In vitro characterization of pralidoxime transport and acetylcholinesterase reactivation across MDCK cells and stem cell-derived human brain microvascular endothelial cells (BC1-hBMECs)

Current therapies for organophosphate poisoning involve administration of oximes, such as pralidoxime (2-PAM), that reactivate the enzyme acetylcholinesterase. Studies in animal models have shown a low concentration in the brain following systemic injection. To assess 2-PAM transport, we studied transwell permeability in three Madin-Darby canine kidney (MDCKII) cell lines and stem cell-derived human brain microvascular endothelial cells (BC1-hBMECs). To determine whether 2-PAM is a substrate for common brain efflux pumps, experiments were performed in the MDCKII-MDR1 cell line, transfected to overexpress the P-gp efflux pump, and the MDCKII-FLuc-ABCG2 cell line, transfected to overexpress the BCRP efflux pump. To determine how transcellular transport influences enzyme reactivation, we developed a modified transwell assay where the inhibited acetylcholinesterase enzyme, substrate, and reporter are introduced into the basolateral chamber. Enzymatic activity was inhibited using paraoxon and parathion. The permeability of 2-PAM is about 2 × 10−6 cm s−1 in MDCK cells and about 1 × 10−6 cm s−1 in BC1-hBMECs. Permeability is not influenced by pre-treatment with atropine. In addition, 2-PAM is not a substrate for the P-gp or BCRP efflux pumps. The low permeability explains poor brain penetration of 2-PAM and therefore the slow enzyme reactivation. This elucidates one of the reasons for the necessity of sustained intravascular (IV) infusion in response to organophosphate poisoning.

2-PAM is an ionic molecule and hence the permeability across the blood-brain barrier has been assumed to be very low [9,10]. Clinical trials have shown that 2-PAM is rapidly cleared from the body [11,12], highlighting the need for continuous infusion to maintain a therapeutic dose [11,[13][14][15]. Based on animal studies, the minimum effective concentration in blood is reported to be around 4 mg L −1 (about 30 µM) [9,15]. Therefore, to assess transport into the brain we studied permeability of 2-PAM in four cell lines: MDCKII, MDCKII-MDR1, MDCKII-FLuc-ABCG2, and BC1-hBMECs. Madin-Darby canine kidney epithelial cells (MDCKs) are widely used for in vitro assessment of brain penetration and the permeability values for a wide range of solutes have been reported [16]. The MDCKII-MDR1 cell line is transfected to express the human P-gp efflux pump, and the MDCKII-FLuc-ABCG2 line is transfected to overexpress BCRP efflux pump. Human brain microvascular endothelial cells (BC1-hBMECs) are derived from humaninduced pluripotent stem cells (hiPSCs) [17][18][19].
All experiments with MDCK cells were performed in Hank's balanced salt solution (HBSS) with 10 mM HEPES (Sigma) and 15 mM glucose (Sigma), pH 7.4. After incubation in media for 2 h, cell monolayers were immersed in fresh HBSS for 30 min to remove any traces of media. Then 100 µM or 10 µM 2-PAM (pralidoxime chloride, Sigma) in HBSS was pipetted into either the apical or basolateral chamber, with HBSS on the receiving side. Cell monolayers with test solutes were incubated at 37 °C with 5 % CO 2 on a rocker to ensure good mixing.
The concentration of 2-PAM was determined by HPLC (1260 Infinity HPLC, Agilent, Santa Clara, CA, USA) with UV Vis detection at 296 nm. Solvents were degassed by sonication for 45 min before use and all samples were run at room temperature. An isocratic flow of 44 vol.% acetonitrile (HPLC grade, Chromasolv, Sigma) and 56 vol.% ammonium acetate (0.03 M; HPLC grade, Sigma), pH 4.5, was used with a PolyCAT A column (100 × 2.1 mm, 5 µM, 300 Å, 102CT05-03, Poly LC Inc, Columbia, MD, USA) [26]. Calibration curves were constructed from standard solutions with concentrations of 0.1, 1, 10 and 100 µM. Due to the simplicity of the procedure, no internal standard was used.

