Polymeric Self-Assemblies with Boron-Containing Near-Infrared Dye

Nov 28, 2016 - ... 150 W xenon light source (MAX-150, Asahi Spectra Co., Ltd., Japan, ...... Lin , M.; Menon , N. V.; Borah , P.; Ng , H. J.; Loong , ...
0 downloads 0 Views 5MB Size
Article pubs.acs.org/Biomac

Polymeric Self-Assemblies with Boron-Containing Near-Infrared Dye Dimers for Photoacoustic Imaging Probes

Downloaded via UNIV OF SOUTH DAKOTA on June 27, 2018 at 09:25:14 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Koji Miki,*,† Akane Enomoto,† Tatsuhiro Inoue,† Tatsuya Nabeshima,‡ Sousuke Saino,‡ Soji Shimizu,§ Hideki Matsuoka,∥ and Kouichi Ohe*,† Departments of †Energy and Hydrocarbon Chemistry and ∥Polymer Chemistry, Graduate School of Engineering, Kyoto University, Katsura, Nishikyo-ku, Kyoto 615-8510 Japan ‡ Graduate School of Pure and Applied Sciences and Tsukuba Research Center for Interdisciplinary Materials Science (TIMS), University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8571, Japan § Department of Chemistry and Biochemistry, Graduate School of Engineering, Kyushu University, Fukuoka 819-0395, Japan S Supporting Information *

ABSTRACT: Polymers containing pyrrolopyrrole aza-BODIPY (PPAB) and thiophene-bridged BODIPY dimers (TBD) having poly(ethylene glycol) (PEG) or PEGylated hyaluronic acid (HA) were prepared by facile conjugation approaches. Self-assemblies consisting of TBD-conjugated polymers more efficiently generated photoacoustic (PA) signals than PPAB−PEG conjugate upon irradiation with near-infrared pulsed laser light. Among dye-conjugated polymers examined, TBD-HA-PEG conjugates efficiently generated photoacoustic signals, 1.49−1.83 times stronger than that of commercially available indocyanine green (ICG). We found that the following two factors are essential to enhance PA signals from self-assemblies: (1) the formation of strongly interacting TBD aggregates and (2) enhancement of the elastic modulus of self-assemblies by conjugating TBDs with HA. TBDconjugated HA derivatives circulated in blood vessels for a longer time (15.6 ± 4.9% injected dose (ID) in blood 24 h after injection) and more specifically accumulated in tumor tissues (17.8 ± 3.5% ID/g in tumor 24 h after injection) than ICGconjugated HA derivatives, visualizing a tumor site more clearly. The cell uptake experiment of dye−HA conjugates indicates that ICG-conjugated polymers internalized into cells or merged with cell walls to emit strong fluorescence, while TBD-conjugated polymers were not internalized into cells. Because the disassembly of the TBD-conjugated HA derivatives is suppressed, aggregated TBDs emit weak fluorescence but efficiently generate strong PA signals in tumor tissues.



INTRODUCTION Photoacoustic (PA) imaging is one of the most powerful molecular imaging techniques for visualizing disease targets.1−4 For high-contrast detection of target tissues, a variety of compounds and nanomaterials have been investigated as a probe agent for PA imaging. PA contrast agents for in vivo use need to have the following properties: (i) high molar extinction coefficient in the near-infrared (NIR) region between 700 and 900 nm (optical window) and (ii) efficient conversion of excited photoenergy for thermoelastic expansion. Dyes endowed with such properties have been intensively investigated to develop a PA contrast agent for practical use.5−9 Over the last half-century, indocyanine green (ICG, Figure 1a), which has a high molar extinction coefficient (∼1.0 × 105 M−1 cm−1 at 780 nm), has been used as an approved contrast agent to visualize the bloodstream and diseases, especially in the liver.10,11 Because of its low fluorescence quantum yield, © 2016 American Chemical Society

especially in aqueous media, ICG has recently been applied as a probe molecule of PA contrast agents.12−14 Recently, we have developed amphiphilic hyaluronic acid (HA) derivatives containing ICG for tumor imaging.15 Self-assemblies of HA derivatives with an ICG-condensed hydrophobic core, which can carry a large number of ICG molecules, could successfully accumulate in tumor tissues through passive tumor targeting,16 visualizing the tumor site with high-contrast using optical and PA tumor-imaging devices. However, because of its rapid photobleaching property,17 contrast agents having an ICG moiety are hardly applicable to continuous monitoring of the PA signal. Received: October 23, 2016 Revised: November 18, 2016 Published: November 28, 2016 249

