The first 2 authors contributed equally to this article.
Epidermal growth factor receptor mutations detected by denaturing high-performance liquid chromatography in nonsmall cell lung cancer
Impact on response to therapy with epidermal growth factor receptor-tyrosine kinase inhibitors
Article first published online: 19 AUG 2010
Copyright © 2010 American Cancer Society
Volume 116, Issue 18, pages 4309–4317, 15 September 2010
How to Cite
Cohen, V., Agulnik, J. S., Ang, C., Kasymjanova, G., Batist, G., Small, D., Brandao, G., Chong, G. and Miller, W. H. (2010), Epidermal growth factor receptor mutations detected by denaturing high-performance liquid chromatography in nonsmall cell lung cancer. Cancer, 116: 4309–4317. doi: 10.1002/cncr.25214
- Issue published online: 3 SEP 2010
- Article first published online: 19 AUG 2010
- Manuscript Accepted: 15 DEC 2009
- Manuscript Revised: 19 OCT 2009
- Manuscript Received: 14 JUL 2009
- nonsmall cell lung cancer;
- epidermal growth factor receptor;
- EGFR mutation;
- denaturing high-performance liquid chromatography;
- EGFR-tyrosine kinase inhibitors
Somatic mutations in the epidermal growth factor receptor (EGFR) kinase domain are associated with sensitivity to EGFR-tyrosine kinase inhibitors (EGFR-TKI) in patients with nonsmall cell lung cancer (NSCLC).
The authors tested the possibility that nucleotide sequencing may be poorly suited for detection of mutations in tumor samples and found that denaturing high-performance liquid chromatography (dHPLC) was an efficient and more sensitive method for screening.
These results suggested that some reports based on standard DNA sequencing techniques may have underestimated mutation rates. In the present report, the authors examined the relationship between the presence and type of EGFR mutations detected by dHPLC and various clinicopathologic features of NSCLC, including response to therapy with EGFR-TKI. Among 251 patients with advanced disease, 100 individuals received EGFR-TKI. Those whose tumors harbored a detectable EGFR kinase mutation were much more likely to have a partial response (PR) or stable disease (SD) with EGFR-TKI therapy than patients whose tumor contained no mutation (80% vs 35%; P = .001). Among the individual genotype subgroups, the frequency of a PR or SD was significantly different between patients with an exon 19 deletion compared with those with no detectable mutation (86% vs 35%; P < .001). Furthermore, patients whose tumors expressed an exon 19 mutant EGFR isoform exhibited a trend toward better EGFR-TKI response (86% vs 67%; P = .171) and improved survival compared with patients whose tumors expressed an exon 21 mutation.
Our findings warrant confirmation in large prospective trials and exploration of the biological mechanisms of the differences between mutation types. Cancer 2010. © 2010 American Cancer Society.
The small molecule epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKI) gefitinib and erlotinib have made a major impact on the treatment of advanced nonsmall cell lung cancer (NSCLC). In initial clinical trials of EGFR-TKI in NSCLC, although tumors in the vast majority of patients failed to respond, a minority showed dramatic tumor shrinkage accompanied by symptomatic improvement. Such responses were noted to be more common in East Asians, in women, in nonsmokers, and in patients with adenocarcinomas, especially those with areas of bronchioalveolar carcinoma.1-3 These observations suggested that there may be a molecular mechanism underlying the sensitivity to these drugs, which would be present more frequently in these patients. This ultimately led to the sequencing of the EGFR gene and the identification of EGFR mutations.4-6
Genetic alterations in 2 exons account for approximately 90% of the EGFR mutations reported to date in lung adenocarcinomas. The first is a short in-frame deletion of 9 to 24 nucleotides in exon 19. The other is a point mutation in exon 21 (2573T>G) that results in substitution of leucine by arginine at codon 858 (L858R).4-10 Other much less common mutations have also been described in exons 18, 20, and 21.
