Mechanisms of p185 HER2 Expression in Human Non-Small Cell Lung Cancer Cell Lines Jeffrey A. Kern, Robert A. Robinson, Adi Gazdar, Lisa Torney, and David B. Weiner Departments of Internal Medicine and Pathology, University of Iowa College of Medicine, Iowa City, Iowa; Departments of Medicine and Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; and the National Cancer Institute and Naval Hospital, Bethesda, Maryland
To identify mechanisms that allow p185HER2 expression in lung cancer, we performed Western, Southern, and Northern blot analyses of 14 cell lines derived from human non-small cell lung carcinomas and one cell line derived from a human mesothelioma. Human bronchiole epithelial cells and rat type II pneumocytes were found to express p185HER2 at low to undetectable levels by Western blot technique. In contrast, 13 lung cancer cell lines expressed pI85 H ER2 , and eight of these 13 expressed p185HER2 at levels at least 2-fold higher than that found in normal bronchiole epithelial cells or type II pneumocytes. Genomic Southern analysis showed that amplification of the HER2 gene was present in only one of the eight cell lines that expressed p185HER2 at these higher levels. Increased levels of steady-state HER2 mRNA occurred in the remaining seven cell lines. We conclude that in human non-small cell lung carcinoma cell lines the most common mechanism resulting in increased p185HER2 expression is due to mechanisms that increase HER2 mRNA levels, with HER2 gene amplification occurring less commonly.
Proto-oncogenes represent normal cellular genes that were identified on the basis of their similarity to genetic sequences with known tumorigenic or transforming potential. Considerable circumstantial evidence now exists showing that acquired alterations in either the structure, copy number, or expression of these genes may play a role in the pathogenesis of some human malignancies. One such gene called HER2 or c-erbB-2 is a member of the tyrosine kinase supergene family and encodes a membrane-bound tyrosine kinase glycoprotein (pI85 HER2 ) with structural similarity to the epidermal growth factor receptor (EGFR) (1-4). Based on structural homology, it is postulated to be a growth factor receptor and a specific ligand has recently been isolated that may allow testing of this hypothesis (5). Recently, we reported expression of p185HER2 in 30% of human lung cancer cell lines and biopsy specimens (6). Schneider and associates (7) have also found HER2 mRNA in non-small cell lung cancer cell lines and freshly resected non-small cell lung cancer. The expression of p185H ER2 in lung cancer has important clinical correlates, as we have found that in pulmonary adenocarcinomas it independently identifies an older patient population (66.6 ± 10.1 versus 57.5 ± 10.8 yr of age; P = 0.04) and patients with shorter (Received in original form May 8, 1991 and in revised form September 16, 1991) Address correspondence to: Jeffrey A. Kern, M.D., Department ofInternal Medicine, Room C-33A, GH, University ofIowa College of Medicine, Iowa City, IA 52242. Abbreviations: epidermal growth factor receptor, EGFR; fetal bovine serum, FBS; Hanks' balanced salt solution, HBSS; peripheral blood mononuclear cell(s), PBMC; sodium dodecyl sulfate polyacrylamide gel electrophoresis, SDS-PAGE. ' Am. J. Respir. Cell Mol. BioI. Vol. 6. pp. 359-363, 1992
survival (83.7 ± 94.1 versus 188.5 ± 120 wk; P = 0.01) (8). Therefore, expression in human lung adenocarcinomas may be more predictive for clinical outcome than any other prognosticator presently known, with the exception of surgical stage. Interestingly, p185H ER2 expression in lung tumors is at levels significantly higher than in normal bronchiole epithelium (6, 8). The mechanism(s) allowing increased p185HER2 expression in human non-small cell lung cancer is unknown. In human breast and ovarian cancer, HER2 gene amplification is common, resulting in p185HER2 overexpression (9). But in primary human lung cancers, HER2 gene amplification has rarely been found (7, 9, 10). Therefore, we hypothesized that mechanisms other than HER2 gene amplification occur in human lung cancer to result in increased p185H ER2 expression. To test this hypothesis, we performed Western blot analysis of rat type II pneumocytes, human bronchiole epithelial cells, and SV-40 immortalized tracheal cells to establish normal levels of p185HER2 expression in the lung. To identify mechanisms involved in p185H ER2 overexpression in human lung cancer, Western, Southern, and Northern blot analyses were next performed on human non-small cell lung cancer cell lines. We found that increased p185HER2 expression is associated with increased levels of HER2 mRNA, whereas gene amplification occurs less commonly. These results have important implications for understanding the etiology of lung neoplasia.
