Spontaneous Transfer of Indocyanine Green (ICG) From Liposomes To Albumin Is Inhibited By The Antioxidant α-Tocopherol

Abstract

Indocyanine Green (ICG) is a clinically approved organic dye with near-infrared absorption and fluorescence. Over the years, many efforts to improve the photophysical and pharmacokinetic properties of ICG have investigated numerous nanoparticle formulations, especially liposomes with membrane embedded ICG. A series of systematic absorption and fluorescence experiments, including FRET experiments using ICG as a fluorescence energy acceptor, found that ICG transfers spontaneously from liposomes to albumin protein residing in the external solution with a half-life of ~10 minutes at 37 °C. Moreover, transfer of ICG from liposome membranes to external albumin reduces light-activated leakage from thermosensitive liposomes with membrane embedded ICG. A survey of lipophilic liposome additives discovered that the presence of clinically approved antioxidant, α-tocopherol, greatly increases ICG retention in the liposomes (presumably by forming favorable aromatic stacking interactions), inhibits ICG photobleaching, and prevents albumin-induced reduction of light triggered liposome leakage. This new insight will help researchers with the specific task of optimizing ICG-containing liposomes for fluorescence imaging or phototherapeutics. More broadly, the results suggest a broader design concept concerning light triggered liposome leakage; that is, proximity of the light absorbing dye to the bilayer membrane is a critical design feature that impacts the extent of liposome leakage.

Graphical Abstract

INTRODUCTION

Indocyanine Green (ICG) has been used clinically for many decades as a near-infrared (NIR) fluorescent dye for in-vivo diagnostics and imaging. ICG has high affinity for blood proteins (especially albumin, globulins, α-lipoproteins), which confines the dye to the blood vessels and enables spectroscopic monitoring and fluorescence imaging of blood perfusion in wide range of diseases and injuries.^1,2,3 ICG clears fairly rapidly from the blood via the biliary pathway, and monitoring the change in bloodstream ICG levels after intravenous dosage is a convenient way to assess physiological functions such as liver function or cardiac output.^4,5 The recent discovery that ICG emission extends into the short wave infrared (SWIR) range beyond 1000 nm where the signal to background ratio is improved augurs well for continued development of ICG-based biomedical technologies.^6,7,8

ICG is an amphiphilic molecule with a propensity to self-aggregate at low micromolar concentrations in water, and logP has been measured to be 1.81 when the aqueous phase is phosphate buffer (pH 7.4, 100 mM).⁹ Over the years, efforts to improve the photophysical and pharmaceutical properties of ICG have investigated numerous nanoparticle formulations.^10,11,12,13 A significant fraction of these nanoparticle formulations are ICG-containing liposomes and they have been evaluated for various applications in fluorescence imaging and phototherapeutics. ^{11,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37} Quite a few published studies have assessed the stability of ICG-containing liposomes by monitoring the change in sample absorbance or fluorescence intensity over time. But this type of bulk sample measurement does not unambiguously eliminate the possibility of ICG transfer from the liposomes to albumin and related serum proteins within the same sample. This uncertainty is potentially significant because albumin is the most abundant blood protein (typical concentration in human blood is 35–50 mg/mL),³⁸ and a relatively fast and extensive ICG transfer process could alter the imaging or phototherapeutic performance of the liposomes.

The literature on ICG-containing liposomes includes studies that investigated the dependence of ICG properties of dye concentration, liposome composition, and the location of the dye within the bilayer membrane.^14,21 Molecular dynamics simulations suggest that ICG is buried within the lipophilic core of simple bilayer membranes, but in pegylated liposomes the ICG can move to the membrane surface and interact with the PEG chains.^16,18 Here, we show that ICG transfers readily from standard fluid phase phospholipid liposomes to bovine serum albumin (BSA) in the external solution. BSA has an ICG association constant of 5 ×10⁵ M⁻¹ and is employed here as a cheap, functional substitute for human serum albumin.³⁹ The data includes a collection of related but independent fluorescence intensity and Förster Resonance Energy Transfer (FRET) studies. We find that liposome composition impacts the extent of ICG transfer to BSA and we have discovered that the presence of α-tocopherol within the liposome membrane greatly increases ICG retention and inhibits ICG photobleaching. The potential of this ICG transfer process to diminish phototherapeutic performance was assessed by conducting experiments that measured NIR light triggered leakage of aqueous contents from ICG-containing liposomes. The results will help researchers with the specific task of optimizing ICG-containing liposomes for phototherapeutics, and they also hint at a broader concept concerning light triggered liposome leakage; that is, ICG proximity to the bilayer membrane is a critical design feature that impacts the extent of liposome leakage.

RESULTS AND DISCUSSION

Octanol Partitioning Experiments

Many vendors and academic publications quote a calculated logP value for ICG of 6.05 or similar, reflecting very high lipophilicity.^14,40 Curiously, measured values of logP are reported to be much smaller (e.g., logP = 1.81 ± 0.06 when the aqueous phase is pH 7.4, 100 mM phosphate buffer⁹). Our interest was the change in logP when BSA was present in the aqueous phase. As shown in Figure 1a, a significant difference in 1-octanol partitioning is visible to eye, and quantitative measurements (Figure 1b) determined logP to be 2.05 ± 0.34 when the aqueous phase was 50 mM HEPES buffer, 100 mM NaCl, pH 7.4, and 0.55 ± 0.11 when BSA (0.25 mM) was included in the aqueous phase. Thus, BSA can effectively extract ICG from a 1-octanol phase and into an adjacent aqueous phase.

Figure 1.

