Abstract
To identify the characteristics and sequence of epidermal growth factor receptor (EGFR) abnormalities relevant to the pathogenesis and progression of lung adenocarcinoma, we performed a precise mapping analysis of EGFR mutation, gene copy number, and total and phosphorylated EGFR protein expression for the same tissue sites. We examined normal bronchial and bronchiolar epithelium (NBE) and tumor tissues obtained from 50 formalin-fixed lung adenocarcinomas, including 24 EGFR-mutant primary tumors with nine corresponding lymph node metastases and 26 wild-type primary tumors. NBE in 12 of 24 (50%) mutant and 3 of 26 (12%) wild-type tumors harbored EGFR mutations; these NBE also showed a lack of EGFR copy number increase and frequent EGFR (69%) and phosphorylated EGFR (33%) overexpression. EGFR mutation and protein overexpression were more frequent in NBE sites within tumors than in NBE sites adjacent to and distant from tumors, suggesting a localized field effect. Sites with high and low EGFR copy numbers were heterogeneously distributed in six of nine primary tumors and in one of eight metastases. EGFR protein overexpression was significantly higher in metastasis sites than in primary tumors. We conclude from our findings that EGFR mutations and protein overexpression are early phenomena in the pathogenesis of lung adenocarcinoma and that EGFR mutation precedes an increase in gene copy number. In EGFR-mutant adenocarcinoma metastases, the higher levels of EGFR overexpression and more homogeneously distributed high gene copy numbers suggest tumor progression. Our findings have important implications for the development of new strategies for targeted chemoprevention and therapy in lung adenocarcinoma using EGFR inhibitors.
Epidermal growth factor receptor (EGFR), a tyrosine kinase (TK) member of the ErbB family, has shown frequent abnormalities in non–small cell lung carcinomas. These abnormalities include protein overexpression, gene amplification, and mutation (1–3). Somatic EGFR mutations have been identified in specific subsets of patients with lung adenocarcinoma, including never or light smokers, women, and patients of East Asian descent (4). The mutations cluster in the first four exons (18–21) of the TK domain of the gene, and ∼90% of the mutations are composed of either an in-frame deletion in exon 19 or a specific missense mutation in exon 21 (4). An increase in EGFR gene copy number, including high polysomy and gene amplification shown by fluorescent in situ hybridization (FISH), has been detected in 22% of patients with surgically resected (stages I-IIIA) non–small cell lung carcinomas and correlated with EGFR protein overexpression (2). Higher frequencies (40-50%) of EGFR high copy number have been reported in patients with advanced non–small cell lung carcinomas (5–10). Despite this knowledge, limited information is available on the role of EGFR abnormalities in the early pathogenesis and progression of lung adenocarcinomas.
Recently, we showed that mutation of the EGFR TK domain is an early event in the pathogenesis of lung adenocarcinoma and is detected in histologically normal bronchial and bronchiolar epithelium (NBE) in 43% of patients with EGFR-mutant tumors (11). We found that EGFR mutations were more frequent in normal epithelium within the tumor (43%) than in adjacent sites (24%), suggesting a localized field effect (11). However, no comprehensive information is available regarding the role of EGFR abnormalities, including gene mutation, increased copy number, and protein overexpression in the early pathogenesis and progression of lung adenocarcinomas.
Both EGFR gene mutations and high copy number (gene amplification and high polysomy identified by FISH) have been associated with sensitivity to the small-molecule TK inhibitors gefitinib and erlotinib in patients with lung adenocarcinoma (5–18). However, some of these results have been rather controversial (9, 10, 19, 20). In these studies of gefitinib and erlotinib, most of the EGFR mutation and copy number analyses were done in very small tissue samples or in cytologic specimens obtained from primary tumor and metastasis sites in patients with advanced-stage lung cancer (5–9, 12–16). To date, no studies have been done to identify the characteristics of EGFR gene and protein expression abnormalities at different sites with respect to primary lung adenocarcinomas and in corresponding sites of metastasis, information that might resolve some of the controversy.
To identify the sequence of EGFR abnormalities involved in the pathogenesis and progression of lung adenocarcinoma, we did a precise mapping analysis correlating EGFR mutation, gene copy number, and protein expression in NBE fields, primary tumors, and corresponding lymph node metastases that were obtained from 50 patients with lung adenocarcinomas, including 24 patients with EGFR-mutant primary tumors with nine corresponding lymph node metastasis sites and 26 patients with EGFR–wild-type primary tumors.
Materials and Methods
Case selection
To map EGFR gene and protein expression abnormalities, we obtained formalin-fixed, paraffin-embedded lung adenocarcinoma tissue specimens from the Lung Cancer Specialized Program of Research Excellence Tissue Bank at The University of Texas M. D. Anderson Cancer Center (Houston, TX). The tumor tissue specimens came from 50 patients with surgically resected lung adenocarcinomas (tumor-node-metastasis stage I-IIIA) with known EGFR mutations in exons 18 to 21, as described previously (3, 11). This bank was approved by the M. D. Anderson Cancer Center Institutional Review Board.
Of these 50 patients, 24 patients had lung adenocarcinoma with EGFR mutations in exon 18 (n = 1), exon 19 (n = 13), and exon 21 (n = 10), and 26 patients had EGFR–wild-type lung adenocarcinoma. The patients' clinicopathologic features are summarized in Table 1. All lung adenocarcinomas were of mixed histologic subtype (WHO classification; ref. 21). None of the patients had received cytotoxic and/or targeted therapy. Clinical staging was based on the revised International System for Staging Lung Cancer (22).
