key: cord-0738010-276ffzo4 authors: Dunlop, Rachael Anne; Banack, Sandra Anne; Cox, Paul Alan title: A comparison of the efficiency of RNA extraction from extracellular vesicles using the Qiagen RNeasy MinElute versus Enzymax LLC RNA Tini Spin columns and qPCR of miRNA date: 2021-07-21 journal: Biol Methods Protoc DOI: 10.1093/biomethods/bpab015 sha: cfc0d396dbf7328e6cb2429b752cb730de42ab8f doc_id: 738010 cord_uid: 276ffzo4 One consequence of the current SARS-CoV-2 pandemic is an interruption to the supply of laboratory consumables, particularly those used for RNA extraction. This category includes column-based RNA extraction kits designed to retain short RNA species (defined as having fewer than 200 nucleotides), from small sample volumes, for example exosomes or extracellular vesicles. Qiagen manufactures several kits for the extraction and enrichment of short RNA species, such as microRNA, which contain silica-membrane columns called “RNeasy MinElute Spin Columns”. These kits, which also contain buffers and collection tubes, range in price from USD380 to greater than USD1000 and have been subject to fulfillment delays. Scientists seeking to reduce single-use plastics and costs may wish to order the columns separately, however Qiagen does not sell the RNeasy MinElute Spin Columns (in reasonable quantities) as an individual item. Thus, we sought an alternative product and found RNA Tini Spin columns from Enzymax LLC. We conducted a systematic comparison of the efficiency of RNA extraction for miRNA qPCR using the Qiagen RNeasy MinElute Spin Columns and Enzymax LLC RNA Tini Spin columns and the Qiagen total RNA extraction protocol that enriches for short RNA species. We compared the efficiency of extraction of five spike-in control miRNAs, five sample signal miRNAs, and eight low- to medium-abundance miRNAs by qPCR. The RNA was extracted from extracellular vesicle preparations purified from human plasma using CD81 immunoprecipitation. We report no statistically significant differences in extraction efficiencies between the two columns for any of the miRNAs examined. Therefore, we conclude the Enzymax RNA Tini Spin columns are adequate substitutes for the Qiagen RNeasy MinElute Spin Columns for short RNA species enrichment and downstream qPCR from extracellular vesicles in the miRNAs we examined. exosomes" since L1CAM/CD171 is enriched in the brain (see https://gtexportal.org/home/gene/L1CAM). However, since L1CAM/CD171 is also highly expressed in neurons in the gut, colon, and tibial nerve and "exosome" preparations can be heterogeneous in size and composition, we prefer to use the term neural-enriched extracellular vesicles (NEE) . It naturally follows that we refer to the preparation left behind once the NEE are extracted as total-minus-NEE fraction or T-N. As stated above, exosomes contain a wide range of proteomic and genetic material. miRNAs are small (between 18 and 22 nucleotides), non-coding RNAs that regulate gene expression and protein translation and are enriched in EVs, including exosomes and NEEs (Kim, 2005) . A large body of evidence reporting on disparate pathologies demonstrates that the differential expression of miRNAs can be diagnostic, prognostic, or both for various diseases, including some cancers (He et al., 2021) and neurodegenerative diseases (You and Ikezu, 2019) . For example, we recently reported an eight miRNA fingerprint in NEEs that distinguished ALS patients from healthy controls (Banack et al., 2020) . Conventional RNA extraction and quantitation techniques, such as those that preferentially capture long (greater than 200 nucleotides nt) RNA species including mRNA, are unsuitable for retaining short RNA (fewer than 200 nucleotides) species. Furthermore, when using column-based RNA purification methods, the protocol needs to be designed specifically to retain short RNA species and this requires specialised wash buffers and optimisation of the ethanol volume and concentration added to the column during the washing steps. Qiagen is a leader in RNA extraction and makes kits designed to isolate short RNAs in a total RNA fraction from serum or plasma-derived EVs using silica-based spin columns (ExoRNeasy Midi/Maxi Kit, #77144, 77164, 77023, RNeasy Micro Kit, #74004, RNeasy MinElute Clean-up kit #74204 and miRNeasy microkit #217084) . In the case of Qiagen's RNA ExoRNeasy kit, the kit comes in two parts --Part 1: Vesicle Isolation and Part 2: RNA isolation. This is appropriate for researchers who want to isolate EVs and RNA using the one kit, but for researchers who isolate EVs using a different method, such as CD81 immunoprecipitation, choosing a kit without Part I is appropriate to save costs and plastics. Whilst the buffers for total RNA