Materials and Methods Cell Culture The rat neuroblastoma cell line BI04 or the breast adenocarcinoma cell line SK-BR-3 was utilized as a positive control in Northern blot and Western blot analyses. These lines are
360
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992
known to express p185HERZ to high levels (3, 11). Human lung tumor cell lines were obtained from American Type Culture Collection (ATCC) (Rockville, MD) or established from biopsy or aspiration specimens as part of approved clinical protocols. They included human adenocarcinomas National Cancer Institute H23, H522, H1264, H1466, H358, ATCC CALU-3, A-427, A-549, SK-LU; large-cell carcinoma, NCI H1299; squamous-cell carcinomas NCI H520, ATCC CALU-1, CALU-6; mesothelioma ATCC SK-MES; and adenosquamous carcinoma NCI H596. The cells were cultured in RPMI-1640 media, 10% fetal bovine serum (FBS), or Eagle's modified essential media and 10% FBS and incubated at 37° C in a humidified 5 % COz atmosphere. The BEAS-2B cell line was obtained from Curtis C. Harris (Laboratory of Human Carcinogenesis, National Cancer Institute, Bethesda, MD). It was derived from human tracheal epithelium immortalized by SV-40 viral infection and maintained in culture according to previously published methods (12). SK-BR-3 was obtained from ATCC and maintained in culture according to their recommendations. p185HERz-specific Rabbit Antiserum Polyclonal anti-p185HERz antiserum was prepared by immunizing rabbits with a synthetic peptide corresponding to amino acid residues 1240 through 1255 of the human p185HERZ sequence coupled to keyhole limpet hemocyanin. This antiserum has been previously described in detail (6, 8, 13, 14). Preparation of Rat Type II Pneumocytes and Human Bronchiole and Nasal Epithelial Cells Rat type II cells were derived from pathogen-free male Sprague-Dawley rats by enzymatic digestion of the alveoli and differential adherence as previously described (15). Bronchiole epithelial cells were obtained from human lung tissue removed from patients undergoing thoracotomy and lobectomy for lung cancer. The primary tumor was immediately resected from the specimen, and bronchioles from areas of lung uninvolved with tumor were resected and opened longitudinally. The epithelial cell layer and basement membrane were removed by sharp dissection, washed in Hanks' balanced salt solution (HBSS), recovered by centrifugation, and lysed in protein lysis buffer. A nasal polyp was obtained from a patient undergoing polypectomy. Immediately after resection, the polyp was sectioned under sterile conditions to remove the epithelial layer and basement membrane from the underlying stroma. The epithelial cells were washed in calcium-free HBSS and cultured on 100-mm plates in RPMI and 10% FBS at 37° C in a humidified 5% COz atmosphere. Upon reaching confluency, the cells were lysed in protein lysis buffer. Western Blot Analysis Cells were disrupted in protein lysis buffer (20 mM Tris [pH 8], 0.33 M sucrose, 0.5 mM EGTA, 25 J.'M leupeptin, 1.5 J.'M aprotinin, 5 J.'M n-ethyl-maleimide, 0.5% Triton X-100). A total of 100 J.'g of total protein from each cell line was subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (7.5% gel, reducing condition) and electrically transferred to a nitrocellulose membrane. The nitrocellulose membrane was probed with DBW-2, washed, and incubated with ['z5I]protein A (sp act, 10 J.'Ci/p.g; New England Nuclear, Boston, MA). The membrane was then
washed with phosphate-buffered saline and exposed to Kodak XAR-5 film (Eastman Kodak, Rochester, NY). RNA Isolation and Northern Blot Analysis Total cellular RNA was isolated by a guanidium isothiocyanate method with cesium chloride modification (16, 17). RNA was size-fractionated under denaturing conditions as previously described (8). The RNA was transferred to nylon membranes and hybridized under standard conditions with 1Q6 cpm/lane of a heat-denatured 3zP-Iabeled plasmid DNA (18) with a full-length HER2 cDNA insert (4). Blots were washed under high stringency and exposed at -70° C to Kodak XAR-5 film. To ensure equal loading and transfer of RNA in all lanes, radiolabeled probes were removed from the membranes by incubation with 0.2 % SDS, 10 mM Tris at 85° C for 60 min, and blots were reprobed using a {j-actin control probe (19). Preparation of DNA and Southern Blot Analysis Approximately 5 x 107 cells were washed in PBS and genomic DNA extracted from proteinase K-digested cells by ethanol precipitation after deproteinization (20). Isolation of high-molecular-weight DNA was verified by examination of the DNA on a 0.3 % agarose gel. DNA was subjected to restriction endonuclease digestion with EcoRI, and DNA fragments resolved by electrophoresis in a 0.8% agarose gel using tx TAE (40 mM Tris-acetate, 1 mM EDTA) running buffer. DNA was transferred to nylon membranes after depurination according to the method of Southern (21). Membranes were hybridized with a 3zP-Iabeled HER2 cDNA (4, 18), washed twice in zx SSC, 0.1% SDS at 25° C for 30 min and twice in 0.1% SSC, 0.1% SDS at 55° C for 30 min, and exposed to Kodak XAR-5 film. To ensure equal loading and transfer of DNA among lanes, radiolabeled probes were removed from the membranes by incubation with 0.2 % SDS, 10 mM Tris at 85° C for 60 min and reprobed using the human EGFR cDNA (22).