Clinicopathologic features of patients with lung adenocarcinomas examined for EGFR abnormalities in tumors and adjacent normal epithelium
Features/samples . | EGFR status . | . | . | |||
---|---|---|---|---|---|---|
. | Mutant (n = 24) . | Wild-type (n = 26) . | Total (n = 50) . | |||
Mean age (y) | 61.3 | 62.7 | 62.1 | |||
Gender | ||||||
Female | 19 (79%) | 13 (50%) | 32 | |||
Male | 5 (21%) | 13 (50%) | 18 | |||
Ethnicity | ||||||
East Asian | 13 (54%) | 9 (35%) | 22 | |||
Not East Asian | 11 (56%) | 17 (65%) | 28 | |||
Smoking history | ||||||
Never | 16 (67%) | 9 (35%) | 25 | |||
Former | 7 (29%) | 10 (38%) | 17 | |||
Current | 1 (4%) | 7 (27%) | 8 | |||
Stage of disease | ||||||
I | 11 (46%) | 15 (58%) | 26 | |||
II | 5 (21%) | 4 (15%) | 9 | |||
IIIA | 8 (33%) | 7 (27%) | 15 |
Features/samples . | EGFR status . | . | . | |||
---|---|---|---|---|---|---|
. | Mutant (n = 24) . | Wild-type (n = 26) . | Total (n = 50) . | |||
Mean age (y) | 61.3 | 62.7 | 62.1 | |||
Gender | ||||||
Female | 19 (79%) | 13 (50%) | 32 | |||
Male | 5 (21%) | 13 (50%) | 18 | |||
Ethnicity | ||||||
East Asian | 13 (54%) | 9 (35%) | 22 | |||
Not East Asian | 11 (56%) | 17 (65%) | 28 | |||
Smoking history | ||||||
Never | 16 (67%) | 9 (35%) | 25 | |||
Former | 7 (29%) | 10 (38%) | 17 | |||
Current | 1 (4%) | 7 (27%) | 8 | |||
Stage of disease | ||||||
I | 11 (46%) | 15 (58%) | 26 | |||
II | 5 (21%) | 4 (15%) | 9 | |||
IIIA | 8 (33%) | 7 (27%) | 15 |
EGFR abnormality mapping
We retrospectively reviewed H&E-stained histology sections of primary tumor, lymph node metastases, and adjacent normal lung tissue specimens to identify tissue foci available for EGFR abnormality analyses. The EGFR abnormalities included EGFR mutations in exons 18 and 21, as shown by microdissection and PCR-based sequencing; EGFR copy number, as shown by FISH; and total EGFR and phosphorylated EGFR (pEGFR), as shown by immunohistochemical analyses. Representative examples of these molecular changes are illustrated in Fig. 1.
A representative case of EGFR-mutant lung adenocarcinoma: EGFR gene and protein expression abnormalities in NBE (A–E), primary tumor (F–J), and lymph node metastasis (K–O) sites. Histologic characteristics (A, F, and K) of tissue sections stained with H&E (magnification, ×100). PCR-based EGFR sequencing (B, G and L) of the same EGFR mutation in exon 21 (L858R, black arrowhead) in NBE (B), primary tumor (G), and lymph node metastasis sites (L). EGFR FISH analysis (C, H and M) of low trisomy (low copy number) in the NBE sample (C), high polysomy in the primary tumor site (H), and gene amplification (M) in the metastasis site. Immunohistochemical analysis (D, I, N, E, J, and O) of high EGFR and pEGFR expression in the membrane and cytoplasm in all three types of samples.
A representative case of EGFR-mutant lung adenocarcinoma: EGFR gene and protein expression abnormalities in NBE (A–E), primary tumor (F–J), and lymph node metastasis (K–O) sites. Histologic characteristics (A, F, and K) of tissue sections stained with H&E (magnification, ×100). PCR-based EGFR sequencing (B, G and L) of the same EGFR mutation in exon 21 (L858R, black arrowhead) in NBE (B), primary tumor (G), and lymph node metastasis sites (L). EGFR FISH analysis (C, H and M) of low trisomy (low copy number) in the NBE sample (C), high polysomy in the primary tumor site (H), and gene amplification (M) in the metastasis site. Immunohistochemical analysis (D, I, N, E, J, and O) of high EGFR and pEGFR expression in the membrane and cytoplasm in all three types of samples.
We used serial 5-μm-thick histology sections for the tissue microdissection, FISH, and immunohistochemical analyses. We identified a total of 316 noncontiguous tumor and epithelial foci from among 142 NBE specimens (obtained from 50 patients; 2.84 sites/patient), 144 primary tumors (from 50 patients; 2.88 sites/patient), and 30 lymph node metastases (from 9 patients; 3.3 sites/patient). We examined NBE and primary tumors in both EGFR-mutant and EGFR–wild-type cases and metastasis sites in EGFR-mutant cases only. All epithelial foci consisted of normal or mildly hyperplastic epithelia that harbored small bronchi (65 sites) and bronchioles (77 sites).
The NBE specimens were obtained from three different locations based on their relationship to the tumors: within the tumor (47 sites), ≤5 mm from the tumor margin (adjacent to tumor; 63 sites), and >5 mm from the tumor margin (“distant” lung; 32 sites). We did not detect squamous metaplastic or dysplastic lesions in the bronchial structures or atypical adenomatous hyperplasias in the alveolar tissue. We identified small bronchi on the basis of well-defined smooth muscle and discontinuous cartilage layers. Bronchioles were defined as small conducting airways lacking well-defined smooth muscle wall or cartilage layers. We assessed the location of the small bronchial and bronchiolar respiratory epithelium examined for EGFR abnormalities based on the epithelia's location in relation to the tumor tissue in the corresponding histology sections, as previously described (11).
