3-(4,5-Dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT), vinblastine sulfate (VLB, 96%), cyclosporine A (CsA, 98.5%), trifluoperazine dihydrochloride (TFP, 99.0%), histopaque 1077, rhodamine 123 (Rho 123), and lectin from Phaseolus vulgaris (PHA) were purchased from Sigma-Aldrich Co (St Louis, MO). Doxorubicin hydrochloride (DOX, 99.8%, Synbias Pharma Ltd.) was obtained from Nanox Release Technology (Buenos Aires, Argentina) and was used dissolved in bi-distilled water. Verapamil hydrochloride 98% was provided by Parafarm (Buenos Aires, Argentina) and was used as the reference P-gp inhibitor dissolved in ethanol at 30 μM. RPMI-1640 and Gibco® cell culture reagents were purchased from Invitrogen Life Technologies (Carlsbad, CA). Sterile plastic material was purchased from Greiner Bio-One (Frickenhausen, Germany). All solvents were HPLC grade.

FITC mouse anti-human (P-gp) was purchased from BD (BD Biosciences, USA). MDR1 PREDEASY™ ATPase assay kit was obtained from Solvo Biotechnology (Szeged, Hungary). Flow cytometry was performed in a Becton Dickinson (BD) FACSCanto II flow cytometer (BD Biosciences, USA).

Compounds 1–15 (Figure 1) were previously isolated in the laboratory from native and naturalized plants from Argentina. They were tested at 90% or higher purity, determined by HPLC. The compounds assayed were: vanillin (1) and, 4-hydroxy-3-methoxycinnamaldehyde (2), both isolated from Melia azedarach (Carpinella et al., 2003), ilicol (3) isolated from Flourensia oolepis (Diaz Napal and Palacios, 2013), scopoletin (4) obtained from M. azedarach (Carpinella et al., 2005), (Z,Z)-5-(trideca-4,7-dienyl)-resorcinol (5) isolated from Lithrea molleoides (Carpinella et al., 2011), 2′,4′-dihydroxychalcone (6) and (-)-pinocembrin (7) both obtained from F. oolepis (Diaz Napal et al., 2009; Joray et al., 2015), naringenin (8) isolated from Baccharis salicifolia (Céspedes et al., 2006; del Corral et al., 2012), dalenin (9) obtained from Dalea elegans (Chiari et al., 2011), apigenin (10) isolated from B. salicifolia (del Corral et al., 2012), quercetin (11), 3-O-methylquercetin (12), and 23-methyl-6-O-desmethylauricepirone (13) obtained from Achyrocline satureioides (Joray et al., 2011, 2013) and (±)-pinoresinol (14) and meliartenin (15) both obtained from M. azedarach (Carpinella et al., 2002, 2003). 1-acetoxy-(+)-pinoresinol (16) (Figure 1) was purchased from Chem Faces (Wuhan, PR). The ¹H, ¹³C, 2D-NMR spectra, and HPLC chromatograms of the compounds are available upon request from the authors.

The K562 human CML cell line (Rumjanek et al., 2013) and its MDR counterpart, Lucena 1, were used (Moreira et al., 2014). Lucena 1 was chosen as a well-characterized resistant cell line in which P-gp overexpression is the exclusive mechanism of acquired resistance (Moreira et al., 2014). Real time quantitative PCR analysis for P-gp and MRP1 showed no statistical differences between cells maintained with vincristine and those maintained with doxorubicin. Both lines were routinely maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 U mL⁻¹ penicillin and 100 μg mL⁻¹ streptomycin in a 5% CO₂ humidified atmosphere at 37°C. Cells were subcultured twice a week and used when under 20th passage from frozen stocks. Lucena 1 cells were cultured in the presence of 60 nM of DOX to maintain P-gp overexpression, until 4 days before the experiments, when they were transferred to drug-free medium. These cells displayed a higher superficial P-gp expression than parental K562 cells (see Figure 2). All experiments were performed with the cells in the logarithmic growth phase with cell viabilities above 90% determined by staining with trypan blue.