Results p185 HER2 Is Expressed at Low Levels in Normal Human Bronchiole Epithelial Cells We first sought to determine the level of p185HERZ expression in normal pulmonary epithelial cells. Previous data using immunocytochemical techniques have shown that p185HERZis expressed at low levels in normal bronchiole epithelium and type II pneumocytes (6, 8). To confirm this, ciliated bronchiole epithelial cells were isolated from two human lung specimens obtained at thoracotomy, and p185HERZ was quantitated by Western analysis. In addition, rat type II pneumocytes were isolated from rat lungs and similarly evaluated for p185HERZ expression. As shown in Figure 1, rat type II pneumocytes expressed only 5 % of the amount of p185HERZ as compared with the B104 cell line, whereas human bronchiole epithelium expressed undetectable amounts. These low levels of biochemically defined p185HERZ expression agree with our previous immunocytochemical findings in normal pulmonary epithelial cells. In addition, the SV-40 immortalized, but nontransformed, tracheal cell line BEAS-2B and cultured nasal polyp epithelial cells had undetectable levels of p185HERz. We conclude from these studies that p185HERZ expression in nontransformed
Kern, Robinson, Gazdar et al.: pI85 HER2 Expression in Non-Small Cell Lung Cancer
kD 200-
proximately 5,000 D higher than pI85 HER2 identified in the B104 cell line . This difference has previously been ascribed to interspecies glycosylation differences. Interestingly, in some cell lines the reactive protein appeared as a doublet (H522, H1299, A-427, CALU-I, CALU-6) separated by a molecular mass of roughly 5,000 D. A 160-kD incompletely glycosylated form of pI85 HER2 was also identified in cell lines that expressed high levels of p185 HER2 (B104, HI 466, SK-LU, CALU-3) (1). No difference in pI85 HER2 molecular mass was found when examined under reducing or nonreducing SDS-PAGE conditions. To quantitate the level of pI85 HER2 expression among the cell lines, the autoradiogram hybridization signals were digitally quantitated and normalized to the signal obtained from the B104 cell line. All values obtained by digital quantitation were confirmed by quantitative dot blot analysis of serial dilutions of cellular proteins (data not shown). As shown in Table 1, expression ranged from undetectable levels (H520, SK-MES) to levels greater than that found in the B104 cell line (CALU-3). Eight of the 13 pI85 HER2 expressing cell lines expressed levels of pI85 HER2 that were at least 2-fold greater than the highest level found in normal human or rat tissue (Table 1). Indeed, one cell line (CALU-3) expressed levels of p185 HER2 higher than the rat cell line B104 and equal to the breast carcinoma cell line SK-BR-3 known to have an amplified HER2 gene (Figure 1) (11).
92.5 69-
43-
r M
...
0
iii
M
Ii:
::J
:: 1Il..J eta za-
a:
lD
N
rh
et
UJ
lD
= III UJUJ
III
a-I» I-U I-a et:::li
a:=> UJ
lD
Z
Antibody : DBW· 2
a-
Figure1. p185HER2 expression in normal human lung. p185HER2 was identified in 100 JA.g oftotal cell protein from uncultured bronchiole epithelial cells, cultured epithelial cells from a nasal polyp, an SV40 immortalized tracheal cell line (BEAS-2B) , uncultured rat type II pneumocytes, a breast adenocarcinoma cell line (SK-BR-3), and a lung adenocarcinoma cell line (CALU-3 ). This figure is representative of two Western blots.
Amplification of the HER2 Proto-oncogene in Human Lung Carcinoma Cell Lines Is Uncommon To identify mechanisms that allow pI85 HER2 to be expressed in the cell lines at levels higher than that found in normal bronchiole epithelium, we first assessed HER2 gene copy number in all cell lines by genomic Southern blot analysis . Autoradiogram hybridization . signals were quantitated by digital analysis and compared to a DNA sample prepared from peripheral blood mononuclear cells (PBMC) of a normal volunteer. EcoRI digests of genomic DNA derived from PBMC or the lung cancer cell lines resulted in four restriction fragments (14, 8.1, 6.6, 4.7 kb) (Figure 3) . One cell line, CALU-3, was found to have an 8-fold amplification of the HER2 gene (Figure 3, Table 1). The remaining cell lines had no amplification and therefore contain two HER2 genes. To control for equal DNA loading among lanes, all blots were probed
bronchiole epithelial cells is very low, at most 5 % of the expression levels observed in B104, the rodent tumor cell line . p185HERl Is Variably Expressed in Human Non-Small Cell Lung Cancer Cell Lines Fourteen cell lines derived from human non-small cell lung cancers and one mesothelioma cell line were subjected to Western blot analysis to identify pI85 HER2 expression. Thirteen of the 14 lung cancer cell lines expressed varying levels of pI85 HER2 , while one squamous cell carcinoma and the mesothelioma cell line did not express pI85 HER2 (Figure 2). The relative molecular mass of pI85 HER2 was 190 kD, ap-
2
Figure 2. p185HER2 expression
kD
in human lung cancer cell lines. p185HER2 was identified in 100 J1.g of total cell protein from each cell line. The presence of a DBW-2 185-kD reactive doublet can be seen in cell lines H522, H1299, A-427, CALU-l, and CALU-6. The 160-kD form is present in B104, H1466, SK-LU , and CALU-3. This figure is representative of 10 Western blots performed for p185HER2 •
200.0 -
3
4
5
6
7
8
9
10
11
92.5-
69.0 -
69.0 C!i
U :::li lD a-
M
N
:I:
:ll l'3 C'l :I:
~
l::l
II>
:I:
2
kD 200.0 -
92.5-
iii
361
;Z N
i:
.'"'"
i:
'" ;1; .(
=>
III
It: III
:::li
..J
UJ
~
C!i
iii
U :::li lD a-
3
4
5
:§:
'" '"
.(
II>
:I:
N
i:
... ...