Microdissection and DNA extraction
Approximately 1,000 cells were precisely microdissected from 8-μm-thick, H&E-stained, formalin-fixed, paraffin-embedded histology sections for each site using laser capture microdissection (Arcturus Engineering Laser Capture Microdissection System; MDS Analytical Technologies), as previously described (11). To prevent the nonspecific binding of the mutant cells to the microdissection cap film, the microdissected tissue samples were redissected from the film under stereomicroscope visualization using fine needles (25-gauge 5/8-inch needles). We then extracted the DNA using 25 μL of PicoPure DNA Extraction solution containing proteinase K and incubated the DNA at 65°C for 20 h. Subsequently, proteinase K was inactivated by heating samples at 95°C for 10 min.
EGFR mutation analysis
Exons 18 and 21 of EGFR were PCR-amplified using DNA extracted from microdissected NBE and tumor cells, as previously described (3, 11). Each PCR was done using HotStarTaq Master Mix (Qiagen) for 40 cycles at 94°C for 30 s, 63°C for 30 s, and 72°C for 30 s, followed by a 7-min extension at 72°C. PCR products were directly sequenced using the Applied Biosystems PRISM dye terminator cycle sequencing method (Perkin-Elmer Corp.). We confirmed all sequence variants by independent PCR amplifications from at least two independent microdissections and sequenced the variants in both directions.
EGFR FISH analysis
We analyzed the gene copy number per cell using the LSI EGFR SpectrumOrange/CEP 7 SpectrumGreen Probe (Abbott Molecular), as previously described (5). Histology sections were incubated at 56°C overnight and deparaffinized by washing in CitriSolv (Fisher Scientific). After incubation in 2× SSC buffer (pH 7.0) at 75°C for 15 to 25 min, the histology sections were digested with proteinase K (0.25 mg/mL in 2× SSC) at 37°C for 15 to 25 min, rinsed in 2× SSC (pH 7.0) at room temperature for 5 min, and dehydrated using ethanol in a series of increasing concentrations (70%, 85%, 100%). We applied the EGFR SpectrumOrange/CEP 7/SpectrumGreen probe set (Abbott Molecular) onto the selected area, according to the manufacturer's instructions, on the basis of the tumor foci seen on each slide. We then covered the hybridization area with a glass coverslip and sealed the coverslip with rubber cement. The slides were incubated at 80°C for 10 min for codenaturation of chromosomal and probe DNA and then placed in a humidified chamber at 37°C for 20 to 24 h to allow hybridization to occur. Posthybridization washes were done in 1.5 mol/L of urea and 0.1× SSC (pH 7.0-7.5) at 45°C for 30 min and in 2× SSC for 2 min at room temperature. After the samples were dehydrated in a series of increasing ethanol concentrations, 4′,6′-diamidino-2-phenylindole (0.15 mg/mL in Vectashield Mounting Medium; Vector Laboratories) was applied for chromatin counterstaining. FISH analysis was done independently by two authors (M. Varella-Garcia and A.C. Xavier), who were blinded to the patients' clinical characteristics and all other molecular variables. Patients were classified into six FISH strata according to the frequency of cells with the EGFR gene copy number and referred to the chromosome 7 centromere, as follows: (a) disomy ( ≥3 copies in <10% of cells); (b) low trisomy (3 copies in 10% to 40% of the cells, ≥4 copies in <10% of cells); (c) high trisomy (3 copies in ≥40% of cells, ≥4 copies in <10% of cells); (d) low polysomy (≥4 copies in 10–40% of cells); (e) high polysomy (≥4 copies in ≥40% of cells); and (f) gene amplification (ratio of EGFR gene to chromosome ≥2, presence of tight EGFR gene clusters and 15 copies of EGFR per cell in 10% of the analyzed cells). The high polysomy and gene amplification categories were considered to be high EGFR copy number, and the other categories were considered to be nonincreased EGFR copy number, as previously published (5). Analysis was done in approximately 50 nuclei per tumor and epithelial site, and the section of the area was guided by image captured in the H&E-stained section.
Immunohistochemical staining
Tissue histology sections for immunohistochemical analyses were deparaffinized, hydrated, heated in a steamer for 10 min with 10 mmol/L of sodium citrate (pH 6.0) for antigen retrieval, and washed in Tris buffer. Peroxide blocking was done with 3% H2O2 in methanol at room temperature for 15 min, followed by 10% bovine serum albumin in TBS with Tween 20 for 30 min at room temperature. For the EGFR analysis, tissue sections were incubated for 2 h with primary antibodies against the EGFR clone 31G7 (1:100 dilution; Zymed) and pEGFR Tyr 1086 (1:100 dilution; Invitrogen). Tissue sections were then incubated for 30 min with the secondary antibody (EnVision+ Dual Link labeled polymer; DAKO), after which diaminobenzidine chromogen was applied for 5 min. The slides were then counterstained with hematoxylin and topped with a coverslip. For EGFR and pEGFR expression, antibody specificity was confirmed using blocking peptide and phosphatase incubation experiments. For the control experiments, we used formalin-fixed and paraffin-embedded pellets from lung cancer cell lines with confirmed EGFR and pEGFR overexpression. Thyroid transcription factor-1 (TITF-1) antibody (1:100 dilution, Cell Marque) was used for the identification of TITF-1–positive cells. All four antibodies were incubated for 1.5 h at room temperature. Immunohistochemistry results were scored jointly by two authors (X. Tang and I.I. Wistuba), who were blinded to clinical and other molecular variables. Immunostaining of the cell membrane and cytoplasm for EGFR and pEGFR was evaluated by light microscopy (magnification, ×20). A semiquantitative approach was used to generate a score for each tissue site, as previously described (2, 23, 24). Membrane and cytoplasm stains were recorded separately. We defined the intensity score as follows: 0, no appreciable staining in the NBE or malignant cells; 1, barely detectable staining in NBE or malignant cells compared with the stromal elements; 2, readily appreciable staining; 3, dark brown staining of cells; and 4, very strong staining of cells. The score was also based on the fraction of cells showing a given staining intensity (0-100%). We calculated the immunohistochemical scores by multiplying the intensity and extension, and the scores ranged from 0 to 400. For the statistical analyses, scores of 0 to 200 signified negative/low expression, and scores >200 indicated positive/overexpression, as previously reported (2, 23, 24). For the evaluation of nuclear TITF-1 immunohistochemical expression, 200 epithelial cells were quantified by light microscopy (magnification, ×20), and a score (range, 0-100) expressing the percentage of positive cells was obtained.
