Extracellular Vesicle Capture by AnTibody of CHoice and Enzymatic Release (EV-CATCHER): A customizable purification assay designed for small-RNA biomarker identification and evaluation of circulating small-EVs

Abstract

Circulating nucleic acids, encapsulated within small extracellular vesicles (EVs), provide a remote cellular snapshot of biomarkers derived from diseased tissues, however selective isolation is critical. Current laboratory-based purification techniques rely on the physical properties of small-EVs rather than their inherited cellular fingerprints. We established a highly-selective purification assay, termed EV-CATCHER, initially designed for high-throughput analysis of low-abundance small-RNA cargos by next-generation sequencing. We demonstrated its selectivity by specifically isolating and sequencing small-RNAs from mouse small-EVs spiked into human plasma. Western blotting, nanoparticle tracking, and transmission electron microscopy were used to validate and quantify the capture and release of intact small-EVs. As proof-of-principle for sensitive detection of circulating miRNAs, we compared small-RNA sequencing data from a subset of small-EVs serum-purified with EV-CATCHER to data from whole serum, using samples from a small cohort of recently hospitalized Covid-19 patients. We identified and validated, only in small-EVs, hsa-miR-146a and hsa-miR-126-3p to be significantly downregulated with disease severity. Separately, using convalescent sera from recovered Covid-19 patients with high anti-spike IgG titers, we confirmed the neutralizing properties, against SARS-CoV-2 in vitro, of a subset of small-EVs serum-purified by EV-CATCHER, as initially observed with ultracentrifuged small-EVs. Altogether our data highlight the sensitivity and versatility of EV-CATCHER.

Keywords: TEM; exosome purification; extracellular vesicles; micro‐RNA profiling; sequencing.

FIGURE 1

Strategy developed for selectively purifying small‐EVs from human biofluids. a. Schematic representation of the EV‐CATCHER assay designed for purification of small‐EVs from biofluids, which relies on binding of a degradable dsDNA‐linker (uracilated 5′‐azide oligonucleotide annealed to complementary uracilated 3′‐biotin oligonucleotide) to a DBCO‐activated antibody and to a streptavidin‐coated platform. b. Acrylamide gel migration of single stranded oligonucleotides and double stranded (ds) hybridized DNA‐linker. Single stranded (ss) biotin labelled DNA (lane 2); ssDNA azide (lane 3), dsDNA linker (lane 4), and UNG digested dsDNA linker (lane 5) were separated on a 15% non‐denaturing PAGE gel. c. Monoclonal human anti‐CD63 antibody (Ab) activation, for selection of small‐EVs expressing CD63, and dsDNA‐linker conjugation. Representative image of Coomassie stained non‐denaturing PAGE gel (12%) with migration of protein ladder (lane a), 1 μg native anti‐CD63 Ab (lane b); dsDNA‐linker conjugated to 1 μg DBCO‐activated anti‐CD63 antibody using increased amounts of dsDNA‐linker (lanes c‐e), dsDNA‐linker‐Ab (500 ng ‐ 1 μg) digested by UNG (lane f). Experiments were replicated and imaged on a ChemiDoc MP imager