N
6
7
- '"
:) ..J et U
:) ..J et U
8
M
:) ..J et U
362
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992
TABLE I
specific phenomenon. Furthermore, the amount of pi 85HER2 expressed by the cell lines did not correlate with HER2 gene amplification (r = 0.41, P = 0.10). This suggests that mechanisms in addition to gene amplification are involved in overexpression of pi 85HER2 • A 3.4-kb EcoRI fragment was found in three of the cell lines (CALU-l , CALU-6, H522) and PBMC DNA from one normal volunteer. This fragment may represent a DNA polymorphism in the human HER2 gene in addition to other polymorphisms that have previously been described (7, 11).
P185HER2 expression in non-small cell lung cancer cell lines pI85 HER'*
HER2 mRNAt
(%)
(%)
P N
0 100
0 100
P
5
Tumor Type
Cell Lines PBMC B104 Rat type II pneumocytes Bronchiole epithelium NCI-H1264 NCI-H23 A-549 NCI-358 A-427 SK-LU NCI-H522 NCI-H1466 CALU-3 NCI-H520 CALU-6 CALU-I NCI-H596 NCI-H1299 SK-MES
P A A BA BA BA A A A A S S S AS LC M
0 4 5 12 15 18 27 32 70 113 0 7 9
10
8 10 21 54 27 20 47 59 220 6 19 22 12
13 0
12
HER2 Gene Amplification;
HER2 mRNA Levels Steady-state levels of HER2 mRNA were next analyzed by Northern blotting to determine if differences in HER2 mRNA resulted in the elevated levels of pl85 HER2 expression. The signals obtained by autoradiography were digitally analyzed and normalized to the B104 control. Digital quantitation was confirmed by quantitative dot blot analysis of serial dilutions of total cellular RNA derived from each cell line (data not shown). The full-length human HER2 cDNA probe identified a single 4.8-kb mRNA transcript in the lung cancer cell lines (Figure 4). HER2 mRNA was not found or was present at very low levels in pl85 HER2 nonexpressing cell lines (H520, SK-MES) and in PBMC (Table 1). A large accumulation of HER2 mRNA was found in the CALU-3 cell line which correlated with its degree of HER2 gene amplification and protein expression. In the remaining cell lines, there was a good correlation between pl85 HER2 levels and HER2 mRNA accumulation, but not between HER2 gene amplification and mRNA accumulation, suggesting that changes in HER2 gene expression and not gene amplification result in increased p185HER2 expression in lung cancer cell lines.
1 1 1 1 1 I 1 1 8 1 1 1 1 1 1
13
Definition of abbreviations: PBMC = peripheral blood mononuclear cells; P = primary cultures; N = neuroblastoma; A = adenocarcinoma; BA = bronchioloalveolar cell carcinoma; S = squamous cell carcinoma; AS = adenosquamous; LC = large cell carcinoma; M = mesothelioma. * Western blot analysis was performed as detailed in MATER1ALS AND METHODS. The values reported represent the results of 10 separate gels and six individual experiments. t Northern blot analysis was performed as detailed in MATERlALS AND METHODS. The values reported represent the results of six separate gels and four individual experiments. t Genomic Southern blot analysis was performed as detailed in MATERIALS AND METHODS. Values reported represent the results of four separate experiments. Amplificationwas defined as a 2-fold or greater increase in densitometric intensity of the hybridization signal as referenced to PBMC.
Discussion Oncogenic mechanisms for pl85 HER2 include point mutations resulting in dramatic biochemical and enzymatic changes (14) and HER2 gene amplification with protein overexpression (9). We have previously reported that the HER2 proto-oncogene protein product, pi 85HER2 , is quantitatively abnormal in human non-small cell lung cancer (6, 8). In cell lines derived from human non-small cell lung cancer, we now show that: (1) p185HER2 is biochemically similar to pl85 HER2 present in other human cancers; (2) by Western blot analysis, p185HER2 is overexpressed as compared with normal pulmonary epithelial cells; and (3) the quantitative abnormality in p185HER2 expression is best correlated with increased level of HER2 mRNA and not HER2 gene amplification. Because p185HER2 is normally expressed in the human lung (6, 8), the level of expression in lung cancer becomes
with a related human tyrosine kinase cDNA, the EGFR (Figure 3). All HER2 values were then corrected for the level of EGFR DNA as densitometrically determined from the autoradiogram. After this analysis, no additional cell lines were found to have amplification of the HER2 gene, and the same cell line remained amplified to the same degree. As the EGFR gene may be amplified in squamous cell carcinomas, this correction could underreport HER2 gene amplification in this histologic type oflung cancer. However, Southern analysis showed no amplification of the EGFR gene in any of the cell lines studied . Therefore, amplification of the HER2 gene in the CALU-3 cell line appears to be a 2
kb 14 .0 8.1 6.6 4.7 -
0
::;; aJ
a.
