include the possibility to perform highthroughput and real-time kinetic measurements as well as its ease of use. Although we measured fluorescence only every 3 min, it is possible to measure continuously with temporal resolution limited only by the speed of the plate reader. In addition, due to kinetic measurements, each well can serve as its own internal control, i.e., baseline uptake can be established prior to drug exposure. The use of such internal controls reduces variation due to e.g. differences in transporter expression and/ or cell numbers in different wells.
Moreover, kinetic measurements offer the possibility to determine the IC 50 at different time points, thus allowing for more sophisticated experiments. For example, it is possible to pharmacologically modulate drug-induced inhibition in search for a possible antidote that may be applied in the treatment of intoxicated patients, which would not be possible if endpoint measurements like radiometric assays are used. HEK cells, lacking neuron-specific machinery, are an excellent model to investigate these direct effects on neurotransmitter function. Intoxicated patients might also benefit from this assay as that it can help to determine the most relevant mode of action of emerging NPS, providing some information on possible symptoms.
Since Ki values are hardly reported in literature, we compared the IC 50 values we obtained with the fluorescence-based assay to those Table 2 Estimated brain concentrations of commonly used illicit drugs, NPS and fluoxetine compared to their potency to inhibit monoamine transporters. Estimated brain concentrations were calculated using human serum concentrations and brain partitioning factors (BPF) found in literature. All human serum concentrations were obtained from recreational use doses (voluntary intake, driving under the influence or accidental non-fatal intoxications), except for 5-APB. Serum concentration of α-PVP is based on blood concentrations (small caps). BPFs were based on serum and brain concentrations of rat (bold) or mouse (italic) data, or human post mortem blood (underlined) or serum (strikethrough) values compared to brain concentrations. Estimated brain concentrations for 5-APB and 25B-NBOMe were based on the observation that most BPFs are > 1. Grey blocks indicate that IC 50 values obtained using the fluorescent assay are not in the estimated brain concentration range. Bold IC 50 values indicate values within the estimated brain concentration range.