Effect of BSA on ICG partitioning in 1-octanol-buffer at 22°C. (a) Color photographs of ICG (10 μM) partitioned between (left) 1-octanol and HEPES buffer (50 mM, 100 mM NaCl, pH 7.4) and (right) 1-octanol and HEPES buffer + BSA (0.25 mM). (b) LogP values (22 °C) for ICG; average of triplicate measurements ± SD.

Studies that Treat ICG Liposomes with BSA

Summarized in Scheme 1 are the various liposome compositions that were evaluated in this study (the chemical structures are provided in Scheme S1), and analysis of the liposomes by Dynamic Light Scattering (Table 1 with raw data in Figures S1–S3) indicated negligible variation in liposome size or morphology. In Figure 2a–b are reference absorbance and fluorescence spectra of free ICG, ICG + BSA, and ICG Liposomes. The data highlights the known solvatochromism of ICG;^41,42 that is, the wavelength of the maxima band is red-shifted in less polar media, which is what occurs upon ICG binding within the hydrophobic pocket of BSA.⁸ Moreover, the fluorescence red-shift effect is even larger when the ICG is incorporated within liposomal membranes.^14,16,21,31 In Figure 2c–d are representative absorbance and fluorescence spectra acquired during a titration experiment that added aliquots of BSA to ICG Liposomes (complete titration data set in Figure S4). The BSA addition induced a decrease in ICG absorptivity and a blue-shift in ICG fluorescence maxima along with deceased fluorescence brightness. A comparison with the reference spectra in Figure 2a–b suggests that a large fraction of the ICG transfers from the liposomes to the added BSA, and independent experimental proof for this conclusion is provided in the following sections.

Scheme 1.

Liposome compositions (molar %) for FRET Liposomes, (1) POPC:DiI₂₀₍₅₎:ICG (98:1:1), (2) POPC:DSPE-PEG₂₀₀₀:DiI₂₀₍₅₎:ICG (96:2:1:1), (3) POPC:DSPE-PEG₂₀₀₀:DOTAP:DiI₂₀₍₅₎:ICG (92:2:4:1:1), (4) POPC:DSPE-PEG₂₀₀₀:Cholesterol:DiI₂₀₍₅₎:ICG (94:2:2:1:1), (5) POPC:DSPE-PEG₂₀₀₀:Cholesterol:DiI₂₀₍₅₎:ICG (56:2:40:1:1), (6) POPC:DSPE-PEG₂₀₀₀:Cholesterol: DOTAP:DiI₂₀₍₅₎:ICG (90:2:2:4:1:1), (7) POPC:DSPE-PEG₂₀₀₀:Cholesterol: α-tocopherol:DiI₂₀₍₅₎:ICG (74:2:2:20:1:1), (8) POPC:DSPE-PEG₂₀₀₀:Cholesterol: α-tocopherol:DiI₂₀₍₅₎:ICG (56:2:20:20:1:1); ICG Liposomes, (2’) POPC:DSPE-PEG₂₀₀₀:ICG (97:2:1), (5’) POPC:DSPE-PEG₂₀₀₀:Cholesterol:ICG (57:2:40:1), (7’) POPC:DSPE-PEG₂₀₀₀:Cholesterol: α-tocopherol:ICG (75:2:2:20:1), (8’) POPC:DSPE-PEG₂₀₀₀:Cholesterol: α-tocopherol:ICG (56:2:20:20:1), (9’) POPC:DSPE-PEG₂₀₀₀:α-tocopherol:ICG (87:2:10:1), (10’) POPC:DSPE-PEG₂₀₀₀: α-tocopherol:ICG (77:2:20:1); Thermosensitive FRET Liposome (1_t’) DPPC:DiI₂₀₍₅₎:ICG (98:1:1); and Thermosensitive ICG Liposomes, (2_t”) DPPC:DSPE-PEG₂₀₀₀ (98:2), (2_t’) DPPC:DSPE-PEG₂₀₀₀:ICG (97:2:1), (5_t’) DPPC:DSPE-PEG₂₀₀₀:Cholesterol:ICG (57:2:40:1), (7_t’) DPPC:DSPE-PEG₂₀₀₀:Cholesterol: α-tocopherol : ICG (75:2:2:20:1). Insert shows the spectral overlap of liposomal DiI₂₀₍₅₎ as a FRET donor (ex: 640 nm, em: 665 nm) and liposomal ICG as a FRET acceptor (ex: 780 nm, em: 840 nm).

Table 1.

Dynamic Light Scattering data. Z-average and the Polydispersity Index of the different compositions of FRET Liposomes and ICG Liposomes. Each value represents the average of triplicate experiments ± SD. Liposome compositions are listed in Scheme 1.

Lipid Composition	Z- average (nm)	Polydispersity index	Lipid Composition	Z- average (nm)	Polydispersity index
1	134.2 ± 5.6	0.12 ± 0.01	8	179.0 ± 5.7	0.22 ± 0.06
2	135.3 ± 4.9	0.15 ± 0.02	2’	160.0 ± 3.2	0.24 ± 0.01
3	140.0 ± 3.8	0.14 ± 0.01	5’	155.8 ± 4.6	0.25 ± 0.01
4	148.4 ± 1.5	0.12 ± 0.03	7’	160.4 ± 3.0	0.14 ± 0.05
5	141.8 ± 1.5	0.14 ± 0.01	8’	166.9 ± 4.5	0.10 ± 0.01
6	153.5 ± 4.5	0.19 ± 0.03	9’	170.2 ± 6.6	0.12 ± 0.01
7	156.0 ± 3.0	0.13 ± 0.03	10’	172.7 ± 3.5	0.11 ± 0.02

Figure 2.

