Effect of fatty acids on the permeability barrier of model and biological membranes

Accepted Manuscript Title: Effect of fatty acids on the permeability barrier of model and biological membranes Author: Ahmad Arouri Kira E. Lauritsen Henriette L. Nielsen Ole G. Mouritsen PII: DOI: Reference:

S0009-3084(16)30123-2 http://dx.doi.org/doi:10.1016/j.chemphyslip.2016.10.001 CPL 4494

To appear in:

Chemistry and Physics of Lipids

Received date: Revised date: Accepted date:

18-8-2016 26-9-2016 4-10-2016

Please cite this article as: Arouri, Ahmad, Lauritsen, Kira E., Nielsen, Henriette L., Mouritsen, Ole G., Effect of fatty acids on the permeability barrier of model and biological membranes.Chemistry and Physics of Lipids http://dx.doi.org/10.1016/j.chemphyslip.2016.10.001 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Effect of fatty acids on the permeability barrier of model and biological membranes Ahmad Arouri a,b, Kira E. Lauritsen a,§, Henriette L. Nielsen a,§, Ole G. Mouritsen a,b* a

MEMPHYS-Center for Biomembrane Physics, Department of Physics, Chemistry, and Pharmacy, University of Southern Denmark, Odense, Denmark. b The Lundbeck Foundation Nanomedicine Research Center for Cancer Stem Cell Targeting Therapeutics (NanoCAN), University of Southern Denmark, Odense, Denmark. §

These authors contributed equally.

Address for Correspondence: MEMPHYS-Center for Biomembrane Physics, Department of Physics, Chemistry, and Pharmacy, University of Southern Denmark, Campusvej 55, DK-5230 Odense, Denmark. Tel.: +45 6550 3528. Fax: +45 6550 4048. E-mail addresses: [email protected]; [email protected] (Ahmad Arouri), [email protected] (Ole G. Mouritsen). URL: http://www.memphys.dk, http://www.nanocan.org.

1

Graphical abstract

Highlights  In general, membrane perturbing effect of saturated fatty acids increases with acyl chain length.  C12, C14 and C16 saturated fatty acids possess the highest cytotoxicity on MCF-7 cell line.  Saturated fatty acids in sub-cytotoxic concentrations cannot reduce the permeability barrier of cell membranes.  The membrane perturbing effect of fatty acids on model membranes cannot simply be carried over to biological membranes of live cells.

2

Abstract Because of the amphipathicity and conical molecular shape of fatty acids, they can efficiently incorporate into lipid membranes and disturb membrane integrity, chain packing, and lateral pressure profile. These phenomena affect both model membranes as well as biological membranes. We investigated the feasibility of exploiting fatty acids as permeability enhancers in drug delivery systems for enhancing drug release from liposomal carriers and drug uptake by target cells. Saturated fatty acids, with acyl chain length from C8 to C20, were tested using model drug delivery liposomes of 1,2- dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and the breast cancer MCF-7 cell line as a model cell. A calcein release assay demonstrated reduction in the membrane permeability barrier of the DPPC liposomes, proportionally to the length of the fatty acid. Differential scanning calorimetry (DSC) and dynamic light scattering (DLS) experiments revealed that C12 to C20 fatty acids can stabilize DPPC liposomal bilayers and induce the formation of large structures, probably due to liposome aggregation and bilayer morphological changes. On the other hand, the short fatty acids C8 and C10 tend to destabilize the bilayers and only moderately cause the formation of large structures. The effect of fatty acids on DPPC liposomes was not completely transferrable to the MCF-7 cell line. Using cytotoxicity assays, the cells were found to be relatively insensitive to the fatty acids at apoptotic sub-millimolar concentrations. Increasing the fatty acid concentration to few millimolar substantially reduced the viability of the cells, most likely via the induction of necrosis and cell lysis. A bioluminescence living-cell-based luciferase assay showed that saturated fatty acids in sub-cytotoxic concentrations cannot reduce the permeability barrier of cell membranes. Our results confirm that the membrane perturbing effect of fatty acids on

3

model membranes cannot simply be carried over to biological membranes of live cells.

Abbreviations DPPC: 1,2- Dipalmitoyl-sn-glycero-3-phosphocholine DSC: Differential scanning calorimetry DLS: Dynamic light scattering CMC: Critical micelle concentration MP: Melting point

Keywords: membrane perturbation, liposome, permeability enhancing effect, fatty acid, cytotoxicity, dye release, drug delivery.

4

1. Introduction The overexpression of secretory phospholipase A2 (sPLA2) in several cancer types has opened new avenues for active and selective liposomal drug delivery to cancer tissues [1]. The exploitation of the lipolytic sPLA2 to degrade liposomes and trigger drug release at the target site has shown promising results [1-3]. The sPLA2-catalyzed lipid hydrolysis will release an equimolar amount of lysolipids and fatty acids (Figure 1). Due to their intrinsic membrane perturbing properties, it has been postulated that the in-situ sPLA2-generated fatty acids and lysolipids can reduce the permeability barrier of the liposomal drug carrier as well as of the target cell membrane, thereby allow for enhanced drug release or drug uptake, respectively [2, 4]. However, fatty acids and lysolipids are themselves cytotoxic [5-7]. Despite the pronounced permeability enhancing effects shown with model membranes [8, 9], preliminary experiments with living cells suggest that fatty acids and lysolipids in concentrations below their cytotoxicity limit can barely enhance the intracellular uptake of anticancer drugs [4, 7, 10]. Fatty acids are amphipathic molecules with non-cylindrical molecular geometry; therefore, they form micelles in solutions and can incorporate readily into lipid membranes, which will increase the curvature stress within the lipid bilayer [4, 11, 12]. Consequently, free fatty acids can efficiently create instabilities in the lipid membrane and lower the membrane permeability barrier [9, 13]. The membrane perturbing effect of fatty acids is a complex and concentration-dependent process, which is strongly modulated by the chemical structure of the fatty acid, e.g., chain length and degree of unsaturation, as well as the phase state of the lipid bilayer [8, 9, 14]. A comparison between the influence of a selection of saturated and unsaturated fatty acids on the permeability barrier of DPPC, POPC, and POPC/cholesterol

