2.1.

Materials

ICG is a kind gift from Serb Laboratories. Suppocire NC™ is purchased from Gattefossé (Saint-Priest, France). Myrj™ 52, polyethylene glycol 40 stearate, and Super Reffined Soybean Oil are from Croda Uniqema (Chocques, France). QTracker^®705, DiO, DiI, DiD, and DiR dyes (Fig. 1) are purchased from Invitrogen. The organic dyes are purified by column chromatography on silica gel, using cyclohexane/ethyl acetate gradient for elution, before use. Lipoid S75 (soybean lecithin at >75% phosphatidylcholine) is purchased from Lipoid (Germany), and other chemical products are purchased from Sigma Aldrich (Saint-Quentin Fallavier, France).

Fig. 1

Lipidot structure. (a) Lipidots are dye-loaded oily droplets dispersed in aqueous buffer, whose diameter can be adjusted between 30 and 120 nm. Their fluorescent properties are conferred by lipophilic or amphiphilic cyanine dyes (b), encapsulated in the lipid core and/or the surfactant layer.

2.2.

Lipidot Preparation

80 μl of a 10 mM dye solution in CH₂Cl₂ (EtOH for ICG) are poured in a 5 ml vial. The solvent is then evaporated under vacuum before an oil premix is added. The amount of oil, Suppocire NC™, and lecithin used in the oil premix are, respectively, 25, 75, and 150 mg for formulation of 30 nm diameter lipidots (F30), 85, 255, and 65 mg for 50 nm (F50), 102, 308, and 50 mg for 80 nm (F80), 144, 431, and 50 mg for 100 nm (F100), and 150, 450, and 45 mg for 120 nm (F120). After homogeneization at 60°C, the continuous aqueous phase, composed of 350 mg of Myrj™ 52 for F30, 345 mg for F50, 300 mg for F80, 325 mg for F100, 215 mg for F120, and of the appropriate amount of aqueous medium (154 mM NaCl if not stated otherwise, qsp 2 ml), is introduced. The vial is placed in a 60°C water bath and the mixture is sonicated for 5 min using a VCX750 ultrasonic processor (power output 190 W, 3-mm probe diameter, Sonics). Lipidots are dialyzed overnight at room temperature against 1000 times their volume in the appropriate aqueous buffer (12–14,000 Da MW cut off membranes, ZelluTrans). Finally, the lipidot dispersion is filtered through a 0.22 μm Millipore membrane for sterilization before characterization and/or injection.

2.3.

Size and Zeta Potential Measurements

Dynamic light scattering is used to determine the particle hydrodynamic diameter and zeta potential (Zeta Sizer Nano ZS, Malvern Instrument). All lipidots are diluted to 2 mg/ml of lipids in sterile PBS 0.1X and are transferred in Zeta Sizer Nano cells (Malvern Instrument) before each measurement, performed in triplicate.

2.4.

Optical Properties

The absorbance and fluorescence measurements are performed in 1X PBS, respectively, using a Cary 300 Scan UV-Visible spectrophotometer (Varian) and a Perkin Elmer LS50B fluorimeter. Previous to calculations, scattering contribution is subtracted from the absorbance spectrum using the equation:

2.5.

Photobleaching

Fluorescence intensities of capillaries (Ringcaps, Herschmann, internal diameter 450 μm) containing DiD-lipidots (equivalent 17 μM of DiD), 17 μM of Cy5 or 1.3 μM QT705 in water are recorded using a homemade system adapted from the commercially available Aequoria™ system from Hamamatsu. The excitation device is composed of 10 light-emitting diode (LED) emitting at 633 nm (adapted from the LuxiFlux™ device available from Hamamatsu). It also includes an interference bandpass filter (470DF20 from Schott) for a light illumination power of 200 μW/cm². The filtered fluorescence signal (filter OG515, from Schott) is measured by a cooled CCD camera (Lumenera), placed at 160 mm from the imaging field, with 30 ms exposure time. The LuCam Capture^® software is used to drive the setup and for image processing. Fluorescence intensities are measured from the recorded images using the ImageJ^® software.

2.6.

Animal Experiments

All animal procedures are in compliance with the guidelines of the European Union (regulation n°86/609), taken in the French law (decree 87/848) regulating animal experimentation. All efforts are made to minimize animal suffering and to reduce the number of animals used. All animal manipulations are performed with sterile techniques and are approved by the Rhône-Alpes Animal Care and Use Committee (France).

2.7.

