Design of Over-1000 nm Near-Infrared Fluorescent Polymeric Micellar Nanoparticles by Matching the Solubility Parameter of the Core Polymer and Dye

Polymeric micellar nanoparticles (PNPs) encapsulating over-thousand-nanometer (OTN) near-infrared (NIR) fluorescent dye molecules in block polymers having hydrophobic and hydrophilic chains are promising agents for the dynamic imaging of deep tissue. To achieve OTN-NIR fluorescent PNPs (OTN-PNPs) having high brightness, it is crucial to increase the affinity between the core polymer and dye molecules by matching their polarities; thus, criteria and methods to evaluate the affinity are required. In this study, we used the Hansen solubility parameter (HSP), including the polarity term, to evaluate the affinity between the two substances. HSP values of the OTN-NIR fluorescent dye IR-1061 and four core polymers, poly(lactic-co-glycolic acid) (PLGA), poly(lactic acid) (PLA), poly(ε-caprolactone) (PCL), and polystyrene (PSt), were calculated using the Hansen solubility sphere method and molecular group contribution method, respectively. The relative energy density between IR-1061 and each core polymer calculated using their HSP values revealed that the affinities of PLGA and PLA for IR-1061 are higher than those of PCL and PSt. Therefore, OTN-PNPs composed of PLGA, PLA, and PCL core polymers were prepared and compared. The OTN-PNPs having PLGA and PLA cores could be loaded with larger amounts of IR-1061, had higher photoluminescence intensities, and showed higher stability in phosphate buffered saline than those having PCL cores. Moreover, the OTN-PNPs having PLGA or PLA cores were used for the dynamic imaging of live mice. Thus, matching the solubility parameters of the core polymer and dye molecule is a useful approach for designing high-performance OTN-NIR fluorescent probes.


■ INTRODUCTION
Dynamic live imaging in the over-thousand-nanometer (OTN) near-infrared (NIR) optical window, the so-called second biological window (NIR-II), 1 allows significantly higher spatial resolution with deeper imaging depth 2−4 and lower autofluorescence 5 than imaging in the traditional visible (400−700 nm) and first NIR window (NIR-I, 700−900 nm). Among the various OTN-NIR fluorescent imaging probes, including rareearth-doped ceramics, 6−8 single-walled carbon nanotubes, 9−12 and quantum dots, 13−15 organic-dye-based fluorophores exhibit great potential for biomedical research and clinical use because of their high biocompatibility, designable structures, and tunable optical properties. 16−18 Polymethine molecules such as IR-26, IR-1048, IR-1051, and IR-1061 are commercially available OTN-NIR fluorophores. They need to be conjugated or incorporated into biocompatible polymers for use in in vivo imaging because they are insoluble in water. While polymer formulations of various dyes such as phthalocyanines 35 and heptamethines 36,37 have been investigated, the encapsulation of IR-1061 in biocompatible polymer nanoparticles (PNPs) for use in in vivo imaging has also been reported. 19,20 Specifically, IR-1061 (see Figure 1) is a polar and hydrophobic molecule that undergoes aggregation and fluorescence quenching in environments with mismatched polarities. For example, recently, we reported highly emissive OTN-NIR polystyrene (PSt)-based nanoparticles having core polymers with tuned polarities, which was achieved by changing the monomer ratio (styrene to acrylic acid), to increase the affinity for IR-1061. 21 However, the quantitative evaluation of the affinity of dyes for the core polymers remains challenging, and this has limited the rational and efficient design of OTN-NIR fluorescent agents.