5

liposomes was addressed in an earlier study [9]. It has been shown that fluid liposomes are more sensitive to the effect of fatty acids than gel-state liposomes. The effect of unsaturated fatty acids (e.g., oleic acid, OA) on the permeability barrier of liposomes was in general more pronounced than saturated fatty acids (e.g., palmitic acid, C16). In fluid membranes (POPC), the rate of dye release observed with OA was slightly higher for C16, both achieving more than 90% dye release after 15 minutes. In gel-state membranes (DPPC), the dye release induced by OA was more than 70% after 25 minutes compared with less than 20% with C16. The influence of free fatty acids on model and biological membranes, though addressed in some earlier studies, is still not well-understood [4]. Therefore, we carried out a systematic study to understand the membrane perturbing effect and cytotoxicity of a homologous series of saturated fatty acids (C8 to C20) on model large unilamellar liposomes (LUV) of 1,2- dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and the breast cancer MCF-7 cell line, respectively. In this context, DPPC liposomes serve as a model drug delivery system. The chemical structure and some of the properties reported in literature of the saturated fatty acids tested in this study are listed in Table 1. The critical micelle concentration (CMC) of fatty acids is inversely proportional to the length of the acyl chain, where increasing the acyl chain by two carbon atoms reduces the CMC by approximately five fold [15]. The CMC of the fatty acids used in this study ranges from 384 mM for caprylic acid (C8) to 0.08 mM for arachidic acid (C20) [15]. In contrast to the CMC, the melting point (MP) of a fatty acid from the solid to the liquid state increases with the chain length, and it ranges from 15°C for C8 to 75°C for C20 [16]. The partition coefficient (K) and molar enthalpy of partitioning (HP) for the partitioning of C10 to C16 fatty acids into gel-state DPPC liposomes are also shown

6

in Table 1; whereas K increases with the acyl chain length of the fatty acid, HP is determined by the underlining events that accompany the partitioning process [12]. In the present study, calcein-release assays were used to determine quantitatively the fatty acid-induced reduction in the membrane permeability barrier of DPPC liposomes. The influence of the fatty acids on the phase state and size distribution of DPPC liposomes were checked using differential scanning calorimetry (DSC) and dynamic light scattering (DLS), respectively. The cytotoxicity of the fatty acids was assessed using a standard MTT test. To evaluate the effect of the fatty acids on cellular drug uptake, a bioluminescence luciferin assay was developed using MCF-7 breast cancer cells engineered to produce firefly luciferase (luc2 gene) enzyme. Using luciferin as a model drug, any change in luciferin cellular uptake after the co-addition of the fatty acids with luciferin can be followed by monitoring any changes in the emerged luminescence signal. The luciferase-luciferin reaction was found to follow a hyperbolic-Michaelis-Menten-like kinetics with increasing luciferin concentration [3]. The fatty acid concentrations and experimental temperatures used in the different experiments and their relation to the CMC and MP values of the fatty acids are illustrated in Figure 2. Our results confirm the perturbing effect of fatty acids on model membranes, which is dependent on the acyl chain length, as well as show that saturated fatty acids below their cytotoxicity limits do not reduce the permeability barrier of biological membranes.

2. Materials and methods 2.1. Materials. 1,2-Dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) was purchased from Corden Pharma LLC (Switzerland). Luciferin (sodium salt) was purchased from Regis Technologies (USA), and Lipofectamine was obtained from Invitrogen. The

7

fatty acids (C8 – C20), calcein, Triton X-100 (TX-100), HEPES, and all other chemical and solvents were purchased from Sigma Aldrich (Germany). The purity of the fatty acids was > 98%. All substances were used as received without any further purification or modification. The aqueous solutions were prepared using ultrapure deionized water (Milli-Q, resistivity > 18 MΩ⋅ cm). If not otherwise specific, the concentration of the substances was calculated from the weight of the dry materials (weight/volume). 2.2. Calcein-release experiment. DPPC lipid was hydrated using HEPES buffer (10 mM HEPES, 154 mM NaCl, pH 7.4) containing calcein at a self-quenching concentration of 32 mM. The liposomes were prepared by extrusion 15 times through two 100 nm polycarbonate filters using an Avanti Mini Extruder (Avanti Polar Lipids, Alabaster, AL, USA) at a temperature higher than the phase transition temperature DPPC (42°C). Untrapped calcein was removed by dialysis in 5 x 1 L HEPES buffer using dialysis membranes (Spectra/Por Float-A-Lyzer G2, MWCO 3.5 – 5 kD) from Spectrum® Laboratories, Inc. The phospholipid concentration was determined using a procedure adapted from Bartlett’s phosphate assay [17]. The permeability enhancing effect of the fatty acids was measured in triplicates at 25°C using FLUOstar Omega Microplate Reader (BMG LABTECH) in a 96-well microplate (Nunc). The fatty acids were initially dissolved in ethanol before diluting the samples with HEPES buffer. After determining the baseline of calcein-loaded DPPC, 100 µl of the fatty acid solution (100 µM, ethanol < 0.2 vol.%) were injected into the lipid solution (100 µl, 250 µM) to achieve a final lipid concentration of 125 µM and fatty-acid-to-lipid molar ratio of 0.4 (total volume 200 µl). The emitted light was recorded for about 23 minutes at 520 nm using an excitation wavelength of 490 nm. A separate experiment was performed using Triton X-100 (0.2 wt.% final concentration) to determine the

8

100% calcein release value, which was used to calculate the percentage calcein release (%CR) for the fatty acids using the following equation %CRFA = (CRFA/CRTX100) x100%.