FIGURE 2

Identification of a low‐background binding platform and comparison to 11 existing small‐EV purification methods. a. Diagram comparing EV‐CATCHER to 10 commercially available kits and ultracentrifugation for purification of small‐EVs. Three magnetic‐bead purification methods were identified where chemical treatment have been described (1% Formic Acid pH 2.0‐ 4.0, 0.1 M Glycine pH 2.5‐3.0 (Chen et al., 2020; Hisey et al., 2018; Hong et al., 2014), or proprietary buffer (SBI ExoFlow)) for small‐EV release (Barros et al., 2018; Burkova et al., 2019; Gutzeit et al., 2014; Hardin et al., 2018; Huang et al., 2018; Mallegol et al., 2007; Patel et al., 2019; Peterson et al., 2015; Theodoraki et al., 2018; Velandia‐Romero et al., 2020; ). b. Evaluation of non‐specific small‐RNA binding to four commercially different customizable magnetic beads (130 nm to 4.5 μm diameter), including both streptavidin‐coated (3 methods) and Carboxyl‐coated beads (1 method), by comparison to wells used for the EV‐CATCHER assay. A 1xPBS solution containing 100 pg ath‐miR‐159a RNA oligonucleotide was incubated with wells (1‐ EV‐CATCHER in triplicate) or the four different types of customizable magnetic‐beads (2‐ MojoSort, 3‐ Carboxyl‐Dynabeads, 4‐ ExoCap, 5‐ Streptavidin‐coated Dynabeads) in triplicate. Wells and magnetic‐beads were washed 3 times and the eluted RNA was quantified by RT‐qPCR (Left graph). Two additional conditions were tested to prevent non‐specific binding of ath‐miR‐159a including a pre‐treatment with BSA (1%) (middle graph), and a combined pre‐treatment with BSA (1%) and RNase‐A (12.5 μg/ml) (Right graph). c. Evaluation of non‐specific binding of ath‐miR‐159a as a contaminant in 100 μl of human serum for all 10 commercial kits (#2 to #11) and the ultracentrifugation method (#12). All magnetic‐bead kits were tested in their ‘Exosome’ purification commercial forms with antibodies on their surface (anti‐Tim4 for Fujifilm and CD63 for all other techniques) and compared to anti‐CD63 EV‐CATCHER (#1). Following manufacturer's instructions, small‐EVs were eluted and RNA was extracted prior to RT‐qPCR quantification of ath‐miR‐159a. All experiments were repeated three times. The averaged comparative thresholds (Ct) values are added above the standard deviation bars in the graph, for each commercial kit and method. d. Evaluation of non‐specific binding of MCF‐7 small‐EVs to EV‐CATCHER and 4 customizable small‐EV commercial magnetic‐bead purification assays. The experimental collection of non‐specifically bound small‐EVs or small‐EVs remaining in solution after non‐specific capture with EV‐CATCHER and the 4 magnetic bead‐based commercial kits is displayed in the schematic (above). RNase A (12.5 μg/ml) was added prior to all experiments, to remove free‐floating miRNAs (Compare grey bars to black bars in all graphs). The top two graphs represent RT‐qPCR quantifications of hsa‐miR‐21 from wells or magnetic beads exposed to the MCF7 solution containing 10 μg of small‐EVs/ ‘exosomes’(left) or remaining in the solution (right), after RNA extraction (All experiments were performed in triplicate). The lower two graphs represent the same conditions with quantification of hsa‐miR‐200c (Non‐specific binding to wells or magnetic beads (left) or remaining in solution (right). All RT‐qPCR experiments were performed in triplicate for each method. Data is represented as fold change with technical replicates (subtraction of ath‐miR‐159a and 10 μg MCF‐7 control). e. Western blot analyses (8 μg total protein) of small‐EVs purified by CD63 EV‐CATCHER and all commercial kits (Tim4 antibody for Fujifilm, and CD63 antibody for all other magnetic bead‐based methods) and methods (ExoEasy, Nörgen, ExoQuick, ultracentrifugation). Anti‐human ‐ApoB, ‐CD63, ‐Albumin, ‐ApoA, ‐CD9, ‐CD81 antibodies were used to evaluate the small‐EVs purified by each of the 11 selected methods/commercial kits. f. Nanoparticles tracking of released small‐EVs using a Spectradyne NT instrument. Small‐EVs evaluated were obtained by EV‐CATCHER (anti‐CD63), MagCapture (anti‐CD63), ExoFlow (anti‐CD63), and global ExoEasy, Norgen purification kit, ExoQuick, and ultracentrifugation purifications methods. g. Transmission Electron Microscopy (TEM) with direct magnification of 20,000x and scale bars of 100 nm for small‐EVs purified from human serum using EV‐CATCHER (anti‐CD63, duplicated isolations), Exoflow (anti‐CD63), ExoCap (anti‐CD63), Dynabeads T1 MyOne™ beads (anti‐CD63) and ultracentrifugation