A large number of retrospective and prospective studies have confirmed the link between clinical characteristics associated with EGFR-TKI response and EGFR mutations.11-20 The retrospective response rate to EGFR-TKI treatment in mutation-positive cases is 77% (range of 30%-100%, with most series reporting response rates >60%), compared with 10% in mutation negative patients. Prospective mutational studies have yielded remarkably similar findings, despite the use of different EGFR-TKI and the participation of patients from different ethnic populations.21-26 Collectively, they have demonstrated response rates of 62% to 82%. Interestingly, emerging data suggest that specific EGFR genotypes may be predictive of favorable outcomes, specifically that cases with EGFR exon 19 deletions may have increased response rate and survival with EGFR-TKI compared with L858R cases.25, 26 This is in contrast to the reported natural history of such patients, where those with exon 19 deletions appear to have a shorter survival that those with L858R.9
Several conceptual and technical issues may confound the correlative analysis of EGFR mutations and response. First, the combination of infrequent mutations in unselected NSCLC cohorts and common use of fine-needle aspirates for diagnosis that are often insufficient for molecular analysis has resulted in relatively small numbers of mutation-positive patients in most series, limiting the statistical power of most US and European studies. Second, most retrospective studies include tumor samples collected at initial diagnosis, whereas EGFR-TKI therapy may have been administered after multiple courses of chemotherapy. As such, additional mutations leading to EGFR-TKI resistance may have arisen in the interim and may account for some mutation-positive unresponsive cases. Finally, technical differences in assessing EGFR mutations by polymerase chain reaction (PCR) amplification from archival formalin-fixed, paraffin-embedded tissue and the limited reliability of various mutation detection approaches may explain differences in reported mutation frequencies across studies. Our own group tested the possibility that nucleotide sequencing may be poorly suited for mutation screening of tumor samples and found that denaturing high-performance liquid chromatography (dHPLC) was an efficient and more sensitive method for screening.27 These results suggested that some previous reports using standard DNA sequencing techniques may have underestimated mutation rates.
In the present report, clinical specimens from NSCLC patients were analyzed for EGFR mutations with the aim of examining the relationship between the presence and type of EGFR mutations identified by dHPLC and various clinicopathologic features of NSCLC, including response to therapy with EGFR-TKI.
MATERIALS AND METHODS
Patient Selection and Clinical Chart Review
By using an electronic database maintained at the Division of Pulmonary Diseases of the Sir Mortimer B. Davis-Jewish General Hospital, all NSCLC patients with known EGFR mutation status were identified. Variables recorded included patient demographics (sex, disease stage, smoking status, histology, and the presence or absence of an exon 19 deletion or exon 21 mutation), tumor response to gefitinib or erlotinib, and survival outcomes. Staging was performed using American Joint Committee on Cancer staging criteria. Patients were categorized using standard criteria, including former smokers, defined as patients who had quit smoking at least 1 year before their diagnosis of lung cancer, and never-smokers, defined as patients who had smoked <100 cigarettes in their lifetime. Pack years of smoking were calculated by multiplying the number of packs smoked per day by the number of years smoking. Histology was classified according to the World Health Organization classification system. Criteria used for classifying response and progression included the Response Evaluation Criteria in Solid Tumors, radiographs, and computed tomography or bone scans. The study protocol was approved by our institutional review board.
Analysis of EGFR Mutations
Patients and genomic DNA
NSCLC tissues were obtained through protocols approved by our institutional review board. Cases included surgically resected specimens and samples of bronchoscopic and computed tomography-guided needle biopsies. All clinical samples were subjected to dHPLC analyses (Fig. 1). A LightCycler assay was used to confirm the presence of the L858R mutation in some specimens.
DNA extraction and PCR
Paraffin-embedded tumor samples were cut into 10 sections measuring 8 to 10 μm thick and put into Eppendorf tubes. One milliliter of toluene was added, and the samples were incubated at 37°C for 15 minutes, followed by a 5-minute centrifugation at 14,000 rpm at room temperature. The supernatant fluid was decanted, and this was followed by another wash with 1 mL of toluene. Samples were then washed twice with 1 mL of 100% ethanol and centrifuged at 14,000 rpm for 5 minutes. One hundred eighty microliters of ATL buffer (QIAGEN, Valencia, Calif) was added after air-drying the samples at 55°C for 30 minutes. QIAGEN's Tissue Protocol was then used according to the manufacturer's instructions, using reagents from the QIAamp DNA Blood Mini Kit (QIAGEN).
