Laboratory Methods for KRAS Mutation Detection
A variety of laboratory methods have been utilized to detect mutations in the KRAS gene.[21,22] Most methods include the use of PCR to amplify the appropriate region of the KRAS gene, including exons 2 and 3, and then utilize different methods to distinguish wild-type from mutant sequences in key codons, such as 12 and 13. The detection methods include nucleic acid sequencing, allele-specific PCR methods, single-strand conformational polymorphism analysis, melt–curve analysis, probe hybridization and others. Some of the key performance features of the various methods are summarized in Table 4. The main features for consideration for all of these molecular techniques are the ability to distinguish the appropriate spectrum of variants at the codons of interest and the sensitivity or limit of detection (LOD) for mutant alleles. Both of these parameters are important, given the fact that tumors may be very heterogeneous, both with regard to the percentage of tumor cells within a given tissue and the potential for genetic heterogeneity. Genetic heterogeneity may result from tumor cells being hom*ozygous, heterozygous or hemizygous for a KRAS mutation. The potential importance of genetic heterogeneity was shown in a study evaluating multiple areas from a given tumor, including the tumor center and invasion front. In this analysis, up to 20% of the samples showed heterogeneity for KRAS mutation status, with greater heterogeneity found in early-stage cancers.[37] A variety of methods have been developed for KRAS mutation analysis to address these specific issues, related to increased analytical sensitivity, and they include allele-specific PCR using amplification refractory mutation system (ARMS) technology or coamplification at a lower denaturation temperature-PCR methods, pyrosequencing approaches and real-time PCR methods that use specific probe technologies, such as peptide nucleic acids.[29,34,35,38]
The mutation frequency for the KRAS gene in metastatic colon cancers has been reported to be in the range of 30–55%.[1,3,39] This variance in mutation frequency can be attributed to both methodology differences (i.e., codons evaluated, spectrum of mutations identified, LOD and other performance features), as well as differences in the study cohorts from which the estimate of mutation frequency was derived. For positive samples, approximately 80% of the mutations occur in codon 12, with 35G>A (G12D) and 35G>T (G12V) being the most common variants; approximately 15–20% occur in codon 13 with 38G>A (G13D), the most common variant; and a small percentage of mutations (<5%) have also been identified in other codons, such as 61 and 146. These observations obtained from multiple studies were confirmed in a large analysis of over 1000 colon cancers, in which the mutation frequency determined by sequence analysis was 39.3%, with the common variants being those described previously.[23] In addition, this study was able to show that the mutation frequency and distribution of variants was not significantly different in primary versus metastatic lesions. This is an important observation and may have clinical relevance regarding which specimen is evaluated when considering therapeutic options for a given patient. In studies that have evaluated large numbers of clinical specimens submitted for KRAS mutation testing, similar mutation frequencies (40–42%) and allele distributions have been observed.[40,41] These data confirm observations from clinical study cohorts in real-world clinical samples, and support the robustness of technologies, such as allele-specific PCR methods for the routine analysis of clinical samples. In lung cancers, 10–22% of tumor samples contain a KRAS mutation, with the mutation frequency being dependent on a variety of features, including cancer type (adenocarcinoma vs squamous cell carcinoma), smoking history and presence of other genetic lesions, such as EGFR mutations and ALK gene fusions.[2] Approximately 97% of mutations are found in codons 12 and 13, with the remainder in codons 61 and 146.[2,24,25]
Three of the commonly used methods for KRAS mutation detection in clinical samples, include nucleic acid sequencing (dideoxy and pyrosequencing), real-time PCR with melt–curve analysis and allele-specific PCR with various modes used to distinguish mutant from wild-type sequences.[26–29,39] Dideoxy or Sanger sequencing technologies are often considered the gold standard for the molecular characterization of genetic variants because they provide the means to detect all potential variations, including base substitutions, insertions and deletions. The major limitations of dideoxy sequencing are the analytical LOD for mutations within a background of normal genetic sequences, which is often at the 10–20% level for a specific mutation, and variation in methods between laboratories. Pyrosequencing technologies are becoming more commonly used to detect variations in smaller amplicons of 50 nucleotides or less, with this technique providing increased analytical sensitivity, being able to detect less than 5% of a specific mutation in a background of