SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

BRG1 and BRM, two core catalytic subunits in SWI/SNF chromatin remodeling complexes, have been suggested as tumor suppressors, yet their roles in carcinogenesis are unclear. Here, we present evidence that loss of BRG1 and BRM is involved in the progression of lung adenocarcinomas. Analysis of 15 lung cancer cell lines indicated that BRG1 mutations correlated with loss of BRG1 expression and that loss of BRG1 and BRM expression was frequent in E-cadherin-low and vimentin-high cell lines. Immunohistochemical analysis of 93 primary lung adenocarcinomas showed loss of BRG1 and BRM in 11 (12%) and 16 (17%) cases, respectively. Loss of expression of BRG1 and BRM was frequent in solid predominant adenocarcinomas and tumors with low thyroid transcription factor-1 (TTF-1, master regulator of lung) and low cytokeratin7 and E-cadherin (two markers for bronchial epithelial differentiation). Loss of BRG1 was correlated with the absence of lepidic growth patterns and was mutually exclusive of epidermal growth factor receptor (EGFR) mutations. In contrast, loss of BRM was found concomitant with lepidic growth patterns and EGFR mutations. Finally, we analyzed the publicly available dataset of 442 cases and found that loss of BRG1 and BRM was frequent in E-cadherin-low, TTF-1-low, and vimentin-high cases and correlated with poor prognosis. We conclude that loss of either or both BRG1 and BRM is involved in the progression of lung adenocarcinoma into solid predominant tumors with features of epithelial mesenchymal transition and loss of the bronchial epithelial phenotype. BRG1 loss was specifically involved in the progression of EGFR wild-type, but not EGFR-mutant tumors.

Lung cancer is the leading cause of cancer death in many developed countries, including the United States and Japan.[1, 2] The identification of genetic abnormalities, such as epidermal growth factor receptor (EGFR) mutations, KRAS mutations, EML4–ALK translocation, and MET amplifications has revolutionized our understanding of the molecular mechanisms in lung cancer development.[3] However, it has become increasingly apparent that epigenetic alternations play equally important roles in tumorigenesis, and among them, chromatin remodeling factors have attracted much attention recently.[4] Indeed, identification of mutations of chromatin remodeling factors in cancer has been a major hot topic in the past 2 years.[5-8]

BRG1 and BRM, two core catalytic ATPase subunits in human SWI/SNF chromatin remodeling enzymes, have now emerged as bona fide tumor suppressor genes.[9-12] Inactivating mutations of BRG1 have been identified in 35% of non-small cell lung cancer cell lines and a subset of primary lung cancer.[9] In a mouse model of lung cancer, targeted knockout of BRG1 can affect tumor development.[10] In contrast to BRG1, mutations of BRM have rarely been identified and epigenetic silencing of BRM plays a contributory role in some cancers.[4, 11] However, whether loss of BRG1 and BRM affects phenotype and differentiation of lung cancer cells remains unexplored. Furthermore, the previous studies were conducted before the discovery of EGFR mutations, and thus relationship between the EGFR status and loss of BRG1 and BRM is completely unknown.

We have recently demonstrated that lung adenocarcinoma could be classified into two groups based on the patterns of gene expression and genetic abnormalities; bronchial epithelial phenotype tumors and mesenchymal-like phenotype tumors.[13] “Bronchial epithelial phenotype” represents a group of lung adenocarcinomas with high expression of bronchial epithelial markers. This group includes thyroid transcription factor (TTF)-1 positive terminal respiratory unit (TRU) type[14] in addition to TTF-1 negative tumors with high expression of bronchial epithelial markers such as CK7 and MUC1, as detailed in our previous report.[13] Bronchial epithelial phenotype tumor exhibits high phosphorylation of EGFR and MET and frequent mutations or amplifications of EGFR, MET, and HER2. In contrast, mesenchymal-like phenotype tumors were characterized by the absence of the bronchial epithelial phenotype, triple-negative for TTF-1, MUC1, and CK7, showed no or little phosphorylation of EGFR and MET, no mutation or amplification of EGFR, MET, or HER2, and with features of epithelial mesenchymal transition (EMT), such as low E-cadherin and high FGFR1, vimentin, and ZEB1 expressions.[13] The absence of mutations or amplifications of EGFR, MET, or HER2 in mesenchymal-like phenotype tumors suggested to us that other genetic or epigenetic abnormalities may play a role in this group of tumors.