2.3. Differential scanning calorimetry (DSC). DPPC liposomes (average radius 73 nm) were prepared by sonication in HEPES buffer (10 mM HEPES, 154 mM NaCl, pH 7.4). Prior to the DSC experiment, the fatty acid dissolved in ethanol was added to the preformed liposomes to achieve a 1:1 molar ratio (ethanol < 0.2 vol.%). To prepare the premixed DPPC/fatty acid liposomes, DPPC dissolved in chloroform and the fatty acids dissolved in ethanol were mixed in a 1:1 molar ratio, after which the organic solvents were evaporated under vacuum. The premixed lipid films were hydrated in HEPES buffer (10 mM HEPES, 154 mM NaCl, pH 7.4), and the liposomes were produced by sonication in a low power bath sonicator at T > Tm for 30 minutes (average radius 86 nm). The DSC measurements were carried out using N-DSC-II calorimeter (Calorimetry Science, Provo, UT, USA). The lipid samples (1 mM DPPC) were degassed for 10 minutes before being loaded into the DSC cell. A scanning rate of 1 °C min-1 was applied. The reference cell was filled with buffer. The DPPC/Triton X-100 sample contained 0.2 wt.% Triton X-100. Several heating and cooling scans were performed to ensure the reproducibility of the thermograms. The thermograms were normalized to the DPPC content of each sample, and the DSC curves shown correspond to the second heating scan. 2.4. Dynamic light scattering (DLS). The dynamic light scattering measurements were performed by means of DelsaMax Pro (Beckman Coulter, Brea, CA). The system is equipped with a 50 mW diode pumped solid-state (DPSS) single longitudinal-mode laser operating at 532 nm, and an array of 32 photodetectors. The DPPC liposomes were prepared in HEPES buffer (10 mM HEPES, 154 mM NaCl, pH

9

7.4) by extrusion as described before. The titrations were performed in disposable 1 cm PMMA cuvette (Brand GmbH + Co. KG, Wertheim, Germany) at 25°C. The DPPC/fatty acid mixtures contained 50 µM DPPC and varying volumes of 275 µM fatty acid solutions (ethanol < 0.5 vol.%) and were equilibrated for one hour before the DLS measurements. The DPPC/Triton X-100 sample contained 0.2 wt.% Triton X-100. The samples were measured for at least 2 minutes at 25 °C. The particle radius averages were extracted from the unweighted particle size distributions, in which the particle size is correlated with its respective scattering intensity. 2.5. Cytotoxicity. Breast cancer MCF-7 cells were cultured in DMEM media (SigmaAldrich) supplemented with 10% fetal calf serum (Sigma-Aldrich), 1% pen-strep (Sigma- Aldrich), 1% GlutaMAX (Invitrogen), and 1 µg insulin per ml DMEM buffer (Sigma-Aldrich). For the assay, 10,000 cells per well in 50 µl DMEM buffer with supplements were plated in a 96-well microplate (Nunc) and incubated for 24 hours before the fatty acids were added. The fatty acids were initially dissolved in DMSO, and then dilutions were prepared using DMEM media with supplements (final DMSO < 3.2 vol.%). The fatty acid-containing solutions were carefully removed 24 hours after their addition and the cells were incubated at 37 °C in a fresh media with serum for 48 hours. Cytotoxicity was assessed using a standard 3-(4,5-dimethylthiazolyl)2,5-diphenyltetrazolium bromide (MTT) assay (Cell Proliferation Kit I, Roche, Germany). Cell viability is expressed as percentage MTT reduction relative to the untreated cells. The assay was performed in triplicates, and the absorbance was measured using a FLUOstar Omega Microplate Reader (BMG LAB- TECH) at 550 and 700 nm.

10

2.6. Bioluminescence bioassay. The assay was performed using a luciferase (luc2 gene) transfected MCF-7 breast cancer cell line. The cells were cultured in DMEM media (Sigma–Aldrich) supplemented with 10% fetal calf serum (Sigma-Aldrich), 1% pen-strep (Sigma-Aldrich), 1% GlutaMAX (Invitrogen), and 1 µg insulin per ml DMEM media (Sigma-Aldrich). For the assay, 2,500 cells per well in 50 µl DMEM media with supplements were seeded in a sterile white 96-well microplate (Nunc), and the cells were incubated for 24 hours. The fatty acids were initially dissolved in DMSO, and then 300 µM solutions were prepared using DMEM media with supplements (final DMSO < 0.1 vol.%). Prior to the addition of luciferin, 1 µl Lipofectamine (and 99 µl media) or 100 µl of the 300 µM fatty acid solution was added to the cells (final fatty acid concentration 200 µM). The experiments were performed in triplicates, and the luminescence signal from the luciferase-luciferin reaction was recorded in a FLUOstar Omega Microplate Reader (BMG LAB-TECH) thermostated at 37°C. After the equilibration of the luciferin solution and the seeded microplate for a few minutes at 37°C, the assay was started by the addition of 10 µl of a 16 mM luciferin solution (1 mM final concentration), which was also prepared in DMEM media with supplements. The bioluminescence is expressed as percentage relative to the untreated cells.