extraction enriching for short RNAs can be bought separately, the RNeasy MinElute Spin Columns cannot. Enzymax LLC also makes spin columns to extract total RNA including short RNA species (RNA Tini Spin column with collection tube, #EZC1070N) and these can be purchased separately in batches of 50. A significant hurdle in using miRNA to design diagnostic or prognostic tests for human disease is access to patient samples. Clinical trials can be a source for plasma, but samples from such trials are frequently only available in small volumes for discovery research. Thus, the researcher is working with minimal amounts of material to begin with, meaning preferred methods to extract RNA must be efficient in preserving as much RNA as possible. When searching for miRNA biomarkers, the researcher generally carries out a global measure of all short RNA species using next-generation sequencing (NGS) to identify species of interest and then quantitates these findings using real-time quantitative PCR (qPCR). Although qPCR amplifies targeted RNA many times over, it can still be a problem to obtain a robust signal after 40 cycles, especially when working with NEE, which is ~5-10% of the fraction of total EVs (Mustapic et al., 2017) . For this reason, it is critical to preserve all RNA throughout the purification process, which begins with the columnbased purification process. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 We have previously described an eight miRNA fingerprint that distinguishes blood plasma samples from ALS patients and healthy controls (Banack et al., 2020) . Since Qiagen does not sell the RNeasy MinElute Spin Columns separately, we sought an alternative, and decided to test RNA Tini Spin columns manufactured by Enzymax LLC #EZC107N. We compared RNA yield from duplicate T-N aliquots taken from the same EV preparations, and from three individual blood plasma samples (n =3), using the Qiagen RNeasy MinElute Spin Columns and the RNA Tini Spin columns manufactured by Enzymax LLC. We report no significant differences in the mean raw quantitation cycles (Cqs) for any of the miRNAs when comparing RNA extracted from the Qiagen columns versus the Enzymax columns. Thus, we conclude that the Enzymax RNA Tini Spin columns are a suitable substitute for the Qiagen RNeasy MinElute Spin Columns for medium to high-abundance miRNA. For low-abundant miRNA, researchers should check their extraction on a case-by-case basis. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Materials and methods Materials: Normal single donor human plasma, frozen once, was obtained from Innovative Research (10 mls, #IPLA-N-S, Novi, Michigan, USA), and EVs extracted as described previously (Section 2.3. EV Extraction, Banack et al., 2020) . We used the total-minus-neural (T-N) fraction of the EV preparations for these experiments. T-N represents the total heterogeneous EV population minus the EVs positive for L1CAM/CD171 neural surface proteins, designated NEE (Banack et al., 2020 Preparation of spike-in controls for RNA extraction and cDNA synthesis efficiency: Spike-in control templates for RNA extraction (available as a pre-mixed, lyophilized vial of UniSp2, UniSp4, and UniSp5 from the Qiagen miRCURY RNA spike-in kit, for RT #339390) and cDNA synthesis controls UniSp6 (available as a lyophilized vial as part of the miRCURY LNA RT Kit #339340), and cel-miR-39-3p (available as a lyophilized vial as part of the Qiagen miRCURY RNA spike-in kit, for RT #339390) were removed from the freezer (-20˚C) and briefly centrifuged. 80 µL nuclease-free water was added to each of the vials, the vials were incubated for 20 mins on ice, then vortexed, and briefly centrifuged. To avoid repeat freeze-thaw cycles, 5 µL aliquots were stored at -20˚C. RNA extraction from EVs: Extraction of total RNA retaining short RNA species from T-N was conducted following the manufacturer's instructions as described in Qiagen RNeasy Midi Kit Part 2: RNA Isolation (a detailed protocol is available at www.qiagen.com/HB-2630 "exoRNeasy Midi/Maxi Handbook" beginning page 21, step 6) with the following modifications: for each T-N sample to be lysed, 1 µL of a mix of UniSp2, UniSp4, and UniSp5 (see above for preparation instructions) was added to 700 µL of QIAzol lysis reagent (#79306) and mixed by pipetting up and down. 50 µL T-N pellets were removed from -80˚C freezer and thawed on ice, the sample transferred to a fresh RNAse-free 2 mL tube, then 700 µL of the QIAzol/spike-ins mix was added, and each sample mixed by pipetting and briefly vortexed. Tubes were incubated at RT for 5 min, 90 µL chloroform (MP Biomedicals #0219400225) added, then the tube capped securely and shaken vigorously for 15 secs. Tubes were incubated for 5 mins