2-PAM/atropine
To assess whether atropine, which is often co-administered with 2-PAM, modulates the transport of 2-PAM we performed experiments where MDCK cells were pretreated with atropine. After 2 h incubation in media, and 30 min rinse in HBSS, MDCKII monolayers were pretreated with 1 µM atropine for 30 min, rinsed in HBSS for 5 min, and then incubated in 100 µM 2-PAM for permeability measurements using the same procedure as described above. 2-PAM concentrations were measured by HPLC with an isocratic flow of 55 vol.% acetonitrile and 45 vol.% ammonium acetate (0.03 M).

Rhodamine 123
To confirm the up-regulation and polarization of P-gp efflux pumps to the apical face, we measured the permeability of Rhodamine 123, a known P-gp substrate, across MDCKII and MDCKII. MDR1 monolayers [27]. Permeability experiments were performed for 60 min at a concentration of 50 µM. The concentration of Rhodamine 123 (excitation 486 and emission 523) was measured by fluorescence (Fluorolog, Horiba Scientific, Edison, NJ, USA). Calibration curves were generated over the concentration range from 0.001 to 1 µM.

Coupled permeability and acetylcholinesterase reactivation
To assess coupled 2-PAM transport and enzyme reactivation, experiments were performed in a transwell device with acetylcholinesterase in the basolateral chamber. Confluent monolayers of MDCKII cells were formed as described above. Electric eel acetylcholinesterase (AChE, 1U or about 1 µL, >1000 U/mg, Sigma) was placed into the basolateral chamber (24 well plate). A mixture of Ellman's reagent (DTNB), final concentration 300 µM, and acetylthiocholine (ASCh), final concentration 450 µM, dissolved in HBSS, was introduced into the basolateral chamber, to give a final volume of 600 µL. The time-dependent activity of the enzyme was determined from the absorbance of the DTNB reporter at 412 nm using a plate reader (Spectramax M3). Results were normalized to the activity of the uninhibited enzyme, unnormalized data are provided in the Additional file 1 ( Figure S1: Non-normalized reactivation data).
Inhibition was achieved by incubating the enzyme with 0.72 mM parathion (PESTANAL-grade, Sigma) or 4.6 µM paraoxon (PESTANAL-grade, Sigma) for 20 min prior to experiments. In inhibition experiments, the enzyme was inhibited with parathion. Paraoxon, a metabolite of parathion, is about three orders of magnitude more potent as an anticholinesterase inhibitor [28]. The parathion concentration was 157-fold higher than the paraoxon concentration, reflecting their different activities. For reactivation experiments, 2-PAM was introduced into either the apical or basolateral chamber in HBSS. Introducing 2-PAM into the basolateral chamber simulates reactivation alone, whereas introduction of 2-PAM into the apical chamber simulates coupled trans-endothelial transport and reactivation.

Positive control (uninhibited enzyme + substrate)
To assess the kinetics of enzyme interaction with the substrate, uninhibited enzyme (AChE), along with substrate (ASCh), and reporter (DTNB) in HBSS were introduced into the basolateral chamber. An apical transwell chamber with a monolayer of MDCK cells was located on the top of the basolateral chamber to ensure that the control was performed in the same way as the other experiments.

Negative control (inhibited enzyme + substrate)
To assess the efficiency of enzyme inhibition, AChE was mixed for 20 min with concentrated parathion (0.72 mM final concentration) or paraoxon (4.6 µM final concentration) organophosphates (OP). The inhibited enzyme (AChE-OP) was then placed in the basolateral chamber with substrate (ASCh) and reporter (DTNB) in HBSS. An apical transwell chamber was located on the top of the basolateral chamber as described above.