DOI: 10.1021/acs.biomac.6b01568 Biomacromolecules 2017, 18, 249−256

Article

Biomacromolecules

Figure 2. Amphiphilic BODIPY dimers as a contrast agent for PA tumor imaging. of self-assemblies for measurements was prepared by filtration with a syringe filter (0.45 μm pore size, PVDF). UV−vis absorption spectra were recorded on a UV−vis spectrophotometer (V-570, JASCO Corporation, Japan). Fluorescence spectra were measured using a spectrofluorimeter (FluoroMax-3, Horiba Jobin Yvon Inc., France). Fluorescence intensities of solutions of dyes were measured using an IVIS Imaging System Series (PerkinElmer, excitation wavelength, λexc = 745 nm; detection wavelength, λem = 820 nm). Images were analyzed using Living Image 2.50-Igor Pro 4.09 software (PerkinElmer) to quantify the fluorescence intensity, according to the manufacturer’s instructions. Dynamic light scattering (DLS) was measured on an FPAR-1000 (Otsuka Electronics Co., Ltd., Japan) at 20 °C. The sample solutions were diluted in H2O and filtered prior to analysis (syringe filter, 0.45 μm pore size). The concentrations of the analyzed sample solutions were 0.05 g/L. The measurements were performed at scattering angles of 90° at 25 °C. The critical aggregation concentration (cac) values of dye−polymer conjugates were determined using static light scattering (SLS, SLS6000, Otsuka Electronics Co., Ltd., Japan). The measurement was conducted according to the reported method.34 The data are summarized in Figure S8. Transmission electron microscopy (TEM, JEM-1400, JEOL Ltd., Japan) was used to visualize the morphology of the dried selfassemblies. Samples were dropped onto a TEM copper grid covered with a carbon film (200 mesh, Nisshin EM, Japan) and dried for 3 h before measurement. PA signals of samples in H2O, phosphate buffered saline (PBS) solutions (pH 7.4), and 75% fetal calf serum (FCS) solutions (concentration: 1.3 × 10−4 M (dye)) were measured according to the reported procedure.15 Photobleaching Test. Solutions of amphiphilic dye conjugates (1−2 mg) were dissolved in DMF/H2O (v/v = 1:20, 2 mL). To minimize the influence of the background signal, the absorbance of all solutions was adjusted to approximately 0.5−1.5. All solutions were transferred into quartz cells in air and irradiated using a 150 W xenon light source (MAX-150, Asahi Spectra Co., Ltd., Japan, illuminance: 12 μW·cm−2 (above 400 nm) at the sample level, wavelength: 400−800 nm) with a visual light module and an optical filter (cutoff < 400 nm). The time-dependent photobleaching was monitored by measuring the UV−vis absorbance at 780 nm. UV−vis absorption spectra are summarized in Figure S9. Optical Imaging Experiments In Vivo. All animal experiments were approved by the Animal Research Committee of Kyoto University and carried out in accordance with its guidelines. For the current study, a suspension of colon 26 mouse cancer cells (1 × 106, RIKEN) was subcutaneously inoculated into the femurs of 6-week-old nude mice (BALB/c Slc-nu/nu mice, Japan SLC). A solution of TBDconjugated HA derivative 4b in PBS was prepared, filtered (0.45 μm pore size), and kept in the dark at room temperature before injection. The solution (100 μL in PBS, [TBD] = 5.0 × 10−4 M) was intravenously injected into each mouse. Optical imaging experiments to detect the distribution of the fluorescence probes in the mice were carried out with an IVIS Imaging System 200 Series (λexc = 745 nm, λem = 820 nm). In all imaging experiments, the contrast agent in the

Figure 1. (a) ICG, (b) BODIPY, (c) fused-type BODIPY dimer derivative, and (d) thiophene-bridged BODIPY dimer.

Boron dipyrromethene (BODIPY, Figure 1b) is known to be resistant to photobleaching and highly fluorescent in the visible region, thereby being applied as a labeling agent in vitro.18−20 Aiming at its application to photoacoustic imaging in vivo, many researchers have tried to develop BODIPY-based NIR dyes by extending the π-conjugated system.21,22 Recently, Shimizu and Kobayashi reported the π-extended fused-type BODIPY dimers, pyrrolopyrrole aza-BODIPY (PPAB), and their application to organic photovoltaics in the vis region (Figure 1c).23,24 Nabeshima et al. reported thiophene-bridged BODIPY dimers (TBD) exhibiting a remarkable redshift in absorption and emission (Figure 1d).25,26 In this article, we wish to report that NIR dyes having this unique BODIPY dimer structure are suitable as photostable probe molecules for high-contrast PA tumor imaging in vivo. To the best of our knowledge, there are a few examples of BODIPY derivatives as photoacoustic imaging probes in vivo.27−29 In vivo photoacoustic imaging utilizing π-extended BODIPY analogues for sentinel lymph node identification27 and tumor detection28 have recently been developed. Chang, Wu, and co-workers demonstrated that water-soluble BODIPY dimer-loaded bovine serum albumin nanoparticles could be used as a contrast agent for in vivo PA tumor imaging.29 However, in these papers, the relationship between PA signal intensity and dye aggregation in nanoparticles have not been mentioned. Although it is wellknown that the aggregation of BODIPYs leads to their fluorescence quenching,30 the aggregation-induced PA signal enhancement of BODIPYs has not been intensively investigated. We and other research groups found that aggregated NIR dyes emit weak fluorescence but generate strong PA signals.15,31,32 By evaluating the photophysical properties of these two types of BODIPY dimers conjugating with poly(ethylene glycol) (PEG) in H2O, we found that the strongly interacted TBDs in a hydrophobic region of polymeric self-assemblies could generate stronger PA signal compared with weakly interacted PPABs (Figure 2). Furthermore, by conjugating PEG and TBD with HA, which is known as an enhancer of the elastic moduli,33 we discovered that nanoparticles of PEG−TBD−HA polymer generated PA signals efficiently and visualized a tumor site clearly in vivo.