at RT then centrifuged at 12,000 x g for 15 mins, 4˚C. The upper aqueous phase (approximately 400 µL), containing total RNA, was carefully transferred to a new 2 mL collection tube being careful not to touch the interphase layer (this contains DNA). Two x volumes of 100% ethanol were added (approximately 800 µL), and the sample mixed thoroughly by pipetting. The use of 2 volumes of 100% ethanol at this stage selects for retaining short RNA species on the spin column. Up to 700 µL sample was pipetted into either a Qiagen RNeasy RNA MinElute Spin Column or an Enzymax RNA Tini Spin column placed inside a 2 mL RNAse-free collection tube. The tube lid was closed tightly and spun at 8000 x g for 15 seconds at RT. The flow-through was discarded. This step was repeated until all the sample was passed through the spin column. 700 µL Buffer RWT (Qiagen #1067933) was added to each column, the lid closed, then the tube centrifuged for 15 seconds at 8000 x g, RT. The flow-through was discarded. 500 µL Buffer RPE (Qiagen #1018013) was added to each 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 column, the lid closed, and the tubes centrifuged for 15 seconds at 8,000 x g, RT, then the flow-through discarded. This step was repeated, but the second time, the tubes were spun for 2 min and the flowthrough discarded. Finally, the spin column was placed into a 1.5 mL RNAse-free collection tube, and 15 µL nuclease-free water added directly to the center of the column, the column was let to stand for 1 min, RT, then spun for full speed at RT for 1 min. The RNA was immediately placed on ice and used for cDNA synthesis or stored at -80˚C. cDNA synthesis: cDNA was synthesised using the miRCURY LNA RT Kit (Qiagen #339340) according to the manufacturer's instructions. 0.5 µL each of spike-in control UniSp6 and C. elegans cel-miR-39-3p (see above for preparation instructions) was added to each reaction to monitor cDNA synthesis efficiency. To optimize the amount of RNA used in the reaction, we conducted cDNA synthesis with 2 and 4 µL RNA, then ran qPCR of QC and sample signal miRNAs. We used 4 µL RNA for all downstream cDNA reactions. qPCR: real-time PCR QC was conducted using miRCURY LNA miRNA SYBR PCR Assays for the targets listed in Table 1 , according to the manufacturer's instructions. cDNA was diluted 1/30 and 3 µL used in each qPCR reaction. The reaction conditions are described in Table 2 . A melt curve and a notemplate control (NTC) were included in each assay to check for primer specificity and any non-specific amplification. qPCR was conducted on the BioRad CFX96™ Real-Time System in 96-well plates and data acquired in Bio-Rad CFX Manager version 3.1 after 40 cycles. Statistics: Cqs (n = 3) returned for qPCR from RNA extracted using the Qiagen RNeasy MinElute Spin Column or the Enzymax RNA Tini Spin column were compared using the Mann-Whitney test for unpaired non-parametric samples, conducted in GraphPad Prism version 9.1.1. for iOS, where p < 0.05 was set as significant. Given that we were planning to assay some reportedly low-abundance miRNAs, we increased the volume of RNA we would typically use in a cDNA reaction to maximize the likelihood of observing a signal in the qPCR. As standard quality controls (QC), we used RNA spike-ins to check for consistency and efficiency in the RNA extraction (UniSp2, 4, and 5) and cDNA synthesis (UniSp6 and cel-miR-39-3p). We synthesized cDNA using either 2 µL RNA (cDNA1 and cDNA2) or 4 µL RNA (cDNA3 and cDNA4) extracted from T-N EVs using the Qiagen RNeasy MinElute Spin Column. QC spike-ins and five sample signal miRNAs were quantitated using qPCR, as described in materials and methods. As expected, we report raw Cq differences in the signal for the RNA extraction spike-ins since we added twice as much to the cDNA reaction ( Figure 1A and B) . Conversely, but also as expected, we report no difference in the cDNA synthesis spike-ins since these were added directly to the reaction mixture in equal quantities (Figure 1A and B) . To further check the optimal volume of RNA to use in the cDNA reaction, we conducted additional QC targeting five sample signal miRNAs, hsa-miR-142-3p, hsa-miR-451a, hsa-miR-23a-3p, hsa-miR-30c-3p, hsa-miR-103a-3p, and hsa-miR-191-5p. The smallest difference in Cq between 2 µL RNA and 4 µL RNA was 1.10 for hsa-miR-142-3p, and the highest was 1.71 for hsa-miR-191-5p (Figure 2A and B) . To maximize the likelihood of observing a signal for predicted low-abundant miRNAs, we used 4 µL RNA for all downstream cDNA synthesis reactions. Spike-in controls were added at two stages: 1) RNA extraction (UniSp2, 4 and 5), and 2) cDNA synthesis (UniSp6 and cel-miR-39-3p). 