Statistical analysis
All relationships between categorical variables were assessed using χ2 and Fisher's exact tests. P < 0.05 values were considered statistically significant.
Results
EGFR abnormalities in the early pathogenesis of lung adenocarcinomas
Patterns of EGFR mutation in NBE. We previously reported our finding of mutations in exons 19 and 21 of EGFR in at least one site of microdissected NBE obtained from lung cancer specimens from 9 of 21 (43%) patients with EGFR-mutant adenocarcinomas, with no such mutations found in any of 26 respiratory epithelium foci from 16 patients with wild-type tumors (11). In the present study, using the same methodology, we analyzed for EGFR mutation in NBE obtained from an additional 3 patients with an EGFR-mutant and 10 patients with EGFR–wild-type lung adenocarcinomas. Combining both data sets, the overall rate of mutation in NBE from EGFR-mutant tumors was 50%. In the wild-type tumor cases, we detected EGFR exon 19 deletions (15 bp, 746-750) in six sites of small bronchial (n = 4) and bronchiolar (n = 2) NBE obtained from three wild-type tumors (Table 2). Thus, an EGFR mutation was found in NBE in 3 of 26 (12%) wild-type adenocarcinomas and in 8 of 57 (14%) of the microdissected epithelial sites (Table 2).
Frequency of EGFR gene mutation and protein overexpression in histologically normal bronchial and bronchiolar epithelium obtained from EGFR-mutant and wild-type lung adenocarcinomas
EGFR abnormality in NBE . | Cases . | . | . | Sites . | . | . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
. | Mutant . | Wild-type . | Total . | Mutant . | Wild-type . | Total . | ||||||
Mutation by sequencing | ||||||||||||
Number | 24 | 26 | 50 | 85 | 57 | 142 | ||||||
Mutant | 12 (50%)* | 3 (12%)* | 15 (30%) | 22 (26%) | 8 (14%) | 30 (21%) | ||||||
Protein overexpression by immunohistochemistry† | ||||||||||||
Number | 23 | 26 | 49 | 78 | 56 | 134 | ||||||
EGFR | 19 (83%) | 15 (58%) | 34 (69%) | 52 (67%) | 35 (63%) | 87 (65%) | ||||||
pEGFR | 10 (44%) | 6 (23%) | 16 (33%) | 24 (31%) | 12 (21%) | 36 (27%) |
EGFR abnormality in NBE . | Cases . | . | . | Sites . | . | . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
. | Mutant . | Wild-type . | Total . | Mutant . | Wild-type . | Total . | ||||||
Mutation by sequencing | ||||||||||||
Number | 24 | 26 | 50 | 85 | 57 | 142 | ||||||
Mutant | 12 (50%)* | 3 (12%)* | 15 (30%) | 22 (26%) | 8 (14%) | 30 (21%) | ||||||
Protein overexpression by immunohistochemistry† | ||||||||||||
Number | 23 | 26 | 49 | 78 | 56 | 134 | ||||||
EGFR | 19 (83%) | 15 (58%) | 34 (69%) | 52 (67%) | 35 (63%) | 87 (65%) | ||||||
pEGFR | 10 (44%) | 6 (23%) | 16 (33%) | 24 (31%) | 12 (21%) | 36 (27%) |
*P = 0.003.
†Positive immunohistochemical overexpression score >200 (range, 0-400).
The combined data showed that NBE with mutant EGFR was detected in the small bronchi (13 of 64, 20%) and bronchioles (17 of 78, 22%) of both mutant and wild-type tumor cases. Overall, however, the mutation frequency was higher in NBE samples microdissected from within the tumor (13 of 47, 28%) than in samples obtained from adjacent tissue and tissue distant from the tumors (17 of 95, 18%; Table 3).
