Hydrophobic insertion-based engineering of tumor cell-derived exosomes for SPECT/NIRF imaging of colon cancer

Background Tumor cell-derived exosomes (TEx) have emerged as promising nanocarriers for drug delivery. Noninvasive multimodality imaging for tracing the in vivo trafficking of TEx may accelerate their clinical translation. In this study, we developed a TEx-based nanoprobe via hydrophobic insertion mechanism and evaluated its performance in dual single-photon emission computed tomography (SPECT) and near-infrared fluorescence (NIRF) imaging of colon cancer. Results TEx were successfully isolated from HCT116 supernatants, and their membrane vesicle structure was confirmed by TEM. The average hydrodynamic diameter and zeta potential of TEx were 110.87 ± 4.61 nm and –9.20 ± 0.41 mV, respectively. Confocal microscopy and flow cytometry findings confirmed the high tumor binding ability of TEx. The uptake rate of 99mTc-TEx-Cy7 by HCT116 cells increased over time, reaching 14.07 ± 1.31% at 6 h of co-incubation. NIRF and SPECT imaging indicated that the most appropriate imaging time was 18 h after the injection of 99mTc-TEx-Cy7 when the tumor uptake (1.46% ± 0.06% ID/g) and tumor-to-muscle ratio (8.22 ± 0.65) peaked. Compared with radiolabeled adipose stem cell derived exosomes (99mTc-AEx-Cy7), 99mTc-TEx-Cy7 exhibited a significantly higher tumor accumulation in tumor-bearing mice. Conclusion Hydrophobic insertion-based engineering of TEx may represent a promising approach to develop and label exosomes for use as nanoprobes in dual SPECT/NIRF imaging. Our findings confirmed that TEx has a higher tumor-targeting ability than AEx and highlight the potential usefulness of exosomes in biomedical applications.


Background
Colon cancer is an extremely complex and multifactorial disease, causing millions of deaths every year [1]. Although colonoscopy is widely used to diagnose colon cancer, it offers limited sensitivity and specificity for early-stage disease. Molecular imaging is a noninvasive or minimally invasive alternative, providing a detailed insight into the physiological and pathological processes of the human body; hence, molecular imaging is more likely to diagnose cancer at an early stage [2,3]. Each imaging technique offers different spatial resolution, sensitivity, depth of tissue penetration, cost, and time resolution. Multimodality imaging combines the advantages of different imaging technologies, providing comprehensive, three-dimensional information, as well as a more accurate spatial positioning and molecular information ideal for the detection of small lesions [4]. Hence, the development of multimodality molecular imaging agents has gained increasing attention over the last years. The agent requires a suitable carrier with certain intrinsic properties, such as large carrying capacity and facile surface modification. Synthetic nanoparticles, including liposomes, metal nanoparticles, and magnetic nanoparticles, have a broad clinical application in multimodality imaging [5][6][7]. However, most of them are artificial drug carriers possessing potential toxicity, immunogenicity, and inability to penetrate most organs [8].
Exosomes have emerged as promising natural nanocarriers due to their nontoxicity and biocompatibility [9,10]. They are extracellular vesicles of endosomal origin secreted by almost all types of cells [11]. In the past decade, exosomes have emerged as novel nanocarriers in drug delivery systems owing to their suitable particle size (30-150 nm), high stability, and large carrying capacity [12]. Furthermore, exosome-based drug delivery harnesses endogenous mechanisms for uptake, intracellular trafficking, and subsequent delivery of the cargo [10]. To date, different types of exosomes have been developed to deliver drugs to tumors; tumor cell-derived exosomes (TEx), adipose stem cell-derived exosomes (AEx), and epidermal cell-derived exosomes are among the most promising ones [13][14][15][16][17]. Exosomes from different cells have differential properties, and TEx have inherent tumor-targeting capabilities [18]. Additionally, various biomedical imaging modalities have been modified to trace exosomes. These modalities include magnetic resonance (MR), single-photon emission computed tomography (SPECT), positron emission tomography (PET), and optical imaging [19][20][21][22][23][24]; among these, SPECT and optical imaging are currently the most commonly used due to their low cost and wide availability. SPECT can be used to image the whole body and offers excellent penetration; however, SPECT imaging is limited by the relatively long acquisition time, short imaging time window, and low spatial resolution. Near-infrared fluorescence (NIRF, 650-1000 nm) imaging offers real-time and high-resolution tissue structure information [25,26], although tissue penetration is limited. Multimodality SPECT and NIRF imaging can provide complementary insight into disease progression and real-time tumor delineation [27]. To the best of our knowledge, there are currently no exosomebased nanoprobes for multimodality SPECT and NIRF imaging.
To visualize the exosomes, functional molecules should be introduced on exosomes' surfaces to modify exosomes for multimodal imaging. However, it remains a challenge to modify different functional groups on the surface of exosomes due to the small size and complex surface chemistry of exosomes [28]. Here, we propose a hydrophobic insertion strategy to modify TEx with DSPE-PEG 2000 -Cy7 and DSPE-PEG 2000 -HYNIC. The modified TEx (HYNIC-TEx-Cy7) was labeled with 99m Tc, allowing for SPECT and NIRF imaging of tumor-bearing nude mice in vivo. AEx labeled with 99m Tc and Cy7 ( 99m Tc-AEx-Cy7) were used for comparison.