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For the RNA extraction spike-ins, we report a difference in Cqs of 1.05 for UniSp2 and 0.97 for UniSp5 between column types (Figure 3A and B) , where the Qiagen RNeasy MinElute Spin Columns returned a lower Cq (indicating there was more miRNA in the qPCR reaction). In theory, each PCR cycle doubles the amount of amplicon in a reaction. Hence, a Cq difference of one equates to a doubling of material, assuming the qPCR reaction is 100% efficient. However, these differences were not significantly different (see Figure 3B) , and we report no other significant differences for the RNA extraction spike-in controls (Figure 3) . We also reported no significant difference in Cqs between column types for UniSp6 and cel-miR-39-3p (see Figure 3 ) for the cDNA synthesis controls. Although exosomes are reported to contain mRNA, we have had difficulty using standard RNA quality/quantity methods (such as a bioanalyzer or spectrophotometer) to accurately quantitate RNA from our EV samples. Thus, to quantitate the amount of RNA extracted by the different columns, we selected five miRNAs representative of the sample miRNA signal (Banack et al., 2020) . Some of these targets are predicted to be found in low abundance and others in high abundance (see Figure 4) . We then conducted qPCR on the cDNA synthesized from RNA extracted using the Qiagen RNeasy MinElute Spin columns or the Enzymax RNA Tini Spin columns. We report no difference in the mean raw Cqs for the five-sample signal miRNAs between different column types (See Figure 4A and B) , suggesting the columns were equally efficient at extracting these miRNAs. We have previously reported medium-to low-abundance miRNA in NEE (Banack et al., 2020) ; thus, we are particularly interested in determining any differences in the extraction efficiency of these miRNAs between the two column types. We used eight miRNA we have previously examined (Banack et al., 2020) and measured raw Cqs in the cDNA synthesized from RNA extracted using the Enzymax RNA Tini Spin or Qiagen RNeasy MinElute Spin columns. We report no significant differences in the levels of the medium-to low-abundance miRNAs (see Figure 5A and B), suggesting that the two columns are equally efficient at extracting low-to-medium abundance miRNA. The 50 µL T-N samples used in these comparison experiments represent technical replicates aliquoted from the same EV preparation. From three separate human plasma sample EV preparations, we processed two technical replicates from each, for downstream analysis. The six 50 µL aliquots (2 x 3 separate sample preps) were processed in exactly the same manner, except three were processed using the Qiagen RNeasy MinElute Spin columns and the other three, the Enzymax RNA Tini Spin column. For this reason, we consider raw mean Cqs a suitable method for comparison of miRNA levels across the different columns and we did not conduct further normalization of these samples. Owing to the lack of affordability and availability of Qiagen RNeasy MinElute Spin columns independent of a kit, we compared the efficiency of RNA extraction with a purported equivalent, the Enzymax RNA Tini Spin columns. Qiagen RNeasy MinElute Spin columns come in various kits ranging in cost from USD380 to more than USD1000, whereas Enzymax RNA Tini Spin columns can be bought separately in packs of 50 for USD56. Unlike the Qiagen RNeasy MinElute Spin columns, which are not sold separately, the lysis and wash buffers required to conduct the total RNA extraction protocol that retains short RNA and described herein can be purchased separately from Qiagen. We report no significant difference in mean raw Cqs for any of the synthetic extraction spike-ins between the two brands of RNA extraction columns (Figure 3A and B) . Similarly, there was no significant difference in the mean raw Cqs for five sample miRNA signal genes, hsa-miR-142-3p, hsa-miR-451a, hsa-miR-23a-3p, hsa-miR-30c-5p, hsa-miR-103a-3p, and hsa-miR-191-5p (Figure 4A and B) . We compared mean raw Cqs in cDNA samples synthesised using 4 µL RNA extracted from T-N using either the Qiagen RNeasy MinElute Spin columns or Enzymax RNA Tini Spin columns. We report no difference in any of the miRNAs we targeted (Figure 5A and B) . In the experiments described here, we used Qiagen RNeasy MinElute Spin Columns from the more expensive Qiagen exoRNeasy Midi kit (#77144). However, we note that the chemistry of the RNeasy minElute Spin Columns found in other Qiagen kits is not different. For example, the Qiagen RNeasy MinElute clean-up kit (#74204) also comes with 50 RNeasy MinElute Spin Columns (but with different buffers than those used here) and has a list price of USD380. In addition, the miRNeasy micro kit (#217084) also comes with 50 RNeasy MinElute