FIGURE 2

Identification of a low‐background binding platform and comparison to 11 existing small‐EV purification methods. a. Diagram comparing EV‐CATCHER to 10 commercially available kits and ultracentrifugation for purification of small‐EVs. Three magnetic‐bead purification methods were identified where chemical treatment have been described (1% Formic Acid pH 2.0‐ 4.0, 0.1 M Glycine pH 2.5‐3.0 (Chen et al., 2020; Hisey et al., 2018; Hong et al., 2014), or proprietary buffer (SBI ExoFlow)) for small‐EV release (Barros et al., 2018; Burkova et al., 2019; Gutzeit et al., 2014; Hardin et al., 2018; Huang et al., 2018; Mallegol et al., 2007; Patel et al., 2019; Peterson et al., 2015; Theodoraki et al., 2018; Velandia‐Romero et al., 2020; ). b. Evaluation of non‐specific small‐RNA binding to four commercially different customizable magnetic beads (130 nm to 4.5 μm diameter), including both streptavidin‐coated (3 methods) and Carboxyl‐coated beads (1 method), by comparison to wells used for the EV‐CATCHER assay. A 1xPBS solution containing 100 pg ath‐miR‐159a RNA oligonucleotide was incubated with wells (1‐ EV‐CATCHER in triplicate) or the four different types of customizable magnetic‐beads (2‐ MojoSort, 3‐ Carboxyl‐Dynabeads, 4‐ ExoCap, 5‐ Streptavidin‐coated Dynabeads) in triplicate. Wells and magnetic‐beads were washed 3 times and the eluted RNA was quantified by RT‐qPCR (Left graph). Two additional conditions were tested to prevent non‐specific binding of ath‐miR‐159a including a pre‐treatment with BSA (1%) (middle graph), and a combined pre‐treatment with BSA (1%) and RNase‐A (12.5 μg/ml) (Right graph). c. Evaluation of non‐specific binding of ath‐miR‐159a as a contaminant in 100 μl of human serum for all 10 commercial kits (#2 to #11) and the ultracentrifugation method (#12). All magnetic‐bead kits were tested in their ‘Exosome’ purification commercial forms with antibodies on their surface (anti‐Tim4 for Fujifilm and CD63 for all other techniques) and compared to anti‐CD63 EV‐CATCHER (#1). Following manufacturer's instructions, small‐EVs were eluted and RNA was extracted prior to RT‐qPCR quantification of ath‐miR‐159a. All experiments were repeated three times. The averaged comparative thresholds (Ct) values are added above the standard deviation bars in the graph, for each commercial kit and method. d. Evaluation of non‐specific binding of MCF‐7 small‐EVs to EV‐CATCHER and 4 customizable small‐EV commercial magnetic‐bead purification assays. The experimental collection of non‐specifically bound small‐EVs or small‐EVs remaining in solution after non‐specific capture with EV‐CATCHER and the 4 magnetic bead‐based commercial kits is displayed in the schematic (above). RNase A (12.5 μg/ml) was added prior to all experiments, to remove free‐floating miRNAs (Compare grey bars to black bars in all graphs). The top two graphs represent RT‐qPCR quantifications of hsa‐miR‐21 from wells or magnetic beads exposed to the MCF7 solution containing 10 μg of small‐EVs/ ‘exosomes’(left) or remaining in the solution (right), after RNA extraction (All experiments were performed in triplicate). The lower two graphs represent the same conditions with quantification of hsa‐miR‐200c (Non‐specific binding to wells or magnetic beads (left) or remaining in solution (right). All RT‐qPCR experiments were performed in triplicate for each method. Data is represented as fold change with technical replicates (subtraction of ath‐miR‐159a and 10 μg MCF‐7 control). e. Western blot analyses (8 μg total protein) of small‐EVs purified by CD63 EV‐CATCHER and all commercial kits (Tim4 antibody for Fujifilm, and CD63 antibody for all other magnetic bead‐based methods) and methods (ExoEasy, Nörgen, ExoQuick, ultracentrifugation). Anti‐human ‐ApoB, ‐CD63, ‐Albumin, ‐ApoA, ‐CD9, ‐CD81 antibodies were used to evaluate the small‐EVs purified by each of the 11 selected methods/commercial kits. f. Nanoparticles tracking of released small‐EVs using a Spectradyne NT instrument. Small‐EVs evaluated were obtained by EV‐CATCHER (anti‐CD63), MagCapture (anti‐CD63), ExoFlow (anti‐CD63), and global ExoEasy, Norgen purification kit, ExoQuick, and ultracentrifugation purifications methods. g. Transmission Electron Microscopy (TEM) with direct magnification of 20,000x and scale bars of 100 nm for small‐EVs purified from human serum using EV‐CATCHER (anti‐CD63, duplicated isolations), Exoflow (anti‐CD63), ExoCap (anti‐CD63), Dynabeads T1 MyOne™ beads (anti‐CD63) and ultracentrifugation

