Assessing EGFR Mutations
http://www.100md.com
《新英格兰医药杂志》
To the Editor: Recent reports have noted the presence of novel mutations of the epidermal growth factor receptor (EGFR) gene in small samples of DNA extracted from paraffin-embedded sections or laser-microdissected specimens.1,2 Nagahara et al.1 reported EGFR mutations in colon carcinomas. These mutations were novel GA or AG transitions. Similarly, Tsao et al. (July 14 issue)2 report that 24 (53 percent) of the mutations they found in specimens of non–small-cell lung cancers were novel variant mutations; 22 (92 percent) of these mutations were CT/GA or AG/TC transitions. These mutations had not been reported previously in more than 2000 analyses for EGFR mutations performed on DNA extracted from pieces of frozen tumors.
By sequencing multiple polymerase-chain-reaction (PCR) amplification products (10 per sample) from 70 samples of lung-tumor DNA extracted from a single (5 μm) paraffin-embedded section, we found 45 uncommon mutations in exons 18 through 21 of the EGFR gene, including the 22 transitions reported by Tsao et al. and the 4 mutations reported by Nagahara et al. All the uncommon mutations detected were found to be artifacts. Indeed, the mutations were also found in multiple amplifications of paraffin-embedded DNA samples from 50 normal tissues (lymph nodes) obtained from 50 patients who did not have neoplasms; the mutations were identified in a minority of the amplifications from the same sample; and the CT/GA mutations were prevented by adding uracil-N-glycosylase to the DNA template before PCR amplification.
The artifacts we observed can be ascribed to postmortem deamination of cytosine or adenine, resulting in uracil or hypoxanthine residues, respectively.3 When DNA containing such lesions is used as a template for PCR, CT/GA or AG/TC transitions will result. CT/GA mutations can be prevented by treating the template with uracil-N-glycosylase, which removes uracil from DNA, creating a strand break. Artifactual mutations have been described by several authors,3,4,5 and alternative explanations have also been reported.4,5
These artifacts can easily be observed by carrying out multiple PCR amplifications of very small amounts of DNA, particularly if the DNA is isolated from paraffin-embedded tissues3,4,5 (Figure 1). With the use of a low DNA copy number, PCR amplification starts from single or few templates, making possible the detection of alterations present in subsets of DNA molecules that are otherwise undetectable.4
Figure 1. Sequences of Multiple PCR Amplifications of Small Amounts (5 ng) of DNA Extracted from Paraffin-Embedded Tissues.
Dots indicate that the sequence is identical with the consensus sequence of EGFR exon 19, shown at the top of the panels. The numbers correspond to the number of amplifications. Panel A shows DNA from a lung tumor, in which five different mutations are present in four independent amplifications. Panel B shows DNA from normal tissue (lymph node), in which six different mutations are present in five independent amplifications. Multiple mutations can be seen in the same DNA sample and sometimes in the same amplification (e.g., amplification 3 in Panel A and amplifications 6 and 9 in Panel B); furthermore, only one or a few amplifications from a single DNA sample show the same mutation, indicating that base changes are present in subsets of DNA molecules. Bases enclosed in rectangles correspond to alterations that were reported by Nagahara et al.1 and Tsao et al.2 as being novel mutations. PCR denotes polymerase chain reaction, and EGFR epidermal growth factor receptor.
Since EGFR mutations may be used for guidance in the clinical management of lung cancer, extreme care must be taken when working with small amounts of DNA, especially if the samples were extracted from paraffin. The occurrence of artifacts can be prevented with the use of larger amounts of template DNA (at least 1 μg of DNA recovered from paraffin). If very few copies of template molecules must be used, as in the case of DNA obtained from microdissected sections, the examination of multiple amplifications is imperative.
Antonio Marchetti, M.D.
Lara Felicioni, M.D.
Fiamma Buttitta, M.D.
Center for Excellence on Aging
66013 Chieti, Italy
amarchetti@unich.it
References
Nagahara H, Mimori K, Ohta M, et al. Somatic mutations of epidermal growth factor receptor in colorectal carcinoma. Clin Cancer Res 2005;11:1368-1371.
Tsao M-S, Sakurada A, Cutz J-C, et al. Erlotinib in lung cancer -- molecular and clinical predictors of outcome. N Engl J Med 2005;353:133-144.
Hofreiter M, Jaenicke V, Serre D, Haeseler Av A, Paabo S. DNA sequences from multiple amplifications reveal artifacts induced by cytosine deamination in ancient DNA. Nucleic Acids Res 2001;29:4793-4799.