EGFR mutation and protein overexpression in histologically normal epithelium by location
EGFR abnormality in NBE . | Location in relation to the tumor . | . | . | Structure . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|
. | Inside . | Adjacent . | Distant . | Bronchiole . | Small bronchus . | |||||
Mutation | ||||||||||
Mutant tumor | 11/31 (36%)* | 10/35 (29%) | 1/17 (6%)* | 10/43 (23%) | 12/42 (29%) | |||||
Wild-type tumor | 2/15 (13%) | 3/28 (11%) | 1/15 (7%) | 4/34 (12%) | 2/23 (9%) | |||||
All tumors | 13/46 (28%) | 13/63 (21%) | 2/32 (6%) | 14/77 (19%) | 14/65 (22%) | |||||
EGFR overexpression† | ||||||||||
Mutant tumor | 24/29 (83%)‡ | 20/33 (61%)‡ | 8/16 (50%)‡ | 18/38 (47%)§ | 34/40 (85%)§ | |||||
Wild-type tumor | 10/15 (67%) | 17/28 (61%) | 8/13 (62%) | 14/33 (42%)§ | 21/23 (91%)§ | |||||
All tumors | 34/44 (77%) | 37/61 (61%) | 16/29 (55%) | 32/71 (45%)§ | 55/63 (87%)§ | |||||
pEGFR overexpression† | ||||||||||
Mutant tumor | 13/29 (45%)∥ | 5/33 (15%)∥ | 6/16 (38%)∥ | 10/38 (26%) | 14/40 (35%) | |||||
Wild-type tumor | 5/15 (33%) | 5/28 (18%) | 2/13 (15%) | 2/33 (6%)§ | 10/23 (44%)§ | |||||
All tumors | 18/44 (41%) | 10/61 (16%) | 8/29 (28%) | 12/71 (17%)¶ | 24/63 (38%)¶ |
EGFR abnormality in NBE . | Location in relation to the tumor . | . | . | Structure . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|
. | Inside . | Adjacent . | Distant . | Bronchiole . | Small bronchus . | |||||
Mutation | ||||||||||
Mutant tumor | 11/31 (36%)* | 10/35 (29%) | 1/17 (6%)* | 10/43 (23%) | 12/42 (29%) | |||||
Wild-type tumor | 2/15 (13%) | 3/28 (11%) | 1/15 (7%) | 4/34 (12%) | 2/23 (9%) | |||||
All tumors | 13/46 (28%) | 13/63 (21%) | 2/32 (6%) | 14/77 (19%) | 14/65 (22%) | |||||
EGFR overexpression† | ||||||||||
Mutant tumor | 24/29 (83%)‡ | 20/33 (61%)‡ | 8/16 (50%)‡ | 18/38 (47%)§ | 34/40 (85%)§ | |||||
Wild-type tumor | 10/15 (67%) | 17/28 (61%) | 8/13 (62%) | 14/33 (42%)§ | 21/23 (91%)§ | |||||
All tumors | 34/44 (77%) | 37/61 (61%) | 16/29 (55%) | 32/71 (45%)§ | 55/63 (87%)§ | |||||
pEGFR overexpression† | ||||||||||
Mutant tumor | 13/29 (45%)∥ | 5/33 (15%)∥ | 6/16 (38%)∥ | 10/38 (26%) | 14/40 (35%) | |||||
Wild-type tumor | 5/15 (33%) | 5/28 (18%) | 2/13 (15%) | 2/33 (6%)§ | 10/23 (44%)§ | |||||
All tumors | 18/44 (41%) | 10/61 (16%) | 8/29 (28%) | 12/71 (17%)¶ | 24/63 (38%)¶ |
*Comparison of NBE from inside tumor vs. NBE distant (P = 0.02).
†Positive immunohistochemical overexpression score >200 (range 0-400).
‡Comparison of NBE from inside tumor vs. NBE adjacent + distant (P = 0.02)
§Comparison of NBE from bronchiole vs. small bronchus (P < 0.001).
∥Comparison of NBE from inside tumor vs. NBE adjacent + distant (P = 0.038).
¶Comparison of NBE from bronchiole vs. small bronchus (P = 0.006).
In our previously reported comparison of NBE and corresponding tumors (16 specimens), we always observed identical EGFR mutations in both sites examined (11). In this study, we have expanded the number of NBE sites (n = 85) examined for the mutation in patients with EGFR-mutant adenocarcinomas and detected five sites (6%) from three cases in which NBE showed mutations different from the ones detected in the corresponding tumor specimens (data not shown). Importantly, in all cases with a mutation in NBE, an identical mutation was detected in at least one site of the corresponding tumor specimen. Thus, in this expansion of our previous study (11), a relatively more heterogeneous EGFR mutation pattern of the respiratory field was detected in NBE microdissected from mutant lung adenocarcinomas, but most NBE and corresponding tumors shared the same mutation.
EGFR copy number and correlation with gene mutation in NBE. To determine the morphologic stage at which EGFR copy abnormalities arise in EGFR-mutant adenocarcinomas, we did a precise mapping analysis and examined EGFR copy number in 21 NBE sites obtained from nine mutant adenocarcinomas using FISH. All nine tumor specimens showed at least one site with a high copy number. These epithelial sites were also examined in the EGFR mutation analysis. Most NBE (14 of 21, 67%) showed no EGFR FISH abnormalities (disomy), including four EGFR-mutant sites with exon 19 (15 bp) deletions and exon 21 (L858R) point mutations. Trisomy was detected in seven (33%) NBE sites obtained from six (67%) cases. We did not identify any NBE with EGFR amplifications or a high level of polysomy, which have been defined as high gene copy number. In contrast, the nine tumors mapped showed significantly higher frequency of EGFR amplification (11 of 42 sites, 26%; P < 0.018) or a high level of polysomy (22 of 42, 52%; P < 0.001) compared with NBE. Our findings indicate that high EGFR copy number does not occur in peripheral NBE in EGFR-mutant lung adenocarcinomas and that gene mutations precede copy number abnormalities in the sequential pathogenesis of these tumors.
EGFR immunohistochemical expression and correlation with gene mutation in NBE. We evaluated the level of EGFR and pEGFR protein expression in 134 NBEs obtained from EGFR-mutant and wild-type lung adenocarcinomas. Overall, a high level of EGFR (69%) and a moderate level of pEGFR (33%) expression were detected in NBE from patients with tumors (Table 2). However, EGFR and pEGFR were expressed to a greater degree in NBE sites obtained from patients with EGFR-mutant tumors than in patients with wild-type tumors (Table 2), although these differences were not statistically significant. The frequency of EGFR, but not of pEGFR, overexpression was higher in EGFR–wild-type NBE sites (85 of 111, 77%) than in mutant sites (14 of 24, 58%; P = 0.039). Of interest, NBE located inside tumors showed the highest frequency of EGFR and pEGFR overexpression compared with NBE located adjacent to and distant from tumors, especially in EGFR-mutant tumors (Table 3). Small bronchi also showed a higher frequency of overexpression of both markers compared with bronchioles (Table 3). Thus, the overexpression of EGFR and pEGFR is a common event in NBE from patients with lung adenocarcinomas, especially in EGFR-mutant tumors, and shows a localized field phenomenon effect similar to gene mutation.
