Main

Epidermal growth factor receptor (EGFR) gene mutations have been detected frequently in lung adenocarcinomas, especially adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma component,1, 2, 3, 4, 5, 6, 7, 8, 9 and the mutations have been significantly associated with sensitivity to EGFR tyrosine kinase inhibitors such as gefitinib or erlotinib.10, 11 On the other hand, a subset of lung adenocarcinomas has mutations in the K-ras gene, a known downstream signaling molecule in the EGFR signaling pathway.12, 13 K-ras-mutated lung adenocarcinomas, as opposed to EGFR-mutated adenocarcinoma, often show a histopathological feature of mucinous bronchioloalveolar carcinoma,12 and K-ras mutations are found more frequently in patients who are resistant to therapy with EGFR tyrosine kinase inhibitors.14 EGFR and K-ras mutations have been reported to be mutually exclusive in lung adenocarcinomas, and be significantly more associated with never- and ever-smokers, respectively.4, 5, 6, 7, 9 Thus, EGFR- or K-ras-mutated adenocarcinomas are considered to be distinct subsets of lung adenocarcinomas.

Atypical adenomatous hyperplasia is considered to be a precursor lesion of peripheral lung adenocarcinoma; a multistep carcinogenesis where invasive lung adenocarcinoma develops from atypical adenomatous hyperplasia through bronchioloalveolar carcinoma (adenocarcinoma in situ) has been postulated.15 EGFR or K-ras gene mutation seems to have a critical role in the development of lung adenocarcinomas and to be an initiating event, as some atypical adenomatous hyperplasias or pure bronchioloalveolar carcinomas have been found to harbor EGFR or K-ras gene mutation,7, 8, 9, 13, 16, 17 and moreover in transgenic mouse models expressing mutated EGFR or K-ras gene in type II pneumocyte, development of lung adenocarcinomas through atypical adenomatous hyperplasias has been observed.18, 19 Although it is important to know the exact incidence of EGFR and K-ras gene mutations in atypical adenomatous hyperplasias to understand pathogenesis of lung adenocarcinoma, the results obtained so far are controversial.7, 8, 13, 16, 17 We have previously reported that most pure nonmucinous bronchioloalveolar carcinomas have EGFR mutations, whereas mucinous bronchioloalveolar carcinomas are significantly associated with K-ras mutations by means of loop-hybrid mobility shift assay (LH-MSA),9 which is a rapid and sensitive PCR-based method to detect hot spot mutations.20, 21 In the present study, we evaluated 54 atypical adenomatous hyperplasias of the lung for the hot spot mutations in exons 19 and 21 of EGFR gene and in codon 12 of K-ras gene using LH-MSA, (1) to elucidate the frequencies of EGFR and K-ras gene mutations in atypical adenomatous hyperplasias, then (2) to compare those between atypical adenomatous hyperplasias and adenocarcinomas, and finally (3) to discuss whether or not all atypical adenomatous hyperplasias progress to adenocarcinomas with a nonmucinous or mucinous bronchioloalveolar carcinoma component.

Materials and methods

Patients and Tumor Tissues

A total of 54 atypical adenomatous hyperplasias were obtained from 28 patients (Table 1). All the atypical adenomatous hyperplasias were diagnosed by three pathologists (YS, YK, and ET) according to the 2004 WHO classification system; ‘atypical adenomatous hyperplasia is a localized proliferation of mild to moderately atypical cells lining alveoli and sometimes respiratory bronchioles’.15 Mean tumor size, in greatest diameter, was 5±2 mm (range, 2–10 mm). Twenty-eight atypical adenomatous hyperplasias were incidental findings in 13 adenocarcinoma patients and the other atypical adenomatous hyperplasias (n=26) were major findings in the remaining 15 patients. All the 28 patients underwent pulmonary resection at the Kanagawa Cancer Center Hospital, Yokohama, Japan, from July 1996 to January 2006. The 28 patients included 9 men and 19 women with a median age of 59.5 years (range, 42–77 years). There were 20 never-smokers and 8 ever-smokers. Informed consents were obtained for all cases and the institutional review board of the Kanagawa Cancer Center approved this study.

