2.1.

Materials

Suppocire NB™ was purchased from Gattefossé (Saint-Priest, France), Lipoid S75 (soybean lecithin at $>75\%$ phosphatidylcholine) from Lipoid (Ludwigshafen, Germany), Myrj™ S40 (polyethylene glycol 40 stearate), and super refined soybean oil from Croda Uniqema (Chocques, France). ICG, IR780 iodide dye, and other chemicals were purchased from Sigma-Aldrich (Saint-Quentin-Fallavier, France).

2.2.

LipImage™ 815 Preparation

The IR780-lipid dye was synthetized in two steps using commercial IR780 iodide as a starting material. In the first step, a carboxylic acid group was introduced in the fluorophore template using the same procedure as previously described for the IR786 dye.³⁹ A ${\mathrm{C}}_{18}$ lipid chain was then grafted on the fluorophore using oleylamine and standard $-{\mathrm{CO}}_2\mathrm{H}/{\mathrm{NH}}_2$ coupling chemistry activated by benzotriazol-1-yloxy)tris(dimethylamino)phosphonium hexafluorophosphate (BOP).

An oil premix with, respectively, 85, 255, and 65 mg of oil, Suppocire NB™ and lecithin was prepared. IR780-lipid dye solution (200 μL) in ethanol ($10\mathrm{mg}/\mathrm{mL}$) was poured in a 5-mL vial and mixed with the oil premix melted at 50°C. The mixture was homogenized and the solvent was then evaporated under argon flux. After homogenization at 50°C, the continuous aqueous phase, composed of 345 mg of Myrj™ S40 and the appropriate amount of aqueous solution (154 mM NaCl if not stated otherwise, qsp 2 mL), was introduced. The vial was placed in a 50°C water bath and the mixture was sonicated for 5 min using a VCX750 Ultrasonic processor (power output 190 W, 3-mm probe diameter, Sonics). LipImage™ 815 was dialyzed overnight at room temperature against 1000 times their volume in the appropriate aqueous buffer (12 to 14,000 Da MW cut off membranes, ZelluTrans, Carl Roth, France). Finally, the nanoparticle dispersion was 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 (DLS) was used to determine the particle hydrodynamic diameter and zeta potential (Zeta Sizer Nano ZS, Malvern Instrument, Orsay, France). Particle dispersions were diluted to $2\mathrm{mg}/\mathrm{mL}$ of lipids in sterile 0.1X PBS and transferred in Zeta Sizer Nano cells (Malvern Instrument) before each measurement, performed in triplicate.

2.4.

Dye Encapsulation Efficiency and Payload

The encapsulation efficiency of IR780-lipid dye in LipImage™ 815 was determined by spectophotometric titration and calculated by dividing the absorbance measured at 790 nm for freshly purified particle dispersion by the absorbance at 790 nm of the same particle dispersion before purification, by keeping the lipid concentration constant. The amount of dye which had been removed by dialysis was estimated. Absorbance was measured in 1X PBS using a UV-visible spectrophotometer (Cary 300 Scan, Varian, Les Ulis, France). The dye payload into the particles was obtained by calculation considering the initial dye amount in milligrams and the encapsulation efficiency.

2.5.

Optical Properties

The absorbance and fluorescence measurements were performed in 1X PBS, respectively, using a Cary 300 Scan UV-visible spectrophotometer (Varian) and a Perkin Elmer LS50B fluorimeter. No scattering contribution was considered due to nanoparticle size and spectrum range (750 to 850 nm). Fluorescence quantum yields were calculated using ICG in dimethylsulfoxide (DMSO) ($\mathrm{\Phi }=0.13$) as a reference.⁴⁰

2.6.

WST-1 In Vitro Cytotoxicity Assay

In accordance with the International Organization for Standardization (ISO) 10993, NIH-3T3 murine fibroblast cells were chosen to perform classical WST-1 cytotoxic assay (11644807001, Roche, Boulogne-Billancourt, France). The cell line was purchased from the American Type Culture Collection (3T3, CRL 1658, ATCC, Manassas, VA). Experimental cultures were prepared from deep-frozen stock vials and maintained in a subconfluent state in culture Petri Dish ($113{\mathrm{cm}}^2$) under a humidified (90%) atmosphere of 95% air/5% ${\mathrm{CO}}_2$ at 37°C. Cells were grown in Dulbecco’s modified Eagle’s medium high glucose (DMEM, 41966-029, Gibco, Invitrogen, Saint-Aubin, France) supplemented with 10% newborn calf serum (P30-0401, PAN Biotech GmbH, Aidenbach, Germany), and 1% antibiotics (penicillin and streptomycin) (15140-122, Gibco). Culture medium was changed every other day.

