Immunohistochemical detection of mutations in the epidermal growth factor receptor gene in lung adenocarcinomas using mutation-specific antibodies
© Xiong et al.; licensee BioMed Central Ltd. 2013
Received: 18 December 2012
Accepted: 5 February 2013
Published: 18 February 2013
The recent development of antibodies specific for the major hotspot mutations in the epidermal growth factor receptor (EGFR), L858R and E746_A750del, may provide an opportunity to use immunohistochemistry (IHC) as a screening test for EGFR gene mutations. This study was designed to optimize the IHC protocol and the criteria for interpretation of the results using DNA sequencing as the gold-standard.
Tumor sections from fifty lung adenocarcinoma specimens from Chinese patients were immunostained using L858R and E746_A750del-specific antibodies using three different antigen retrieval solutions, and the results were evaluated using three different sets of criteria. The same specimens were used for DNA purification and analysis of EGFR gene mutations.
In this study the optimal buffer for antigen retrieval was EDTA (pH 8.0), and the optimal scoring method was to call positive results when there was moderate to strong staining of membrane and/or cytoplasm in >10% of the tumor cells. Using the optimized protocol, L858R-specific IHC showed a sensitivity of 81% and a specificity of 97%, and E746_A750del-specific IHC showed a sensitivity of 59% and a specificity of 100%, both compared with direct DNA analysis. Additionally, the mutant proteins as assessed by IHC showed a more homogeneous than heterogeneous pattern of expression.
Our data demonstrate that mutation-specific IHC, using optimized procedures, is a reliable prescreening test for detecting EGFR mutations in lung adenocarcinoma.
The virtual slide(s) for this article can be found here: http://www.diagnosticpathology.diagnomx.eu/vs/2059012601872392
KeywordsLung adenocarcinoma Epidermal growth factor receptor Mutation Immunohistochemistry
Somatic mutations within the tyrosine kinase (TK) domain of the epidermal growth factor receptor (EGFR) gene are found in approximately 30% of lung adenocarcinomas in Asian populations. Studies support that some of these activating mutations are not only reliable predictors of response to the small molecule EGFR tyrosine kinase inhibitors (TKIs) gefitinib and erlotinib but also prognostic factors for survival[2–4]. Among numerous TK domain mutations, 85–90% are exon 19 E746_A750 deletions or exon 21 L858R point mutations. A variety of DNA-based molecular methods are used to detect EGFR mutations. These methods have respective advantages and disadvantages, with no consensus on which one is the best. For example, direct sequencing of PCR-amplified genomic DNA can detect all mutations in the regions analyzed, but has limited analytical sensitivity when the tumor cells are not a large fraction of the specimen. The amplification refractory mutation system (ARMS) assay is more sensitive, but detects fewer mutations, usually only one per reaction. In general, direct analysis of DNA is expensive because of the cost of the equipment and reagents. In addition it is technically complex, and usually done in laboratories that specialize in molecular pathology.
Literature review of sensitivity and specificity of mutation-specific immunohistochemistry
Simonetti et al.
Kato Y et al.
Kitamura A et al.
Brevet M et al.
Wu SG et al.
Kozu Y et al.
Hofman P et al.
In the study reported here we optimized the methodology and interpretive aspects of IHC for detection of EGFR mutations, and evaluated the success of this effort by comparison with DNA sequencing. This study investigated the staining protocol, staining pattern, scoring methods, and cut off value to determine the diagnostic power of EGFR mutation-specific IHC in Chinese lung adenocarcinoma patients.
Samples for study were selected according to the following criteria: lung adenocarcinoma, surgically resected, primary, solitary and no preoperative therapy. A total of 50 cases were collected retrospectively and prospectively from the Department of Pathology, Peking University First Hospital during January 2010 to January 2012.
All specimens were dissected and immersed in 10% neutral buffered formalin, then fixed overnight. The number of sections for histology depended on the greatest dimension of tumors, i.e. one section per centimeter. If a tumor was less than 2 cm in greatest dimension, the tumor was totally sampled for microscopic examination. Sectioned tissues were embedded in paraffin routinely.
Informed consent for the use of these specimens for medical studies was obtained before surgery.
50 tissue blocks were cut into 4-μm-thick whole sections. EGFR mutation specific antibodies were Rabbit XP® mAbs obtained from Cell Signaling Technology (Danvers, MA), 6B6 specific for the E746-A750del mutation, and 43B2 for the L858R mutation. The antibodies were diluted 1:100 with antigen retrieval buffer before use. The antigen retrieval buffers tested were sodium citrate (pH 6.0), EDTA (pH 8.0) and EDTA (pH 9.0). Cytokeratin AE1/AE3 IHC was used as a quality control for tissue and protocol. The IHC protocol is described in greater detail in the Additional file 1.