MATERIALS AND METHODS

Materials and Methods. The preparation of PEG−dye conjugates and TBD- and PEG-conjugated HA derivatives is summarized in the Supporting Information. To evaluate the properties and morphology of self-assemblies, amphiphiles were dissolved in H2O (5 mg/mL) and stored at room temperature for at least 30 min in the dark. A solution 250

DOI: 10.1021/acs.biomac.6b01568 Biomacromolecules 2017, 18, 249−256

Article

Biomacromolecules same batch was injected into at least three mice to confirm reproducibility, and representative images are shown. To emphasize the areas where imaging probes specifically accumulated, we set a threshold and represent the intensity of the fluorescence above it. Blood Concentration of Self-Assemblies [%Injected Dose (ID) (Blood)]. A solution of TBD-conjugated HA derivative 4b (100 μL in PBS, [TBD] = 5.0 × 10−4 M) was injected into the tail vein of a female outbred BALB/c Slc-nu/nu mouse (7−9-weeks old, Japan SLC, Inc.). Blood (2 μL) was harvested from the tail vein of the mouse after 24 h from the injection and mixed with 1% Triton X-100 aqueous solution (9 μL) and DMF (9 μL). By measuring the fluorescence intensity of the mixture with an IVIS Imaging System 200 Series (λexc = 745 nm, λem = 820 nm), %ID (blood) with respect to the injected dose was determined. Tumor Specificity of HA Derivative 4b [%ID/g (tumor)]. After in vivo tumor-imaging experiments, mice were euthanized under CO2, followed by the surgical resection of tumor tissues. DMF (20.25 times as heavy as the weight of tumor tissues) was added to a mixture of tumor tissues in 1% Triton X-100 aqueous solution (1.25 times as heavy as the weight of tumor tissues). By measuring the fluorescence intensity of the mixture with an IVIS Imaging System 200 Series (λexc = 745 nm, λem = 820 nm), %ID (tumor) with respect to the injected dose was determined. PA Imaging In Vivo. The preparation of tumor-bearing mice was the same as that shown in “optical imaging experiments in vivo”. PA signals from both the tumor-bearing side and the non-tumor-bearing side were measured with a Nexus 128 (Endra Inc., wavelength of laser: 790 nm) before injection of 4b. A solution of TBD-conjugated HA derivative 4b (100 μL in PBS, [ICG] = 5.0 × 10−4 M) was intravenously injected into the tail vein of tumor-bearing mice. After 24 h from the injection, PA signals were measured. Delivery of Dye-Conjugated HA Derivative to Cells. Colon 26 mouse cancer cells (2 × 106 cells/well, RIKEN) were cultured with ICG (500 μL, 0.10 mM PBS solution) or 4b or 5 (100 μL, 0.50 mM PBS solution) in Dulbecco’s Modified Eagle Medium (DMEM, 10% FCS, 2 mL) at 37 °C for 24 h. We checked that the cells did not suffer serious damage after the treatment (Figure S13). The cells were washed with PBS (1 mL × 2), removed from the culture by treatment with trypsin, and suspended in 10% FCS-DMEM (1 mL). The fluorescence intensity of a suspension (20 μL) of cells was measured (λdet = 800 nm). The fluorescence intensity of the cell lysate, which was prepared by treating a suspension (20 μL) of cells with Triton X100 (2.2 μL, 10% PBS solution), was measured (λdet = 800 nm) to determine the amounts of dyes delivered to cells.

Figure 4. Preparation of PEGylated TBD 3.

PEG chain was also prepared from 2b in a similar manner. The UV−vis absorption and photoluminescence spectra of the obtained PEG−dye conjugates were measured (Figure 5 and Table 1). From the absorption spectrum of PEG−PPAB conjugate 1 in DMF, a strong absorption peak at 772 nm was observed together with a shoulder around 710 nm and a broad peak around 500−600 nm. By measuring in H2O, a strong peak observed in the NIR region in DMF was broadened and bathochromically shifted, although the broad peak in the visible region was not changed. This indicates that PPAB moieties formed aggregates in self-assemblies of 1, but weakly interacted with each other. In the case of PEG−TBD conjugate 3a, a strong peak at 765 nm with a shoulder and a sharp peak at 566 nm were observed in DMF. In H2O, all three absorption peaks were broadened and bathochromically shifted. The absorbance maximum of dimers or H-aggregates at 720 nm was slightly stronger than that of monomeric TBD at 770 nm. These results suggested that a large number of TBD moieties in selfassemblies tightly interacted with each other to form aggregates in a hydrophobic core. In the photoluminescence measurement, no fluorescence emission of 1 and 3a was observed in H2O, although emission around 750−850 nm was observed in DMF (Figure 5c,d). These data pointed out that both weakly and strongly interacted dye aggregates in self-assemblies might efficiently release the excited photoenergy as heat through nonradiative relaxation, thereby generating a strong PA signal. The UV−vis absorption spectra of PEG−TBD conjugate 3b were quite similar to those of 3a (see Figure S2), indicating that TBD moieties were similarly aggregated in self-assemblies in H2O. The diameters of self-assemblies of PEG−dye conjugates 1, 3a, and 3b measured using DLS are summarized in Table 2. The standard deviation of the hydrodynamic diameter of 1 was slightly larger than those of 3, indicating that an aqueous solution of PPAB derivative 1 contains self-assemblies with various sizes. The particle morphology of PEG−dye conjugates was evaluated by TEM. In the TEM observation of PPAB derivative 1, small and large spherical self-assemblies, whose diameters were ∼20 nm and 70−100 nm, respectively, without a layer structure were found (Figure 6a). By considering the theoretical length of 1 (∼20 nm), these are considered to be micelles and multimicellar aggregates, respectively. In contrast, TBD derivatives 3 having one or two PEG chains formed selfassemblies with a slightly rough surface (Figure 6b,c). In light of a layered structure, self-assemblies of 3a are thought to be monolayer capsules in which TBDs interact with each other and align in the shell (Figure 7a). Because it was reported that TBD derivatives form aggregates having head−head and tail− tail interactions,25 similar aggregates like dimers or Haggregates were considered to be formed in the selfassemblies.35,36 Monolayer capsules of 3b seemed to be unstable because of the lower PEG density, therefore, 3b