We now show in this paper that loss of expression of chromatin remodeling factors, BRG1 and BRM, correlated with features of mesenchymal-like phenotype with solid predominant histology. In particular, BRG1 loss occurred exclusively in EGFR wild-type tumors.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

Cell lines and medium

We used 19 non-small cell lung cancer (NSCLC) cell lines; 15 adenocarcinoma cell lines (H23, H358, H441, H522, H1395, H1648, H1650, H1703, H1795, H2087, HCC827, HCC4006, Calu3, A549, PC-3), three large cell carcinoma cell lines (H661, H1299, Lu65), and one adenosquamous cell line (H596). HCC827, H1650, H1975, PC-3, and HCC4006 were EGFR-mutated cell lines. The sources of the cell lines were described in our previous report.[13] All cell lines were maintained in RPMI1640 supplemented with 10% FCS, glutamine, and antibiotics in a humidified atmosphere with 5% CO2 and 95% air.

Genetic and protein analysis of cell lines

The DNA, RNA, and cell lysates were prepared from cell lines by standard procedures. Experimental details of sequencing, copy number analyses, and Western blotting are given in Doc. S1. Antibodies used in western blot analysis were summarized in Table 1.

Table 1. Antibodies used in western blot analysis
AntibodiesCloneSources
BRG1 (sc-17796)Mouse monoclonalSanta Cruz Biotechnology (Santa Cruz, CA, USA)
BRM (A301-016A)Rabbit polyclonalBethyl Laboratory (Montgomery, TX, USA)
TTF-1 (clone 8G7G3/1)Mouse monoclonalDAKO (Glostrup, Denmark)
Cytokeratin 7 (clone OV-TL 12/30)Mouse monoclonalDAKO (Glostrup, Denmark)
Vimentin (clon V9)Mouse monoclonalDAKO (Glostrup, Denmark)
E-cadherin (clone 36)Mouse monoclonalBD Biosciences (Franklin Lakes, NJ, USA)
MUC1 smaller cytoplasmic subunitHamster monoclonalLab Vision (Cheshire, UK)
Anti-rabbit IgG peroxidase conjugate Amersham (Arlington Hights, IL, USA)
Anti-mouse IgG peroxidase conjugate Amersham (Arlington Hights, IL, USA)
Anti-Armenian hamster IgG peroxidase conjugate Jackson Immunoresearch (West Grove, PA, USA)

Patients and tumors

Tumor specimens were obtained from 93 patients who underwent lung cancer surgery at the Jichi medical university hospital during the period from October 2005 to June 2008. The demographic and clinicopathologic details of the patients and tumors are provided in Doc. S1.

Immunohistochemistry and evaluation

Formalin-fixed, paraffin-embedded tumor specimens were analyzed by immunohistochemistry using antibodies to BRG1, BRM, E-cadherin, cytokeratin 7, MUC1, TTF-1, p-EGF, and p-MET. The sources of antibodies, staining procedures, and methods of evaluation, are given in Doc. S1.

Mutation analyses of formalin-fixed, paraffin-embedded tissue sections

Details are shown in Doc. S1.