One possible tool for the numerical estimation of the affinity between two substances is the solubility parameter. There are two common solubility parameters: the Hildebrand solubility parameter and the Hansen solubility parameter (HSP). The Hildebrand solubility parameter condenses all structural information into one representative value, whereas the HSP is composed of three factors: dispersion forces (δ d ), intermolecular dipole interactions (δ p ) and hydrogen-bonding  ACS Nanoscience Au pubs.acs.org/nanoau Article interactions (δ h ). 22 Therefore, in this study, we chose the HSP to enable the better approximation of polar interactions and hydrogen bonding between the two substances, that is, the organic polymer and fluorescent dye. 23−25 We first investigated the HSPs of the polar dye molecule IR-1061 and the hydrophobic parts of the amphiphilic polymers that interact with IR-1061 in the OTN-PNPs. We selected four commercially available amphiphilic polymers: poly(ethylene glycol) (PEG)-block-poly(lactic-co-glycolic acid) (PLGA) (PEG-b-PLGA), PEG-block-poly(lactic acid) (PLA) (PEG-b-PLA), PEG-block-poly(ε-caprolactone) (PCL) (PEG-b-PCL), and PEG-block-PSt (PEG-b-PSt) (see Figure 1 for the corresponding structures). To design OTN-PNPs with a high affinity between IR-1061 and the core polymer, it is necessary to evaluate the solubility of solid IR-1061 and each core polymer quantitatively. Thus, we attempted to determine whether solid IR-1061 and the core polymer could dissolve together by examining the relative energy difference (RED) calculated from the corresponding HSPs. To validate the affinities predicted by HSP evaluation, we confirmed the IR-1061 content encapsulated in the OTN-PNPs, as well as the brightness and stability of the OTN-PNPs. Furthermore, in vivo imaging of mice was performed using OTN-PNPs prepared from PEG-b-PLGA and PEG-b-PLA, which were found to have a high affinity for IR-1061 in our HSP evaluation. were purchased from NOF Corporation (Tokyo, Japan). Acetonitrile (ACN) and Dulbecco's phosphate buffered saline (PBS) were purchased from Fujifilm Wako Pure Chemical Corporation (Osaka, Japan). Bovine serum albumin was purchased from Hoffmann-La Roche (Basel, Switzerland). All reagents were used without further purification.

HSP Determination of IR-1061 and Core Polymers
The HSP of IR-1061 was determined using Hansen Solubility Parameters in Practice (HSPiP) 26 based on the Hansen solubility sphere method. 27 In the Hansen solubility sphere method, the HSP is calculated by testing the solubility of the unknown material in various organic solvents. This method can be used regardless of the state of the sample; therefore, it is suitable for determining the HSP of ionic compounds such as IR-1061. Therefore, we evaluated the affinity of IR-1061 for typical organic solvents whose HSPs are already known. 26 One milliliter of each of the 30 organic solvents listed in Table 1 was added to 2 mg of IR-1061. After sonication for 10 min in a bath sonicator (UT-206, SHARP, Japan), each tube was centrifuged at 21 500 g for 10 min (CF16 RN, HITACHI, Japan). If a precipitate of IR-1061 was observed, the solvent was judged to have a low affinity for IR-1061 with a score of 0, whereas when IR-1061 was completely dissolved, the solvent was judged to have a high affinity with a score of 1. From the solubility data, each organic solvent was plotted in the HSP space, and a fitting procedure was used to determine the HSP sphere of IR-1061. The center coordinates and radius of the HSP sphere were obtained by using software (HSPiP, 4th ed., 4.0.07) from the results of solubility scores of IR-1061 in each solvent. In addition, the radius of the HSP sphere is the interaction radius (R 0 ) of IR-1061, which is a tolerance indicator for the interaction of IR-1061 with other materials. 28 The HSPs of PLGA, PLA, PCL, and PSt were also determined using HSPiP based on the molecular group contribution method. 29

Calculation of RED between IR-1061 and Each Core Polymer
To evaluate whether IR-1061 has a high affinity for each core polymer, the distance between IR-1061 and the core polymer (R a ) was calculated using eq 1.
Here, the subscripts 1 and 2 represent IR-1061 and each core polymer, respectively, and the δ values are the HSP factors discussed earlier.
The RED provides an estimate of whether the two materials are miscible. 30 RED is given by the ratio of R 0 and R a , as shorn in eq 2.
A RED of 1 or less indicates that IR-1061 is expected to have a high affinity for the core polymer, whereas a RED larger than 1 means IR-1061 is expected to have a low affinity for the core polymer.

Preparation of OTN-PNPs
The structure of the OTN-PNPs and the molecular structures of the components are shown in Figure 1. Briefly, PEG-b-PLGA in ACN (11.1 mg/mL, 9 mL) and IR-1061 in ACN (0.1 mg/mL, 1 mL) were mixed in a vial, followed by the addition to distilled water (40 mL). The mixing ratio of the polymer and the dye weights was set at 1000:1 based on the dependence of fluorescence intensity on the ratio investigated using PEG-b-PLGA ( Figure S1). Then, the stirred aqueous solution was left overnight at 20°C to allow the evaporation of the ACN. Finally, the obtained OTN-NIR fluorescent PEG-b-PLGA PNPs (OTN−PLGA-PNPs) were purified using a dialysis membrane (molecular weight cutoff (MWCO) = 2 kDa) overnight and then concentrated to 100 mg/mL using a centrifuge filter (MWCO = 100 kDa, 3000g, 20 min). OTN-NIR fluorescent PEG-b-PLA PNPs (OTN-PLA-PNPs) and PEG-b-PCL PNPs (OTN−PCL-PNPs) were prepared in the same manner, except that the block copolymers were changed.