RESULTS AND DISCUSSION PEGylated PPAB and TBD Derivatives and Their SelfAssemblies. PPAB 1 and TBD 3a conjugated with two PEG chains through aromatic rings were synthesized from the known PPAB derivative24 and TBD derivative 2a,25 respectively (Figures 3 and 4, and see also the SI). For comparison with dyes having two PEG chains, TBD derivative 3b having one

Figure 3. PEGylated PPAB 1. 251

DOI: 10.1021/acs.biomac.6b01568 Biomacromolecules 2017, 18, 249−256

Article

Biomacromolecules

Figure 5. Absorption spectra (0.1 mg/mL) of (a) 1 and (b) 3a in DMF (solid) and H2O (dashed). Photoluminescence spectra (λex = 720 nm) of (c) 1 and (d) 3a in DMF (blue) and H2O (red).

Table 1. UV−Vis Absorbance of PEG−Dye Conjugatesa

Upon irradiating aqueous solutions of 1 and 3 with pulsed laser light at 720 and 790 nm, ultrasonic waves were detected using a piezoelectric element. Representative PA signals generated upon irradiation are shown in Figure S7. To demonstrate clearly the conversion efficiency of absorbed photoenergy to PA signal, we normalized the PA signals to the laser power as well as to the absorbance of sample solutions at the irradiated wavelength. The normalized intensities of PA signals, as well as the relative intensities of PA signals compared with those obtained using commercially available ICG dye, are summarized in Table 3. Both the PPAB and the TBD derivatives generated PA signals more efficiently than ICG. Upon irradiating both at 720 and 790 nm, self-assemblies of 3 having TBD moieties generated stronger PA signals than 1. It is known that the PA signal intensity is proportional to the elastic modulus and the volumetric expansion coefficient of the substance.2,4 The difference in PA signal generation between PPAB and TBD might be caused by the strong interaction of TBDs producing aggregates with high elastic modulus. Although the amounts of conjugated PEG were different, there was no significant difference of PA signal intensities between 3a and 3b. Therefore, the PA signal generation caused by thermal energy transfer to PEG chains from excited TBDs was negligible in these cases. Because PEGylated TBD derivatives formed more stable self-assemblies under dilute conditions and generated a PA signal more efficiently than PEGylated PPAB, we selected TBD as a probe molecule for PA tumor imaging. TBD-Conjugated HA Derivatives and Their SelfAssemblies. PEGylated HA derivatives having ICG dyes could efficiently visualize tumor tissues by a PA imaging device, as reported.15 To enhance PA signal intensities, we next focused our attention on HA derivatives conjugating with TBD

λabs,max (nm) PEG−dye conjugates

in DMF

in H2O

1 3a 3b

772 (1.9), 536 (0.84) 765 (6.5), 566 (6.1) 771 (6.5), 566 (6.5)

779 (1.2), 537 (0.87) 770 (3.5), 722 (3.7), 576 (3.7) 775 (4.2), 722 (4.4), 575 (4.7)

a Sample solutions (0.1 mg/mL) in DMF and H2O were measured. Concentrations were calculated by using their theoretical molecular weights. Molar extinction coefficient ε ( × 104 M−1cm−1) was shown in parentheses.

Table 2. Properties of Self-Assemblies of PEG−Dye Conjugates PEG−dye conjugates

DDLSa (nm)

DTEMb (nm)

cacc (×10−3 g/L)

1 3a 3b

118 ± 59 135 ± 36 135 ± 35

86 ± 22 89 ± 18

6.3 0.066 0.032

a Hydrodynamic diameters determined by DLS. bDiameters determined by TEM. cDetermined by SLS.

would form vesicles rather than monolayer capsules (Figure 7b). The cac value of 1, determined by SLS, was lower than those of 3a and 3b. TBDs were aligned to form a layered structure and each TBD showed strong interaction, preventing their disassembly under dilute conditions. We reported that nanoparticles with low cac values circulate in blood vessels for a longer time and more efficiently accumulate in tumor tissues by the EPR effect.37,38 Based on cac values (Table 2), TBD derivatives were considered to be better tumor-imaging probes than the PPAB derivative. 252