Bioinformatic analyses and statistics

Details are shown in Doc. S1.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

Characteristics of lung adenocarcinoma cell lines with loss of BRG1 and/or BRM

First, we used 15 lung cancer cell lines, for which the mutational status of BRG1 was known, to investigate the molecular features that may characterize lung adenocarcinoma cell lines with loss of BRG1. Of the 15 cell lines, six cell lines harbored BRG1 mutations and nine cell lines did not, according to previous literature,[9, 11] and the Sanger COSMIC database (http://www.sanger.ac.uk/genetics/CGP/cosmic/).

Figure 1 summarizes (i) the genetic status of BRG1, EGFR, MET, HER2, BRAF, and KRAS (upper panel), (ii) gene level expressions of BRG1, BRM, TTF1, MUC1, CK7, E-cadherin, and vimentin (middle panel), and (iii) protein expression levels of BRG1, BRM, TTF1, MUC1, CK7, E-cadherin, and vimentin (lower panel) of the 15 cell lines (the microarray analysis data of 15 cell lines is located in Data S1). All six BRG1-mutated cell lines showed extreme loss of BRG1 at gene and protein levels, EMT features (low E-cadherin and high vimentin), and loss of bronchial epithelial markers (TTF-1, CK7, and MUC1). In contrast, the nine BRG1-wild type cell lines showed high expressions of BRG1, as well as high expressions of bronchial epithelial markers and E-cadherin and low expression of vimentin, at both gene and protein levels. As for gene abnormalities, BRG1-wild type cell lines showed gene abnormalities for EGFR, MET, HER2, BRAF, or KRAS, but BRG1 mutated cell lines showed no such genetic abnormalities, except for KRAS mutations.

image

Figure 1. Genetic status of BRG1, epidermal growth factor receptor (EGFR), MET, HER2, BRAF, and KRAS (upper panel), gene level expressions of BRG1, BRM, TTF1, MUC1, CK7, E-cadherin, and vimentin (middle panel), and protein expression levels of BRG1, BRM, TTF1, MUC1, CK7, E-cadherin, and vimentin (lower panel) of the 15 cell lines. In the upper panel, the grey box means presence of genetic abnormalities and the white box means absence of genetic abnormalities. Color indications in the middle lane are as follows: red means more than or equal to the average of each gene expression; orange: under the average and more than or equal to half the average; and green: under half the average.

Download figure to PowerPoint

We also examined the expressions of BRM in the same cell lines. Of the 15 cell lines, 10 cell lines expressed the BRM protein at modest or high levels, which was largely concordant with gene expression (Fig. 1). As with BRG1, loss of BRM expression was similarly frequent in cell lines with EMT features and loss of the bronchial epithelial phenotype.

These results suggest the following: (i) loss of either or both BRG1 and BRM would be involved in the acquisition of EMT features and loss of the bronchial phenotype; and (ii) loss of BRG1 gene and protein expression correlate with the BRG1 mutation status.

We conducted the same analysis in five EGFR-mutated cell lines (HCC827, H1650, H1975, PC-3, and HCC4006), as shown in Figure S1. Although genetic status of BRG1 was unknown in H1650, H1975, PC-3, and HCC4006, all five EGFR-mutated cell lines showed high expression levels of BRG1, which suggested that loss of BRG1 would be mutually exclusive with EGFR mutations.

Immunohistochemical expression of BRG1 and BRM in primary lung adenocarcinoma tissues

Next, we used 93 cases of primary lung adenocarcinoma cases in our institution to examine the immunohistochemical expressions of BRG1 and BRM and their relationship with (i) histopathological subtypes, (ii) presence or absence of lepidic growth components, (iii) expressions of E-cadherin, TTF-1, CK7, and MUC1, (iv) genetic status of EGFR and KRAS, and (v) activation levels of EGFR and MET.