Normalized (a) absorption and (b) fluorescence (ex: 780 nm) spectra for free ICG (10 μM) in buffer, ICG + BSA (4.59 μM) in buffer, ICG Liposomes in buffer. Normalized (c) absorption and (d) fluorescence (ex: 780 nm) spectra for ICG Liposomes in buffer with different concentrations of BSA. In all cases, the buffer was HEPES (20 mM, 100 mM NaCl, pH 7.4) and the ICG Liposomes were composition 2’ (POPC:DSPE-PEG₂₀₀₀:ICG, 97:2:1, 50 μM total lipid). T = 22 °C.

Studies that Treat FRET Liposomes with BSA

FRET is very sensitive to changes in dye separation when the distances are in the range of 1–10 nm, and biomembrane studies often use FRET to monitor changes in membrane structure and integrity.⁴³ Quite a few studies have employed lipophilic cyanine dyes as a donor/acceptor FRET pair embedded within lipid based nanocarriers,^{44,45,46,47,48} including liposomes.⁴⁹ But there appears to be no reported liposome studies that specifically use ICG as the FRET acceptor.^{10,44,45,50,51,52} As illustrated by boxed insert within Scheme 1, there is considerable spectral overlap of liposomal DiI₂₀₍₅₎ emission as a FRET donor and liposomal ICG absorbance as a FRET acceptor (ex: 780 nm, em: 840 nm). Indeed, a set of comparative fluorescence studies (Figure S5) found that liposomes containing a binary mixture of 1% DiI₂₀₍₅₎ and 1% ICG exhibited strongly quenched DiI₂₀₍₅₎ fluorescence indicating FRET. But there was only a very small increase in the ICG emission intensity, suggesting that some fraction of the ICG forms non-fluorescent ground-state complexes within the liposome membrane (i.e., static quenching of ICG).^44,53,54 Regardless of the exact fluorescence quenching mechanism(s), the large change in the fluorescence intensity ratio at 665/840 nm for liposomes containing DiI₂₀₍₅₎:ICG (which we call FRET Liposomes) makes it experimentally useful as a sensitive ratiometric indicator of spatial proximity between the two dyes. DiI₂₀₍₅₎ is sometimes called DiD in the literature and two key points concerning its chemical structure are its positive charge and its extremely high lipophilicity. These two factors combine to ensure that the DiI₂₀₍₅₎ is strongly retained in the anionic liposome membranes even in the presence of BSA.^55,56,57 Thus, the FRET Liposome experiment illustrated in Figure 3a was designed to produce a decrease in FRET efficiency (i.e., an increase in 665/840 nm fluorescence intensity ratio) if there was molecular transfer of ICG from the liposomes to BSA residing in the external aqueous solution.

Figure 3.

(a) Schematic illustration of FRET experiment with FRET Liposomes of composition 1 (POPC:DiI₂₀₍₅₎:ICG, 98:1:1, 50 μM total lipid). (b) Time dependent change in absorption and fluorescence (ex: 640 nm) spectra for FRET Liposomes in buffer. (c) Time dependent change in absorption and fluorescence (ex: 640 nm) spectra for FRET Liposomes in buffer + BSA (0.25 mM). (d) Plot of normalized FRET efficiency against time. Data is the average of duplicate experiments where error bars represent ± SD. In all cases, the buffer was HEPES (20 mM, 100 mM NaCl, pH 7.4) and T = 22 °C.

In Figure 3b are representative absorbance and fluorescence spectra for an aliquot of FRET Liposomes that was simply diluted in buffer and monitored over 170 minutes. There was very little change in the absorbance or fluorescence profile for the DiI₂₀₍₅₎:ICG pair suggesting negligible dissociation of the ICG from the diluted liposomes. In contrast, there was a substantial ratiometric change over 170 minutes when an identical aliquot of FRET liposomes was diluted into buffer containing BSA (Figure 3c). Specifically, there was a large increase in DiI₂₀₍₅₎ fluorescence intensity at 665 nm suggesting greatly decreased quenching by the ICG as FRET acceptor (see Figure S6 for additional analysis of this specific experiment). Plotted in Figure 3d is a comparison of the changes in normalized FRET efficiency (calculated using the equation in the experimental section) in the presence and absence of BSA. The curve exhibits a half-life of ~30 minutes at 22 °C for spontaneous transfer of ICG from the liposomes to the albumin protein residing in the external solution.

For experimental convenience, all the following studies were conducted at 22 °C, but a repeat of this ICG transfer experiment at the physiologically relevant temperature of 37 °C determined the half-life to ~ 10 minutes (Figure S8). This short time frame for ICG transfer is experimentally significant since many in vivo imaging studies of ICG-containing liposomes extend well beyond 10 minutes.

The FRET Liposome assay was used to screen eight different liposome compositions that are summarized in Scheme 1. In each case, the change in normalized FRET efficiency over 170 minutes was determined for liposomes diluted in buffer and in buffer + BSA. The individual spectral profiles are provided in Figures S7, S9, S10–S14, and S16 and the normalized FRET efficiency values for liposomes in buffer + BSA at time = 0 and 170 minutes are summarized as a bar graph in Figure 4. The data reflects the comparative capacity of a liposome additive to inhibit the transfer of ICG from liposomes to BSA. Notably, the presence of 2% PEGylated DSPE-PEG₂₀₀₀ (composition 2), 4% cationic DOTAP (composition 3),²⁷ or 40 % cholesterol (composition 5) in the liposomes had negligible effect on the extent of ICG transfer to the added BSA. Interestingly, ICG transfer to BSA was greatly inhibited when the liposome membranes included 20 % α-tocopherol (composition 7).⁵⁸

Figure 4.

Normalized FRET efficiency for FRET Liposomes at t = 0 minutes and t = 170 minutes after separate liposome samples were added to buffer + BSA. Data is the average of duplicate experiments where error bars represent ± SD. T = 22 °C. Liposome compositions are listed in Scheme 1.