Cells (${10}^4\text{cells}/\mathrm{mL}$) were seeded in 96-well plates (Nunc, Sigma-Aldich, Saint-Quentin-Fallavier, France). After 24-h incubation at 37°C, different concentrations of lipid nanoparticles, from 50 to $1500\mu \mathrm{g}/\mathrm{mL}$ (${7.2\times 10}^{11}$ to ${2.2\times 10}^{13}\text{particles}/\mathrm{mL}$), were added for 24 h to the culture medium. Each group had sixplicate wells. Cytotoxicity was assessed 24 h following the nanoparticle removal using WST-1 assay, analogue to MTT assay. WST-1 reagent (Roche, Boulogne-Billancourt, France) was added (10%) to the culture medium and kept in the incubator for 3 h. Cells without nanoparticles and those incubated with a solution of ${\mathrm{H}}_2{\mathrm{O}}_2$ 10 mM were, respectively, used as negative and positive controls. Absorbance was then recorded at 450 nm (soluble formazan titration) and 690 nm (background subtraction) using a microplate reader (Infinite M1000, Tecan, Lyon, France). The absorbance difference (450 to 690 nm) was directly proportional to the number of viable cells. Cell viability was expressed as a percentage of viable cells as determined using the following equation: $\text{Viability}(\%)=[(A_S-A_{\mathrm{PC}})/(A_{\mathrm{NC}}-A_{\mathrm{PC}})]\times 100$; where $A_S$, $A_{\mathrm{PC}}$, and $A_{\mathrm{NC}}$ represented absorbances of the sample, the positive control (cells with ${\mathrm{H}}_2{\mathrm{O}}_2$ 10 mM), and the negative control (only cells), respectively. Data were compared among groups by one-way ANOVA followed by Fisher’s protected least significance differences test.

2.7.

In Vivo Imaging Experiments

All animal procedures were 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 were made to minimize animal suffering and to reduce the number of animals used. All animal manipulations were performed with sterile techniques and were approved by the Rhône-Alpes Animal Care and Use committee (France) or by the animal ethics committee of Institut d’Imagerie Biomédicale.

To assess the biodistribution of LipImage™ 815 in comparison to free IR780-lipid dye dispersed in $5\%\text{ethanol}/95\%\mathrm{NaCl}$ 154 mM, healthy FVB female mice (6 weeks old, Janvier, France), weighing 22 to 25 g, were injected through the tail vein in bolus with 100 μL volumes using 29G syringes. The injected doses corresponded to 8 nmol of dye and ${1.3\times 10}^{14}$ particles for LipImage™ 815 (${1.3\times 10}^{15}\text{particles}/\mathrm{mL}$). Mice were anesthetized and sacrificed at different time points (30 min, 4 h, 24 h, $n=3/\text{point}$), organs were harvested and their fluorescence analyzed using a Fluobeam®800 imaging set-up (Fluoptics, Grenoble, France). Fluobeam®800 is a compact and portable open-field NIR imaging system, combining fluorescence excitation by a 785-nm laser source and an optical head with a highly sensitive charge-coupled device (CCD) camera and white light-emitting diodes (LEDs) for field illumination. Dedicated optics scatter the laser and LEDs, enlightening a field of 6 cm in diameter at a distance of 17 cm. The power density of laser irradiation on tissue is $96\mu \mathrm{W}/{\mathrm{mm}}^2$. The fluorescence signal is collected through a high-pass filter over 800 nm. Quantification of organ fluorescence was performed using the ImageJ® software, after subtraction of autofluorescence measured for each organ in control mice.