2.0 -
M N
J:
3
CD It> M
J:
4
0
N It>
J:
5
N N It>
J:
6
'" 0>
It>
J:
7
;l;
8
0> 0>
N
N
i:
i:
9
10
11
. . ..
'" '" i:
....
N
:)
It>
..J
To identify mechanisms that allow p185HER2 expression in lung cancer, we performed Western, Southern, and Northern blot analyses of 14 cell lines derived from human non-small cell lung carcinomas and one cell line derived from a human mesothelioma. Human bronchiole epithelial cells and rat type II pneumocytes were found to express p185HER2 at low to undetectable levels by Western blot technique. In contrast, 13 lung cancer cell lines expressed pI85 H ER2 , and eight of these 13 expressed p185HER2 at levels at least 2-fold higher than that found in normal bronchiole epithelial cells or type II pneumocytes. Genomic Southern analysis showed that amplification of the HER2 gene was present in only one of the eight cell lines that expressed p185HER2 at these higher levels. Increased levels of steady-state HER2 mRNA occurred in the remaining seven cell lines. We conclude that in human non-small cell lung carcinoma cell lines the most common mechanism resulting in increased p185HER2 expression is due to mechanisms that increase HER2 mRNA levels, with HER2 gene amplification occurring less commonly.
Proto-oncogenes represent normal cellular genes that were identified on the basis of their similarity to genetic sequences with known tumorigenic or transforming potential. Considerable circumstantial evidence now exists showing that acquired alterations in either the structure, copy number, or expression of these genes may play a role in the pathogenesis of some human malignancies. One such gene called HER2 or c-erbB-2 is a member of the tyrosine kinase supergene family and encodes a membrane-bound tyrosine kinase glycoprotein (pI85 HER2 ) with structural similarity to the epidermal growth factor receptor (EGFR) (1-4). Based on structural homology, it is postulated to be a growth factor receptor and a specific ligand has recently been isolated that may allow testing of this hypothesis (5). Recently, we reported expression of p185HER2 in 30% of human lung cancer cell lines and biopsy specimens (6). Schneider and associates (7) have also found HER2 mRNA in non-small cell lung cancer cell lines and freshly resected non-small cell lung cancer. The expression of p185H ER2 in lung cancer has important clinical correlates, as we have found that in pulmonary adenocarcinomas it independently identifies an older patient population (66.6 ± 10.1 versus 57.5 ± 10.8 yr of age; P = 0.04) and patients with shorter (Received in original form May 8, 1991 and in revised form September 16, 1991) Address correspondence to: Jeffrey A. Kern, M.D., Department ofInternal Medicine, Room C-33A, GH, University ofIowa College of Medicine, Iowa City, IA 52242. Abbreviations: epidermal growth factor receptor, EGFR; fetal bovine serum, FBS; Hanks' balanced salt solution, HBSS; peripheral blood mononuclear cell(s), PBMC; sodium dodecyl sulfate polyacrylamide gel electrophoresis, SDS-PAGE. ' Am. J. Respir. Cell Mol. BioI. Vol. 6. pp. 359-363, 1992
survival (83.7 ± 94.1 versus 188.5 ± 120 wk; P = 0.01) (8). Therefore, expression in human lung adenocarcinomas may be more predictive for clinical outcome than any other prognosticator presently known, with the exception of surgical stage. Interestingly, p185H ER2 expression in lung tumors is at levels significantly higher than in normal bronchiole epithelium (6, 8). The mechanism(s) allowing increased p185HER2 expression in human non-small cell lung cancer is unknown. In human breast and ovarian cancer, HER2 gene amplification is common, resulting in p185HER2 overexpression (9). But in primary human lung cancers, HER2 gene amplification has rarely been found (7, 9, 10). Therefore, we hypothesized that mechanisms other than HER2 gene amplification occur in human lung cancer to result in increased p185H ER2 expression. To test this hypothesis, we performed Western blot analysis of rat type II pneumocytes, human bronchiole epithelial cells, and SV-40 immortalized tracheal cells to establish normal levels of p185HER2 expression in the lung. To identify mechanisms involved in p185H ER2 overexpression in human lung cancer, Western, Southern, and Northern blot analyses were next performed on human non-small cell lung cancer cell lines. We found that increased p185HER2 expression is associated with increased levels of HER2 mRNA, whereas gene amplification occurs less commonly. These results have important implications for understanding the etiology of lung neoplasia.
