Effect of Cholesterol on Cellular Uptake of Cancer Drugs Pirarubicin

Article pubs.acs.org/JPCB

Effect of Cholesterol on Cellular Uptake of Cancer Drugs Pirarubicin and Ellipticine Lei Zhang,†,‡ W. F. Drew Bennett,§ Tao Zheng,† Ping-Kai Ouyang,† Xinping Ouyang,∥ Xueqing Qiu,∥ Anqi Luo,‡ Mikko Karttunen,*,⊥ and P. Chen*,†,‡ †

Biological and Pharmaceutical Engineering, Nanjing Technology University, 30 Puzhu Road South, Nanjing, Jiangsu, China, 211816 Department of Chemical Engineering and Waterloo Institute for Nanotechnology, University of Waterloo, 200 University Avenue West, Waterloo, Ontario, Canada, N2L 3G1 § Department of Chemistry and Biochemistry, University of California, Santa Barbara, California 93106, United States ∥ School of Chemistry and Chemical Engineering, South China University of Technology, Guangzhou, P.R. China, 510640 ⊥ Department of Mathematics and Computer Science & Institute for Complex Molecular Systems, Eindhoven University of Technology, MetaForum, 5600 MB Eindhoven, The Netherlands ‡

ABSTRACT: The cell membrane is a major barrier for drug transport. Given that many cancer drugs must passively cross the cell membrane, understanding drug− membrane interactions is crucial. We used fluorescence-activated cell sorting to investigate how cholesterol influences the transport of the cancer drugs ellipticine and pirarubicin across cell membranes. We showed that cholesterol depletion helped pirarubicin cross the membranes of nonsmall cell lung carcinoma and Chinese hamster ovary cells. In contrast, the uptake of ellipticine was not strongly influenced by cholesterol depletion. To study the microscopic origins of these observations, atomistic molecular dynamics simulations were performed. Doxorubicin (similar in structure to pirarubicin) and ellipticine were simulated in model membranes of POPC and POPC with 40 mol % cholesterol. Atomistic free energy calculations for the translocation of a single ellipticine and doxorubicin across the lipid bilayers qualitatively matched the experiment results. The free energy barrier for doxorubicin crossing the bilayer was strongly increased when cholesterol was present, while for ellipticine the barrier remained similar with and without cholesterol. Molecular dynamics simulations showed that the different hydrogen-bonding propensities of the two drugs are likely the major factor for the different behaviors. The qualitative agreement between cell experiments and atomistic computer simulations illustrates the potential to link observed biological phenomena and single molecule mechanisms of actions. Our results suggest that the traditional understanding of drug ̈ and needs to be re-examined. permeation and the influence of cholesterol on the small molecule transport is naive adhesion,24,25 and elastic moduli (or stiffness).19,24,26 Specifically for drug transport, cholesterol has been shown to have a pronounced effect on passive transport through membranes.27−32 This is largely due to cholesterol’s ability to reduce the free area and volume of a fluid bilayer,20 and possibly through its involvement in domain formation in membranes.33 It has been shown that permeability of the structurally similar cancer drugs pirarubicin, doxorubicin, and daunorubicin into large unilamellar vesicles decreases with an increasing amount of cholesterol.27 Results from MD simulations also support the cholesterol dependent uptake of doxorubicin, showing that the drug significantly perturbs the membrane structure during transport.28 On the experimental side, Pallarés-Trujillo et al. used two sets of cells, the first sensitive to chemotherapy drugs

1. INTRODUCTION Understanding of the physical mechanisms of how small molecules are engulfed and transported by membranes is important for both protection from environmental agents1 and developing drugs that use the membrane composition for their target selection.2,3 In drug design, octanol−water partition coefficients, especially Lipinski’s “rule of five”,4 are often used to estimate permeation and distribution of drugs within the lipid membrane and body.4,5 However, solubility, formulation, pH, and intestinal enzymology all have a strong influence on drug uptake,6 and the partition coefficient approach is not sufficient.7−9 A better understanding of drug and lipid membrane interactions is one of the key issues in drug discovery and development.2,10−12 Cholesterol is an important and abundant lipid in mammalian plasma membranes. It has a strong influence on membranes’ structural and physical properties,13−18 which has been implicated in cell processes including membrane fluidity and transport,19−22 uptake of small molecules,18,20,22,23 © 2016 American Chemical Society

Received: December 16, 2015 Revised: March 2, 2016 Published: March 3, 2016 3148

DOI: 10.1021/acs.jpcb.5b12337 J. Phys. Chem. B 2016, 120, 3148−3156

Article

The Journal of Physical Chemistry B

Figure 1. Structures of doxorubicin (A), pirarubicin (B), and ellipticine (C).

