Membrane Radiolabelling of Exosomes for Comparative Biodistribution Analysis in Immunocompetent and Immunodeficient Mice - A Novel and Universal Approach

Extracellular vesicles, in particular exosomes, have recently gained interest as novel drug delivery vectors due to their biological origin and inherent intercellular biomolecule delivery capability. An in-depth knowledge of their in vivo biodistribution is therefore essential. This work aimed to develop a novel, reliable and universal method to radiolabel exosomes to study their in vivo biodistribution. Methods: Melanoma (B16F10) cells were cultured in bioreactor flasks to increase exosome yield. B16F10-derived exosomes (ExoB16) were isolated using ultracentrifugation onto a single sucrose cushion, and were characterised for size, yield, purity, exosomal markers and morphology using nanoparticle tracking analysis (NTA), protein measurements, flow cytometry and electron microscopy. ExoB16 were radiolabelled using 2 different approaches - intraluminal labelling (entrapment of 111Indium via tropolone shuttling); and membrane labelling (chelation of 111Indium via covalently attached bifunctional chelator DTPA-anhydride). Labelling efficiency and stability was assessed using gel filtration and thin layer chromatography. Melanoma-bearing immunocompetent (C57BL/6) and immunodeficient (NSG) mice were injected intravenously with radiolabelled ExoB16 (1x1011 particles/mouse) followed by metabolic cages study, whole body SPECT-CT imaging and ex vivo gamma counting at 1, 4 and 24 h post-injection. Results: Membrane-labelled ExoB16 showed superior radiolabelling efficiency and radiochemical stability (19.2 ± 4.53 % and 80.4 ± 1.6 % respectively) compared to the intraluminal-labelled exosomes (4.73 ± 0.39 % and 14.21 ± 2.76 % respectively). Using the membrane-labelling approach, the in vivo biodistribution of ExoB16 in melanoma-bearing C57Bl/6 mice was carried out, and was found to accumulate primarily in the liver and spleen (~56% and ~38% ID/gT respectively), followed by the kidneys (~3% ID/gT). ExoB16 showed minimal tumour i.e. self-tissue accumulation (~0.7% ID/gT). The membrane-labelling approach was also used to study ExoB16 biodistribution in melanoma-bearing immunocompromised (NSG) mice, to compare with that in the immunocompetent C57Bl/6 mice. Similar biodistribution profile was observed in both C57BL/6 and NSG mice, where prominent accumulation was seen in liver and spleen, apart from the significantly lower tumour accumulation observed in the NSG mice (~0.3% ID/gT). Conclusion: Membrane radiolabelling of exosomes is a reliable approach that allows for accurate live imaging and quantitative biodistribution studies to be performed on potentially all exosome types without engineering parent cells.


Fluorescent labelling of exosomes
AlexaFluor 488 NHS Ester (Alexa-NHS) was used to fluorescently label lysine (Lys) residues on the surface of B16F10 exosomes (ExoB16). Alexa-NHS was reacted with ExoB16 at a 1:400 molar ratio (Lys:Alexa-NHSit was assumed that 1 exosome is equivalent to 1 BSA molecule i.e. containing 59 lysine residues) protected from light for 1 hour at room temperature (RT). Labelled ExoB16 (Alexa-ExoB16) was purified from excess unreacted Alexa-NHS using Sepharose® CL-2B columns (self-packed according to the dimensions of the commercially available NAP-5™ columns) optimised such that exosomes elute in the first 2 x 500 µl fractions.

Exosomal surface proteins detection by dot blot following mock radiolabelling
ExoB16 were subjected to a mock membrane-radiolabelling protocol (as described in the Methods section of the main manuscript, but without the addition of 111 InCl3 to the 0.2 M ammonium acetate buffer pH 5.5). 40 µl of 5 x 10 10 p/ml unlabelled and labelled ExoB16 in PBS was spotted onto a nitrocellulose membrane (10 µl at a time) and dried under a nitrogen stream. Non-specific binding on the membrane was blocked by 3% milk in Tris-buffered saline with 0.1% Tween-20 (TBS-T, pH 7.6) at RT for 1 hour. The membrane was incubated with primary antibodies (1:1000 dilution) in 3% milk in TBS-T (i.e. the blocking buffer) at 4°C overnight. Washing was done 3 times with TBS-T for 5 minutes each. The membrane was then incubated with HRP-conjugated rabbit antimouse secondary antibody (1:1000 dilution) in the blocking buffer for 1 hour at RT.
Washing was done 3 times with TBS-T for 5 minutes each as above. The washed membrane was incubated with the ECL substrate for 3 min, and then imaged using the Gel Doc™ system (Bio-Rad, US) under the "Intense Bands" setting. The acquired imaged was processed using Image Lab™ software (Bio-Rad, US).

In vitro uptake of exosomes in cancer cell lines
B16F10 and GL261 cells were seeded in 96-well plates at a density of 30,000 cells per well and left to settle overnight. Each cell line was then treated with 2.5 x 10 10 Alexa-ExoB16 for 1, 4, and 24 hours. Penicillin-Streptomycin (10% of total volume in each well) was added to prevent bacterial growth. After treatment, cells were detached, resuspended in 3% FBS/PBS and run on FACSCalibur under the FL1 channel for detection of AlexaFluor 488 signals. The results were analysed using CellQuest Pro software (BD Biosciences, US). Untreated B16F10 and GL261 cells were used as control.