Frozen tissue samples were ground in cell lysis buffer, and 20 μL of proteinase K was added per milliliter of sample. DNA was extracted according to the Genomic DNA Isolation Kit (Puregene, Gentra Systems, Plymouth, Minn).
PCR reactions were set up as follows: 5 μL of 5× reaction buffer containing 15 mM MgCl2 and 200 nmol deoxynucleotide triphosphates (QIAGEN), 0.5 μL each of 20 μM EGFR exon 19 forward and reverse primers (Invitrogen, Carlsbad, Calif), 0.25 μL of 5U/μL Platinum Taq (Invitrogen), 2 μL of genomic DNA, and sterile water up to 25 μL. A pre-PCR denaturing step was carried out at 95°C for 5 minutes, followed by 35 cycles at 94°C for 20 seconds, 60°C for 30 seconds, and 72°C for 60 seconds. A final elongation was performed at 72°C for 10 minutes, followed by an indefinite hold at 5°C. A 1% agarose check gel was run to confirm amplification.
The dHPLC system used was a Transgenomic Wave Nucleic Acid Fragment Analysis System (Transgenomic, Omaha, Neb). dHPLC was carried out on the WAVE system with a DNASep column (Transgenomic). The mobile phases comprised 0.05% acetonitrile in 0.1 M triethylammonium acetate (TEAA) (eluent A) and 25% acetonitrile in 0.1 M TEAA (eluent B). To detect heteroduplices for exon 21, amplified products were subjected to an additional 3-minute, 95°C denaturing step followed by gradual reannealing from 95 to 65°C over a period of 30 minutes. The PCR-amplified exon 19 products were analyzed without any further treatment. Exon 19 and 21 products were then eluted at a flow rate of 0.9 mL/min. The start- and endpoints of the gradient by mixing eluents A and B and the temperature required for successful resolution of heteroduplex molecules were adjusted by using an algorithm provided with WAVEmaker system control software version 4.1.42 (Transgenomic). Five microliters of PCR products was injected for each run. The flow rate was 0.9 mL/min, and the ultraviolet detector was set to 260 nm. Heterozygous profiles were identified by visual inspection of the chromatograms on the basis of the appearance of additional, earlier eluting peaks. Corresponding homozygous profiles show only 1 peak.
LightCycler EGFR L858R mutation assay
Primers and probes were designed using LCProbe Design 1.0 (Roche Diagnostics, Basel, Switzerland). The forward and reverse primer sequences for exon 21 of EGFR were 5′-GTTTCAGGGCATGAACTAC-3′ and 5′-CTGACCTAAAGCCACCTC-3′, respectively. The sequences of the anchor and sensor probes were CTCTTCCGCACCCAGCAGT-Fluo and LC-640-GGCCCGCCCAAAATCTGT-P, respectively. The sensor probe was complementary to the mutant species. Real-time PCR, followed by melting curve analysis, was performed on a LightCycler 2.0 instrument using LightCycler Software 4.0 (Roche Diagnostics). Distinct melting temperatures were observed for each allele: 56°C for the wild type allele and 66°C for the mutant allele.
Patient characteristics were compared using Fisher exact test or chi-square test when appropriate. Univariate and multivariate logistic regression analyses were used to evaluate the strength of association between mutation status and factors such as sex, histology, and smoking status. Tumor response was categorized based on interval computed tomography scans as complete response, partial response (PR), stable disease (SD), or progressive disease using Response Evaluation Criteria in Solid Tumors. Association between tumor response and EGFR mutation was investigated using Fisher exact test.
Overall survival was defined as the time elapsed from the date of diagnosis of metastatic disease to the date of death or the date of the last follow-up for patients alive at the time of analysis. Overall survival curves were drawn using the Kaplan-Meier method, and the differences among mutation and wild type groups were compared using log-rank and Breslow tests.