Compounds 1–15 and then compound 16, which was selected after demonstrating lower binding energy than its lead compound 14 and other pinoresinol derivatives, were screened by MTT colorimetric assay for their ability to reverse cellular resistance to DOX. Briefly, 5 × 10⁴ Lucena 1 cells per well were seeded in 96-well plates containing RPMI-1640 medium in the presence of DOX alone or in combination with the tested compounds dissolved in ethanol or acetonitrile as appropriate. DOX was added to reach final concentrations of 0.05–431 μM. The compounds were assayed at the highest non-toxic concentration observed in both cell lines (cytotoxicity ≤ 20%, Cytotoxicity (%) = [1-(Optical density treatment-Optical density DMSO)/(Optical density control-Optical density DMSO)] x 100, determined by MTT proliferation assay (Joray et al., 2015); maximum concentration tested 40 μg mL⁻¹). Following the primary screening, compounds 14 and 16 (dissolved in ethanol) were tested at serial dilutions. Negative controls containing only bi-distilled water, ethanol or acetonitrile (1% v/v since no adverse effects were observed at this concentration) were simultaneously run as well as viability controls with no addition of the dissolution solvents. Verapamil was used as a positive control at 30, 0.11 and 0.055 μM. After 48 h incubation at 37°C with 5% CO₂, 20 μL of 5 mg mL⁻¹ MTT solution in sterile PBS was added to each well and incubation continued for 4 h. Subsequently, the supernatants were removed and replaced with 100 μL DMSO to solubilize the resulting purple formazan crystals produced from metabolically viable cells. Absorbance was measured with an iMark micro-plate reader (Bio-Rad, USA) at 595 nm.

Half inhibitory concentrations (IC₅₀) represent the concentrations of DOX required to inhibit 50% of cell proliferation (compared to the solvent controls, which showed no differences in comparison to viability controls) and were calculated from the mean values of data from wells.

The same reversal assay was performed in K562 cells, in order to discard a decrease in the IC₅₀ of DOX due to effects other than P-gp inhibition. The reversal fold (RF) values for the tested compounds, which indicate their capacity to reduce resistance to DOX, were calculated by dividing the IC₅₀ of DOX alone by the IC₅₀ of DOX in the presence of the tested compounds (Xu et al., 2014).

The assays were also performed with VLB, which was added, dissolved in DMSO at 2 × 10⁻⁵ to 44 μM alone or in combination with 14 at 112 μM. The negative control contained DMSO, while verapamil was used as the reference compound.

The effect of 14 or 16 on DOX intracellular accumulation was further studied.

In order to establish the best assay conditions, 2.5 × 10⁵ Lucena 1 or K562 cells mL⁻¹ were seeded onto 24-well plates and pre-incubated in complete RPMI-1640 medium in the presence of 112 μM of 14 (dissolved in ethanol), verapamil or ethanol (1% v/v) from 0 to 48 h at 37°C with 5% CO₂. Following incubation, 5 μM DOX was added and cells were further incubated for 1 h.

Time course of DOX accumulation was investigated by co-incubating 2.5 × 10⁵ Lucena 1 or K562 cells mL⁻¹ with 112 μM of 14 or 1% ethanol in 96-well plates for 1 h at 37°C prior to the addition of 5 μM DOX. The intracellular DOX was measured at increasing periods of time.

To obtain further information about the type of inhibition exerted by compound 14, the kinetic behavior was analyzed by the Lineweaver-Burk double-reciprocal method compared to data obtained in the absence of the inhibitor (ethanol control) (Arnaud et al., 2010). The drug retention rate corresponding to the amount of DOX remaining inside 50,000 Lucena 1 cells after 1 h incubation was plotted vs. DOX concentrations and fitted with the Michaelis-Menten equation, R = R_max [S]/K_m + [S], where R is the drug retention rate, R_max the maximal drug retention rate, [S] is the substrate concentration and K_m the Michaelis constant of DOX efflux (Copeland, 2000; Arnaud et al., 2010).

