2.1.

Materials

All chemicals were purchased from Sigma-Aldrich (Milwaukee, Wisconsin) and used without further purification unless otherwise noted. The free radical initiator $2,2^{\prime }$ -azo-bis(isobutylnitrile) (AIBN) was recrystallized from methanol prior to use. Styrene was purified by distillation under reduced pressure. Maleic anhydride was recrystallized from chloroform. Solvents used were ACS grade and obtained from Sigma-Aldrich with the exception of ethyl ether (EMD, Gibbstown, New Jersey).

2.2.

Polymer Synthesis

Poly(styrene-alt-maleic anhdyride)-block-poly(styrene) (PSMA-b-PSTY) copolymers were prepared by a thermally initiated, two-step reversible addition-fragmentation chain transfer (RAFT) polymerization using the chain transfer agent (CTA) S-benzyldithiobenzoate (BDTB). BDTB was prepared from S-(thiobenzoyl)thioglycolic acid and benzylmercaptan as previously described.²⁷ In the first step, equimolar amounts of styrene and maleic anhydride $\left(0.03\mathrm{mol}\right)$ were polymerized for $18\mathrm{h}$ at $60°\mathrm{C}$ with AIBN as initiator and BDTB $\left(0.00075\mathrm{mol}\right)$ as chain-transfer agent in p-dioxane $(50\mathrm{wt}\%)$ . The BDTB:AIBN ratio was 10:1, and the polymerization vessel was subject to three cycles of freeze-vacuum-thaw before polymerization. The PSMA macro-CTA (Mn $9700\mathrm{kDa}$ , $\mathrm{PDI}=1.34$ ) was isolated by precipitation in ethyl ether, filtered, and dried under vacuum at room temperature. In the second step, $1.9130\mathrm{g}$ of macro-CTA were combined with $0.033\text{-}\mathrm{mol}$ styrene and $3.23\text{-}\mathrm{mg}$ AIBN (10:1 macro-CTA:initiator ratio) in dimethylformamide (DMF) $(36\mathrm{wt}\%)$ in a round bottom flask. The solution was degassed by three cycles of freeze-vacuum-thaw, and polymerization was carried out for $24\mathrm{h}$ in a $60°\mathrm{C}$ oil bath. The block copolymer PSMA-b-PSTY (Mn 21,600 kDa, $\mathrm{PDI}=1.31$ ) was recovered by precipitation in ethyl ether, followed by filtration, and dried under vacuum at room temperature.

The molecular weights of the PSMA macro-CTA and PSMA-b-PSTY block copolymers were determined by size exclusion chromatography using Tosoh TSK-GEL $\alpha \text{-}3000$ and $\alpha \text{-}4000$ columns (Tosoh Bioscience, Montgomeryville, Pennsylvania) connected in series to a Viscotek GPCmax VE2001 and refractometer VE3580 (Viscotek, Houston, Texas). High pressure liquid chromatography (HPLC)-grade DMF containing $0.1\mathrm{wt}\%$ LiBr was used as the mobile phase. The molecular weights of the synthesized copolymers were determined using a series of poly(methyl methacrylate) standards.

Following synthesis of the PSMA-b-PSTY block copolymer, the PSMA block was derivatized with butylamine as previously described,²⁸ yielding a pH-sensitive hydrophilic PSMA derivative through aminolysis of 60% of the anhydride groups in the PSMA block. Briefly, a solution of PSMA-b-PSTY block copolymer $(13\mathrm{wt}\%)$ was prepared in DMF. Subsequently, $1\mathrm{mL}$ of a $0.33\text{-}\mathrm{M}$ solution of butylamine in DMF was added drop-wise while stirring, and the solution was allowed to react for $24\mathrm{h}$ at room temperature. The polymer was recovered by precipitation in ethyl ether, and remaining anhydrides were hydrolyzed by dissolving the polymer in $1\text{-}\mathrm{N}$ NaOH. After $3\mathrm{h}$ , the solution was extensively dialyzed against deionized water using SpectraPor Slide-a-Lyzer cassettes ( $\mathrm{MWCO}=3500$ , Pierce, Rockford, Illinois) to remove sodium salts and neutralize the solution prior to lyophilization.

2.3.

