Characterization of Retinol Stabilized in Phosphatidylcholine Vesicles with and without Antioxidants

Abstract

Retinol stability has been reported to be improved by encapsulation in liposomes, both with and without cholesterol. However, this improvement is limited because of lipid peroxidation. In this study, we compare the stability of retinol in phosphatidylcholine liposomes under ultraviolet (UV) light or standard room air, with and without the addition of antioxidants. Both butylated hydroxytoluene (BHT) and a proprietary mix (StoppOx) improved the shelf stability from <10 to over 30 d. The addition of cholesterol had no effect. Fluorescence imaging showed a heterogeneous distribution of retinol within the vesicles, including within the aqueous layer. Fluorescence lifetimes were equally heterogeneous. Under UV irradiation, StoppOx protected retinol for significantly longer than BHT and via different mechanisms. This suggests that natural antioxidants work well to improve the retinol stability, but that further work to determine the optimal vesicle structure remains to be performed.

Introduction

Retinol (vitamin A) is used in pharmaceuticals and cosmetics for dermal applications to correct hyperpigmentation, reduce the appearance of fine lines, and treat acne and atopic dermatitis. Both prescription and nonprescription topical creams have shown to result in clinically significant improvement of the appearance of aged and photodamaged skin.¹⁻⁵ Creating stable formulas for consumer use, however, remains a barrier as retinol is highly sensitive to light and oxygen. UV light causes degradation into a variety of potentially harmful products. Photoirradiation of retinol by UV light (254 nm) in an oxygenated ethanolic solution has been shown to produce retinal, ROH 5,6-epoxide, 5,8-epoxyretinol, and 13,14-epoxyretinol. Irradiated retinol can also activate photosensitizers, such as chlorophyll, rose bengal, or riboflavin, with drastic alteration of the resulting photoproducts.⁶ Depending upon the conditions, reactive oxygen species (ROS), including singlet oxygen, may be produced.⁷⁻⁹ Endogenous photosensitizers in cells result in the production of a large number of photoproducts not predicted from simple solutions; these products may damage DNA and cell membranes directly or indirectly (Scheme 1). Because these effects are well known, most retinoid preparations are intended as “night creams” with use during the day not recommended. Preparations must be protected from light, refrigerated, and protected with an antioxidant preservative, usually butylated hydroxytoluene (BHT). Most retinol preparations sold to cosmetics manufacturers are preserved with BHT.

Scheme 1. Reaction Pathways of Liposome-Encapsulated Retinol.

Photoexcitation may lead to direct breakdown into retinal, epoxides, or other products, or may lead to excited-state retinol which interacts with oxygen to produce ROS, including singlet oxygen and superoxide; these ROS may break down retinol and lipids; lipoperoxyl radicals further degrade retinol. The ellipses and italicized labels indicate where the degradation reactions may be inhibited: oxygen scavengers interact directly with molecular oxygen and prevent the formation of ROS, while antioxidants react with ROS and other radicals to compete with their reaction with retinol.

Recent concern over possible adverse health effects of BHT has led to its removal from food products by major manufacturers. General Mills removed BHT from its packaged cereals in 2015, and Kellogg’s has offered a variety of BHT-free products, although they have planned to continue using BHT in their cereals. BHT and butylated hydroxyanisole are widely used antioxidant food additives that have been extensively studied for their safety; although they have not been associated with cancer,¹⁰ consumer concerns persist. Several studies have reported pulmonary toxicity in mice resulting from oral or topical exposure to BHT.^11,12 In 2007, a study suggested that BHT may aggravate allergic responses.¹²

Although concentrations found in cosmetics are reported to be safe for topical use,¹³ there is increasing interest in BHT-free topical retinol products. Methods of stabilization include natural antioxidants, encapsulation, or some combination of both. Encapsulation materials include different types of lipids, chitosan, pectin, and zein particles.¹⁴⁻²⁰ Lipid particles have been shown to improve the skin uptake of retinol compared to control solutions,^21,22 but simple lipid encapsulation can increase the shelf life only to a limited degree because the lipids themselves are vulnerable to oxidation. Lipid microcapsules oxidize more rapidly than bulk lipids because of the increased surface area and preferential localization of lipid hydroperoxides at the oil–water interface; lipid oxidation may cause increased degradation of encapsulated products.^23,24 Retinoids can photosensitize the formation of lipid hydroperoxides in a UV dose-dependent fashion.²⁵