With the aim of determining the minimum effective concentration (MEC) of 14 or 16, 2.5 × 10⁵ Lucena 1 or K562 cells mL⁻¹ were incubated with 7–112 or 0.027–14 μM of each compound, respectively, for 1 h at 37°C with 5% CO₂. Afterwards, 5 μM solution of DOX was added and the plates were further incubated for 1 h. Verapamil or ethanol (1%) were run as positive and negative controls, respectively. Viability controls were included in all the experiments.

After DOX incubation in all the assays described, cells were placed on ice to stop the reactions, followed by washing twice with ice-cold PBS. The intracellular DOX-associated fluorescence of 50,000 individual cells was determined by FACScan flow cytometry. DOX was excited at 488 nm and the emitted light was collected with a 585/42 nm bandpass filter. Dead cells and cell debris were excluded by forward and side scatter gating. Mean fluorescence intensities (MFI) were analyzed with Flowjo software (Tree Star, Inc. Ashland, OR).

The data collected was compared to that obtained with 30 μM verapamil (which was considered as the maximum inhibition) and expressed as the published equation (Huang et al., 2013):

With the aim of discarding fluorescence by the compounds themselves, flow cytometry of 14 or 16 was performed at the maximum concentration tested in the absence of DOX. No fluorescence due to these compounds was observed.

To determine the effect on DOX efflux, 2.5 × 10⁵ Lucena 1 or K562 cells mL⁻¹ were first pre-incubated in 24-well plates in the presence of 14, dissolved in ethanol at 112 μM, verapamil or ethanol (1% v/v) for 1 h at 37°C with 5% CO₂. After incubation, DOX 5 μM was added to the reaction mixture and further incubated for 1 h. Subsequently, cells were washed with cold PBS and incubated for different times with 14, verapamil or ethanol in DOX-free medium to allow dye extrusion. With the aim of determining the minimum effective concentration (MEC) of 14, 2.5 × 10⁵ Lucena 1 cells were incubated with 7–112 μM of the lignan, for 1 h at 37°C with 5% CO₂. Afterwards, 5 μM solution of DOX was added and the plates were further incubated for 1 h. After washing, cells were incubated for 30 min with 14. Verapamil or ethanol (1%) were run as positive or negative controls, respectively. Cells were then washed with ice-cold PBS and the intracellular DOX-associated MFI of 50,000 individual cells was determined by flow cytometry as previously described.

Rho 123, a P-gp fluorescent substrate, is frequently employed as an indicator of P-gp activity and it was therefore used in an additional efflux study. Briefly, Lucena 1 or K562 cells at a density of 2.5 × 10⁵ mL⁻¹ were pre-incubated with 14 dissolved in ethanol at 112 μM (2 × MEC), TFP (8 μM), verapamil or ethanol (1% v/v) for 1 h at 37°C with 5% CO₂. Subsequently, cells were incubated with 500 ng mL⁻¹ of Rho 123 for 30 min. After incubation, the medium was removed and the cells were incubated in Rho 123-free medium with 14, verapamil, TFP and ethanol, for a further 30 min to allow Rho 123 extrusion. After washing twice with cold PBS, the amount of Rho 123 remaining in 10,000 individual cells after the efflux period was quantified by FACScan flow cytometry. Excitation was performed with a laser operating at 488 nm and the emitted fluorescence was collected through a 530/30 nm pass filter.

The ATPase activity of P-gp was determined using the PREDEASY™ ATPase assay kit as per the manufacturer's recommendation. In the activation assay, increasing concentrations of 14 (0.27–600 μM) or verapamil (0.14–300 μM), both dissolved in DMSO (2% final concentration), were pre-incubated in ATPase assay buffer with membrane preparations from Spodoptera frugiperda ovarian cells (Sf9) containing human P-gp and 10 mM MgATP for 10 min at 37°C. For monitoring the inhibition of the substrate stimulated-ATPase activity, verapamil (40 μM) was added to the incubation mixture as an ABC transporter-related ATPase activator. The ATPase reaction was subsequently stopped and the inorganic phosphate (Pi) produced was measured colorimetrically at 630 nm. DMSO was used as a solvent control and CsA (final concentration of 40 μM) was used as reference inhibitor in the inhibition assay.