For histology and long-term biodistribution studies, 10 immunocompetent 6-week old BALB/cJ RJ mice were used. Nanoparticles (200 μL of solution containing $2\times {10}^{15}\text{nanoparticles}/\mathrm{mL}$ loaded with 300 μM of fluorophore, i.e., representing a dose of 60 nmol of dye, $4.0\times {10}^{14}\text{particles}$) were injected in the tail vein and animals were sacrificed at the indicated time points. The organs were then imaged with the fluoSTIC system for fluorescence detection. FluoSTIC is a previously described two-dimensional (2-D)-FRI system.⁴¹ Excitation was performed using a sub miniature A (SMA) coupled 500 mW, 740-nm laser diode from Power Technology (model# IQ1A) as NIR light source, and a SMA coupled 7-mW cold (5600K) white LED from Thorlabs (model# MCWHF1). This system provides continuous NIR excitation of the body of the mouse because it produces a homogeneous lightened field (5 cm in diameter) with an illumination power of $10\mathrm{mW}/{\mathrm{cm}}^2$. The fluorescence signal is collected by a CM-030GE-RH (Jai, Denmark), 17-mm diameter camera with a $1/3$ in monochrome CCD sensor and a resolution of $656\times 494\text{pixels}$. The filtration scheme of FluoSTIC accommodates an excitation filter centered at 740 nm that excites ICG with a 40% efficiency as well as an emission filter from 772 to 857 nm that collects most ICG emitted photons. All filters are interference filters purchased from Chroma (Bellows Falls, VT). A calibration curve was prepared from serial dilutions of particles (data not shown). Intensities of fluorescence were indicated as reference light units $(\mathrm{RLU})/\text{pixels}/\mathrm{ms}$.

For plasmatic distribution, 6-week old BALB/cJ RJ mice received the same dose of nanoparticles in the tail vein (60 nmol of fluorophores, ${4.0\times 10}^{14}\text{particles}$), and blood samples were collected at the indicated times from the tail vein. The blood was then centrifuged at 4°C at 2000 rpm, and the plasma was used to detect the presence of fluorescence. A drop of 20 μL was exposed to the fluoSTIC system and the amount of fluorescence was quantified as $\mathrm{RLU}/\text{pixel}/\mathrm{ms}$.

In the tumor model study, male athymic Swiss nude mice (Janvier, Le Genest-Isle, France) were used and maintained under specific pathogen-free conditions. A total of ${10}^7$ human prostate cancer cells (PC3) in 200 μL phosphate-buffered saline were injected subcutaneously. When tumor diameter ranged between 5 and 8 mm (4 to 5 weeks), animals were injected intravenously with 200 μL of solution containing $2\times {10}^{15}\text{nanoparticles}/\mathrm{mL}$ loaded with 300 μM of fluorophore (60 nmol).

3.1.

Optical Properties of LipImage™ 815

The commercial IR780 fluorophore, similarly to the IR-820 dye, is a heptacyanine dye presenting in its polymethine backbone a C6 ring for structure rigidification and improvement of optical and stability properties, as well as a chlorine group, which can be substituted by ${\mathrm{SN}}_1$ substitution reactions as extensively described in the literature.^39,42 Presenting suitable NIR optical properties for in vivo imaging (798-nm absorption, 819-nm emission in DMSO, Table 1), we reinforced its lipophilic character by substituting the chlorine group with an oleyl chain to favor its anchoring in the lipid core of the particles.

Table 1

Optical properties of LipImage™ 815 and ICG-Lipidots® in PBS 1X (n=120  dyes/particle), in comparison to free dyes in DMSO and ICG in glycosylated water. λmax(abs), λmax(em) are, respectively, the dyes absorption and emission maxima, ε is the molar extinction coefficient at the absorption maximum (/dye), Φ the fluorescence quantum yield, b the brightness. Brightness is defined as b=ε×Φ for free dyes, b=n×ε×Φ for nanoparticles (n=120  dyes/particle).

ICG-Lipidots® and LipImage™ 815 lipid nanoparticles encapsulating, respectively, ICG and the newly designed IR780 lipid dye, were formulated following standard previously described procedures,^{34,37,38,43,44} leading to transparent green nanoparticle dispersions in aqueous buffer. DLS was used to assess ICG-Lipidots® and LipImage™ 815 hydrodynamic diameter, polydispersity, and zeta potential as well as the stability of the particle dispersions with time. The mean hydrodynamic diameter was 50 nm ($\text{polydispersity}=0.13$) for both formulations and was not affected by dye payload. Nanoparticles were stable for more than 6 months while the formulations were stored at 4°C (Fig. 1). The same behavior was observed for zeta potential, which was $-2.5\pm 0.5\mathrm{mV}$ for both the nanovectors. This neutral charge in NaCl or PBS buffer was accounted for by the dense PEG coating of the particle surface, which also ensured a high colloidal stability for the particles.

Fig. 1

Evolution of the hydrodynamic diameter and polydispersity (PDI) of the nanoparticle dispersions during storage at 4°C in the dark.