Spectrum of Mutations in NSCLC Patients
Tumor specimens suitable for genetic analysis were available from 373 patients. The majority of tumors examined (77%) were adenocarcinomas or bronchioalveolar carcinomas. Overall, 88 (24%) of the 373 patients evaluated had activating mutations of EGFR; 64 patients exhibited exon 19 deletions, and 24 exhibited exon 21 point mutations. No tumors had an activating mutation in >1 exon. Of note, the clinical characteristics of our patients with mutations were similar to those of patients in published studies. The majority were female never-smokers with adenocarcinoma pathology (Table 1). In a multivariate analysis, smoking status, histology, and disease stage were the only significant predictors for the presence of an EGFR mutation (Table 2).
|Characteristic||Mutation Positive||Mutation Negative, No. (%)|
|Female||63 (29)||157 (71)|
|Male||25 (16)||128 (84)|
|Early: I, II, and III||15 (12)||107 (88)|
|Advanced: IIIB (effusion) and IV||73 (29)||178 (71)|
|Former/current smokers||50 (18)||236 (82)|
|Never-smokers||38 (44)||49 (56)|
|Adenocarcinoma||79 (27)||209 (73)|
|Other||9 (11)||76 (89)|
|Surgery||20 (14)||124 (86)|
|Combined chemoradiation||8 (17)||40 (83)|
|Doublet chemotherapy||17 (27)||47 (73)|
|Single-agent chemotherapy||13 (34)||25 (76)|
|Palliative radiotherapy||10 (38)||16 (62)|
|Best supportive care||20 (38)||33 (62)|
|First line||7 (33)||14 (67)|
|Second line||20 (42)||28 (58)|
|Third line||13 (42)||18 (58)|
|None||48 (18)||225 (82)|
|Gefitinib||9 (39)||14 (61)|
|Erlotinib||31 (40)||46 (60)|
|Type of biopsy|
|Surgical||22 (19)||118 (81)|
|Bronchoscopic||62 (30)||142 (70)|
|CT-guided needle||4 (20)||16 (80)|
|Factor||Univariate Analysis||Multivariate Analysis|
Correlation of Mutational Status With Clinical Response
Among 251 patients with advanced disease (stage IIIB pleural effusion or stage IV), 100 (40%) received EGFR-TKI therapy at some point in time during their treatment course. Of these, 40 (40%) had mutations of EGFR. No patient had a complete response. Nine (23%) patients had a PR as their best response and 23 (58%) had SD, for a total disease control rate of 80%. These figures are lower than those widely reported in the literature. Several factors may have contributed to this observation in our mutated cohort. First, our study was a retrospective analysis of highly selected patients from a single institution treated in a real-world environment in which clinical trial standards were not applied. Second, this study included tumor samples collected at initial diagnosis, whereas EGFR-TKI therapy was administered in many cases after multiple courses of chemotherapy. The majority of our patients (>80%) received EGFR-TKI in the second- or third-line setting. As such, additional mutations leading to EGFR-TKI resistance may have arisen in the interim and may account for some mutation-positive unresponsive cases. Finally, technical aspects in assessing EGFR mutations may have impacted our results. The low response rate to EGFR-TKI among mutation-positive patients raises the possibility of false-positive labeling with the dHPLC method. dHPLC can indeed give a false-positive screening result. The majority of these are often the consequence of overcautious scoring of dHPLC traces, as samples that give a weak or atypical dHPLC trace are typically recorded as positive, as is appropriate for a screening technique. However, we do not believe that this is the case in our study. Our own group previously reported on dHPLC (the same assay used in this study) for detection of EGFR mutations, showing that this method was an efficient and more sensitive method for screening compared with sequence analysis.27 In that study, all amplicons with shifts detected by dHPLC were because of a DNA sequence variation; the false-positive rate of dHPLC in that analysis was 0%.
Response data for the various genotypes are listed in Tables 3. Patients whose tumors harbored a detectable EGFR kinase mutation were much more likely to have a PR or SD with EGFR-TKI therapy than patients whose tumor contained no mutation (80% vs 35%; P = .001). Among the individual genotype subgroups, the frequency of a PR or SD was significantly different between patients with an exon 19 deletion mutation compared with those with no detectable mutation (86 vs 35%; P < .001). Patients whose tumors expressed an exon 19 mutant EGFR isoform exhibited a nonsignificant trend toward better EGFR-TKI response compared with patients whose tumors expressed an exon 21 mutation (83% vs 67%; P = .171). Those patients harboring an exon 21 mutation also appeared to respond better to EGFR-TKI compared with wild type cases; however, because of the small number of L858R cases, this difference did not reach statistical significance. Furthermore, treatment with EGFR-TKI was associated with a significant improvement in survival in all patients (18.3 months; 95% confidence interval [CI], 13.4-23.2 vs 11.0; 95% CI, 8.1-13.9 months; Breslow test P = .001).