Centrifugation Studies

To independently prove that the ICG:BSA complex was spatially distant from the liposomes, we devised a centrifugation experiment that separated the liposomes from the BSA and enabled spectral analysis of each separated fraction. As illustrated in Figure 5a, the centrifugation experiment used high density liposomes filled with sucrose which promoted liposome sedimentation as a pellet.⁵⁹ Centrifugation at 10,000 g for 45 minutes created a pellet containing a large fraction (but not all) of the liposomes with virtually all of the BSA (MW = 67 kDa) in the supernatant. The first set of centrifugation experiments examined ICG Liposomes with compositions 2’ (i.e., POPC liposomes containing 2% DSPE-PEG₂₀₀₀), 5’ (i.e., same liposomes also containing 40% cholesterol), or 7’ (i.e., same liposomes also containing 20% α-tocopherol). For each liposome composition, two identical samples of the liposomes were incubated for 3 hours, one in buffer and the other in buffer + BSA. In each case, the sample was subsequently centrifuged, the liposome pellet was removed, and a fluorescence spectrum of the supernatant was acquired. In the case of ICG Liposomes with composition 2’ or 5’ the presence of BSA in the supernatant raised the ICG fluorescence intensity by more than a factor of two, indicating extensive transfer of ICG from the liposomes to the BSA. In the homologous case of ICG Liposomes with composition 7’ the presence of BSA in the supernatant did not change the ICG fluorescence intensity in the supernatant, consistent with the conclusion that the α-tocopherol in the liposomes strongly inhibited ICG transfer from the liposomes to the BSA (Figure 5b and S18).

Figure 5.

(a) Summary of the centrifugation experiment. High density, sucrose loaded ICG Liposomes or FRET Liposomes were centrifuged to produce a pellet containing a large fraction of the liposomes and a supernatant containing all the BSA and the remaining fraction of liposomes. (b) Normalized fluorescence of the supernatant for four separate experiments using ICG Liposomes. (c) Relative FRET efficiency of the supernatant for four separate experiments using FRET Liposomes. In all cases, the buffer was HEPES (20 mM, 100 mM NaCl, pH 7.4; or the same buffer + 0.25 mM BSA) and the initial total lipid composition was 100 μM. Data is the average of duplicate experiments where error bars represent ± SD. ICG Liposome compositions are 2’ (POPC:DSPE-PEG₂₀₀₀:ICG, 96:2:1), 5’ (POPC:DSPE-PEG₂₀₀₀:Cholesterol:ICG, 57:2:40:1), or 7’ (POPC:DSPE-PEG₂₀₀₀:Cholesterol: α-tocopherol:ICG, 74:2:2:20:1). FRET Liposome compositions are 2 (POPC:DSPE-PEG₂₀₀₀:DiI₂₀₍₅₎:ICG, 96:2:1:1), 5 (POPC:DSPE-PEG₂₀₀₀:Cholesterol:DiI₂₀₍₅₎:ICG, 56:2:40:1:1), or 7 (POPC:DSPE-PEG₂₀₀₀:Cholesterol: α-tocopherol:DiI₂₀₍₅₎:ICG, 74:2:2:20:1:1).

Additional support for these conclusions was gained by repeating the centrifugation experiment with high density FRET Liposomes with compositions 2 (i.e., POPC liposomes containing 2% DSPE-PEG₂₀₀₀), 5 (i.e., same liposomes also containing 40% cholesterol), or 7 (i.e., same liposomes also containing 20% α-tocopherol). For each liposome composition, two identical samples of the liposomes were incubated for 3 hours, one in buffer and the other in buffer + BSA. After centrifugation and removal of the pellet (which contained most of the liposomes), the fluorescence spectrum of the supernatant was acquired with 640 nm excitation and the relative FRET efficiency was determined (calculated according to the equation in the experimental section). The bar graph in Figure 5c (spectral data in Figure S19) shows that the relative FRET efficiency for FRET Liposomes with composition 2 (i.e., POPC liposomes containing 2% DSPE-PEG₂₀₀₀) or 5 (i.e., same liposomes also containing 40% cholesterol), was greatly reduced when the supernatant contained BSA, indicating that a significant amount of ICG had transferred from the FRET Liposomes to the BSA. In comparison, the homologous experiment using FRET liposomes with composition 7 (i.e., same liposomes also containing 20% α-tocopherol) produced very little change in relative FRET efficiency indicating very little ICG transfer from the FRET liposomes to the BSA.

α-Tocopherol, one of the vitamin E group of naturally-occurring molecules, is an amphiphilic compound and its location within bilayer membranes has been studied extensively by different experimental methods and molecular dynamics simulations.^60,61,62,63 It is known that α-tocopherol has cholesterol-like capacity to act as a membrane plasticizer and depress or eliminate the gel-to-fluid transition within bilayer membranes.⁶⁰ The new finding of this current study is that α-tocopherol has much greater capacity than cholesterol to inhibit ICG transfer from liposomes to BSA in the external solution. Consider the experiment described in the preceding paragraph; that is, addition of BSA to FRET liposomes with composition 7 (i.e., liposomes containing ICG and 20% α-tocopherol). Even though α-tocopherol has moderate affinity for BSA (association constant is 7 × 10⁵ M⁻¹)^64,65 it is extremely lipophilic (logP of 12.2)⁶⁶ and so does not transfer from liposomes to the added BSA. Rather, it remains in the liposome membrane and helps to retain the ICG by providing stabilizing interactions. A number of literature studies have investigated non-covalent association of α-tocopherol with various biomembrane amphiphiles, and several different supramolecular interactions have been identified.^60,67,68,69 Since aromatic rings are a common structural feature within α-tocopherol and ICG (and absent in cholesterol), we suggest that α-tocopherol helps to retain ICG within the membrane by forming favorable intermolecular aromatic stacking interactions (π-π interactions).^70,71 The capabilities of both molecules to engage in aromatic stacking within lipophilic microenvironments is well-documented.^4,72,73

