Comparison of photodynamic efficiency of cholesterol, selected cholesterol esters, metabolites and oxidation products in lipid peroxidation processes*

Cholesterol (Ch) is one of the most important compo- nents of biological membranes which has a significant impact on their biophysical properties. As a key compo- nent of lipid membranes, along with other unsaturated lipids present in a biological membrane Ch undergoes an oxidation reaction during oxidative stress. Cholesterol oxidation products, cholesteryl esters and metabolites are also localized in the lipid membranes, where they may modify the membrane properties. In this work, the impact on lipid peroxidation (induced by a photody- namic action) of cholesterol, selected cholesteryl esters, cholesterol oxidation products and metabolites has been studied using EPR oximetry and direct detection of singlet oxygen phosphorescence at 1270 nm. The obtained rate constant values of interaction of selected lipids and sterols with singlet oxygen indicate that the tested compounds are not efficient singlet oxygen quenchers. Nev- ertheless, to different extents, presence of sterols modi-fies the oxygen photoconsumption rate in peroxidisable liposomes.


INTRODUCTION
Cholesterol (Ch) is one of the most important components of biological membranes and its presence and concentration have a significant impact on their biophysical properties, such as fluidity, permeability or hydrophobicity (Subczynski et al., 2017) (Fig. 1). As a key component of lipid membranes, Ch may be exposed to a stream of free radicals or singlet oxygen during oxidative stress and along with other unsaturated lipids present in a membrane serves as a substrate for oxidation reaction (Geiger et al., 1997;Vila et al., 2001;Kulig et al., 2016). Cholesterol oxidation products, such as 7-ketocholesterol (7KCh) or 5,6-epoxy cholesterol, are also hydrophobic, and therefore they remain in the cell membranes where they may interact with other membrane components and modify the membrane properties (Kulig et al., 2016;Wnętrzak et al., 2017). 7KCh is one of the most extensively studied oxysterols and is considered to be the most cytotoxic (Anderson et al., 2020). Increased levels of 7KCh have been noted in the human atherosclerotic plaque (Rao et al., 2014) and in ocular tissues where it accumulates with age and after extensive exposure to light (Rodriguez & Fliesler, 2009;Rodriguez et al., 2014). Atherosclerosis development is accompanied by accumulation of cytoplasmic lipid droplets mainly composed of cholesteryl oleate (ChO) (Brown & Goldstein, 1983) (Fig. 1). The source of cholesteryl oleate for lipid droplets are foam cells derived from macrophages that took up oxidized low-density lipoprotein (LDL) (Enomoto et al., 1987;Nagano et al., 1991). LDL consists of a hydrophobic core of neutral lipids, mainly cholesteryl esters, surrounded by a polar surface shell and serves as the major cholesterol transporter in human plasma (Koivuniemi et al., 2009).
Cholestanol is a saturated sterol existing at a very low physiological concentration in animals (Dotti et al., 2001) (Fig. 1). However, an increase in its level  occurs along with the development of a hereditary disease, the cerebrotendinous xanthomatosis (CTX) (Dotti et al., 2001;Seyama, 2003). Cholestanol deposition causes neurological dysfunction, mental retardation and juvenile cataract (Bhattacharyya et al., 2007). It has been proposed that the mechanism responsible for cell damage in CTX involves an imbalance of the cholesterol/cholestanol ratio in the plasma membrane and structural changes caused by cholestanol, which in turn disturb the calcium channel function of the membrane (Seyama, 2003). It seems that physiologically relevant sterols localized in the lipid membranes exert a significant impact on their biophysical properties and function. In this work, the effect of cholesterol, selected cholesteryl esters, cholesterol oxidation products and metabolites on lipid peroxidation (induced by a photodynamic action) has been studied.
Sodium phosphate, potassium phosphate, sodium chloride, potassium chloride, liquid chromatography grade benzene, chloroform, and methanol were purchased from Avantor Performance Materials Poland (Gliwice, Poland) and used as supplied.

Preparation of multilamellar liposomes
Multilamellar liposomes were prepared by the film deposition method as previously described (Nayar, 1989;MacDonald et al., 1991;Subczynski et al., 2007). Briefly: selected synthetic lipids were dissolved in chloroform which was then evaporated with a stream of neutral gas (nitrogen or argon). The lipid film formed on the bottom of the glass tube was completely dried under reduced pressure for 2-4 h. A phosphate buffer saline (PBS, 10 mM, pH 7.4), previously incubated for at least 24 h with chelex to remove a trace of transition metal ions, was added to the dried lipid film at room temperature, in small portions to the final volume. During preparation, the samples were vortexed vigorously to completely remove the lipid film from the test tubes.
In these studies, liposomes with two slightly different lipid compositions were prepared. One type of liposomes comprised of POPC (16 mM), DMPC (2 mM) and Ch (2 mM) or its selected esters, oxidation products or metabolites. The second type of liposomes tested in this work was composed of partly oxidised (16:0)(22:6) PC (final concentration 9 mM) , DMPC (2 mM) and Ch or its selected esters, oxidation products or metabolites (1 mM).
The final concentrations of synthetic lipids in both suspensions of multilamellar liposomes (MLV) were 20 mM and 12 mM, respectively. All preparations were performed in darkness or under dim light and, where possible, under nitrogen.

