Indocyanine green-incorporated exosomes for improved in vivo imaging of sentinel lymph node

For improved physical stability and fast accumulation in sentinel lymph nodes in vivo, indocyanine green (ICG) was incorporated into exosomes (EXOs) to form a novel ICG-based near-infrared fluorescence probe. Compared with free ICG, prepared ICG-TBAI-encapsulated EXOs (ICG/EXOs) smaller than 50 nm had greater physicochemical stability in aqueous solution when exposed to light at 37 °C. In addition, sentinel lymph nodes were visualized within minutes with ICG/EXOs at a low ICG dose (0.05 mg/kg) owing to its well-controlled size distribution. Taken together, the results show that ICG/EXOs are effective biocompatible and sensitive near-infrared probes for real-time sentinel lymph node imaging in vivo.


Introduction
Biocompatible probes with strong fluorescence intensities (FIs) and low background signals at sites of detection are essential for optical diagnosis and imaging in vivo. Indocyanine green (ICG) is a widely used FDA-approved nearinfrared fluorescence probe for sentinel lymph node biopsy during surgery (Desmettre et al. 2000;Chi et al. 2013;Watkin and La-Touche 2014). However, external factors such as light and temperature can lessen the FI of ICG in vivo (Engel et al. 2008). Accordingly, ICG has been formulated into nanosized particles and hydrogels for enhanced stability in aqueous solution (Mok et al. 2012;Kim et al. 2014).
Exosomes (EXOs) have a natural phospholipid bilayer composed of cholesterol, sphingolipids, glycerophospholipids, and ceramide (Caradec et al. 2014). They are released from cells into diverse body fluids such as plasma, saliva, and urine via exocytosis (Tan et al. 2013;Caradec et al. 2014;Johnsen et al. 2014). Exosomal proteins and genes have crucial biological functions, including immune modulation, tumor progression, induction of angiogenesis, cell invasion, and metastasis (Tan et al. 2013;Caradec et al. 2014;Villarroya-Beltri et al. 2014). Recently, EXOs with hydrodynamic sizes of approximately 30-100 nm have been harnessed not only as nature-derived nanomaterials for the delivery of small molecules and small interfering RNAs but also as diagnostic biomarkers (Sun et al. 2010;Alvarez-Erviti et al. 2011;Johnsen et al. 2014). Given that the hydrodynamic size of particles is critical for in vivo distribution and successful localization in sentinel lymph nodes, EXOs may be appropriate carriers for the efficient delivery of therapeutic molecules and imaging agents to lymphatic systems and tumor tissues (Cabral et al. 2011;Noh et al. 2012). However, to the best of our knowledge, EXO-based probes have not yet been studied for lymph node imaging.
In this study, ICG was incorporated into EXOs, and the combination was evaluated for real-time imaging of sentinel lymph nodes. Optimal loading efficiency and hydrodynamic size were determined for ICG-loaded EXOs. In addition, fluorescence stability and optimal hydrodynamic size for long-term incubation were examined in aqueous solutions. Furthermore, ICG-tetrabutylammonium iodide & Hyejung Mok hjmok@konkuk.ac.kr (TBAI) and ICG-TBAI-encapsulated EXOs (ICG/EXOs) were administered to mice forepaws via transdermal injection, and fluorescence imaging was monitored in realtime with an in vivo imaging instrument.

Materials
Fetal bovine serum (FBS) was purchased from Gibco BRL (USA

Preparation of ICG/EXOs
The EXOs in this study were isolated from FBS according to the manufacturer's protocol. FBS (1 mL) was mixed with ExoQuick-TC EXO precipitation solution (200 lL) via inversion. After incubation at 4°C overnight, the mixture was centrifuged at 15009g for 30 min. After additional centrifugation at 15009g for 5 min, the pellet was resuspended with phosphate-buffered saline (PBS) solution and stored at 4°C. The amount of protein in the EXOs was determined using the BCA protein assay kit according to the manufacturer's protocol. For facile encapsulation of ICG by EXOs, hydrophobic ICG-TBAI salt was prepared as previously reported (Kim et al. 2010). ICG (16 mg) was mixed with TBAI (46 mg) at an ICG/ TBAI molar ratio of 1/6 in methanol. After evaporation of methanol, ICG-TBAI complex was dissolved in DMSO for further experiments.