Materials and Methods Cell Culture The rat neuroblastoma cell line BI04 or the breast adenocarcinoma cell line SK-BR-3 was utilized as a positive control in Northern blot and Western blot analyses. These lines are
360
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992
known to express p185HERZ to high levels (3, 11). Human lung tumor cell lines were obtained from American Type Culture Collection (ATCC) (Rockville, MD) or established from biopsy or aspiration specimens as part of approved clinical protocols. They included human adenocarcinomas National Cancer Institute H23, H522, H1264, H1466, H358, ATCC CALU-3, A-427, A-549, SK-LU; large-cell carcinoma, NCI H1299; squamous-cell carcinomas NCI H520, ATCC CALU-1, CALU-6; mesothelioma ATCC SK-MES; and adenosquamous carcinoma NCI H596. The cells were cultured in RPMI-1640 media, 10% fetal bovine serum (FBS), or Eagle's modified essential media and 10% FBS and incubated at 37° C in a humidified 5 % COz atmosphere. The BEAS-2B cell line was obtained from Curtis C. Harris (Laboratory of Human Carcinogenesis, National Cancer Institute, Bethesda, MD). It was derived from human tracheal epithelium immortalized by SV-40 viral infection and maintained in culture according to previously published methods (12). SK-BR-3 was obtained from ATCC and maintained in culture according to their recommendations. p185HERz-specific Rabbit Antiserum Polyclonal anti-p185HERz antiserum was prepared by immunizing rabbits with a synthetic peptide corresponding to amino acid residues 1240 through 1255 of the human p185HERZ sequence coupled to keyhole limpet hemocyanin. This antiserum has been previously described in detail (6, 8, 13, 14). Preparation of Rat Type II Pneumocytes and Human Bronchiole and Nasal Epithelial Cells Rat type II cells were derived from pathogen-free male Sprague-Dawley rats by enzymatic digestion of the alveoli and differential adherence as previously described (15). Bronchiole epithelial cells were obtained from human lung tissue removed from patients undergoing thoracotomy and lobectomy for lung cancer. The primary tumor was immediately resected from the specimen, and bronchioles from areas of lung uninvolved with tumor were resected and opened longitudinally. The epithelial cell layer and basement membrane were removed by sharp dissection, washed in Hanks' balanced salt solution (HBSS), recovered by centrifugation, and lysed in protein lysis buffer. A nasal polyp was obtained from a patient undergoing polypectomy. Immediately after resection, the polyp was sectioned under sterile conditions to remove the epithelial layer and basement membrane from the underlying stroma. The epithelial cells were washed in calcium-free HBSS and cultured on 100-mm plates in RPMI and 10% FBS at 37° C in a humidified 5% COz atmosphere. Upon reaching confluency, the cells were lysed in protein lysis buffer. Western Blot Analysis Cells were disrupted in protein lysis buffer (20 mM Tris [pH 8], 0.33 M sucrose, 0.5 mM EGTA, 25 J.'M leupeptin, 1.5 J.'M aprotinin, 5 J.'M n-ethyl-maleimide, 0.5% Triton X-100). A total of 100 J.'g of total protein from each cell line was subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (7.5% gel, reducing condition) and electrically transferred to a nitrocellulose membrane. The nitrocellulose membrane was probed with DBW-2, washed, and incubated with ['z5I]protein A (sp act, 10 J.'Ci/p.g; New England Nuclear, Boston, MA). The membrane was then
washed with phosphate-buffered saline and exposed to Kodak XAR-5 film (Eastman Kodak, Rochester, NY). RNA Isolation and Northern Blot Analysis Total cellular RNA was isolated by a guanidium isothiocyanate method with cesium chloride modification (16, 17). RNA was size-fractionated under denaturing conditions as previously described (8). The RNA was transferred to nylon membranes and hybridized under standard conditions with 1Q6 cpm/lane of a heat-denatured 3zP-Iabeled plasmid DNA (18) with a full-length HER2 cDNA insert (4). Blots were washed under high stringency and exposed at -70° C to Kodak XAR-5 film. To ensure equal loading and transfer of RNA in all lanes, radiolabeled probes were removed from the membranes by incubation with 0.2 % SDS, 10 mM Tris at 85° C for 60 min, and blots were reprobed using a {j-actin control probe (19). Preparation of DNA and Southern Blot Analysis Approximately 5 x 107 cells were washed in PBS and genomic DNA extracted from proteinase K-digested cells by ethanol precipitation after deproteinization (20). Isolation of high-molecular-weight DNA was verified by examination of the DNA on a 0.3 % agarose gel. DNA was subjected to restriction endonuclease digestion with EcoRI, and DNA fragments resolved by electrophoresis in a 0.8% agarose gel using tx TAE (40 mM Tris-acetate, 1 mM EDTA) running buffer. DNA was transferred to nylon membranes after depurination according to the method of Southern (21). Membranes were hybridized with a 3zP-Iabeled HER2 cDNA (4, 18), washed twice in zx SSC, 0.1% SDS at 25° C for 30 min and twice in 0.1% SSC, 0.1% SDS at 55° C for 30 min, and exposed to Kodak XAR-5 film. To ensure equal loading and transfer of DNA among lanes, radiolabeled probes were removed from the membranes by incubation with 0.2 % SDS, 10 mM Tris at 85° C for 60 min and reprobed using the human EGFR cDNA (22).