Table 1 Clinicopathological findings and EGFR and K-ras mutational analyses of 54 atypical adenomatous hyperplasias obtained from 28 patients
Full size table

DNA Extraction and Sequencing Analyses of the EGFR and K-ras Genes

Formalin-fixed, paraffin-embedded tissue sections were used for DNA isolation from the all atypical adenomatous hyperplasia lesions as described previously.20 Mutational analyses of EGFR gene exons 19 and 21, and K-ras gene codon 12 were performed by LH-MSA and sequencing as described previously.20

Immunohistochemistry

Immunohistochemical evaluations were performed using the avidin-biotin-peroxidase complex method with 3 μm-thick sections of the formalin-fixed, paraffin-embedded specimens. A monoclonal antibody against the Ki-67 antigen (MIB-1; MBL, Nagoya, Japan; 1:100 dilution) was used to assess the proportion of proliferating tumor cells. The Ki-67 labeling index (LI) was defined as the ratio of MIB-1-stained tumor cells to all tumor cells counted, multiplied by 100. To evaluate the Ki-67 LI, stained tumor cells were counted in at least three high-power fields that showed the highest positivity for each section.

Statistical Analyses

Values are shown as mean±s.d. Fisher's exact test was used to assess the association of EGFR mutations with clinical factors including patients' age, sex, and smoking history. Associations between histological subtypes and EGFR or K-ras mutations were evaluated by Fisher's exact test. Maximum size and Ki-67 LI of the atypical adenomatous hyperplasias with and without EGFR mutations were compared by the Mann-Whitney's U test. A P-value of <0.05 were considered to be statistically significant.

Results

EGFR Mutations in Atypical Adenomatous Hyperplasias

EGFR mutations were detected in 17 of the 54 atypical adenomatous hyperplasias (32%) examined; 10 tumors had a deletion mutation at exon 19, and the others (n=7) had a point mutation at exon 21 (L858R) (Table 1).

Correlation of Clinicopathological Characteristics with EGFR Mutations in Atypical Adenomatous Hyperplasias

Although atypical adenomatous hyperplasias of younger patients, women, and never-smokers had EGFR mutations more frequently, no statistically significant associations were found between the EGFR mutation status and either age, sex, or smoking history (Table 2).

Table 2 Relationship between clinicopathological characteristics and frequency of EGFR mutations in the 28 atypical adenomatous hyperplasia patients
Full size table

Both mean sizes of EGFR-mutated and wild-type atypical adenomatous hyperplasias were ∼5 mm; no statistically significant association between EGFR mutation and tumor size was observed. Mean Ki-67 LI of EGFR-mutated atypical adenomatous hyperplasias was 1.7%, whereas that of atypical adenomatous hyperplasias with wild-type EGFR was 1.8%; no statistically significant relationship between EGFR mutation and Ki-67 LI was observed (Table 3). We observed no apparent histological differences between atypical adenomatous hyperplasias with and without EGFR mutations (Figure 1).

Table 3 Comparison of clinicopathological data between atypical adenomatous hyperplasias with and without EGFR gene mutations
Full size table
Figure 1
figure 1

Histopathology of atypical adenomatous hyperplasias with and without EGFR mutations. All of these tumors were found in female patients who had never smoked. Both (a) atypical adenomatous hyperplasias with EGFR mutation (No. 18) and (b) without EGFR mutation (No. 27b) have relatively high cellularity and show almost the same histology. There is no obvious histological difference between atypical adenomatous hyperplasias with relatively low cellularity (c and d), but (c and d) show atypical adenomatous hyperplasias with EGFR mutation (No. 10) and without EGFR mutation (No. 15), respectively.

Full size image

Characteristics of the Patient having a K-ras-Mutated Atypical Adenomatous Hyperplasia

Genomic DNAs of K-ras in five atypical adenomatous hyperplasias were not amplified sufficiently by PCR. Thus, we evaluated the remaining 49 atypical adenomatous hyperplasias for K-ras mutations. K-ras mutation was observed in only 1 of the 49 atypical adenomatous hyperplasias (2%) examined; the mutation is G12S (GGT>AGT). The patient having K-ras-mutated atypical adenomatous hyperplasia (Case No. 3) was a 66-year-old man and had smoking history. Interestingly, he had 10 atypical adenomatous hyperplasias; no mutation was observed in 8 of the 10 atypical adenomatous hyperplasias, and the other 2 had a K-ras or EGFR (deletion at exon 19) mutation.