concentration of 1 mg/mL. Aliquots of THP−THF and EPT− THF were then transferred to a 4 mL glass vial and dried under a gentle stream of nitrogen for 10 min. Two stock solutions, EPT and THP, were prepared by adding DMSO 2% (v/v) and HyPure water into the vial. THP and EPT were prepared at 120 and 50 μM for cell treatment, and diluted in F-12K medium to reach the final concentration before treatment. All solutions were adjusted to pH 7.4 using 0.1 M hydrochloric acid and 0.1 M sodium hydroxide. 2.3. Cell Culture. Both A549 and CHO-K1 cells were cultured in F-12K medium supplemented with 10% FBS. All of the cells were incubated at 37 °C in a humidified atmosphere containing 5% CO2. For fluorescence-activated cell sorting (FACS) measurement, cells were cultured in a 24-well cell culture plate. 2.4. Cholesterol Depletion and Drug Treatment Protocols. Cholesterol content in the membranes was reduced by treating them with MβCD, which extracts cholesterol in a concentration and incubation time dependent manner.38 MβCD solutions (5 mM) were prepared in F-12K medium without FBS. For depleting cholesterol, we used MβCD incubation times of 30 min. Next, 4% PFA solution was prepared in PBS. CHO-K1 cells (40 000/well) were plated in a 24-well cell culture plate in F-12K medium with 10% FBS. After 24 h of incubation, the medium was removed and washed three times with PBS. After removing the PBS, 200 μL of the MβCD solution was added to the cells. For the negative and positive controls, we added 200 μL F-12K medium without FBS. MβCD was removed after 30 min, and the cells were washed three times with 300 μL of PBS to remove the residual MβCD. After that, 200 μL of F-12K medium and the drug THP (or EPT) (vol/vol = 3:1) were added. For positive control, 200 μL of F-12K medium and drug complex (vol/vol = 3:1) were also added. For negative control, only 200 μL of F-12K medium was used. The solutions were then incubated for 1 h. After that, the F-12K−drug solution was removed, and the cells were washed three times with PBS. We then added 100 μL of trypsin− ETDA, and the solution was allowed to incubate for 8 min. The cells were then suspended using 400 μL of PFA and collected. Minimal lighting was used during preparation to prevent damage. Solutions were stored in a 4 °C fridge until needed for FACS measurement. Aluminum foil was used for preventing light exposure during storage. Shorter incubation times of 30 min for THP (or EPT) were used for comparisons with the 60 min. Both the nontreated MβCD and THP (or EPT) served as the negative controls. Nontreated MβCD with THP (or EPT) served as the positive control. To investigate the temperature dependence uptake for THP (or EPT), we placed the 24-well plates by only treating THP

and the other resistant, to show that cholesterol depletion increases the uptake of vincristine.31 Recently, Weber et al. found that cholesterol depletion can help doxorubicin to enter Michigan Cancer Foundation-7 breast cancer cells.32 The general conclusion from the above studies is that cholesterol changes the mechanical properties of the cell membrane, and this affects the uptake of small molecules. The mechanical properties of cells are controlled by the microscopic level molecular interactions. Computer simulations are able to provide valuable insight into the molecular level membrane defects created by charged/polar small molecules, electric fields, lipid flip-flop, and mechanical stress; see, e.g., refs 15 and 34−37. In this study, we show how cholesterol affects the uptake of cancer drugs pirarubicin and ellipticine (structures are shown in Figure 1) on lung cancer cell line A549 (nonsmall cell lung carcinoma cells) and noncancerous cell line CHO-K1 (Chinese hamster ovary cells). Our main finding is that, although cholesterol acts as a barrier for pirarubicin uptake, it has a negligible effect on ellipticine uptake. Atomistic molecular dynamics (MD) simulations qualitatively match the experimental findings and provide a plausible explanation for the unique effects of cholesterol on drug transport. This is due to a fundamentally different mechanism of doxorubicin (and likely pirarubicin) transport across bilayers: Without cholesterol, the POPC bilayer is more fluid; that is, doxorubicin maintains its hydration even at the bilayer center. Hydration is essential for transport of doxorubicin. With cholesterol added, permeation is slowed down or blocked, since the energetic cost for deforming the cholesterol containing bilayer to maintain hydration is too high. Ellipticine permeation does not depend on hydration. Instead, due to its hydrophobicity (Figure 1), it prefers to remain desolvated and thus is not influenced by cholesterol concentration.

2. MATERIALS AND METHODS 2.1. Materials. The anticancer agents ellipticine (EPT, 99.0% pure) and pirarubicin (THP, 99.0% pure) were purchased from EMD Biosciences Inc. (La Jolla, CA, USA). Methyl-β-cyclodextrin (MβCD), paraformaldehyde (PFA), dimethyl sulfoxide (DMSO, spectral grade 99.0%), and tetrahydrofuran (THF, reagent grade 99.0%) were obtained form Sigma-Aldrich (Oakville, ON, Canada). CHO-K1 and A549 cells were purchased from American Type Culture Collection (ATCC, Washington, DC, USA). F-12 Kaighn’s modification (F-12K) medium, fetal bovine serum (FBS), trypsin-ETDA, phosphate buffer saline (PBS), and HyPure molecular biology grade water (HyPure water) were purchased from HyClone Laboratories Inc. (Utah, USA). 2.2. Pirarubicin and Ellipticine Solution Preparation. THP and EPT were dissolved in THF to obtain a final 3149


Article


Figure 2. Cellular uptake of pirarubicin into A549 cells after cholesterol depletion at concentrations of 7.5, 15.0, and 30.0 μM after (A) 30 min and (B) 60 min of incubation. Before incubation, the A549 cells were depleted of cholesterol using MβCD for 30 min (M30).

Table 1. Mean Fluorescence Intensities for the Results in Figure 2a

a

uptake (μM)

THP30

M30+THP30

THP60

M30+THP60

7.5 15.0 30.0

55.0 ± 1.4 88.5 ± 5.0 136.5 ± 21.9

115.0 ± 14.1 200.5 ± 13.1 309.0 ± 1.4

72.0 ± 2.8 199.5 ± 20.5 251.5 ± 10.6

448.0 ± 29.7 773.5 ± 120.9 921.5 ± 34.7

Results are given as mean values ± SD, n = 3.