A number of studies use antioxidants, including BHT and metal chelates, to improve the stability of lipid-encapsulated pharmaceuticals.²⁶⁻²⁸ Other studies have looked at oxygen scavengers, which interact directly with oxygen rather than with secondary products of oxidation reactions. Oxygen scavengers include sodium sulfite (Na₂SO₃), a combination of glucose oxidase and catalase (GOx + catalase) and ascorbic acid (AA).²³

The purpose of this study was to characterize retinol encapsulated in phosphatidylcholine (PC) multilamellar vesicles. We characterized the liposomes using confocal microscopy and fluorescence lifetime imaging (FLIM). Imaging demonstrated an unusual pattern of retinol distribution throughout the liposomes, including within the aqueous phase. The distribution of retinol fluorescence was found to be diffuse and dispersed, not overlapping the staining seen using membrane-targeting dyes such as FM 1-43. Retinol fluorescence lifetimes varied from liposome to liposome and among domains in individual liposomes, reflecting the environmental effects of liposome size and lamellarity on fluorescence lifetimes.

Retinol was encapsulated in PC with and without preservatives (BHT or a “natural” lipophilic preparation [StoppOx]). Using the native fluorescence of retinol, the stability was monitored for several weeks under ambient conditions, or for 20 min under UV irradiation. Liposomes with and without 10% cholesterol were also tested under these conditions because addition of cholesterol has been reported to improve the stability of liposome-encapsulated retinol.²⁹

Retinol photoluminescence has been well studied and is an established method of determining retinol concentration.^30,31 The absorption band at 325 nm with an emission at ∼490 nm is due to a π–π transition of the retinoid’s polyene chain. As the retinol degrades, this fluorescence disappears. Multilamellar vesicles were more effective than unilamellar vesicles in protecting retinol, so raw vesicles without extrusion or freeze–thaw, which makes vesicles uniform and unilamellar,³² were selected for a long-term study. We found that StoppOx was as effective as BHT in preserving retinol under both ambient conditions and UV light. Liposome preparations containing StoppOx were stable for at least 1 month under ambient light and temperature. A biphasic-modified Gompertz growth model was used to fit the decay. When exposed to UV light, retinol fluorescence degraded over a time course of 20 min, with liposomes containing StoppOx being more stable than liposomes with BHT. The photodegradation curves could be fit to a mix of first- and second-order processes, which indicated that StoppOx protected the liposomes from first-order photobleaching, while BHT did not.

Results and Discussion

Encapsulation Efficiency

High-performance liquid chromatography (HPLC) was used only to estimate the encapsulation efficiency. The supernatant from retinol-containing liposomes had retinol concentrations below the limits of detection (<5 IU/mL retinol), indicating an encapsulation efficiency of ≥98.9%. Previous studies have seen only small changes in efficiency (<1%) due to pH changes (between 3 and 11) and increased mass loading,^14,18 so we did not vary these parameters for our studies.

Vesicle Characterization

The vesicle shape, size, and lamellarity were investigated by confocal microscopy, with and without an additional membrane dye (N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino)styryl)pyridinium dibromide [FM 1-43] or N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)phenyl)hexatrienyl)pyridinium dibromide [FM 4-64]) (Figure 1). PC liposomes containing retinol had a mean (±standard deviation) diameter of 4 ± 3 μm (range: 1.2–56 μm, N = 909), as measured from the microscopic images. With 10% cholesterol, the mean size was reduced slightly, with a marked reduction in the largest vesicles (mean diameter: 3 ± 3 μm, range: 1.2–24 μm, N = 553). FM 4-64 emits in the far-red and was compared with the greener FM 1-43 to ensure that there was no channel bleed-through from retinol emission (see Figure S1 for spectra). Retinol emission appeared diffuse, mostly near the edges of the liposome but between layers as well, in contrast to the sharp membrane labeling seen with the FM dyes (Figure 1A–C). Zoomed images of individual vesicles showed some retinol fluorescence limited to lipid bilayers, but a substantial amount occurring throughout subcompartments of the liposome or even the entire liposome (Figure 1D,E). Colabeling with FM dyes illustrated the thickness of the bilayer and emphasized the presence of the retinol between layers (Figure 1F–L). The addition of cholesterol led to slightly smaller vesicles but no qualitative change in the fluorescence distribution (Figure 1M). This appearance is consistent with the ability of retinol to rapidly transfer between and among lipid bilayers, as has been observed by bulk spectroscopy and radiolabeling. A time constant of <30 s has been reported for transfer between bilayers, indicating that despite its hydrophobic nature, retinol is able to dissociate from lipids and travel through the aqueous phase without the assistance of shuttle proteins.³³⁻³⁵

Figure 1.