|Mutation||Response to TKI||Total|
|Exon 19a,b||6 (22%)||18 (64%)||4 (14%)||28|
|Exon 21||3 (25%)||5 (42%)||4 (33%)||12|
|Wild type||5 (8%)||16 (27%)||39 (65%)||60|
Correlation of Tumor Genotype With Survival Outcomes
On the basis of Kaplan-Meier analysis, the median overall survival for the entire population was 15.6 months. Median survival for patients with advanced disease was 13.3 months. Patients with stage IIIB pleural effusion or stage IV disease whose tumors expressed an exon 19 mutant EGFR isoform exhibited a trend toward improved survival compared with patients whose tumors expressed an exon 21 mutation or who had no detectable mutation (19.9 vs 13.5 vs 14.5 months, respectively, P = .55). There was no significant difference in survival in favor of the exon 21 mutation subgroup compared with those patients with no EGFR mutation (Fig. 2).
In recent years, EGFR-TKI have assumed a prominent role in the management of NSCLC alongside traditional cytotoxic chemotherapeutic agents. As a result, identification of predictive markers has become increasingly relevant in helping to identify those most likely to benefit from these agents to maximize responses and spare patients unnecessary toxicities. Although several important demographic and clinical factors are associated with treatment response, EGFR somatic mutations have emerged as the best predictors for EGFR-TKI sensitivity. Direct sequencing analyses remain the current gold standard for detecting EGFR mutations; however, these methods are labor-intensive and time-consuming. Furthermore, in comparison to newer assays such as dHPLC, the sensitivity of sequencing was remarkably lower at 30%.28, 29 Recently, our group demonstrated that dHPLC is capable of extracting DNA from significantly smaller tumor sample volumes than required for sequencing.27 This is particularly advantageous given that pathological and molecular analysis must often be performed on less than optimal tissue specimens. It is interesting to note that in our study population, 26% of evaluable tissue specimens were found to contain mutation, which is higher than the 10% to 15% incidence typically reported in the general North American and Western European populations.5 It is possible that the enhanced sensitivity of the dHPLC might have contributed to our higher mutation rate. The higher rate among our cohort may also be attributable to an imbalance in the pathological type or other patient characteristics.
In this series of NSCLC patients, the clinicopathologic profiles and outcomes of subjects who tested positive for EGFR mutations were confirmed. Consistent with previous reports, patients who are female, never-smokers, and who have bronchioalveolar or adenocarcinoma histology were significantly more likely to harbor mutations.30-32 Smoking status was the strongest predictor of harboring an EGFR mutation.
For patients with advanced disease, EGFR mutations were associated with an improved response to treatment with EGFR-TKI agent (either in form of partial response or disease stabilization). Also observed in our cohort was a significant prolongation of survival in patients with EGFR mutations compared with mutation-negative patients after adjusting for age, diagnosis, and stage of disease. This was a heterogeneous population, receiving many treatment modalities and therapeutic regimens; therefore, this finding suggests that EGFR mutation is a favorable prognostic factor regardless of treatment. This conclusion has been drawn by other investigators, including the TRIBUTE investigators, where mutation-positive patients had a prolonged survival regardless of treatment arm.33 However, a prospective randomized study of an EGFR-targeted agent compared with another systemic treatment, in EGFR mutation-positive patients exclusively, is required to determine whether mutations may be both prognostic and predictive of increased survival. The recently reported IPASS (Iressa Pan-Asia Study) study has provided us with 1 of the first opportunities to directly compare chemotherapy to an EGFT-TKI and may help clarify the prognostic and predictive relevance of EGFR mutations. IPASS was an open-label, randomized, parallel-group study that assessed the efficacy, safety, and tolerability of gefitinib versus doublet chemotherapy (carboplatin and paclitaxel) as first-line treatment in a clinically selected population of patients from Asia. Results demonstrated superior progression-free survival for gefitinib compared with doublet chemotherapy in the overall population of clinically selected patients with advanced NSCLC in Asia. The EGFR mutation status of a patient's tumor was a strong predictor of benefit with gefitinib or chemotherapy. Preplanned subgroup analyses showed that progression-free survival was significantly longer for gefitinib than doublet chemotherapy in patients with EGFR mutation-positive tumors.34
The distribution of EGFR mutation types identified in our cohort was consistent with other series, with most mutations being in-frame exon 19 deletions and the L858R point mutations in exon 21.20, 25, 26, 35 These 2 most frequent EGFR mutations have been consistently associated with EGFR-TKI responsiveness.36, 37 Among the individual genotype subgroups in our patient population, the frequency of a PR or SD was significantly higher among patients with an exon 19 deletion mutation compared with those with exon 21 point mutations or no detectable mutation. Those patients harboring an exon 21 mutation also appeared to respond better to EGFR-TKI compared with wild type cases; however, because of the small number of L858R cases, this difference did not reach statistical significance.