(or EPT) without cholesterol depletion into the 37 °C incubator or 4 °C fridge, respectively. 2.5. Fluorescence-Activated Cell Sorting Measurement. Cellular uptake of THP and EPT was studied using BD FACS Calibur Flow Cytometry (BD 208 Biosciences, Mississauga, Canada). For THP, the emission and excitation wavelengths are 560 and 480 nm, respectively,39 and they are 550 and 300 nm for neutral EPT.40 2.6. Molecular Dynamics Simulations. Atomistic MD simulations were conducted for EPT and doxorubicin (DOX) in pure POPC and cholesterol:POPC mixtures of 0 and 40 mol % using the GROMACS simulation suite version 4.41 The GROMOS force field was used for the POPC,42 cholesterol,43 EPT,44 DOX,28 and the SPC water model.45 The availability of the force field parameters motivated our choice of simulating DOX instead of THP, but given the similarity in chemical structure (Figure 1) and results reported in the literature,27,28,32 we do not expect a large difference in their mechanism of membrane permeation; this choice is justified by the fact that we are interested in the qualitative behavior of THP which has been experimentally confirmed to be similar for DOX and THP. Quantitative details are different, but to study them, a full force field parametrization and its verification against experiments should be performed for THP and that is beyond the scope this study. We would also like to point out that using structurally similar molecules as reporters (typically due to fluorescence) is a commonly used experimental method. For example, fluorescent dehydroergosterol is often used as a cholesterol reporter for membrane studies46 (adding fluorescent groups can, however, lead to subtleties47). We used a 2 fs time step, neighbor list updates every 10 steps, and LINCS for bond length constraints.48 Berendsen semi-isotropic pressure coupling was used with a coupling constant of 2.5 ps and compressibility of 4.5 × 10−5 bar−1.49 A temperature of 310 K was maintained with the v-rescale thermostat with a 0.1 ps time constant.50 The particle mesh Ewald method with a real space cutoff of 1.0 nm was used for electrostatics51 for accurate

inclusion of long-range effects.52 Lennard-Jones interactions were shifted between 0.9 and 1.0 nm, and the long-range dispersion correction was used. The free energy for transferring a single DOX or EPT molecule from water to the center of the bilayer was calculated using umbrella sampling.53 The drug molecule was inserted at each umbrella position by slowly turning on the molecule Lennard-Jones and electrostatic interactions with the rest of the system during a 500 ps simulation. A harmonic position restraint of 3000 kJ/mol nm2 was applied between the center of mass of the drug molecule and the center of mass of the bilayer in the directions normal to the plane of the membrane. Along the z-axis, 41 independent simulations were performed with the position restraint varied from z = 0 to z = 4.0 nm with a 0.1 nm increment between adjacent simulations. Each simulation was run for 300 ns for DOX and 400 ns for EPT. The weighted histogram analysis method was used for computing the potential of mean force, and error bars were estimated using the bootstrap method with 200 bootstraps.53,54 Hydrogen bonds were calculated with a geometric criterion, with a 0.35 nm distance cutoff between the acceptor and donor and a 30° angle between the hydrogen, donor, and acceptor.

3. RESULTS 3.1. Uptake of Pirarubicin and Ellipticine after Cholesterol Depletion of Cell Membrane. Figure 2A shows the uptake of THP into A549 cells at concentrations of 7.5, 15.0, and 30.0 μM. Cholesterol was first depleted for 30 min using MβCD, and then, the cells were incubated for another 30 min. Increasing the incubation time to 60 min (Figure 2B) showed the same trend for cholesterol dependent uptake. The mean intensity of fluorescence (MIF) values at the three concentrations and two incubation times are shown in Table 1. For the 30 μM THP (60 min incubation), the intensity increases by a factor of ∼3.7 when cholesterol was depleted compared with the positive control without MβCD treatment. 3150


Article


Figure 3. Cellular uptake of ellipticine into A549 cells after cholesterol depletion at concentrations of 3.1, 9.4, and 12.5 μM after (A) 30 min and (B) 60 min of incubation. Before incubation, the A549 cells were depleted of cholesterol using MβCD for 30 min (M30).

respectively (Figure 5C). Similar temperature independent uptake was found for EPT (Figure 5B and D). 3.3. Molecular Dynamics Simulations. Free energy profiles for the passive transport of EPT and DOX across the model lipid bilayer are shown in Figure 6. Both EPT (ca. 30 kJ/ mol) and DOX (ca. 60 kJ/mol) have free energy troughs near the water/bilayer interface. This is expected given that both molecules are amphipathic. The position of the trough is shifted outward when cholesterol is in the bilayer because cholesterol increases the membrane thickness.15,20 The depth of the free energy minima for both EPT and DOX is similar for the bilayers with and without cholesterol. EPT and DOX have free energy barriers at the center of the lipid bilayer. For EPT, the free energy barrier is roughly the same for the POPC and the 40 mol % cholesterol bilayer. The barrier for DOX is ca. 20 kJ/ mol higher in the cholesterol bilayer compared to the POPC bilayer. This larger barrier for the 40 mol % cholesterol bilayer would substantially reduce the permeability of DOX. The free energy differences we observed for DOX but not for EPT are due to different physical mechanisms of transfer, originating from the molecules’ different chemical structures. DOX is much more polar than EPT and can form, on average, over 15 hydrogen bonds in water, compared to ca. 2.5 for EPT (we only have simulated the neutral, deprotonated state of EPT) (Figure 7). Moving the drugs into the hydrophobic bilayer core means that these hydrogen bonds are lost, which costs energysignificantly more in the case of DOX. There is an energetic balance between the water−lipid interface deforming to maintain water−drug and lipid headgroup−drug hydrogen bonds and other polar interactions. For EPT, there are fewer hydrogen bonds to lose, so the molecule is completely desolvated near the center of both bilayers (Figure 8A and B). Even at the very center of the pure POPC bilayer, DOX maintains nearly five hydrogen bonds with water and lipids (Figure 7), which induces a large water defect in the POPC bilayer, with POPC head groups penetrating into the center of the bilayer (Figure 8C). For the 40 mol % cholesterol bilayer, the cost of deforming the bilayer is higher due to the rigidifying effect of cholesterol, so DOX loses all hydrogen bonds and is desolvated near the bilayer center (Figure 8D).

