The MSCs were supplied by Lonza (Lonza, Basel, Switzerland) and cultured in the presence of MSC basal medium (MSCBM; Lonza). MSC-derived EVs were collected from the supernatant of MSCs cultured overnight in RPMI-1640 (Lonza) supplemented with 0.5% of BSA (Sigma-Aldrich, St. Louis, MO, USA). The cell supernatant was centrifuged twice at 3,000 × g for 20 min to remove cell debris and then ultracentrifuged at 100,000 × g (Beckman Coulter Optima L-90K ultracentrifuge; Beckman Coulter, Brea, CA, USA) for 1 h at 4ºC. EVs were stored in serum-free RPMI-1640 supplement with 1% DMSO at −80ºC. EV protein content was quantified by the Bradford method (Bio-Rad, Hercules, CA, USA).

Two labeling protocols were used: i) Cells were stained in suspension and incubated with 5 μM Vybrant Cell Tracers DiD [excitation (Ex), 640 nm; emission (Em), 700 nm] or DiI (Ex, 530 nm; Em, 580 nm) (Molecular Probes, Eugene, OR, USA) solution without serum for 20 min at 37ºC. Cells were then washed in complete medium by centrifugation and cultivated for 24 h prior to supernatant collection. i) EVs were isolated by ultracentrifugation as previously described (LCD-EVs) (4,13). ii) EVs were directly labeled with 1 μM Vybrant Cell Tracers DiI or DiD during the ultracentrifugation procedure (DL-EVs) and then washed twice by ultracentrifugation in 1X phosphate-buffered saline (PBS) (28).

EVs labeled with the two methods were characterized by cytofluorimetric analysis using FITC- or PE-conjugated antibody against CD44, CD105, CD90 and α5-integrin. FITC or PE mouse non-immune isotypic IgG (Miltenyi Biotec, Bergisch Gladbach, Germany) were used as the controls. Briefly, EVs (10 μg) were incubated for 15 min at 4ºC with antibodies in 100 μl and then diluted in 300 μl and immediately acquired. FACS analysis was performed using a guava easyCyte Flow Cytometer (Millipore, Billerica, MA, USA) and analyzed with InCyte software, as previously described (29,30).

The size distribution of the LCD-EVs and DL-EVs was analyzed using a NanoSight LM10 instrument (NanoSight Ltd., Amesbury, UK) equipped with the nanoparticle tracking analyses (NTA) 2.0 analytic software.

Human renal proximal tubular epitheial cells (PTECs) were labeled following the manufacturer’s instructions with the CFSE green dye (Vybrant CFDA SE Cell Tracer kit; Molecular Probes). Cells were incubated for 5 h at 37ºC with 50 μg/ml DL-EVs or LCD-EVs and after washing the cells were fixed in 3.5% paraformaldehyde containing 2% sucrose. Confocal microscopy analysis was performed using a Zeiss LSM 5 Pascal model confocal microscope (Carl Zeiss, Oberkochen, Germany). Hoechst 33258 dye (Sigma-Aldrich) was added for nuclear staining.

Studies were conducted in accordance with the national guidelines and regulations and were approved by the Ethics Committee of the University of Torino. Male CD1 nude mice (6–8 weeks old) (Charles River Laboratories, Lyon, France), were fed for 1 week with a special diet (AIN 79; Mucedola, Settimo Milanese, Italy) to reduce tissue autofluorescence. AKI was induced, as previously described (16), by an intramuscular injection of glycerol (7 ml/kg body weight of 50% glycerol solution) into the inferior hind limbs. At 3 days post-injury, the mice were injected intravenously (i.v.) with 200 μg of DiD-labeled EVs. Sixteen nude mice with AKI were treated with LCD-EVs and were sacrificed after 5 h (n=9) and 24 h (n=7). Eleven nude mice with AKI were treated with DL-EVs and were sacrificed after 5 h (n=6) and 24 h (n=5). The same amount of LCD- and DL-EVs was i.v. injected in 12 and 6 healthy mice, respectively. The animals were sacrificed after 5 h (LCD-EV, n=6; DL-EV, n=3) and 24 h (LCD-EV, n=6; DL-EV, n=3).

In vitro experiments were performed using the IVIS 200 small animal imaging system (PerkinElmer, Waltham, MA, USA) using the Ex filter at 640 nm and the Em filter at 700 nm. Background fluorescence was measured and subtracted by setting up a background measurement (Ex filter, 530 nm). EV samples were placed in a non-fluorescent black container and the fluorescence intensity of increasing concentrations of LCD-EVs and DL-EVs (15, 30, 50 and 100 μg) in the same volume was evaluated. Image analysis involved the designation of regions-of-interest (ROI) as the circular area of the well containing the EV concentrations to obtain the average intensity ± standard deviation (SD), as previously described (31).

In vivo fluorescence imaging was performed with the same wavelength as described for in vitro acquisition. Identical illumination settings, such as exposure time (2 sec), binning factor (factor of 4), f/stop (set to 2) and 12 fields of view, were used for acquiring all images, and fluorescence Em was normalized to photons per second per centimeter squared per steradian (p/sec/cm²/sr). The color image represents the spatial distribution of fluorescence within the animal overlaid on black and white photographs of the mice, collected at the same time. Images were acquired and analyzed using Living Image 4.0 software (PerkinElmer), as previously described (32).

The mice were anesthetized with 2.5% isoflurane (Merial, Lyon, France) and images were acquired in the prone and supine position after 15 min, 5 and 24 h post-EV injection. The mice with AKI and the healthy mice treated with PBS were used as blank controls for the fluorescence signal of EVs in the AKI and healthy groups, respectively. The fluorescence signal was quantified in the kidney region and in the abdominal area, in ROI draw freehand. The relative mean fluorescence intensity of each ROI was obtained by subtracting the mean fluorescence intensity of the corresponding ROI on the blank mouse from the measured mean fluorescent intensity, as previously described (22,33). Data were expressed as the average radiance ± SD.