Akbari M, Hansen MD, Halgunset J, Skorpen F, Krokan HE. Low copy number DNA template can render polymerase chain reaction error prone in a sequence-dependent manner. J Mol Diagn 2005;7:36-39.
Williams C, Pontén F, Moberg C, et al. A high frequency of sequence alterations is due to formalin fixation of archival specimens. Am J Pathol 1999;155:1467-1471.
The authors reply: Marchetti et al. have highlighted the difficulties in performing an EGFR mutation analysis on small, formalin-fixed, paraffin-embedded biopsy specimens, which commonly are the only samples available from patients with advanced lung cancer and are insufficient to permit repeated assays. We recently completed mutational assays on 45 additional samples from the BR.21 trial and successfully obtained results for exons 19 and 21 in 25 samples (success rate, 56 percent). Among the results, we identified four new cases with exon 19 deletions or exon 21 L858R mutations. Given the uncertainty surrounding the identity or functional significance of novel mutations, we have reanalyzed our data, focusing only on exon 19 deletions and exon 21 L858R mutations for which the functional significance is clearer. Including the new samples, 24 cases that could be evaluated had these classic mutations, among 201 patients, for a prevalence rate of 11.9 percent, which is similar to mutation rates reported by other investigators. Among patients with classic mutations who were treated with erlotinib, the response rate was 30.0 percent (3 of 10 patients), as compared with 7.5 percent (8 of 106) for the remaining 106 patients (P=0.05). The hazard ratio for death associated with erlotinib, as compared with placebo, was 0.52 (95 percent confidence interval, 0.21 to 1.31; P=0.16) among those with classic mutations (Figure 1A), whereas it was 0.75 (95 percent confidence interval, 0.53 to 1.07; P=0.11) among those with wild-type EGFR or other variant mutations of indeterminate significance (Figure 1B). The interaction P value for treatment by classic-mutation status was 0.45. These findings, although based on small numbers, are consistent with our previous results showing no significant difference in the survival benefit achieved with erlotinib between patients with classic mutations and patients with wild-type EGFR. We therefore continue to believe that testing for EGFR mutations is not necessary to identify patients for treatment with erlotinib, since patients with wild-type EGFR and those with mutant genotypes both have the potential to benefit from therapy.
Figure 1. Kaplan–Meier Estimates of Survival.
Panel A shows the results for patients with exon 19 deletions or exon 21 L858R mutations, and Panel B shows the results for patients with wild-type EGFR or other mutations of indeterminate significance. P values were calculated with the use of the log-rank test. CI denotes confidence interval.
Ming-Sound Tsao, M.D.
Suzanne Kamel-Reid, Ph.D.
Frances A. Shepherd, M.D.
Princess Margaret Hospital
Toronto, ON M56 2M9, Canada
By sequencing multiple polymerase-chain-reaction (PCR) amplification products (10 per sample) from 70 samples of lung-tumor DNA extracted from a single (5 μm) paraffin-embedded section, we found 45 uncommon mutations in exons 18 through 21 of the EGFR gene, including the 22 transitions reported by Tsao et al. and the 4 mutations reported by Nagahara et al. All the uncommon mutations detected were found to be artifacts. Indeed, the mutations were also found in multiple amplifications of paraffin-embedded DNA samples from 50 normal tissues (lymph nodes) obtained from 50 patients who did not have neoplasms; the mutations were identified in a minority of the amplifications from the same sample; and the CT/GA mutations were prevented by adding uracil-N-glycosylase to the DNA template before PCR amplification.
The artifacts we observed can be ascribed to postmortem deamination of cytosine or adenine, resulting in uracil or hypoxanthine residues, respectively.3 When DNA containing such lesions is used as a template for PCR, CT/GA or AG/TC transitions will result. CT/GA mutations can be prevented by treating the template with uracil-N-glycosylase, which removes uracil from DNA, creating a strand break. Artifactual mutations have been described by several authors,3,4,5 and alternative explanations have also been reported.4,5
These artifacts can easily be observed by carrying out multiple PCR amplifications of very small amounts of DNA, particularly if the DNA is isolated from paraffin-embedded tissues3,4,5 (Figure 1). With the use of a low DNA copy number, PCR amplification starts from single or few templates, making possible the detection of alterations present in subsets of DNA molecules that are otherwise undetectable.4
Figure 1. Sequences of Multiple PCR Amplifications of Small Amounts (5 ng) of DNA Extracted from Paraffin-Embedded Tissues.