3. Results 3.1. Permeability assay The influence of the fatty acids (C8 – C12) on the permeability barrier of DPPC bilayer was assessed using liposomes loaded with calcein in a self-quenching concentration (32 mM). The percentage of calcein release (%CR) values for the tested fatty acids are shown in Figure 3, which were determined at a fatty-acid-to-lipid

11

molar ratio of 0.4 and normalized to the effect of 0.2 wt.% Triton X-100 (%CRFA = (CRFA/CRTX-100) x100%). Except for C8, the fatty-acid induced calcein leakage appear to increase with the length of the fatty acid, ranging from 0% for C10 to 25% for C20.

3.2. Differential scanning calorimetry (DSC) DSC was used to elucidate the effect of the fatty acids on the main phase transition of DPPC. The DSC thermograms as well as the extracted main phase transition temperature (Tm), and transition enthalpy (H) values are shown in Figure 4A, B and C, respectively. Two types of experiments were performed: (A) fatty acids were added to preformed DPPC liposomes, and (B) fatty acids were premixed with DPPC prior to the preparation of the liposomes. In contrast to the former case, the mixtures in the latter case were initially dissolved in organic solvents. The fatty-acid-to-lipid molar ratio was 1.0, and the DSC curves presented are the second heating scans. The data points in Figure 4A show the melting temperature of the fatty acids. A fully hydrated DPPC bilayer undergoes the main phase transition at around 42°C. The small differences in the thermogram of pure DPPC between “post-addition” and “premixing” samples (see Figure 4A) are caused by the different preparation procedures. The use of ethanol (EtOH) to dissolve fatty acids in the “post-addition” samples (final ethanol concentration < 0.2 vol.%) barely affected the phase transition of DPPC, which was completely abolished upon the addition of Triton X-100 detergent. As shown in Figure 4A, the short fatty acids (C8 and C10) appear to destabilize the DPPC bilayer, more pronouncedly in case of C10, which downshifted Tm (Figure 4B) and reduced H (Figure 4C). Unlike the short fatty acids C8 and C10, the fatty acids C12 to C20 substantially stabilized the DPPC bilayer increasing both Tm and H. The membrane stabilization was proportional to the length of the fatty

12

acid chains and reached a plateau at C16 (See Figure 4B). Interestingly, the premixing of DPPC with C12 totally eliminated the main phase transition of DPPC. In addition to increasing the DPPC lipid chain packing and stabilizing the bilayer, the DSC data show that C18 and C20 broaden the phase transition probably due to the formation of domains. This could be an indication of demixing of the fatty acids (i.e., C18 and C20) and DPPC as a result of the mismatch between the two components [18]. The influence of the fatty acids on the phase transition of DPPC was less pronounced in the case of the “premixed” samples, especially regarding the changes in H. This could be due to the preparation procedure, during which the lipid and fatty acid molecules could arrange to minimize the stress in the formed mixed bilayer. In addition, in the “premixed” liposomes the fatty acids have the possibility to interact with both lipid leaflets, which could further reduce the destabilizing effects of the fatty acids.

3.3. Dynamic light scattering (DLS) The DLS data for different fatty-acid-to-lipid ratios (0.5, 1 and 5) are shown in Figure 5. Unlike the detergent Triton X-100, which could efficiently solubilize DPPC liposomes, the C8 - C20 fatty acids were not found to be efficient membrane solubilizers in the investigated concentration range. At a molar ratio of 0.5, all fatty acids slightly increased the average size of the liposomes, most probably as a result of their incorporation into the lipid bilayer [19]. Large structures were observed with higher fatty-acid-to-lipid molar ratios. Expect for C8, which induced the formation of large structures, the DLS data clearly indicate that the formation of large structures was also dependent on the fatty acid length.

13

3.4. Cytotoxicity assay A standard viability assay (MTT test) using MCF-7 breast cells was utilized to test the cytotoxicity of the fatty acids. As shown in Figure 6, the addition of 0.25 mM fatty acids only slightly affected the viability of the cells, except for C20 that reduced the cell viability to about 67%. Resistance to the toxic effects of fatty acids has been reported before in literature [20, 21]; therefore the cells were challenged with higher fatty acid concentrations. Interestingly, by applying 1.3 mM fatty acids, only C12 showed substantial cytotoxicity, reducing the cells viability to around 37% of the control. A higher fatty acid concentration of 6.4 mM pronouncedly reduced the viability of the cells for all fatty acids. The fatty acids C12, C14, and C16 caused the maximum cell death (around 10% cell viability), whereas the other fatty acids (C8, C10, C18, and C20) were less cytotoxic retaining 35 – 48% cell viability.