Micelle Formation

Modified PSMA-b-PSTY block copolymers $\left(37.3\mathrm{mg}\right)$ were dissolved in $6\text{-}\mathrm{mL}$ DMF and $360\text{-}\mu \mathrm{L}$ $1\text{-}\mathrm{N}$ NaOH $(6\mathrm{vol}\%)$ . After $30\min$ of stirring, $6\mathrm{mL}$ of nanopure ${\mathrm{H}}_2\mathrm{O}$ (pH 9) was added by syringe pump at a rate of $20\mu \mathrm{L}∕\min$ . Following injection, the solution was transferred to a dialysis cassette (MWCO 3500, Pierce) and dialyzed extensively against nanopure ${\mathrm{H}}_2\mathrm{O}$ . Following dialysis, the solution was filtered through a $0.2\text{-}\mu \mathrm{m}$ filter.

2.4.

Indocyanine Green Encapsulation

A solvent evaporation method was employed to encapsulate ICG into the micelles after complexing ICG with tetrabutylammonium iodide to form a hydrophobic ICG-tetrabutylamine salt. Briefly, a $1\text{-}\mathrm{mM}$ solution of ICG was prepared in chloroform with a six-fold molar excess of tetrabutylammonium iodide. The solution was sonicated for $30\min$ , then diluted in chloroform to obtain a final ICG concentration of $260\mu \mathrm{M}$ . The chloroform ICG solution was then added to a stirring micelle solution $(2.1\mathrm{mg}∕\mathrm{mL})$ , yielding a final ICG concentration of $10\text{-}\mu \mathrm{M}$ ICG. The chloroform was removed by evaporation, causing ICG to partition into the hydrophobic cores of the PSMA-b-PSTY micelles. Free ICG was then separated from the ICG-micelle solutions using Amicon regenerated cellulose centrifuge filters (MWCO $100\mathrm{kDa}$ , Millipore, Billerica, Massachusetts). Following ultrafiltration, the micelle-ICG solution was rinsed three times using nanopure ${\mathrm{H}}_2\mathrm{O}$ . The carryover volume between rinses was less than $100\mu \mathrm{L}$ . Purified micelles were resuspended in the original volume of nanopure ${\mathrm{H}}_2\mathrm{O}$ . The final micelle concentration after filtering was determined by ${}^1\mathrm{H}\text{-}\mathrm{NMR}$ . Lyophilized samples were resuspended in deuterated DMF containing 1% trimethylsilane (TMS). Concentration was determined by comparing the styrenic peak area to a series of known standards using the TMS peak as an internal reference standard.

2.5.

Characterization

2.6.

Formulation and Stability

2.7.

In Vitro Study—Cytotoxicity and ${\mathit{IC}}_{\mathit{50}}$ Value Determination

The ${\mathrm{IC}}_{50}$ (concentration of material that induces 50% cell death) of PSMA-b-PSTY micelles encapsulating ICG was determined using the MTS tetrazolium salt (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium salt) metabolic assay on NIH-3T3 cells. Cells were seeded at 12,000 cells/well in a 96-well tissue culture polystyrene plate (Costar Corning Incorporated Lowell Massachusetts). The cells were allowed to adhere to plates overnight prior to addition of micelle solutions. Stock micelle solutions at $2.1\mathrm{mg}∕\mathrm{mL}$ were sterilized by filtration through a $0.2\text{-}\mu \mathrm{m}$ syringe filter, diluted in cell culture media to final concentrations of $0.3$ , 0.6, 1.1, and $1.7\mathrm{mg}∕\mathrm{mL}$ , then added to cells. The cells were incubated in the presence of micelle solutions for $4.5\mathrm{h}$ in the incubator. The media was then replaced with fresh media, and the cells were returned to the incubator for $24\mathrm{h}$ . At the conclusion of the $24\mathrm{h}$ incubation, the cells were rinsed and treated with MTS reagent following manufacturer’s specifications (CellTiter 96® ${\mathrm{AQ}}_{\text{ueous}}$ MTS Reagent Powder, Promega Corporation Madison, Wisconsin). Cell viability was evaluated by recording the absorbance of each well at $490\mathrm{nm}$ using a microplate reader. Positive controls consisted of cells exposed to the same water/media concentration as each micelle treatment, and a negative control consisted of cells exposed to $0.5\text{-}\mathrm{mg}∕\mathrm{mL}$ branched-polyethylenimine (b-PEI, MW 25K, Sigma, Saint Louis, Missouri).