The dye encapsulation efficiency was assessed by spectrophotometric titration at the maximum of absorption comparing the optical absorption of freshly prepared solutions before and after dialysis. The difference allowed determining the amount of dye still encapsulated in the particle core. The entrapment efficiency of ICG-Lipidots® varied with the loading ratio (40% for low loading $[\mathrm{ICG}]<10\mathrm{dye}/\text{particle}$, 80% for higher loading $10<[\mathrm{ICG}]<50\mathrm{dye}/\text{particle}$ and decreased for very high loading $>50\mathrm{dye}/\text{particle}$) for 30 nm nanoparticles.³¹ The entrapment efficiency of the IR780-lipid fluorophore in the nanoparticles remained above 70% for dye loading up to $330\text{dyes}/\text{particle}$ (Fig. 2).

Fig. 2

Influence of dye loading on encapsulation efficiency for LipImage™ 815.

The optical features of LipImage™ 815 and ICG-Lipidots® (50-nm diameter particles loaded with $120\text{dyes}/\text{particle}$) dispersed in 1X PBS buffer in comparison to those of the nonencapsulated dyes in DMSO, and presently clinically used ICG in glycosylated water are presented in Table 1. The normalized absorbance and emission spectra of the particles are represented in Fig. 3. IR780-lipid dye displayed similar optical properties to ICG in DMSO, with absorbance maximum at 798 versus 795 nm for ICG, emission maximum at 819 versus 821 nm for ICG, fluorescence quantum yield of 0.12 versus 0.13 for ICG. Freshly prepared ICG in glycosylated water presented blue-shifted absorption and emission as well as reduced fluorescence quantum yield (0.048).³¹ Molar extinction coefficient was lower for IR780-lipid dye than that of ICG, $\mathrm{150,000}\mathrm{L}{\mathrm{mol}}^{-1}{\mathrm{cm}}^{-1}$ versus $215,000\mathrm{L}{\mathrm{mol}}^{-1}{\mathrm{cm}}^{-1}$, leading to brightness for the free dye in DMSO 1.5-fold lower than for ICG. LipImage™ 815 and ICG-Lipidots® showed optical properties close to those of the free dyes in DMSO: the absorbance and emission maxima were slightly blue shifted (5 nm) for IR780-lipid dye, showing a low positive solvatochromism from the apolar oily core of the nanoparticle to highly polar DMSO. This effect was not observed for free and encapsulated ICG where the absorbance maximum was slightly red shifted. For both the dyes, molar extinction coefficients were not influenced or were slightly influenced by encapsulation. Fluorescence quantum yields slightly decreased upon encapsulation, from 0.13 to 0.06 for ICG-Lipidots® and from 0.12 to 0.08 for LipImage™ 815 (Table 1). However, fluorescence quantum yields of the LipImage™ 815 and ICG-Lipidots® dispersions remained unchanged over 3 months during storage at 4°C in the dark in aqueous buffer (Fig. 4), whereas glycosylated ICG, with lower fluorescence quantum yield, aggregated and became nonfluorescent in a few hours.¹⁹ The IR780-lipid dye was not soluble at all in aqueous buffer. The fluorescence quantum yield of LipImage™ 815 was stable for dye payloads up to $330\text{dyes}/\text{particle}$ at least, allowing the achievement of highly bright nanoparticles (Fig. 5). The stability of the fluorescence quantum yield over the increasing dye loading evidenced an absence of self-quenching and aggregation of the dye in the lipid core. Overall, LipImage™ 815 and ICG-Lipidots® displayed similar optical properties in terms of absorption and emission wavelengths and brightness. Encapsulation of the dyes in the lipid formulation ensured their prolonged preservation of both colloidal and optical properties, while stored in ready-to-use conditions (solvent-free saline buffer).

Fig. 3

Normalized absorption and emission spectra of LipImage™ 815 and ICG-Lipidots® in 1X PBS (excitation at 750 nm).

Fig. 4

Time evolution at 4°C in darkness of the fluorescence quantum yields of LipImage™ 815 and ICG- Lipidots® in 1X PBS (excitation at 750 nm).

Fig. 5

Influence of dye loading on fluorescence quantum yield and particle brightness for LipImage™ 815.

3.2.

In Vitro Cytotoxicity

The cytotoxicity of different nanoparticle dispersions, naked Lipidots®, ICG-Lipidots® and the novel NIR probe LipImage™ 815, was assessed on NIH 3T3 fibroblasts through a cell viability assay (WST-1) (Fig. 6). The toxicity profiles of these nanoformulations were relatively similar, all of them exhibiting a half-maximal inhibitory concentration close to the high lipid concentration of $1\mathrm{mg}/\mathrm{mL}$, corresponding to $1.5\times {10}^{13}\text{particles}/\mathrm{mL}$. This indicated that (1) lipid nanoparticles were very well tolerated by fibroblasts in culture for an incubation time of 24 h and (2) the incorporation of a NIR dye for in vivo imaging purpose (either ICG or IR780-lipid dye) did not modify the toxicity index of the basic formulation.

