Although human lung tumor-derived cell lines play an important role in the investigation of lung cancer biology and genetics, there is no comprehensive study comparing the genotypic and phenotypic properties of lung cancer cell lines with those of the individual tumors from which they were derived. We compared a variety of properties of 12 human non-small cell lung carcinoma (NSCLC) cell lines (cultured for a median period of 39 months; range, 12–69) and their corresponding archival tumor tissues. There was, in general, an excellent concordance between the lung tumor cell lines and their corresponding tumor tissues for morphology (100%), the presence of aneuploidy (100%), immunohistochemical expression of HER2/neu (100%) and p53 proteins (100%), loss of heterozygosity at 13 chromosomal regions analyzed (97%) using 37 microsatellite markers, microsatellite alterations (MAs, 75%), TP53 (67%), and K-ras (100%) gene mutations. In addition, there was 100% concordance for the parental allele lost in all 115 comparisons of allelic losses. Some discrepancies were found; more aneuploid subpopulations of cells were detected in the cell lines as well as higher incidences of TP53 mutations (4 of 10 mutations not found in the tumors) and microsatellite alterations (two cell lines with MAs not detected in the tumors). Similar loss of heterozygosity frequencies by chromosomal regions and mean fractional allelic loss index were detected between successfully cultured and 40 uncultured lung tumors (0.45 and 0.49, respectively), indicating that both groups were similar. Our findings indicate that the NSCLC cell lines in the large majority of instances retain the properties of their parental tumors for lengthy culture periods. NSCLC cell lines appear very representative of the lung cancer tumor from which they were derived and thus provide suitable model systems for biomedical studies of this important neoplasm.

Lung cancer is the most frequent cause of cancer deaths in both men and women in the United States (1), and tobacco smoking is accepted as the major cause of lung cancer (2). Lung cancer is classified into two major types, SCLC3 and NSCLC (3). Squamous cell carcinoma, adenocarcinoma, and large cell carcinoma are the major histological types of NSCLC (3). Several studies have revealed that multiple genetic changes (estimated to be at least 10 to 20) are found in clinically evident lung cancers involving known tumor suppressor genes and several dominant oncogenes, including myc family members, K-ras, TP53, RB, p16-CDKN2, and candidate tumor suppressor gene regions on chromosomes 3p, 5q22 (APC-MCC region), 8p, 11p, and others (4).

Cell lines established from human tumors provide an unlimited, self-replicating source of malignant cells free of contaminating stromal cells and can be studied by investigators throughout the world. Such permanent cultures derived from human lung tumors have been widely used to investigate almost every aspect of lung cancer biology (5). Despite the pivotal role played by human tumor cell lines in biomedical research, there is a widespread belief in the scientific community that they have substantially changed from the individual tumors from which they were derived. Lung carcinomas cell lines have extensive chromosomal rearrangements, oncogene mutations, and multiple sites of allelic loss and gene amplification (6). Thus, many investigators presume, without direct evidence, that change in phenotypic properties and additional molecular changes develop during the prolonged time required for cell culture establishment and subsequent passage. Our recent studies have indicated that this supposition is not correct. Although we found that only a subset of primary breast carcinomas can be successfully cultured, these breast cancer cell lines faithfully retained the properties of their parental tumors for lengthy culture periods (up to 60 months; Refs. 7 and 8), and thus they provide suitable model systems for studying a subset of breast cancers. Basically, we discovered that the breast cancer specimens that yielded tumor cell lines (11%) came from clinically advanced tumors with advanced genetic changes and poor prognosis (7, 8). It was the initial tumor specimen rather than changes in culture that generated breast tumor cell lines with multiple genetic alterations. Thus, we were interested to see if the same findings were true for lung cancer. As far as we could determine, other than myc amplification in SCLC tumors (9), no detailed comparison of the properties of human lung cancer cell lines with those of the tumors from which they were derived has been published. In the present study, we compared the morphological, phenotypic, and genetic changes in 12 NSCLC cell lines with those present in their corresponding tumor tissues. We found that, for the most part, lung cancer cell lines are very representative of the tumor specimen from which they were derived.

