Substrates of the Human Brain Proton-Organic Cation Antiporter and Comparison with Organic Cation Transporter 1 Activities

Many organic cations (OCs) may be transported through membranes by a genetically still uncharacterized proton-organic cation (H + OC) antiporter. Here, we characterized an extended substrate spectrum of this antiporter. We studied the uptake of 72 drugs in hCMEC/D3 cells as a model of the human blood–brain barrier. All 72 drugs were tested with exchange transport assays and the transport of 26 of the drugs was studied in more detail concerning concentration-dependent uptake and susceptibility to specific inhibitors. According to exchange transport assays, 37 (51%) drugs were good substrates of the H + OC antiporter. From 26 drugs characterized in more detail, 23 were consistently identified as substrates of the H + OC antiporter in six different assays and transport kinetic constants could be identified with intrinsic clearances between 0.2 (ephedrine) and 201 (imipramine) mL × minute−1 × g protein−1. Excellent substrates of the H + OC antiporter were no substrates of organic cation transporter OCT1 and vice versa. Good substrates of the H + OC antiporter were more hydrophobic and had a lower topological polar surface area than non-substrates or OCT1 substrates. These data and further research on the H + OC antiporter may result in a better understanding of pharmacokinetics, drug–drug interactions and variations in pharmacokinetics.


Introduction
The membrane transport of most organic cations is not well understood. Roughly estimated, one-third of all low-molecular weight drugs and other substances of natural or synthetic origin are organic cations. As such, they are majorly or almost completely positively charged at the typical pH in the human body. Particularly when they are more hydrophilic, such positively charged substances may hardly cross biological membranes by diffusion. More lipophilic substances may not only pass membranes by nonionic diffusion but also with these substances, influx and efflux transport may be significantly accelerated by membrane transporters. For several organic cations, organic cation transporters such as OCT1, OCT2, OCT3, OCTN1, OCTN2, THTR1, THTR2, MATE-1, MATE-2K and several others have been identified as relevant transporters [1][2][3]. However, these known transporters do not contribute in a relevant manner to the membrane transport of many cationic drugs. For many of these drugs, the H + OC antiporter may mediate the transport. Among these drugs are several first-generation antihistaminergic drugs, nicotine, varenicline, oxycodone and clonidine [4]. This H + OC antiporter was first described more than 20 years ago [5,6] and has a broad cell and tissue expression. In particular, its expression in the blood-brain barrier was studied in humans or human cell lines, but the H + OC antiporter is also expressed in the kidneys, liver, gut, lung and placenta [7][8][9][10][11][12]. Many studies on the H + OC The upper 26 substances were characterized concerning transport kinetics, trans-stimulation with diphenhydramine and MDAI as exchange substrates, inhibition with diphenhydramine, imipramine and memantine and pH gradient dependency using FCCP. The lower 46 were tested with trans-stimulation assays only. MW, molecular weight; TPSA, topological polar surface area, also known as polar surface area, which is the surface sum of all polar atoms (mostly oxygen and nitrogen; * quaternary amines are always positively charged in human body fluids and were arbitrarily labelled with a pka of 14. The chemical features of all substrates were determined using Instant JChem (ChemAxon, Budapest, Hungary).
In addition, we tested whether or not substrates of the H + OC antiporter might also be substrates of the organic cation transporter 1 (OCT1). This was motivated by our observation that the typical substrates of the H + OC antiporter were not or only poor substrates of OCT1, indicating specific differences in the pharmacophores of the H + OC antiporter and OCT1.