ICG Liposome Photobleaching Studies

α-Tocopherol is a naturally-occurring inhibitor of lipid oxidation,⁷⁴ and a scavenger of photogenerated singlet oxygen,^75,76,77 but it has apparently not been previously investigated as an additive to inhibit photobleaching of liposome-bound near-infrared dyes such ICG.^13,78,79 This is surprising since the photochemical instability of ICG has been known for long time.^14,15,80,33 We conducted a photobleaching experiment that irradiated a comparative set of ICG Liposomes containing three different amounts of α-tocopherol (0 %, 10 %, or 20 %) and monitored the change in the intensity of ICG absorption and fluorescence signal. The data in Figure 6 (spectral data in Figure S20) shows that ICG photobleaching was substantially attenuated when the liposomes contained 10 % or 20% α-tocopherol (compositions 9’ or 10’). This result suggests that α-tocopherol might be broadly useful as a dye-photostabilizing additive within many types of fluorescent liposomes and related self-assembled nanoparticles.¹³

Figure 6.

Normalized photobleaching profiles of ICG Liposomes with different molar % of α-tocopherol. (a) changes in absorption and (b) emission (ex: 780 nm) intensity for three different ICG Liposomes containing 0%, 10% or 20% α-tocopherol irradiated for 50 minutes with a Xenon Arc lamp (150 W) equipped with a 620 nm long pass filter. In all cases, the buffer was HEPES (20 mM, 100 mM NaCl, pH 7.4) and the ICG Liposomes were 50 μM total lipid with composition 2’ (POPC:DSPE-PEG₂₀₀₀:ICG, 97:2:1), composition 9’ (POPC:DSPE-PEG₂₀₀₀:α-tocopherol:ICG (87:2:10:1), or composition 10’ (POPC:DSPE-PEG₂₀₀₀: α-tocopherol:ICG, 77:2:20:1).

Light-Triggered Leakage from ICG Liposomes

Although this study focused on ICG Liposomes, we note that ICG transfer to albumin has been observed for a related solid nanoparticle formulation¹³ and thus is likely to occur with many types of nanoscale ICG formulations.^10,11 With regard phototherapeutics, ICG nanoparticles have been investigated for photothermal therapy, ^26,27 photodynamic therapy,^{12, 81} and light triggered drug release.^10,11 There is literature evidence that the geometric location of a photosensitizer relative to a liposome membrane (i.e., photosensitizer embedded in the membrane or protruding into the aqueous solution) can affect the rate and extent of light-triggered liposome leakage,⁸² and we were curious to learn if ICG transfer to albumin would diminish the extent of NIR light activated membrane disruption and subsequent contents leakage. Therefore, we conducted experiments that measured NIR light triggered leakage of aqueous contents from 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) liposomes that also contained ICG as the light absorbing trigger.^{16,17,18,24,25,30,32,83} DPPC liposomes are thermally responsive and they leak when heated close to their transition temperature of ~41 °C. Not only does the embedded ICG provide a source of photothermal heating when irradiated with NIR light, it also acts as a photosensitizer to generate reactive oxygen species that can induce phospholipid oxidation and increased liposome membrane permeability.^{12,26,82,84,85}

First, we confirmed that BSA extracted ICG from thermosensitive FRET Liposomes, composed of DPPC:DiI₂₀₍₅₎:ICG, 98:1:1 (composition 1’_t) with a half-life of ~3 minutes at 22 °C (Figure S17). Then, we prepared thermosensitive ICG Liposomes that were comprised of DPPC with 1 mol% of ICG (the added ICG does not alter the DPPC phase transition temperature¹⁶) and conducted light triggered leakage of the encapsulated fluorescent dye calcein as a measure of aqueous contents release (Figure 7a). The experimental protocol warmed the liposomes to a starting temperature of 37°C and then irradiated them with light from a 785 nm diode laser. The first set of experiments compared the calcein leakage from the thermosensitive ICG Liposomes (liposome composition 2_t’ or 5 _t’) that were in buffer or in buffer + BSA. As shown in Figure 7b, light induced leakage of calcein from the liposomes in buffer + BSA was only 60% of the calcein leakage observed when the same liposomes were in buffer. When the experiment was repeated with thermosensitive ICG Liposomes that also contained 20% α-tocopherol (liposome composition 7_t’), there was no analogous decrease in leakage when liposomes were irradiated in buffer + BSA (spectral data given in Figure S21). These results are consistent with a picture of attenuated light-triggered liposome leakage due to transfer of the light absorbing ICG from the thermosensitive ICG Liposomes (composition 2_t’ or 5 _t’) to the BSA residing in the external solution. Moreover, this deleterious effect could be prevented when the liposomes contained α-tocopherol (composition 7_t’) which helps retain the ICG in the liposome membrane. This α-tocopherol additive approach to optimizing thermosensitive ICG Liposomes is complementary to the literature idea of coating the liposomes with a surrounding layer of biopolymer such as hyaluronic acid¹⁷ or chitosan.²⁰

Figure 7.