FIGURE 2

Identification of a low‐background binding platform and comparison to 11 existing small‐EV purification methods. a. Diagram comparing EV‐CATCHER to 10 commercially available kits and ultracentrifugation for purification of small‐EVs. Three magnetic‐bead purification methods were identified where chemical treatment have been described (1% Formic Acid pH 2.0‐ 4.0, 0.1 M Glycine pH 2.5‐3.0 (Chen et al., 2020; Hisey et al., 2018; Hong et al., 2014), or proprietary buffer (SBI ExoFlow)) for small‐EV release (Barros et al., 2018; Burkova et al., 2019; Gutzeit et al., 2014; Hardin et al., 2018; Huang et al., 2018; Mallegol et al., 2007; Patel et al., 2019; Peterson et al., 2015; Theodoraki et al., 2018; Velandia‐Romero et al., 2020; ). b. Evaluation of non‐specific small‐RNA binding to four commercially different customizable magnetic beads (130 nm to 4.5 μm diameter), including both streptavidin‐coated (3 methods) and Carboxyl‐coated beads (1 method), by comparison to wells used for the EV‐CATCHER assay. A 1xPBS solution containing 100 pg ath‐miR‐159a RNA oligonucleotide was incubated with wells (1‐ EV‐CATCHER in triplicate) or the four different types of customizable magnetic‐beads (2‐ MojoSort, 3‐ Carboxyl‐Dynabeads, 4‐ ExoCap, 5‐ Streptavidin‐coated Dynabeads) in triplicate. Wells and magnetic‐beads were washed 3 times and the eluted RNA was quantified by RT‐qPCR (Left graph). Two additional conditions were tested to prevent non‐specific binding of ath‐miR‐159a including a pre‐treatment with BSA (1%) (middle graph), and a combined pre‐treatment with BSA (1%) and RNase‐A (12.5 μg/ml) (Right graph). c. Evaluation of non‐specific binding of ath‐miR‐159a as a contaminant in 100 μl of human serum for all 10 commercial kits (#2 to #11) and the ultracentrifugation method (#12). All magnetic‐bead kits were tested in their ‘Exosome’ purification commercial forms with antibodies on their surface (anti‐Tim4 for Fujifilm and CD63 for all other techniques) and compared to anti‐CD63 EV‐CATCHER (#1). Following manufacturer's instructions, small‐EVs were eluted and RNA was extracted prior to RT‐qPCR quantification of ath‐miR‐159a. All experiments were repeated three times. The averaged comparative thresholds (Ct) values are added above the standard deviation bars in the graph, for each commercial kit and method. d. Evaluation of non‐specific binding of MCF‐7 small‐EVs to EV‐CATCHER and 4 customizable small‐EV commercial magnetic‐bead purification assays. The experimental collection of non‐specifically bound small‐EVs or small‐EVs remaining in solution after non‐specific capture with EV‐CATCHER and the 4 magnetic bead‐based commercial kits is displayed in the schematic (above). RNase A (12.5 μg/ml) was added prior to all experiments, to remove free‐floating miRNAs (Compare grey bars to black bars in all graphs). The top two graphs represent RT‐qPCR quantifications of hsa‐miR‐21 from wells or magnetic beads exposed to the MCF7 solution containing 10 μg of small‐EVs/ ‘exosomes’(left) or remaining in the solution (right), after RNA extraction (All experiments were performed in triplicate). The lower two graphs represent the same conditions with quantification of hsa‐miR‐200c (Non‐specific binding to wells or magnetic beads (left) or remaining in solution (right). All RT‐qPCR experiments were performed in triplicate for each method. Data is represented as fold change with technical replicates (subtraction of ath‐miR‐159a and 10 μg MCF‐7 control). e. Western blot analyses (8 μg total protein) of small‐EVs purified by CD63 EV‐CATCHER and all commercial kits (Tim4 antibody for Fujifilm, and CD63 antibody for all other magnetic bead‐based methods) and methods (ExoEasy, Nörgen, ExoQuick, ultracentrifugation). Anti‐human ‐ApoB, ‐CD63, ‐Albumin, ‐ApoA, ‐CD9, ‐CD81 antibodies were used to evaluate the small‐EVs purified by each of the 11 selected methods/commercial kits. f. Nanoparticles tracking of released small‐EVs using a Spectradyne NT instrument. Small‐EVs evaluated were obtained by EV‐CATCHER (anti‐CD63), MagCapture (anti‐CD63), ExoFlow (anti‐CD63), and global ExoEasy, Norgen purification kit, ExoQuick, and ultracentrifugation purifications methods. g. Transmission Electron Microscopy (TEM) with direct magnification of 20,000x and scale bars of 100 nm for small‐EVs purified from human serum using EV‐CATCHER (anti‐CD63, duplicated isolations), Exoflow (anti‐CD63), ExoCap (anti‐CD63), Dynabeads T1 MyOne™ beads (anti‐CD63) and ultracentrifugation