TITF-1 immunohistochemical expression and EGFR mutation in NBE. Recently, on the basis of immunohistochemical findings of higher levels of nuclear TITF-1 expression, a crucial transcription factor of the lung, in EGFR-mutant lung adenocarcinomas compared with in wild-type tumors, it has been suggested that EGFR-mutant lung adenocarcinoma originates from the terminal respiratory unit (25), which is composed of alveolar cells and nonciliated bronchiolar epithelium. Its characteristics are highlighted by the expression of TITF-1 (25). We therefore investigated the correlation between EGFR mutation and TITF-1 nuclear expression in tumor and normal epithelium sites. EGFR-mutant lung adenocarcinomas (18 of 20 cases, 90%) showed higher expression of TITF-1 than did wild-type tumors (10 of 26 cases, 38%; P < 0.001). However, in immunohistochemical studies, we did not see a significant difference in the frequency of TITF-1 expression between EGFR-mutant (11 of 25 sites, 44%) and EGFR–wild-type (34 of 105 sites, 33%; P = 0.273) respiratory epithelia. Our findings therefore indicate that NBE cells expressing TITF-1 are not the exclusive precursors of EGFR-mutant adenocarcinomas. From these results, it is clear that these tumors do not originate exclusively from terminal respiratory unit structures.
EGFR abnormalities in the progression of lung adenocarcinomas
EGFR mutation pattern in primary tumors and corresponding metastasis. To identify the characteristics of EGFR abnormalities in the progression of mutant lung adenocarcinomas, we examined EGFR gene mutation, gene copy number, and protein expression in primary tumors and corresponding metastases by performing a detailed mapping analysis of tumor specimens. For this study, we selected nine lung adenocarcinomas with known EGFR mutations in exon 19 (n = 5) and exon 21 (n = 4), and with lymph node metastases for which there was sufficient tissue to perform our mapping analysis.
For the mutation analysis of EGFR exons 19 and 21, we did precise tissue microdissection from noncontiguous primary tumor foci (n = 56 sites, 6.2 sites/tumor; range 2-11 sites) containing at least 1,000 cells. Surprisingly, four of the nine primary tumors examined showed mixed EGFR gene patterns (Fig. 2A): three showed two or more types of mutations, and one showed five sites with exon 19 (15 bp, 746-750) deletion and two sites with the wild-type EGFR gene. EGFR mutation analysis of 30 corresponding lymph node metastasis sites from the nine EGFR-mutant cases (3.3 sites/case; range 1-6 sites) detected only one type of EGFR mutation in all tumor sites in each case, and the mutation was always present in at least one site of the corresponding primary tumor. Similar to the corresponding primary tumor, one metastasis case showed EGFR–wild-type (five sites) and EGFR-mutant [one site, exon 19 (15 bp, 746-750) deletion] tumor sites (Fig. 2A). All these findings were confirmed by sequencing analyses of independently microdissected samples. In summary, our findings showed a relatively high level of heterogeneity for the EGFR mutation, and several tumor cell clones had mutation patterns in the primary tumor specimens that differed from the mutation patterns in the lymph node metastasis sites.
A, EGFR mutation pattern in 56 primary tumor and 30 lymph node metastasis sites obtained from nine patients with EGFR-mutant lung adenocarcinomas. A homogeneous mutation pattern was detected in five primary tumors (cases 2, 3, 4, 7, and 9) and all but one (case 6) metastasis case. Case 6, mixed wild-type and mutant sites in both primary tumor sites and corresponding metastases. B, EGFR copy number pattern shown by FISH in 42 primary tumor and 29 lymph node metastasis sites obtained from nine patients with EGFR-mutant lung adenocarcinomas. Different FISH copy number categories (low vs. high) were found in six of nine primary tumors and in one of eight corresponding metastases. Positive EGFR FISH expression included high polysomy and gene amplification, and negative EGFR FISH expression included disomy and trisomy.
A, EGFR mutation pattern in 56 primary tumor and 30 lymph node metastasis sites obtained from nine patients with EGFR-mutant lung adenocarcinomas. A homogeneous mutation pattern was detected in five primary tumors (cases 2, 3, 4, 7, and 9) and all but one (case 6) metastasis case. Case 6, mixed wild-type and mutant sites in both primary tumor sites and corresponding metastases. B, EGFR copy number pattern shown by FISH in 42 primary tumor and 29 lymph node metastasis sites obtained from nine patients with EGFR-mutant lung adenocarcinomas. Different FISH copy number categories (low vs. high) were found in six of nine primary tumors and in one of eight corresponding metastases. Positive EGFR FISH expression included high polysomy and gene amplification, and negative EGFR FISH expression included disomy and trisomy.
EGFR copy number abnormalities in primary tumors and corresponding metastasis. We used FISH to investigate the EGFR gene copy number abnormalities in 42 primary tumor sites (2.1 sites/case; range 2-7 sites) and 29 metastasis sites (3.2sites/case; range 1-6 sites), which were also examined for the mutation analysis. Overall, all primary tumors and corresponding metastases showed at least one site of high gene copy number (high polysomy or gene amplification; Fig. 2B). However, six (67%) primary tumor cases and one (11%) metastasis case showed at least one site without high copy number (disomy in one primary tumor site, high trisomy in one metastasis site, and low polysomy in seven primary and three metastasis sites; Fig. 2B). Thus, EGFR copy number heterogeneity was higher in primary tumor sites than in corresponding metastasis sites.
EGFR immunohistochemical expression in primary tumors and corresponding metastasis sites. In the nine EGFR-mutant lung adenocarcinoma cases mapped for EGFR abnormalities, we examined both primary tumors and the corresponding lymph node metastases for EGFR and pEGFR immunohistochemical expression. For both tumor locations combined, 96 distinct tumor sites were examined (n = 65 primary tumor sites, 7.2 sites/case; and n = 31 metastasis sites, 3.4 sites/case). Significantly higher levels of EGFR and pEGFR expression were detected in metastasis sites compared with primary tumor sites (Table 4). No correlation between EGFR and pEGFR expression and EGFR copy number status by FISH was detected.