Discussion

We have shown that 17 of the 54 (32%) atypical adenomatous hyperplasias examined had a hot spot mutation (deletion mutation at exon 19 or L858R at exon 21) of the EGFR gene. To date, EGFR mutations in atypical adenomatous hyperplasias have been reported in three previous studies where it was observed that 2 of 7, 0 of 5, or 1 of 35 atypical adenomatous hyperplasias had EGFR mutations.7, 8, 17 One of these groups reported that 1 of 35 (3%) atypical adenomatous hyperplasias and 4 of 37 (11%) nonmucinous bronchioloalveolar carcinomas examined had EGFR mutations by PCR-direct sequencing of formalin-fixed materials.17 Our results that EGFR mutations were detected in 32% of atypical adenomatous hyperplasias (in this study) and 88% of nonmucinous bronchioloalveolar carcinomas (in a previous study9) using LH-MSA sharply contrasts to their study in terms of EGFR mutational frequency. These discrepancies may be due to the methodology used in studies. As LH-MSA, which is a modified heteroduplex technique, is able to detect mutations in samples containing as little as 5% mutated DNA,20 whereas direct sequencing has been reported to require at least 30% of mutated DNA in the sample,22 LH-MSA is considered to be more sensitive than direct sequencing20 and to have an advantage especially in analyzing tissues such as atypical adenomatous hyperplasias, which often contains not so many tumor cells as compared with the normal counterparts.

As atypical adenomatous hyperplasias and nonmucinous bronchioloalveolar carcinomas probably represent continuum of progression of alveolar intraepithelial neoplasia, differential diagnosis between atypical adenomatous hyperplasias and small-sized nonmucinous bronchioloalveolar carcinomas is often difficult.15 The possibilities that some atypical adenomatous hyperplasias in this study may be diagnosed as ‘nonmucinous bronchioloalveolar carcinomas’ by other pathologists cannot be ruled out. However, we consider that EGFR mutations were not observed only in atypical adenomatous hyperplasias with relatively high cellularity and severe cytological atypia, which may correspond to BACs for some pathologists, from our following observations: (1) both atypical adenomatous hyperplasias with and without EGFR mutations were almost the same in maximum size and proliferation index (Table 3) and (2) variability in cellularity and cytological atypia was similarly observed in EGFR-mutated atypical adenomatous hyperplasias as well as atypical adenomatous hyperplasias without the mutation (Figure 1). These results lead us to conclude that EGFR mutations are already present in a subset of atypical adenomatous hyperplasias with relatively low cellularity and mild cytological atypia, which would be diagnosed as atypical adenomatous hyperplasias by any pathologists. Our measurement that EGFR mutations were observed in about 30% of atypical adenomatous hyperplasias examined would be reasonable, although further studies are needed to validate the mutational frequency.

We had previously reported that 15 of 17 (88%) pure nonmucinous bronchioloalveolar carcinomas (adenocarcinoma in situ) and 49 of 65 (75%) invasive adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma component had EGFR mutations.9 In addition to these high frequencies of EGFR mutations in lung adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma feature, it is known that (1) most human lung adenocarcinomas expressing mutated EGFR respond well to EGFR tyrosine kinase inhibitors10, 11 and (2) transgenic mice expressing mutated EGFR in type II pneumocytes develop atypical adenomatous hyperplasias, bronchioloalveolar carcinomas, and invasive adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma component in the lung in a time-dependent manner.19 These observations imply that persistent EGFR signaling due to activating EGFR mutations would be essential for development and maintenance in lung adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma component. Therefore, we started the present study with a working hypothesis that the EGFR mutational frequency in atypical adenomatous hyperplasias is similar to those of more advanced lesions. However, the frequency of EGFR mutations in atypical adenomatous hyperplasias (32%) is significantly lower (Fisher's exact test, P<0.0001) than those in nonmucinous bronchioloalveolar carcinomas (88%) and invasive adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma component (75%) (hereafter we call these tumors adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma feature) (Figure 2).9 Assuming that all adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma feature come from atypical adenomatous hyperplasias, atypical adenomatous hyperplasias with EGFR mutations are likely to progress to adenocarcinomas much more frequently than those without the mutations. In other words, atypical adenomatous hyperplasias without EGFR mutations are unlikely to progress to adenocarcinomas. This speculation matches the clinical observation that about half of pure ground-glass opacities, identified on high-resolution computed tomography and usually diagnosed as atypical adenomatous hyperplasia or nonmucinous bronchioloalveolar carcinoma on histopathological examination, do not increase in size and density for more than 2 years.23 However, we did not observed significant differences in histology, tumor maximum size, and proliferation activity (estimated by Ki-67 LI) between atypical adenomatous hyperplasias with and without EGFR mutations (Figure 1; Table 3). Further studies are needed to determine the biological differences between atypical adenomatous hyperplasias with and without EGFR mutations.