All experiments were performed in the absence or presence of 1.2 mM of sodium orthovanadate, an ABC transporter-related ATPase inhibitor, in order to measure the vanadate-sensitive portion of the total ATPase activity. ATPase activities were determined as Activity (%) = (A − B) – (E − F) × 100/(C − D) – (E − F), where A is the activity in the presence of 14 alone or with verapamil as activator in the inhibition study, B is the activity in the presence of 14 alone or with verapamil in the inhibition study, in the presence of vanadate (background activity), C is the maximum activation value due to verapamil and D is the same value as C but with the addition of vanadate (background activity), E is the activity of control with DMSO while F is the activity of DMSO in the presence of vanadate. Membranes from Sf9 cells expressing defective P-gp were used as controls.

The structural model of the P-gp used for the docking studies was a homology model of the human P-gp previously proposed in Jara et al. (2013). In this work, this model was subject to stochastic molecular dynamics for obtaining average inter-residue distances in good agreement with the experimental distances and challenged to reproduce the right activity order of a pool of known inhibitors, as well as to find binding sites in good agreement with those proposed on computational and experimental bases in the literature (Jara et al., 2013).

The following probe ligands were docked in the structural models of the P-gp:

pinoresinol derivatives, such as 1-acetoxy-(+)-pinoresinol (16), 1-acetoxy-(−)-pinoresinol (16a), 1-hydroxy-(+)-pinoresinol, phylligenin, pinoresinol diglucoside and (+)-pinoresinol 4-glucoside submitted to docking to establish whether their activity improved with respect to 14.

The structures of these ligands were obtained by performing a conformational search (when relevant) and a full geometry optimization at the semiempirical PM6 level of theory, characterizing the structures as minima by diagonalizing the Hessian matrix and ensuring the absence of negative eigenvalues; next, a refinement was made at the PBE0/6-31G^* level using the Gaussian 09 (Rev. B01) package (http://www.gaussian.com). The Autodock 4.2.6 package (Morris et al., 2009) was used for the docking simulations, precomputing a grid in the interior of the whole TMD. Considering the large size of the docking region, 2,000 runs of Lamarckian genetic algorithm were performed for each ligand (Morris et al., 1998, 2009). The population was set at 150 individuals, up to 10⁵ generations with 1 survivor per generation and a limit of 6 × 10⁶ energy evaluations and the remaining algorithm control parameters set to program defaults. The cluster analysis was made with 2.5Å of RMSD. Molecular graphics rendering was performed using VMD 1.9.2 (Humphrey et al., 1996).

Besides the main grid box including the transmembrane region, molecular docking on the NBDs was carried out to obtain preliminary information about the presence of direct binding of 14 to the NBDs.

To determine the effect of 14 on P-gp surface expression, Lucena 1 or K562 cells at a density of 5 × 10⁴ cells mL⁻¹ were incubated in the presence of 14, dissolved in ethanol at 112 μM (maximum non-cytotoxic concentration, in order to ensure the presence of the effect) or 1% ethanol (control) for 24 and 48 h. Then, cells were washed with cold PBS and labeled for 30 min at 4°C in the dark with FITC-conjugated mouse anti-human P-gp antibody, which binds to an external epitope of P-gp, according to the manufacturer's instructions. Finally, cells were washed and suspended with ice-cold PBS and fluorescence intensity was determined in 10,000 individual cells by FACScan flow cytometry. FITC fluorescence was measured at an excitation wavelength of 488 nm and emitted light was collected with a 530/30 nm bandpass filter.

The cytotoxicity of 14 and 16 on peripheral blood mononuclear cells (PBMC) was evaluated by MTT assay (Joray et al., 2015). PBMC were collected from fresh heparinized blood and separated by density gradient centrifugation (Ficoll®) as described by Rennó et al. (2008). As the current study required samples from healthy human volunteer donors, ethical approval was provided by the Catholic University of Córdoba Research Ethics Board. Signed informed consents were obtained from donors. For the cytotoxicity assay, 1 × 10⁵ PBMC/well were incubated in duplicate in 96-well plates with PHA 10 μg mL⁻¹, in the presence of increasing concentrations of 14 (28–560 μM), 16 (3.5–315 μM) (both dissolved in ethanol) or 1% ethanol for 48 h. Absorbance (Abs) and percentage of cytotoxicity were determined as described above and the IC₅₀ values were calculated. The experiment was carried out in two separate stages.