Testing 26 Possible Substrates by Applying 6 Criteria for H + OC Antiporter Substrates
In the first step, we performed a detailed analysis of 26 drugs concerning typical parameters indicating transportation via the H + OC antiporter, namely the (a) antiport with two prior established H + OC antiporter substrates (diphenhydramine, 5,6-Methylenedioxy-2-aminoindane), (b) inhibition by typical inhibitors of the H + OC antiporter (imipramine, diphenhydramine and memantine) and (c) modulation (reduction expected) of the transport by a disruptor of the cell-membrane proton gradient (FCCP). The 26 drugs were selected from medically relevant drugs by chemical criteria typical for organic cation (solute carrier, SLC) transporter substrates (pka > 7.4, a molecular weight < 600, and a LogD at pH 7.4 < 2). The ratios indicating transport or inhibition of transport are summarized in Table 2. * Trans DPH, trans-stimulation (antiport) with diphenhydramine (DPH) as probe substrate, given as mean (M) and standard deviation (SD); activity was measured as ratio of DPH remaining within the cells after 2 min incubation with each of the 26 substrates over control incubation without substrate; Trans MDAI, trans-stimulation with 5,6-Methylenedioxy-2-aminoindane as a probe substrate, and parameters and procedures were the same as with DPH; Cis DPH, inhibition of the influx of each of the 26 substrates by the known H + OC antiporter substrate diphenhydramine, data given as ratio of intracellular substrate concentration after 2 min of incubation with inhibitor over substrate concentration without inhibitor; Cis Imip, inhibition with imipramine, parameters as above; Cis Mem, inhibition with memantine, parameters as above; FCCP, effect of FCCP on influx transport given as mean (M) ratio of intracellular substrate concentration with FCCP co-incubation over substrate concentration without FCCP co-incubation. NM, not meaningful; -, not measured or not measured in replicates.
Most tested substances, except for morphine and pyrazinamide, were subject to extensive exchange transport during the 2-min incubation, as indicated by ratios significantly below unity when using DPH or MDAI as antiporter substrates (Table 2).
Also in the inhibition experiments, morphine and pyrazinamide were the only two substances where the uptake of which was not inhibited by the three inhibitors used here (Table 2, Figure 1B,C). As indicated in Figure 1, inhibition by DPH and MDAI showed a strong correlation and showed that imipramine was generally a stronger inhibitor while memantine was a weaker inhibitor, corresponding also with KM values of 20 µM and 29 µM identified in this study for imipramine and memantine, respectively.
As indicated by the summarizing view in the radar plot, all substances except for morphine and pyrazinamide were identified as possible substrates of the H + OC antiporter. The latter two were also not affected at all by the protonophore FCCP ( Figure 1D, upper right square). However, the effects of FCCP correlated generally less well with inhibition and trans-stimulation parameters, probably due to other off-target interactions of FCCP with the transport process, such as alterations in intracellular energy metabolism and/or other effects on activity, e.g., by binding to the transporter protein.
We then measured the concentration-dependent uptake in hCMEC/D3 cells and calculated the transport kinetic parameters (K M , V max , intrinsic clearance (Cl int )). The concentration-uptake curves are shown in Figure 2. As observed, for 24 of the 26 organic cations tested, there was a saturable concentration-dependent uptake into hCMEC/D3 cells.
Transport kinetic constants are summarized in Table 3. Bupropion, citalopram, imipramine, duloxetine and memantine were transported with highest affinity (K M below 30 µM). Imipramine, duloxetine and clozapine had the highest intrinsic clearance (above 100 mL/min/g protein). Ephedrine stood out with a very low maximal transport rate. However, with the relatively low K M , the H + OC antiporter might still be relevant in vivo.
In conclusion, of these experiments with 26 possible substrates of the H + OC antiporter, only 2 substrates, pyrazinamide and morphine, did clearly not fulfill the criteria of H + OC antiporter substrate (see also Figure 1). For a third substrate, metoprolol, no saturated transport could be seen in concentration-dependent uptake measurements, indicating that metoprolol may mainly be taken up by simple (not transporter mediated) diffusion. Although metoprolol might be a substrate of the H + OC antiporter according to the transstimulation and inhibition criteria ( Table 2), we did not classify metoprolol as an H + OC antiporter substrate because there was no saturated transport.

Testing Further 46 Possible Substrates by Antiport (Trans-Stimulation) Assays
From the measurements shown above but also from published data we concluded, antiport and inhibition measurements are highly correlated and, therefore, redundant for characterizing a substance as a substrate of the H + OC antiporter. Therefore, we tested another set of 46 substances only with the antiport assays using diphenhydramine and MDAI (Tables 1 and 4). As seen in Figure 3, which shows the correlation between the two trans-stimulation assays with prototypic exchange substrates diphenhydramine and MDAI, both assays correlated well.
While there was no visually apparent cut-off, the 72 substances tested here could be differentiated into 53 (74%) substrates of both exchange partners (using any ratio below 1 as indicative for antiport) and 37 (51%) good substrates (using a ratio below 0.7 as indicative for good antiport).