Three sets of criteria were used for interpretation of the IHC results, referred to as Score A, B and C, respectively, in this study. A positive result using score A was moderate to strong staining of membrane and/or cytoplasm in >10% tumor cells. A positive result using score B was membrane staining in >10% tumor cells with any intensity. A positive result using score C was membrane and/or cytoplasmic staining in >50% of the tumor cells with any intensity. In this study all 50 specimens were analyzed using Score A, B and C separately, so as to evaluate the validity of these scoring methods by comparing to the results of DNA sequencing. Both the intensity and percentage of stained cells were assessed at low magnification (objective magnification ×10). The distribution of staining, membrane or cytoplasm, was assessed at high magnification (objective magnification ×40). Four experienced pathologists (Yan Xiong, Ying Dong, Lin Nong and Jing Zhao) reviewed all of the slides independently, and then replicated the analysis 16 to 18 weeks later. The intra- and inter-observer reliability was analyzed.
H&E stained sections of FFPE tissue were reviewed for each sample and those with greater than 50% tumor volume were selected for molecular testing. Genomic DNA was extracted using the QIAamp DNA FFPE Tissue Kit (Qiagen, Valencia, CA) according to the manufacturer’s protocol.
Mutant-enriched PCR for EGFR exon 19 and 21
Sequences of oligodeoxyribonucleotides
DNA sequencing for EGFR exon 19 and 21
The amplicons were treated with ExoSap (Amersham Biosciences, Piscataway, NJ) to remove the primers and dNTPs; then 1 μL was sequenced using the M13 tail primers as sequencing primers and Applied Biosystems (ABI, Foster City, CA) BigDye Terminator v.3.1 chemistry. The sequencing reactions were purified using the CleanSeq system (Agencourt Bioscience, Beverly, MA) and then resolved by capillary electrophoresis on the ABI 3100 Prism Genetic Analyzer.
Statistical analysis was done using the statistics software SPSS V16.0 (SPSS Inc., Chicago, IL). Fleiss’ Kappa was used to determine inter-observer agreement. Cohen’s Kappa was used to determine intra -observer agreement and agreement of IHC and DNA sequencing. A Kappa value between 0.81 and 1.0 was defined as nearly perfect agreement, between 0.61 and 0.8 as substantial agreement, between 0.41 and 0.60 as moderate agreement, between 0.21 and 0.40 as fair agreement, between 0.00 and 0.20 as slight agreement. For each Kappa, the 95% confidence interval (CI) was calculated. Difference was considered significant (P < 0.05), if the lower and upper boundary of the 95% CI showed no overlap.
All experiments above have been performed with the approval of Peking University First Hospital Ethics Committee.
In the total cohort of 50 samples L858R was identified in 16 cases, a deletion in exon 19 in 17 cases, and neither of them in 17 cases. Of the 17 cases with exon 19 deletion, 14 had a p.E746_A750del (c.del2235_2249 on the DNA level), one had a p.L747_T751del (c.2240_2254del), one had a p.L747_P753delinsS (c.2240_2257del), and one had a p.L747_T751delinsPT (c.2239_2253delinsCCAACG) that had not been previously reported. In our study, of all 33 cases with EGFR mutations, L858R and E746_A750del together comprised 90% (30/33) and the others, including L747_T751del, L747_P753delinsS and L747_T751delinsPT, comprised 10%, which was concordant with other studies. From this point of view L858R and E746_A750del are recognized as the most common mutations and the other mutation types are described as uncommon mutations.
Evaluation of antigen retrieval buffer
Intra- and inter-observer agreement based on slides treated with different antigen retrieval buffers
Antigen retrieval buffer
EDTA (pH 8.0)
Na Citrate (pH 6.0)
EDTA (pH 9.0)
Based on different scoring systems, the percentage of positive cases was different too. For L858R-specific IHC it was 28% (14/50) on Score A, 16% (8/50) on Score B, and 40% (20/50) on Score C; for E746_A750del-specific IHC it was 20% (10/50) on Score A, 14% (7/50) on Score B, and 24% (12/50) on Score C.