Summary of EGFR abnormalities by sites in nine primary lung adenocarcinomas and corresponding lymph node metastases
EGFR abnormality/number of sites . | Primary tumor . | Metastases . | ||
---|---|---|---|---|
Mutation | ||||
Number of sites examined | 56 | 30 | ||
Mutation positive | 54 (96%)* | 25 (83%)* | ||
Copy no. | ||||
Number of sites examined | 42 | 29 | ||
Low copy no. | 9 (21%) | 4 (14%) | ||
High copy no. | 33 (79%) | 25 (86%) | ||
High polysomy | 22 (52%) | 18 (62%) | ||
Gene amplification | 11 (26%) | 7 (24%) | ||
Protein overexpression† | ||||
Number of sites examined | 65 | 31 | ||
EGFR | 42 (65%)‡ | 30 (97%)‡ | ||
pEGFR | 9 (14%)§ | 21 (68%)§ |
EGFR abnormality/number of sites . | Primary tumor . | Metastases . | ||
---|---|---|---|---|
Mutation | ||||
Number of sites examined | 56 | 30 | ||
Mutation positive | 54 (96%)* | 25 (83%)* | ||
Copy no. | ||||
Number of sites examined | 42 | 29 | ||
Low copy no. | 9 (21%) | 4 (14%) | ||
High copy no. | 33 (79%) | 25 (86%) | ||
High polysomy | 22 (52%) | 18 (62%) | ||
Gene amplification | 11 (26%) | 7 (24%) | ||
Protein overexpression† | ||||
Number of sites examined | 65 | 31 | ||
EGFR | 42 (65%)‡ | 30 (97%)‡ | ||
pEGFR | 9 (14%)§ | 21 (68%)§ |
*The same case harbored two primary tumor and five metastasis sites with EGFR-wild-type sequence.
†Positive immunohistochemical expression score >200 (range 0-400).
‡Primary tumor vs. metastasis (P = 0.02).
§Primary tumor vs. metastasis (P = 0.00001).
Discussion
Using a detailed molecular pathology mapping strategy, we determined the sequence of EGFR abnormalities in the early pathogenesis of EGFR-mutant lung adenocarcinomas and identified the pattern of EGFR changes in the progression of EGFR-mutant lung adenocarcinomas from primary tumors to lymph node metastasis. First, we showed that EGFR mutations precede gene copy number abnormalities in the pathogenesis of these tumors and that EGFR and pEGFR immunohistochemical protein expressions are frequent events in histologically normal peripheral bronchial and bronchiolar epithelium adjacent to lung adenocarcinomas. Second, our data indicated that although primary lung adenocarcinomas show some degree of EGFR gene copy number heterogeneity, this phenomenon is rare in metastases. Although these findings can be considered tumor progression phenomena, they also have important clinical implications from the standpoint of making decisions regarding the use of EGFR TK inhibitor therapy on the basis of the finding of EGFR gene abnormalities.
Despite the evidence showing that atypical adenomatous hyperplasia is a precursor of peripheral lung adenocarcinomas (26), there is consensus that the pathogenesis of most adenocarcinomas is unknown. Our previously reported findings of an EGFR mutation in NBE in 9 of 21 (43%) patients with EGFR-mutant adenocarcinomas indicated that the EGFR gene mutation is an early event in the pathogenesis of lung adenocarcinoma (11). In this study, we have investigated normal epithelium from additional patients with EGFR-mutant or wild-type lung adenocarcinomas and specifically have two new findings in this study; (a) we detected an EGFR mutation (exon 19, 15 bp deletion, 746-751) in six sites of small bronchial and bronchiolar epithelium obtained from three patients with wild-type adenocarcinoma, and (b) whereas an identical mutation was detected in the majority of specimens of mutant normal epithelium compared with the corresponding invasive tumor (75% of cases and 94% of sites), we found few normal epithelium sites (6%) in three of 12 cases (25%) of EGFR-mutant tumors, demonstrating the existence of a different mutation pattern between normal epithelium and the corresponding invasive tumor. All these data reinforce the concept of a field effect phenomenon in EGFR mutations in lung adenocarcinoma pathogenesis that affects histologically normal bronchial and bronchiolar respiratory epithelia.
We have previously shown that molecular abnormalities occur in a stepwise fashion in the sequential pathogenesis of squamous cell carcinoma of the lung, with molecular changes commencing in histologically normal bronchial epithelium in smokers and in patients with lung cancer (27, 28). Our findings suggest that EGFR abnormalities also occur sequentially in the early pathogenesis of lung adenocarcinoma, with a mutation commencing in histologically normal epithelium and a high EGFR copy number appearing at the invasive tumor stage. A recent report (29) of selective gene amplification of the shorter allele of the EGFR intron 1 polymorphism CA simple sequence repeat 1, which is the allele more frequently mutated in tumors harboring an EGFR mutation, also suggests that mutations occur earlier than copy number abnormalities in the pathogenesis of lung adenocarcinoma. Our findings of frequent EGFR (69%) and pEGFR (33%) protein overexpression in normal distal bronchial and bronchiolar epithelium from patients with either EGFR-mutant or EGFR–wild-type lung adenocarcinomas indicate a field phenomenon in the peripheral airway. A relatively high frequency of EGFR protein expression has also been reported in centrally located, histologically normal (42%) and hyperplastic (54%) bronchial epithelium from smokers (23). In addition, our data indicate that the mechanisms of protein overexpression seem to be unassociated with high gene copy number and mutation in NBE. Other mechanisms can explain EGFR overexpression in normal epithelial cells, including ligand-dependent up-regulation and activation, as well as inhibition of endocytosis-related protein down-regulation in the cell membrane (30).