Results p185 HER2 Is Expressed at Low Levels in Normal Human Bronchiole Epithelial Cells We first sought to determine the level of p185HERZ expression in normal pulmonary epithelial cells. Previous data using immunocytochemical techniques have shown that p185HERZis expressed at low levels in normal bronchiole epithelium and type II pneumocytes (6, 8). To confirm this, ciliated bronchiole epithelial cells were isolated from two human lung specimens obtained at thoracotomy, and p185HERZ was quantitated by Western analysis. In addition, rat type II pneumocytes were isolated from rat lungs and similarly evaluated for p185HERZ expression. As shown in Figure 1, rat type II pneumocytes expressed only 5 % of the amount of p185HERZ as compared with the B104 cell line, whereas human bronchiole epithelium expressed undetectable amounts. These low levels of biochemically defined p185HERZ expression agree with our previous immunocytochemical findings in normal pulmonary epithelial cells. In addition, the SV-40 immortalized, but nontransformed, tracheal cell line BEAS-2B and cultured nasal polyp epithelial cells had undetectable levels of p185HERz. We conclude from these studies that p185HERZ expression in nontransformed
Kern, Robinson, Gazdar et al.: pI85 HER2 Expression in Non-Small Cell Lung Cancer
kD 200-
proximately 5,000 D higher than pI85 HER2 identified in the B104 cell line . This difference has previously been ascribed to interspecies glycosylation differences. Interestingly, in some cell lines the reactive protein appeared as a doublet (H522, H1299, A-427, CALU-I, CALU-6) separated by a molecular mass of roughly 5,000 D. A 160-kD incompletely glycosylated form of pI85 HER2 was also identified in cell lines that expressed high levels of p185 HER2 (B104, HI 466, SK-LU, CALU-3) (1). No difference in pI85 HER2 molecular mass was found when examined under reducing or nonreducing SDS-PAGE conditions. To quantitate the level of pI85 HER2 expression among the cell lines, the autoradiogram hybridization signals were digitally quantitated and normalized to the signal obtained from the B104 cell line. All values obtained by digital quantitation were confirmed by quantitative dot blot analysis of serial dilutions of cellular proteins (data not shown). As shown in Table 1, expression ranged from undetectable levels (H520, SK-MES) to levels greater than that found in the B104 cell line (CALU-3). Eight of the 13 pI85 HER2 expressing cell lines expressed levels of pI85 HER2 that were at least 2-fold greater than the highest level found in normal human or rat tissue (Table 1). Indeed, one cell line (CALU-3) expressed levels of p185 HER2 higher than the rat cell line B104 and equal to the breast carcinoma cell line SK-BR-3 known to have an amplified HER2 gene (Figure 1) (11).
92.5 69-
43-
r M
...
0
iii
M
Ii:
::J
:: 1Il..J eta za-
a:
lD
N
rh
et
UJ
lD
= III UJUJ
III
a-I» I-U I-a et:::li
a:=> UJ
lD
Z
Antibody : DBW· 2
a-
Figure1. p185HER2 expression in normal human lung. p185HER2 was identified in 100 JA.g oftotal cell protein from uncultured bronchiole epithelial cells, cultured epithelial cells from a nasal polyp, an SV40 immortalized tracheal cell line (BEAS-2B) , uncultured rat type II pneumocytes, a breast adenocarcinoma cell line (SK-BR-3), and a lung adenocarcinoma cell line (CALU-3 ). This figure is representative of two Western blots.
Amplification of the HER2 Proto-oncogene in Human Lung Carcinoma Cell Lines Is Uncommon To identify mechanisms that allow pI85 HER2 to be expressed in the cell lines at levels higher than that found in normal bronchiole epithelium, we first assessed HER2 gene copy number in all cell lines by genomic Southern blot analysis . Autoradiogram hybridization . signals were quantitated by digital analysis and compared to a DNA sample prepared from peripheral blood mononuclear cells (PBMC) of a normal volunteer. EcoRI digests of genomic DNA derived from PBMC or the lung cancer cell lines resulted in four restriction fragments (14, 8.1, 6.6, 4.7 kb) (Figure 3) . One cell line, CALU-3, was found to have an 8-fold amplification of the HER2 gene (Figure 3, Table 1). The remaining cell lines had no amplification and therefore contain two HER2 genes. To control for equal DNA loading among lanes, all blots were probed
bronchiole epithelial cells is very low, at most 5 % of the expression levels observed in B104, the rodent tumor cell line . p185HERl Is Variably Expressed in Human Non-Small Cell Lung Cancer Cell Lines Fourteen cell lines derived from human non-small cell lung cancers and one mesothelioma cell line were subjected to Western blot analysis to identify pI85 HER2 expression. Thirteen of the 14 lung cancer cell lines expressed varying levels of pI85 HER2 , while one squamous cell carcinoma and the mesothelioma cell line did not express pI85 HER2 (Figure 2). The relative molecular mass of pI85 HER2 was 190 kD, ap-
2
Figure 2. p185HER2 expression
kD
in human lung cancer cell lines. p185HER2 was identified in 100 J1.g of total cell protein from each cell line. The presence of a DBW-2 185-kD reactive doublet can be seen in cell lines H522, H1299, A-427, CALU-l, and CALU-6. The 160-kD form is present in B104, H1466, SK-LU , and CALU-3. This figure is representative of 10 Western blots performed for p185HER2 •
200.0 -
3
4
5
6
7
8
9
10
11
92.5-
69.0 -
69.0 C!i
U :::li lD a-
M
N
:I:
:ll l'3 C'l :I:
~
l::l
II>
:I:
2
kD 200.0 -
92.5-
iii
361
;Z N
i:
.'"'"
i:
'" ;1; .(
=>
III
It: III
:::li
..J
UJ
~
C!i
iii
U :::li lD a-
3
4
5
:§:
'" '"
.(
II>
:I:
N
i:
... ...