Singlet Oxygen Quenching Measurements
To determine the rate constants of the interactions of singlet oxygen with cholesterol and its selected esters, oxidation products or metabolites, time-resolved direct detection of singlet oxygen ( 1 O 2 , 1 ∆ g ) phosphorescence at 1270 nm was used. In these measurements, changes in singlet oxygen lifetime were monitored as a function of the sterol concentration. Tetraphenylporphyrine (TPP) was used as an efficient singlet oxygen generator, with quantum yield up to 0.6 in benzene (Bonnett et al., 1988;Redmond & Gamlin, 1999). Absorbance of TPP in a chloroform solution was adjusted to 0.05 at the excitation wavelength (λ=645 nm). Sensitizer solution was placed in a quartz fluorescence cuvette (QA-1000; Hellma, Mullheim, Germany) and excited with 645 nm light generated by an integrated nanosecond DSS Nd:YAG laser system (NT242-1k-SH/SFG; Ekspla, Vilnius, Lithuania). The near-infrared singlet oxygen phosphorescence was measured perpendicularly to the excitation beam in a photon-counting mode using a thermoelectric cooled NIR PMT module (H10330-45; Hamamatsu Photonics, Hamamatsu, Japan). Measurements were repeated with increasing concentrations of cholesterol, its selected esters, oxidation products or metabolites. Data analysis, including first-order 1 O 2 ( 1 ∆ g ) phosphorescence decay fitted by the Levenberg-Marquardt algorithm, was performed by custom-written software.

Statistical Analysis
Statistical analysis was performed by Student's ttest, and linear regression was performed by the method of least squares.

Green-light induced oxygen consumption in the presence of Rose Bengal
The EPR oximetry measurements enabled to compare the impact of cholesterol, selected cholesteryl esters, cholesterol oxidation products and cholesterol metabolites on the rate of oxygen consumption (d[O 2 ]/dt) in liposome samples of two types in respect to their composition and peroxidisability, undergoing oxidation induced by two different agents. In the first type of liposomes, comprised of POPC, DMPC and cholesterol or its derivative or analogue, oxygen consumption was induced by irradiation with green light (516-586 nm; 44 mW/ cm 2 ) in the presence of Rose Bengal (10 µM). Results of oxygen photo-uptake measurements in liposomes, subjected to RB-photosensitized oxidation are presented in Fig. 2 and Table 1.
The rate of oxygen consumption (d[O 2 ]/dt ) measured in the dark was very low and did not exceed 4×10 -4 (mM/min), indicating that oxygen was hardly consumed, and lipids did not undergo significant oxidation. Similar values were obtained for green light irradiated samples without a photosensitizer (not shown). The oxygen uptake rates measured in irradiated samples containing a sensitizer were the highest in liposomes comprised of cholestanol (0.401±0.094 mM/min), 7-ketocholesterol (0.353±0.038 mM/min), cholesterol 5α,6α -epoxide  Table 1). RB is an efficient, water soluble photosensitizer that can generate both, the singlet oxygen and free radicals upon irradiation with visible/green light (Neckers, 1989;Różanowska et al., 1995;DeRosa & Crutchley, 2002). It has been shown that quantum yield of singlet oxygen generation by RB in PBS solutions reaches up to 0.75 (Redmond & Gamlin, 1999). Although RB, as a moderately hydrophobic compound (Crandon et al., 2020), easily associates with the lipid membrane (Lambert & Kochevar, 1997), it cannot incorporate into it because it exists as a dianion in polar solvents (Stockett et al., 2020). Thus, RB generates reactive oxygen species (ROS) in the aqueous phase at the surface of the membrane, and not in the membrane itself. On the other hand, the substrates for the reaction with ROS generated by RB, peroxidizable lipids, are localised inside the lipid membrane of liposomes suspended in this aqueous solution. The lifetime of a singlet oxygen ( 1 O 2 ( 1 ∆ g )) in water is All bars represent green light-induced (516-586 nm; 44 mW/cm 2 ) initial rates of oxygen uptake (mM/min) in the presence of Rose Bengal (10 µM). Oxygen consumption rate in the dark-incubated samples did not exceed 4x10 -4 mM/min (not shown). Differences in the oxygen photo-uptake rate between 7KCh and each of cholesteryl esters are statistically significant (P<0.05). only 4 μs (Rodgers & Snowden, 1982), and the sites of the 1 O 2 ( 1 ∆ g ) attack, the double bonds in the lipid molecules forming the liposome membrane, are located inside the membrane. Location and depth of immersion of cholesterol and its derivatives/analogues in the membrane depends on lipid composition of the membrane, the length of fatty acid chains esterified in phospholipids, and the degree of their unsaturation (Subczynski et al., 2012;Yang et al., 2016). The highest values of d[O 2 ]/dt were observed in liposomes containing cholesterol, cholestanol, ketocholesterol and cholesterol epoxide (Fig. 2, Table 1). All of these molecules contain a hydroxyl group, which due to its location close to the membrane surface, may increase the probability of ROS to reach inside the lipid bilayer. In the case of cholesterol esters, in which the hydroxyl group is "blocked" by a fatty acid or succinate esterified at this site, and in the case of cholestane, which is devoid of the hydroxyl group, the rates of oxygen consumption were significantly lower. There was almost a two-fold decrease in the rate of oxygen consumption in comparison to the cholesterol-containing membranes. These observations are confirmed by the results obtained in liposomes of the second type containing partially peroxidised (16:0)(22:6) PC, DMPC and Ch or its derivative/analogue and irradiated with blue light (404-515 nm; 55 mW/cm 2 ).