3.5. Bioluminescence assay In order to assess the purported permeability enhancing effect of the fatty acids, we exploited a bioluminescence assay developed in our lab using living MCF-7 breast cancer cells engineered to produce firefly luciferase (luc2 gene) enzyme. A bioluminescence signal will be generated upon the addition of free luciferin substrate, which will be oxidized by the luciferase enzyme in a complex ATP-driven reaction emitting yellow-green light as a byproduct [3, 22]. Under well-defined experimental conditions, including number of cells and luciferin concentration, the generated signal is controlled by the amount of luciferin that enter the cells as well as the viability and health/activity status of the cells. In these experiments, sub-cytotoxic fatty acids concentration of 200 mM was used in addition to Lipofectamine, which served as a positive control. Lipofectamine is a mixture of cationic lipids that is widely used as a

14

transfection reagent for polynucleotides, and cationic-lipid-polynucleotide complex is believed to enter cells via endocytosis [23]. Free luciferin (1mM) was administered to the cells in the presence and absence of fatty acids (200 µM) or Lipofectamine, and the emergence of bioluminescence was monitored. The fatty acids were initially dissolved in DMSO, and the concentration used (< 0.1 vol.%) did not influence the assay (see Figure 7). Unlike Lipofectamine, which increased the bioluminescence signal to 160%, the fatty acids did not affect the cellular uptake of luciferin. The longer fatty acids, i.e., from C12 to C20, slightly reduced the bioluminescence to 65 – 79% of the control, implying their modest cytotoxicity at this concentration.

4. Discussion 4.1. Effect on model membrane In general, the interaction of fatty acids with lipid membranes is non-specific, poorly selective, and modulated by the physico-chemical characteristics of the fatty acid and the lipid bilayer in question. The incorporation of fatty acids into lipid membranes and their flip-flop across the bilayer occur almost instantaneously [12, 24], and this in turn will increase the curvature stress within the lipid bilayer due to the fatty acid’s non-cylindrical geometry [25, 26]. The local accumulation of the fatty acids in the membrane will create small-scale heterogeneities and membrane defects, possibility accompanied by toroidal pores and membrane morphological changes, all of which will destabilize the membrane and reduce its permeability barrier [4, 11]. At high enough fatty acid concentrations, the membrane may eventually disintegrate into bicelles and micelles, in a detergent-like manner [19, 27-29].

15

Our results clearly indicate that the fatty-acid induced perturbations are proportional to the acyl chain length of the fatty acid. This concurs well with earlier work in our lab, where the rate and efficiency of dye release from catansome vesicles were found proportional to fatty acid concentration and acyl chain length [14]. Whereas short fatty acids tend to destabilize the lipid bilayer, longer fatty acids can stabilize gel-state membranes rendering them more rigid [30] as well as efficiently create membrane instabilities and morphological changes. The large structures observed by DLS can result from liposome aggregation, tubular formation, and structural changes into nonvesicular forms [19, 31]. Since the fatty-acid concentrations used in the present assay were below the corresponding CMC values (see Figure 2), the higher membrane permeabilizing effect of the longer fatty acids (Figure 3) can be explained by their stronger partitioning into the lipid bilayer. For instance, an earlier thermodynamic study on the interaction between saturated fatty acids (C10, C12, C14, and C16) and gel-state DPPC liposomes (at 20°C) revealed that the degree of partitioning into the lipid bilayer is strongly linked to the chain length of the fatty acid (the partition coefficient (K) and molar enthalpy of partitioning (HP) values are listed in Table 1) [12]. The relationship between the fatty acid chain length and the partition coefficient (K) was found to be linear; increasing the fatty acid length by two carbon atoms increased K by almost one order of magnitude. The incorporation of the short fatty acid C10 into DPPC gel-state bilayer was associated with positive (exothermic) enthalpy of partitioning (HP) indicating high favorable entropic forces. The partitioning of longer fatty acids was coupled with stronger exothermic events, with a maximum of HP attained for C14. The major exothermic driving forces for the incorporation of fatty acids, and amphipathic molecules in general, are the van der Waals bonding and

16

membrane partitioning, whereas molecule-desolvation and membrane perturbations are endothermic events [32]. The higher-than-anticipated effect of C8 (See Figures 3 and 5) can be attributed to fact that, in contrast to all fatty acids, C8 is in the liquid (oil) state at 25°C (see Table1 and Figure 2), which could have increased the induced membrane perturbations. Another possible explanation is the higher mismatch between C8 and DPPC (C16 acyl chains), which could have reduced their miscibility [15].

4.2. Effect on biological membranes Fatty acids are natural components that play an important role in many physiological cell functions [33-35]. However, aberrant levels of fatty acids, especially saturated fatty acids, can accompany cellular dysfunction and damage [36-38]. Fatty acids can exert their biological and pathological effects via receptor-mediated signaling as well as via altering the physical and mechanical properties of lipid bilayers [39]. The cytotoxicity of fatty acids is mainly derived from their ability to perturb the integrity and function of cell membranes, which can lead to programmed cell death (apoptosis) as well as to necrosis and cell lysis at high enough concentrations [5, 4042]. It has also been noted that necrosis occurs directly after the exposure to fatty acids, whereas apoptosis takes place in a subsequent stage [43]. It seems from earlier studies that susceptibility to apoptosis depends on the cell type [21], whereas necrosis (and cell lysis) is not cell-type dependent [44]. In general, the fatty acid concentration required for inducing apoptosis is in the sub-millimolar range [21, 45, 46], and millimolar concentrations are needed to lyse the cells [44]. The cytotoxicity of fatty acids is highly reduced in the presence of serum proteins to which fatty acids strongly bind [6, 43].

17

The MCF-7 cell line in our study appears to be resistant to the apoptotic pathway, as the viability was only significantly reduced upon the use of micromolar concentrations of the fatty acids, probably mainly due to the disruption of the lipid part of the cell membrane and the induction of necrosis and cell lysis. Generally, the cytotoxic/cell-lysis effect of saturated fatty acids increase with their length, which agrees with earlier reports [46, 47]. Still, the particularly high cytotoxicity of C12, C14, and C16 fatty acids is still unclear and awaits further investigations. Unlike observations with DPPC liposomes, sub-cytotoxic fatty acid concentrations did not enhance the cellular uptake of luciferin. Although the fatty acid concentration used in the bioluminescence assay (200 µM) might not be physiologically relevant, it is not expected that smaller and more physiologically relevant concentrations will lead to different outcome.