DOI: 10.1021/acs.biomac.6b01568 Biomacromolecules 2017, 18, 249−256

Article

Biomacromolecules

Figure 6. TEM images of (a) 1, (b) 3a, and (c) 3b.

conjugated with 36% of azide side chain, and others were conjugated with cross-linker 2a. Both the UV−vis and photoluminescence spectra of 4 in DMF and H2O were quite similar to those of 3, suggesting that TBDs were successfully conjugated with HA and interacted with each other in self-assemblies (Figures S4). To check the stability against photobleaching, we measured the UV−vis absorbance of TBD- and ICG-conjugates after continuous irradiation with vis−NIR light. As expected, TBD conjugates 4 showed excellent photostability, while ICG conjugate 5 underwent gradual photobleaching (Figures 10 and S9). The hydrodynamic diameters of 4a and 4b measured by DLS are shown in Table 4. Although ICG-conjugated HA derivative 5 formed multimicellar aggregates in H2O,15 the morphology of self-assemblies of 4 observed by TEM seems to be monolayered capsules, which are quite similar to those of 3 (Figures S5). Despite the similar morphology, the cac values of 4 were much larger than those of 3. This indicates that TBD dyes interacted with each other, but the interaction was not as strong as that in self-assemblies of 3 (Figure 11). HA main chains might prevent maintaining a proper distance between TBDs for interaction. Because 4a contains both intermolecular and intramolecular cross-linkers, slightly disordered selfassemblies were formed, producing a higher cac value than for 4b. Interestingly, PA signal intensities of 4 at both 720 and 790 nm were much stronger than those of ICG itself and HA derivative 5 conjugating with ICG moieties. Because it is known that the formation of a cross-linked core in HA cryogel scaffolds enhanced their elastic moduli,33 HA derivatives 4 would form self-assemblies with a high elastic modulus, generating a strong PA signal. To enhance the PA signal of dye-conjugated polymeric self-assemblies, two factors are essential: (a) to select an NIR dye that tends to form dimers or H-aggregates and can readily generate heat through photoirradiation and (b) to use a polymer, such as hyaluronic acid, that tends to form selfassemblies with a high elastic modulus. As HA derivative 4b generated PA signals more efficiently than PEGylated TBDs 3 and HA derivative 4a, we next examined in vitro and in vivo experiments utilizing 4b. PA Tumor Imaging Utilizing 4b and Its In Vivo Kinetics. Before tumor imaging in vivo, we evaluated the stability of 4b under physiological conditions. Both the UV−vis absorption spectra and PA signal intensities in PBS (pH 7.4) and FCS measured at 790 nm were unchanged after 24 h (see the SI). From these data, HA derivative 4b was expected to keep its self-assembled structure and efficiently generate a PA signal in vivo. The fluorescence images of tumor-bearing mice 24 h after injection of 4b were collected (Figure 12a). Because 4b does not emit fluorescence in H2O efficiently, 4b could visualize the tumor tissues (colon 26 cells), which were subcutaneously

Figure 7. Plausible interaction of TBDs in self-assemblies of 3.

Table 3. Relative Intensities of PA Signals of PEG−Dye Conjugates Compared with ICG Dyea PA contrast agent

PA (V/J·abs) at 720 nm

PA/PAICG at 720 nm

PA′ (V/J·abs) at 790 nm

PA′/PA′ICG at 790 nm

1 3a 3b ICG

47 55 56 38

1.23 1.44 1.47 1

49 63 53 45

1.08 1.40 1.17 1

a PA signals (PA and PA′) of dyes in H2O (concentration (dye) = 1.0 × 10−4 ∼ 1.0 × 10−5 M) were measured by irradiating a sample solution with pulse laser light at 720 and 790 nm, respectively. PA and PA′ were normalized by the power of pulse laser and the absorbance of dyes. The PA signals of commercially available ICG dye at 720 nm (PAICG) and 790 nm (PA′ICG) were similarly measured.

dyes and PEG moieties. PEG- and TBD-conjugated HA derivatives 4 were synthesized from the known HA having azido moieties with TBD 2 through a copper-mediated [3 + 2] cycloaddition reaction (Figure 8), according to the preparation methods for PEG- and ICG-conjugated HA derivative 515 (Figure 9). The conjugation of difunctionalized TBD 2a with HA having azido moieties gave 4a in low yield because of the formation of highly cross-linked insoluble polymers, in which TBD 2a acts as either intermolecular or intramolecular crosslinker. In contrast, HA derivative 4b conjugating with monofunctionalized TBD 2b was obtained in good yield. The conjugating efficiencies of TBD to PEG moieties were determined by comparing the absorbance (λmax = 780 nm) of the TBD moieties in the HA derivatives in DMF with that of 2 as a standard compound. In the case of 4a, PEG moieties were 253

DOI: 10.1021/acs.biomac.6b01568 Biomacromolecules 2017, 18, 249−256

Article

Biomacromolecules

Figure 8. PEG- and TBD-conjugated HA derivatives 4.