In Vivo OTN-NIR Fluorescence Imaging The HSP of IR-1061 was determined using the Hansen solubility sphere method. Table 1 shows the solubilities of IR-1061 in each organic solvent. N,N-Dimethylformamide, dimethyl sulfoxide, γ-butyrolactone, methylene dichloride, Nmethyl-2-pyrrolidone, and propylene carbonate satisfied the criteria for dissolving IR-1061 at a concentration of 2 mg/mL. The Hansen solubility sphere plotted from the results of the solubility evaluation of IR-1061 is shown in Figure 2, in which the blue spheres and red cubes indicate the solvents that meet the solubility criteria (that is, those that dissolved IR-1061 at a concentration of 2 mg/mL) and those that do not, respectively.  (Table 2). This means that PLGA and PLA, which have RED values of less than 1, can "dissolve" IR-1061; that is, they are highly compatible with IR-1061. In contrast, PCL and PSt have signifficantly different δ p values from IR-1061. The obtained RED values indicate that the polarity of the core polymer is an important factor in designing or selecting a polymer that is highly compatible with IR-1061, which is consistent with our previous reports. 21 In Vitro Properties of OTN-PNPs Containing PLGA, PLA, and PCL To validate the affinities predicted by HSP evaluation, we prepared three types of IR-1061-loaded polymer micelles having different core polymers, PLGA, PLA, and PCL. The OTN-PNPs were obtained by adding dropwise a mixture of IR-1061 and PEG-b-PLGA, PEG-b-PLA, or PEG-b-PCL in ACN to water, followed by stirring to remove the ACN and achieve micellization. The experiment using PSt, which showed the lowest result in the HSP prediction (Table 2), was not performed because we already reported that PSt alone has a low compatibility with IR-1061. 21  Particularly, PLGA and PLA have larger solubility parameters than PCL due to the difference in polarity and hydrogen bonding. This difference is expected to be related to the compatibility with IR-1061, which is similarly polar but insoluble in water. Thus, the amounts of IR-1061 encapsulated in the OTN-PLGA-PNPs and OTN-PLA-PNPs were significantly higher than that encapsulated in the OTN-PCL-PNPs, which is consistent with the differences in the affinities between IR-1061 and the core polymers based on HSP evaluation. It should be noted that the fluorescence intensity and absorption spectra used to calculate the amount of encapsulated dye are influenced by the surrounding molecules such as the core matrix. Therefore, the fluorescence intensity is not completely proportional to the dye amount calculated by the absorbance. Although the quantum yield of these PNPs was not evaluated, it was estimated from the NIR absorption data for OTN-PLGA-PNPs and OTN-PLA-PNPs to be similar to our previously reported polystyrene-based probes (approximately 0.6−0.7%). 21 Interestingly, in the absorption spectrum of the OTN-PCL-PNPs, there is a peak at approximately 780 nm, suggesting that the IR-1061 dye was exposed to water (Figure 3b) 21 and, thus, the PCL had not encapsulated IR-1061 well. In addition, the photoluminescence intensities of the OTN-PLGA-PNPs and OTN-PLA-PNPs are approximately 4 times higher than those of the OTN-PCL-PNPs, suggesting the usefulness of HSP prediction in designing highly emissive OTN-NIR fluorescent probes.  The in vitro stabilities of the OTN-PLGA-PNPs, OTN-PLA-PNPs, and OTN-PCL-PNPs were assessed by analyzing the time-dependent changes in the OTN-NIR photoluminescence intensity of each OTN-PNP dispersed in PBS or a solution of 4% albumin dissolved in PBS at 37°C for 7 days. The buffer and solution were used as in vitro models to test the stability of PNP in vivo, but it was tested for a longer period of time (up to 7 days) in the present study because these models were simpler than the actual in vivo environment and may have milder effects. The OTN-NIR photoluminescence intensities of the OTN-PLGA-PNPs, OTN-PLA-PNPs, and OTN-PCL-PNPs were maintained at 41 (±6.0)%, 29 (±5.9)%, and 5.9 (±3.1)%, respectively, after incubation in PBS for 4 days at 37°C (Figure 4a). After a 7 day incubation, the retention rates were  On the other hand, in a solution of 4% albumin dissolved in PBS, the photoluminescence intensity decreased to approximately 25% after 4 days, even for the OTN-PLGA-PNPs and OTN-PLA-PNPs (Figure 4c). Albumin is the dominant protein in blood serum; thus, it was chosen as a model protein to estimate the influence of blood proteins on the stability of the OTN-PNPs. Free IR-1061 is completely quenched when it was just mixed with aqueous albumin solution owing to the influence of water molecules on IR-1061 (unpublished work). Because albumin interacts with the polymer micelles and, thus, reduces micelle stability, 31,32 the albumin enhanced the leakage of IR-1061 from the core polymer. As above, the stability of OTN-PNPs was found to be different depending on the core polymer. Further studies are currently underway to analyze the kinetics of degradation and dye leakage by investigating their properties at more time points and to increase the stability of OTN-PNPs in the physiological environments.