FIGURE 2

Identification of a low‐background binding platform and comparison to 11 existing small‐EV purification methods. a. Diagram comparing EV‐CATCHER to 10 commercially available kits and ultracentrifugation for purification of small‐EVs. Three magnetic‐bead purification methods were identified where chemical treatment have been described (1% Formic Acid pH 2.0‐ 4.0, 0.1 M Glycine pH 2.5‐3.0 (Chen et al., 2020; Hisey et al., 2018; Hong et al., 2014), or proprietary buffer (SBI ExoFlow)) for small‐EV release (Barros et al., 2018; Burkova et al., 2019; Gutzeit et al., 2014; Hardin et al., 2018; Huang et al., 2018; Mallegol et al., 2007; Patel et al., 2019; Peterson et al., 2015; Theodoraki et al., 2018; Velandia‐Romero et al., 2020; ). b. Evaluation of non‐specific small‐RNA binding to four commercially different customizable magnetic beads (130 nm to 4.5 μm diameter), including both streptavidin‐coated (3 methods) and Carboxyl‐coated beads (1 method), by comparison to wells used for the EV‐CATCHER assay. A 1xPBS solution containing 100 pg ath‐miR‐159a RNA oligonucleotide was incubated with wells (1‐ EV‐CATCHER in triplicate) or the four different types of customizable magnetic‐beads (2‐ MojoSort, 3‐ Carboxyl‐Dynabeads, 4‐ ExoCap, 5‐ Streptavidin‐coated Dynabeads) in triplicate. Wells and magnetic‐beads were washed 3 times and the eluted RNA was quantified by RT‐qPCR (Left graph). Two additional conditions were tested to prevent non‐specific binding of ath‐miR‐159a including a pre‐treatment with BSA (1%) (middle graph), and a combined pre‐treatment with BSA (1%) and RNase‐A (12.5 μg/ml) (Right graph). c. Evaluation of non‐specific binding of ath‐miR‐159a as a contaminant in 100 μl of human serum for all 10 commercial kits (#2 to #11) and the ultracentrifugation method (#12). All magnetic‐bead kits were tested in their ‘Exosome’ purification commercial forms with antibodies on their surface (anti‐Tim4 for Fujifilm and CD63 for all other techniques) and compared to anti‐CD63 EV‐CATCHER (#1). Following manufacturer's instructions, small‐EVs were eluted and RNA was extracted prior to RT‐qPCR quantification of ath‐miR‐159a. All experiments were repeated three times. The averaged comparative thresholds (Ct) values are added above the standard deviation bars in the graph, for each commercial kit and method. d. Evaluation of non‐specific binding of MCF‐7 small‐EVs to EV‐CATCHER and 4 customizable small‐EV commercial magnetic‐bead purification assays. The experimental collection of non‐specifically bound small‐EVs or small‐EVs remaining in solution after non‐specific capture with EV‐CATCHER and the 4 magnetic bead‐based commercial kits is displayed in the schematic (above). RNase A (12.5 μg/ml) was added prior to all experiments, to remove free‐floating miRNAs (Compare grey bars to black bars in all graphs). The top two graphs represent RT‐qPCR quantifications of hsa‐miR‐21 from wells or magnetic beads exposed to the MCF7 solution containing 10 μg of small‐EVs/ ‘exosomes’(left) or remaining in the solution (right), after RNA extraction (All experiments were performed in triplicate). The lower two graphs represent the same conditions with quantification of hsa‐miR‐200c (Non‐specific binding to wells or magnetic beads (left) or remaining in solution (right). All RT‐qPCR experiments were performed in triplicate for each method. Data is represented as fold change with technical replicates (subtraction of ath‐miR‐159a and 10 μg MCF‐7 control). e. Western blot analyses (8 μg total protein) of small‐EVs purified by CD63 EV‐CATCHER and all commercial kits (Tim4 antibody for Fujifilm, and CD63 antibody for all other magnetic bead‐based methods) and methods (ExoEasy, Nörgen, ExoQuick, ultracentrifugation). Anti‐human ‐ApoB, ‐CD63, ‐Albumin, ‐ApoA, ‐CD9, ‐CD81 antibodies were used to evaluate the small‐EVs purified by each of the 11 selected methods/commercial kits. f. Nanoparticles tracking of released small‐EVs using a Spectradyne NT instrument. Small‐EVs evaluated were obtained by EV‐CATCHER (anti‐CD63), MagCapture (anti‐CD63), ExoFlow (anti‐CD63), and global ExoEasy, Norgen purification kit, ExoQuick, and ultracentrifugation purifications methods. g. Transmission Electron Microscopy (TEM) with direct magnification of 20,000x and scale bars of 100 nm for small‐EVs purified from human serum using EV‐CATCHER (anti‐CD63, duplicated isolations), Exoflow (anti‐CD63), ExoCap (anti‐CD63), Dynabeads T1 MyOne™ beads (anti‐CD63) and ultracentrifugation

