Comparison of serum exosome isolation methods on co-precipitated free microRNAs

Sample collection

Human blood samples were collected into serum collection tubes from the Shanghai Chest Hospital. Blood samples were centrifuged at 3000 rpm at 4 °C for 10 min for serum collection, followed by an additional centrifugation step at 3,000 g for 15 min. The supernatants of different samples were mixed and then equally divided into several portions (500 µL) for subsequent experiments.

This study was approved by the Ethical Committee of the School of Biomedical Engineering, Shanghai Jiao Tong University (Registration No. 2019045) and carried out in accordance with the Declaration of Helsinki. All participants provided written informed consent.

Exogenous miRNA addition

The serum (500 µL) were added with 5 µL 7 ×10⁹ copies/µL cel-miR-39 or 5 µL 7 ×10⁶ copies/µL cel-miR-39 or 5 µL ddH₂O. The sequence of cel-miR-39 is: forward 5′-UCACCGGGUGUAAAUCAGCUUG-3′.

Exosome isolation

Exosome identification

RNase A and RNase Inhibitor treatment

The exosome pellet was suspended in 500 µL PBS and a determined volume of RNase A (Thermo Scientific, MAN0012003) was added in the RNase A treatment groups to reach a final concentration of 10 µg/mL, 20 µg/mL or 30 µg/mL. And RNase inhibitor (Thermo Scientific, EO0381) was added in the RNase inhibitor treatment groups to reach a final concentration of 1 U/µL. Then the mixture was incubated at 37 °C for 1 h.

RNA isolation

Exosomal RNA was extracted by miRNeasy Mini Kit (Qiagen, 217004) according to the manufacturer’s recommendations. The final elution volume was 35 µL.

Reverse Transcription Polymerase Chain Reaction (RT-PCR)

The cDNA of cel-miR-39, miR-126 and miR-152 were synthesized with the Revert Aid First Strand cDNA Synthesis Kit (Thermo Scientific, K1622). Briefly, 5 µL total RNA from each sample were mixed with 5 µL of RT-Primer, 2 µL of 10 ×T4 Buffer, 2 µL of 10mM dNTP Mix, 1 µL of RNase inhibitor, 1 µL of Transcriptase, 0.5 µL of T4 Polynucleotide Kinase, 0.25 µL of T4 Kinase, and 3.25 µL of nuclease-free water. Then the 20 µL of mixture was incubated at 16 °C for 30 min, 42 °C for 30 min, 85 °C for 5 min followed by cooling at 4 °C in the PCR instrument (LongGene, A100).

Note that 5 µL of 35 µL extracted RNA were used in RT-PCR, so if all the exogenous cel-miR-39 had been co-isolated with exosomes, the current concentration of cel-miR-39 in this procedure would be 10⁹ copies/µL or 10⁶ copies/µL. Therefore, 10⁹ copies/µL or 10⁶ copies/µL cel-miR-39 were used as a standard control.

Here, stemloop RT primers were used for cDNA synthesis. The sequence of miR-126 is: 5′-UCGUACCGUGAGUAAUAAUGCG-3′. The RT primer of miR-126 is: 5′-GATGAGGAGTGTCGTGGAGTCGGCAATTTCCTCATCACGCATTA-3′. The sequence of miR-152 is: 5′-AGGUUCUGUGAUACACUCCGACU-3′. The RT primer of miR-152 is: 5′- GATGAGGAGTGTCGTGGAGTCGGCAATTTCCTCATCAAGTCGGAG-3′. The RT primer of cel-miR-39 is: 5′- GATGAGGAGTGTCGTGGAGTCGGCAATTTCCTCATCCAAGCTG -3′.