Overall, in the large majority of cases (>80%), nuclear staining for BRG1 and BRM was observed in cancer cells (Figs 2, 3). Stromal cells constantly stained positive for BRG1 and BRM, and thus served as excellent internal positive controls. Using the criteria described in the Methods (Doc. S1), 11 cases (12%) were judged as showing low expression levels of BRG1 and 16 cases (17%) as showing low expression levels of BRM. Five cases (5%) showed low expression levels of both BRG1 and BRM. Most of the BRG1-low cases were either completely negative or showed only scattered positive staining for BRG1. In contrast, BRM showed a more heterogeneous staining pattern, typically showing strong positive staining in lepidic growth components, while showing negative or weak staining in invasive high-grade components (Fig. 4A–C).

image

Figure 2. BRG1 expressions in lung adenocarcinomas. Overall, more than 80% of cases showed positivity for BRG1. (A) Lepidic growth components showed strong immunoreactivity for BRG1. (B) Moderately differentiated adenocarcinomas, such as papillary adenocarcinoma, frequently showed positivity for BRG1. (C) Solid adenocarcinomas with mucin were frequently negatively stained for BRG1. (D) Some cases with papillary or acinar morphology showed negative staining for BRG1. Note BRG1 positivity in stromal cells.

Download figure to PowerPoint

image

Figure 3. BRM expressions in lung adenocarcinomas. Overall, more than 80% of cases showed positivity for BRM. (A) Lepidic growth components showed strong immunoreactivity for BRM. (B) Moderately differentiated adenocarcinomas, such as papillary adenocarcinoma, also often show positivity for BRM. (C) Solid adenocarcinoma with mucin frequently showed negative or weak staining for BRM. (D) Some cases with papillary or acinar morphology show negative staining for BRM. Note BRM positivity in stromal cells.

Download figure to PowerPoint

image

Figure 4. Heterogeneous BRM expression in lung adenocarcinoma. (A) High-power field of lepidic growth components, which show strong positivity for BRM. (B) High-power field of invasive acinar components, which show negative positivity for BRM. (C) Low-power field of invasive adeno-carcinoma with lepidic growth components; left side shows lepidic growth components, and right side shows invasive acinar components.

Download figure to PowerPoint

Table 2 shows correlations between the expression levels of BRG1 and BRM and histopathological subtypes. All cases of well differentiated adenocarcinomas, that is, adenocarcinoma in situ (AIS) and minimally invasive adenocarcinoma (MIA), showed positive immunostaining for both BRG1 and BRM (20 of 20, 100%; Figs 2A,3A). Moderately differentiated adenocarcinomas, that is, acinar or papillary adenocarcinoma, also frequently showed positive immunostaining for both BRG1 and BRM (37 of 46, 80%; Figs 2B,3B), while some of them showed loss of either BRG1 or BRM (9 of 46, 20%; Figs 2D,3D). In contrast to these well- to moderately-differentiated tumors, poorly-differentiated adenocarcinomas (solid adenocarcinomas) frequently showed loss of expression of either BRG1 or BRM (12 of 13, 92%; Figs 2C,3C). Most cases (4 of 5, 80%) with loss of both BRG1 and BRM showed solid morphology (Table 2). One of three cases (33%) of invasive mucinous adenocarcinoma showed loss of BRG1.

Table 2. Correlations between expression levels of BRG1 and BRM and histopathological subtypes of primary lung adenocarcinomas
 BRG1 high BRM highBRG1 low BRM highBRG1 high BRM lowBRG1 low BRM lowTotal
Non-mucinous adenocarcinoma in situ80008
Minimally invasive adenocarcinoma1200012
Invasive adenocarcinoma, lepidic predominant90009
Invasive adenocarcinoma, acinar predominant802111
Invasive adenocarcinoma, papillary predominant2933035
Invasive mucinous adenocarcinoma21003
Colloid adenocarcinoma10001
Invasive adenocarcinoma, micropapillary predominant10001
Invasive adenocarcinoma, solid predominant126413
Total71611593

We also examined correlations between the expression levels of BRG1 and BRM and the presence or absence of lepidic growth components (Table 3). Strikingly, all cases with BRG1 loss were devoid of lepidic growth components, while 6 of 16 cases with BRM loss showed lepidic growth components (Table 3).