Tumour-associated macrophages (TAMs) population analysis
Female C57Bl/6 mice and male NOD SCID gamma (NSG) mice (~20 g, 6-8 weeks old) were inoculated with B16F10 cells (1 x 10 6 cells in 100 µl PBS) subcutaneously into the left and right rear flanks of the mice to establish subcutaneous (SC) B16F10 tumours. The mice were monitored closely post-inoculation and were culled when the tumours reached ~200-300 mm 3 . The tumours were excised, and each tumour was chopped into fine pieces in 1 ml serum-free media, which was then topped up to 4 ml with serum-free media. The following enzymes were added to the tumour suspension at their respective final concentrations: Collagenase Type IV (2 mg/ml) and DNase I (150 µg/ml). The tumours were digested for 45 minutes at 37°C in a shaking water bath (320 rpm), vortexing every 15 minutes. The reaction was stopped by adding 10 ml 1% BSA solution (prepared in PBS), and the cell suspension was passed through a 70 µm cell strainer. Tumour-derived cells in the filtrate were washed once with PBS and subjected to a red blood cell lysis step by adding 5 ml 1X RBC lysis buffer to the cell pellet post-washing and incubated for 5 minutes at RT. The reaction was stopped by adding 25 ml PBS. Cells were pelleted (500 g, 5 min) and resuspended in 400 µl PBS. Cells were single-and triple-stained with rat anti-mouse CD45-APC, F4/80-FITC Biochemical analysis of B16F10 exosomes. Histograms illustrate the detection of CD81 and CD9 using flow cytometry on exosomes isolated from B16F10 cells. Exosomes were coupled to aldehyde/sulphate latex beads prior to detection. Exo-beads complex were subsequently stained using a 2-step labelling (anti-CD81 or anti-CD9 1° ab/Cy5-conjugated 2° ab). Exo-beads complex stained with Cy5-conjugated 2° ab only was used as control. The shift in FL4 signals from the control indicates positive expression of the markers on the exosomes.

Fig. S2
Exosome and protein elution profile by gel filtration. B16F10 exosomes and BSA solution were loaded separately onto Sepharose® CL-2B columns self-packed according to the dimensions of the commercially available NAP-5 columns. 3.5 x 10 11 exosomes (100 µl) were loaded onto the column. (A) 80 µg/ml and (B) 800 µg/ml BSA solution was loaded onto the column (100 µl for both concentrations). Eight 500 µl fractions were collected and the particle number and protein concentration in each fraction was measured using Nanosight and microBCA assay respectively. Values are expressed as mean ± SD, where n=3.  ExoB16 were subjected to a mock membrane-radiolabelling protocol (as described in the Methods section of the manuscript, but without the addition of 111 InCl3 to the 0.2 M ammonium acetate buffer). Equal numbers (40 µl from 5 x 10 10 p/ml stock) of both unlabelled and labelled of ExoB16 were then blotted on a nitrocellulose membrane, followed by blocking in 3% milk solution. The membrane is then incubated with primary rabbit anti-mouse CD9 and CD63 antibodies, followed by incubation with HRP-linked goat anti-rabbit secondary antibodies. After addition of the ECL substrate, the membrane was imaged using Gel Doc™ system (Bio-Rad, USA) and the image processed using Image Lab™ software (Bio-Rad, USA).

Fig. S9 Uptake of fluorescently-labelled exosomes in B16F10 cells in vitro.
B16F10-(ExoB16) and GL261-derived exosomes (ExoGL261murine glioma) were fluorescently labelled with AlexaFluor®488-NHS dye. B16F10 cells were then incubated with 2.5 x 10 10 fluorescently-labelled ExoB16 and ExoGL261 for 1, 4 and 24 h, after which the cells were collected and analysed using flow cytometry under the FL1 channel. Degree of exosome uptake are expressed as the fold difference in median fluorescence intensity (MFI) from that of untreated cells. MFI fold difference value of at least 1.5 (dashed line) is regarded as uptake in the B16F10 cells. Values are expressed as mean ± SD, where n=3, and Student's t-test was used for statistical analysis (p*< 0.05).

Fig. S10 CD47 expression on exosomes derived from cancer and non-cancer cell lines.
Histograms illustrate the detection of CD47 on exosomes isolated from B16F10 (ExoB16murine melanoma), IENS (ExoIENSmurine glioblastoma stem cells), HEK293 (ExoHEK -human epithelial) and Panc1 (ExoPanc1human pancreatic adenocarcinoma) cell lines using flow cytometry. Exosomes were coupled to aldehyde/sulphate latex beads prior to detection. Exo-beads complex were then stained with APC-conjugated anti-CD47 or their respective isotype controls (rat anti-mouse for murine exosomes, or mouse anti-human for human exosomes). The shift in FL4 signals indicates positive expression of CD47 on the exosomes.

Fig. S11 Comparison of tumour-associated macrophages (TAMs) population between tumours developed in immunocompetent and immunodeficient mice. (A)
Representative gating strategy used to identify TAM population among cells isolated from tumours developed in both C57Bl/6 and NSG mice. Cells were triple-stained with APCconjugated anti-CD45, FITC-conjugated anti-F4/80 and PE conjugated anti-CD11b antibodies. TAMs are identified as CD45+ F4/80+ CD11b+. First, viable cells which are CD45+ are selected, followed by cells which are both F4/80+ CD11b+. (B) Population of TAM isolated from tumours developed in C57Bl/6 and NSG mice, expressed as the percentage of CD45+ F4/80+ CD11b+ cells in total cells isolated from the tumours. Values are expressed as mean ± SD, where n=3, and Student's t-test was used for statistical analysis (p* < 0.05).