|Home | About | Journals | Submit | Contact Us | Français|
The broad involvement of miRNAs in critical processes underlying development, tissue homoeostasis and disease has led to a surging interest among the research and pharmaceutical communities. To study miRNAs, it is essential that the quantification of microRNA levels is accurate and robust. By comparing wild-type to small RNA deficient mouse embryonic stem cells (mESC), we revealed a lack of accuracy and robustness in previous published multiplex qRT-PCR techniques. Here, we describe an optimized method, including purifying away excessive primers from previous multiplex steps before singleplex real time detection, which dramatically increases the accuracy and robustness of the technique. Furthermore, we explain how performing the technique on a microfluidic chip at nanoliter volumes significantly reduces reagent costs and permits time effective high throughput miRNA expression profiling.
(Following the manufacturer's protocol using TRIzol.)
Device: Microcon Ultracell YM-3, Cutoff SS and DS Nucleotides: 10
Volume: Desired volume of re-concentrated RNA solution is dependent on experimental design and should include control experiments.
Note: the final concentration of the stem-loop primers should be 1-5 nM (here 2 nM).
*List of reverse stem-loop primers can be found at http://urology.ucsf.edu/blellochlab/protocols.htm
Reaction Volume: 27.5μl (22μl MM + 5.5μl RT-Product)
Note: Pre-Amplification is necessary to enhance signal strength, especially when starting concentration of RNA is limited. The input RNA level determines the optimal number of Pre-PCR cycles. Given the same relative efficiency each PCR cycle should double the concentration of the final product. Since some miRNAs will be more highly expressed, avoid over cycling. However, under-amplification can result in a loss of detection sensitivity, especially for microRNAs expressed at low levels. In our hands 12 pre amplification cycles, using 100ng of total RNA, appeared sufficient. Using serum as starting material, 12 cycles results in detectable miRNA levels, but 15 cycles appeared to be superior. Finally, starting with 3 oocytes (roughly 400 pg total RNA per oocyte3), 16 pre-amplification cycles successfully allowed us to screen for microRNA expression levels4.
* List of forward primers can be found at http://urology.ucsf.edu/blellochlab/protocols.htm
Note: Desired volume of concentrated cDNA Solution is dependent on experimental design and should include control experiments.
Note: In our hands the exclusive enzymatic clean-up successfully removed primers but also led to partial degradation of the pre-amplified product, possibly due to partial denaturing of the short products as temperature rises up to 80 °C during the inactivation step of the enzyme. Avoiding heat inactivation and directly purifying the PCR products from the enzymatic reaction using spin columns removes any primer without product degradation.
Note: Chip needs to be used 24 hours after opening the package. Make sure not to spill the control line fluid since control line fluid on the chip or in the inlets makes the chip unusable. Chip needs to be loaded within 60 minutes of priming.
In a disposable plastic tube, combine the components in the table below to make the Sample-Pre-Mix (Volumes per inlet).
Note: Final Volume is 2.75 μl, as overage for pipette dispense losses
Note: Vortex and centrifuge 96 well plate briefly to collect fluid. Final Volume per inlet 5 μl, take pipetting loss into account by preparing slightly higher volume.
* List of forward primers can be found at http://urology.ucsf.edu/blellochlab/protocols.htm
# List of TaqMan Probes can be found at http://urology.ucsf.edu/blellochlab/protocols.htm
Note: In order to allow for correct loading of samples and assays ensure the correct positioning of the chip. It is convenient to align the notch to the top left hand corner.
Ensure all assay and sample solutions are mixed before loading the chip. It is convenient to do this in a 96 well plate and spin the plate down before loading the chip. Be aware of the chip pipetting map for later reference.
Unused sample inlets should be filled with 2.75 μl sample Mix and 2.25 μl DNA free water.
Unused assay inlets should be filled with 2.5 μl assay loading reagent and 2.5 μl water.
Do not go past the first stop while pipetting to avoid introducing air bubbles.
After RT- PCR the proprietary software provides amplification curves, heat maps and Ct values for each well. Please refer to the software manual for data analysis.
Figure 1. Diagrammatic representation of the experimental workflow. Shown is a step-by-step description of the multiplex qRT-PCR protocol, including the purification (Step E) of excessive primers from multiplex reverse transcription (Step C) and multiplex preamplification (Step D) before real time detection (Step F), resulting in a purified cDNA template for qRT-PCR. FP: microRNA specific forward primer, URP: universal reverse primer, R-SLP: microRNA specific reverse stem-loop primer. Click here for a larger image.
Figure 2. Polyacryamide gel purification of the qRT-PCR template after pre-PCR. Bands of pre-PCR product size are exclusively seen in WT samples, but not in microRNA deficient DGCR8-/- or Dicer-/- samples (arrows) after 96-plex RT and 12 cycles of preamplification. Note that non-specific products of unknown origin (arrowheads) and excessive primers are found in all samples (stars).