Direct interaction of 2-PAM (inhibited enzyme + substrate + reactivator)
To assess the kinetics of direct reactivation of inhibited enzyme, 100 µM 2-PAM was introduced in the basolateral chamber with the inhibited enzyme (AChE-OP), substrate (ASCh) and reporter (DTNB) in HBSS. An apical transwell chamber was located on the top of the basolateral chamber as described above.

Coupled transport of 2-PAM and reactivation
To evaluate the coupled transport of 2-PAM across a cell monolayer and reactivation of inhibited enzyme, 100 µM of 2-PAM was introduced into the apical chamber, with inhibited enzyme (AChE-OP), reporter (DTNB), and substrate (ASCH) in the basolateral chamber.

Statistics
Permeability, activity, and reactivation half-time represent the mean ± standard deviation. Statistical significance was determined using a student's t test (two-tailed with unequal variance) with p < 0.01 ** and p < 0.001 ***. The average permeability values for the MDCK cell lines were calculated from analysis of all of the replicates. Due to variations between differentiations, the average permeability across the BC1-hBMECs was calculated from the average values from each differentiation. Similarly, the efflux ratio was calculated from the average value obtained from each differentiation.

Influence of atropine on 2-PAM permeability
To determine whether atropine, co-administered with 2-PAM, modulates the permeability of 2-PAM, we measured the permeability of 2-PAM following pretreatment of the MDCKII monolayer with atropine. The permeability of 2-PAM was 2.54 ± 0.33 × 10 −6 cm s −1 , which was not significantly different to the value of 2.99 ± 1.12 × 10 −6 cm s −1 obtained without pre-treatment with atropine ( Table 1).

Table 1 Permeability of pralidoxime (2-PAM), rhodamine 123 R123) and Lucifer yellow (Ly) across MDCKII, MDCKII-MDR1, MDCKII-FLuc-ABCG2, and BC1-hBMEC monolayers
A→B represents apical-to-basolateral permeability, and B→A represents basolateral-to-apical permeability. Permeability values are reported as mean ± standard deviation. The efflux ratio is the ratio of basolateral-to-apical permeability divided by the apical-to-basolateral permeability. For MDCK cells, permeabilities and efflux ratios were calculated from the total number of replicates (N). Data were obtained from at least three independent experiments each with two or more replicates. For the BC1-hBMECs, the permeabilities and efflux ratios were calculated from the average of each differentiation, where N represents the number of independent differentiations rhodamine in MDCKII and MDCKII. MDR1 cells were significantly different (p = 0.02) supporting upregulation and polarization of P-gp efflux pumps to the apical side of the MDCKII.MDR1 cells. Apical-to-basolateral permeabilities of 0.83 × 10 −6 and 0.89 × 10 −6 cm s −1 , with corresponding efflux ratios of 9 and 115, have been reported for transport of 5 µM rhodamine 123 across MDCKII. MDR1 cells [27,29]. The reported efflux ratio for rhodamine in BC1-hBMEC cells is approximately 4 [19].