Confocal images of PC-retinol liposomes. Scale bars = 5 μm. The colors represent a pseudocolor code as follows: blue, excitation 405 nm, emission 420–480 nm; yellow, excitation 488 nm, emission 495–550 nm; and red, excitation 543 nm and emission 680 nm longpass. (A) Schematic of some possible configurations of multilamellar liposomes. (B) Image of retinol-only liposomes. (C) Image of PC liposomes containing retinol and labeled with FM 1-43, showing the dye-only and lack of bleed-through in the yellow pseudo-colored channel. (D,E) Zoom of single liposomes with retinol only. (F–I) Zoom of single liposomes with retinol and FM 1-43. (J–L) Zoom of single liposomes with retinol and FM 4-64. (M) Single liposome containing 10% cholesterol, with retinol and FM 4-64.

FLIM images also showed significant heterogeneity in the liposome size, lamellarity, and retinol content. The fluorescence lifetime of retinol alone excited at 730 nm was 3.4 ± 0.3 ns in our experiments, with significant variation among vesicles and different layers of the same vesicle (Figure 2A). In the presence of 10% cholesterol, areas of longer lifetime were apparent in the images; the lifetime showed two distinct distributions and could not be fit to a single decay. One component was equivalent to that of the PC-only vesicles (3.5 ± 0.4 ns), while the other was longer (4.5 ± 0.3 ns). Many vesicles appeared similar to the PC-only vesicles, where a few were clearly dominated by the longer lifetime, suggesting a heterogeneous distribution of the cholesterol (Figure 2B). Closeups of single vesicles showed homogeneous lifetimes within and between bilayers of PC-only vesicles (Figure 2C), while longer lifetime domains could be seen in cholesterol-containing vesicles (Figure 2D). Lifetime histograms showed a single distinct lifetime for PC-only vesicles, but two distinct components for cholesterol-containing vesicles; the strength of this second component varied from sample to sample, with the data shown representing one chosen field of view (Figure 2E). These findings are consistent with previous results, which have shown that retinol photophysical properties are sensitive to the molecular conformation and environment,³⁶ with reported lifetimes of 2.3 ns in glycerol, 4.2 ns in hexane, and 6–8 ns when bound to proteins such as albumin or retinol-binding protein.³⁷

Figure 2.

Fluorescence lifetime images of retinol in vesicles. Excitation was 730 nm and emission was 500–550 nm. Images are color-coded according to the lifetime, as shown in the color bar. (A) PC-retinol. (B) PC-retinol-cholesterol. (C) Zoom in of a single PC-retinol vesicle. (D) Zoom in of a single PC-retinol-cholesterol vesicle, showing areas suggesting domains. (E) Histogram of lifetimes of all vesicles in panels A (PC only) and B (+Chol), showing a single discrete peak in the PC-only liposomes and two peaks in the liposomes containing cholesterol.

All imaging experiments were performed on fresh samples. As retinol degraded, it lost its fluorescence. No emission was observed from vesicles stored for 30 d or exposed to UV light for 20 min (not shown). The addition of KI did not quench the vesicle fluorescence (Figure S2).

Because of the high degree of heterogeneity, we also prepared batches of unilamellar vesicles by freeze–thaw. These vesicles were uniform in size and lamellarity (data not shown). However, retinol in these vesicles was less stable than that in the multilamellar vesicles, so their use was abandoned. Their fluorescence was also not substantially affected by KI (Figure S3).

Spectroscopy and Stability Analysis

Retinol in PC vesicles showed a broad fluorescence emission peak at 490 nm when excited at 348 nm (Figure 3A). Both BHT and StoppOx showed some degree of fluorescence at this excitation wavelength, but it was significantly blue-shifted relative to that of retinol, so that a wavelength could be chosen that would indicate retinol only in both fresh and degraded solutions (505 nm). It was important to monitor the fluorescence changes of the antioxidant-only controls as solutions degraded because the antioxidant compounds showed shifts in emission with time and light exposure (Figure 3B).