Blue-light induced oxygen consumption in the presence of highly peroxidisable (16:0)(22:6)PC
In these samples, the lowest oxygen consumption rates were also observed in the presence of cholesterol esters. The values of d[O 2 ]/dt for HSCh, PCh and OCh were as follows: (6.9±0.2) µM/min, (8,5±0.2) µM/min and (11.8±2.0) µM/min (Fig. 3, Table 2). The highest rates of oxygen photoconsumption were observed in the presence cholestanol, cholestane, cholesterol epoxide and ketocholesterol: (19.3±2.0) µM/min, (18.3±0.04) µM/min, (18.8±2.0) µM/min and (17.7±1.0) µM/min, respectively. Only a slightly lower value of d[O2]/dt was recorded in the irradiated control sample, i.e. liposomes comprised of (16:0)(22:6)PC and DMPC only: (14.3±1.5) µM/min (Fig. 3, Table 2). Low, but noticeable level of oxygen uptake was observed in all samples incubated in the dark, possibly due to autooxidation of highly peroxidizable (16:0)(22:6)PC. Polyunsaturated docosahexaenoic acid (DHA) esterified at the SN-2 position of this lipid plays a very important role in the protection of photoreceptor membranes in the retina of the eye (Jeffrey et al., 2001). This acid is highly susceptible to free radical peroxidation, the products of which may affect degeneration of the retina (Liu et al., 2014). Recently, it has been shown that oxidation of (16:0)(22:6)PC extends its absorption spectrum to the visible region of the electromagnetic spectrum and makes it photoreactive (Różanowska et al., 2021). Oxidized (16:0)(22:6)PC, or rather a mixture of its oxidation products, shows significant photoreactivity when exposed to short wavelength visible light. Oxidized (16:0)(22:6)PC efficiently generates singlet oxygen and free radicals under blue light illumina- All black bars represent oxygen uptake rate measured in samples incubated in the dark. Differences in oxygen photo-uptake rate between ChP-or HSCh-containing samples and the other samples are statistically significant (P<0.05). Table 2. Oxygen uptake rate (µM/min) in highly peroxidisable liposomes containing (16:0)(22:6)PC (9 mM), DMPC (2 mM) and cholesterol or selected cholesteryl esters, cholesterol oxidation products or cholesterol metabolites (1 mM). Oxygen photo consumption induced in liposomes irradiated with blue light (404-515 nm; 55 mW/cm 2 ) or incubated in the dark. tion (Różanowska et al., 2021). It is not known which oxidation products of (16:0)(22:6)PC are responsible for the observed photoreactivity and if they are present in the retina, but similar results were obtained in case of ex vivo oxidised lipids extracted from bovine retinas (Koscielniak et al., 2017;Pawlak et al., 2019). In liposomes containing partially peroxidized (16:0) (22:6)PC, the reactive oxygen species were likely generated inside the membrane. And the main substrate for ROS generated by partially oxidised (16:0)(22:6)PC in the tested samples was mainly the phospholipid itself. In the control sample, which contained (16:0)(22:6)PC and DMPC, the only substrate for oxidation was the unsaturated lipid. Unfortunately, it was not possible to monitor the rate of oxidation/uptake of (16:0)(22:6)PC during the irradiation of the tested samples. In the presence of cholesteryl esters, a significant decrease in d[O 2 ]/dt was observed when compared to the control liposomes. Although partial inhibition of oxygen photoconsumption observed in the presence of cholesteryl esters may indicate their antioxidant role, it seems that structural effects induced by cholesteryl esters in the lipid membranes are responsible for the observed reduction in oxygen consumption rate in liposomes containing (16:0)(22:6)PC. Indeed, ESR studies employing appropriate spin labels indicated that the ester bond was directed towards the surface of the membrane, while the tetracyclic part of Ch and the rest of the esterified fatty acid were buried in the hydrophobic interior of the bilayer (Grover et al., 1979). Atomistic simulation studies of the location of cholesteryl oleate in LDL indicated that the oleate chains in most of ChO molecules, especially in the LDL core, existed mainly in extended conformations relative to the ring structures. However, those molecules, located close to the lipoprotein surface, were in a kinked conformation, where the angle between the fatty acid chain and the cholesterol molecule was below 60° (Heikelä et al., 2006). Esters lacking a hydroxyl group are less polar and therefore localize deeper in the membrane. This causes the molecules of other lipids to slide apart, and the change is even more profound. The rest of the oleic acid has a double bond in the chain. This bond occurs in the cis conformation, so the chain of this ester is bent at an angle of 133° (Weijers, 2016) and takes up more space. Due to changes in the arrangement of lipid molecules in the membrane bilayer, forced in a way by the cholesteryl ester molecules introduced between them, water molecules can penetrate deeper into the membrane. Considering that water dramatically reduces the singlet oxygen lifetime, when compared to the lipid environment (Baier et al., 2005), presence of water molecules in the membrane could reduce oxidation of the membrane lipids and the corresponding rate of oxygen uptake. It cannot be ruled out that a similar effect was responsible for the reduced oxygen consumption observed in liposomes irradiated in the presence of RB. The presence of Ch slightly accelerated the rate of oxygen uptake in the tested system during exposure to blue light in comparison to the control sample. It should be remembered that cholesterol, also easily oxidized, may act as an additional substrate for ROS generated by partially oxidised (16:0)(22:6)PC exposed to blue light. Interestingly, the lowest d[O 2 ]/dt in irradiated samples containing partially oxidized (16:0)(22:6)PC was observed in the presence of cholesterol derivatives characterised by the highest interaction rate constant with singlet oxygen.