The results are expressed as mean ± SE. Data were analyzed using Student's t-test or two-way analysis of variance (ANOVA) using GraphPad Prism software (Graphpad Prism 5.0, Graphpad Software, Inc., CA, USA), with p ≤ 0.05 as statistically significant. All experiments were performed in duplicate or triplicate at least three times. The 50% inhibitory concentrations (IC₅₀) were calculated by log-Probit analysis responding to at least seven concentrations at the 95% confidence level with upper and lower confidence limits. The curves from the ATPase assays were fitted to the relative activity vs. compound 14 concentrations plot using non-linear regression. Top (maximal response) and bottom (maximally inhibited response) values were not constrained to constant values 100 and 0, respectively.

Since there is no common pharmacophore determining that a compound behaves as a P-gp reverser (Yuan et al., 2012) and due to the wide diversity of chemical structures that interact with this transporter (Robert and Jarry, 2003) as well as the lack of information on metabolites from Argentinian flora as P-gp chemosensitizers, we decided to investigate this effect in 15 plant-derived compounds belonging to different chemical families obtained from plants from central Argentina (Carpinella et al., 2002, 2003, 2005; Diaz Napal et al., 2009; Chiari et al., 2010, 2011; Joray et al., 2011, 2013, 2015; del Corral et al., 2012; Diaz Napal and Palacios, 2013). Compounds 1–15 were studied by an MTT assay, in order to evaluate the potentiation of DOX cytotoxicity in insensitive Lucena 1 cells, which were 35-fold more resistant to this drug [IC₅₀ = 40.78 (16.60–100.18) μM] than their parental cell line K562 [IC₅₀ = 1.16 (0.52–2.58) μM]. The combination of the maximum non-toxic concentrations of the tested compounds with DOX showed the flavonoid 9 and the lignan 14 as the most effective, enabling the IC₅₀ value of DOX in Lucena 1 cells to be decreased by a factor of 3.2 and 9.4, respectively (Table 1). The latter value is similar to that obtained with verapamil 30 μM (p > 0.05). When 14 was also tested at 28 μM, it caused a 3.4-fold increase in Lucena 1 sensitivity to DOX (Table 1), indicating that this compound was as active as 9 (p > 0.05). Given the highest effect reached by 14 when tested at a higher concentration, due to its low cytotoxicity, we selected 14 as the first compound to be further studied. As shown in Table 1, compound 14 was able to potentiate DOX toxicity in a dose-dependent manner (b = 0.070; p = 0.013; CI 95% = 0.016 to 0.124) from 7 μM (see also Figure 3). The comparison of the dose-response curves of DOX alone and in combination with 14 clearly showed the enhancement in DOX cytotoxicity (Figure 3). It is worth noting that 14 did not increase DOX sensitivity in K562 (Table 1 and Figure 3), while compounds 6 and 9 caused a decrease in DOX IC₅₀, with the former also showing reversing properties in Lucena 1 (Table 1). These results further support the need to first investigate the activity of compound 14.

We also studied the reversal property of 14 to VLB resistance in Lucena 1 [VLB IC₅₀ values of 2.49 μM (0.34–18.20) and 0.24 μM (0.04–1.89) in Lucena 1 and K562, respectively]. When 14 was co-administered with VLB, it significantly restored the sensitivity of the MDR cell line to VLB with a reversal fold activity of 60.20 (p < 0.0001), while the RF value in K562 was 9.63 (p < 0.05). Verapamil induced a decrease in the VLB IC₅₀ of 49.17 (p < 0.0001) and 2.35-fold (p < 0.05) in Lucena 1 and K562, respectively.

Based on the results of the reversal assay, we evaluated the capacity of 14 to inhibit the function of P-gp by measuring the intracellular accumulation of DOX by flow cytometry.