FIGURE 3

Evaluation of small‐EV purification reproducibility using EV‐CATCHER with different biological fluids. a. Western blot evaluation of small‐EV surface markers using anti‐Alix, ‐CD63, ‐CD9, and ‐CD81 antibodies. EV‐CATCHER purification of CD63⁺ small‐EVs from MCF‐7 tissue culture media (left) and human plasma (middle) with decreasing total protein inputs before purification. For left and middle Western blots, 3 μg (lane 1), 2.75 μg (lane 2), 2.5 μg (lane 3), 2.25 μg (lane 4), 2 μg (lane 5), 1.75 μg (lane 6), and 1.5 μg (lane 7) total protein was used before EV‐CATCHER purification. EV‐CATCHER was also used for purification of CD63⁺ small‐EVs from a serum sample, and validated by the four different surface protein antibodies b. Western blot evaluation of ApoB, Albumin, and ApoA1 proteins from MCF7, human plasma, and human serum CD63⁺ small‐EVs purified with EV‐CATCHER. c. Transmission electron microscopy (TEM) of MCF‐7 small‐EV stock (left, MCF‐7 Stock) and anti‐CD63 EV‐CATCHER purified CD63⁺ small‐EVs from MCF‐7 small‐EV stock (MCF‐7), CD63⁺ small‐EVs from human plasma (Plasma), and CD63⁺ small‐EVs from human serum (Serum). Direct magnification of 20,000x and scale bars of 200 nm are represented on the TEM images. c. Representative particle size distribution characterized by nanoparticle tracking (Spectradyne nCS1 equipped with TS400 microfluidic cartridges) for anti‐CD63 EV‐CATCHER‐purified small‐EVs from MCF‐7 small‐EV stock (left graph), human plasma (middle graph), and human serum (right graph). Peak filtering performed for diameter < 65 nm and transit times > 80 μs. The concentrations are representative of particles detected between 65 and 150 nm