TEx and AEx isolation and characterization
TEx and AEx were successfully isolated from the supernatant of tumor cells and adipose stem cells (ASCs), respectively. Expectedly, TEx and AEx appeared as membrane vesicles under a TEM (Fig. 1a, b). The average hydrodynamic diameters of TEx and AEx were 110.87 ± 4.61 nm and 136.47 ± 2.50 nm, respectively. The zeta potentials of TEx and AEx were -9.20 ± 0.41 mV and -7.22 ± 0.60 mV, respectively (Fig. 1c). The average hydrodynamic diameter and zeta potential of TEx did not change significantly for up to 4 days, indicated the excellent stability (Fig. 1d, e). A CCK-8 assay was performed to evaluate the cell cytotoxicity of exosomes to HCT116 colon cancer cells. HCT116 cancer cells or adipose stem cell were co-incubated with TEx and AEx at various concentrations (up to 200 μg/mL) and different time periods (up to 72 h). The results showed that the survival rate of cells in each group was greater than 90% ( Fig. 1f-i). TEx and AEx had no obvious toxicity to HCT116 colon cancer cell and adipose stem cell.

In vitro tumor cell binding
After a 12-h incubation of FITC-TEx with HCT116 cell, tumor cells were analyzed by flow cytometry. The fluorescence intensity of tumor cells was increased with increasing concentrations of FITC-TEx, reaching a maximum when cells were co-cultured with 20 μg/mL of FITC-TEx (Fig. 2a). As exhibited in fluorescence images, the uptake of Cy5-labeled TEx in HCT116 cell was increased over time, but the uptake did not change significantly after 12 h (Fig. 2b). The quantification of the fluorescent intensity was consistent with images ( Fig. 2c). Confocal microscopy of HCT116 cells incubated with Cy5-labeled TEx revealed strong fluorescence signals in the cell membrane and cytoplasm (Fig. 2d).

In vivo NIRF imaging
NIRF imaging was performed 1, 6, 12, 18, and 24 h after the administration of 99m Tc-TEx-Cy7 and 99m Tc-AEx-Cy7, and the changes in the biodistribution of the multimodality nanoprobes were observed over time. NIRF imaging revealed that 99m Tc-TEx-Cy7 (Fig. 4a) were taken up by tumor cells at a higher rate than 99m Tc-AEx-Cy7 (Fig. 4b). The quantification of the fluorescent intensity was consistent with NIRF images (Fig. 4c). The most appropriate time for NIRF imaging was 18 h after the injection of 99m Tc-TEx-Cy7.