Cell Line and Tissue Samples.

Twelve human lung cancer lines (prefix HCC for Hamon Cancer Center) established in our laboratory during the period 1992–1997 from invasive lung carcinomas by published methods (10) and their corresponding archival paraffin-embedded tumor tissue were studied. All cell lines from which corresponding archival materials were available were used. Of these, 11 cell lines were established from primary lung tumors from patients undergoing curative intent resections and one from a malignant pleural effusion. The histological diagnoses were as follows: 5 adenocarcinomas, 3 squamous cell, 2 adeno-squamous cell, 1 large cell carcinoma, and 1 spindle plus giant cell carcinoma. The patients ranged in age from 43 to 69 years (median, 56.5). Analysis of the Tumor-Node-Metastasis staging indicated that these patients were suffering from varying degrees of tumor extent, ranging from stage I (six patients) through stage IV (one patient). Clinicopathological information of the cases is presented in Table 1. The median time period from date of initial culture until the cell lines were studied was 39 months (range, 12–69 months; Table 1). We also compared the findings from the allelotyping analysis with those of a test set of 40 archival NSCLCs from patients undergoing curative intent surgical resections that we did not attempt to culture (referred to as “sporadic” cancers). The latter set consisted of 20 adenocarcinomas and 20 squamous cell carcinomas. The sporadic lung cancer patients ranged in age from 30 to 84 years (median, 59.5), and they presented varying degrees of tumor extent (32 stage I, 4 stage II, 3 stage III, and 1 stage IV).

Morphology, Ploidy, and Immunohistochemical Analyses.

For morphology, ploidy, and immunohistochemical analyses, suspension cultures were harvested during log phase, whereas adherent cultures were harvested when semiconfluent, fixed briefly in formaldehyde, pelleted and suspended in agar (1.5%), and then embedded in paraffin. Five-μm sections were prepared from paraffin-embedded cell lines and corresponding archival tumor tissues. The histological sections of the cell pellets and their corresponding archival tumor tissues were analyzed for structural features including cell morphology, nuclear atypia, and presence of duct-like structures. For ploidy analysis, the sections were Feulgen-stained using Schiff’s thionin reagent. DNA ploidy was determined by image analysis using a Roche Pathology work station (Roche Image Analysis, Burlington, NC). A minimum of 200 cells was analyzed for ploidy using stromal cells as an internal control. The PI was calculated for each cell line and tumor tissue sample. The ploidy status was categorized as diploid (PI = 0.90–1.1) or aneuploid (PI ≥ 1.1).

Immunohistochemical analyses using primary mouse polyclonal antibodies for HER2/neu (Cerb-2; DAKO; dilution, 1:14,000) and p53 protein (DAKO; dilution, 1:80) expression were performed. Immunostaining was performed using a standard avidin-biotin immunoperoxidase method. A minimum of 15 microscopy fields or up to 30,000 μm2 of nuclear area was quantified for immunohistochemical expression by image analysis using the same Roche Pathology work station as for ploidy determination. Immunostained sections were scored, evaluating the incidence of positive-stained cells as follows: negative (−); low positivity (+) if expression was restricted to 0–30% of the cells; moderate positivity (++) for 30–70% positive cells; and strong positivity (+++) if >70% of the cells were positive.

TP53 and K-ras Gene Mutation Analyses.