Characterization of ICG/EXOs
EXOs (protein amount, 40 lg) were mixed with ICG-TBAI at various ICG amounts of 0, 2, 4, and 20 lg (ICG/exosomal protein weight ratios of 0, 0.05, 0.1, and 0.5, respectively) at 37°C for 2 h. After incubation, free ICG-TBAI was removed with a desalting column (MWCO, 40 kDa To examine changes in EXO and ICG/EXO particle sizes, we incubated two types of EXOs in PBS solution for 4 days at 4°C and used DLS to measure particle size at predetermined times.

Sentinel lymph node imaging
For in vivo monitoring of ICG/EXOs, six-week-old female Balb-c nude mice were anesthetized with an intraperitoneal injection of Rompun/Zoletil according to the manufacturer's protocol. Then, free ICG-TBAI and ICG/EXOs were injected into the forepaws of the mice at a low dose (0.05 mg/kg) and a high dose (0.3 mg/kg) of ICG. After the injection of the ICG-based samples and incubation for predetermined times (1, 5, 10, 30, 60, and 120 min and 24 h), fluorescence images of each mouse were obtained with the IVIS imaging system. The total flux of whole mice forepaws was also quantitatively analyzed with the IVIS imaging system.

Results and discussion ICG/EXOs preparation
After the isolation of EXOs from FBS, the amount of isolated exosomal protein was quantified with a BCA assay (Twu et al. 2013). The total amount of protein in the EXOs in PBS solution with detergent (1 % Tx-100) was similar to that in EXOs in solution without detergent, which suggests that the total exosomal proteins in the exterior and interior regions of the EXOs were measured with the BCA assay.
For facile entrapment of ICG within EXOs, ICG was complexed with TBAI via ionic interactions (Kim et al. 2010). As shown in Fig. 1, exosomal protein was incubated with ICG-TBAI for 2 h at 37°C, entrapping ICG within EXOs via hydrophobic interactions. After column purification of the resulting solution to remove free ICG-TBAI, the ICG/EXOs were administered to mice forepaws to examine the fluorescence signals in sentinel lymph nodes. In this study, we hypothesized that EXOs would be excellent vehicles for the delivery of ICG to sentinel lymph nodes due to their hydrodynamic size of approximately 50 nm with narrow size distribution after in vivo administration.

Characterization of ICG/EXOs
To determine the optimal weight ratio of ICG to exosomal protein, we incubated ICG with exosomal protein at various ICG/exosomal protein weight ratios. The hydrodynamic sizes of the resultant ICG/EXOs were examined with DLS. As shown in Fig. 2A, the EXOs exhibited a narrow size distribution with a size of 35.8 ± 3.1 nm. Although the size distribution of the ICG/EXOs at low ICG/exosomal protein weight ratios was similar to that of EXOs alone, large aggregation with broad size distribution was observed in ICG/EXOs at an ICG/exosomal protein weight ratio of 0.5. Figure 2B shows that compared with that of EXOs, the hydrodynamic size of ICG/EXOs up to an ICG/exosomal protein weight ratio of 0.2 was much larger. The sizes of ICG/EXOs at ICG/exosomal protein weight ratios of 0.5 and 0.2 were 82.4 ± 45.5 and 46.2 ± 11.1 nm, respectively. As the ICG/exosomal protein weight ratio deceased, the hydrodynamic size of the resulting ICG/EXOs decreased accordingly. The sizes of ICG/EXOs at ICG/exosomal protein weight ratios of 0.1 and 0.05 were 40.9 ± 1.8 and 46.3 ± 14.3 nm, respectively.
The entrapment efficiency of ICG in EXOs was also examined at various ICG/exosomal protein weight ratios (Fig. 2C). The entrapment efficiencies of ICG within EXOs at ICG/exosomal protein weight ratios of 0.05, 0.1, and 0.2 were 5.4 ± 2.6, 14.0 ± 4.3, and 13.0 ± 4.2 %, respectively. Taken together, ICG/EXOs were prepared at an ICG/exosomal protein weight ratio of 0.1 for further experiments due to the small hydrodynamic size and high entrapment efficiency at this ratio.