FIGURE 4

Small‐RNA sequencing of small‐EVs purified from biofluids. a. Next‐generation sequencing and heatmap representation of the top expressed circulating miRNAs in human serum. The heat map includes the top 49 expressed miRNAs and their detectability using decreasing amounts of total RNA extracted from human serum (6 ng, 3 ng, and 1.5 ng, in duplicate). b. Heat map representation displaying miRNA expression differences between total RNA extracted from mouse RAWS264.7 tissue culture EVs (first two rows), commercially available whole human plasma (third and fourth rows; same plasma as the one analysed in Figure 3), and small EVs obtained by ultracentrifugation from the same commercially available whole human plasma sample (fifth and sixth rows). Duplicate libraries were obtained using newer (Repeats#1) and older (Repeats#2) 3′ barcoded adapters. c. EV‐CATCHER specific capture of mouse RAWS264.7 CD63⁺ small‐EVs, which represent a subset of all CD63⁺ small‐EVs, from the RAWS264.7 total small‐EVs that were spiked into human plasma (same plasma as the one evaluated in Figure 3). The heatmap represents miRNA expression differences between total RNA from human plasma (columns I and II) and total RNA from mouse RAWS264.7 total small‐EVs (columns V and VI) and include the top 58 human plasma‐specific miRNAs and the top 50 mouse RAWS264.7 small‐EV specific transcripts differentially expressed between total RNA from human plasma and total RNA from mouse RAWS264.7 total small‐EVs, respectively. Small‐RNA sequencing profile of CD63⁺ mouse RAWS264.7 small‐EVs (1 μg) purified from human plasma (100 μl) using the CD63⁺ EV‐CATCHER assay harbouring a mouse‐specific anti‐CD63 antibody (columns III and IV). Duplicate libraries were obtained using newer (Repeats#1) and older (Repeats#2) 3′ barcoded adapters. Data was analysed using dedicated Bioconductor packages in the R platform and heatmaps were generated using the ‘NMF’ package (a heatmap function)

FIGURE 5

The EV‐CATCHER assay allows for identification of miRNA expression differences between total RNA from small‐EVs and whole serum. a. Individual box plot analyses of the 10 differentially expressed miRNAs identified from total RNA extracted from CD63⁺/CD9⁺/CD81⁺ small‐EVs purified from serum with EV‐CATCHER (green square; hsa‐miR‐146a, hsa‐miR‐126‐3p, hsa‐miR‐424, hsa‐miR‐151‐3p, hsa‐miR‐126‐5p, hsa‐miR‐627‐5p, hsa‐miR‐145, hsa‐miR‐205, and hsa‐miR‐200c) and 2 differentially expressed miRNAs identified from total RNA from whole serum (orange square; hsa‐miR‐550‐5p, hsa‐miR‐629*) of the same mildly and severely ill Covid‐19 hospitalized patients (first set of Covid‐19 serum samples), by next‐generation small‐RNA sequencing. b. Box‐plot of miRNA amount representation (in zeptomoles (10^–21 mole) of the top 10 differentially expressed miRNAs identified from small‐EVs purified with EV‐CATCHER from the serum of mildly (n = 13) and severely ill Covid‐19 hospitalized patients (n = 17), per sample. c. Box plots representation of the miRNA integrative signature (including the top 10 miRNAs identified in (a.)) between the two small‐RNA libraries prepared with RNA extracted from small‐EVs purified from mild and severely ill patients (library #3 includes 7 mild cases and 8 severe cases, and library #4 includes 6 mild cases and 9 severe cases), displaying significant differences between the two patient groups. d. Box plot representation of the integrative miRNA signature (10 miRNAs) between total RNA extracted from CD63⁺/CD9⁺/CD81⁺ small‐EVs (subset of all CD63⁺/CD9⁺/CD81⁺ small‐EVs in the serum samples) purified with EV‐CATCHER and total RNA extracted from whole sera between mildly and severely ill patients. e. Box plot RT‐qPCR validations of the top 4 differentially expressed miRNAs identified between mild and severely ill Covid‐19 hospitalized patients by small‐RNA sequencing, with RNA extracted from small‐EVs purified with EV‐CATCHER. Data are represented as ΔΔCt with technical triplicates performed for each individual sample. Differential expression was assessed with DESeq2 (R/Bioconductor package) for sequencing results and t‐tests for miRNA score and qPCR results. Median comparative threshold (Ct) for all samples from each group (mildly or severely ill hospitalized patients) is presented in boxes below box‐plots