PA imaging in vivo was carried out before and 24 h after intravenous injection of 4b. Before injection, the blood vessels around the tumor site were imaged by irradiation with pulsed laser light at 790 nm because of PA signal generation by hemoglobin (Figure 12c,d). Although weak background signals from blood vessels were also detected, PA signals generated from the tumor site were enhanced after 24 h. PA signals around the tumor site detected 24 h after injection of 4b were 2.5 ± 0.8 times stronger than those detected before injection. PA signals generated from the tumor-bearing side of mice were 2.0 ± 1.4 times as strong as those from the non-tumor-bearing side 24 h after injection. These results indicate that 4b efficiently circulated in blood vessels for a long time and accumulated at the tumor site through passive tumor targeting. To check this, we next estimated the tumor specificity and blood concentration using an optical imaging device. The tumor specificity, accumulated ratio of 4b in the tumor site with respect to the total ID, was determined as 17.8 ± 3.5%ID/g (tumor) by measuring fluorescence intensities of 4b wextracted into DMF from the harvested tumor tissues. By measuring fluorescence intensities of 4b in blood harvested 24 h after injection, its blood concentration was determined as 15.6 ± 4.9%ID (blood). As reported, the tumor specificity and blood concentration of 5 are 11.8 ± 2.7%ID/g (tumor) and 15.0 ± 4.0%ID (blood), respectively.15 Both self-assemblies 4b and 5 could efficiently avoid trapping from the reticuloendothelial system thanks to a PEG shell, while 4b showed more efficient accumulation in tumor tissues through passive tumor targeting. Because both of cac values and diameters of nanoparticles were quite similar, we considered that the influence of these two factors, stability and size of nanoparticles, on tumor specificity would be small. Fox, Szoka, and Fréchet reported that flexible cyclic polymers can more readily deform to pass through a pore than rigid spherical polymers.39 Because of their monolayered structure with a slightly rough surface, self-assemblies of 4b were considered to be more flexible than multimicellar aggregates consisting of 5, showing better tumor specificity. As described above, self-assemblies of 4b are good PA contrast agents, which could accumulate in tumor tissues efficiently, but they were not so effective in optical imaging. Self-assemblies of 5 could be applicable to both PA and optical imaging. To gain

Figure 9. PEG- and ICG-conjugated HA derivative 5.

Figure 10. Time-dependent decay of absorbance at 780 nm for solutions of 4b (circle), 5 (cross), and ICG (square).

Table 4. Properties of Self-Assemblies of TBD-Conjugated HA Derivatives 4 PEG-dye-HA conjugates

DDLSa (nm)

cacb (×10−3 g/L)

PA/PAICGc at 720 nm

PA′/PA′ICGc at 790 nm

4a 4b 5d

204 ± 94 138 ± 59 157 ± 38

2.4 0.59 0.79

1.60 1.83 1.20

1.49 1.67 1.25

a

Hydrodynamic diameters determined by DLS. bDetermined by SLS. PA/PAICG and PA′/PA′ICG were similarly determined according to the footnote written in Table 3. dThe data were reported in ref 15. c

inoculated in the lower back (Figure 12b), with low contrast by optical imaging in vivo (signal-to-noise ratio (S/N): 1.5 ± 0.1). In contrast, HA derivative 5 visualized tumor tissues with higher contrast (S/N: 2.2 ± 0.4), as reported.15 254

DOI: 10.1021/acs.biomac.6b01568 Biomacromolecules 2017, 18, 249−256

Article

Biomacromolecules

Figure 11. Plausible interaction of TBDs in self-assemblies of 4.

Figure 13. Conceptual illustration of the behavior of (a) 4b and (b) 5 in tumor tissues.

both optical and PA tumor imaging, while TBD-conjugated HA derivative 4b is suitable for PA tumor imaging. Although the reason why the cell internalization behavior is different is not clear, the difference in the morphology of self-assemblies might influence their fate in living cells.

Figure 12. (a) NIR fluorescence images (cps) of tumor-bearing mice 24 h after injection of 4b (100 μL of a PBS solution, [TBD] = 5.0 × 10−4 M). Color code: low intensity black, high intensity yellow. (b) PA images were taken from a right side (non-tumor-bearing side) and a left side (tumor-bearing side). Representative PA images of tumorbearing mice before and 24 h after injection of 4b taken from (c) a right side and (d) a left side. Arrows point out tumor sites. Thresholds were appropriately established. Mouse colon cancer cells (colon 26, 1 × 106, RIKEN) were subcutaneously inoculated into the lower back of nude mice 1 week before the imaging experiment. Because of the specifications of PA imaging apparatus, the law data shown were obtained as inverted images.



