Mol Biol Rep. 2016 Oct; 43(10):1165-78.
Identification of extracellular miRNA in archived serum samples by next-generation sequencing
Aarti Gautam1, Raina Kumar2, George Dimitrov2, Allison Hoke3, Rasha Hammamieh1 and Marti Jett4
1US Army Center for Environmental Health Research, 568 Doughten Drive, Fort Detrick, 21702-5010, MD, USA.
2Advanced Biomedical Computing Center, Frederick National Laboratory for Cancer Research/Leidos-Biomedical Inc., Frederick, MD, 21702, USA.
3The Geneva Foundation, US Army Center for Environmental Health Research, Fort Detrick, MD, 21702, USA.
4US Army Center for Environmental Health Research, 568 Doughten Drive, Fort Detrick, 21702-5010, MD, USA. email@example.com.
miRNAs act as important regulators of gene expression by promoting mRNA degradation or by attenuating protein translation. Since miRNAs are stably expressed in bodily fluids, there is growing interest in profiling these miRNAs, as it is minimally invasive and cost-effective as a diagnostic matrix. A technical hurdle in studying miRNA dynamics is the ability to reliably extract miRNA as small sample volumes and low RNA abundance create challenges for extraction and downstream applications. The purpose of this study was to develop a pipeline for the recovery of miRNA using small volumes of archived serum samples. The RNA was extracted employing several widely utilized RNA isolation kits/methods with and without addition of a carrier. The small RNA library preparation was carried out using Illumina TruSeq small RNA kit and sequencing was carried out using Illumina platform. A fraction of five microliters of total RNA was used for library preparation as quantification is below the detection limit. We were able to profile miRNA levels in serum from all the methods tested. We found out that addition of nucleic-acid based carrier molecules had higher numbers of processed reads but did not enhance the mapping of any miRBase annotated sequences. However, some of the extraction procedures offer certain advantages: RNA extracted by TRIzol seemed to align to the miRBase best; extractions using TRIzol with carrier yielded higher miRNA-to-small RNA ratios. Nuclease-free glycogen can be the carrier of choice for miRNA sequencing. Our findings illustrate that miRNA extraction and quantification is influenced by the choice of methodologies. Addition of nucleic acid-based carrier molecules during extraction procedure is not a good choice when assaying miRNA using sequencing. The careful selection of an extraction method permits the archived serum samples to become valuable resources for high-throughput applications.
- PMID: 27510798; DOI:10.1007/s11033-016-4043-6
The Department of Defense Serum Repository (DoDSR) contains serum samples that are available for research with the appropriate approvals. These serum samples are associated with the Defense Medical Surveillance System (DMSS), which provides information about the military and medical experiences and also longitudinal data of the service members. Serum separated from blood contains microRNAs (miRNAs) which are small, non-protein coding RNAs. miRNAs are known to be well conserved across species and are important in gene regulation. miRNAs circulating in the body, such as those found in serum, are thought to influence the circulatory and immune systems. Serum miRNAs could serve as a source of novel biomarkers for understanding diseases and are ideal for studying regulation of targeted gene expression. miRNAs remain stable for years, even in severe conditions such as freeze-thaw cycles, suboptimal storage conditions, and changes in pH levels. Experiments on serum miRNAs, however, are hindered by low concentrations and poor quantification.
In our pilot study, we used some test samples from the DoDSR that were stored at sub-optimal temperatures after processing. Our aim was to determine the feasibility of using a larger sample set from this repository. The purpose of this study was to test different extraction methods to isolate RNA that can be processed down-stream for next-generation sequencing technology (NGS). The sample preparation for sequencing used an Illumina procedure for ligating bait sequences known as DNA adaptors at both ends of small RNAs, followed by conversion to cDNA sequences using reverse transcription, and then standard PCR amplification. Next, the small RNA libraries were size selected and quantified using standard procedures. These were then sequenced on the Illumina platform using the polymerase-based sequencing by synthesis method.