Quantitative real-time PCR (qRT-PCR)

To validate the quantity of cel-miR-39, miR-126 and miR-152 in the RNA samples, qPCR was performed using Fluorescence Polymerase Chain Reaction (PCR) Detection Kit for the Analysis of Human miRNA Gene Expression (Jiangsu MicroDiag Biomedicine Technology Co. Ltd., China) according to the manufacturer’s recommendations. Briefly, the synthesized cDNA of cel-miR-39, miR-126 and miR-152 from each sample were mixed with 10 × Buffer, 25 mM MgCl2, 25 mM dNTP Mix, F-primer, R-primer, miRNA-specific probe, HS Tag, UDG and ddH2O. Meanwhile, 10² - 10⁷ copies of cel-miR-39, miR-126 and miR-152 were amplificated in this procedure to get a curve for calculating miRNA copy numbers in each sample.

The PCR reaction mixture was incubated at 37 °C for 5 min, and then 94 °C for 5 min, followed by 50 cycles of 94 °C for 15 s, 60 °C for 60 s, and cooling at 50 °C for 30 s, which was performed using a real-time fluorescent quantitative PCR instrument (Roche, LightCycler 480).

Here, the miRNA-specific probes were used. The F-primer of miR-126 is: 5′-CTCCGTCAGGGTCGTACCGTGAGTAA-3′. The R-primer of miR-126 is: 5′- CTCAAGTGTCGTGGAGTCGGC-3′. The specific probe of miR-126 is: 5′-FAM-TTTCCTCATCACGCATTA-MGB-3′. The F-primer of miR-152 is: 5′-CTCCGTCAGGGAGGTTCTGTGATACA-3′. The R-primer of miR-152 is: 5′-CTCAAGTGTCGTGGAGTCGGC-3′. The specific probe of miR-152 is: 5′-FAM-TTTCCTCATCAAGTCGGAG-MGB-3′. The F-primer of cel-miR-39 is: 5′-CGTATGAGCG TCACCGGGTGTAAATCA-3′. The R-primer of miR-152 is: 5′-CTCAAGTGTCGTGGAGT CGGCAA-3′. The specific probe of miR-152 is: 5′-FAM-TTTCCTCATCCAAGCTG-MGB-3′.

Statistical analysis

Each experiment was repeated at least three times. Data were analyzed and all statistical graphs were generated by GraphPad Prism 6.0 (GraphPad Software Inc, La Jolla, CA, USA). Differences between two groups were analyzed using one-way or two-way ANOVA and the probability value below 0.05 was considered significant.

Workflow of this research and the identification of exosomes

The experimental procedure of this research is summarized (Fig. 1), with details given in subsequent sections. First, 5 µL 7 ×10⁹ copies/µL or 7 ×10⁶ copies/µL exogenous cel-miR-39 was added in 500 µL human serum while same volume of ddH₂O was added in another group as blank control (step I). Next, the exosomes were isolated by ExoQuick or UC method (step II) and characterized using TEM, NTA and western blot (Figs. 2A–2E). TEM showed exosomes’ “cup-shaped” morphology and co-isolated impure proteins in both methods (Figs. 2A, 2D). NTA analysis showed exosomes’ similar particle size distribution mostly between 30-200 nm (Figs. 2B, 2E), however, ExoQuick-isolated exosomes tended to be a little smaller than those from UC, which may be caused by more co-precipitated impure proteins from ExoQuick method. Furthermore, exosomes from both methods obtained the exosome-associated protein marker, CD63, TSG101, CD9 (Figs. 2C, 2F) (Thery et al., 2018). Following, to clear up the co-isolation free miRNA in exosome pellets, 10 µg/mL, 20 µg/mL or 30 µg/mL RNase A were added in RNase A group and incubated with exosomes at 37 °C for 1 h. Moreover, 1 U/µL RNase inhibitor were used to eliminate the effect of environmental RNase; the control group was a blank control with no enzyme addition, which was also placed at 4 °C for 1 h (step III). Then, total RNA including miRNA of exosomes was extracted by QIAGEN miRNeasy Mini Kit and cel-miR-39, miR-126 or miR-152 in exosome samples were detected using RT-PCR followed by qPCR. 5 µL 10⁹ copies/µL or 10⁶ copies/µL cel-miR-39 was used as standard control in PCR procedure (step IV, V and VI).