In Vivo Imaging

Two Nude mice (Charles River, France) receive a 10 μl intradermic injection of nanoparticles (46 dyes/particle) in each paw, resulting in doses of 4 pmol of DiO-lipidots (184 pmols of dyes), 2 pmol of DiI-lipidots (92 pmol of dyes), and 1 pmol of DiD-lipidots and ICG-lipidots (46 pmol of dyes), respectively, in the front left, front right, rear left, and rear right extremity. For DiD-lipidot/QTracker^®705 comparison, two Nude mice receive a 10 μl intradermic injection of 2 pmol of DiD-lipidots and 20 pmol of QTracker^®705, respectively, in the rear left and rear right extremity. In vivo images are recorded on an IVIS^® system (Xenogen) and monitored using LivingImage^® software. Mice are scanned using 10 excitation (435, 465, 500, 535, 575, 605, 640, 675, 710, and 745 nm) and four emission filters, respectively, centered on the emission wavelengths of DiO, DiI, DiD, and ICG. For each emission filter, three-components spectral unmixing is carried out using the images acquired at the different excitation wavelengths (two-components for the GFP filter).

2.8.

Culture of NIH-3T3 Murine Fibroblasts

In accordance with the International Organization for Standardization 10993, NIH-3T3 murine fibroblast cells are chosen to perform classical WST-1 cytotoxic assay. The cell line is purchased from the American Type Culture Collection. Experimental cultures are prepared from deep-frozen stock vials, and maintained in a sub-confluent state in culture Petri dish (113 cm²) under a humidified (90%) atmosphere of 95% air/5% CO₂ at 37 °C. Cells are grown in Dulbecco's modified Eagle's medium high glucose (Gibco) supplemented with 10% newborn calf serum (PAN Biotech Gmb) and 1% antibiotics (penicillin and streptomycin) (Gibco). Culture medium is changed every other day.

2.9.

Cell Viability Assay

Cells (5 × 10⁴ cells/ well) are seeded in 6-well plates (Nunc). After 24 h incubation at 37°C, different concentrations of nanoparticles (DiD-lipidots or QTracker^®705), from 1 to 75 nM, are added for 24 h to the culture medium followed by two washes. Each group has triplicate wells. Cytotoxicity is assessed 24 h following the nanoparticle treatment using WST-1 assay, analog to MTT assay. WST-1 reagent (Roche) is added (10%) to the culture medium and kept in the incubator for 4 h. Cells without nanoparticles and incubated with a solution of H₂O₂ 10 mM are used as negative and positive controls, respectively. Absorbance is then recorded at 450 nm (soluble formazan titration) and 690 nm (background substraction) using a microplate reader (Tecan). The absorbance difference (450 to 690 nm) is directly proportional to the number of viable cells.

2.J.

Statistical Analysis

Cell viability is expressed as a percent of viable cells calculated from the negative control. Data are compared among groups by one-way analysis of variance followed by Fisher's protected Least Significance Differences test.

3.1.

Design and Colloïdal Properties of Dye-Loaded Lipid Nanoparticles (Lipidots)

Complex mixtures of soluble and insoluble species, both in the lipid core [soybean oil (25% w/w) and Suppocire NC™ wax (75% w/w)] and the surfactant shell [soybean lecithin (phospholipids) and PEG-based co-surfactants (Myrj™ 52)] are used for lipidot design, using high-energy process to reach sub-100 nm diameter.^{22, 25} As previously described,²² the heterogeneity of the lipids and surfactants mixtures constituting the nanodroplet favors its colloidal stability. In this study, typical results are presented using 50 nm diameter lipidots, but other nanoparticle sizes can easily be achieved modulating the surfactant to lipid ratio.^{22, 25}

Different lipophilic (Di family, log P > 10, P: octanol/water partition coefficient) or amphiphilic (ICG, only near-infrared dye approved for human use until now, log P = 2.1) cyanine dyes can be loaded in the oily phase, before addition of the aqueous phase and particle processing by ultrasonication (Fig. 1). Lipidots were then purified using dialysis, eliminating any nonencapsulated dyes and surfactants. Dye loading efficiency was calculated comparing the dispersion absorbance and fluorescence before and after dialysis. The results for 50 nm lipid nanoparticles loaded with a local dye concentration of ≈ 1.2 mM are displayed in Table 1. All 4 highly lipophilic indocyanines (DiO, DiI, DiD, and DiR) display encapsulation efficiency above 90% (respectively, 93%, 96%, 95%, and 92%). A lower loading value is obtained for ICG (around 77%, final local concentration of 0.9 mM), probably due to the presence of sulfonate moieties rendering this fluorophore much more hydrophilic. The approximate number of fluorescent molecules per 50-nm diameter lipidot is estimated as 44 for hydrophobic cyanines, and 36 for ICG.