Concordance analysis of IHC and DNA sequencing
Diagnostic power of L858R-specific IHC on score A, B and C
0.810 (95% CI: 0.701, 0.919)
0.470 (95% CI: 0.332, 0.608)
0.397 (95% CI: 0.261, 0.533)
Diagnostic power of E746_A750del-specific IHC on score A, B and C
0.653 (95% CI: 0.521, 0.785)
0.480 (95% CI: 0.342, 0.618)
0.472 (95% CI: 0.334, 0.610)
The studies that established the relationship between mutations in the EGFR gene and response to the small molecule EGFR TKIs gefitinib and erlotinib were done using analysis of DNA extracted from the tumor. The recent availability of antibodies that are specific for the mutations most clearly associated with response to EGFR TKIs, L858R and E746_A750del, create the opportunity to exploit an alternative method to evaluate NSCLC for EGFR mutations to aid decisions with regard to EGFR TKI therapy.
IHC has the advantage of being a method that is routinely applied in solid tumor diagnosis in pathology. Also, it can be used on specimens that are not optimal for DNA analysis such as small tissue samples or individual cells obtained from body fluids, bronchial washings, and fine needle aspirates. Although some studies have shown that EGFR gene sequencing could be successfully applied to cytological specimens, it is still a problem to get enough DNA for sequencing from such samples in routine practice. Thus, the development of antibodies that specifically detect mutant EGFR protein by IHC could be a valuable addition to the current methods used to diagnose and predict response to treatment of lung cancer.
In 2009, Yu et al. reported generating two mAbs from New Zealand rabbits, one against the E746_A750del and the other against the L858R point mutation, and evaluated them by Western blotting, immunofluorescence and IHC. They tested these antibodies in a series of cell lines and in tumor tissues from patients with primary NSCLC, with known and unknown EGFR mutations, comparing the IHC results with DNA sequencing. They found that IHC with these mutation-specific antibodies showed a sensitivity of 92% and a specificity of 99%. Recently, several studies examined the presence of EGFR mutations in NSCLC by IHC using the same two antibodies and the reported sensitivity ranged from 24% to 100% and specificity ranged from 77% to 100%[8–15]. IHC is known to sometimes suffer from high inter-laboratory variability in assay performance, and high inter-observer variability in assay interpretation. These drawbacks may explain the variability in results of the studies described above. There is still much work to be done before IHC can be considered an adequate substitute for direct analysis of mutations in the EGFR gene in NSCLC.
In our study we found that slides treated by EDTA (pH 8.0) showed the best histological pictures with strongly specific staining and minimal background. As a result when pathologists reviewed these slides the intra- and inter-observer agreement was better than those treated by sodium citrate (pH 6.0) and EDTA (pH 9.0). The difference was statistically significant (P < 0.05). In conclusion, EDTA (pH 8.0) is the preferred buffer for antigen retrieval for IHC using EGFR mutation specific antibodies.
Scoring is the final step involved in the IHC protocol, but is not the least one, because the scoring system plays a critical role in obtaining a reliable result. In our study, we compared three scoring systems that have been used by other investigators, using DNA sequencing as the gold standard[10, 13, 15]. For L858R-specific IHC the agreement with DNA sequencing using Score A was superior to Score B and C. The difference was statistically significant. For E746_A750del-specific IHC the agreement with DNA sequencing was good for Score A, which was superior to Score B and C, but the difference was not statistically significant. In conclusion our study showed that Score A is the most appropriate way to interpret the EGFR mutation-specific IHC.
Based on Score A the specificity of EGFR mutation-specific IHC was very high, 100% for exon 19 deletions and 97% for L858R, while sensitivity was lower, 81% for L858R and 59% for exon 19 deletions. In another words, mutation-specific IHC demonstrated extremely high specificities, but much lower sensitivity. The low sensitivity of the exon 19 del IHC is mostly due to the presence of several exon 19 deletions other than the most common E746_A750del, which is the target of the exon 19 del antibodies. We conclude, based on our work, that NSLC cases positive by IHC could be selected as candidates for EGFR-TKI, while negative cases should be referred for further testing by DNA analysis.
In our study the majority of cases were either negative or positive in 100% of the tumor cells. The staining pattern showed characteristics of homogeneity more than heterogeneity. Consequently, we expect that evaluation of the mutation status by IHC should be reliable using small biopsy specimens or tissue microarray.
Our study also showed that all of the normal alveolar epithelial cells were completely negative and the intensity of mutation-specific immunostaining was much fainter in tumor cells with a lepidic pattern comparing to other patterns. This demonstrates that the EGFR mutations are tumor-specific, and likely an initiating event in lung cancer tumorigenesis.