Figure 3. Purification after multiplex preamplification vastly improves accuracy of qRT-PCR results. a) Comparison of relative expression levels of WT mESC to Dgcr8-/- (canonical microRNA deficient) mESC reveals 1) the detection of more microRNAs and 2) the loss of false positive signals in Dgcr8-/- backgrounds after purifying away primers of the pre-PCR product. b) After purification, relative expression levels of both DGCR8 -/-/WT and Dicer -/-/WT show a loss of false positive signals and allow for proper categorization of rare Dgcr8 independent /Dicer dependent small RNAs (miR-320, -484, -877).5
Figure 4. Fluidigm high-throughput qRT-PCR miRNA profiling. Example screen shot of 96.96 Fluidigm qRT-PCR mircoRNA profiling to screen for alteration in miRNA levels of prostate cancer patient sera. The real-time qPCR analysis software provides amplification curves, color-coded heat maps and cycle threshold (Ct).
miRNAs are short (18-24 nucleotides), non-coding RNAs, which regulate gene expression post-transcriptionally by both destabilizing messenger RNAs (mRNA) and inhibiting their translation.6 The fact that at least one-third of human genes contain conserved miRNA binding-sites in their 3'UTR and the evident interactions of miRNAs with genes for pluripotency, proliferation and apoptosis suggests a critical contribution in cell fate decisions, tissue homeostasis and diseases such as cancer.6,7 Therefore accurate microRNA expression profiles are of broad interest.
To test the published multiplex qRT-PCR miRNA expression profiling protocol2 we used small RNA deficient mESC as negative controls. DGCR8-/- cells are deficient of all canonical microRNAs, while Dicer-/- cells lack both canonical and non-canoncial micoRNAs.8,9
Comparing levels in wild type mES- to DGCR8-/- cells, we uncovered a lack of accuracy as several expressed10 microRNAs were not detected in the wild-type cells11. Some even showed lower expression levels relative to the knockout cells (Figure 3a). We hypothesized that the lack of accuracy might be caused by the massive primer carry over of the two consecutive multiplex steps (RT and Pre-PCR) before singleplex RT-quantification. However, multiplex pre-amplification is necessary to enhance signal strength, especially when starting concentration of RNA is limited. By purifying away pre-PCR products from primers by size selection on native polyacrylamide gels (Figure 2), we could demonstrate a substantial improvement in accuracy by detecting more microRNAs and a loss of false positive signals in both Dgcr8-/- and Dicer-/- backgrounds (Figures 3a and 3b) In addition the modified qRT-PCR approach allowed for the proper categorization of rare Dgcr8 independent / Dicer dependent small RNAs (Figure 3b)11. Therefore an additional step to purify excessive primers away from pre-PCR product is clearly advantageous, especially when using large multiplex primer sets per sample.
The optimal number of pre-amplification cycles is determined by the input RNA concentrations. The balance to not under- or overamplify should be guided by particulars of the specific experiment.
With more than a thousand known microRNAs, use of standard 384-well plates may not be optimal for extensive microRNA profiling, especially when comparing multiple samples. The Fluidigm Dynamic Array IFC enables, with a significant decrease of pipetting steps and needed chemistry, testing up to 96 individual samples against 96 different microRNAs in a single experiment (9216 reactions) at nanoliter scale (6.7 nL). For each run the analysis software provides amplification curves, color-coded heat maps and cycle threshold values (Ct). The simultaneous large-scale profiling reduces experimental variances and allows for mean expression value normalization, which out-performs other normalization strategies that make use of small RNA controls.12
Compared to commercially available 384 well platforms with pre-assigned TaqMan probes, the combination of a high-throughput profiling platform with custom-made primer sets offers high experimental flexibility.
The optimized multiplex qRT-PCR approach in combination with the Dynamic Array platform successfully allowed us to simultaneously screen 48 prostate cancer patient sera for alterations in levels of 384 miRNA (Figure 4) and to normalize data without spike in controls (i.e. synthetic microRNAs).11
Despite the increase of steps from samples preparation to profiling results, the described approach is a time- and cost-effective high throughput method to profile large sample sets for miRNA expression levels, even with limited starting material.
Alan Mir is an employee of Fluidigm Corporation. Otherwise, we have no financial interests to disclose.
We would like to thank the Blelloch lab for commenting on the text. This work was supported by funds to RB from NIH (K08 NS48118 and R01 NS057221), California Institute of Regenerative Medicine (CIRM) (Seed Grant RS1-00161, New Faculty Award RN2-00906) and the Pew Charitable Trust and to F.M. from the Wissenschaftlich Urologische Gesellschaft eV.