In vivo SPECT imaging
SPECT imaging was performed 6, 12, 18, and 24 h after the administration of 99m Tc-TEx-Cy7 and 99m Tc-AEx-Cy7. Tumor-bearing mice exhibited an accumulation of the multimodality nanoprobe in the abdominal cavity. Compared with the 99m Tc-AEx-Cy7 group, the 99m Tc-TEx-Cy7 group exhibited a higher tumor uptake of the tracer (Fig. 5a, b). The most appropriate time for SPECT imaging was 18 h after the injection of 99m Tc-TEx-Cy7.

In vivo toxicity studies
BALB/c mice (n = 5) received an i.v. injection of 200 μL of PBS, or PBS containing 99m Tc-TEx-Cy7 or 99m Tc-AEx-Cy7 to evaluate the in vivo potential toxicity. No significant hepatic or renal toxicity was observed from the indicating normal values of liver and kidney function markers, including ALT, AST, ALP, BUN and CRE ( Fig. 7a-d). Also, we did not observe significantly evidence of major organ damage from the H&E stained sections (Fig. 7e).

Discussion
In this study, we employed a hydrophobic insertion method to label TEx with a radionuclide ( 99m Tc) and NIRF dye (Cy7) and develop a multimodality imaging nanoprobe ( 99m Tc-TEx-Cy7) targeting colon cancer. Our in vivo and in vitro findings indicated the high affinity of the probe for tumor cells. To the best of our knowledge, this is the first study to use TEx as a nanocarrier for multimodal SPECT and NIRF imaging.  TEx offer exhibited several favorable characteristics as natural nanocarriers, such as suitable nanoparticle sizes (110.87 ± 4.61 nm), negative zeta potential (-9.20 ± 0.41 mV), no obvious toxicity and high biocompatibility, making them ideal for various biomedical applications. As revealed by the results of DLS, the average hydrodynamic diameter and zeta potential remained similar for up to 4 days, indicating excellent stability. Flow cytometry analysis, fluorescence imaging and confocal imaging revealed that the nanoprobe had good tumor cell binding ability and that a major porpertion nanoprobes were internalized by tumor cells. The nanoprobe exhibited no obvious cytotoxicity, as shown by in vitro and in vivo toxicity studies.
Several modification strategies have been proposed to modify exosomes; these approaches include antigenantibody binding, genetic engineering, loading, and hydrophobic interaction [29]. Hydrophobic interactions are ideal for incorporating various functional groups on the surface of exosomes, offering rapid reaction, simplicity, low cost, and high yield. Additionally, hydrophobic insertion can be used to engineer virtually all exosome types without affecting the morphology and biological properties of exosomes [28]. Hydrophobic and amphiphilic materials can penetrate lipid bilayers, offering an ideal platform for membrane modification. The lipid analog DSPE-PEG is widely used as a modification material to insert functional molecules on the surface of exosomes based on hydrophobic interactions with exosome membrane lipids [28]. In addition, the use of PEG can endow exosomes with the so-called "stealth" properties to reduce the protein adsorption [30,31]. Taking account of the limitations of the device channel, TEx were modified with DSPE-PEG 2000 -FITC (TEx-FITC) and subsequently used for in vitro tumor cell uptake analysis. We found that the fluorescence signals of tumor cells were elevated with increasing concentrations of TEx-FITC. TEx were also modified with DSPE-PEG 2000 -Cy5 (TEx-Cy5) and co-incubated with tumor cells; confocal imaging revealed strong fluorescence signals in tumor cell membranes and the cytoplasm. These findings suggest the successful modification of TEx. HYNIC and Cy7 were also inserted into the surface of TEx using the same method.
The use of 99m Tc-HMPAO and 99m Tc-tricarbonyl to label exosomes has been previously reported [19,32]. However, labeling with 99m Tc-HMPAO and 99m Tc-tricarbonyl requires is elaborate, requiring expensive and complex radioactive precursors. In this study, we used a simple labeling approach that provided radiochemical purities of > 85% for both 99m Tc-TEx-Cy7 and 99m Tc-AEx-Cy7. The radiolabeled exosomes were stable, with more than 80% remaining intact after incubation in FBS for 6 h at 37 °C.
SPECT imaging provides excellent penetration and sensitivity, whereas NIRF imaging offers high temporal resolution, spatial resolution, and real-time tumor delineation. In this study, the nanoprobe's in vivo biodistribution was assessed using SPECT imaging, and the tumor boundaries were identified using NIRF imaging. Multimodality imaging with SPECT and NIRF can combine the advantages of SPECT and NIRF. The most appropriate SPECT and NIRF imaging time was determined to be 18 h after the injection of the nanoprobe, as a maximum tumor uptake (1.46 ± 0.06% ID/g) and tumor-to-muscle ratio (8.22 ± 0.65) was observed 18 h after injection of 99m Tc-TEx-Cy7. Exosomes derived from tumor cells inherently possess a high tumor-targeting ability. Our in vivo and in vitro analyses demonstrated that 99m Tc-TEx-Cy7 had a high affinity for tumor cells. TEx and AEx exhibited similar average hydrodynamic diameters and zeta potentials. Nevertheless, compared with 99m Tc-AEx-Cy7, 99m Tc-TEx-Cy7 showed higher tumor cell uptake at all the time points tested and more profound tumor accumulation in tumor-bearing mice. These results suggest that TEx offer better tumor-targeting ability than AEx.
There are several limitations to this study. TEx accumulation was observed in the liver, spleen, and kidneys, impacting the quality of imaging. A pre-targeting approach for nuclear imaging could be employed to reduce the uptake in the liver and spleen. Furthermore, the production and isolation of exosomes still remain challenges. New methods for isolation of Ex are research hotspot, such as microfluidic methods.