Figure 3.

Fluorescence emission spectra of retinol, antioxidants, and lipids with excitation at 348 nm. (A) Freshly prepared solutions. Retinol dominates the emission; to avoid the StoppOx signal, peaks were measured at 505 nm for decay curves. (B) In degraded solutions, the emission of StoppOx dominates, although it is less than that seen in fresh solutions.

Long-term stability analyses showed that both BHT and StoppOx prevented retinol degradation in liposomes for at least 30 days before the loss of emission began. There was no significant difference between BHT and StoppOx (Figure 4A). Liposomes containing retinol only, without antioxidants, were stable for ∼8 d before the signal began decaying exponentially (Figure 4A).

Figure 4.

Retinol stability in liposomes with or without BHT or StoppOx. (A) Relative emission. The data points show fluorescence emission at 505 nm normalized to time t = 0. For the first 30 d, no significant decay of retinol in BHT- or StoppOx-containing liposomes was seen. The points show relative emission ± SEM for two independent studies, each run in triplicate; the lines are linear fits with the slope insignificantly different from 0. Retinol-only showed a plateau for 8.0 ± 0.6 d, followed by an exponential decay with k = 0.17 ± 0.03. (B) Natural log of emission normalized to its maximum value. The emission of antioxidant-protected liposomes relative to retinol-only liposomes showed stability for 30 d followed by slow decay, which could be fit to the biphasic model discussed in the text.

After 30 d, the antioxidant-containing liposomes began to show a loss of signal in a nonexponential fashion (Figure 4B). The fluorescence of StoppOx-containing liposomes normalized to that of the controls could be fit to a modified Gompertz equation that identifies a turning point t_max where the sigmoidal growth decays.³⁸ The full equation for the model may be written as

1

where I/I_max is the fluorescence intensity normalized to its maximum value; t is the time; t_max is the time at which the maximum signal occurs; k is a rate constant that determines the rate at which the growth rate approaches the final net growth rate, μ_min, at long times. The advantage of this model is that it is an equation with a single nonlinear parameter, k. This fit then yields a meaningful interpretation of the curve. It may also be seen that at values of t close to t_max, the curve becomes a parabola

2

allowing for the choice of initial fit values of the parameters or for an approximate fit of I to a Gaussian function. Approximation of the function as a Gaussian did not change the quality of the fit for our data, making this a very simple model. The fit values for two independent 42-day studies were consistent, and are plotted together in Figure 4B. Fit values were t_max = 31.5 ± 0.5 days and kμ_min = 0.0033 ± 0.0002 days^–2.

Photolysis

Liposomes with retinol only and retinol plus BHT or StoppOx were exposed to UV irradiation at 900 and 600 μW/cm². Two independent preparations of vesicles were prepared in 6-fold replicates each, and fluorescence emission was measured every 2 min for 20 min. The results of the two independent trials were consistent; one is shown in Figure 5A. It can be seen that while StoppOx protected retinol from photolysis under these conditions, the presence of BHT led to increased degradation at long times (>10 min for 900 W/cm² and >15 min for 600 W/cm²).

Figure 5.

Fluorescence decay as a function of time under UV irradiation. (A) High power irradiation, fluorescence intensity normalized to time 0 for retinol only, retinol + BHT, and retinol + StoppOx. (B) Reciprocal of normalized intensity for high power irradiation with fit to eq 4a (fit parameters in Table 1). (C) Low power irradiation, fluorescence intensity normalized to time 0 for retinol only, retinol + BHT, and retinol + StoppOx. (D) Reciprocal of normalized intensity for low power irradiation with fit to eq 4a (fit parameters in Table 1).

These data could be fit to a model of photobleaching, as described previously for photosensitizers and other fluorophores.³⁹⁻⁴¹ The photobleaching reactions are expressed as a composite of reactions with a first-order term and a second-order term

3a

This may be rewritten as

3b

The integrated rate equation is

which can be written as

4a

or

4b

which is a fit with only two parameters, C₁ and C₂R₀. The observed fluorescence intensity is proportional to the light flux Φ as well as to [R] in accordance with the equation

where Q is the quantum yield of the molecule, ε is the molar absorptivity, and b is the pathlength. Thus, at a given intensity of excitation, the observed signal can be considered to be proportional to [R].