5. Conclusions The efficient membrane perturbing effect saturated fatty acids exert on model membrane was not completely transferable to biological cell membranes. It is understandable that simple model lipid membranes under thermodynamic equilibrium conditions do not fully represent the composition complexity and dynamic nature of real biological membranes as well as the fact that biological membranes are intrinsically non-equilibrium systems. In contrast to simple model membranes, cell membranes provide a much larger lipid pool thereby limiting the effect of fatty acids. In addition, the surface of cell membranes is crowded with floating bulky macromolecules, which will restrict the membrane surface area exposed to exogenous compounds. These distinct characteristics of cell membranes can explain the differences between the tendencies observed with the model lipid membranes and the

18

cell assays. We believe the present comparative study should pave the way for further understanding of the membrane perturbing effect amphipathic molecules show on model and biological membranes.

Acknowledgments Dr. Steffen Schmidt and Prof. Jan Mollenhauer (Institute for Molecular Medicine, University of Southern Denmark) are gratefully acknowledged for providing the luciferase-transfected MCF-7 breast cancer cell line. This work was supported by The Lundbeck Foundation Center of Excellence NanoCAN (Nanomedicine Research Center for Cancer Stem Cell Targeting Therapeutics) and the Danish Council for Independent Research-Technology and Production Sciences (DFF-FTP)

19

References [1] A. Arouri, A.H. Hansen, T.E. Rasmussen, O.G. Mouritsen, Lipases, liposomes and lipid-prodrugs, Curr. Opin. Colloid Interface Sci. 18 (2013) 419-431. [2] T.L. Andresen, S.S. Jensen, K. Jørgensen, Advanced strategies in liposomal cancer therapy: Problems and prospects of active and tumor specific drug release, Prog. Lipid Res. 44 (2005) 68-97. [3] A. Arouri, J. Trojnar, S. Schmidt, A.H. Hansen, J. Mollenhauer, O.G. Mouritsen, Development of a cell-based bioassay for phospholipase A2-triggered liposomal drug release, PLoS One 10 (2015) e0125508. [4] A. Arouri, O.G. Mouritsen, Membrane-perturbing effect of fatty acids and lysolipids, Prog. Lipid Res. 52 (2013) 130-140. [5] J.Y. Kong, S.W. Rabkin, Palmitate-induced apoptosis in cardiomyocytes is mediated through alterations in mitochondria: prevention by cyclosporin A, Biochim. Biophys. Acta 1485 (2000) 45-55. [6] C.C. Hsieh, M.H. Yen, H.W. Liu, Y.T. Lau, Lysophosphatidylcholine induces apoptotic and non-apoptotic death in vascular smooth muscle cells: in comparison with oxidized LDL, Atherosclerosis 151 (2000) 481-491. [7] A. Arouri, O.G. Mouritsen, Phospholipase A2-susceptible liposomes of anticancer double lipid-prodrugs, Eur. J. Pharm. Sci. 45 (2012) 408-420. [8] J. Davidsen, O.G. Mouritsen, K. Jørgensen, Synergistic permeability enhancing effect of lysophospholipids and fatty acids on lipid membranes, Biochim. Biophys. Acta 1564 (2002) 256-262. [9] H. Jespersen, J.H. Andersen, H.J. Ditzel, O.G. Mouritsen, Lipids, curvature stress, and the action of lipid prodrugs: free fatty acids and lysolipid enhancement of drug transport across liposomal membranes, Biochimie 94 (2012) 2-10. [10] N. Rasmussen, J.H. Andersen, H. Jespersen, O.G. Mouritsen, H.J. Ditzel, Effect of free fatty acids and lysolipids on cellular uptake of doxorubicin in human breast cancer cell lines, Anticancer Drugs 21 (2010) 674-677. [11] O.G. Mouritsen, Lipids, curvature, and nano-medicine, Eur. J. Lipid Sci. Technol. 113 (2011) 1174-1187. [12] P. Høyrup, J. Davidsen, K. Jørgensen, Lipid membrane partitioning of lysolipids and fatty acids: effects of membrane phase structure and detergent chain lengths, J. Phys. Chem. B 105 (2001) 2649-2657. [13] B.D. McKersie, J.H. Crowe, L.M. Crowe, Free fatty-acid effects on leakage, phase properties and fusion of fully hydrated model membranes, Biochim. Biophys. Acta 982 (1989) 156-160. [14] K.R. Rosholm, A. Arouri, P.L. Hansen, A. González-Párez, O.G. Mouritsen, Characterization of fluorinated catansomes: A promising vector in drug-delivery, Langmuir 28 (2012) 2773-2781. [15] S.S. Tzocheva, P.A. Kralchevsky, K.D. Danov, G.S. Georgieva, A.J. Post, K.P. Ananthapadmanabhan, Solubility limits and phase diagrams for fatty acids in anionic (SLES) and zwitterionic (CAPB) micellar surfactant solutions, J. Colloid Interface Sci. 369 (2012) 274-286. [16] G. Knothe, R.O. Dunn, A Comprehensive Evaluation of the Melting Points of Fatty Acids and Esters Determined by Differential Scanning Calorimetry, J. Am. Oil Chem. Soc. 86 (2009) 843-856. [17] G.R. Bartlett, Phosphorus assay in column chromatography, J. Biol. Chem. 234 (1959) 466-468.