Spin Columns, plus buffers and Qiazol lysis reagent and has a list price of USD424. We learned that a combination of wash buffers and the ethanol volume and concentration used in the extraction protocol determines whether short RNA species are retained on the column. Specifically, diluting the RNA aqueous phase collected after phase separation using Qiazol (Step 11 in the exoRNeasy Midi/Maxi Handbook) with two volumes of 100% ethanol adjusts the binding conditions of the column to retain short RNA species. However, we recommend that researchers conduct their own QC to assess the suitability of these columns for their specific target short RNA species on a case-by-case basis. Data that supports the findings of this study will be made available upon reasonable request from the corresponding author, RAD. The Brain Chemistry Labs thanks Brian and Wetonnah McCoy for their generous support of our exosome research. The Brain Chemistry Labs have submitted a patent application for the miRNA fingerprint described in Banack et al., 2020 . hsa-miR-126-5p YP00206010 339306 Table 1 : miRCURY LNA miRNA PCR Assays used for the analysis of quality control (QC) and low-to medium-abundance miRNA. Figure 1 : Quality control assays were run to optimize the volume of RNA for use in the cDNA reaction. UniSp2, 4, and 5 are RNA extraction spike-in controls (added to the lysis buffer before RNA extraction and to check for RNA extraction efficiency), and UniSp6 and cel-miR-39-3p are cDNA synthesis controls (added to the reaction mixture before cDNA synthesis and used to check the efficiency and reproducibility of the cDNA synthesis reaction). As expected, we report no difference in the mean raw Cqs for the cDNA synthesis spike-ins (UniSp6 and cel-miR-39-3p), but a Cq difference of approximately one was observed for the RNA extraction Cqs. The reason for this difference is because twice as much UniSp2, 4, and 5 templates were added to the cDNA reaction (contained in 2 µL RNA or 4 µL RNA), whereas the same amount of template was added to the cDNA reaction for either 2 µL or 4 µL RNA. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Figure 2 : Quality control assays were run to optimize the volume of RNA to use in the cDNA reaction. hsa-miR-142-3p, hsa-miR-451a, hsa-miR-23a-3p, hsa-miR-30c-5p, hsa-miR-103a-3p and hsa-miR-191-5p are used to measure each sample miRNA signal. We report raw Cqs greater than 1 (which corresponds to a doubling of template in a 100% efficient reaction) between cDNA synthesised with 2 µL RNA versus 4 µL RNA. Since we planned to assay some reported low-abundance miRNA, we selected 4 µL RNA for all downstream cDNA reactions. cDNA was synthesized using the miRCURY LNA RT Kit (Qiagen #339340), and qPCR of the sample signal miRNAs was run using miRCURY LNA miRNA SYBR individual PCR Assays according to the conditions described in materials and methods. These miRNAs were chosen to allow for comparison to previous experiments (Banack et al., 2020) . cDNA1 and cDNA2 were synthesized using 2 µL RNA, and cDNA3 and cDNA4 were synthesized using 4 µL RNA. Each raw Cq data point represents the miRNA Figure 5: EVs were purified as described previously (see material and methods), and RNA was extracted from 50 µL of the T-N fraction and used for column comparison. Eight miRNAs with medium to low abundance were analyzedd. Following cDNA synthesis, target miRNA Cqs were measured using individual miRCURY LNA SYBR green assays and a single qPCR analysis. Mean raw Cqs are reported. No-template controls were run concurrently for each target miRNA and returned a mean raw Cq of zero (data not shown). Each mean Cq represents three separate extracellular vesicle preparations and RNA extractions from three different plasma samples (n = 3). We report no significant differences between the mean Cqs for any miRNA, suggesting the Qiagen and Enzymax columns were equally efficient at extracting low-to medium-abundant miRNA. Unpaired non-parametric Mann-Whitney tests were conducted using GraphPad Prism 9.11 where p < 0.05 was set as statistically significant. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Microglial Extracellular Vesicles as Vehicles for Neurodegeneration Spreading A Comprehensive Review on Exosomes and Microvesicles as Epigenetic Factors An miRNA fingerprint using neural-enriched extracellular vesicles from blood plasma: Towards a biomarker for amyotrophic lateral sclerosis/motor neuron disease Role of exosomes in malignant glioma: MicroRNAs and proteins in pathogenesis and diagnosis Plasma-Derived Exosomal microRNA-130a Serves as a Noninvasive Biomarker for Diagnosis and Prognosis of Oral Squamous Cell Carcinoma MicroRNA biogenesis: Coordinated cropping and dicing Exosomes in