Table 3. Correlations between expression levels of BRG1 and BRM and clinico-pathological factors
 BRG1 expressionBRM expression
HighLowP-valueHighLowP-value
  1. a

    Pathological N-factors were not determined for three cases of stage IV patients with pleural dissemination. Underlined values are < 0.05.

Pathological stage
I6070.508257100.3499
II + III + IV224206
T-stage
T15250.25084890.6492
T2, T3, T4306297
Nodal involvementa
Positive2010.36081740.7381
Negative6185811
Lymphatic invasion
Positive2210.20041850.5066
Negative60105911
Vessel invasion
Positive2350.23731711 0.0002
Negative596605
Pleural invasion
Positive216 0.0471 2160.4122
Negative6155610
Dissemination
Positive32 0.0449 410.8648
Negative7997315
Pulmonary metastasis
Positive400.4540400.3514
Negative78117316
Lepidic growth
Present650 <0.0019 596 0.0019
Absent17111810
Smoking Index
≤6002660. 13442210 0.0093
>600565556

Table 4 shows correlations between the expression levels of BRG1 and BRM and that of bronchial epithelial markers (TTF-1, CK7, and, MUC1) and E-cadherin. The expressions of TTF-1, CK7, MUC1 (membranous expression), and E-cadherin were frequently reduced in cases with loss of BRG1 and BRM (shown in Fig. S2). In particular, loss of E-cadherin and TTF-1 was remarkably correlated with loss of BRG1; all but one case of E-cadherin-low tumors showed BRG1 loss and all cases with BRG1 loss showed low expression levels of TTF-1. Depolarized expression of MUC1 was also frequent in cases with loss of BRG1 and BRM.

Table 4. Correlations between expression levels of BRG1 and BRM and genetic status of EGFR and KRAS and expression levels of E-cadherin, TTF-1, CK7, MUC1, phopsho-MET, and phospho-EGFR
 BRG1 expressionBRM expression
HighLowP-valueHighLowP-value
  1. Underlined values are < 0.05.

EGFR mutations
Positive450 0.0006 3960.3382
Negative37113810
KRAS mutations
Posivite520.1537430.0615
Negative7797313
E-cadherin
High813 <0.0001 7212 0.0227
Low1854
TTF-1
High620 <0.0001 575 0.0010
Low20112011
CK7
High744 <0.0001 67110.0707
Low87105
MUC1(membranous)
High633 0.0007 624 <0.0001
Low1981512
MUC1(depolarized)
High730.059646 0.0001
Low7987310
Phospho-EGFR
High694 0.0003 676 <0.0001
Low1371010
Phospho-MET
High1920.71021920.2892
Low6392814

Table 4 also shows correlations between the expression levels of BRG1 and BRM and genetic status of EGFR and KRAS. Mutually exclusive correlations were observed between EGFR mutations and BRG1 loss (= 0.0006), but no significant correlations between EGFR mutations and BRM loss (= 0.3382). KRAS mutations were sometimes harbored by cases with loss of BRG1 or BRM.

We also examined the expressions of phospho-EGFR and phospho-MET and compared them with the expressions of BRG1 and BRM (Table 4). Low phosphorylation levels of EGFR were significantly correlated with loss of BRG1 and BRM (= 0.0003, < 0.0001, respectively). Phosphorylation of MET tended to be low in cases with loss of BRG1 and BRM, but the difference was not significant.

BRM loss was significantly more frequent in heavy smokers and in cases with vessel invasion (= 0.0093 and = 0.0002, respectively; Table 3). BRG1 loss was significantly correlated with pleural invasion and pleural dissemination (= 0.0471 and = 0.0449, respectively; Table 3).