Relationship between H + OC Antiporter Substrates and OCT1 Substrates
Another question of this study was whether there may be an overlap or distinction of substrates of the H + OC antiporter and substrates of the well-characterized organic cation transporter OCT1. As illustrated in Figure 4, when using diphenhydramine as a probe substrate, only four of the substances tested were good substrates of both the H + OC antiporter and OCT1 (dobutamine, milnacipran, desmethylvenlafaxine and meptazinol). When using MDAI as antiporter probe substrate (right panel in Figure 4), there was almost a mutual exclusion concerning good antiporter substrates and good OCT1 substrates. All substances with their diphenhydramine and MDAI trans-stimulation ratios and their OCT1 uptake ratios are provided in Table 4. As indicated by the summarizing view in the radar plot, all substances except for morphine and pyrazinamide were identified as possible substrates of the H + OC antiporter. The latter two were also not affected at all by the protonophore FCCP ( Figure 1D, upper right square). However, the effects of FCCP correlated generally less well with inhibition and trans-stimulation parameters, probably due to other off-target interactions of FCCP with the transport process, such as alterations in intracellular energy metabolism and/or other effects on activity, e.g., by binding to the transporter protein.  Table 2). (B) Correlation between diphenhydramine and imipramine inhibition, (C) correlation between diphenhydramine and memantine inhibition and (D) correlation between diphenhydramine trans stimulation ratio and FFCP inhibition effect. and imipramine inhibition, (C) correlation between diphenhydramine and memantine inhibition and (D) correlation between diphenhydramine trans stimulation ratio and FFCP inhibition effect.
We then measured the concentration-dependent uptake in hCMEC/D3 cells and calculated the transport kinetic parameters (KM, Vmax, intrinsic clearance (Clint)). The concentration-uptake curves are shown in Figure 2. As observed, for 24 of the 26 organic cations tested, there was a saturable concentration-dependent uptake into hCMEC/D3 cells.     Substrates are ordered with decreasing exchange activity against diphenhydramine. The diphenhydramine ratio refers to intracellular diphenhydramine after two-minute incubation with the substrate over intracellular diphenhydramine without substrate addition, and SD is the corresponding standard deviation. MDAI ratio and SD are the corresponding parameters with 5,6-methylenedioxy-2-aminoindane. The OCT1 ratio is the ratio of cell uptake in OCT1 overexpressing cells over uptake in mock transfected cells. In the table, we differentiated between ratios below 0.7 with diphenhydramine antiport, possibly indicating good substrates; below 1.0, possibly indicating any substrates; and above 1.0 indicating non-substrates. With OCT1, any statistically significant ratio above 1.5 also indicates substrates of OCT1, whereas ratios above 3 may indicate good substrates.
From the measurements shown above but also from published data we concluded, antiport and inhibition measurements are highly correlated and, therefore, redundant for characterizing a substance as a substrate of the H + OC antiporter. Therefore, we tested another set of 46 substances only with the antiport assays using diphenhydramine and MDAI (Tables 1 and 4). As seen in Figure 3, which shows the correlation between the two trans-stimulation assays with prototypic exchange substrates diphenhydramine and MDAI, both assays correlated well. Figure 3. Correlation between the trans-stimulation (antiport) with diphenhydramine (DPH) and 5,6-Methylenedioxy-2-aminoindane (MDAI). Cells were pre-incubated for 30 min with either DPH or MDAI and the percentages refer to the percentage remaining within the cells 2 min after addition of the substrates. While there is no clearly apparent cutoff between substrates and non-substrates, all substances above 1 are per definition non-substrates and all substances below 0.7 may be considered as good substrates. Therefore, these reference lines are shown in the graph together with the linear regression line. As observed, the regression line is very similar to the line of identity.
While there was no visually apparent cut-off, the 72 substances tested here could be differentiated into 53 (74%) substrates of both exchange partners (using any ratio below 1 as indicative for antiport) and 37 (51%) good substrates (using a ratio below 0.7 as indicative for good antiport).

Relationship between H + OC Antiporter Substrates and OCT1 Substrates
Another question of this study was whether there may be an overlap or distinction of substrates of the H + OC antiporter and substrates of the well-characterized organic cation transporter OCT1. As illustrated in Figure 4, when using diphenhydramine as a probe substrate, only four of the substances tested were good substrates of both the H + OC antiporter and OCT1 (dobutamine, milnacipran, desmethylvenlafaxine and meptazinol). When using MDAI as antiporter probe substrate (right panel in Figure 4), there was almost a mutual exclusion concerning good antiporter substrates and good OCT1 substrates. All substances with their diphenhydramine and MDAI trans-stimulation ratios and their OCT1 uptake ratios are provided in Table 4.   Table 4).    Table 4).