The same protocol as above was used to study the uptake of EPT into A549 cells. The effect of cholesterol on EPT uptake was found to be different from THP. Figure 3 shows EPT uptake at concentrations of 3.1, 9.4, and 12.5 μM EPT for (A) 30 min and (B) 60 min incubation time after cholesterol depletion by MβCD for 30 min. After cholesterol depletion, the peak did not shift; i.e., there was no significant change in uptake. In Table 2, the MIF value shows the same trend as the Table 2. Mean Fluorescence Intensities for the Results in Figure 3a uptake (μM)

EPT30

M30+EPT30

EPT60

M30+EPT60

3.1 9.4 12.5

14.4 ± 0.6 35.9 ± 2.9 63.8 ± 5.6

9.2 ± 2.8 31.1 ± 0.6 61.2 ± 4.7

29.5 ± 1.7 53.9 ± 2.8 84.0 ± 6.1

28.1 ± 1.5 45.7 ± 4.9 81.2 ± 2.8

a

Results are given as mean values ± SD, n = 3.

peaks. For 12.5 μM (60 min incubation time), the MFI has no significant change after cholesterol depletion for 30 min. Contrary to the effect cholesterol has on THP uptake, EPT uptake did not change after cholesterol depletion. We also measured the effect of cholesterol on THP and EPT uptake into CHO-K1 cells. Figure 4 shows that cholesterol inhibits THP from crossing the cell membrane but has no effect on EPT. These results are consistent with the uptake of THP and EPT into A549 cells. 3.2. Temperature Independent and Concentration Dependent Uptake of Pirarubicin and Ellipticine. FACS was used to measure the concentration dependence of THP and EPT uptake on A549 cells. As shown in Table 1, the MIF increases with increasing THP concentration (7.5, 15.0, and 30.0 μM were used). When increasing the incubation time from 30 to 60 min for the 30.0 μM THP, uptake increases with the MIF increasing ∼1.8-fold. A similar trend was found on THP at 7.5 μM (1.3-fold) and 15.0 μM (2.3-fold). For EPT, as shown in Figure 3 and Table 2, concentration and incubation time dependent uptake was found at concentrations of 3.1, 9.4, and 12.5 μM for 30 and 60 min incubation times, respectively. Uptake of both EPT and THP into A549 cells was found to be temperature independent. As shown in Figure 5, the peaks of 15 μM THP do not shift (Figure 5A) at 4 and 37 °C with the MIF values being 198.1 ± 3.4 and 193.6 ± 11.8,

4. DISCUSSION Adsorption and distribution of drug molecules are central to their effectiveness. Cellular membranes are a problematic barrier for passive transport due to their unique chemistry: an 3151


Article


Figure 4. Cellular uptake of (A) pirarubicin into CHO-K1 cells after cholesterol depletion at concentrations of 7.5 μM after 60 min of incubation. Before incubation, the CHO-K1 cells were depleted of cholesterol using MβCD for 30 min (M30). Cellular uptake of (B) ellipticine into CHO-K1 cells after cholesterol depletion at concentrations of 9.4 μM after 60 min of incubation. Before incubation, the CHO-K1 cells were depleted of cholesterol using MβCD for 30 min (M30). Parts C and D show the mean fluorescence intensities for the results in parts A and B. Results are given as mean values ± SD, n = 3.

∼5 nm, two molecules thick, hydrophobic, liquid-phase sheet. Understanding drug partitioning, including the effects from lipid diversity and asymmetry between the bilayer leaflets, and in particular cholesterol, is important in order to be able to reduce drug candidate attrition. Given that many cancer cells have different lipid compositions compared to healthy cells,55 with such knowledge, one might be able to tailor future drugs for targeted uptake. We show that cancer cell uptake of THP is cholesterol dependent, but the uptake of EPT is not. With MD simulations, we provide a molecular level explanation for the physical mechanism likely responsible for the differences. Our results also suggest that the currently used octanol−water partition coefficients4,5 and Lipinski’s “rule of five”4 may be insufficient for screening potential drugs. The uptake of drugs can be directly tuned by changing their chemistry in a lipid dependent manner. For example, one can speculate that, if we wanted to impart EPT with cholesterol sensitive permeation, our results would suggest adding some hydrogen bonding groups to the rings. Whether this would disrupt EPT’s cellular function would, of course, need to be addressed by medicinal chemists. The potential to control cellular uptake based on drug structure is extremely important but will require extensive research of fundamental microscopic mechanisms in drug− membrane interactions. Previous work has shown that cholesterol has an influence on the permeability of small molecules.18 The permeability coefficients of both pirarubicin and doxorubicin into large

unilamellar vesicles have previously been observed to decrease with increasing amounts of cholesterol.27 Our findings on living cells A549 and CHO-K1 show similar behavior. In contrast, cholesterol did not inhibit the uptake of the more hydrophobic drug EPT (Figure 1). When we depleted cholesterol with MβCD treatment for 30 min, no significant change in the uptake of EPT into A549 cells was observed (Figure 3). The reason that EPT has similar uptake with and without cholesterol is that in both cases EPT loses contact with water and its permeation is well described by the inhomogeneous solubility−diffusion model in both bilayers.56 The simplest model to describe small molecule passive membrane transport is the solubility−diffusion model, where the permeability can be calculated from the molecule’s solubility in water and the hydrophobic lipid environment, and its diffusion coefficient in both environments. The inhomogeneous solubility−diffusion model is a variant that was developed on the basis of atomistic MD. It shows that the solute’s free energy and diffusion coefficient strongly depend on its position in the membrane.56 Our free energy profiles illustrate this concept: For example, EPT’s deep free energy minima at the membrane interface would affect its permeability, and this is accounted for if we assume inhomogeneous solubility−diffusion permeation but not if we use the simpler solubility−diffusion model. The mechanism for the passive transport of DOX, and likely similar drugs like THP, changes with cholesterol content. For the POPC bilayer, the passive transport of DOX does not proceed 3152