Dots indicate that the sequence is identical with the consensus sequence of EGFR exon 19, shown at the top of the panels. The numbers correspond to the number of amplifications. Panel A shows DNA from a lung tumor, in which five different mutations are present in four independent amplifications. Panel B shows DNA from normal tissue (lymph node), in which six different mutations are present in five independent amplifications. Multiple mutations can be seen in the same DNA sample and sometimes in the same amplification (e.g., amplification 3 in Panel A and amplifications 6 and 9 in Panel B); furthermore, only one or a few amplifications from a single DNA sample show the same mutation, indicating that base changes are present in subsets of DNA molecules. Bases enclosed in rectangles correspond to alterations that were reported by Nagahara et al.1 and Tsao et al.2 as being novel mutations. PCR denotes polymerase chain reaction, and EGFR epidermal growth factor receptor.
Since EGFR mutations may be used for guidance in the clinical management of lung cancer, extreme care must be taken when working with small amounts of DNA, especially if the samples were extracted from paraffin. The occurrence of artifacts can be prevented with the use of larger amounts of template DNA (at least 1 μg of DNA recovered from paraffin). If very few copies of template molecules must be used, as in the case of DNA obtained from microdissected sections, the examination of multiple amplifications is imperative.
Antonio Marchetti, M.D.
Lara Felicioni, M.D.
Fiamma Buttitta, M.D.
Center for Excellence on Aging
66013 Chieti, Italy
amarchetti@unich.it
References
Nagahara H, Mimori K, Ohta M, et al. Somatic mutations of epidermal growth factor receptor in colorectal carcinoma. Clin Cancer Res 2005;11:1368-1371.
Tsao M-S, Sakurada A, Cutz J-C, et al. Erlotinib in lung cancer -- molecular and clinical predictors of outcome. N Engl J Med 2005;353:133-144.
Hofreiter M, Jaenicke V, Serre D, Haeseler Av A, Paabo S. DNA sequences from multiple amplifications reveal artifacts induced by cytosine deamination in ancient DNA. Nucleic Acids Res 2001;29:4793-4799.
Akbari M, Hansen MD, Halgunset J, Skorpen F, Krokan HE. Low copy number DNA template can render polymerase chain reaction error prone in a sequence-dependent manner. J Mol Diagn 2005;7:36-39.
Williams C, Pontén F, Moberg C, et al. A high frequency of sequence alterations is due to formalin fixation of archival specimens. Am J Pathol 1999;155:1467-1471.
The authors reply: Marchetti et al. have highlighted the difficulties in performing an EGFR mutation analysis on small, formalin-fixed, paraffin-embedded biopsy specimens, which commonly are the only samples available from patients with advanced lung cancer and are insufficient to permit repeated assays. We recently completed mutational assays on 45 additional samples from the BR.21 trial and successfully obtained results for exons 19 and 21 in 25 samples (success rate, 56 percent). Among the results, we identified four new cases with exon 19 deletions or exon 21 L858R mutations. Given the uncertainty surrounding the identity or functional significance of novel mutations, we have reanalyzed our data, focusing only on exon 19 deletions and exon 21 L858R mutations for which the functional significance is clearer. Including the new samples, 24 cases that could be evaluated had these classic mutations, among 201 patients, for a prevalence rate of 11.9 percent, which is similar to mutation rates reported by other investigators. Among patients with classic mutations who were treated with erlotinib, the response rate was 30.0 percent (3 of 10 patients), as compared with 7.5 percent (8 of 106) for the remaining 106 patients (P=0.05). The hazard ratio for death associated with erlotinib, as compared with placebo, was 0.52 (95 percent confidence interval, 0.21 to 1.31; P=0.16) among those with classic mutations (Figure 1A), whereas it was 0.75 (95 percent confidence interval, 0.53 to 1.07; P=0.11) among those with wild-type EGFR or other variant mutations of indeterminate significance (Figure 1B). The interaction P value for treatment by classic-mutation status was 0.45. These findings, although based on small numbers, are consistent with our previous results showing no significant difference in the survival benefit achieved with erlotinib between patients with classic mutations and patients with wild-type EGFR. We therefore continue to believe that testing for EGFR mutations is not necessary to identify patients for treatment with erlotinib, since patients with wild-type EGFR and those with mutant genotypes both have the potential to benefit from therapy.
Figure 1. Kaplan–Meier Estimates of Survival.
Panel A shows the results for patients with exon 19 deletions or exon 21 L858R mutations, and Panel B shows the results for patients with wild-type EGFR or other mutations of indeterminate significance. P values were calculated with the use of the log-rank test. CI denotes confidence interval.
Ming-Sound Tsao, M.D.
Suzanne Kamel-Reid, Ph.D.
Frances A. Shepherd, M.D.
Princess Margaret Hospital
Toronto, ON M56 2M9, Canada