Group Drug
Serum concentration (µM) Brain partitioning factor (BPF) Estimated brain concentration ( reported measured with radio-labelled ligands. Notably, IC 50 values are mostly comparable between methods (Table 3). This is in line with previous studies that investigated the potency of antidepressants to inhibit transporter function and applied both the radiometric method and the fluorescent substrate. No difference in inhibition potencies was observed between both methods when experimental conditions like temperature or cell attachment were kept similar (Tsuruda et al., 2010;Jørgensen et al., 2008). For (illicit) drugs or NPS, only one article determined Ki values on hNET uptake using both a radiometric method ( 3 H-NE) and an analogous fluorescent substrate, ASP + . Comparable Ki values were determined with both methods for amphetamine, cocaine and MDMA to inhibit hNET uptake (Haunsø and Buchanan, 2007). The IC 50 values, obtained using a fluorescent substrate, for the inhibition of hSERT for the phenethylamines amphetamine, 4-FA, MDMA and PMMA differ strongly compared to results obtained by others using radiometric methods (for references see Table 3). Differences in IC 50 values for transporter inhibition by drugs between our study and others could be due to the use of disparate experimental setups, such as cell type, cells in suspension versus attached cells, and measurements at room temperature versus at 37°C. Notably, in contrast to radio-labelled ligands, our single-cell imaging experiments, suggest that the fluorescent substrate is not subject to reverse transport (Fig. 6). The fluorescent substrate is therefore ideally suited to study inhibition of transporters as results are not confounded by reverse transport. Consequently, the fluorescence levels below zero at high drug concentrations (Fig. 4) were likely not caused by reverse transport and/or passive dye leakage. This effect is likely due to other factors, such as limited dye bleaching and/or limited sequestration of the dye in intracellular compartments with different ionic/pH conditions that somewhat attenuate the fluorescence of the dye. The fluorescence levels fluctuation around 100% at low drug concentrations are likely simply reflecting biological variation at a no-effect level.
Additional differences between our assay and those using radio-labelled substrates, relate to the experimental conditions. In our study, we aimed at mimicking in vivo conditions as closely as possible. Therefore, the cells were pre-incubated with fluorescent substrate prior to drug exposure providing intracellular levels of 'endogenous' substrate, which is more comparable to the in vivo situation. On the other hand, most studies using radio-labelled substrates pre-incubate the cells with drugs prior to adding the substrates. It has previously been suggested that pre-incubation of cells with 'slow binding' drugs results in lower IC 50 values (Tsuruda et al., 2010). Our data indicate that preincubating cells with drugs or pre-incubating cells with substrate has limited effects on the hSERT IC 50 values ( Fig. 5 and Supplemental Table 2). Those particular experiments should be considered as proof of principle and were therefore limited to one transporter and two drugs. hSERT was chosen, since it showed the highest difference between IC 50 data measured using the fluorescence-based and radiometric-based assays, especially with exposure to phenethylamines. Therefore, a drug of the phenethylamines class (MDMA) and a non-phenethylamine drug (cocaine) were chosen.
Since uptake and binding for hSERT is known to be temperature dependent (Tsuruda et al., 2010;Elfving et al., 2001;Saldaña and Barker, 2004;Oz et al., 2010), our experiments were performed at a physiological temperature (37°C), whereas most other studies performed experiments at room temperature. Notably, we observed a~3 fold increase in MDMA potency on hSERT when temperature was lowered from 37°C to room temperature (data not shown). Thus, the physiological temperature used in our study likely accounts for some of the observed differences in IC 50 values. Table 3 Inhibition of monoamine transporter uptake (IC 50, μM) by illicit drugs, NPS and fluoxetine compared to literature. All articles reported in this table used radio-labelled substrates. Potency for uptake inhibition was determined using transfected HEK293 cells (e, g, h, i, k, l, m, o, p, q), rat brain synaptosomes (a, b, f, g, j, n, italic), human platelets (c, SERT, underlined), C6 glial cells (c, DAT + NET, underlined with stripes), or JAr cells (d, underlined with dots). Almost all studies used cells in suspension, except for b, d, g, h, k and c, DAT + NET (bold).  Marona-Lewicka et al., 1995;b Hyttel, 1982;c Cozzi et al., 1999;d Martel and Keating, 2003;e Meltzer et al., 2006;f Nagai et al., 2007;g Jørgensen et al., 2008;h Yoon et al., 2009;i Eshleman et al., 2013;j Baumann et al., 2013;k Rosenauer et al., 2013;l Simmler et al., 2013;m Simmler et al., 2014;n Marusich et al., 2014;o Rickli et al., 2015a; Rickli et al., 2015b;q Rickli et al., 2015c. A. Zwartsen et al. Toxicology in Vitro 45 (2017 Furthermore, we used attached cells in our experiments, in contrast to most other studies that used cells in suspension (Table 3). To obtain cells in suspension, trypsin is often used. This process can cause changes in cell morphology and damage to membrane proteins, resulting in cellular dysfunction and stress responses (Huang et al., 2010), which may increase the sensitivity of cells to the effect of drugs. Only one study used attached cells at 37°C to investigate illicit drugs and NPS using radio-labelled substrates. In line with our data, they also reported lower potencies for amphetamine, 4-FA and MDMA to inhibit hSERT (IC 50 ± 100 μM; Rosenauer et al., 2013). Simmler and Liechti (2016) also reported drugs to be less potent releasers when attached cells were used compared to cells in suspension.

Group
Alternatively, the difference between our data and data found in literature may be explained by interaction of the fluorescent substrate with the binding site of phenethylamines at hSERT, since for nonphenethylamine drugs effects on hSERT were comparable to those reported with radiometric assays. However, since the manufacturer would not provide chemical details about the substrate, this could not be assessed.
Thus, many differences in the experimental approach could explain the difference between the IC 50 values of phenethylamines on hSERT inhibition measured using the fluorescent substrate and radioactively labelled monoamines. Even though our fluorescence measurements are closer to physiological conditions, which of both methods derives correct IC 50 values of phenethylamines on hSERT remains unclear, resulting in the risk to over-or underestimate the potency of illicit drugs and NPS to inhibit monoamine transporters.
In summary, our data demonstrate that the novel fluorescent-based method detects drug-induced inhibition of hDAT, hNET and hSERT. This high-throughput kinetic assay discriminates between a variety of commonly used illicit drugs and NPS that concentration-dependently inhibit the reuptake of monoamines with high reproducibility between experiments. For most drugs, IC 50 values are in the range of estimated brain concentrations, indicating that inhibition of monoamine transporters contributes to the psychological effects. Compared to radiometric assays, IC 50 values are comparable for hDAT, hNET and hSERT, with the exception of phenethylamines on hSERT, which show higher IC 50 values in the fluorescence-based assay. These differences however might be explained by experimental differences. The fluorescencebased assay has several advantages compared to the use of radio-labelled monoamines, including the possibility to measure effects kinetically (providing temporal information about transporter regulation and function), at physiological conditions (cell integrity, endogenous neurotransmitter concentration, temperature) and being less laborious without requiring specific laboratory facilities.

Conflict of interest
The authors do not have any conflict of interest.