Stability of ICG/EXOs
ICG is extremely vulnerable to external physicochemical conditions such as light, pH, and temperature (Saxena et al. 2004). The relative FIs of ICG and ICG/EXOs were examined after incubation at 37°C in the presence of light for predetermined times (Fig. 3A). Although both free ICG and ICG-TBAI exhibited severe FI loss, the FI of ICG/ EXOs was not decreased during incubation. This result indicates that the incorporation of ICG into EXOs may protect ICG from physicochemical damage from light exposure in solution.
The stability of ICG/EXOs was also determined by measuring their sizes after incubation at 4°C for 4 days, as shown in Fig. 3B. No noticeable change in the size of the ICG/EXOs was observed for 2 days, whereas the hydrodynamic size of the ICG/EXOs gradually increased after 3 days. However, no change in the size of the EXOs occurred during incubation at 4°C for 4 days. The nonspecific aggregation of ICG/EXOs during extended incubation might be attributable to the presence of ICG on EXO surfaces. Accordingly, freshly prepared ICG/EXOs were used for the in vivo imaging of sentinel lymph nodes.  In previous studies, ICG and formulated ICG have been administered via transdermal injection at a concentration of 0.08-1.25 mg/kg (Robinson et al. 2013). In the present study, ICG-TBAI and ICG/EXOs containing a low dose of ICG (0.05 mg/kg) were injected into mice forepaws. Fluorescence signals in sentinel lymph nodes were then monitored in real-time, as shown in Fig. 4A. As a control, ICG-TBAI containing a high dose of ICG (0.3125 mg/kg) was also administered. ICG-TBAI with low-dose ICG had a poor fluorescence signal in sentinel lymph nodes, whereas ICG/EXOs with low-dose ICG had a rapid and strong fluorescence signal. ICG-TBAI with high-dose ICG showed fluorescent signals comparable to those of ICG/ EXOs with low-dose ICG. The FIs of each lymph node were also analyzed quantitatively (Fig. 4B). The total flux measurements of gated sentinel lymph nodes were 0.4 ± 0.3, 1.9 ± 0.9, and 2.2 ± 0.8 (910 9 ) photons/s after treatment with ICG-TBAI (low-dose ICG), ICG/EXOs (low-dose ICG), and ICG-TBAI (high-dose ICG), respectively, and incubation for 2 min. These results demonstrated that ICG/EXOs provide strong and fast imaging of sentinel lymph nodes with a low dose of ICG. Interestingly, mice treated with free ICG-TBAI (high-dose ICG) showed increased FI after 120 min while no change in FI was observed in mice treated with ICG/EXOs. It might be attributed to fast translocation of ICG/EXOs into sentinel lymph nodes within several minutes, compared to free ICG-TBAI. In a previous study, particles with sizes of 10-50 nm were effective for fast translocation into sentinel lymph nodes and extended accumulation (Noh et al. 2012). ICG/EXOs with wellcontrolled sizes will likely allow favorable and fast localization to lymph nodes and result in clear visualization of sentinel lymph nodes at a low ICG concentration.
In this study, ICG/EXOs were fabricated as novel ICGbased near-infrared probes for real-time monitoring of sentinel lymph nodes. ICG/EXOs with a size of *50 nm exhibited stable FI in the presence of exterior light. In addition, ICG/EXOs containing low-dose ICG provided fast and strong visualization of sentinel lymph nodes after transdermal injection owing to their well-controlled size distribution. Taken together, the results of this study indicate that these nanosized near-infrared ICG/EXO probes could allow fast and excellent real-time monitoring of not only sentinel lymph node but also other organs and tissues, including tumors with low vascularization.