Chemical Characteristics of H + OC Antiporter Substrates
Next, we compared if non-substrates, poor substrates and good substrates of the H + OC antiporter might differ in some of the most relevant chemical properties and we also compared this with the substrate patterns of OCT1. As shown in Figure 5, neither H + OC antiporter substrates nor OCT1 substrates can be distinguished by molecular weight, but substances with a molecular weight above 400 Dalton are rarely substrates of the antiporter or of OCT1. Interestingly, the smaller the topological polar surface area (TPSA), the higher the likelihood it is a substrate of the H + OC antiporter. This is also compatible with the fact that a TPSA below 70 Å is a typical feature of substances passing well through the blood-brain barrier [19]. Only one substrate, trimethoprim, had a TPSA above 100. Trimethoprim may not really be a good H + OC antiporter substrate, since it was classified as such only according to the diphenhydramine antiport assay results but not with MDAI (Table 4). Correspondingly, substrates of the H + OC antiporter were significantly more hydrophobic compared with non-substrates or with substrates of OCT1 ( Figure 5). While OCT1 substrates only rarely have a logD value above 1.5, several H + OC antiporter substrates have logD values around and even above 2. High lipophilicity may constitute a general upper limit of SLC influx transporters simply due to the then prevailing passive and non-ionic diffusion of very lipophilic substrates. However, as shown here, this limit concerning lipophilicity depends on the transporter ( Figure 5). Finally, because the question about the genetic identity of the H + OC antiporter is still unanswered, we looked if selected typical substrates of the H + OC antiporter might be substrates of one of the four well-characterized proton-dependent membrane transporters with antiport activity, namely the two OCTN transporters and the two high-affinity thiamine transporters, THTR1 and THTR2. Although there were indeed minor transport activities found for several of these other antiporters (Figure 6), the ratio of the Finally, because the question about the genetic identity of the H + OC antiporter is still unanswered, we looked if selected typical substrates of the H + OC antiporter might be substrates of one of the four well-characterized proton-dependent membrane transporters with antiport activity, namely the two OCTN transporters and the two highaffinity thiamine transporters, THTR1 and THTR2. Although there were indeed minor transport activities found for several of these other antiporters (Figure 6), the ratio of the cell uptake into overexpressing cells over empty-vector transfected cells almost always was below 2, a cutoff frequently used to define a medically relevant transport activity.

Discussion
Since the protein conferring H + OC antiport activity is still not identified, there is no completely unequivocal assay for identifying H + OC antiporter substrates. Moreover, currently, we cannot exclude that the apparently antiporter-mediated uptake of more than 40 substrates that are not mediated by one but by several proteins. Nevertheless, from the first set of extensively tested 26 substrates, 23 consistently turned out to be H + OC substrates according to trans-stimulation, inhibition, pH gradient dependence and concentration-dependent (saturated) transport. From the entire set of 72 drugs, based on a cutoff of 0.7 in the diphenhydramine antiport assay, 42 turned out to be good substrates (Table 4). We have preselected our substrates for more hydrophilic cationic substances with a molecular weight below 600 Dalton, but it is nevertheless remarkable that for this large subset of drugs, about 50% were transported by the H + OC antiporter.
Here, we specifically looked at the correlation between transport via the organic cation uniporter OCT1 and the H + OC antiporter. Although not being absolutely mutually exclusive, the correlation showed that apparently very good OCT1 substrates are often not H + OC antiporter substrates and vice versa ( Figure 4). As known, OCT1 substrates

Discussion
Since the protein conferring H + OC antiport activity is still not identified, there is no completely unequivocal assay for identifying H + OC antiporter substrates. Moreover, currently, we cannot exclude that the apparently antiporter-mediated uptake of more than 40 substrates that are not mediated by one but by several proteins. Nevertheless, from the first set of extensively tested 26 substrates, 23 consistently turned out to be H + OC substrates according to trans-stimulation, inhibition, pH gradient dependence and concentration-dependent (saturated) transport. From the entire set of 72 drugs, based on a cutoff of 0.7 in the diphenhydramine antiport assay, 42 turned out to be good substrates (Table 4). We have preselected our substrates for more hydrophilic cationic substances with a molecular weight below 600 Dalton, but it is nevertheless remarkable that for this large subset of drugs, about 50% were transported by the H + OC antiporter.