Article


Figure 5. Cellular uptake of (A) 15.0 μM pirarubicin and (B) 3.1 μM ellipticine into A549 cells after 60 min of incubation at 4 or 37 °C. Parts C and D show the mean fluorescence intensities for the results in parts A and B. Results are given as mean values ± SD, n = 3.

Figure 6. Free energy profiles for transferring a single drug molecule from water (∼3.5 nm) to the center of the lipid bilayer. The reaction coordinate is the distance between the center of mass of the lipid bilayer and the center of mass of the drug in the direction normal to the plane of the bilayer.

Figure 7. Average number of hydrogen bonds formed between the drug and the water and lipids, as the drug moves across the membrane. POPC is the pure POPC bilayer, and the CHOL bilayer is 40 mol % cholesterol with POPC. Hydrogen bonds are calculated with a geometric criterion, with a 0.35 nm distance cutoff between the acceptor and donor and a 30° angle between the hydrogen, donor, and acceptor.

via the inhomogeneous solubility−diffusion mechanism. This is because the permeation depends on the POPC bilayer deforming to maintain DOX hydration. The energetic barrier and permeation rate depend on the collective properties of the lipid bilayer, and this is not captured by a one-dimensional reaction coordinate. For DOX permeation in POPC, a water defect would have to form on the opposite leaflet first. A pore across the membrane may be the transition state. Cholesterol content increases the energy cost for membrane bending, so a DOX molecule at the bilayer center does not cause defect formation. DOX’s transport across a bilayer with high

cholesterol content is well described by the inhomogeneous solubility−diffusion model. These critical molecular details are needed to properly characterize drug permeation. It is also important to note that both drugs have relatively deep free energy troughs at the bilayer interface. Reducing the affinity of a drug for the membrane interface could make it more effective and less prone to transport out of cells by multidrug resistant proteins. 3153


Article


system investigations62 but are computationally very costly, and hence, we chose to simulate only the neutral form of EPT, as experiments indicate that it deprotonates upon crossing the membrane.63 We would also like to point out that it is also possible that the drugs might interact with each other during membrane permeation, meaning that collective effects may become important. In that case, more elaborate reaction coordinates are likely needed.

5. CONCLUSIONS Cholesterol inhibits the uptake of THP but does not influence the uptake of EPT. The intensity of uptake for THP is increased up to about a factor of 4 by incubating MβCD for 30 min to deplete cholesterol. Contrary to THP, the uptake of EPT is not affected by cholesterol depletion. We also confirmed that both EPT and THP uptake are temperature independent (in the range of studied temperatures) but concentration dependent. Atomistic MD simulations and free energy calculations for the passive transport of a single drug molecule across POPC bilayers with and without cholesterol provided mechanistic insight into the influence of cholesterol on membrane drug transport. The free energy barrier is lower for DOX in the POPC bilayer compared to the bilayer with cholesterol because the mechanism changes from DOX remaining hydrated throughout the POPC bilayer to becoming dehydrated at the center of the cholesterol bilayer. For EPT, the free energy variation is similar for the bilayer with and without cholesterol because in both bilayers EPT is dehydrated at the bilayer center. These differences are due to the increased hydrogen bonding propensity for DOX compared to EPT.

Figure 8. Snapshots of the EPT and DOX molecules at the center (corresponds to 0 nm in Figure 6) of the POPC and 40 mol % cholesterol bilayer: (A) EPT-POPC; (B) EPT-40% CHOL; (C) DOXPOPC; (D) DOX-40% CHOL. Water is shown with red and white, the POPC tails are gray lines, phosphorus red balls, nitrogen yellow balls, cholesterol thick gray lines, and EPT or DOX are colored by atom type: green, carbon; blue, nitrogen; oxygen, red.

■

There are a number of technical issues for both simulations and experiments that makes direct quantitative comparison not possible in this study. Most importantly, the FACS experiments are on real cells, while our simulations are simple model phospholipid bilayers. Control experiments show that the mechanism of uptake for both types of drugs is passive diffusion, because uptake is concentration dependent, and temperature independent. There are, however, other phenomena that may affect the drug distribution obtained from FACS. These include the effects of multidrug transporters and nonspecific binding to other proteins and biomolecules.9 It should also be noted that MβCD can lead to redistribution of cholesterol between raft-like and non-raft-like regions. It also interacts with other phospholipids, which can then influence its distribution and cholesterol binding affinity.38 Using FACS experiments along with other biophysical characterization of permeability27,57 would allow an even more direct comparison between simulations and experiments. There are a number of challenges for computer simulations to calculate free energies and rates in complex systems. Due to the slow diffusion of lipids, it is likely that the error in our free energy calculations may be somewhat underestimated.58,59 Longer simulations, kinetic and diffusion analysis, multidimensional free energy calculations, and multicomponent lipid mixtures are possible approaches for more quantitative predictions of drug permeation in the future. We have only simulated the deprotonated and neutral EPT because classical MD simulations cannot model chemical bond changes. Constant pH simulations60,61 for EPT transport would be desirable, particularly with respect to pH dependent permeation. Constant pH methods are starting to be used for lipid