At the end of the experiments (5 or 24 h post-EV injection), the mice were sacrificed and dissected tissues (kidneys, spleen, liver and lungs) were imaged immediately. The mean fluorescence of each tissue sample was obtained by subtracting the fluorescence intensity of corresponding tissue from the blank mouse, as previously described (33).

Mice were sacrificed at 5 and 24 h and confocal microscopy analysis (Leica TSC SP5 II) was performed on frozen sections for localization of DiD-labeled EVs in the kidneys. Hoechst 33258 dye (Sigma-Aldrich) was added for nuclear staining. Images were analyzed using ImageJ software.

The results are generally expressed as the means ± SD. Statistical analysis was performed by ANOVA with Dunnet’s multi-comparison test or the Newman-Keuls multi-comparison or by the Student’s t-test where appropriate. A p-value of <0.05 was considered to indicate a statistically significant difference.

The OI of EVs obtained with the two following methods was compared: i) DL-EVs were labeled with DiD after their production and purification; ii) LCD-EVs were obtained by the supernatant of MSCs previously labeled with DiD and then cultured for 24 h prior to EV collection. OI images were acquired after EV dilution, ranging from 15 to 100 μg of EV proteins, in 100 μl of PBS and the average intensity within the entire circle area of each well was calculated. The fluorescence signal correlated linearly with the EV concentration for both labeling methods. DL-EVs were brighter compared with LCD-EVs (Fig. 1).

EVs labeled with two methods showed the same phenotype as unlabeled EVs. Cytofluorimetric analyses showed their fluorescent signal in the NIR region (Fig. 2A and B) and the presence of several antigens typically expressed by MSCs and by their EVs (13), such as CD44, CD105, CD90 and α5-integrin (Fig. 2C). Regarding size distribution analyzed with NanoSight, LCD-EVs showed a size range of 180±73 nm not significantly different from the size distribution of unlabeled EVs (145±57 nm). DL-EVs showed a size of 250±89 nm with a second peak of larger size probably due to some aggregations occurring during the labeling procedure (Fig. 3A).

To evaluate the ability of labeled EVs to be incorporated by PTECs, 50 μg/ml of EVs labeled with the red dye, DiI, following the same procedure described above, were added to the cells. DL-EVs and LCD-EVs were equally incorporated within PTECs, as observed by confocal microscopy after 5 h of incubation (Fig. 3B).

The ability of labeled EVs to be visualized by OI on the whole body of live mice was assessed using an IVIS 200 system in a model of AKI induced by an intramuscular glycerol injection, as previously described (13). Three days after the glycerol injection, blood urea nitrogen (131±16 mg/dl) and creatinine (0.9±0.2 mg/dl) levels were significantly increased compared with the healthy controls (28±10 and 0.2±0.1 mg/dl, respectively) and were associated with diffuse tubular epithelial injury, characterized by tubular hyaline casts, vacuolization and widespread necrosis of the proximal and distal tubular epithelium, loss of brush border and denudation of the basal membrane (data not shown). The 200 μg of DL-EVs or LCD-EVs was inoculated i.v. 3 days following the induction of AKI, when functional and morphological damage had reached its peak (13). Healthy mice were treated with the same amount of EVs in order to evaluate whether the accumulation of EVs was specific for the site of injury. Fig. 4 shows a fluorescent signal in the region of kidneys of AKI mice treated with LCD-EVs and DL-EVs and analyzed posterior after 15 min and 5 h. However, the intensity of fluorescence in the DL-EV-treated mice with AKI was still present at 24 h after the i.v. injection and was significantly higher than the fluorescence signal generated by the kidneys of mice with AKI injected with LCD-EVs. In the healthy mice, we observed a fluorescent signal only in the mice treated with DL-EVs after 5 h in the left dorsal region that may correspond to spleen accumulation; however, the signal decreased rapidly and 24 h after the i.v. injection was not detectable (Fig. 4). Analyzing the signal in the abdominal area, a higher fluorescence intensity was observed in the AKI groups compared with the healthy groups. Nevertheless, the DL-EV-treated mice with AKI displayed a major increase in fluorescence intensity in comparison with the LCD-EV-treated AKI group, possibly due to the accumulation of EVs in the liver and spleen (Fig. 5).

For each experimental group, the mice were sacrificed at 5 and 24 h after the EV injection and the fluorescent signal from freshly dissected tissues was quantified immediately by OI. The fluorescence intensity of the kidneys of mice with AKI treated with LCD-EVs and DL-EVs was significantly higher after 5 h compared with the kidneys of mice with AKI treated with PBS [AKI CTL (control)], as shown in Fig 6. Nevertheless, the increase in fluorescence was significantly maintained for 24 h following treatment in the DL-EV group, whereas it decreased in the LCD-EV group (Fig. 6A). No signal was detected in the kidneys of healthy mice treated with LCD-EVs and DL-EVs, suggesting a specific accumulation of EVs at the site of injury.

The fluorescence signal of DL-EVs was also detected in the spleen and particularly in the liver with high variability. The fluorescence signal of DL-EVs was of low intensity in the lungs of both the AKI and healthy groups. LCD-EVs were detectable only in the injured kidneys.

The presence of LCD-EVs and of DL-EVs within injured kidneys was confirmed by confocal analysis using the appropriate wavelength (Fig. 7). In the kidney sections derived from healthy mice, the presence of fluorescent EVs was almost absent.