Coupled transport and acetylcholinesterase (AChE) reactivation
To assess the coupled transcellular transport and AChE reactivation, we performed transwell experiments with a monolayer of MDCKII cells and inhibited enzyme in the basolateral chamber (Fig. 2a). Acetylcholinesterase was inhibited with an organophosphate (parathion or paraoxon) for 20 min and then introduced into the basolateral chamber of a transwell device, along with acetylthiocholine (ASCh) and the colorimetric reporter (DTNB). Control experiments were performed to confirm the activity of the enzyme and effectiveness of the inhibitor.
In the absence of 2-PAM, the activity of the inhibited enzyme (AChE-OP) increased very slowly during the 2 h experiment (Fig. 2b). When 2-PAM was introduced into the basolateral chamber with inhibited enzyme, reactivation occurred much more quickly (Parathion: inhibited enzyme and direct reactivation p < 0.001, Paraoxon: inhibited enzyme and direct reactivation p < 0.01). However, when 2-PAM was introduced into the apical chamber, reactivation of the inhibited enzyme was slowed considerably due to the coupled transport and reactivation (Parathion: direct reactivation and transport reactivation p < 0.001, paraoxon: direct reactivation and transport reactivation p < 0.01).
The activity of the enzyme following transcellular transport of 2-PAM across MDCKII monolayers (3.49 × 10 −4 abs s −1 ) was about four fold lower than when the inhibited enzyme was directly exposed to 2-PAM (1.35 × 10 −3 abs s −1 ) (Fig. 2c). Similarly, the halftime for reactivation of the substrate increased sixfold from 680 s for direct reactivation to 4100 s following transcellular transport (Fig. 2d).
MDCK cells are widely used to assess brain penetration of small molecules. Although MDCK cell lines are epithelial in origin and not human, they express tight junction proteins, which limit paracellular transport. Variants such as MDCKII-MDR1 can be used to determine whether a solute is an efflux pump substrate. The stem cell derived BC1-hBMECs exhibit high transendothelial electrical resistance (TEER > 1000 Ω cm 2 ), low permeability to solutes such as Lucifer yellow, and express tight junction proteins (e.g. claudin-5), transporters (e.g. LAT-1), and efflux pumps (e.g. P-gp) [17,19]. The permeability of the stem cell derived BC1-hBMECs (p = 1.12 ± 0.80 × 10 −6 cm s −1 ) was slightly lower than values obtained in MDCK cells, but in the range that is consistent with slow accumulation in the brain.
High permeability values are usually associated with small molecular weight and moderate lipophilicity [31,32]. While 2-PAM has a molecular weight under 500 Da (172 Da), fewer than 5 hydrogen bond donors (1), and fewer than 10 hydrogen bond acceptors (2), the charge results in a low lipophilicity and hence 2-PAM is not expected to have a high permeability.
There was no significant difference between apical-andbasolateral and basolateral-to-apical permeabilities in MDCK cells, indicating that 2-PAM is not a substrate of the P-gp or ABCG2 pumps. To confirm the polarized expression and activity of the P-gp efflux pumps, we determined efflux ratios of 10.7 and 20.3 for the MDCKII and MDCKII. MDR1 cell lines for the known P-gp substrate rhodamine 123.
Treatment for organophosphate poisoning involves co-administration of 2-PAM and atropine. The permeability of 2-PAM was the same in MDCKII cells and cells pretreated with atropine, showing that atropine does not modulate the permeability of 2-PAM.

Coupled transcellular transport and enzyme reactivation
To study coupled transcellular transport of the neurotoxin antidote 2-PAM with enzyme reactivation, we developed a modified transwell assay with inhibited enzyme (AChE-OP), substrate (ASCh), and reporter (DTNB) in the basolateral chamber. When 2-PAM was introduced into the apical chamber of the transwell device, the activity of the enzyme decreased four fold compared to the case where 2-PAM was introduced directly into the basolateral chamber. Similarly, the halftime for reactivation of the enzyme increased six fold when coupled to transcellular transport. These results highlight the difficulty in maintaining a therapeutic dose when the permeability is low.

Conclusions
The permeability of the nerve agent reactivator 2-PAM is 1 × 10 −6 -2 × 10 −6 cm s −1 and is not influenced by pre-treatment with atropine. In addition, 2-PAM is not a substrate for the P-gp or BCRP/ABCG2 efflux pumps. Similar permeability values were obtained for human brain microvascular endothelial cells derived from induced pluripotent stem cells. In a modified transwell assay to couple transcellular transport and enzyme reactivation, we showed that transcellular transport decreased enzymatic activity four fold and increased the reactivation half-time six fold. The low permeability explains poor brain penetration of 2-PAM and the necessity for sustained IV infusion in response to organophosphate poisoning.