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *Phone: +1 519 888 4567, ext. 35586. E-mail: p4chen@ uwaterloo.ca. Author Contributions

The manuscript was written through contributions of all authors. L.Z., W.F.D.B., M.K., and P.C. planned the study and designed the analyses; L.Z. and A.L. performed FACS experiments; W.F.D.B. performed molecular simulation experiments; L.Z., W.F.D.B., T.Z., P.-K.O., X.O., and X.Q. analyzed data. L.Z., W.F.D.B., M.K., and P.C. wrote the manuscript. All authors have given approval to the final version of the manuscript. L.Z. and W.F.D.B. contributed equally. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Natural Sciences and Engineering Research Council (NSERC) of Canada and NSERC’s Sir Frederick Banting Fellowship Program for funding (WFDB) and the Program of Scientific Innovation Research of College Graduates in Jiangsu Province (No.CXZZ13_0455) for funding. Compute Canada and SharcNet provided the computational resources.

■

REFERENCES

(1) Nisoh, N.; Karttunen, M.; Monticelli, L.; Wong-Ekkabut, J. Lipid Monolayer Disruption Caused by Aggregated Carbon Nanoparticles. RSC Adv. 2015, 5, 11676−11685.

3154


Article

The Journal of Physical Chemistry B (2) Giacomini, K. M.; Huang, S.-M.; Tweedie, D. J.; Benet, L. Z.; Brouwer, K. L. R.; Chu, X.; Dahlin, A.; Evers, R.; Fischer, V.; Hillgren, K. M. Membrane Transporters in Drug Development. Nat. Rev. Drug Discovery 2010, 9, 215−236. (3) Yıldırım, M. A.; Goh, K.-I.; Cusick, M. E.; Barabási, A.-L.; Vidal, M. DrugTarget Network. Nat. Biotechnol. 2007, 25, 1119−1126. (4) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Experimental and Computational Approaches to Estimate Solubility and Permeability in Drug Discovery and Development Settings. Adv. Drug Delivery Rev. 2012, 64, 4−17. (5) Leo, A.; Hansch, C.; Elkins, D. Partition Coefficients and Their Uses. Chem. Rev. 1971, 71, 525−616. (6) Burton, P. S.; Goodwin, J. T.; Vidmar, T. J.; Amore, B. M. Predicting Drug Absorption: How Nature Made It a Difficult Problem. J. Pharmacol. Exp. Ther. 2002, 303, 889−895. (7) Dobson, P. D.; Kell, D. B. Carrier-Mediated Cellular Uptake of Pharmaceutical Drugs: An Exception or the Rule? Nat. Rev. Drug Discovery 2008, 7, 205−220. (8) Sugano, K.; Kansy, M.; Artursson, P.; Avdeef, A.; Bendels, S.; Di, L.; Ecker, G. F.; Faller, B.; Fischer, H.; Gerebtzoff, G. Coexistence of Passive and Carrier-Mediated Processes in Drug Transport. Nat. Rev. Drug Discovery 2010, 9, 597−614. (9) Kell, D. B.; Oliver, S. G. How Drugs Get into Cells: Tested and Testable Predictions to Help Discriminate between TransporterMediated Uptake and Lipoidal Bilayer Diffusion. Front. Pharmacol. 2014, 5, 10.3389/fphar.2014.00231. (10) Minchinton, A. I.; Tannock, I. F. Drug Penetration in Solid Tumours. Nat. Rev. Cancer 2006, 6, 583−592. (11) Mu, Q.; Jiang, G.; Chen, L.; Zhou, H.; Fourches, D.; Tropsha, A.; Yan, B. Chemical Basis of Interactions between Engineered Nanoparticles and Biological Systems. Chem. Rev. 2014, 114, 7740− 7781. (12) Peetla, C.; Stine, A.; Labhasetwar, V. Biophysical Interactions with Model Lipid Membranes: Applications in Drug Discovery and Drug Delivery. Mol. Pharmaceutics 2009, 6, 1264−1276. (13) Qi, M.; Liu, Y.; Freeman, M. R.; Solomon, K. R. CholesterolRegulated Stress Fiber Formation. J. Cell. Biochem. 2009, 106, 1031− 1040. (14) Simons, K.; Ikonen, E. How Cells Handle Cholesterol. Science 2000, 290, 1721−6. (15) Róg, T.; Pasenkiewicz-Gierula, M.; Vattulainen, I.; Karttunen, M. Ordering Effects of Cholesterol and Its Analogues. Biochim. Biophys. Acta, Biomembr. 2009, 1788, 97−121. (16) Róg, T.; Vattulainen, I. Cholesterol, Sphingolipids, and Glycolipids: What Do We Know About Their Role in Raft-Like Membranes? Chem. Phys. Lipids 2014, 184, 82−104. (17) de Meyer, F.; Smit, B. Effect of Cholesterol on the Structure of a Phospholipid Bilayer. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 3654− 3658. (18) Sun, D.; Lin, X.; Gu, N. Cholesterol Affects C 60 Translocation across Lipid Bilayers. Soft Matter 2014, 10, 2160−2168. (19) Needham, D.; Nunn, R. S. Elastic Deformation and Failure of Lipid Bilayer Membranes Containing Cholesterol. Biophys. J. 1990, 58, 997−1009. (20) Falck, E.; Patra, M.; Karttunen, M.; Hyvö nen, M. T.; Vattulainen, I. Lessons of Slicing Membranes: Interplay of Packing, Free Area, and Lateral Diffusion in Phospholipid/Cholesterol Bilayers. Biophys. J. 2004, 87, 1076−1091. (21) Jeon, J.-H.; Monne, H. M.-S.; Javanainen, M.; Metzler, R. Anomalous Diffusion of Phospholipids and Cholesterols in a Lipid Bilayer and Its Origins. Phys. Rev. Lett. 2012, 109, 188103. (22) Kaiser, H.-J.; Orłowski, A.; Róg, T.; Nyholm, T. K. M.; Chai, W.; Feizi, T.; Lingwood, D.; Vattulainen, I.; Simons, K. Lateral Sorting in Model Membranes by Cholesterol-Mediated Hydrophobic Matching. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 16628−16633. (23) Saito, H.; Shinoda, W. Cholesterol Effect on Water Permeability through Dppc and Psm Lipid Bilayers: A Molecular Dynamics Study. J. Phys. Chem. B 2011, 115, 15241−15250.