Although our analysis suggests that these 2 mutations are clinically different, it remains to be confirmed whether these 2 mutations or other activating EGFR mutations have different clinical impact. In the largest retrospective cohort of patients from 2 US centers that followed EGFR-mutant NSCLC patients given EGFR-TKI (gefitinib or erlotinib) as first- to third-line therapy, it was observed that patients with exon 19 deletion had a significantly improved time to progression and overall survival when compared with L858R patients.25, 26 In both reports, patients with exon 19 in-frame deletions had at least double the progression-free and overall survival of the group with L858R mutation. An explanation for the differential response of the 2 EGFR mutation types to EGFR-TKI remains elusive. The secondary EGFR mutation, T790M, results in the substitution of methionine for threonine in the tyrosine kinase domain of EGFR, disrupting normal binding of erlotinib.38 Jackman et al have hypothesized that T790m mutations may occur more frequently with L585R mutations than with exon 19 deletions.25 In 2 Japanese prospective trials, the response rates for both mutations were not significantly different. Inou et al reported response rates of 67% and 86% for exon 19 deletions and L858R mutants, respectively.21 Asahina et al reported response rates of 83% for exon 19 deletions and 67% for L858R patients.39 The time to progression in the latter trial is similar for both mutant cohorts. Other studies have searched for disparities in the efficacy of EGFR-TKI inhibition between mutation types but have not yielded any definitive findings.
Our study is unique in its description of a large series of patients undergoing EGFR mutation testing using dHPLC as part of standard clinical practice, but it is subject to some limitations. One obvious disadvantage is our sample size, which was limited to only 78 mutation-positive patients. We therefore lacked the power to obtain definitive results from the examination of differences in outcome between various subtypes of EGFR mutations.
In conclusion, we have demonstrated that EGFR mutation screening using dHPLC can be incorporated into clinical care of NSCLC patients. We have confirmed that individuals whose tumors harbored a detectable EGFR kinase mutation were much more likely to have a PR or SD with EGFR-TKI therapy than patients whose tumor contained no mutation. Among the individual genotype subgroups, the frequency of PR or SD was significantly different between patients with an exon 19 deletion mutation compared with those with exon 21 point mutations or no detectable mutation. Furthermore, patients whose tumor expressed an exon 19 mutant EGFR isoform exhibited a trend toward improved survival compared with patients whose tumor expressed an exon 21 mutation. Our findings warrant confirmation in large prospective trials and exploration of the biological mechanism of the differences between mutation types.
CONFLICT OF INTEREST DISCLOSURES
This work was supported in part by a research grant from Roche Canada.
- 24iTARGET: a phase II trial to assess the response to gefitinib in epidermal growth factor receptor (EGFR)-mutated non-small cell lung cancer (NSCLC) tumors [abstract]. J Clin Oncol. 2007; 25: 7504., , , et al.
- 27Evaluation of denaturing high-performance liquid chromatography as a rapid detection method for identification of epidermal growth factor receptor mutations in nonsmall-cell lung cancer. Cancer. 2006; 107: 2858-2865., , , et al.
- 28The detection by denaturing high performance liquid chromatography of epidermal growth factor receptor mutation in tissue and peripheral blood from patients with advanced non-small cell lung cancer [in Chinese]. Zhonghua Jie He He Hu Xi Za Zhi. 2008; 31: 891-896., , , et al.