Table 1

Physicochemical characterization of 50 nm diameter lipidots.

Normalized emission and absorbance spectra of the different dye-loaded 50 nm diameter particles are displayed in Fig. 2. These spectra are similar to those obtained in organic solvent (CH₂Cl₂ for hydrophobic indocyanines, dimethyl sulfoxide for ICG) with a bathochromic shift inferior to 5 nm. Detailed spectral data are displayed in Table 1. The versatility of these formulations allows the production of nanoparticles with different sizes, by controlling the mass ratio of core lipids and surfactants.^{22, 25} Whatever the nanoparticle size, dye loading has no influence on the hydrodynamic diameter, size distribution, or zeta potential (Figs. 3 and 4). Similarly, the particle size has no effect on absorption and emission spectra of the dyes.

Fig. 2

Absorption (b) and emission (c) spectra of 50 nm lipidots (a) absorbing/emitting in the visible and near-infrared range. Measures performed in 154 mM NaCl aqueous solution, using lipidots loaded with 42 dyes /particle.

Fig. 3

Size distribution of different DiO- and DiD-lipidot formulations [mean hydrodynamic diameters: 120 (a), 100 (b), 80 (c), 50 (d), and 25 (e) nm, 1.2 mM local dye concentration].

Fig. 4

Mean hydrodynamic diameter (a) and zeta potential (b) of empty and dye-loaded lipidots measured in 0.1X PBS (50 nm dispersions, 1.2 mM local dye concentration).

3.2.

Lipidots Long-Term Stability

Colloidal stability of lipidot dispersions in 154 mM NaCl is assessed by the control of the particle hydrodynamic size, polydispersity index, and zeta potential over time. The nanoparticle diameter seems to slightly increase over time during 4°C storage: approximately 7 nm after 6 months [Fig. 4a]. Zeta potential does not vary significantly during this period, remaining close to neutrality [0 < ξ < 10 mV, Fig. 4b]. Coalescence and Ostwald ripening are the two nanoemulsion destabilization processes. Particle coalescence is prevented by the important droplet PEG coating.²² PEG neutral coating is also a key feature known to prevent in vivo opsonization of blood-circulating nanocarriers.^{26, 27, 28} The use of complex mixtures for core and shell composition, as well as the suspension monodispersity, limit Ostwald ripening.²²

Chemical and loading stability of the fluorophores in the lipidots is assessed by performing another dialysis of the formulations after 1 year storage in 154 mm NaCl at 4°C, in dark (Table 1). DiI and ICG (despite its amphiphilic character) loaded particles display very good stability, since ≈ 90% of the dyes are still encapsulated while DiO and DiD seem to partly leak outside the particles (≈ 50% loss).

Figure 5 compares the fluorescence quantum yields of encapsulated Di dyes to their commercial hydrophilic counterparts (Cy dyes, supplier data). We observe that Di-lipidot fluorescence quantum yields are 1.6 to 3.2 times higher than those of their hydrophilic counterparts (Fig. 5, Table 1). Optical properties of ICG in aqueous solution, at concentration below 200 μm to prevent dimer formation,²⁹ and ICG-lipidot dispersions can be directly compared. ICG-lipidot fluorescence quantum yield is low (0.06), but this value is still a great improvement in comparison to those observed for the free dye in aqueous solution (0.003 in water and 0.01 in whole blood).²⁹

Fig. 5

Lipidot fluorescence quantum yields in 154 mM NaCl just after formulation and at different time points during storage at 4°C in dark.

Even though molar extinction coefficients of the encapsulated dyes are generally lower than those encountered in organic solvents (Table 1), the high dye local concentration per nanoparticle and important fluorescent quantum yields result in highly bright objects. Brightness, which reflects both the fluorophore light absorption and emission abilities, well describes the quality of the optical properties of a fluorescent probe. Calculated brightness for the different lipidots loaded with ≈ 40 dyes/particle are displayed in Table 1. Furthermore, by increasing internal dye concentrations, brightness score up to 1.6 × 10⁷ M⁻¹ cm⁻¹ has been achieved with DiD-loaded lipidots containing ∼350 dyes per particle [Fig. 6b]. Such high brightness can be achieved thanks to the limited dye self-quenching observed when increasing the fluorophore payload [Fig. 6a].