The data were excellent fits to eq 4a (Figure 5A,B) with fit values given in Table 1. Several observations may be made: one, C₁ appears to depend upon light fluence only when StoppOx is present. C₁ is overall reduced by StoppOx but is increased by BHT. Both BHT and StoppOx increase C₂, which is <0. This means that C₂ does not represent second-order bleaching but second-order reconstitution of the fluorophore.

Table 1. Fit Parameters for Photolysis to eq 4a^a.

	C1	C2R0	R2	change in C1	change in C2
retinol only: high	0.156 ± 0.009	–0.32 ± 0.01	0.9936	N/A	N/A
retinol only: low	0.149 ± 0.005	–0.58 ± 0.04	0.9975	N/A	N/A
retinol + StoppOx: high	0.149 ± 0.004	–0.74 ± 0.02	0.9981	NS	P < 0.0005
retinol + StoppOx: low	0.098 ± 0.008	–0.56 ± 0.08	0.9923	P < 0.0005	NS
retinol +BHT: high	0.26 ± 0.01	–0.74 ± 0.06	0.9941	P < 0.0005	P < 0.0005
retinol +BHT: low	0.211 ± 0.007	–0.83 ± 0.02	0.9971	P < 0.0005	P < 0.0005

^aThe errors indicate uncertainty in the fit parameters. The last two columns indicate the significance of the change in each parameter relative to retinol-only, as measured by the extra sum-of-squares F test. N/A = not applicable; NS = not significant.

Studies with varying quenchers and oxygen concentrations have determined that the second-order reactions are due to the interaction of the fluorophore with species produced by neighboring fluorophores and then diffusing to the site of action. These species are unlikely to be singlet oxygen, which has a very short lifetime and thus radius of diffusion, but represent other radicals.³⁹ In this case, retinol may be regenerated by direct reduction or interaction with molecular oxygen. Ignoring triplet interactions, the possible reactions are

5

where X are nonfluorescent products.

BHT appears to increase the regeneration of retinol by second-order, light fluence-dependent processes. However, it increases direct, first-order photobleaching, rendering the molecule less stable to light after a period of several minutes. StoppOx reduces first-order photobleaching and makes this process more directly responsive to light fluence.

Conclusions

Retinol native fluorescence emission may be used to quantify the stability and to visualize the encapsulation in vesicles, with or without other dyes. A commercial blend of antioxidants, StoppOx, works as well as BHT for stabilizing retinol in liposomes for 30 days under ambient conditions; it is superior to BHT under UV light. Retinol degrades over a period of 30–40 days under ambient conditions and over a period of 20 min when exposed to 900 W/cm² 312 nm light, with a mixture of first- and second-order processes. Further experiments with time-resolved spectra and deoxygenated solutions will elucidate these mechanisms further.

Although a significant number of studies have been done on the stability of retinol, the use of its native fluorescence emission for imaging is rare. The images here show some remarkable features. Retinol is found throughout the vesicles, including in the aqueous phase. Our finding that unilamellar vesicles were less stable than multilamellar vesicles may be related to this finding; retinol in unilamellar vesicles may concentrate predominantly in the aqueous phase and degrade more rapidly. The addition of aqueous antioxidants, such as AA, may further increase the stability in both unilamellar and multilamellar preparations.

Experimental Procedures

High-Performance Liquid Chromatography

Using the standard method for testing vitamin A, HPLC-ultraviolet (UV) was performed by Intertek (Champaign, IL) (details are proprietary). For tests of encapsulation efficiency, liposomes were sedimented and the unencapsulated retinol remaining in the supernatant was quantified after dilution with vegetable oil. Encapsulation efficiency was estimated as

6

where IU is the international unit used to measure vitamin quantities, 3.33 × 10³ mg is the conversion unit from IU to mg, 1.3 mg is the total amount of retinol encapsulated in the lipid solution, and the numeral 10 refers to the 10 mL of supernatant released into oil. The results were presented as IU/serving where a serving is 1 g/mL. The instrument limit of detection was 5 IU/mL.

Chemicals

StoppOx (“Mixture from Sunflower Oil cold pressed organic, Tocopherols, bavarian hop extract and Ethylferulate”) was purchased from All Organic Treasures GmbH. Crystalline all trans retinol, PC (from egg yolk, ≥40% enzymatic), chloroform, methanol, sodium chloride, glycine, cholesterol, pentetic acid (DTPA), potassium iodide (KI), and BHT were all purchased from Sigma-Aldrich and used as delivered. Glass beads, 5 mm, case no. S800242, were purchased from Fisher Scientific.