3.1.

Polymer Synthesis and Micelle Characterization

Block copolymers of poly(styrene-alt-maleic anhydride)-block-poly(styrene) (PSMA-b-PSTY) were successfully prepared with controlled architectures and low polydispersities using a two-step RAFT polymerization. Following polymerization, the maleic anhydride groups of the PSMA block were subject to ring-opening aminolysis using butylamine to obtain an amphiphilic copolymer as previously described (Fig. 1 ).²⁸ 60% of the maleic anhydride residues were opened by aminolysis with butylamine, followed by base-catalyzed hydrolysis to open the remainder of the maleic anhydride moieties. These modifications yield a block copolymer with a hydrophobic poly(styrene) component and hydrophilic poly(styrene-alt-maleic anhydride) component. The copolymers self-assemble in aqueous environments to form polymeric micelles with hydrodynamic diameters of $55\pm 2.7\mathrm{nm}$ as determined by dynamic light scattering. In addition, the overall charge of the micelles was determined to be near-neutral by zeta potential measurements (Table 1 ).

Fig. 1

Schematic of poly(styrene)-block-poly(styrene-alt-maleic anhydride) (PSMA-b-PSTY) diblock copolymer synthesis, the building block for the polymeric micelle to encapsulate ICG.

Table 1

CMC, zeta potential, and particle sizing of unloaded and ICG-loaded micelles.

3.2.

Indocyanine Green Encapsulation and Characterization

ICG was converted to the hydrophobic tetrabutylamonium salt and loaded into the hydrophobic micelle core by solvent evaporation. The efficiency of ICG loading was determined by computing the ratio of ICG concentration after loading to that of total ICG loaded, and was found to be $87\pm 5\%$ . Next, the absorbance and fluorescence spectra of ICG were compared to the spectra of micelle-encapsulated ICG [Figs. 2a and 2b , respectively]. Slight red-shifts were observed in both the absorbance and emission spectra of encapsulated ICG (25 and $10\mathrm{nm}$ , respectively).

Fig. 2

Absorbance/emission spectra of (a) ICG and (b) micelle-encapsulated ICG. For fluorescence spectra, ICG and micelle-encapsulated ICG were excited at their respective maximum absorption wavelengths.

The average diameter of ICG-loaded micelles was determined to be $30\pm 3.1\mathrm{nm}$ by dynamic light scattering. The overall charge of ICG-loaded micelles remained near-neutral (zeta potential of $-0.11\pm 2.2\mathrm{mV}$ ). Particle sizing experiments revealed excellent salt stability of the empty and ICG-loaded micelles with minimal size change observed after $50\min$ of incubation in phosphate buffered saline containing $150\text{-}\mathrm{mM}$ NaCl. The particle characterization data are reported in Table 1. Micelle size and morphology were also determined by AFM after ICG loading. The AFM imaging of these particles showed an average diameter of $30.8\mathrm{nm}\pm 2.5\mathrm{nm}$ (Fig. 3 ). Atomic force microscopy also revealed an overall spherical morphology, though it appeared the micelles flattened on the mica surface with the height of each individual particle smaller $(4.40\mathrm{nm}\pm 0.34\mathrm{nm})$ than its width along the surface. Based on the combined AFM and DLS data, it is reasonable to speculate that the micelles deform during the drying process before AFM imaging.

Fig. 3

Atomic force microscopic (AFM) image of polymeric micelles encapsulating indocyanine green (ICG). Flattened surface image is displayed at right, and at left, the image is displayed at a height of $10\mathrm{nm}$ .

The critical micelle concentration (CMC) is an indication of micelle stability, a critical parameter for micelle-based therapeutics. Generally, the lower the CMC, the more stable the micelle. The CMC of the PSMA-b-PSTY micelles was determined both before and after loading ICG by measuring the solvatochromic shift of PRODAN fluorescence emission as a function of polymer concentration (Fig. 4 ). Above the CMC, PRODAN partitions from aqueous solution into the hydrophobic cores of polymeric micelles, undergoing a significant blueshift relative to its emission in water.²⁹ Based on this method, the CMC of the PSMA-b-PSTY micelles alone was determined to be $1\mathrm{mg}∕\mathrm{L}$ . The CMC for the ICG loaded micelles was also determined to be $1\mathrm{mg}∕\mathrm{L}$ ; thus, ICG loading does not affect the CMC of PSMA-b-PSTY micelles (Table 1).