Figure 2
figure 2

EGFR mutational frequencies in atypical adenomatous hyperplasias, nonmucinous bronchioloalveolar carcinomas, and invasive adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma component. AAH, atypical adenomatous hyperplasia; nm-BAC, nonmucinous bronchioloalveolar carcinoma; AD with nm-BAC, invasive adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma component.

Full size image

We identified only one K-ras mutation in the 49 atypical adenomatous hyperplasias (2%) in this study. This frequency is considerably lower than those (15–39%) of previous studies analyzing atypical adenomatous hyperplasias.13, 16, 24 Although the reason for this marked discrepancy is unclear, we consider that our methodology for mutational analyses is reliable based on the facts as follows: (1) several prior reports, examining DNAs extracted from fresh tumor tissues in Japan, have clarified that about 10% of Japanese lung adenocarcinomas have K-ras mutation,4, 12 and (2) our previous study identified the K-ras mutation of almost the same frequency (9%) in 118 Japanese adenocarcinomas by using LH-MSA.9 In addition, we compared the K-ras mutational frequencies between atypical adenomatous hyperplasias (2%) in this study and adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma feature (2 of 82, 2%) or a mucinous bronchioloalveolar carcinoma component (6 of 9, 67%) in our previous study9 and found that the K-ras mutations in atypical adenomatous hyperplasias or in adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma feature were much less frequent (Fisher's exact test, P<0.0001) than in adenocarcinomas with a mucinous bronchioloalveolar carcinoma component (Figure 3). As atypical adenomatous hyperplasias are strikingly similar to nonmucinous bronchioloalveolar carcinomas, but quite different from mucinous bronchioloalveolar carcinomas in K-ras mutational frequency as well as histopatholology, we surmise that atypical adenomatous hyperplasias do progress sequentially to nonmucinous bronchioloalveolar carcinomas, but not to mucinous bronchioloalveolar carcinomas.

Figure 3
figure 3

K-ras mutational frequencies in atypical adenomatous hyperplasias, adenocarcinoma with a nonmucinous bronchioloalveolar carcinoma feature, and adenocarcinoma with a mucinous bronchioloalveolar carcinoma component. AAH, atypical adenomatous hyperplasia; AD with nm-BAC, adenocarcinoma with a nonmucinous bronchioloalveolar carcinoma feature; AD with m-BAC, adenocarcinoma with a mucinous bronchioloalveolar carcinoma component.

Full size image

In conclusion, about one-third atypical adenomatous hyperplasias examined in this study had EGFR mutations, but few atypical adenomatous hyperplasias had K-ras mutations. We surmise that atypical adenomatous hyperplasias with EGFR mutation would more frequently progress to invasive adenocarcinomas through nonmucinous bronchioloalveolar carcinomas than atypical adenomatous hyperplasias without EGFR mutations. Genetic alterations other than EGFR gene mutations seem to be needed for the development of EGFR-mutated atypical adenomatous hyperplasias to adenocarcinomas with a nonmucinous bronchioloalveolar carcinoma feature. Clarification of the molecular mechanisms involved in the progression process would enable us to discover new targeting molecules, other than EGFR, in lung adenocarcinoma therapy.