FIGURE 6

The EV‐CATCHER assay releases functional small‐EVs. a. Six convalescent serum samples (Second set of covid‐19 serum samples) distributed in two groups based on presence of immunoglobulins (IgG) targeting the Receptor Binding Domain (RBD) of SARS‐CoV‐2 spike protein as antigen. Three high anti‐spike IgG titer (left) and three below level of quantification (BLQ) anti‐spike IgG titer serum samples were quantified by ELISA (Dr. Perlin's Laboratory). High anti‐spike IgG titers were estimated with activity remaining at a 10,000x (fold) dilution, while below level of quantification (BLQ) IgG serum samples did not contain detectable IgG against the spike protein RBA region of SARS‐CoV‐2 at the same dilution. b. Western blot analyses of CD63⁺ small‐EVs purified with anti‐CD63 EV‐CATCHER from sera of high anti‐spike (RBD region) IgG (lanes 1–3) and BLQ IgG (lanes 4–6) serum samples using anti‐ApoB, ‐Alix, ‐CD63, ‐Albumin, ‐ApoA1, ‐CD9, and ‐CD81 antibodies. c. Representative Transmission electron microscopy (TEM) images of small‐EVs isolated from high anti‐spike IgG (CDI ‐001, Figure 6b ‐ Lane 1) and BLQ anti‐spike IgG (CDI‐004, Figure 6b ‐ Lane 6) convalescent sera using the anti‐CD63 EV‐CATCHER assay. Direct magnification was 20,000x and scale bars are for 200 nm. d. Nanoparticle size distribution characterized using the Spectradyne nCS1 instrument with a TS400 microfluidic cartridges of CD63⁺ small‐EVs purified from all six serum donors using the anti‐CD63 EV‐CATCHER assay. e. Schematic representation of the experimental procedure used to obtain high‐purity small‐EVs from convalescent sera by ultracentrifugation. f. In vitro assessment of SARS‐CoV‐2 infection using Vero E6 cells treated with whole sera or small‐EVs purified from the different sera. Healthy Vero E6 cells Control (bar 1, no virus) were subjected to mNG SARS‐CoV‐2 (bar 2, virus), small‐EVs ultracentrifuged (EV control) from serum with high anti‐spike IgG (bar 3) and serum with BLQ anti‐spike IgG (bar 4), mNG SARS‐CoV‐2 with whole sera with high anti‐spike IgG (bars 5, 7 and 9) or BLQ anti‐spike IgG (bars 11, 13 and 15), mNG SARS‐CoV‐2 after Vero E6 cells were treated with CD63⁺ small‐EVs purified with anti‐CD63 EV‐CATCHER from high anti‐spike IgG sera (bars 6, 8, and 10), or with CD63⁺ small‐EVs purified with anti‐CD63 EV‐CATCHER from BLQ anti‐spike IgG sera (bars 12, 14, and 16), and finally UNG digest buffer (bar 17) and UNG digest buffer with mNG SARS‐CoV‐2 (bar 18) for 72 h. Statistical analyses were performed using one‐way ANOVA with Bonferroni post hoc correction. All results are presented as mean ± SEM (n = 3), and **** = P < 0.0001 vs. virus. g. Fluorescent imaging of Vero E6 cells infected with the mNeonGreen SARS‐CoV‐2 reporter virus. Hoechst 33342 and mNeonGreen was visualized using fluorescent Celigo Cell Imaging. Representative fluorescent images display healthy Vero E6 cells (no virus control), infected with mNG SARS‐CoV‐2 and treated with high anti‐spike IgG whole serum (high anti‐spike IgG serum), infected with mNG SARS‐CoV‐2 and treated with CD63⁺ small‐EVs purified with the anti‐CD63 EV‐CATCHER assay from high anti‐spike IgG serum (High anti‐spike IgG patient# CDI‐001), infected with mNG SARS‐CoV‐2 (virus control), infected with mNG SARS‐CoV‐2 with BLQ anti‐spike IgG whole serum (BLQ IgG serum; Patient # CDI‐004), infected with mNG SARS‐CoV‐2 and treated with CD63⁺ small‐EVs purified with the anti‐CD63 EV‐CATCHER assay from BLQ anti‐spike IgG serum (small‐EVs (BLQ serum)). Scale bars on images was at 500 μM