Fig. 4

The ratio of hydrophobic:hydrophilic PRODAN fluorescence emission (in nanometers) as a function of polymer concentration. Above the CMC, micelle formation causes an increase in the emission ratio as PRODAN partitions from aqueous solution into the hydrophobic micelle cores, undergoing a significant solvatochromic blueshift in fluorescence emission $\left(436\text{to}518\mathrm{nm}\right)$ .

3.3.

Solution Stability, Thermal Stability, and Release of Encapsulated Indocyanine Green

To determine whether encapsulation in polymeric micelles could protect ICG from degradation, the aqueous stability of ICG in micelle formulations was investigated. The fluorescence emission of micelle-encapsulated ICG and free ICG was examined over a period of three weeks with measurements repeated approximately every $24\mathrm{h}$ for the first two weeks and every $100\mathrm{h}$ within the third week. The emission profile of ICG over time provides a strong indicator of its overall stability in a given environment, and is a key parameter for imaging applications. In these studies, the fluorescence emission of free ICG decreased rapidly over time, while the fluorescence emission intensity of micelle-encapsulated ICG remained constant throughout the two week duration of the study (Fig. 5 ). The emission of free ICG decreased to 50% of its original value within $96\mathrm{h}$ , and was reduced by more than 90% within $12\text{days}$ . These findings are consistent with the rapid instability of free ICG reported in the literature.^{2, 24, 25} In contrast, no change in ICG emission was observed when ICG was encapsulated in the micelle core.

Fig. 5

Relative fluorescence emission of free ICG and micelle-encapsulated ICG over a three week period at room temperature (RT on legend) and $37°\mathrm{C}$ (37 on legend). Solutions were maintained at the respective temperature and covered from light. Emission intensities were normalized to maximum fluorescence intensity within each sample at $0\mathrm{h}$ .

Free ICG also degrades rapidly at elevated temperatures, another significant limitation for in vivo imaging applications. To evaluate the thermal stability of ICG-encapsulated polymeric micelles compared to free ICG, fluorescence emission of ICG at physiological temperatures was recorded from samples stored in a $37°\mathrm{C}$ oven over a period of three weeks. Emission of ICG was recorded at the same intervals as in the solution stability study. At $37°\mathrm{C}$ , the emission of free ICG decreased rapidly to 17% within $96\mathrm{h}$ and 0% within six days (Fig. 5). In contrast, ICG encapsulated in micelles was relatively protected from thermal degradation, maintaining 97% of original emission for $96\mathrm{h}$ . In addition, after three weeks of incubation at $37°\mathrm{C}$ , encapsulated ICG still showed 41% of its initial emission. These results demonstrate the stabilization of micelle-encapsulated ICG at both room temperature and physiological temperature.

For in vivo imaging applications, the micelles must retain ICG until they have reached the appropriate targeted tissue, and must protect ICG from degradation and clearance until imaging is complete. The release of ICG from polymeric micelles was examined by measuring the retention of ICG within dialysis tubing as a function of time. Solutions of micelle-encapsulated ICG and free ICG were dialyzed against a MWCO 25K membrane in deionized water for $24\mathrm{h}$ at $4°\mathrm{C}$ . At various time points, solutions were removed from the tubing, weighed, and concentrations of ICG remaining determined by absorbance readings at $775\mathrm{nm}$ . By determining the concentration of ICG remaining in the dialysis tubing at a given time point relative to the initial absorbance, the mass loss of ICG was calculated as a percentage of the initial loading (Fig. 6 ). After $24\mathrm{h}$ , more than $89\pm 6.8\%$ of the micelle-encapsulated ICG was retained in the tubing (i.e., 11% release), compared to $27\pm 3.6\%$ remaining of free ICG (i.e., 73% release). This result indicates that the rate of ICG release from the micelles is sufficiently slow for in vivo imaging applications.

Fig. 6

Fraction of ICG released from dialysis bags over a $24\text{-}\mathrm{h}$ period. Solutions of free ICG and micelle-encapsulated ICG were evaluated.

3.4.