Fig. 6

Cell viability assay (WST-1) on NIH 3T3 fibroblasts after 24-h incubation of different particle dispersions: naked Lipidots® (black diamond), ICG-Lipidots® (white square), and LipImage™ 815 (black triangle).

3.3.

In Vivo Imaging

First, the biodistribution of LipImage™ 815 and of the free IR780-lipid dye was assessed in FVB healthy mice (Fig. 7). Fluorescence was relatively homogeneously distributed through the different organs after LipImage™ 815 injection, even if more pronounced in liver, and to a lesser extent in steroid-rich organs (adrenals, ovaries). On the contrary, injection of the IR780-lipid dye dispersed in water/ethanol mixture led to massive fluorescence signal in spleen and liver in a few hours after injection.

Fig. 7

Fluorescence biodistribution 24 h after injection of IR780-lipid dye (a) or LipImage™ 815 (b) in FVB mice. At, adipose tissue; Ag, adrenal glands; Br, brain; He, heart; In, intestines; Ki, kidney; Li, liver; Lu, lung; Mu, muscle; Ov, ovaries; Pa, pancreas; Sg, salivary gland; Sp, spleen; Ut, uterus.

The absence of toxicity consecutive to a systemic administration of LipImage™ 815 was further checked in long-term experiments. For this purpose, 10 mice receive 200 μL of the LipImage™ 815 formulation (60 nmol of fluorophores, $4.0\times {10}^{14}\text{particles}$), and two of them were sacrificed at 24 h after injection, two others at 8 days and two others at 15 days. The remaining four mice were followed till day 21. No signs of toxicity were noticed, i.e., no weight loss, no obvious reaction was visible on the live mice. In addition, after a histologic examination using HES coloration (hematoxylin, eosin, safran) of frozen sections of liver, spleen, lungs, or kidneys, all tissues appeared normal (data not shown). At days 1, 8, 15, and 21, the main organs were also exposed to the NIR camera in order to detect the presence of fluorescence. Biodistribution was similar to that reported in Fig. 7, with a global decrease of fluorescence in all organs starting at day 1 (except liver for which a slight increase was observed between day 1 and day 8). However, the fluorescence signal associated with the LipImage™ 815 injection was still very elevated at day 8, until day 21.

Another series of mice was used for the evaluation of the blood clearance of ICG-Lipidots® and LipImage™ 815. Three mice received LipImage™ 815 intravenously, and three others the same dosage of ICG-Lipidots®. Blood samples were collected at 15 min, 30 min, 1 h, 3 h, and 5 h after systemic administration of ICG-Lipidots®, while we also collected a sample of blood at 24 h for LipImage™ 815. As can be seen in Fig. 8, the fluorescence signal present in the plasma was very low and disappeared extremely rapidly when ICG-Lipidots® were injected (0.6% of the injected dose at 30-min postinjection), while 7.5% of the injected dose was present at 30-min postinjection for LipImage™ 815, 2.5% at 5 h and still 0.5% one day later. This indicated that LipImage™ 815 remained in circulation in the blood stream for extended periods of time before being completely eliminated.

Fig. 8

Fluorescence in plasma after injection of ICG-lipidots® or LipImage™ 815 in BALB/cJ RJ mice.

This long lasting circulation was associated with a very good tumor enhanced permeability and retention (EPR) effect, as can be seen in Fig. 9(a). In two nude mice bearing PC3 subcutaneous prostate tumors, fluorescence was detected noninvasively at the tumor site till 22 days after injection of LipImage™ 815. At this time, the signal was still very elevated and easily detectable using 20 ms exposure time, as shown for mouse #1 (mouse #2 was similar). We then reproduced this experiment in three other mice with the same prostate tumors. As presented in Figs. 9(b) and 9(c), the passive, EPR-guided, accumulation of LipImage™ 815 in the tumors provided tumor/skin ratio ranging between 2 and 3.4. The best signal/noise ratio was obtained at day 9, demonstrating again the long-term persistence of the signal in the tumors although the intensity of the signal was gradually declining over time.

Fig. 9

Fluorescence accumulation in tumor after LipImage™ 815 injection in athymic Swiss nude mice with implanted human prostate cancer cells (PC3). (a) In vivo fluorescence images of the tumor at days 1, 4, 8, 11, 18, 22 and (b) quantification of the fluorescence signal in tumor and skin. (c) Tumor over skin fluorescence ratio.