ExoQuick co-isolates a small proportion of free miRNAs

As described in Fig. 1, higher concentration (7 ×10⁹ copies/µL) of cel-miR-39 was added in 500 µL serum and then exosomes were extracted by ExoQuick method. The ratio of cel-miR-39 in exosome samples to standard cel-miR-39 sample (10⁹ copies/µL) showed that about 2% exosome-independent cel-miR-39 were co-isolated with exosomes extracted by ExoQuick method (Fig. 3A). To measure the effect of RNase A to remove exosome-independent miRNAs, 10 µg/mL, 20 µg/mL, 30 µg/mL RNase A and 1 U/µL RNase inhibitor were respectively added in the same exosome samples, and the results of PCR showed that RNase A can decrease the quantity of co-isolated cel-miR-39 in exosomes but there are no significance difference between different concentrations of RNase A and RNase inhibitor groups (Fig. 3A). Moreover, RNase A decreased the quantity of exosomal miR-126 and miR-152 at the same time compared with control and RNase inhibitor groups (Fig. 3B). Interestingly, RNase inhibitor reduced the amount of exosomal miR-126 but not reduced miR-152 (Fig. 3B).

To verify if the co-isolation is due to the high concentration of cel-miR-39, low concentration (7 ×10⁶ copies/µL) of miR-39 were exogenously added in serum. Similar to the results of high concentration cel-miR-39 experiment, nearly 2% cel-miR-39 were co-isolated with exosomes and RNase A decreased the amount of exosomal miR-126 and miR-152 (Figs. 3C–3D). And RNase inhibitor reduced the amount of exosomal miR-126, not the miR-152 in low concentration cel-miR-39 experiment as well (Fig. 3D). These results indicated that ExoQuick method can co-isolate 2% free miRNA in serum regardless of the concentrations of miRNAs and RNase A not only can reduce the number of exosome-independent miRNAs, but also reduce the miRNAs in the exosomes.

Ultracentrifugation co-isolates almost no proportion of free miRNAs

To investigate if UC can co-isolate free miRNAs in serum with exosomes as ExoQuick method did, higher concentration (7 ×10⁹ copies/µL) of cel-miR-39 was added in 500 µL serum and then exosomes were extracted by UC method. The results showed that only 0.03% exogenous cel-miR-39 were co-isolated with exosomes extracted by UC method (Fig. 4A). Meanwhile, exosomal miR-126 and miR-152 were decreased by RNase A and miR-126 rather than miR-152 was reduced by RNase inhibitor (Fig. 4B). These results indicated that UC method hardly precipitate any exosome-independent miRNAs with exosomes and RNase A can reduce the miRNAs in exosomes no matter which method is used.

Incubation at 37 °C for 1 h reduces the amount of co-isolated free miRNAs

In the previous experiments, we noticed that RNase inhibitor and RNase A both reduced the amount of exogenous cel-miR-39. These two reagents both require a 37 °C incubation with exosomes for 1 h. We speculated if incubation at 37 °C for 1 h could cause the reduction of cel-miR-39. Here, we isolated the exosomes by ExoQuick method, and the results showed that incubation at 37 °C for 1 h reduced the amount of cel-miR-39 as much as the RNase A did (Fig. 5A). Consistent with the previous results in Figs. 3B, 3D and 4B, RNase A reduced the amount of exosomal miR-126 and miR-152, but incubation at 37 °C for 1 h reduced the amount of miR-126 but not miR-152 like the RNase inhibitor did (Fig. 5B). These results indicated that incubation at 37 °C for 1 h without RNase inhibitor and RNase A can remove the co-isolated free miRNAs from exosome pellets, but RNase A can reduce the number of miRNAs in exosomes as well.