Here, we specifically looked at the correlation between transport via the organic cation uniporter OCT1 and the H + OC antiporter. Although not being absolutely mutually exclusive, the correlation showed that apparently very good OCT1 substrates are often not H + OC antiporter substrates and vice versa ( Figure 4). As known, OCT1 substrates are rarely if ever highly lipophilic [1,21], but as illustrated in Figure 5, lipophilicity alone does not explain the difference between OCT1 and H + OC antiporter substrates. A simple argument why lipophilic substances are not substrates of OCT1 may be that passive diffusion in either directions may mask the effect of transporter activities. In this respect, the so-called trans-stimulation assay (data for all 72 substances tested in Table 4) may be less dependent on lipophilicity. Indeed, lipophilic substances such as amitriptyline or its metabolite nortriptyline, which are not OCT1 substrates [22], had a reasonable intrinsic clearance as measured by influx into the hCMEC/D3 cells (Table 3). These data also indicate that overrunning carrier-mediated transports by simple diffusion in both directions cannot be the only explanation why OCT1 substrates are typically not highly lipophilic.
Since one may argue that primarily OCT2 and OCT3, but not OCT1, are the extrahepatic transporters, we also looked at that correlation between transport by the antiporter (DPH antiport and MDAI antiport; see Table 4) and OCT2 and OCT3 transport ratios in a subset of 22 of the substances tested here (detailed data not shown). However, the picture with OCT2 and OCT3 was essentially the same as with OCT1 ( Figure 4). Thus, good substrates of the H + OC antiporter were not or only poor substrates of OCT2 or OCT3.
While the majority of ABC and SLC transporters has currently been identified and characterized concerning at least of some of their substrates [3,23], the genetic identity of the H + OC antiporter has still not been disclosed. One approach to this question might be testing the substrates of the H + OC antiporter as substrates of well-defined SLC transporters. This has been studied earlier by several authors using specific chemical inhibitors and siRNA-based knockdown [24,25]. Here, we also studied the transport of prototypic H + OC antiporter substrates (clonidine, cocaine, diphenhydramine, MDMA, memantine, oxycodone, varenicline and MDAI) as substrates of two human high-affinity thiamine transporters and the two OCTN transporters, all four of which are expressed at a moderate level in the hCMEC/D3 cells [26]. However, based on the ratios of cell uptake with overexpressing cells over the uptake in empty-vector transfected cells, there was either no transport or only low transport ratios ( Figure 6). Of course, one might test all the 72 substances characterized here with the more than 12 organic cation transporters known at present, but this was beyond what was achievable in our present project.
Since we contrasted here the activity of the primarily hepatically expressed OCT1 to the activity of the H + OC antiporter, one may ask whether OCT1 contributed to transports in the hCMEC/D3 cells. Indeed, OCT1 may be expressed in the hCMEC/D3 cells [26] but only at quite low concentrations. Moreover, as demonstrated here (Figure 4, Table 4), OCT1 is mostly not a good transporter for H + OC antiporter substrates.
Currently, we cannot even answer whether the H + OC antiporter is one protein or many, so we are at a similar stage as our colleagues in studying drug metabolism who asked 50 years ago, "Cytochrome P-450 of liver microsomes-one pigment or many" [27]. However, the generally rather high correlation between the different assays used here ( Figure 1) might indicate that it is, indeed, mostly one protein, because with different transport proteins, more variability in the results of the different assays between different substrates might exist.
The medical impact of the finding, that about 50% of the drugs tested here are substrates of the H + OC antiporter, may be significant. For instance, all substrates of the H + OC antiporter may exhibit significant drug-drug interactions between each other when passing through the blood-brain barrier and considering that, in many fields of medicine, polymedication with five or more drugs is widely practiced, and drug concentrations achieved in the brain may be very different in polymedication compared with monother-apy. However, of course, this aspect has to be investigated in further experimental and clinical studies.