Prognostic significance of the expressions of BRG1 and BRM

Lastly, we performed hierarchical cluster analysis using the publicly available data of 442 primary lung adenocarcinoma cases,[15] based on the gene expressions of BRG1, BRM, TTF-1, MUC1, CK7, E-cadherin, and vimentin. Results are shown in Figure 5(A). Principally, primary lung adenocarcinoma cases could be classified into two groups: (i) tumors showing high expression levels of TTF-1, MUC1 and E-cadherin, and low expression levels of vimentin, and (ii) tumors showing low expression levels of TTF-1, MUC1 and E-cadherin, and high expression levels of vimentin (Fig. 5A). High expression levels of both BRG1 and BRM were frequently seen in the former, while low expression levels of either of or both BRG1 and BRM were frequently seen in the latter (Fig. 5A). These results confirm and reinforce data from cancer cell lines and primary lung adenocarcinoma cases in our institution.

image

Figure 5. Analysis of the publicly available data of 442 primary lung adenocarcinoma cases. (A) Hierarchical cluster analysis using the gene expressions of vimentin, CK7, BRG1, BRM, TTF-1, E-cadherin, and MUC1. (B) Patient survival according to the expression levels of BRG1 and BRM. Lung adenocarcinoma cases were separated into three groups based on gene expression levels of BRG1 and BRM: (i) cases with high expression (more than or equal to the average); (ii) cases with moderate expression (under the average, and more than or equal to half the average); and (iii) cases with low expression (under half the average). Left panel shows patient survival curves with high expression levels of BRG1 (BRG1-High), moderate expression levels of BRG1 (BRG1-Mod), and low expression levels of BRG1 (BRG1-Low). Right panel shows patient survival curves with high expression levels of BRM (BRM-High), moderate expression levels of BRM (BRM-Mod), and low expression levels of BRM (BRM-Low). (C) Patient survival according to the expression pattern of BRG1 and BRM. Patients were separated into four groups according to the expression pattern of BRG1 and BRM as follows: cases with high expression levels of both BRG1 and BRM (BRG1-High, BRM-High), cases with high expression levels of BRG1 and moderate or low expression levels of BRM (BRG1-High, BRM-Low), cases with moderate or low expression levels of BRG1 and high expression levels of BRM (BRG1-Low, BRM-High), the cases with moderate or low expression levels of both BRG1 and BRM (BRG1-Low, BRM-Low).

Download figure to PowerPoint

To ascertain the prognostic significance of the expressions of BRG1 and BRM in lung adenocarcinoma, we undertook a survival analysis using the Kaplan–Meier method. We separated 442 lung adenocarcinoma cases into three groups based on the gene expression levels of BRG1 and BRM; (i) cases with high expression (more than or equal to the average); (ii) cases with moderate expression (under the average, and more than or equal to half the average); and (iii) cases with low expression (under half the average). Results are shown in Figure 5(B). High expression of BRG1 and BRM both correlated significantly with better prognosis. Figure 5(C) also shows patient survival curves for the four groups: (i) BRG1-high and BRM-high; (ii) BRG1-high and BRM-low; (iii) BRG1-low and BRM-high; and (iv) BRG1-low and BRM-low. The BRG1-low and BRM-low group showed significantly poorer prognosis than the other groups.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

This is, to our knowledge, the first report demonstrating the tight correlation between loss of BRG1 and BRM and EMT in cancer. Results of this study also confirm and reinforce our previous data that loss of the bronchial epithelial phenotype occurs in lung adenocarcinomas with EMT features.[13]

Recent studies show that loss of another component of the SWI/SNF chromatin remodeling complex, BAF250A (ARID1A), was frequent in high-grade endometrial carcinomas and clear cell carcinomas of the ovary[16] and that loss of the BAF250A protein correlates with the ARID1A mutation status.[17, 18] Interestingly, there appears to be a similarity between loss of BRG1 and that of ARID1A; both tend to occur in high-grade tumors or in tumors with an altered epithelial phenotype.