Fig. 6

Influence of dye-loading ratio on 50 nm lipidot emission properties. Standard formulations with 1.2 mM local dye concentration are indicated with a vertical solid line. (a) Fluorescence quantum yields of DiD- (black squares) and DiO- (white triangles) loaded nanoparticles. (b) Calculated brightness of DiD-loaded lipidots. Inset: absorbance for 30 nm particle dispersions.

Figure 7 presents the photobleaching rates in water of different DiD-loaded lipidots, along with those of Cy5 and QTracker^® 705, displaying emission properties in the same wavelength range. Photobleaching rates of DiD-loaded particles are slightly higher than that of Cy5. DiD-lipidots photobleaching rate is about 10 times higher than for QTracker^®705.

Fig. 7

Photobleaching rates of Qtracker^®705, Cy5, 50 nm lipidots loaded with increasing local concentrations of DiD (corresponding to 46, 150, or 350 DiD molecules per particle), and DiD-loaded lipidots with increasing diameters (respectively 7, 46, and 557 dyes per particle for an equivalent local dye concentration of 1.2 mM).

3.3.

In Vivo Imaging with Lipidots

The potential of lipidots for in vivo multichannel fluorescence imaging is explored by injecting “standard” (i.e., 46 dyes/particle) DiO-lipidots (4 pmol particles), DiI-lipidots (2 pmol particles), DiD-lipidots (1 pmol particles), and ICG-lipidots (1 pmol particles) intradermally in the paws of Nude mice. Minimal cross-talk between the different channels is observed. Fluorescence signal shows that nanoparticles are mainly restricted to the paw surface area, indicating that lipidots remain relatively confined to the interstitial space after intradermal injection. This can promote their progressive drainage through lymphatic system. When injection sites are masked with black tape, migration and accumulation of the DiD- and ICG-loaded lipidots in lymph nodes is evidenced [Fig. 8a]. Popliteal, then caudal lymph nodes, and eventually liver, become and remain fluorescent up to a few weeks with DiD [Fig. 8b], while ICG fluorescence has disappeared 24 h post-injection [Fig. 8c]. As expected, no signal is observed with the more blueshifted fluorophores (DiO, DiI) despite higher injection doses (2 to 4 pmol), further demonstrating the importance of near-infrared wavelengths for in vivo experiments.³⁰

Fig. 8

Lymph node detection on Ivis^® system after spectral unmixing (L: liver, Cln: caudal lymph node, Pln: popliteal lymph node). (a) DiD-lipidots (pseudo-colored blue) and ICG-lipidots (pseudo-colored green) are injected, respectively, in the left rear and right rear paws of the same animal, and are easily differentiated from autofluorescence background (white). Probably due to release from the core of the nanoparticles while metabolization, the dyes display very different kinetic behaviors: DiD emission in the lymph nodes (b) reaches its maximum intensity in ≈24 to 48 h and is cleared much more slowly than the ICG signal (c). (Color online only.)

Based on these results, lymph node imaging using DiD-loaded lipidots or QTracker™705 is then compared. Figure 9 shows that injection of 10 times less lipidots (2 pmol particles) than QTracker^®705 (20 pmol particles) leads to a comparable lymph node fluorescence signal. Both DiD and QTracker^®705 signals tend to last over a long period of time (up to 5 weeks).

Fig. 9

Lymph node detection on Ivis^® system after spectral unmixing (Cy5.5 emission filter). 2 pmol of DiD-lipidots (pseudo-colored blue) and 20 pmol of QTtracker^® 705 (pseudo-colored yellow) are injected intradermally, respectively, in the rear left and rear right paw of the mice (L: liver, Cln: caudal lymph node, Pln: popliteal lymph node). (Color online only.)

3.4.

In Vitro Cytotoxicity Assessment

In vitro cytotoxicity of lipidots is compared to that of QTracker^®705 on NIH-3T3 cell line (mouse fibroblast) both through a WST-1 assay [Fig. 10a] and microscopy [Fig. 10b]. Even for the highest concentration of DiD-loaded nanoparticles investigated (75 nM), there is still 70% of cell viability 24 h after incubation. On the contrary, IC₅₀ is reached between 25 and 35 nM for QTracker^®705, with cell viability of approximately 15% when the particle concentration reaches 35 nM. Moreover, microscopy evidences that cytotoxicity occurs through different mechanisms: lipid nanoparticles modify the cell morphology and their adhesion to the plastic walls is hindered, while QDs induce cell destruction, with much cell debris visible on the images.

Fig. 10

Cell viability (WST-1 assay after a 24 h contact, (a) and microscopy imaging (b) of NIH-3T3 fibroblasts incubated in the presence of increasing concentrations of 50 nm diameter DiD-lipidots or QTracker^® 705.