Light triggered liposome release assay. (a) Light triggered release of the encapsulated fluorescent dye, calcein, from thermosensitive ICG Liposomes induced by 785 nm laser irradiation of separate samples, each with starting temperature of 37 °C. (b) Percent calcein release of different thermosensitive ICG Liposomes in either buffer or buffer + BSA (0.25 mM) after 785 nm laser irradiation (10 W/cm²). In all cases, the buffer was HEPES (20 mM, 100 mM NaCl, pH 7.4) and ICG Liposomes were 50 μM total lipid with composition 2_t” (DPPC:DSPE-PEG₂₀₀₀, 98:2), composition 2_t’ (DPPC:DSPE-PEG₂₀₀₀:ICG, 96:2:1), composition 5_t’ (DPPC:DSPE-PEG₂₀₀₀:Cholesterol:ICG, 57:2:40:1), or composition 7_t’ (DPPC:DSPE-PEG₂₀₀₀:Cholesterol: α-tocopherol:ICG, 74:2:2:20:1). Bar graphs indicate average of duplicate experiments where error bars represent ± SD.

CONCLUSIONS

The major findings of this study are:

1. The ICG embedded within liposomes (ICG Liposomes) transfers spontaneously to albumin protein residing in the external solution. The half-life for ICG transfer from fluid phase POPC liposomes is ~30 minutes at 22 °C, and ~10 minutes at 37°. The half-life for ICG transfer from gel phase DPPC liposomes is ~3 minutes at 22 °C. This ICG transfer process has been observed with lipid core nanoparticles^13,57 but it seems to have not been considered by most published studies (perhaps all of them) that use ICG liposomes. ^{11,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37}

2. Unlike ICG, the classic lipophilic cyanine dyes such as DiI₂₀₍₅₎ with two long hydrocarbon chains do not transfer from liposomes, or related lipid core nanoparticles, to external BSA; ^55,56,57 thus, it is not accurate to assume that ICG exhibits the same phase transfer behavior as the classic lipophilic cyanine dyes.⁴⁸

3. The presence of 20 % α-tocopherol in ICG Liposomes helps retain the ICG in the membrane (presumably by forming favorable aromatic stacking interactions), and inhibits ICG photobleaching. It is worth noting that the maximum liposome loading capacity is approximately 33% α-tocopherol.⁸⁶

4. Transfer of ICG from DPPC liposome membranes to external albumin reduces the extent of light-triggered leakage from these thermosensitive ICG Liposomes.

The capacity of blood proteins, such as albumin, to extract ICG from liposomes raises uncertainty concerning the integrity of ICG Liposomes for in vivo imaging and phototherapeutics. For example, NIR fluorescence images of biodistribution may not specifically indicate the location of ICG Liposomes but might also include the location of ICG:albumin complex.⁵⁷ Moreover, imaging artifacts might arise due to variation in albumin levels with anatomical position, or by albumin fluctuations caused by changes in disease state.³⁸ Our results suggest that fluorescence imaging performance with fluid phase ICG Liposomes can be enhanced by including clinically approved α-tocopherol as a liposome additive that will reduce ICG transfer and inhibit ICG photobleaching. Further studies are needed to determine if spontaneous ICG transfer to albumin occurs when the ICG is embedded in gel phase liposomes with high transition temperatures,^53,54,87 or non-phospholipid vesicles.^88,89,90

With regard phototherapeutics using thermosensitive ICG Liposomes, our light-triggered liposome release results highlight a key concept underlying the efficiency of light-sensitive liposomes. That is, liposome release is more effective when the light absorbing dye (ICG in this case) is embedded within or located very close to the liposome membrane.^91,89 While most of the literature on light-triggered leakage of ICG-DPPC liposomes focuses on photothermal heating, there is also a significant photodynamic chemistry pathway (as reflected by the ICG photobleaching in Figure 6) that undoubtedly disrupts membrane integrity.^{12,26,81,83,92} The impact of both membrane disruption effects (photothermal and photodynamic) depends on the physical distance between the ICG and the liposome membrane; thus, transfer of the ICG from the liposome bilayer to the albumin residing in external solution increases the distance and attenuates the extent of membrane disruption.^82,93,94 It will be important to learn if other classes of NIR dyes undergo spontaneous transfer from liposomes and related nanoparticles to blood proteins, especially the newly developed amphiphilic dyes that exhibit absorption/emission wavelengths in the SWIR range of 900 – 1700 nm.^95,96

EXPERIMENTAL

General Materials

Reagents and solvents were purchased from Sigma Aldrich, Avanti Polar Lipids, Molecular Targeting Technologies and used without further purification unless stated otherwise. Lipid/additive stock solutions are as follows (stored at −20 °C unless stated otherwise): 25 mg/mL 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) in chloroform, 25 mg/mL 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) in chloroform, 25 mg/mL 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycol)-2000] (ammonium salt) (DSPE-PEG₂₀₀₀) in chloroform, 10 mg/mL 1,2-dioleoyl-3-trimethylammonium-propane (chloride salt) (DOTAP) in chloroform, 10 mg/mL Cholesterol in chloroform, 25 mg/mL α-tocopherol in chloroform (stored at 2–8 °C). BSA Stock: 1 mg/mL BSA in doubly distilled water (stored at 2–8 °C). Dye Stocks (stored at −20 °C): 1 mg/mL DiIC₂₀₍₅₎ (iodide salt) in DMSO, 1 mg/mL ICG in methanol. Buffers (stored at room temperature): HEPES buffer (20 mM, 100 mM NaCl, pH 7.4), BSA-containing HEPES buffer (0.25 mM BSA, 20 mM HEPES, 100 mM NaCl, pH 7.4), sucrose-containing HEPES buffer (180 mM, 5 mM HEPES, 20 mM NaCl, pH 7.4), calcein-containing HEPES buffer (0.6 mM calcein, 5 mM HEPES, 20 mM NaCl, pH 7.4). All absorption and fluorescence spectra were acquired using an Evolution 201 UV-Vis Spectrometer (Thermo Fischer Scientific, WA, USA) and Fluoromax Plus C (Horiba Instruments, Kyoto, Japan), respectively.