To determine the desirable period of incubation, the uptake profile of DOX was first investigated at different periods of pre-incubation with 14. As seen in Figure 4A, a significant increase (p < 0.01) in DOX accumulation by a factor of 1.3-fold was observed in Lucena 1 cells in the absence of pre-incubation compared with ethanol control cells, which showed the same rate of accumulation as untreated cells. The increase in DOX-associated MFI remained till 48 h of pre-incubation achieving a 1.4-fold increase. These values compared favorably with the 1.4 and 1.8-fold increase observed with verapamil at 0 and 48 h, respectively (p > 0.05). The highest difference in DOX accumulation with respect to the negative control was observed at 1 h pre-incubation (p < 0.001), at which time the highest percentage of inhibition related to verapamil was obtained (79% inhibition). Given these results, the selected pre-incubation time for further analysis was set at 1 h. No such increase in MFI was observed in K562 cells treated with 14 or with verapamil (Figure 4B).

The time course of DOX accumulation was further investigated. DOX- associated MFI increased in a time-dependent response in Lucena 1 treated with 14 with significantly higher values than that of Lucena 1 ethanol-treated cells (Figure 5). At 20 min, it was observed that 14 fully restored the presence of DOX within Lucena 1 cells (p < 0.05), reaching the MFI values observed in K562 (p > 0.05). This tendency was maintained over the whole period of time. The highest difference in MFI values in relation to the Lucena 1 ethanol control was observed at 1 h (Figure 5), determining that 1 h incubation with DOX was adequate.

To further understand the mechanism of P-gp inhibition, we studied the behavior of 14, monitoring the intracellular retention of DOX by flow cytometry. Increasing the concentrations of 14, the values of MFI remained the same, while the Michaelis-Menten constant (K_m) increased, as observed in the family of straight lines passing through the same point of the vertical axis (Figure 6). This plot indicated that 14 was a competitive inhibitor.

Following primary screening, 14 was assayed at serial dilutions to determine its minimum effective concentration (MEC). It significantly increased the intracellular accumulation of the substrate in Lucena 1 from 56 μM onwards (p < 0.05), with no differences with respect to verapamil (p > 0.05) (data not shown). No dose-dependency was observed. At the concentration mentioned and up to the maximum concentrations evaluated, no increase was observed in DOX accumulations in K562 (p > 0.05) (data not shown).

Since it was determined that 14 is capable of reversing DOX accumulation deficit, it was investigated whether this effect can be maintained after removing DOX from the medium. Figure 7A shows that 14 clearly increased DOX retention in Lucena 1 cells. At 30 min, the DOX-associated MFI increased 1.8-fold in comparison with control cells (p < 0.05) with similar activity to that observed with verapamil (p > 0.05). The effectiveness of 14 was maintained until the end of the experiment (Figure 7A). Neither 14 nor verapamil showed any significant effect on DOX efflux in K562 cells (Figure 7B).

The minimum effective concentration (MEC) corresponded to 56 μM (p < 0.01), with significant differences with respect to verapamil (p < 0.01; Figure 8). A dose-dependent effect was observed (b = 0.037; p ≤ 0.0001; CI 95% = 0.023 to 0.051).

The ability of compound 14 to prevent the extrusion of another P-gp substrate, Rho 123, was further evaluated by flow cytometry. As shown in Figure 9A, exposure of Lucena 1 cells to 14 enhanced intracellular Rho 123 intensity 1.5-fold. This effect was significantly different from that observed with verapamil and with TFP, another known P-gp inhibitor (p < 0.01), but to a lesser extent with the latter. Compound 14 did not induce any increase in Rho 123-associated MFI in K562 (Figure 9B).

To further characterize the P-gp-resistance reversal properties of 14, its influence on basal and verapamil-stimulated ATPase activity was determined. As observed in Figure 10, compound 14 mildly stimulated P-gp-ATPase activity with a maximum of 29.49 ± 4.55% at 66.7 μM, a lower value than that observed with verapamil. It was also observed that 14 antagonized the verapamil-stimulated ATP hydrolysis with an IC₅₀ of 20.89 μM (17.39–23.61).

