Competitive amplification of differentially melting amplicons (CADMA) improves KRAS hotspot mutation testing in colorectal cancer

Abstract

Background: Cancer is an extremely heterogeneous group of diseases traditionally categorized according to tissue of origin. However, even among patients with the same cancer subtype the cellular alterations at the molecular level are often very different. Several new therapies targeting specific molecular changes found in individual patients have initiated the era of personalized therapy and significantly improved patient care. In metastatic colorectal cancer (mCRC) a selected group of patients with wild-type KRAS respond to antibodies against the epidermal growth factor receptor (EGFR). Testing for KRAS mutations is now required prior to anti-EGFR treatment, however, less sensitive methods based on conventional PCR regularly fail to detect KRAS mutations in clinical samples.

Methods: We have developed sensitive and specific assays for detection of the seven most common KRAS mutations based on a novel methodology named Competitive Amplification of Differentially Melting Amplicons (CADMA). The clinical applicability of these assays was assessed by analyzing 100 colorectal cancer samples, for which KRAS mutation status has been evaluated by the commercially available TheraScreen® KRAS mutation kit.

Results: The CADMA assays were sensitive to at least 0.5% mutant alleles in a wild-type background when using 50 nanograms of DNA in the reactions. Consensus between CADMA and the TheraScreen kit was observed in 96% of the colorectal cancer samples. In cases where disagreement was observed the CADMA result could be confirmed by a previously published assay based on TaqMan probes and by fast COLD-PCR followed by Sanger sequencing.

Conclusions: The high analytical sensitivity and specificity of CADMA may increase diagnostic sensitivity and specificity of KRAS mutation testing in mCRC patients.

Figure 1

The analytical sensitivity and specificity of the CADMA assays performed using the Rotorgene 6000. Ten wild-type replicates were run together with a standard dilution series of mutant alleles from cell lines carrying the relevant mutations in a wild-type background (50%, 10%, 1%, and 0.5%) in triplicates. The three replicates of the 0.5% standard could all be distinguished from ten wild-type replicates in all assays. A. The c.34 G > A CADMA assay. B. The c.38 G > A CADMA assay. C. The c.35 G > A CADMA assay. D. The c.34 G > C CADMA assay. E. The c.35 G > T CADMA assay. F. The c.34 G > T CADMA assay.

Figure 2

Examples from screening of mCRC samples using the c.34 G > T CADMA assay. A. Real-time amplification data. High background fluorescence can be observed for sample ID 69. B. The derivative of the raw melting data (melt curve analysis). Sample ID 56 carry the c.34 G > T mutation. Sample ID 2, 69 and 72 were negative for the c.34 G > T mutation. Sample ID 69 and 72 may carry another KRAS mutation as small deviations from the wild-type replicates can be observed. Sample ID 69, which gave high fluorescence during the PCR amplification, has deviating melt curves from 82 to 88°C. C. Normalized HRM difference graph. Of the mCRC samples shown only one (sample ID 56) deviates more from the wild-type replicates than the standard containing 1% mutant alleles.

Figure 3

Examples from screening of mCRC samples using the c.38 G > T CADMA assay. The wild-type mCRC samples showed more variation in the c.38 G > A CADMA assay compared to any of the other CADMA assays. A. The derivative of the raw melting data (melt curve analysis). No heteroduplexes, which melt between 71 and 74°C, can be observed in any of the wild-type mCRC samples. B. Normalized HRM difference graph. Sample ID 2 deviates more from the wild-type replicates than the standard containing 1% mutant alleles from approximately 79 to 81°C. This should not be interpreted as a c.38 G > T mutation, since no heteroduplexes are present.