Cytotoxicity and ${\mathrm{IC}}_{50}$ Value Determination

A cytotoxicity study was performed to determine cell viability when exposed to increasing concentrations of micelle-encapsulated ICG. The ${\mathrm{IC}}_{50}$ , or concentration of polymer resulting in 50% cell death, was determined as an indicator of the toxicity of the polymeric micelles. Cell viability was determined by the MTS metabolic activity assay. No significant cytotoxicity compared to control treatments was observed at concentrations of 0.3, 0.6, and $1.1\mathrm{mg}∕\mathrm{mL}$ of PSMA-b-PSTY. Increasing toxicity with polymer concentration between $1.1\text{to}1.7\mathrm{mg}∕\mathrm{mL}$ (the highest concentration provided) was observed with cell viability decreasing from 88 to 33%, respectively, indicating an ${\mathrm{IC}}_{50}$ of $\sim 1.5\text{-}\mathrm{mg}∕\mathrm{mL}$ PSMA-b-PSTY (Fig. 7 ).

Fig. 7

Percent NIH-3T3 cell viability in response to exposure of cells to increasing concentrations of micelles encapsulating ICG.

4.1.

Polymer Synthesis and Characterization

Previous studies have demonstrated the feasibility of one-pot synthetic procedures to prepare PSTY-b-PSMA block copolymers for the formation of micelles and cross-linked nanoparticles.^{30, 31} However, a two-step synthesis was employed here for ease of characterization of the molecular weight and polydispersity of each block in the copolymer. These block copolymers self-assemble to form polymeric micelles. Advantages of these specific micelles include the PSTY glassy core that contributes to micelle stability (low CMC), and the presence of functional groups in the corona that can be used for future derivitization with targeting ligands. The formulated micelles also have ideal sizes for tumor imaging applications. The desired size of nanoparticles for delivery from leaky tumor vasculature is less than $400\text{to}600\mathrm{nm}$ .²⁶ In addition, particles less than $100\mathrm{nm}$ have decreased recognition by the reticuloendothelial system, resulting in longer circulation half lives.³² Prolonged circulation and reduced clearance will improve passive targeting to tumor tissue via the EPR effect. Unloaded micelles have an average diameter of $55\pm 2.7\mathrm{nm}$ , whereas ICG-loaded micelles have an average diameter of $30\pm 3.1\mathrm{nm}$ . The difference in size for the ICG-loaded micelles as opposed to the unloaded micelles is due to the process of ICG loading, not filtration, as was discovered in a separate study (data not shown). It is possible that the presence of chloroform during ICG loading allows reorganization of the micelle core. In physiologic salt concentrations, the sizes of both unloaded and loaded micelles were shown to be stable. The decrease in micelle hydrodynamic diameter after salt addition may be attributed in part to the salt shielding the anionic charges of PSMA polymers in the micelle corona, thereby decreasing electrostatic repulsion between the PSMA chains. In an in vivo setting, the micelles should be capable of maintaining their overall structure and thus retain cargo for circulation to the targeted site.

4.2.

Indocyanine Green Encapsulation and Characterization

Tumor imaging using ICG is limited by the poor biodistribution and instability of the free dye. Recently, micelles have garnered great interest as carriers for therapeutic molecules because of their ability to improve the biodistribution and retention of small molecules, and their ability to protect fragile molecules during systemic circulation.³² By loading ICG into the cores of polymeric micelles, we hypothesized that the dye would be protected from rapid aqueous and thermal degradation without compromising the attractive spectral properties of the dye. To investigate this hypothesis, micelles derived from modified PSMA-b-PSTY block copolymers were loaded with ICG using a solvent evaporation method, which allowed ICG to partition into the hydrophobic micelle core. Using this method, ICG loading efficiencies of 87% were obtained from $10\text{-}\mu \mathrm{M}$ ICG loading solutions.

In aqueous solutions, free ICG is known to self-quench at concentrations above $5\mu \mathrm{M}$ .^{12, 33} Therefore, self-quenching of ICG in the micelle core was a potential concern. To investigate the effects of ICG concentration on ICG fluorescence in the polymeric micelle, various concentrations of ICG were loaded into the micelles, and the fluorescence emission intensity was compared (data not shown). These studies showed micelle-ICG fluorescence increased as the ICG concentration in the loading solution was changed from $5\text{to}10\mu \mathrm{M}$ , but decreased with further increases in ICG concentration. Based on these results, an initial ICG loading of $10\mu \mathrm{M}$ into polymeric micelles was taken as the optimal concentration.