normal sequences. This technology involves the release and detection of pyrophosphate moieties produced during the incorporation of a specific nucleotide into the synthesized DNA complement of the region of interest. In the process the pyrophosphate is converted into a detectable entity via a series of enzymatic reactions. The resulting pyrogram allows the detection of the specific nucleic acid sequence for the target region. Real-time PCR methods with melt–curve analysis to distinguish nucleotide variations in the amplified targets have been standard laboratory methods, with a variety of applications in genetics, infectious disease and oncology. The melt–curve technologies utilize fluorescent probes complementary to the target amplicon, which can be used to distinguish genetic variants by the differences in the melting temperature needed to dissociate probe from target. Differences in melting temperature are detected based on the loss of fluorescence as a function of increasing temperature. Allele-specific PCR methods are also common laboratory methods used to characterize simple genetic variants, such as point mutations. Allele-specific PCR methods use oligonucleotide primers that allow the specific amplification of mutant versus wild-type sequence through the differential binding and extension of the primer sequences to the target template. The amplification products may be detected by a variety of methods ranging from gel electrophoresis to real-time PCR detection. The PCR primer specificity provides a mechanism to enrich for mutation detection, and thereby provides enhanced sensitivities (LOD of 1–5% for mutant sequences) compared with other methods. Table 4 provides a summary of these and other methods used for analysis of KRAS mutation status. Owing to differences in performance features among the various assays, such as those described in Table 4, combinations of assays may provide the best means to assess clinical samples, particularly those that have low tumor content or from which DNA quality may not be optimal.
In addition to the several examples of laboratory-developed tests (LDTs) for KRAS mutations, there are also a series of kits developed by in vitro diagnostic (IVD) manufacturers for both research purposes and for use in clinical diagnostics. The TheraScreen® assay (DxS, Manchester, UK) is a CE-marked kit intended for the detection and qualitative assessment of seven somatic mutations in the KRAS gene, to aid clinicians in the identification of colorectal cancer patients who may benefit from anti-EGFR therapies, such as panitumumab and cetuximab. This assay uses two technologies, an amplification refractory mutation system (ARMS), which is a version of allele-specific PCR, and detection of amplification products with Scorpion™ probes.[30,33,35] Using ARMS technology, specific mutations can be amplified with great sensitivity in a heterogeneous sample based upon the principle that when an amplification primer is fully matched with the target sequence, the PCR process will proceed with very high efficiency. When there is not a specific match, only low or background levels of amplification will occur. Scorpion™ probes are bifunctional molecules having a PCR primer linked to a probe containing both a fluorophor and a quencher molecule. When these probes bind to the PCR amplicon the quencher and fluorophor become separated, resulting in a fluorescent signal. The fluorscent signals produced with each round of PCR are measured using a real-time PCR format and the appropriate instrumentation (i.e., Light Cycler® 480, ABI 7500). KRAS allele determinations are made based upon established cycle thresholds (Ct). The combination of these technologies provides very good analytical sensitivity with a LOD of 1–5% mutant sequence detectable in a background of wild-type sequences. There have been multiple studies confirming the performance of this technology in both colon and lung cancers.[25,33] The PyroMark® KRAS assay (Qiagen, Valencia, CA, USA) is a research-use assay that detects all variants at codons 12, 13 and 61 using a combination of PCR amplification with pyrosequencing technologies.[31] In this assay, two sets of amplification primers are used to amplify the exon 2 (containing codons 12 and 13) and exon 3 (containing codon 61) regions of the gene, with sequence analysis of the captured products achieved using primers that flank the areas of interest. The resulting pyrograms allow the ascertainment of the specific nucleotide sequences at codons 12, 13 and 61. Published studies with this particular kit show good analytical performance for formalin-fixed paraffin-embedded (FFPE) samples from colon cancer tissues.[35] The Signature® KRAS/BRAF assay (Asuragen, Inc., Austin, TX, USA) is a research use assay that detects 12 variants for codons 12 and 13 of the KRAS gene, and a single variant (V600E) within the BRAF gene. The technologies utilized in this assay include a multiplex PCR amplification of the respective regions of the two genes, probe hybridization and detection using the Luminex® bead-based platform. Bead-bound probes for the different target sequences allow for the delineation of the specific sequences for each gene. The performance of this assay has also been shown to provide adequate analytical results on FFPE samples.[32]