20

[18] A. Arouri, M. Dathe, A. Blume, Peptide induced demixing in PG/PE lipid mixtures: A mechanism for the specificity of antimicrobial peptides towards bacterial membranes?, Biochim. Biophys. Acta 1788 (2009) 650-659. [19] S. Wang, J. Huang, Q. Song, H. Fu, Characterizing assembly morphology changes during solubilization process of dimyristoyl phosphocholine vesicles by ndodecyl triethylammonium bromide, J. Colloid Interface Sci. 311 (2007) 296-302. [20] M.C. Kim, M.G. Kim, Y.S. Jo, H.S. Song, T.I. Eom, S.S. Sim, Effects of C18 fatty acids on intracellular Ca(2+) mobilization and histamine release in RBL-2H3 cells, Korean J. Physiol. Pharmacol. 18 (2014) 241-247. [21] L.N. Andrade, T.M. de Lima, R. Curi, A.M. Castrucci, Toxicity of fatty acids on murine and human melanoma cell lines, Toxicol. in Vitro 19 (2005) 553-560. [22] A.H. Hansen, O.G. Mouritsen, A. Arouri, Enzymatic action of phospholipase A2 on liposomal drug delivery systems, Int. J. Pharm. 491 (2015) 49-57. [23] I. Mortimer, P. Tam, I. MacLachlan, R.W. Graham, E.G. Saravolac, P.B. Joshi, Cationic lipid-mediated transfection of cells in culture requires mitotic activity, Gene Ther. 6 (1999) 403-411. [24] F. Kamp, D. Zakim, F. Zhang, N. Noy, J.A. Hamilton, Fatty acid flip-flop in phospholipid bilayers is extremely fast, Biochemistry 34 (1995) 11928-11937. [25] R.S. Cantor, Lipid composition and the lateral pressure profile in bilayers, Biophys. J. 76 (1999) 2625-2639. [26] N. Fuller, R.P. Rand, The influence of lysolipids on the spontaneous curvature and bending elasticity of phospholipid membranes, Biophys. J. 81 (2001) 243-254. [27] B. Bechinger, K. Lohner, Detergent-like actions of linear amphipathic cationic antimicrobial peptides, Biochim. Biophys. Acta 1758 (2006) 1529-1539. [28] H. Heerklotz, Interactions of surfactants with lipid membranes, Q. Rev. Biophys. 41 (2008) 205-264. [29] M.N. Jones, Surfactants in membrane solubilisation, Int. J. Pharm. 177 (1999) 137-159. [30] T. Inoue, S. Yanagihara, Y. Misono, M. Suzuki, Effect of fatty acids on phase behavior of hydrated dipalmitoylphosphatidylcholine bilayer: saturated versus unsaturated fatty acids, Chem. Phys. Lipids 109 (2001) 117-133. [31] A. De la Maza, J.L. Parra, Vesicle-micelle structural transition of phosphatidylcholine bilayers and Triton X-100, Biochem. J. 303 (1994) 907-914. [32] A. Arouri, M. Dathe, A. Blume, The helical propensity of KLA amphipathic peptides enhances their binding to gel-state lipid membranes, Biophys. Chem. 180181 (2013) 10-21. [33] M.A. Gijon, C.C. Leslie, Regulation of arachidonic acid release and cytosolic phospholipase A2 activation, J. Leukocyte Biol. 65 (1999) 330-336. [34] P.D. Berk, D.D. Stump, Mechanisms of cellular uptake of long chain free fatty acids, Mol. Cell. Biochem. 192 (1999) 17-31. [35] J. Lunn, H.E. Theobald, The health effects of dietary unsaturated fatty acids, Nutr. Bull. (BNF) 31 (2006) 178-224. [36] H. Shimano, Novel qualitative aspects of tissue fatty acids related to metabolic regulation: lessons from Elovl6 knockout, Prog. Lipid Res. 51 (2012) 267-271. [37] J.P. Little, J.M. Madeira, A. Klegeris, The saturated fatty acid palmitate induces human monocytic cell toxicity toward neuronal cells: exploring a possible link between obesity-related metabolic impairments and neuroinflammation, J. Alzheimers Dis. 30 (2012) S179–S183.

21

[38] J.Y. Lee, L. Zhao, D.H. Hwang, Modulation of pattern recognition receptormediated inflammation and risk of chronic diseases by dietary fatty acids, Nutr. Rev. 68 (2010) 38-61. [39] S.C. Frasch, D.L. Bratton, Emerging roles for lysophosphatidylserine in resolution of inflammation, Prog. Lipid Res. 51 (2012) 199-207. [40] G.C. Sparagna, D.L. Hickson-Bick, L.M. Buja, J.B. McMillin, A metabolic role for mitochondria in palmitate-induced cardiac myocyte apoptosis, Am. J. Physiol. Heart Circ. Physiol. 279 (2000) H2124-2132. [41] H. Finstad, M. Myhrstad, H. Heimli, J. Lomo, H. Blomhoff, S. Kolset, C. Drevon, Multiplication and death-type of leukemia cell lines exposed to very longchain, Leukemia 12 (1998) 921-929. [42] T. Martins de Lima, M.F. Cury-Boaventura, G. Giannocco, M.T. Nunes, R. Curi, Comparative toxicity of fatty acids on a macrophage cell line (J774), Clin. Sci. 111 (2006) 307-317. [43] M. Cnop, J. Hannaert, A. Hoorens, D. Eizirik, D. Pipeleers, Inverse relationship between cytotoxicity of free fatty acids in pancreatic islet cells and cellular triglyceride accumulation, Diabetes 50 (2001) 1771-1777. [44] J.A. Lapre, D.S. Termont, A.K. Groen, R. Van der Meer, Lytic effects of mixed micelles of fatty acids and bile acids, Am. J. Physiol. 263 (1992) G333-337. [45] M. Artwohl, M. Roden, W. Waldhausl, A. Freudenthaler, S.M. BaumgartnerParzer, Free fatty acids trigger apoptosis and inhibit cell cycle progression in human vascular endothelial cells, FASEB J. 18 (2004) 146-148. [46] E. Diakogiannaki, S. Dhayal, C.E. Childs, P.C. Calder, H.J. Welters, N.G. Morgan, Mechanisms involved in the cytotoxic and cytoprotective actions of saturated versus monounsaturated long-chain fatty acids in pancreatic beta-cells, J. Endocrinol. 194 (2007) 283-291. [47] R.A. Lovstad, Fatty acid induced hemolysis. Protective action of ceruloplasmin, albumins, thiols and vitamin C, Int. J. Biochem. 18 (1986) 771-775.