The ICG loading efficiency obtained in our studies compares favorably with previously published results. In other systems, drug loading efficiencies in polymeric micelles have been reported between 10 and 90%.^{34, 35, 36, 37, 38} In this study, the hydrophobic tetrabutylamine salt of ICG was loaded with 87% efficiency. A recent study by Saxena examined the stability of ICG when incorporated into PLGA nanoparticles, and calculated a $74.5\pm 2.2\%$ loading efficiency.²⁴ Also, Yu reported loading efficiencies between 90.0 and 97.1% when ICG was incorporated into much larger $0.6\text{to}2\text{-}\mu \mathrm{m}$ polymer nanoparticle-assembled capsules.²⁵ In comparison to these previous studies, we can conclude that the ICG loading method we describe is highly efficient. Unlike previous investigations with ICG, we used tetrabutylammonium iodide to increase the hydrophobicity and micelle loading of ICG. The amount of tetrabutylammonium iodide used to form the ICG salt is 2300-fold below the ${\mathrm{LD}}_{50}$ (oral in rat is $1990\mathrm{mg}∕\mathrm{kg}$ ), and should not pose complications for in vivo imaging applications.³⁹

The stability of the polymeric micelles was determined by measuring the CMC before and after encapsulating ICG. Highly stable polymeric micelles (those with low CMCs) are required for in vivo applications, as micelles are subjected to significant dilution following intravenous injection.³² If injection reduces the block copolymer concentration below the CMC, polymeric micelles become thermodynamically unstable and begin to dissociate into component unimers. Dissociation is a kinetic process that occurs at different rates depending on the plasticity of the micelle core and other factors, but complete dissociation results in a loss of the therapeutic cargo. The CMC of the polymeric micelles described here was determined to be $1\mathrm{mg}∕\mathrm{L}$ for both unloaded micelles and ICG loaded micelles. This value is lower⁴⁰ or comparable^{41, 42} to that reported by other micelles derived from various amphiphilic block copolymers. The stability of the polymeric micelle system we describe is a result of the high glass transition temperature $T_g$ of the styrenic core, as has been well documented in previous work.³² For dosing in a mouse model, typical administration of ICG ranges from $0.2\text{to}0.4\mathrm{mg}∕\mathrm{kg}$ body weight.^{19, 43} This value corresponds to $69\mathrm{mg}$ of micelles per kg or $1.2\mathrm{mg}\text{per}\mathrm{mL}$ blood volume (assuming mouse blood volume is 5 to 6% body weight). Therefore, the target dose will be above the CMC of $1\mathrm{mg}∕\mathrm{L}$ by almost three orders of magnitude.

4.3.

Solution Stability, Thermal Stability, and Release of Encapsulated Indocyanine Green

A recent study examined the stability of ICG when incorporated into PLGA nanoparticles.²⁴ ICG encapsulated in PLGA nanoparticles was stable after a four-day incubation in distilled water, showing 60% decrease in ICG fluorescence compared with free ICG (97.8% decrease).²⁴ In the system presented here, micellar formulations of ICG resulted in sustained stabilization of ICG based on peak fluorescence for more than two weeks without any significant degradation. This reflects the polymeric micelles’ ability to stabilize ICG in an aqueous environment, thus allowing for easier formulation, longer shelf life, and a greatly enhanced diagnostic/therapeutic window. The micellar formulations of ICG also exhibited stability at $37°\mathrm{C}$ ; fluorescence intensity decreased by only 59% after three weeks incubation in solution at physiological temperatures.

Over time, ICG is expected to release from the micelles as it diffuses out of the polystyrene cores. However, because of the glassiness of the poly(styrene) cores, we anticipated that the rate of ICG release would be very slow. The actual release rate of ICG from the micelle cores was determined to be 11% over $24\mathrm{h}$ . It should be noted that for the free ICG control, $\sim 60\%$ of ICG released as a burst from the dialysis tubing within the first six hours, followed by a period of slower release, resulting in a 73% total loss after $24\mathrm{h}$ . The plateau of release observed is likely due to fouling of the dialysis membrane by the released ICG. These results indicate the micelle-ICG formulation is sufficiently stable to allow for reasonable storage times and long imaging windows following in vivo administration.