Technical factors involved in the measurement of circulating microRNA biomarkers for the detection of colorectal neoplasia

Abstract

Background: Circulating miRNAs are emerging as promising blood-based biomarkers for colorectal and other human cancers; however, technical factors that confound the development of these assays remain poorly understood and present a clinical challenge. The aim of this study was to systematically evaluate the effects of factors that may interfere with the accurate measurement of circulating miRNAs for clinical purposes.

Methods: Blood samples from 53 subjects, including routinely drawn serum samples, matched plasma from 30 subjects, and matched serum samples drawn before and after bowel preparation for colonoscopy from 29 subjects were collected. Additionally, 38 serum specimens stored in the clinical laboratory for seven days were used to test the stability of miRNAs. Hemolysis controls with serial dilutions of hemoglobin were prepared. RNA was extracted from serum, plasma or hemolyzed controls with spiked-in cel-miR-39, and levels of miR-21, miR-29a, miR-125b and miR-16 were examined by real-time RT-PCR. Hemolysis was measured by spectrophotometry.

Results: The expression levels of miR-16 and the degree of hemolysis were significantly higher in plasma than in serum (P<0.0001). Measured miR-21, miR-29a, miR-125b and miR-16 expression increased with hemoglobin levels in hemolyzed controls. The degree of hemolysis in serum samples correlated significantly with the levels of miR-21 (P<0.0001), miR-29a (P = 0.0002), miR-125b (P<0.0001) and miR-16 (P<0.0001). All four miRNAs showed significantly lower levels in sera that had been stored at 4°C for seven days (P<0.0001). Levels of miR-21 (P<0.0001), miR-29a (P<0.0001) and miR-16 (P = 0.0003), and the degree of hemolysis (P = 0.0002) were significantly higher in sera drawn after vs. before bowel preparation.

Conclusions: The measured levels of miRNAs in serum and plasma from same patients varied in the presence of hemolysis, and since hemolysis and other factors affected miRNA expression, it is important to consider these confounders while developing miRNA-based diagnostic assays.

Figure 1. Study subjects and blood samples used in this study.

The chart provides the description of patients and healthy volunteers from whom blood samples were collected under various conditions during the course of this study.

Figure 2. Measured miRNAs in matched serum and plasma.

Expression levels of miR-21, miR-29a, miR-125b and miR-16 (A) and the degree of hemolysis (B) are illustrated. Absorbance of each serum sample at 560 nm, 576 nm and 592 nm was measured by spectrophotometry and the degree of hemolysis was calculated by the following formula: estimated hemoglobin level = 2*OD^{576 nm}−OD^{560 nm}−OD^{592 nm}. Differences between serum and plasma were analyzed by the Wilcoxon signed-rank test.

Figure 3. MiR-21, miR-29a, miR-125b and miR-16 in serial dilutions of hemolyzed control samples.

A: Image of serial dilution of hemolyzed control samples. B: Correlation between dilutions of hemolyzed control samples and the degree of hemolysis is shown (Pearson’s correlation coefficient = R²). C: Levels of miR-21, miR-29a, miR-125b and miR-16 were elevated from baseline in hemolyzed control with 1/64^th dilution, and their levels increased along with the presumed hemoglobin concentration.

Figure 4. The degree of hemolysis in human serum samples.

A: The relationship between the visual hemolysis score and the degree of hemolysis which was determined by spectrophotometry. The hemolysis score of each serum sample was visually designated as follows: 0, no sign of hemolysis; 1, slight hemolysis cannot be ruled out because of dark yellow discoloring; 2, hemolysis is strongly suspected by orange to pink discoloring; 3, evident hemolysis with dark pink to red discoloring. B: Correlation between the degree of hemolysis and serum levels of miR-21, miR-29a, miR-125b and miR-16. The Spearman’s rank correlation coefficient (ρ) is shown. C: Correlation between the degree of hemolysis and serum levels of miR-21, miR-29a, miR-125b and miR-16 in visually non-hemolysed sera (hemolysis score of 0 and 1). The Spearman’s rank correlation coefficient (ρ) is presented.

Figure 5. Comparison between freshly prepared and stored serum samples.

A–D: Expression levels of miR-21 (A), miR-29a (B), miR-125b (C) and miR-16 (D) in freshly prepared and stored serum samples. E–H: Levels of four miRNAs were compared between freshly prepared sera from patients without known colorectal lesions and in stored sera. I: Ct values of spiked-in cel-miR-39 in freshly prepared and stored serum samples are shown. Differences were analyzed by the Mann-Whitney U test.

Figure 6. Comparison between serum samples drawn before and after a colonoscopy prep.

Expression levels of miR-21, miR-29a, miR-125b and miR-16 (A) and the degree of hemolysis (B) are shown. Differences between serum samples drawn before (pre-prep) and after (post-prep) colonoscopy prep were analyzed by the Wilcoxon signed-rank test.