It is worth mentioning that 14 had no effect on the ATPase activity of control membranes up to 600 μM and that a complete inhibition of verapamil-stimulated ATPase hydrolysis was observed with CsA (data not shown).

The estimated free energy of binding and the corresponding inhibition constants obtained for the (+)- and (−)-pinoresinol (14) are summarized in Table 2, together with the values obtained for the model substrate DOX, compound 2, as the negative control and the mild and powerful known inhibitors verapamil and tariquidar. As observed, there is a small difference in the estimated binding energies (0.14 kcal/mol) between (+)- and (−)-pinoresinol. Both enantiomers went to the same site, with slight differences in the pose. The binding mode of the subject compound 14 is shown on Figure 11, revealing as main contacts S222, A223, K234, F303, Y307, V310, L339, and F343, most of which are residues from TMH 4, 5 and 6. Almost all of these are also involved in binding the reference inhibitor verapamil and part of them also to the powerful inhibitor tariquidar. The K_I obtained for tariquidar in the sub-nM order (Table 2) is close to that experimentally found [58], while the value obtained for the negative control, 2 is almost 100 μM, which is inactive in practice, matching that observed experimentally.

On the other hand, it was observed that the affinity of 14 for the NBD region was substantially lower than that for the region in the TMD.

As also observed in Table 2, compound 16 was found to have the most favorable binding energy among the set of pinoresinol derivatives that were evaluated with the intention of improving the activity of 14. Compound 16 was also found to be 0.45 kcal/mol more favored than its own enantiomer 16a [derivative of the (−)-pinoresinol], showing a still mild but clearer stereospecificity than for the case of 14. This result encouraged the further experimental study of compound 16.

The poses of 14, 16 and tariquidar are superimposed on Figure 12C. Compounds 14 and 16 overlap with the binding region of tariquidar in one of the homolog halves at the top of the inverted “V” formed by the TMHs. This region overlaps with a secondary DOX binding site (0.5 kcal/mol less stable than the main site i.e., the one with the more negative binding energy, reported in Table 2; see also Figures 12A–C). As previously discussed (Jara et al., 2013), the binding of the subject compound seems to be just the first step of a complex process involving long range conformational changes induced in the transporter. In this context, the fact of finding coincidences in the main contacts compared to those observed for known powerful inhibitors would be a promising feature, besides the binding energies themselves.

P-gp expression at the membrane surface was lower in Lucena 1 cells treated with 14 compared to the control group measured at 48 h (Figure 13A).

No suppression of P-gp presence was observed in the outer membrane in K562 cells at the study times (Figure 13B).

The pinoresinol derivative 16, which showed the most promising activity in silico (in terms of binding energy, Table 2), showed an increase in DOX cytotoxicity by a factor of 22.7 at 14 μM and 1.4 at 0.11 μM (Table 1). There was no significant difference in relation to verapamil at 30 and 0.11 μM with 16 applied at 14 and 0.11 μM, respectively (p > 0.05). At 14 μM, no potentiation was observed in DOX activity in K562 cells (p > 0.05). When tested at the maximum non-toxic concentration (28 μM), it caused a 96.8-fold sensitization to DOX in Lucena 1, while a 17.3-fold activity was observed in K562, meaning that the shift in the IC₅₀ of DOX was not only due to P-gp inhibition. For this reason, this concentration was not used in further assays.

In the accumulation assay, 16 significantly increased the DOX-associated MFI in Lucena 1 from 0.87 μM (p < 0.05) with no difference with respect to verapamil (p > 0.05; Figure 14). As also occurred with 14, no dose-dependency and no increase in DOX accumulation were observed in K562 (p > 0.05) (data not shown) at the MEC and up to the maximum concentration evaluated.

To determine whether 14 or 16 were toxic on non-tumoral human cells, PBMC were obtained. These cells were considered the most adequate in relation to the type of tumor cells evaluated in this study. An MTT assay showed that 14 and 16 displayed IC₅₀ values of 216.25 μM (125.93–371.35) and 102.57 μM (48.60–216.45), respectively.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.