Immunohistochemistry using mutation-specific mAbs is demonstrated to be a reliable test for detecting EGFR mutations in adenocarcinoma of the lung in our study. The diagnostic power of EGFR mutation-specific IHC is influenced by the antigen retrieval and scoring methods. Based on our study EDTA (pH 8.0) is better than sodium citrate (pH 6.0) and EDTA (pH 9.0) as the antigen retrieval buffer. A practical and reliable scoring method, i.e. positive is interpreted as moderate to strong staining of membrane and/or cytoplasm in >10% tumor cells, is recommended. However its final validation depends on strict quality control of the whole protocol, including antibody manufacture, IHC method, scoring system, criteria for interpretation, and the proper way to integrate with molecular methods, etc.
The specificity of EGFR mutation-specific IHC was very high, 100% for exon 19 deletions and 97% for L858R, while sensitivity was relatively lower, 81% for L858R and 59% for Exon 19 deletions. Considering the use of IHC has the advantage of being a method routinely applied in solid tumor diagnosis in pathology, EGFR mutation-specific IHC could be used as a prescreening method for selecting EGFR-TKI candidates. The positive cases by IHC could be selected as candidates for EGFR-TKI, while negative cases should be referred for DNA analysis. Additionally, as the staining pattern showed characteristics of homogeneity more than heterogeneity, it should be reliable to evaluate the mutation status of biopsy specimens or tissue microarray using IHC. Furthermore, it may be possible to use IHC as a substitute when the quantity of the sample DNA is not sufficient for molecular methods, e.g., small tissue samples or individual cells obtained from body fluids, bronchial washings, and fine needle aspirates etc.
Written informed consent was obtained from the patient for publication of this report and any accompanying images.
Yan Xiong, MD, Associated Professor in Department of Pathology, Peking University First Hospital, Beijing, China. Ting Li, MD, Full professor and Chair in Department of Pathology, Peking University First Hospital, Beijing, China.
Amplification refractory mutation system
Clamps for exons 19
Clamps for exons 21
Epidermal growth factor
Formalin-fixed paraffin- embedded
Locked nucleic acid
Non-small cell lung cancer
Negative predictive value
Positive predictive value
Tyrosine kinase inhibitor.
We thank Ying Wang, from department of pathology, Peking University First Hospital, for cutting numbers of slides.
- Tanaka T, Matsuoka M, Sutani A, Gemma A, Maemondo M, Inoue A, Okinaga S, Nagashima M, Oizumi S, Uematsu K: Frequency of and variables associated with the EGFR mutation and its subtypes. Int J Cancer. 2010, 126 (3): 651-655. 10.1002/ijc.24746.View ArticlePubMed
- Cheng L, Alexander RE, Maclennan GT, Cummings OW, Montironi R, Lopez-Beltran A, Cramer HM, Davidson DD, Zhang S: Molecular pathology of lung cancer: key to personalized medicine. Mod Pathol. 2012, 25 (3): 347-369. 10.1038/modpathol.2011.215.View ArticlePubMed
- Bonanno L, Schiavon M, Nardo G, Bertorelle R, Bonaldi L, Galligioni A, Indraccolo S, Pasello G, Rea F, Favaretto A: Prognostic and predictive implications of EGFR mutations, EGFR copy number and KRAS mutations in advanced stage lung adenocarcinoma. Anticancer Res. 2010, 30 (12): 5121-5128.PubMed
- Casorzo L, Corigliano M, Ferrero P, Venesio T, Risio M: Evaluation of 7q31 region improves the accuracy of EGFR FISH assay in non small cell lung cancer. Diagn Pathol. 2009, 4: 36-10.1186/1746-1596-4-36.PubMed CentralView ArticlePubMed
- da Cunha SG, Shepherd FA, Tsao MS: EGFR mutations and lung cancer. Annu Rev Pathol. 2011, 6: 49-69. 10.1146/annurev-pathol-011110-130206.View Article
- Pirker R, Herth FJ, Kerr KM, Filipits M, Taron M, Gandara D, Hirsch FR, Grunenwald D, Popper H, Smit E: Consensus for EGFR mutation testing in non-small cell lung cancer: results from a European workshop. J Thorac Oncol. 2010, 5 (10): 1706-1713. 10.1097/JTO.0b013e3181f1c8de.View ArticlePubMed