Singlet oxygen quenching measurements
Results of the determined singlet oxygen quenching rate by cholesterol, its selected esters, oxidation products and metabolites are presented in Table 3. Saturated lipids interact with singlet oxygen only physically, while unsaturated lipids, including cholesterol and its peroxidisable derivatives/analogues, interact with 1 O 2 ( 1 Δ g ) both chemically and physically (Bacellar & Baptista, 2019). A condition for chemical interaction with singlet oxygen is the presence of a double bond which the singlet oxygen molecule can attack. Among the examined cholesterol derivatives, the double bond was present only in Ch itself, cholesteryl esters and 7KCh. The determined rate constant of the interaction of Ch with singlet oxygen, kq is 3.64×10 4 M -1 s -1 . This value does not differ substantially from the literature data (Broniec et al., 2011). Vever-Bizet et al., determined (by the same method) the rate constant for the reaction of cholesterol with 1 O 2 in benzene to be 5.7×10 4 M -1 s -1 (Vever-Bizet et al., 1989). As expected, the rate constants of HSCh and PCh interactions' with singlet oxygen were almost identical to that determined for Ch alone and were as follows: 3.63×10 4 M -1 s -1 and 3.53×10 4 M -1 s -1 , respectively. In the case of the cholesteryl oleate molecule, the constant of the interaction rate with 1 O 2 was almost twice as high and amounted to 6.46×10 4 M -1 s -1 . This is because an additional double bond appeared in the molecule of this cholesteryl ester, located in the esterified oleic acid (18: 1) chain. The determined rate constant of the interaction of OCh with 1 O 2 also did not differ from the literature kq values determined for the oleic acid and oleic acid methyl ester, which were: 5.3×10 4 M -1 s -1 and 7.4×10 4 M -1 s -1 , respectively (Vever-Bizet et al., 1989).
The obtained rate constants of the selected lipids and sterols interaction with singlet oxygen led to a conclusion that the tested compounds were not efficient singlet oxygen quenchers. The obtained values (~10 4 M -1 s -1 ) were much lower than those reported for typical for biological antioxidants (10 8 -10 10 M -1 s -1 ) (Cantrell et al., 2003;Gruszka et al., 2008).