We examined the lung cancer cell lines and corresponding tumor tissue genomic DNA for mutations in exons 5–8 of the TP53 gene by nested PCR methodology and SSCP analysis as described previously (11), followed by sequencing of both strands by automated cycle sequencing using an ABI PRISM 377 DNA Sequencer (Perkin-Elmer, Branchburg, NJ). For K-ras gene mutation analysis, we used a designed RFLP method, using nested PCR methodology, followed by sequencing, as described previously (12). Using the designed RFLP method, we screened for mutations in codons 12 and 13 of the K-ras gene. For the first step, amplification of DNA corresponding to exon 1 of the K-ras gene was performed by the nested PCR using mismatched primers in the second PCR to introduce a restriction site into the PCR product derived from the wild-type allele. The second PCR products thus obtained were digested using appropriate restriction enzymes. Wild-type alleles were digested and yielded a smaller product than mutant forms, which were digestion resistant. K-ras gene mutations so discovered were confirmed by direct sequencing.

LOH and MA Analyses.

PCR-LOH and MAs analyses were performed on genomic DNA extracted from cell lines with normal genomic DNA obtained from archival specimens. From archival paraffin-embedded tumor tissue samples, invasive lung carcinoma areas were identified and precisely dissected under microscopic visualization using laser capture microdissection (Arcturus Engineering, Inc., Mountain View, CA) from noncoverslipped, H&E-stained slides (13), and DNA extractions were performed as described previously (14). From multiple sections of each tissue sample, 1000–2000 sectioned cells were microdissected. Lymphocytes obtained from corresponding, resected, metastasis-free lymph nodes or stromal lung tissue provided a source for constitutional DNA for all analyses. Five μl of the proteinase K-digested samples, containing DNA from at least 100 sectioned cells, were used for each multiplex PCR reaction.

To evaluate LOH, we used primers flanking 37 microsatellite repeat polymorphisms spanning 13 chromosomal regions. The microsatellite markers and the chromosomal regions analyzed were as follows: 3p12 (D3S1274, D3S1284, and D3S1511), 3p14.2 (D3S4103 and D3S1300), 3p14–21 (D3S1766), 3p21 (D3S1573, D3S1447, Luca 2.2, ITIH-1, D3S1478, D3S1029, and D3S1582), 3p22–24 (D3S1612, D3S2432, D3S1351, and D3S1537), 3p25 (D3S1597), 5q22 (L5.71 in the APC-MCC region), 8p21 (D8S258, D8S133, D8S136, D8S137, and NEFL), 8p22 (D8S1145, D8S602, D8S254, D8S261, and LPL), 8p23 (D8S277, D8S1130, and D8S1106), 9p21 (IFNA and D9S1748), 13q14 (dinucleotide repeat at the RB gene), and 17p13 (TP53 dinucleotide and pentanucleotide repeats). Primer sequences can be obtained from the Genome Database, with five exceptions that have been previously published and referenced (15, 16). For all samples, multiplex PCR (up to six markers) was performed in the first amplification, followed by uniplex PCR for individual microsatellite markers as described previously (16). LOH was scored by visual detection of complete absence of the one tumor allele in heterozygous (i.e., “informative”) cases. MAs were detected by shifts in the mobility of one or both alleles (Fig. 4).

Data Analysis.

To compare the total frequencies of LOH between microdissected tumor tissue samples and cell lines, we used the FAL index. The FAL index is a measure of LOH for all of the informative chromosomal loci (maximum 37 loci) per tumor or cell line sample. The FAL index was calculated as follows:

\[FAL\ index\ {=}\ \frac{Total\ number\ of\ chromosomal\ loci\ with\ LOH}{Total\ number\ of\ informative\ loci\ examined}\]

To compare the total frequencies of MAs between microdissected tumor tissue samples and cell lines, we used the MA index. The MA index is a measure of MAs for all of the chromosomal loci (maximum, 37 loci) per tumor or cell line sample. The MA index was calculated as follows:

\[MA\ index\ {=}\ \frac{Total\ number\ of\ chromosomal\ loci\ with\ MA}{Total\ number\ of\ loci\ examined}\]

Statistical analyses was performed using the nonparametric Wilcoxon and Fisher exact tests. Probability values of P < 0.05 were regarded as statistically significant.

Morphological Features and Ploidy of the Cell Lines and Their Corresponding Tumor Tissues.

