The U87 human glioblastoma cell line was purchased from the American Type Culture Collection (Manassas, VA, USA) and cultured in Dulbecco's modified Eagle's medium (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% (v/v) foetal bovine serum (Invitrogen; Thermo Fisher Scientific, Inc.), 100 U/ml of penicillin and 100 µg/ml of streptomycin (both from Invitrogen; Thermo Fisher Scientific, Inc.). Cultures were maintained in a humidified atmosphere of 5% CO₂ at 37°C. Total RNA was isolated from 2×10⁶ U87 cells using 1 ml TRIzol Reagent (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. RNA was quantified and evaluated for purity with a spectrophotometer based on A₂₆₀ and A₂₈₀ values, followed by visualization on a 1.0% (w/v) agarose gel stained with ethidium bromide.

Stem-loop RT and RT-qPCR primers were synthesized by Genewindows Biotech. Co. Ltd. (Guangzhou, China) and are listed in Table I. cDNA was generated by reverse transcription using 1 µg RNA as template with ReverTra Ace-α-Transcriptase (Toyobo Life Science, Osaka, Japan). For amplification synthetic DNA, ~0.1 ng of each let-7 member were mixed and used as template. Briefly, stem-loop and U6 RT primers were combined to transcribe the total RNA. cDNA was prepared in a series of dilutions and thus a concentration gradient was generated prior to RT-qPCR assay. RT-qPCR was performed using the SYBR Premix Ex Taq II kit (Tli RNaseH Plus; Takara Bio, Inc., Otsu, Japan), following the manufacturer's protocol. qPCR was performed with a LightCycler 480 Real-Time PCR system (Roche Applied Science, Rotkreuz, Switzerland) and quantitation cycle (Cq) values were normalized to U6 small nuclear RNA (16), which was used as an internal control. The amplification profile was: 1 cycle of denaturation at 95°C for 30 sec; followed by 45 cycles of denaturation at 95°C for 5 sec, annealing and extension at 60°C for 1 min; and fluorescence intensity was measured at 75°C. For melting curve analysis, the following parameters were used: Denaturation at 95°C for 5 sec, annealing at a rate of −2.8°C/sec until 60°C, fluorescence intensity were measured throughout the process. All experiments were performed in triplicate. Analysis was performed using Microsoft Excel 2013 (Microsoft Corporation, Redmond, WA, USA) and Origin 8.0 (OriginLab, Wellesley, MA, USA) software.

Poly(A) tails were added to total RNA 3′-ends using the miRNA First-Strand cDNA Synthesis kit (ComWin Biotech Co., Ltd., Beijing, China). Subsequently, RNA was reverse transcribed with anchor primer consisting of oligo(dT) for binding poly(A), a specific sequence for the combination with the reverse primer used in subsequent RT-qPCR, and a binucleotide VN (V indicates any nucleotide except for dT whereas N refers to any nucleotide). qPCR forward primers and thermocycling profiles were similar to the aforementioned stem-loop RT-qPCR.

Cq values and melting curves were obtained using the LightCycler 480 software V1.5, and Origin 8.0 software (OriginLab) was exploited to draw standard curves. The Cq values were plotted against the log₂ of total RNA input, and the amplification efficiency was evaluated through correlation analysis between these two values using Origin 8.0 software. Amplification efficiency (E) was calculated as: E=2^−1/a-1; where a is the slope of the standard curve line. The Clustal algorithm (Bio-Edit software V7.0.5) was applied in multiple sequence alignment, whereas primer designation schemes were produced using Adobe Illustrator CS4 software (Adobe Systems Incorporated, San Jose, CA, USA).

All data were presented as the mean ± standard deviation. Significant differences between samples were calculated by Student's t-test. All tests were performed at least in triplicate. All statistical analyses were carried out using the Origin 8.0 (Additive GmbH, Friedrichsdorf, Germany) and Excel software (Microsoft Corporation, Redmond, WA, USA). P<0.05 was considered to indicate a statistically significant difference.

Total RNA was extracted from U87 cells and reverse transcribed with a mixture of distinct stem-loop RT primers specifically designed for let-7 family members and U6 primers. Melting curves that are representative of amplification specificity combined with amplification efficiency of let-7 and U6 were also measured. Melting curve analysis of the amplified products demonstrated a single, sharp peak, which indicated good specificity (Fig. 2). Except for let-7i, which had a low concentration and therefore cannot be detected at a high dilution ratio, other amplification efficiencies ranged between 0.9288 and 1.1403. In addition, good linear relationships were indicated between PCR Cq values and log₂ of RNA input (R²>0.9; Fig. 3). These data suggested that by performing only one miRNA qPCR reaction, it was feasible to simultaneously detect and quantify all nine let-7 family members, and to further discriminate between let-7a, let-7b and let-7c, as well as between let-7i and let-7g. Notably, the mixture containing our modified RT primers and U6 primer was still able to efficiently and specifically quantify each of the let-7 congeners and U6 internal control, respectively, contributing to a considerable decrease in time and cost.

The expression profiles of let-7 family members in U87 cells were also assessed by the poly(A) tailing method (Fig. 4). This method was initially established to detect miRNAs for large-scale and high-throughput screening, so the forward primers in the stem-loop method and the poly(A) tailing method designed in the present study were the same. There is only a one base difference between mature let-7 miRNA family members, therefore, if the stem-loop section of primers used to reverse transcribe in different let-7 members were identical, the melting temperature values of these distinct let-7 members were alike (Fig. 2). However, in the poly(A) method, cross-reactions between let-7a, let-7b and let-7c or between let-7g and let-7i cannot result in a difference of >5°C in the melting temperature. Therefore, multiple peaks that occurred in the melting curves in the poly(A) tailing method were caused by other reasons.

However, the melting curves of each amplification reaction that were generated using the poly(A) method contained multiple peaks representing different products (Fig. 4), regardless of the specific forward primers of let-7d and let-7f. These may because miRNA precursors were also amplified, since the linear primers used in poly (A) tailing method lack specificity.

The ability of the modified stem-loop RT-qPCR assay to discriminate between miRNAs that differ by as little as a single nucleotide was examined using synthetic templates that contained a certain amount of each let-7 member and the differential Cq values were analysed. Different concentrations of synthetic let-7a, let-7b and let-7c were mixed and reverse transcribed with a mixture of corresponding primers and RT-qPCR was performed with let-7a primers alone. The results demonstrated that the addition of equal concentrations of let-7b or let-7c did not affect the detection of let-7a (Fig. 5). qPCR reactions devoid of let-7a had significantly higher Cq values compared with those that also contained synthetic let-7a. Although 5-fold excess concentration of let-7b or let-7c slightly interfered with let-7a detection, this interference was not statistically significant.