Another quite interesting fact is that the pharmacokinetics and pharmacodynamics of several substrates of the H + OC antiporter show extremely large variations between individuals. At the very least, there are a number of examples for that. For instance, varenicline, an H + OC antiporter substrate [28], has blood concentrations varying by more than 100-fold in published studies [29]. Similarly, oxycodone is a well-established substrate of the H + OC antiporter [13] and pharmacokinetics and pharmacodynamics of oxycodone also showed an extreme interindividual variation [30]. Clonidine may serve as a third example of a H + OC antiporter substrate with highly variable pharmacokinetics and pharmacodynamics [12]. Notably, the variation of these three drug examples is not explained by known variations in genetically polymorphic drug metabolizing enzymes such as CYP2D6 since these enzymes play no or only a minor role here. Thus, it would be interesting to know whether there is significant genomic variation in the gene coding for the H + OC antiporter or if the activity of the antiporter is significantly modulated by other endogenous substances. However, since we do not know the gene yet, we cannot assess whether this variation is due to acquired or inherited factors.
In conclusion, depending on the functional test and the cutoff value used, up to 50% of drugs tested here that have a pKa > 7.4 and a molecular weight below 500 Dalton may be substrates of the proton-organic cation antiporter. Earlier, more than 60 substrates of the H + OC antiporter have been identified [31], and this number is significantly increased by the substrates identified in this publication. This knowledge may allow designing drugs with better blood-brain barrier penetration [32]. The H + OC antiporter is also expressed in many other tissues. It may be that this H + OC antiporter expressed in many other tissues is coded by the same gene(s) as the antiporter(s) expressed in the hCMEC/D3 cell line used in our experiments. In that case, our data may allow a better understanding of pharmacokinetics, drug-drug interactions and variations in pharmacokinetics and drug effects. Finally, the study illustrates that further efforts to identify the gene(s) coding for the H + OC antiporter(s) are very important.

Materials and Methods
All transport measurements were performed in hCMEC/D3 cells [14], which were cultured in an RPMI 1640 medium (Sigma-Aldrich, Taufkirchen, Germany) supplemented with 10% (v/v) fetal bovine serum, 1% penicillin/ streptomycin and 1 ng/mL bFGF as described earlier [16]. The cells were cultured for a maximum of 30 passages in rat collagen I (R&D System, Minneapolis, MN, USA)-coated 75 mm flasks at 37 • C at 95% relative humidity and 5% CO2. Transport assays were performed in collagen I-coated 12-well plates, in which 0.3 million cells/well were seeded and grown for 72 h to reach more than 90% confluence. All compounds used were purchased with purities > 95% from renowned manufacturers (Sigma-Aldrich, Taufkirchen, Germany; Toronto Research Chemicals, Toronto, ON, Canada; Santa Cruz Biotechnology, Darmstadt, Germany; Tocris, Bioscience, Bristol, UK).
For concentration-dependent uptake, cells were washed with pre-warmed HBSS+ (Hanks' balanced salt solution supplemented with 10 µM HEPES, pH 7.4) and then incubated with increasing concentrations (0.1-1000 µM) of pre-warmed substrate solution prepared in HBSS+ for two minutes at 37 • C. The uptake was stopped by adding ice-cold HBSS+, and cells were immediately washed twice with ice-cold HBSS+. After washing, the cells were lysed with a lysis buffer (80% (v/v) acetonitrile, LGC Standards GmbH, Wesel, Germany) containing an individual internal standard (Table 5). For lysis, cells were incubated for 10 min on a shaking device at a slow shaking rate. Intracellular substrate accumulation was measured by high-performance liquid chromatography-coupled mass spectrometry (HPLC-MS/MS). For protein measurements from each cell culture plate, two wells were not used for transport assays. These wells were also treated with a lysis buffer for ten minutes. These lysates were then transferred to a 96-cell plate and protein concentrations were measured using a bicinchoninic acid assay [33].
For trans-stimulation experiments, cells were preloaded with the antiport probe drugs (diphenhydramine or MDAI, 1 µM final concentration) for 30 min and then quickly washed once with pre-warmed HBSS+ and then incubated with each substrate (250 µM final concentration) for two minutes. The exchange transport was stopped by adding ice-cold HBSS+ and cells were further processed as described above for analysis. Results were expressed as the ratio of the intracellular antiport probe drug concentration by adding a substrate for 2 min over the concentration without adding a substrate. Any ratio significantly below unity indicates an antiport.