CONCLUSIONS We have developed PEG-conjugated BODIPY dimer derivatives, bearing PPAB and TBD skeletons, as well as TBD- and PEG-conjugated HA derivatives employing facile conjugation methods. Both PPAB- and TBD-conjugated amphiphiles formed self-assemblies and more efficiently generated PA signals by the irradiation of NIR pulsed laser light, than the clinically approved ICG dye. The following two factors are essential to enhance PA signals from polymeric self-assemblies: (1) the formation of strongly interacted TBD aggregates (dimers or H-aggregates) and (2) enhancement of the elastic modulus of self-assemblies by conjugating TBDs with HA. The conjugation ratio of TBDs with HA, the volume of a hydrophilic shell, and the structure of BODIPY dimers can influence these two factors. From the in vitro and in vivo experiments of HA derivatives, we showed that ICG-conjugated self-assemblies internalized into cells or merged to cell walls to emit strong fluorescence, while the TBD-conjugated selfassemblies were not internalized because of their self-assembled structure. The difference in cell internalization behavior significantly affects their fluorescence intensities and the contrast in optical tumor imaging in vivo. Because the disassembly of TBD-conjugated HA derivative is suppressed, aggregated TBDs efficiently generate a strong PA signal in tumor tissues. By precisely controlling the aggregation of dyes as well as their self-assembled structure in tumor tissues, probes with more practical performance in bimodal imaging will be accessible. Research involving the enhancement of PA signals by tuning these factors is currently under investigation in our laboratory.

insight into the optical and PA active behavior, we next examined the cell uptake of each self-assembly. Mouse colon cancer cells were incubated with ICG (control) and each self-assemblies of 4b and 5. By measuring the fluorescence intensities of cell lysates, 4b was scarcely internalized into cells during 24 h incubation (cell uptake: 0.10%; Figure S12a). In contrast, 16% of treated amounts of 5 were internalized into cells or merged with cell walls. Interestingly, strong fluorescence from 5 was detected from cells suspended by treatment with trypsin after incubation, while weak fluorescence of 4b was detected (Figure S12b). Based on these results, we supposed that the behavior of selfassemblies of 4b and 5 was different after accumulation in tumor tissues (Figure 13). After the permeation of selfassemblies into tumor tissues by the EPR effect,16 selfassemblies of 4b cannot be internalized into cells and retain their nonemissive self-assembled structure. On the other hand, a large number of self-assemblies of 5 accumulated in tumor tissues can be internalized into cells or merged with cell walls, thereby being disassembled emissive form in tumor cells. Therefore, ICG-conjugated HA derivative 5 is applicable to 255

DOI: 10.1021/acs.biomac.6b01568 Biomacromolecules 2017, 18, 249−256

Article

Biomacromolecules



(19) Ni, Y.; Wu, J. Org. Biomol. Chem. 2014, 12, 3774−3791. (20) Kamkaew, A.; Lim, S. H.; Lee, H. B.; Kiew, L. V.; Chung, L. Y.; Burgess, K. Chem. Soc. Rev. 2013, 42, 77−88. (21) Li, H.; Zhang, P.; Smaga, L. P.; Hoffman, R. A.; Chan, J. J. Am. Chem. Soc. 2015, 137, 15628−15631. (22) Frenette, M.; Hatamimoslehabadi, M.; Bellinger-Buckley, S.; Laoui, S.; La, J.; Bag, S.; Mallidi, S.; Hasan, T.; Bouma, B.; Yelleswarapu, C.; Rochford, J. J. Am. Chem. Soc. 2014, 136, 15853− 15856. (23) Shimizu, S.; Iino, T.; Araki, Y.; Kobayashi, N. Chem. Commun. 2013, 49, 1621−1623. (24) Shimizu, S.; Iino, T.; Saeki, A.; Seki, S.; Kobayashi, N. Chem. Eur. J. 2015, 21, 2893−2904. (25) Saino, S.; Saikawa, M.; Nakamura, T.; Yamamura, M.; Nabeshima, T. Tetrahedron Lett. 2016, 57, 1629−1634. (26) Sakamoto, N.; Ikeda, C.; Yamamura, M.; Nabeshima, T. Chem. Commun. 2012, 48, 4818−4820. (27) Akers, W. J.; Kim, C.; Berezin, M.; Guo, K.; Fuhrhop, R.; Lanza, G. M.; Fischer, G. M.; Daltrozzo, E.; Zumbusch, A.; Cai, X.; Wang, L. V.; Achilefu, S. ACS Nano 2011, 5, 173−182. (28) Hu, W.; Ma, H.; Hou, B.; Zhao, H.; Ji, Y.; Jiang, R.; Hu, X.; Lu, X.; Zhang, L.; Tang, Y.; Fan, Q.; Huang, W. ACS Appl. Mater. Interfaces 2016, 8, 12039−12047. (29) Ni, Y.; Kannadorai, R. K.; Peng, J.; Yu, S. W.-K.; Chang, Y.-T.; Wu, J. Chem. Commun. 2016, 52, 11504−11507. (30) Lin, W.; Zhang, W.; Sun, T.; Gu, J.; Xie, Z.; Jing, X. Langmuir 2016, 32, 9575−9581. (31) Dragulescu-Andrasi, A.; Kothapalli, S.-R.; Tikhomirov, G. A.; Rao, J.; Gambhir, S. S. J. Am. Chem. Soc. 2013, 135, 11015−11022. (32) Zhang, D.; Zhao, Y.-Z.; Qiao, Z.-Y.; Mayerhöffer, U.; Spenst, P.; Li, X.-J.; Würthner, F.; Wang, H. Bioconjugate Chem. 2014, 25, 2021− 2029. (33) Seidlits, S. K.; Khaing, Z. Z.; Petersen, R. R.; Nickels, J. D.; Vanscoy, J. E.; Shear, J. B.; Schmidt, C. E. Biomaterials 2010, 31, 3930−3940. (34) Ghosh, A.; Yusa, S. I.; Matsuoka, H.; Saruwatari, Y. Langmuir 2011, 27, 9237−9244. (35) Tleugabulova, D.; Zhang, Z.; Brennan, J. D. J. Phys. Chem. B 2002, 106, 13133−13138. (36) Okada, D.; Nakamura, T.; Braam, D.; Dao, T. D.; Ishii, S.; Nagao, T.; Lorke, A.; Nabeshima, T.; Yamamoto, Y. ACS Nano 2016, 10, 7058−7063. (37) Miki, K.; Kimura, A.; Oride, K.; Kuramochi, Y.; Matsuoka, H.; Harada, H.; Hiraoka, M.; Ohe, K. Angew. Chem., Int. Ed. 2011, 50, 6567−6570. (38) Miki, K.; Hashimoto, H.; Inoue, T.; Matsuoka, H.; Harada, H.; Hiraoka, M.; Ohe, K. Small 2014, 10, 3119−3130. (39) Fox, M. E.; Szoka, F. C.; Fréchet, J. M. J. Acc. Chem. Res. 2009, 42, 1141−1151.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biomac.6b01568. Synthesis of contrast agents, details in tumor imaging, NMR spectra, and other characterization details (PDF).