There have been several comparative studies evaluating the performance of different methods used to accurately characterize KRAS gene status in a variety of specimen types. One study evaluated the performance of a variety of methods, including dideoxy sequencing, single-strand conformational polymorphism analysis, pyrosequencing, melt–curve analysis and a real-time allele-specific PCR method using ARMS.[33] The performance of these testing methods was evaluated on a series of both frozen and FFPE samples. For this study, the DNA was extracted by a single laboratory, and then distributed to the other laboratory sites to perform mutation analysis using their specific methodologies. In general, there was good concordance between the majority of methods, with a 96% concordance among the five best-performing methods. Factors that contributed to discordance included the quality of the DNA from FFPE samples – which was problematic for some methods (i.e., one of the real-time PCR methods) – and low tumor content. In a similar study, a series of consecutive colon cancer FFPE specimens were evaluated for KRAS mutations by dideoxy sequencing, linear-probe array analysis, melt–curve analysis and pyrosequencing.[34] In the tissue set evaluated, approximately 40% of the samples were positive for a KRAS mutation, with the combination of different methods. In this study, all methods were performed in the same laboratory and provided highly concordant results. It was noted that if the tumor content was less than 10%, the mutation frequency was much lower (<5%) than samples with a greater tumor area (˜40% positive samples), highlighting the need for tumor enrichment in the molecular characterization of FFPE tissue samples. The robustness of the different methodologies was demonstrated by comparable performance on samples submitted to the testing site from multiple clinical locations without providing specific control for preanalytical variables, such a fixation time and storage conditions, including age of the tissue blocks. An additional comparative study involved sending a set of samples characterized for KRAS mutation status by dideoxy sequencing to five different commercial laboratories for an assessment of assay performance. The testing laboratories utilized different methodologies including allele-specific PCR amplification with ARMS technology in combination with Scorpion probes for detection; direct DNA sequencing; allele-specific primer extension; allele-specific oligonucleotide hybridization of PCR products for codons 12 and 13; and a picotiter plate pyrosequencing assay. The methods that demonstrated the best performance were the ARMS PCR assay, the allele-specific primer-extension assay, and the pyrosequencing assay. In addition to demonstrating some level of variability in the analytical performance of these methods, other factors considered important for assay performance were differences in sample processing and DNA-enrichment techniques, along with heterogeneity in tumor content between samples. In general, all three of these comparative studies demonstrate that a variety of methods provide comparable results to gold-standard dideoxy-sequencing assays for the detection of KRAS mutations. With all of these methods the sensitivity and specificity of the assays are influenced by the percentage of tumor cells in the samples, the quality of the DNA extracted from the sample, and the performance limitations of the specific method.
To help evaluate the reliability of KRAS mutation testing, the European Society of Pathology (ESP) has established a quality-assurance program for KRAS mutation analysis in colorectal cancers.[101] Participation is scheduled to be available internationally for laboratories that register for the program, beginning in 2011. Ten samples of invasive colorectal carcinomas will be used in the analysis, with three sections provided to the testing sites for hematoxylin eosin staining and KRAS mutation analysis. Assessment of tumor content and KRAS mutation genotype will be evaluated, along with information provided in the standard laboratory report.[22,42] Programs like this are necessary to provide an external mechanism of quality-assurance for laboratories performing molecular pathology-based testing.
Table 5 provides a summary of several key considerations for any molecular assay used for KRAS mutation analysis. Considerations include the preanalytical areas that have been discussed in the section earlier, as well as the spectrum of mutations to be tested, testing methodology, assay performance characteristics and laboratory operations. The ability to perform testing on a variety of specimen types including formalin-fixed paraffin embedded tissues is also very important. The need to characterize key features of the tumor, such as tumor content and percentage, can dramatically impact the analytical performance features and, therefore, it is critical to have appropriate pathology evaluation before molecular testing is performed.