Inhibition is also often termed cis-inhibition in the study of the H + OC antiporter acknowledging that the substrate and inhibitor act from the same side of the cell. For these inhibition experiments, the known inhibitors memantine [28], diphenhydramine and imipramine [17] were used at 250 µM and substrates were used at a 3 µM final concentration prepared in HBSS+. Both substrates and inhibitors were pre-mixed and pre-warmed before the incubation for two minutes at 37 • C. The uptake was stopped by adding ice-cold HBSS+ and the cells were further processed as described above for analysis of the intracellular substrate accumulation. Results were expressed as the ratio of the intracellular concentration of the probe drug with inhibition over its intracellular concentration without inhibition.
Carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) was used to investigate proton gradient dependency for uptake. Substrates were incubated for two minutes simultaneously either in the presence of FCCP (25 µM final concentration) or in the absence. Further steps were performed as described above. The final concentration of ethanol (solvent of the 10 mM FCCP stock solution) did not exceed 0.2% (v/v) per reaction. Results were expressed as the ratio of intracellular substrate concentration with FCCP added over the intracellular concentration without FCCP.
Intracellular concentrations of probe substrates and substrates-in-question were performed using HPLC-MS/MS analyses as essentially as described earlier [1,2,16]. Briefly, the cell lysate was centrifuged at 13,000 rpm in a tabletop centrifuge for 15 min to pellet potential cell debris. Supernatant was then transferred to a deep-well collection plate and samples were evaporated at 37 • C under constant nitrogen flow until dry and subsequently reconstituted in 0.1% (v/v) formic acid. HPLC was conducted using a Shimadzu Nexera HPLC system with a LC-30AD pump, a SIL-30AC autosampler, a CTO-20AC column oven and a CBM-20A controller (all Shimadzu, Kyoto, Japan). Liquid chromatography was performed on a Brownlee SPP RP-Amide column (4.6 × 100 mm inner dimension with 2.7 µm particle size; PerkinElmer, Waltham, MA, USA) with a C18 pre-column at an oven temperature of 40 • C. The aqueous mobile phase contained 0.1% (v/v) formic acid and either 3, 8 or 20% (v/v) organic additive (acetonitrile:methanol 6:1 (v/v), LGC Standards GmbH, Wesel, Germany). The separation of samples (injection volume: 10 µL) was carried out with a flow rate of 300 µL/min. An API 4000 tandem mass spectrometer (AB SCIEX, Darmstadt, Germany) was used in a multiple reaction monitoring (MRM) mode for the detection of substances after positive ionization. Internal standards and masses are provided in Table 5. The Analyst software, version 1.6.2, from AB SCIEX was used for the quantification. For the quantification of intracellular concentrations, a calibration curve with known concentrations consisting of at least seven data points was measured for each substance and experiment.
Data analysis: Here, we assume that any influx or efflux transport is the sum of at least two types of transport, passive diffusion and carrier-mediated transport [7]. Since the gene encoding the H+/OC antiporter is still unidentified, we cannot perform uptake experiments overexpressing the H+/OC antiporter as performed with cell lines with and without overexpression of well-identified transporters. Therefore, the net transport mediated by the overexpressed transporters could not be calculated by subtracting the cellular uptake with an empty-vector cell line. However, assuming that the increase in concentration-dependent cell uptake between the two highest substrate concentrations is mostly due to passive diffusion, we subtracted the concentrations corresponding to this slope from the measured intracellular concentrations and used the corresponding curve for the determination of the transport kinetic constants, as described earlier [16].
K M and V max values with corresponding standard errors were calculated from at least three independent experiments by nonlinear regression analysis using the Michaelis-Menten equation (v = (V max × [S])/(K M + [S]) and using the least squares fit method in GraphPad software. The intrinsic clearance Clint was calculated as the ratio of Vmax over K M .
The uptake ratio for cis-inhibition, trans-stimulation and proton-dependent uptake experiments was calculated relative to the respective control experiment as uptake ratio = intracellular concentration/intracellular concentration in the control. The differences between the test and the control experiment in cis-inhibition, trans-stimulation and proton-dependent uptake experiments were tested for statistical significance using the Student's t-test. Transport kinetic parameters were calculated in the GraphPad Prism software version 5.01 for Windows (GraphPad Software, La Jolla, CA, USA). The chemical features of all substrates were determined using Instant JChem (ChemAxon, Budapest, Hungary) processing the individual isomeric SMILES from PubChem [34].