Conclusions
In this study, a novel exosomes-based nanoprobe was successfully engineered for multimodal SPECT/ NIRF imaging of colon cancer. The use of hydrophobic interactions provides possibility for engineering exosomes-based multimodal imaging agents. Our data verified that DSPE-PEG 2000 functionalized groups can be inserted on the surface of exosomes using this approach. This research also proved that exosomes from tumor cells are potential high-quality nanocarriers for multimodal imaging and have broad application prospects.

Cell culture
The study protocols were approved by the Ethics Committee at the Tongji Medical College of Huazhong University of Science and Technology. Human adipose stem cells were isolated from subcutaneous fat. The human colon cancer cell line HCT116 was maintained in our laboratory (Hubei Province Key Laboratory of Molecular Imaging) in RPMI-1640 medium (Gibco, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, USA). The cells cultured at 37 °C in a humidified atmosphere containing 5% CO2.

Isolation of TEx and AEx
Exosomes from serum-free culture supernatants of HCT116 cells and ASCs were obtained by differential ultracentrifugation. Dead cells and cell fragments were removed by centrifugation at 3000×g for 30 min. The supernatants were centrifuged at 13,000×g for 70 min to remove microvesicles. Subsequently, the supernatants were concentrated using an Amicon ® Ultra-15 Centrifugal Filter Devices (100 kDa molecular weight, Millipore, USA). The supernatants were centrifuged at 120,000×g for 70 min and passed through a 0.22 μm filter to obtain TEx and AEx.

TEx and AEx characterization
For transmission electron microscopy (TEM), TEx and AEx were resuspended at 1.0 mg/mL in phosphate buffer saline (PBS; Gibco, USA), placed on 200-mesh carboncoated copper grids for 2 min, and subjected to negative staining using phosphotungstic acid. The hydrodynamic diameters and zeta potential values were identified by dynamic light scattering (DLS; Malvern Instruments Ltd., Worcestershire, UK). The changes in hydrodynamic diameters were monitored for 4 days by DLS to test the stability of TEx in vitro.