In Vivo Imaging Using OTN-NIR Fluorescence As a demonstration of the in vivo OTN-NIR fluorescence imaging capacity of the OTN-PLGA-PNPs and OTN-PLA-PNPs, we performed imaging of a cancer model lesion, which was established by subcutaneously inoculating colon-26 cells (a model cell line of murine colon carcinoma) into the abdomen. Immediately following the intravenous injection of the OTN-PNPs via the tail vein and excitation by light at 980 nm, the blood vessels under the skin could be observed with high clarity over the whole mouse body ( Figure 5). Furthermore, imaging of the blood vessels was still possible 4 h after injection, indicating the high retention of the OTN-PLGA-PNPs and OTN-PLA-PNPs in the circulating blood. The OTN-NIR fluorescence emission was observed from the liver as a blood-rich organ as well as blood vessels until 4 h postinjection ( Figure 5). Moreover, at 24 h after injection, the accumulation of OTN-PLGA-PNPs and OTN-PLA-PNPs in the tumors was observed ( Figure 5), probably owing to the enhanced permeability and retention (EPR) effect. 33 A decrease in image brightness over time after the injection of the OTN-PNPs was observed, but this is consistent with the ACS Nanoscience Au pubs.acs.org/nanoau Article results of the in vitro stability tests using albumin, and, as reported previously, 31,32 OTN-PNPs typically decompose faster in vivo than in solutions containing a single protein because blood serum is a complex mixture of proteins, including albumins and globulins, as well as small molecular compounds.
In this study, the imaging analysis was performed with a single mouse per group as a demonstration to show the potential angiography and cancer imaging using our designed OTN-PNPs. For cancer imaging, the stability of encapsulated dye with PNPs should be controlled by design of the PNPs, because their long retention in the blood is generally required following intravenous injection. Thus, the polymer that encapsulates IR-1061 must not only be hydrophobic but also have a certain polar. This paper showed the OTN-PNPs with properties that enable cancer imaging can be designed by matching the compatibility of the dye with core polymers. On the basis of the in vivo study, the OTN-PNPs containing commercially available dyes and block copolymers designed based on their HSP can be used for both conventional vascular imaging 20,34 and tumor imaging.

■ CONCLUSION
We have reported a method to design OTN-NIR fluorescentdye-loaded polymeric micellar nanoparticles (OTN-PNPs) by matching the solubility parameter of the core polymers to that of hydrophobic dye, which, in this study, was IR-1061. Using the Hansen solubility parameter, HSP, as an evaluation index, we found that PLGA and PLA have higher affinity for IR-1061 than PCL and PSt do. Further, the OTN-PNPs composed of PEG-b-PLGA and PEG-b-PLA showed higher IR-1061 encapsulation efficiencies, brightness, and stabilities in PBS compared to the OTN-PNPs composed of PEG-b-PCL, suggesting that, in the former cases, the dye molecules are etained in the PNP structure because of the higher affinity of IR-1061 for the core polymer. Crucially, using the OTN-PNPs composed of PEG-b-PLGA and PEG-b-PLA, the in vivo imaging of live mice could be performed, and blood vessels and tumor tissue could be imaged. Therefore, we proposed that matching the solubility parameters of dyes and core polymers is a useful approach for designing high-performance fluorescent polymer nanoparticles containing hydrophobic dyes.