22

Figure captions

Figure 1. (Top) The sPLA2-catalyzed hydrolysis products of phospholipids, i.e., fatty acids and lysolipids. (Bottom) The effect of fatty acids and lysolipids, separately or in combination, on the curvature stress of lipid bilayers. The figure was adapted from Ref. [4].

23

Figure 2. Critical micelle concentration (CMC) and melting point (MP) of the fatty acids listed in Table 1, shown as a function of chain length, with indications of the temperature and fatty acid concentration used in the various experiments.

24

Figure 3. Maximum calcein release from calcein-loaded DPPC liposomes after the addition of various fatty acids (C8 – C20). The values represent the percentage of the calcein release (CR%) induced by 0.2 wt.% Triton X-100, i.e., %CRFA = (CRFA/CRTX100)

x100%. The measurements were carried out at 25°C using 125 µM liposomes

(100 nm in diameter) loaded with 32 mM calcein and a fatty-acid-to-lipid ratio of 0.4. The solutions were prepared in HEPES buffer (10 mM HEPES, 154 mM NaCl, pH 7.4). The error bars represent the standard deviation of three different measurements (n = 3).

25

Figure 4. Heating DSC scans (A), main phase transition temperature T m (B), and transition enthalpy H (C) of DPPC liposomes before and after the addition of fatty acids (C8 – C20) with fatty-acid-to-lipid ratio of 1.0. The data points in (A) show the melting temperature of the fatty acids, and the line was inserted as a guide to the eye. In post-addition experiments, the fatty acids were added to already prepared liposomes, whereas in the premixing experiments the fatty acids were premixed with DPPC before the preparation of the liposomes. Control experiments were performed to show the effect of ethanol (EtOH) and Triton X-100 (TX-100) on DPPC phase transition. The mixtures contained 1 mM DPPC, and the solutions were prepared in HEPES buffer (10 mM HEPES, 154 mM NaCl, pH 7.4). The DSC curves shown here correspond to the second heating scan, and the baseline of the curves was shifted for clarity.

26

27

28

Figure 5. Average hydrodynamic radii obtained from DLS measurements of DPPC liposomes before and after the addition of fatty acids (C8 – C20) with fatty-acid-tolipid ratio of 0.5, 1, and 5. Triton X-100 was also added as a control solubilizing agent with similar ratios. The DPPC/fatty acid mixtures contained 50 µM DPPC and 25, 50, and 250 µM fatty acids. The solutions were prepared in HEPES buffer (10 mM HEPES, 154 mM NaCl, pH 7.4) and the measurements were carried out at 25°C.

29

Figure 6. Effect of the fatty acids (C8 – C20) on the viability of MCF-7 cells, presented as percentage of the untreated control, at different fatty acid concentration (0.25, 1.3 and 6.4 mM) as determined using a standard MTT test. The error bars represent the standard deviation of three different measurements (n = 3).

30

Figure 7. Effect of Lipofectamine and fatty acids (C8 – C20) on the maximum bioluminescence, presented as percentage of the untreated control, generated from luciferase-luciferin reaction after the addition of 1 mM luciferin to MCF-7 cells expressing luciferase enzyme in the presence of 1µl Lipofectamine or 200 µM fatty acid. Lipofectamine serves as a positive control. The error bars represent the standard deviation of three different measurements (n = 3).

31

Table 1. The code name, chemical structure, critical micelle concentration (CMC) of the sodium salt (at 25°C), and melting point (MP) of the fatty acids (FA) used in the present study. The partition coefficient (K) and molar enthalpy of partitioning (HP) are for the partitioning of the fatty acids into DPPC vesicles at 20°C. CMC and MP values were adapted from refs. [15] and [16], respectively. K and HP were adapted from ref. [12]. Data are not available in some cases (–). Fatty acid

Code

Structure

CMC

/

Na-FA

MP (°C)

K (M-1)

HP (kcal mol-1)

(mM)

Caprylic acid

C8

CH3(CH2)6COOH

384

15

–

–

Capric acid

C10

CH3(CH2)8COOH

93

31

65

2.1

Lauric acid

C12

CH3(CH2)10COOH

23

43

740

-2.7

Myristic acid

C14

CH3(CH2)12COOH

5.5

54

7000

-7.2

Palmitic acid

C16

CH3(CH2)14COOH

1.3

62

103000

-2.9

Stearic acid

C18

CH3(CH2)16COOH

0.3

69

–

–

Arachidic acid

C20

CH3(CH2)18COOH

0.08

75

–

–

32