Octanol Partition Experiments

A biphasic mixture of either 1-octanol (0.50 mL) and HEPES buffer (0.50 mL, 50 mM, 100 mM NaCl, pH 7.4) or 1-octanol (0.50 mL) and HEPES buffer supplemented with 0.25 mM BSA (0.50 mL) was charged with a 5 μL aliquot of ICG stock solution (1 mM in methanol). If all the ICG was in the octanol layer, the concentration would be 10 μM. Each biphasic mixture was vigorously shaken for 15 seconds and centrifuged at 1000 g for 2 minutes to cleanly separate the two phases. After obtaining color photographs, a 250 μL aliquot form each phase was transferred to a 96-well plate, and the absorbance at 780 nm was measured. LogP was determined using the formula P = log(A_octanol/A_buffer) where A_octanol = absorbance of ICG in octanol phase and A_buffer = absorbance of ICG in buffer phase at 780 nm. All measurements were performed in triplicate.

Liposome Preparation

FRET Liposomes and ICG Liposomes.

Stock solutions of FRET Liposomes with compositions 1, 2, 3, 4, 5, 6, 7, or 8 and ICG Liposomes with compositions 2’, 5’, 7’, 8’, 9’, or 10’ were prepared using thin-film hydration method followed by mechanical extrusion. The desired lipid-additive-dye formulations were mixed in a test tube. Organic solvents of the mixture were removed using a high vacuum system overnight until a thin lipid film was achieved. Lipid films were rehydrated with HEPES buffer (20 mM, 100 mM NaCl, pH 7.4) to give a 1 mM total lipid concentration. The hydrated film was subjected to four freeze/thaw cycles which consisted of freezing in liquid nitrogen and melting in a water bath at a temperature above the melting temperature of the phospholipids in the mixture. The resulting multilamellar vesicles were extruded 21 times through a polycarbonate membrane (200 nm pore size, 19 mm diameter, Whatman Nucleopore Track-Etched Membranes) using a mini-extruder (Avanti Polar Lipids, Alabama, USA) at room temperature to produce a unilamellar vesicle suspension.

Sucrose-loaded liposomes.

Stock solutions of sucrose-loaded FRET Liposomes with compositions 2, 5, or 7 and sucrose-loaded ICG Liposomes with compositions 2’, 5’, or 7’ were prepared using the above procedure, except a sucrose-containing HEPES buffer (180 mM, 5 mM HEPES, 20 mM Nacl, pH 7.4) was used for lipid thin-film rehydration. To remove the unincorporated sucrose and dye molecules, the liposomes were filtered were passed through a Sephadex G-25 Superfine Gel Column (PD-10, GE HealthCare, WI, USA) using HEPES (20 mM, 100 mM NaCl, pH 7.4) buffer as the eluent.

Calcein-loaded liposomes.

Stock solutions of calcein-loaded thermosensitive ICG Liposomes with compositions 2_t”, 2_t’, 5_t’, or 7_t’ were prepared using the above procedure. The lipid films were rehydrated with a calcein-containing HEPES buffer (60 mM calcein, 5 mM HEPES, 20 mM NaCl, pH 7.4) that was pre-heated to ~60 °C giving a 1 mM total lipid concentration. The hydrated film was further incubated for 30 minutes in a water bath heated to ~60 °C by gently stirring the test tube at every 10 minutes. The formulation was subjected to four freeze/thaw cycles which consist of freezing in liquid nitrogen and melting in a water bath at ~60 °C respectfully. The resulting multilamellar vesicles were further incubated for additional 15 minutes in the ~60 °C water bath and extruded 21 times through a polycarbonate membrane (200 nm pore size, 19 mm diameter, Whatman Nucleopore Track-Etched Membranes) using a mini-extruder (Avanti Polar Lipids, Alabama, USA) at ~60 °C to produce a suspension of unilamellar vesicles. The resulting formulation was quickly cooled down to room temperature and passed through a Spehadex G-25 Superfine Gel Column (PD-10, GE HealthCare, WI, USA) using HEPES (20 mM, 100 mM NaCl, pH 7.4) buffer as the eluent. All liposome formulations were stored in the dark at 2–8 °C.

Dynamic Light Scattering Studies

Dynamic Light Scattering experiments were carried out using a Zetasizer Nano ZS (Malvern Panalytical, Malvern, UK) to obtain the hydrodynamic diameters and the polydispersity index of the liposomes at pH 7.4 (He-Ne laser wavelength at 633 nm). Measurements were obtained in a low volume disposable sizing cuvette (1 mL, 1 cm path length) with filtered HEPES buffer (20 mM, 100 mM NaCl, pH 7.4, 25 mm syringe filter with 0.45 μM Acrodisc Supor membrane, Pall Corporation, NY, USA) at a backscattering angle of 173° and a temperature of 25 ° C. The reported data is the average of three replicates for each liposome formulation.

Studies Using ICG Liposomes

Normalized absorption and emission (ex: 780 nm) spectra of 10 μM ICG in HEPES buffer (20 mM, 100 mM NaCl, pH 7.4), 10 μM ICG + 4.09 μM BSA in HEPES buffer, and ICG Liposomes with composition 2’ (50 μM total lipid) in HEPES buffer were obtained in a quartz cuvette (1 mL). For BSA titrations, different BSA concentrations (0.06 μM-4.09 μM) were added to a solution of ICG Liposomes with composition 2’ (50 μM total lipid) in HEPES buffer. Absorption and emission (excitation wavelength: 780 nm, slit width: 3 nm) spectra of ICG liposomes were recorded 3 minutes after each BSA addition to allow for equilibration. In all cases, T = 22 °C.