(24) Fu, C.; He, J.; Li, C.; Shyy, J. Y. J.; Zhu, Y. Cholesterol Increases Adhesion of Monocytes to Endothelium by Moving Adhesion Molecules out of Caveolae. Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 2010, 1801, 702−710. (25) Norman, L. L.; Oetama, R. J.; Dembo, M.; Byfield, F.; Hammer, D. A.; Levitan, I.; Aranda-Espinoza, H. Modification of Cellular Cholesterol Content Affects Traction Force, Adhesion and Cell Spreading. Cell. Mol. Bioeng. 2010, 3, 151−162. (26) Yang, Y.-T.; Liao, J.-D.; Lin, C.-C. K.; Chang, C.-T.; Wang, S.H.; Ju, M.-S. Characterization of Cholesterol-Depleted or -Restored Cell Membranes by Depth-Sensing Nano-Indentation. Soft Matter 2012, 8, 682−687. (27) Frézard, F.; Garnier-Suillerot, A. Permeability of Lipid Bilayer to Anthracycline Derivatives. Role of the Bilayer Composition and of the Temperature. Biochim. Biophys. Acta, Lipids Lipid Metab. 1998, 1389, 13−22. (28) Yacoub, T. J.; Reddy, A. S.; Szleifer, I. Structural Effects and Translocation of Doxorubicin in a Dppc/Chol Bilayer: The Role of Cholesterol. Biophys. J. 2011, 101, 378−385. (29) Khajeh, A.; Modarress, H. The Influence of Cholesterol on Interactions and Dynamics of Ibuprofen in a Lipid Bilayer. Biochim. Biophys. Acta, Biomembr. 2014, 1838, 2431−2438. (30) Khajeh, A.; Modarress, H. Effect of Cholesterol on Behavior of 5-Fluorouracil (5-Fu) in a Dmpc Lipid Bilayer, a Molecular Dynamics Study. Biophys. Chem. 2014, 187, 43−50. (31) Pallarés-Trujillo, J.; Domenech, C.; Grau-Oliete, M. R.; RiveraFillat, M. P. Role of Cell Cholesterol in Modulating Vincristine Uptake and Resistance. Int. J. Cancer 1993, 55, 667−671. (32) Weber, P.; Wagner, M.; Schneckenburger, H. Cholesterol Dependent Uptake and Interaction of Doxorubicin in Mcf-7 Breast Cancer Cells. Int. J. Mol. Sci. 2013, 14, 8358−8366. (33) Simons, K.; Ikonen, E. Functional Rafts in Cell Membranes. Nature 1997, 387, 569−572. (34) Shigematsu, T.; Koshiyama, K.; Wada, S. Effects of Stretching Speed on Mechanical Rupture of Phospholipid/Cholesterol Bilayers: Molecular Dynamics Simulation. Sci. Rep. 2015, 5, 15369. (35) Bennett, W. F. D.; Tieleman, D. P. The Importance of Membrane DefectsLessons from Simulations. Acc. Chem. Res. 2014, 47, 2244−2251. (36) Gurtovenko, A. A.; Anwar, J.; Vattulainen, I. Defect-Mediated Trafficking across Cell Membranes: Insights from in Silico Modeling. Chem. Rev. 2010, 110, 6077−6103. (37) Rog, T.; Stimson, L. M.; Pasenkiewicz-Gierula, M.; Vattulainen, I.; Karttunen, M. Replacing the Cholesterol Hydroxyl Group with the Ketone Group Facilitates Sterol Flip-Flop and Promotes Membrane Fluidity. J. Phys. Chem. B 2008, 112, 1946−1952. (38) Zidovetzki, R.; Levitan, I. Use of Cyclodextrins to Manipulate Plasma Membrane Cholesterol Content: Evidence, Misconceptions and Control Strategies. Biochim. Biophys. Acta, Biomembr. 2007, 1768, 1311−1324. (39) Sureau, F.; Moreau, F.; Millot, J.-M.; Manfait, M.; Allard, B.; Aubard, J.; Schwaller, M.-A. Microspectrofluorometry of the Protonation State of Ellipticine, an Antitumor Alkaloid, in Single Cells. Biophys. J. 1993, 65, 1767. (40) Sadatmousavi, P.; Chen, P. Self/Co-Assembling Peptide, Ear8-Ii, as a Potential Carrier for a Hydrophobic Anticancer Drug Pirarubicin (Thp)Characterization and in-Vitro Delivery. Int. J. Mol. Sci. 2013, 14, 23315−23329. (41) Hess, B.; Kutzner, C.; Van Der Spoel, D.; Lindahl, E. Gromacs 4: Algorithms for Highly Efficient, Load-Balanced, and Scalable Molecular Simulation. J. Chem. Theory Comput. 2008, 4, 435−447. (42) Poger, D.; Van Gunsteren, W. F.; Mark, A. E. A New Force Field for Simulating Phosphatidylcholine Bilayers. J. Comput. Chem. 2010, 31, 1117−1125. (43) Höltje, M.; Förster, T.; Brandt, B.; Engels, T.; von Rybinski, W.; Höltje, H.-D. Molecular Dynamics Simulations of Stratum Corneum Lipid Models: Fatty Acids and Cholesterol. Biochim. Biophys. Acta, Biomembr. 2001, 1511, 156−167. 3155