- Yu J, Kane S, Wu J, Benedettini E, Li D, Reeves C, Innocenti G, Wetzel R, Crosby K, Becker A: Mutation-specific antibodies for the detection of EGFR mutations in non-small-cell lung cancer. Clin Cancer Res. 2009, 15 (9): 3023-3028. 10.1158/1078-0432.CCR-08-2739.View ArticlePubMed
- Simonetti S, Molina MA, Queralt C, de Aguirre I, Mayo C, Bertran-Alamillo J, Sanchez JJ, Gonzalez-Larriba JL, Jimenez U, Isla D: Detection of EGFR mutations with mutation-specific antibodies in stage IV non-small-cell lung cancer. J Transl Med. 2010, 8: 135-10.1186/1479-5876-8-135.PubMed CentralView ArticlePubMed
- Kato Y, Peled N, Wynes MW, Yoshida K, Pardo M, Mascaux C, Ohira T, Tsuboi M, Matsubayashi J, Nagao T: Novel epidermal growth factor receptor mutation-specific antibodies for non-small cell lung cancer: immunohistochemistry as a possible screening method for epidermal growth factor receptor mutations. J Thorac Oncol. 2010, 5 (10): 1551-1558. 10.1097/JTO.0b013e3181e9da60.PubMed CentralView ArticlePubMed
- Kitamura A, Hosoda W, Sasaki E, Mitsudomi T, Yatabe Y: Immunohistochemical detection of EGFR mutation using mutation-specific antibodies in lung cancer. Clin Cancer Res. 2010, 16 (13): 3349-3355. 10.1158/1078-0432.CCR-10-0129.View ArticlePubMed
- Brevet M, Arcila M, Ladanyi M: Assessment of EGFR mutation status in lung adenocarcinoma by immunohistochemistry using antibodies specific to the two major forms of mutant EGFR. J Mol Diagn. 2010, 12 (2): 169-176. 10.2353/jmoldx.2010.090140.PubMed CentralView ArticlePubMed
- Wu SG, Chang YL, Lin JW, Wu CT, Chen HY, Tsai MF, Lee YC, Yu CJ, Shih JY: Including total EGFR staining in scoring improves EGFR mutations detection by mutation-specific antibodies and EGFR TKIs response prediction. PLoS One. 2011, 6 (8): e23303-10.1371/journal.pone.0023303.PubMed CentralView ArticlePubMed
- Kozu Y, Tsuta K, Kohno T, Sekine I, Yoshida A, Watanabe S, Tamura T, Yokota J, Suzuki K, Asamura H: The usefulness of mutation-specific antibodies in detecting epidermal growth factor receptor mutations and in predicting response to tyrosine kinase inhibitor therapy in lung adenocarcinoma. Lung Cancer. 2011, 73 (1): 45-50. 10.1016/j.lungcan.2010.11.003.View ArticlePubMed
- Hofman P, Ilie M, Hofman V, Roux S, Valent A, Bernheim A, Alifano M, Leroy-Ladurie F, Vaylet F, Rouquette I: Immunohistochemistry to identify EGFR mutations or ALK rearrangements in patients with lung adenocarcinoma. Ann Oncol. 2012, 23 (7): 1738-1743. 10.1093/annonc/mdr535.View ArticlePubMed
- Kawahara A, Yamamoto C, Nakashima K, Azuma K, Hattori S, Kashihara M, Aizawa H, Basaki Y, Kuwano M, Kage M: Molecular diagnosis of activating EGFR mutations in non-small cell lung cancer using mutation-specific antibodies for immunohistochemical analysis. Clin Cancer Res. 2010, 16 (12): 3163-3170. 10.1158/1078-0432.CCR-09-3239.View ArticlePubMed
- Lamy PJ, Jacot W: Worldwide variations in EGFR somatic mutations: a challenge for personalized medicine. Diagn Pathol. 2012, 7: 13-10.1186/1746-1596-7-13.PubMed CentralView ArticlePubMed
- Otto C, Csanadi A, Fisch P, Werner M, Kayser G: Molecular modeling and description of a newly characterized activating mutation of the EGFR gene in non-small cell lung cancer. Diagn Pathol. 2012, 7: 146-10.1186/1746-1596-7-146.PubMed CentralView ArticlePubMed
- Mok TS, Wu YL, Thongprasert S, Yang CH, Chu DT, Saijo N, Sunpaweravong P, Han B, Margono B, Ichinose Y: Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma. N Engl J Med. 2009, 361 (10): 947-957. 10.1056/NEJMoa0810699.View ArticlePubMed
- Savic SGB, Rufle A, Bihl M, Barascud A, Herzog M, Tapia C, Terracciano L, Feichter G, Bubendorf L: EGFR-gene analysis on cytological specimens of non-small-cell lung cancers. Diagn Pathol. 2007, 2 (Suppl 1): 1-10.1186/1746-1596-2-S1-S1.View Article
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.