Studies Using FRET Liposomes

Absorption and emission (ex: 640 nm) spectra of FRET liposomes with compositions 1, 2, 3, 4, 5, 6, 7, or 8 (50 μM total lipid) in HEPES buffer (20 mM HEPES, 100 mM NaCl, pH 7.4) were obtained in a quartz cuvette (1 mL, 1 cm path length). Absorption and emission were recorded every 10 minutes for 170 minutes. A second data set was acquired using a HEPES buffer supplemented with BSA (0.25 mM). Normalized FRET efficiency at each time point was calculated using the following equation (Normalized FRET Efficiency = (F_t^{800 nm}/F_t^{700 nm})/(F₀^{800 nm}/F₀^{700 nm}) where F₀^{800 nm} is the acceptor emission and F₀^{700 nm} is the donor emission at t = 0 minutes. F_t^{800 nm} is the acceptor emission and F_t^{700 nm} is the donor emission at the time point. FRET efficiencies were normalized to t = 0 minutes for buffer and buffer + BSA separately. Relative FRET efficiency was calculated at t = 0 minutes and t = 170 minutes using the following equation (Relative FRET efficiency = F _t^{800 nm}/(F_t ^{700 nm} + F_t ^{800 nm}) where F_t^{800 nm} is the acceptor emission and F_t^{700 nm} is the donor emission at t = 0 minutes and t = 170 minutes. All experiments were conducted in duplicate and T = 22 °C. For compositions 1 and 7, the emission (ex: 640 nm) spectra were repeated at T = 37 °C.

Centrifugation Studies

Separate 500 μL solutions of sucrose-loaded ICG Liposomes with compositions 2’, 5’, or 7’ (100 μM total lipid) in HEPES buffer (20 mM HEPES, 100 mM NaCl, pH 7.4) were loaded in a 1.5 mL Eppendorf tube and incubated at 25 °C for 5 minutes, then centrifuged at 10,000g for 45 minutes using a LSE high-speed microcentrifuge (Corning Life Sciences, NC, USA). The resulting supernatant was separated from the pellet and transferred into an Eppendorf tube (0.5 mL). A second data set was acquired using HEPES buffer supplemented with BSA (0.25 mM). For analysis, each fluorescence emission reading at 833 nm was normalized to the liposomes with composition 2’ in HEPES buffer. The entire procedure was repeated using sucrose-loaded FRET Liposomes with compositions 2, 5, or 7, and ex: 640 nm for the fluorescence acquisition. For analysis, relative FRET for each formulation was calculated using the following equation (Relative FRET = F^{800 nm}/(F ^{700 nm} + F ^{800 nm}) where F^{800 nm} is the acceptor emission and F^{700 nm} is the donor emission. All experiments were conducted in duplicate and T = 22 °C.

Photobleaching Studies

Photobleaching studies used ICG Liposomes with compositions 2’, 9’, or 10’ containing 0, 10 or 20 molar % of α-tocopherol, respectively. The liposomes were prepared using the above film hydration procedure and 1 mL of each formulation (50 μM total lipid) in HEPES buffer (20 mM HEPES, 100 mM NaCl, pH 7.4) was placed in a quartz cuvette (1 cm path length). While exposing to air, the cuvette was continuously irradiated for 50 minutes at a distance of 5 cm using a 150 W Xenon Arch Lamp equipped with a 620 nm long-pass filter. Absorption and emission (excitation wavelength: 780 nm, slit width: 3 nm) spectra were recorded at different time points up to 50 minutes. Normalized maximum absorbance and emission intensities were plotted against the time.

Calcein Release Studies

Light triggered release from calcein-loaded thermoresponsive ICG Liposomes with compositions 2_t”, 2_t’, 5_t’, or 7_t’ was measured by monitoring the increase in calcein fluorescence. Separate 1 mL dispersions of each calcein-loaded ICG Liposome formulation (50 μM total lipid) in HEPES buffer (20 mM HEPES, 100 mM NaCl, pH 7.4) were placed in a quartz cuvette and pre-heated to ~37 °C in the dark. Maintaining the cuvette at ~37 °C, the solution was irradiated from above at a 3 cm distance with a 785 nm diode laser (10 W/cm², Thorlabs, NJ, USA). A calcein fluorescence spectrum was recorded at 515 nm (ex: 490 nm) before and after irradiation of each liposome formulation. Finally, 1% (w/v) Triton X was added to lyse the liposomes and completely release the encapsulated calcein. A preliminary set of experiments with the composition 2_t’ determined that calcein leakage increased with irradiation time up to a period of 2 minutes where it was >60%, thus 2 minutes was chosen as standard irradiation time for experiments that compared compositions 2_t”, 2_t’, 5_t’, and 7_t’. A second data set was acquired using HEPES buffer supplemented with BSA (0.25 mM). The percent calcein release was calculated using the following equation. (% calcein release. = [(F_after−F_before)/(F−F_before)]*100) where F_after is the emission after irradiation. F_before is the emission before irradiation, and F is the emission after Triton X addition. The temperature increase of the cuvette during the 2 minutes irradiation was recorded using a thermocouple (OMEGA, CT, USA) every 2 sec. All experiments were conducted in duplicate.

ABBREVIATIONS

ICG

Indocyanine Green

NIR

near-infrared

SWIR

short wave infrared

BSA

bovine serum albumin

FRET

Förster Resonance Energy Transfer

POPC

1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine

DPPC

1,2-dipalmitoyl-sn-glycero-3-phosphocholine

DSPE-PEG₂₀₀₀

1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycol)-2000] (ammonium salt)

DOTAP

1,2-dioleoyl-3-trimethylammonium-propane (chloride salt)