In vitro tumor cell binding
Fluorescence intensity of TEx was analyzed by flow cytometry to determine the level of internalization. The modification of exosomes was adapted from a literature reported method [28]. Briefly, we introduced 1 mg DSPE-PEG per 1 mg exosomes. TEx were incubated with 1 mg DSPE-PEG 2000 -FITC (Ruixi, Xi'an, China) at room temperature for 30 min, and FITC-TEx were passed through a centrifugal filter device (100 kDa molecular weight, Amicon ® Ultra-15) as previously described [28]. For internalization assay, HCT116 cells were seeded in 10 cm cell culture dishes and treated with different concentrations of FITC-TEx (0, 5, 10, 20 μg/mL). After incubation at 37℃ for 12 h, cells were digested and dissolved in 200 µL PBS for flow cytometry analysis (FACSort, BD, USA). The tumor binding ability of TEx was assessed by confocal microscopy. TEx and DSPE-PEG 2000 -Cy5 (Ruixi, Xi'an, China) were incubated at room temperature for 30 min, and Cy5-labeled TEx (TEx-Cy5) were obtained. TEx-Cy5 (100 μg /mL) were added onto HCT116 cells grown in a confocal dish and incubated at 37℃ for different time periods (6 h, 12 h, 24 h, 48 h). The cell nuclei were counterstained with 4′,6-Diamidino-2-phenylindole (DAPI) (Boster, Wuhan, China). Cells were fixed with paraformaldehyde and observed under a Fluorescence microscope. Then TEx-Cy5 (100 μg /mL) were added onto HCT116 cells grown in a confocal dish and incubated for 12 h. The cytoskeleton of tumor cells was stained with FITC-phalloidin, and cell nuclei were counterstained with DAPI (Boster, Wuhan, China). Cells were fixed with paraformaldehyde and observed under a confocal microscope (LSM 880, ZEISS).

Cellular uptake of 99m Tc-TEx-Cy7
To assess cell uptake, we incubated 1 × 10 6 of HCT116 cells with RPMI-1640 medium supplemented with 10% fetal bovine serum containing 99m Tc-labeled TEx (37 kBq/well) at 37 °C for 0.5, 1, 2, 3, 6, 12, 24, 36 and 48 h. HCT116 cells incubated with 99m Tc-labeled AEx were used as a control. After incubation, the supernatants were removed and washed with cold PBS. The remaining cells were lysed in 0.1 M NaOH and rinsed with cold PBS. Cell lysates and supernatants were collected. Radioactivity was measured using a γ-counter (PerkinElmer, USA), and the cellular uptake of 99m Tc-TEx-Cy7 was calculated as the radioactivity in the cells divided by the total added radioactivity and multiplied with 100 to get the percentage. Experimental conditions were performed in triplicate.

Tumor-bearing nude mouse models
All mouse experimental procedures were reviewed and approved by the Animal Care Committee of Tongji Medical College, Huazhong University of Science and Technology. HCT116 cells (5 × 10 6 ) suspended in 100 μL PBS were subcutaneously injected into the upper right leg of BALB/C nude mice (male, 4 weeks old; Beijing HFK Bioscience co., Ltd, China). After the tumor volume reached approximately 50 mm 3 , tumor-bearing mice were used for imaging.

NIRF imaging 99m
Tc-TEx-Cy7 were injected into tumor-bearing mice (n = 3 per group) via the tail vein for NIRF imaging. Mice were anesthetized with 2% isoflurane, and NIRF imaging was performed at different time points (1, 6, 12, 18, and 24 h after injection). Static NIRF images were acquired with 750 nm excitation and 790 nm emission filters using an IVIS Spectrum imaging system (In-Vivo FX PRO, Bruker, Germany). NIRF images were analyzed using Bruker MI (Bruker, Germany).

Biodistribution analysis
To determine the metabolic characteristics of TEx, we assessed the biodistribution of 99m Tc-TEx-Cy7 in HCT116 tumors. HCT116 tumor-bearing mice were injected with 29.6 MBq 99m Tc-TEx-Cy7. Animals were sacrificed 6, 12, 18, and 24 h after injection (n = 3 mice per time point). Tissues were excised, weighed, and analyzed using a γ-counter. The radioactivity in organs and tissues was calculated as the percentage of injected dose per gram of tissue (% ID/g) and corrected for radioactive decay.

Statistical analysis
Data are shown as the mean ± standard deviation (SD). Comparisons between groups were evaluated with the unpaired Student's t-test. p < 0.05 was considered to be statistically significant. Statistical analysis was conducted using GraphPad Prism v8.0 software.