Another interesting finding of this study was that features of TRU type lung adenocarcinomas,[14] that is, lepidic growth features, high expression levels of the TTF-1 protein, and EGFR mutations were absent in all cases with loss of the BRG1 protein. In tumors with BRG1 loss, BRG1 protein expression was typically absent in almost all cancer cells. These results suggest that loss of BRG1 occurs at an early step of carcinogenesis of lung adenocarcinoma with the mesenchymal-like phenotype, that is, a subset of non-TRU type lung adenocarcinomas with EMT features.

All cases with concomitant loss of BRG1 and BRM were devoid of lepidic growth components, harbored no EGFR mutations, and correlated more with solid adenocarcinoma morphology than a single loss of BRG1 or BRM, which suggested that loss of BRM may also occur in a subset of the mesenchymal-like phenotype, simultaneously with, or subsequently to, loss of BRG1, and may accelerate poorer differentiation and EMT and lead to the more malignant phenotype. The survival analysis of Shedden's data, which showed that cases with concomitant loss of BRG1 and BRM had poorer prognosis than cases with a single loss of BRG1, supports this hypothesis.

BRM expression was positive in lepidic growth components, but was weak or absent in invasive poorer differentiated lesions, such as solid components. In contrast to BRG1 loss, BRM loss may occur during the progressions of lung adenocarcinomas with the bronchial epithelial phenotype. Figure 6 shows our hypothetical schemes for BRG1 and BRM loss in the development of two types of lung adenocarcinomas: lung adenocarcinomas showing the bronchial epithelial phenotype and those showing the mesenchymal-like phenotype.

image

Figure 6. Hypothetical schemes of BRG1 and BRM loss in the development of two types of lung adenocarcinomas: lung adenocarcinoma with the bronchial epithelial phenotype (upper panel) and lung adenocarcinoma with mesenchymal-like phenotype (lower panel).

Download figure to PowerPoint

BRG1 and BRM regulate a broad range of genetic programs, including cell differentiation and proliferation, and it has been suggested that SWI/SNF complexes may dictate lineage-specific chromatin remodelling functions and act as master regulators of the master regulators.[4] Thus, although the exact mechanism by which loss of BRG1 and BRM leads to tumor development and EMT is unknown, loss of BRG1 and BRM may cause uncontrolled cellular proliferation and disrupt the differentiation program of bronchial epithelial cells,[19] resulting in formation of tumors with loss of expression of CK7, MUC1, and TTF-1. Why BRG1 loss occurs exclusively in the progression of EGFR wild-type tumor is currently unknown. One speculation could be that the simultaneous presence of EGFR mutation and BRG1 loss is for some reason incompatible with survival of cancer cells. Finally, we speculate that epigenetic therapy aiming to restore the functions of BRG1 and BRM would be a possible new therapy for treating tumors with EMT features.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

This study was supported in part by the Smoking Research Foundation, the Foundation for Development of Community, the Vehicle Racing Commemorative Foundation, and the Ministry of Health, Labor and Welfare, and the Ministry of Education, Culture, Sports, Science and Technology of Japan. This publication was subsidized by JKA through its promotion funds from KEIRIN RACE. We thank Dr S. Matsukuma and Dr E. Tsuchiya, Molecular Pathology and Genetics Division, Kanagawa Cancer Center Research Institute, for technical advice on the loop-hybrid method.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
cas12065-sup-0001-DataS1.pdfPDF document7KData S1. Gene expression data of 19 cell lines.
cas12065-sup-0002-DocS1.docWord document60KDoc. S1. Supporting information about materials and methods.
cas12065-sup-0003-FigS1.pdfPDF document171KFig. S1. Gene and protein level expressions of BRG1, BRM, etc. of EGFR-mutated cell lines.
cas12065-sup-0004-FigS2.pdfPDF document269KFig. S2. Immunostaining for BRG1, BRM and E-cadherin in serial sections.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.