To reduce the interpersonal differences among different samples, we first created a pooled sample by combing multiple samples into a homogenous pool for comparison to unpooled samples. Each procedure was performed in triplicate. Our efforts also focused on a method allowing us to isolate DNA as well as RNA simultaneously from the samples. The extraction methods included different commercially available procedures such as TRIzol Reagent (Invitrogen, Life Technologies, Grand Island, NY, USA), the mirVana miRNA Isolation Kit (Ambion, Life Technologies, Thermo Fisher Scientific, Waltham, MA, USA), the miRNeasy Mini Kit (Qiagen, Valencia, CA, USA), and the ExoQuick Exosome precipitation (System Biosciences, Inc., Mountain View, CA, USA). The utility of a few carriers such as glycogen (Invitrogen), bacteriophage MS2 (Roche, Basel, Switzerland), and yeast tRNA (Invitrogen) in NGS was also investigated. Nucleic acid-based carriers have a drawback as they can lead to inaccurate quantification of the extracted material and may also interfere with the NGS applications. TRIzol-chloroform extraction has been shown to be the most efficient overall for large-scale experiments, but it has also been reported to fail to isolate RNA that is low in guanine and cytosine. Many of the commercial extraction kits use filters that minimize the carry-over of inhibitors during the extraction process. The quantification of miRNAs in blood is problematic because of the high protein content and low RNA concentrations in plasma and serum. Instruments such as the Agilent 2100 Bioanalyzer have the capability to obtain information on miRNA concentrations by capillary electrophoresis using their small RNA kit. The small RNA quantification isn’t within the limit of detection using spectrophotometric devices, and our study found that the presence of extracellular carriers made this impossible.
Multiple reports have attempted to detect plasma/serum miRNA using microarrays and quantitative real time PCR, but we focused on using the NGS approach and the Illumina method for sample preparation. Our data showed that small RNA was preserved in the DoDSR archived samples, and while each procedure used was successful in extracting miRNA, some of the extraction procedures performed better than others.
The essential findings after a review of processed reads indicated that the addition of carrier to TRIzol extraction lead to inconsistency in the processed reads between pooled and unpooled samples, especially with the use of nucleic acid-based carriers such as bacteriophage MS2 or yeast tRNA. The use of glycogen produced comparable reads between the pooled and unpooled samples and requires further investigation. We observed that extraction of pooled and unpooled samples with TRIzol only yielded similar numbers of processed reads, but low ratios of miRNA to small RNA. The use of the carrier during TRIzol extraction did increase the miRNA to small RNA ratio, except when we used carrier yeast tRNA for the pooled samples. Exosome precipitation prior to RNA extraction yielded a lower number of processed reads and did not produce greater amounts of miRNA as compared to the TRIzol procedure alone. The miRNeasy method yielded similar numbers of processed reads from pooled and unpooled samples, and also seemed to yield a consistent amount of miRNA extracted across all corresponding samples.
We also looked into the read distribution plots from these samples and found that the addition of carrier resulted in a larger range for read lengths that could be attributed to RNA degradation, carrier degradation, or the presence of piwi-interacting RNA (piRNA), which is known to be longer in length than miRNA. On aligning these reads to miRBase, TRIzol extractions gave the maximum percentage of aligned reads, whereas the miRNeasy extractions produced the least amount of aligned reads. Though addition of carrier yielded the maximum numbers of processed reads, these had the lowest percentage of sequences aligned to miRBase. We further investigated the counts of RNA species obtained by mapping the reads to the human reference genome assembly version 19, and we observed that the vast majority of the reads from the unpooled and pooled samples were coding RNA species. We observed that all of the replicates among each procedure were positively correlated with each other for pooled samples, as well as for unpooled samples.. There was a weak correlation of miRNA counts for the extractions that used yeast tRNA as compared to the other procedures. We investigated the common miRNAs among different procedures, and in both datasets, TRIzol extraction with MS2 carrier yielded the maximum miRNA overlap than those isolated by the other procedures. Since the pooled samples were derived from multiple samples, we expected this observed increase in the number of common miRNAs.
One of the main challenges associated with NGS technology was the need to computationally sift through the millions of reads generated to identify both known and novel miRNAs. A further complication was the specific mapping of the short reads without compromising the quality of reads mapping to known miRNA mature reference sequences. To address these challenges, we customized known tools to fine-tune the mapping algorithm parameters and to trim the reads in a specific way to ensure filtered reads mapped to the reference sequences with higher accuracy. To identify the other small RNA species in this study, several computational steps were performed to exhaustively filter the read sequences that do not represent the true sequences, such as filtering all the known contaminants and sequencing platform artifacts, and filtering for different RNA species at each step.
We determined the efficiencies of different extraction methods, and we were able to successfully sequence miRNAs from the DoDSR, which can be a valuable resource for studying military relevant diseases.
The contents of this article have been tailored from following journal article.
Aarti Gautam, Raina Kumar, George Dimitrov, Allison Hoke, Rasha Hammamieh and Marti Jett. Identification of extracellular miRNA in archived serum samples by next-generation sequencing.Molecular Biology Reports 43, 1165-1178.
The views, opinions, and/or findings contained in this report are those of the authors and should not be construed as official Department of the Army position, policy, or decision, unless so designated by other official documentation. Citations of commercial organizations or trade names in this report do not constitute an official Department of the Army endorsement or approval of the products or services of these organizations.