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Phone: +81 75 383 2497. *E-mail: [email protected]. Phone: +81 75 383 2495. ORCID

Kouichi Ohe: 0000-0001-8893-8893 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by a Grant-in-Aid for Young Scientists (A) (No. 26708024), a Grant-in-Aid for Scientific Research on Innovative Areas “New Polymeric Materials Based on Element-Blocks (No. 2401)” (15H00723, 15H00739, and 15H00756) of The Ministry of Education, Culture, Sports, Science, and Technology, Japan. We acknowledged Prof. Teruyuki Kondo and Associate Prof. Yu Kimura in Kyoto University for TEM measurement.



REFERENCES

(1) Zackrisson, S.; van de Ven, S. M. W. Y.; Gambhir, S. S. Cancer Res. 2014, 74, 979−1004. (2) Kim, C.; Favazza, C.; Wang, L. V. Chem. Rev. 2010, 110, 2756− 82. (3) Ntziachristos, V.; Razansky, D. Chem. Rev. 2010, 110, 2783−94. (4) Wang, L. V.; Wu, H.-i. Biomedical Optics: Principles and Imaging; Wiley-Interscience, Hoboken, NJ, 2007. (5) Luo, S.; Zhang, E.; Su, Y.; Cheng, T.; Shi, C. Biomaterials 2011, 32, 7127−7138. (6) Gao, F.; Bai, L.; Feng, X.; Tham, H. P.; Zhang, R.; Zhang, Y.; Liu, S.; Zhao, L.; Zheng, Y.; Zhao, Y. Small 2016, 12, 5239−5244. (7) Chen, Q.; Liu, X.; Zeng, J.; Cheng, Z.; Liu, Z. Biomaterials 2016, 98, 23−30. (8) Cash, K. J.; Li, C.; Xia, J.; Wang, L. V.; Clark, H. A. ACS Nano 2015, 9, 1692−1698. (9) Sreejith, S.; Joseph, J.; Lin, M.; Menon, N. V.; Borah, P.; Ng, H. J.; Loong, Y. X.; Kang, Y.; Yu, S. W.-K.; Zhao, Y. ACS Nano 2015, 9, 5695−5704. (10) Schaafsma, B.; Mieog, J. S. D.; Hutteman, M.; van der Vorst, J. R.; Kuppen, P. J. K.; Löwik, C. W. G. M.; Frangioni, J. V.; van de Velde, C. J. H.; Vahrmeijer, A. L. J. Surg. Oncol. 2011, 104, 323−332. (11) Kokudo, N.; Ishizawa, T. Liver Cancer 2012, 1, 15−21. (12) Hu, D.; Liu, C.; Song, L.; Cui, H.; Gao, G.; Liu, P.; Sheng, Z.; Cai, L. Nanoscale 2016, 8, 17150−17158. (13) Wang, H.; Liu, C.; Gong, X.; Hu, D.; Lin, R.; Sheng, Z.; Zheng, C.; Yan, M.; Chen, J.; Cai, L.; Song, L. Nanoscale 2014, 6, 14270− 14279. (14) Hannah, A.; Luke, G.; Wilson, K.; Homan, K.; Emelianov, S. ACS Nano 2014, 8, 250−259. (15) Miki, K.; Inoue, T.; Kobayashi, Y.; Nakano, K.; Matsuoka, H.; Yamauchi, F.; Yano, T.; Ohe, K. Biomacromolecules 2015, 16, 219−227. (16) Matsumura, Y.; Maeda, H. Cancer Res. 1986, 46, 6387−6392. (17) Holzer, W.; Mauerer, M.; Penzkofer, A.; Szeimies, R.-M.; Abels, C.; Landthaler, M.; Bäumler, W. J. Photochem. Photobiol., B 1998, 47, 155−164. (18) Kowada, T.; Maeda, H.; Kikuchi, K. Chem. Soc. Rev. 2015, 44, 4953−4972. 256

DOI: 10.1021/acs.biomac.6b01568 Biomacromolecules 2017, 18, 249−256