br Development and validation of sensitive miRNA quantificat
3.3. Development and validation of sensitive miRNA quantification assay combining DNA-Peptide dendrimer probe and LC-MS/MS
To quantify target miRNA, all the miRNA samples were bio-tinylated in advance and then immobilized on the streptavidin agarose beads. In our previous study, the maximum loading ca-pacity of streptavidin agarose beads for biotinylated miRNAs was ~2.1 biotin molecules/streptavidin , so almost all the miRNAs in samples can be immobilized using excessive agarose beads.
Under the optimized conditions for DPD probe and LC-MS/MS, the signal intensity of miR-21 was ~8 times greater compared to that without signal amplification . Then, standard curves were constructed from the relative peak area ratio between the reporter peptide and the stable isotope-labeled internal standard across the concentration (1/x2 weighting). From Fig. 5A, it can be seen that the area ratio is sensitive to the concentration of miR-21, the fitting
range is from 0.2 pM to 10 nM, and the LOQ was 0.2 pM (Fig. S9). In addition, QC samples prepared in tRNA library yeast were used to evaluate the precision and the accuracy of the assay. Six replicates of QC samples prepared at four different miR-21 con-centration levels were evaluated in three validation runs. The re-sults are listed in Table 1S. Both accuracy and precision were
±15% (LLOQ, ±20%). In particular, the %RSD and %bias measured on the LLOQ was much less than 5% and generally better than 3%, indicating the good accuracy and precision at low concentrations . This feature may bring the benefits of early diagnosis and increase the diagnostic and theranostic accuracy.
3.4. Quantification of miR-21 in breast Mitomycin C and tissue samples
Fig. 5. (A) Representative calibration curve (0.2 pMe10 nM) for the miR-21 standards. Relative peak area ratio of the reporter peptide and the stable isotope-labeled internal standard was plotted against concentration. (B) Quantification of miR-21 in 102 matched pairs of breast tissue samples using the DPD probe combined with LC-MS/MS.
Fig. 6. (A) Passing-Bablok regression analysis and (B) the corresponding Bland-Altman plot for qRT-PCR vs our assay termed as DPD-MS. The solid line corresponds to the regression line. Dashed lines represent the 95% confidence interval for the regression line.
difference was statistically significant (p < 0.05). For comparison, qRT-PCR was conducted on the same samples (Fig. S10). The ob-tained level of miR-21 was slightly higher than that determined by our assay, but the difference was not significant. Such bias might be attributed to the nonspecific amplification existing in qRT-PCR experiment as described earlier.
copies/mg) in normal tissues and
in tumors. A Mann-Whitney two-way test showed that tumor samples have a significantly higher level of miR-21 than normal samples (p < 0.0001). The average increase was about 3.5 fold. Furthermore, qRT-PCR was also performed on these samples and Passing-Bablok regression analysis was conducted for method comparison using MedCalc® software. As shown in Fig. 6, our assay was correlated well with the qRT-PCR assay (y ¼ 1.03 x - 0.54, p ¼ 0.54). Bland-Altman analysis demonstrated that there was a concentration-dependent bias, particularly at value below 109 copies/mg (Fig. 6). This observation was probably due to false-positives caused by nonspecific amplification, leading to the higher value of miR-21 determined by qRT-PCR, which could be more pronounced at low concentrations .
Furthermore, miR-21 expression level in tumors was evaluated according to the histopathological features and molecular subtypes of breast cancer. Correlations of miR-21 expression with tumor size, tumor grade, lymph node status and molecular subtypes of breast cancer are summarized in Table 1. As shown, high level of miR-21 appears to be associated with advanced clinical stage, distant me-tastases and large tumor size of breast cancer, which was in accordance with previous reports . However, no correlation was observed with hormone receptor status. In addition, the patients were separated into two groups according to the median level of miR-21, miR-21 low expressers (n ¼ 51) and miR-21 high expressers (n ¼ 51). The 5-year survival rate in patients with low miR-21 expression was 82.4%, which was significantly higher than those with high miR-21 expression (45.1%; p < 0.001, log-rank test). Kaplan-Meier survival analyses revealed that the breast cancer patients with high miR-21 expression had a significantly poorer prognosis compared to those with low miR-21 expression (Fig. 7). These observations suggest that miR-21 is dynamically regulated by certain molecular and cellular activities within the tumor.