Liposome Preparation

The method used for preparing liposomes was adapted from the literature.²⁹ For stability studies, a 13 mL solution of chloroform and methanol (2:1 v/v) was prepared with 130 mg of PC in a 25 mL round-bottom flask. For trials assessing antioxidant properties, 13 μL of StoppOx or 1.3 mg of BHT was added. If needed, 10% cholesterol was added to the solution during this step as well. The organic solvents were removed under vacuum while being heated at 30 °C until only a lipid film was left in the flask. A 13 mL aqueous buffer (10 mM glycine and 0.115 M NaCl, adjusted to pH 9 with NaOH) was added to the round-bottom flask along with glass beads to help with homogeneity. The flask was then placed under nitrogen and 1.3 mg of retinol was added. The flask was then removed from nitrogen and the solution was mixed until homogeneous. The solution was stored in a closed 15 mL falcon tube at room temperature for immediate use or at 4 °C, wrapped in a foil, for long-term storage.

For microscopy analyses and ultraviolet irradiation studies, retinol was added to the chloroform/methanol solution instead of the aqueous solution.

Absorbance and Emission

Fluorescence spectra, with excitation at 350 nm and emission at 400–704 nm, were collected on a CLARIOstar plate reader (BMG Labtech) using epi-illumination in a 96-well black plate (Greiner). For quenching studies, KI was added from a 3 M stock solution to the desired concentration.

Long-Term Stability

Liposome solutions were prepared and left at room temperature under ambient fluorescent room light. The fluorescence spectra were obtained for all the solutions in triplicates every 1–3 days, where the fluorescence emission was normalized to retinol-only liposome solutions.

UV Irradiation

Six-fold replicate samples in a 96-well quartz plate were exposed to a 312 nm UV transilluminator (Fisher Scientific, Pittsburgh, PA). Power at the sample was measured at 312 nm using a Coherent FieldMaster power meter with an IM2 UV sensor and was ∼900 μW/cm². The power was modulated by placing the sample on a 3 mm thick glass plate, which reduced the power to ∼600 μW/cm². The fluorescence spectrum with excitation at 348 nm and emission at 380–740 nm was recorded after successive 2 min UV exposure times.

Confocal Microscopy and FLIM

Confocal and lifetime images were obtained on a ZEISS LSM 880 confocal system with an Airyscan detector. All images shown were obtained using a 63× oil immersion objective (NA = 1.4). Airyscan achieves super-resolution images by capturing 32 images of a single field of view across a specialized detector. Each image contains positional information that enables the reconstruction of a single super-resolution image.^28,42 Channels collected were pseudo-colored as follows: Blue, excitation 405 nm (diode laser), emission 420–480; Green, excitation 488 nm (Ar ion), emission 495–550 nm; and Red, excitation 543 nm (He–Ne), emission 680 nm longpass. FLIM microscopy was performed on the same microscope using a Becker & Hickl SPC-150 FLIM module with a BIG2 detector. The dichroic in the BIG2 detector collected emissions in two channels: Channel, 1 500–550 nm (green) and Channel 2, 575–610 (red). Excitation for FLIM was provided by a Chameleon Ultra Ti:Sapphire laser (Coherent) operating at a repetitive rate of 80 mHz. The two-photon wavelength used for excitation was 730 nm based upon the literature.^43,44 FLIM data were analyzed using SPCImage (Becker & Hickl). Size distributions were obtained from confocal images using Fiji⁴⁵ “analyze particles.”

Curve Fitting

Graphing, calculation of standard deviations and standard errors, and curve fitting were performed using GraphPad Prism 8. Statistical differences between parameters were measured using the extra sum-of-squares F test and the AIC model.

Glossary

Abbreviations

PC

phosphatidylcholine

FLIM

fluorescence lifetime imaging microscopy

BHT

butylated hydroxytoluene

SEM

standard error of the mean

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.0c02102.

• Spectra of lipids with dyes and results of KI quenching studies (PDF)

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

This work was supported by a contract from Marie-Veronique Inc. to Portland State University.

The authors declare no competing financial interest.