Article

The Journal of Physical Chemistry B (44) Malde, A. K.; Zuo, L.; Breeze, M.; Stroet, M.; Poger, D.; Nair, P. C.; Oostenbrink, C.; Mark, A. E. An Automated Force Field Topology Builder (Atb) and Repository: Version 1.0. J. Chem. Theory Comput. 2011, 7, 4026−4037. (45) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; Hermans, J. Interaction Models for Water in Relation to Protein Hydration. Intermolecular Forces; Pullman, B., Ed.; Springer: Dordrecht, The Netherlands, 1981; pp 331−342. (46) Pourmousa, M.; Rog, T.; Mikkeli, R.; Vattulainen, l.; Solanko, L. M.; Wustner, D.; List, N. H.; Kongsted, J.; Karttunen, M. Dehydroergosterol as an Analogue for Cholesterol: Why It Mimics Cholesterol So Wellor Does It? J. Phys. Chem. B 2014, 118, 7345− 7357. (47) Repáková, J.; Holopainen, J. M.; Karttunen, M.; Vattulainen, I. Influence of Pyrene-Labeling on Fluid Lipid Membranes. J. Phys. Chem. B 2006, 110, 15403−15410. (48) Hess, B. P-Lincs: A Parallel Linear Constraint Solver for Molecular Simulation. J. Chem. Theory Comput. 2008, 4, 116−122. (49) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; DiNola, A.; Haak, J. R. Molecular Dynamics with Coupling to an External Bath. J. Chem. Phys. 1984, 81, 3684−3690. (50) Bussi, G.; Donadio, D.; Parrinello, M. Canonical Sampling through Velocity Rescaling. J. Chem. Phys. 2007, 126, 014101. (51) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. A Smooth Particle Mesh Ewald Method. J. Chem. Phys. 1995, 103, 8577−8593. (52) Cisneros, G. A. s.; Karttunen, M.; Ren, P.; Sagui, C. Classical Electrostatics for Biomolecular Simulations. Chem. Rev. 2014, 114, 779−814. (53) Kumar, S.; Rosenberg, J. M.; Bouzida, D.; Swendsen, R. H.; Kollman, P. A. The Weighted Histogram Analysis Method for FreeEnergy Calculations on Biomolecules. I. The Method. J. Comput. Chem. 1992, 13, 1011−1021. (54) Hub, J. S.; De Groot, B. L.; Van Der Spoel, D. G_Whama Free Weighted Histogram Analysis Implementation Including Robust Error and Autocorrelation Estimates. J. Chem. Theory Comput. 2010, 6, 3713−3720. (55) Michel, V.; Bakovic, M. Lipid Rafts in Health and Disease. Biol. Cell 2007, 99, 129−140. (56) Marrink, S.-J.; Berendsen, H. J. Simulation of Water Transport through a Lipid Membrane. J. Phys. Chem. 1994, 98, 4155−4168. (57) Girard, P.; Pécréaux, J.; Lenoir, G.; Falson, P.; Rigaud, J.-L.; Bassereau, P. A New Method for the Reconstitution of Membrane Proteins into Giant Unilamellar Vesicles. Biophys. J. 2004, 87, 419− 429. (58) Neale, C.; Bennett, W. F. D.; Tieleman, D. P.; Pomès, R. Statistical Convergence of Equilibrium Properties in Simulations of Molecular Solutes Embedded in Lipid Bilayers. J. Chem. Theory Comput. 2011, 7, 4175−4188. (59) Filipe, H. A.; Moreno, M. J. o.; Róg, T.; Vattulainen, I.; Loura, L. M. How to Tackle the Issues in Free Energy Simulations of Long Amphiphiles Interacting with Lipid Membranes: Convergence and Local Membrane Deformations. J. Phys. Chem. B 2014, 118, 3572− 3581. (60) Chen, W.; Shen, J. K. Effects of System Net Charge and Electrostatic Truncation on All-Atom Constant pH Molecular Dynamics. J. Comput. Chem. 2014, 35, 1986−1996. (61) Donnini, S.; Tegeler, F.; Groenhof, G.; Grubmüller, H. Constant pH Molecular Dynamics in Explicit Solvent with λDynamics. J. Chem. Theory Comput. 2011, 7, 1962−1978. (62) Bennett, W. D.; Chen, A. W.; Donnini, S.; Groenhof, G.; Tieleman, D. P. Constant pH Simulations with the Coarse-Grained Martini ModelApplication to Oleic Acid Aggregates. Can. J. Chem. 2013, 91, 839−846. (63) MacCallum, J. L.; Bennett, W. D.; Tieleman, D. P. Distribution of Amino Acids in a Lipid Bilayer from Computer Simulations. Biophys. J. 2008, 94, 3393−3404.

3156


Effect of Cholesterol on Cellular Uptake of Cancer Drugs Pirarubicin

Recommend Documents