Based on findings of higher levels of immunohistochemical expression of nuclear TITF-1, a crucial transcription factor of the lung, in EGFR-mutant lung adenocarcinomas compared with wild-type tumors, it has been suggested that EGFR-mutant lung adenocarcinoma originates from the terminal respiratory unit (25). We found EGFR mutations in microdissected histologically normal epithelial cells from small bronchi and bronchioles, which supports the concept of adenocarcinomas arising from the peripheral lung airway. Our findings indicate that NBE cells expressing TITF-1 are not the exclusive precursors of EGFR-mutant adenocarcinomas. From this finding, it is clear that these tumors do not originate exclusively from terminal respiratory unit structures. In addition, we cannot exclude the possibility that common stem or progenitor cells for both bronchial and bronchiolar epithelium bear EGFR mutations.
It has been suggested that activating TK EGFR mutations are a potent oncogenic event by which mutant tumor cells become physiologically dependent on the continued activity of the phosphorylated protein for the maintenance of their malignant phenotype (31). Our detailed mapping analysis of the EGFR gene mutation and copy number of multiple precisely microdissected sites in nine mutant primary tumors and corresponding lymph node metastases showed an identical or monoclonal pattern of mutation in most (n = 5) primary tumors and all metastases. These findings corroborate the monoclonal concept of tumor development and the monoclonal evolution of metastases (32, 33). However, two primary tumors lacking identical or monoclonal EGFR-mutant patterns harbored different sizes of exon 19 deletions (12 versus 15 bp and 15 versus 18 bp deletions). This finding could be explained by a tumor progression phenomenon in which the deletion size changed during the evolution of the malignancy. However, two very interesting primary tumors in our study exhibited findings that challenged the concept of the monoclonal evolution of tumors. One case showed a single site with an exon 19 (15 bp) deletion, whereas the remaining eight sites lacked the deletion but showed a point mutation (TTA747CCA) in the same exon. Of interest, the three metastasis sites examined harbored the most frequent mutation detected in the primary lung tumor. The other case showed areas of wild-type and mutant EGFR in both primary tumors and metastases, a phenomenon that is difficult to explain and suggests that molecular events other than an EGFR mutation may be responsible for tumor development in lung adenocarcinomas. These findings were confirmed by the sequencing of independently microdissected samples. In the latter case, the finding of a high EGFR copy number (high polysomy) in wild-type tumor sites raises the possibility of an alternative explanation—that the wild-type allele is preferentially amplified in some tumor cells. As a result, the mutant allele is underrepresented and is not detectable by our current sequencing methodology.
Retrospective studies have provided data suggesting that a high EGFR gene copy number shown by FISH is associated with treatment response, time to progression, and survival in patients with advanced non–small cell lung carcinoma treated with EGFR TK inhibitors (5-7, 10, 17). In these studies, high EGFR copy number as shown by FISH was defined as true gene amplification or high polysomy with equal to or more than four EGFR copies in ≥40% of cells (5, 34). Our mapping analysis of primary tumors and corresponding lymph node metastases in which we used the same EGFR FISH criteria showed that a frequent high copy number in mutant tumors was the most frequent pattern detected. Despite the fact that most primary tumor sites and nearly all metastasis sites showed high copy numbers, high polysomy and gene amplification were heterogeneously distributed in both tumor locations. More importantly, five of nine (56%) primary tumors and one metastasis (13%) showed one or more sites without an increased copy number (FISH negative). Similarly, EGFR and pEGFR immunohistochemical expression was less heterogeneous in primary tumors and more frequent in metastases. Taken together, these data suggest that EGFR copy number analyzed by FISH and protein expression analyzed by immunohistochemistry in small core biopsy or fine-needle aspiration specimens obtained from primary tumors, and more rarely from metastases, could miss these molecular changes, especially if only a small number of malignant cells are available for examination. In addition, if the suggested presence of EGFR high copy number correlates with sensitivity to EGFR TK inhibitors (5–7, 17), it is likely that metastases will show a better response to therapy than will primary tumors. This is an important consideration, in light of the fact that the site of origin (primary versus metastasis) of the tumor specimen was not reported and factored into the biomarker analyses in any of the clinical trials testing the efficacy of EGFR TK inhibitors in patients with advanced non–small cell lung carcinoma in whom EGFR copy number determined by FISH was examined as a predictor of response and prognosis (5–7). Our results show that a better understanding of the pattern of molecular abnormalities and their corresponding biomarker expression, including primary tumors and the frequent metastases seen for this tumor type, is important in lung cancer.
In summary, our data suggest that gene mutations and protein overexpression are the earliest phenomena in EGFR-mutant lung adenocarcinoma, occurring at the NBE stage, and that this is followed by the development of a focal increase in copy number at the tumor stage (Fig. 3). At the metastasis sites, however, all three abnormalities were more frequent than they were in the primary tumors and were homogeneously distributed throughout the malignant cells.
Proposed sequence of EGFR abnormalities occurring in the early pathogenesis and progression of EGFR-mutant lung adenocarcinomas. NBE field, primary tumor, and metastasis sites. Small circles, NBE, which acquires EGFR mutations and EGFR protein (total and phosphorylated) overexpression (gray circles). In the primary tumor stage, the EGFR copy number increases (high polysomy and gene amplification) in small tumor foci (striped ovals). In the metastasis site, tumor cells show both EGFR mutation and high copy number throughout most of the lesion.
Proposed sequence of EGFR abnormalities occurring in the early pathogenesis and progression of EGFR-mutant lung adenocarcinomas. NBE field, primary tumor, and metastasis sites. Small circles, NBE, which acquires EGFR mutations and EGFR protein (total and phosphorylated) overexpression (gray circles). In the primary tumor stage, the EGFR copy number increases (high polysomy and gene amplification) in small tumor foci (striped ovals). In the metastasis site, tumor cells show both EGFR mutation and high copy number throughout most of the lesion.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.