Highly sensitive detection of DNA hypermethylation in melanoma cancer cells

Abstract

Aberrant hypermethylation of CpG islands in the promoter region of tumor suppressor genes is a promising biomarker for early cancer detection. This methylation status is reflected in the methylation pattern of ctDNA shed from the primary tumor; however, to realize the full clinical utility of ctDNA methylation detection via liquid biopsy for early cancer diagnosis, improvements in the sensitivity and multiplexability of existing technologies must be improved. Additionally, the assay must be cheap and easy to perform in a clinical setting. We report the integration of methylation specific PCR (MSP) to melt curve analysis on giant magnetoresistive (GMR) biosensors to greatly enhance the sensitivity of our DNA hybridization assay for methylation detection. Our GMR sensor is functionalized with synthetic DNA probes that target methylated or unmethylated CpG sites in the MSP amplicon, and measures the difference in melting temperature (T_m) between the two probes (ΔT_m), giving an analytical limit of detection down to 0.1% methylated DNA in solution. Additionally, linear regression of ΔT_m's for serial dilutions of methylated:unmethylated mixtures allows for quantification of methylation percentage, which could have diagnostic and prognostic utility. Lastly, we performed multiplexed MSP on two different genes, and show the ability of our GMR assay to resolve this mixture, despite their amplicons' overlapping T_m's in standard EvaGreen melt analysis. The multiplexing ability of our assay and its enhanced sensitivity, without necessitating deep sequencing, represent important steps toward realizing an assay for the detection of methylated ctDNA in plasma for early cancer detection in a clinical setting.

Keywords: Array; Cancer; DNA melting; GMR sensors; Methylation specific PCR; PCR.

Figure 1:

(a) GMR sensors manufactured by MagArray, with enhanced image of the 80-sensor array. (b) GMR chip mounted on custom-made temperature controlled cartridge for melt analysis. (c) Schematic of melt analysis on GMR sensors: ramping the temperature causes DNA denaturation and attenuation of GMR signal due to loss of bound MNPs.

Figure 2:

EvaGreen melt analysis of single-plex RARB methylation specific PCR for EST164/EST094 dilution series at annealing temperatures T_a = 59 °C (a), 62 °C (b), and 64 °C (c). The peak at T_m = 81 °C corresponds to ^MRARB, the peak at T_m = 77.5 °C corresponds to ^uMRARB.

Figure 3:

DNA melting curves measured on GMR sensors. Target DNA hybridized to surface tethered probes was denatured with a temperature ramp from 20 to 85 °C. Curves measured for (a) 100% ^mRARB (b) 1% ^mRARB and (c) 0% ^mRARB. ΔT_m is defined as the difference between melting curve of the DNA hybridized to the M and uM probes. ΔT_m varies with the content of methylated target. Melt curves for 10% and 0.1% methylated RARB are not pictured, values of ΔT_m can be found in Figure 4.The colored regions represent one standard deviation interval from the mean for n = 6 sensors.

Figure 4:

Plot of ΔT_m versus semi-logarithmic percentage of initial methylated DNA prior to MSP for annealing temperatures Ta = 59 °C, 62 °C, and 64 °C. Error bars are given as standard error of the mean for n = 6 sensors. Linear regression of these results is shown for T_a = 64 °C.

Figure 5:

(a) GMR binding curves for multiplexed RARB/KIT methylation specific PCR on 50 ng EST045 (^MRARB/^MKIT) at T_a = 59 °C. Synthetic DNA probes corresponding to methylated and unmethylated RARB/KIT were spotted on one GMR chip. (b) Normalized GMR signal from the binding curves of 5a was measured while ramping temperature from 20 °C to 85 °C. Inset: EvaGreen melt analysis of the multiplexed RARB/KIT methylation specific PCR gives one peak at T_m = 81 °C. The colored regions represent one standard deviation interval from the mean for n = 6 sensors.