N
6
7
- '"
:) ..J et U
:) ..J et U
8
M
:) ..J et U
362
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992
TABLE I
specific phenomenon. Furthermore, the amount of pi 85HER2 expressed by the cell lines did not correlate with HER2 gene amplification (r = 0.41, P = 0.10). This suggests that mechanisms in addition to gene amplification are involved in overexpression of pi 85HER2 • A 3.4-kb EcoRI fragment was found in three of the cell lines (CALU-l , CALU-6, H522) and PBMC DNA from one normal volunteer. This fragment may represent a DNA polymorphism in the human HER2 gene in addition to other polymorphisms that have previously been described (7, 11).
P185HER2 expression in non-small cell lung cancer cell lines pI85 HER'*
HER2 mRNAt
(%)
(%)
P N
0 100
0 100
P
5
Tumor Type
Cell Lines PBMC B104 Rat type II pneumocytes Bronchiole epithelium NCI-H1264 NCI-H23 A-549 NCI-358 A-427 SK-LU NCI-H522 NCI-H1466 CALU-3 NCI-H520 CALU-6 CALU-I NCI-H596 NCI-H1299 SK-MES
P A A BA BA BA A A A A S S S AS LC M
0 4 5 12 15 18 27 32 70 113 0 7 9
10
8 10 21 54 27 20 47 59 220 6 19 22 12
13 0
12
HER2 Gene Amplification;
HER2 mRNA Levels Steady-state levels of HER2 mRNA were next analyzed by Northern blotting to determine if differences in HER2 mRNA resulted in the elevated levels of pl85 HER2 expression. The signals obtained by autoradiography were digitally analyzed and normalized to the B104 control. Digital quantitation was confirmed by quantitative dot blot analysis of serial dilutions of total cellular RNA derived from each cell line (data not shown). The full-length human HER2 cDNA probe identified a single 4.8-kb mRNA transcript in the lung cancer cell lines (Figure 4). HER2 mRNA was not found or was present at very low levels in pl85 HER2 nonexpressing cell lines (H520, SK-MES) and in PBMC (Table 1). A large accumulation of HER2 mRNA was found in the CALU-3 cell line which correlated with its degree of HER2 gene amplification and protein expression. In the remaining cell lines, there was a good correlation between pl85 HER2 levels and HER2 mRNA accumulation, but not between HER2 gene amplification and mRNA accumulation, suggesting that changes in HER2 gene expression and not gene amplification result in increased p185HER2 expression in lung cancer cell lines.
1 1 1 1 1 I 1 1 8 1 1 1 1 1 1
13
Definition of abbreviations: PBMC = peripheral blood mononuclear cells; P = primary cultures; N = neuroblastoma; A = adenocarcinoma; BA = bronchioloalveolar cell carcinoma; S = squamous cell carcinoma; AS = adenosquamous; LC = large cell carcinoma; M = mesothelioma. * Western blot analysis was performed as detailed in MATER1ALS AND METHODS. The values reported represent the results of 10 separate gels and six individual experiments. t Northern blot analysis was performed as detailed in MATERlALS AND METHODS. The values reported represent the results of six separate gels and four individual experiments. t Genomic Southern blot analysis was performed as detailed in MATERIALS AND METHODS. Values reported represent the results of four separate experiments. Amplificationwas defined as a 2-fold or greater increase in densitometric intensity of the hybridization signal as referenced to PBMC.
Discussion Oncogenic mechanisms for pl85 HER2 include point mutations resulting in dramatic biochemical and enzymatic changes (14) and HER2 gene amplification with protein overexpression (9). We have previously reported that the HER2 proto-oncogene protein product, pi 85HER2 , is quantitatively abnormal in human non-small cell lung cancer (6, 8). In cell lines derived from human non-small cell lung cancer, we now show that: (1) p185HER2 is biochemically similar to pl85 HER2 present in other human cancers; (2) by Western blot analysis, p185HER2 is overexpressed as compared with normal pulmonary epithelial cells; and (3) the quantitative abnormality in p185HER2 expression is best correlated with increased level of HER2 mRNA and not HER2 gene amplification. Because p185HER2 is normally expressed in the human lung (6, 8), the level of expression in lung cancer becomes
with a related human tyrosine kinase cDNA, the EGFR (Figure 3). All HER2 values were then corrected for the level of EGFR DNA as densitometrically determined from the autoradiogram. After this analysis, no additional cell lines were found to have amplification of the HER2 gene, and the same cell line remained amplified to the same degree. As the EGFR gene may be amplified in squamous cell carcinomas, this correction could underreport HER2 gene amplification in this histologic type oflung cancer. However, Southern analysis showed no amplification of the EGFR gene in any of the cell lines studied . Therefore, amplification of the HER2 gene in the CALU-3 cell line appears to be a 2
kb 14 .0 8.1 6.6 4.7 -
0
::;; aJ
a.
2.0 -
M N
J:
3
CD It> M
J:
4
0
N It>
J:
5
N N It>
J:
6
'" 0>
It>
J:
7
;l;
8
0> 0>
N
N
i:
i:
9
10
11
. . ..
'" '" i:
....
N
:)
It>
..J