There was a complete concordance between the morphological features of the primary tumors and their corresponding cell lines (Fig. 1). All of the tumors demonstrated poorly differentiated features, as did the cell blocks of the tumor cell lines. Although the corresponding cell lines grew as monolayers of epitheloid cells, some gland-like formations and squamous differentiation were seen in most adenocarcinomas and squamous cell carcinoma cell lines, respectively (Fig. 1). The cultures consisted of medium- or large-sized cells with high-grade nuclear atypia and the presence of occasional multinucleated cells. The spindle and giant cell carcinoma demonstrated both cell components in the tumor as well as corresponding cell culture (Fig. 1).

Although all of the tumors and cell lines were aneuploid (Table 1), variation was noted in the PIs of the cell lines and their corresponding tumors (Table 1). Although a single population of aneuploid cells was detected in most (10 of 12; 83%) of the tumor tissues, seven cell lines (71%) demonstrated more than one population of aneuploid cells (e.g., polyploid).

Immunohistochemical Profile, TP53 and K-ras Mutations of the Cell Lines, and Their Corresponding Tumor Tissues.

Immunohistochemical expressions of the HER2/neu and p53 proteins (Fig. 2) demonstrated a complete concordance between tumors and cell lines for both markers (Table 2). Although all cell lines and their corresponding tumors overexpressed p53 protein, only three (25%) of the tumors demonstrated immunostaining for HER2/neu, two adenocarcinomas and one adeno-squamous cell carcinoma. HER2/neu immunostaining was focal in the tumor and corresponding cell lines in all positive cases. For both tumors and cell lines, overexpression of the p53 protein was localized in all cases to the cell nucleus, whereas HER2/neu expression was localized predominantly to the outer cell membrane (Fig. 2).

TP53 gene mutations were detected in 10 of the 12 (83%) lung cancer cell lines analyzed (Table 3). Identical mutations were detected in six cell lines and their corresponding tumor tissues (60%), and only wild-type forms were present in the corresponding normal tissues (Fig. 3,A). In four cell lines, TP53 mutations were found, but we failed to detect these mutations in the tumor specimen, despite their positive immunostaining for p53. K-ras gene mutations at codon 12 were detected in only 2 of 12 (17%) lung cancer cell lines analyzed, and identical mutations were detected in their corresponding tumor tissues (Table 3 and Fig. 3 B). Both of these cases were adenocarcinomas.

LOH Analysis of Cell Lines and Their Corresponding Tumor Tissues.

Because analyses were performed on archival tumor tissue samples that had been carefully microdissected and on cultures that were pure tumor cell populations, allelic losses for both sample types were detected as complete loss of one of the two alleles present in constitutional DNA of informative cases (Fig. 4). As expected from prior work, high frequencies of allelic loss (range, 44–100%) were detected at most of the 13 chromosomal regions analyzed in the archival tumor specimens (Table 2). The highest frequencies for allelic losses were at one or more 8p regions (100%), 9p21 region (78%), TP53 gene on 17p.13 (78%), and at one or more 3p regions (67%). In 145 of 150 (97%) comparisons for the 13 different chromosomal regions, the LOH or heterozygosity retention was identical in the cell line and corresponding tumor (Table 2). Likewise, in comparing results for the individual 37 markers, there was concordance in 239 of 269 (89%) of cases (Table 4). There was 100% concordance for retention of heterozygosity (124 of 124 cases). The discrepancies came in LOH, where there was 79% (115 of 145) concordance. This always represented additional LOH in the cell lines not detected in the tumors. Almost half of these discrepancies came in 2 of the 12 paired cell lines and tumor cases (HCC461 and HCC1359). However, in all 115 cases in which allelic losses of individual markers were present in both cell lines and corresponding tumors, the same parental allele was lost in the tumor and cell line (Table 4).

MA Analysis of Cell Lines and Their Corresponding Tumor Tissues.

Although lung cancer cell lines demonstrated a higher incidence of MAs than the corresponding tumor tissues, similar mean MAs indices were detected in both groups of lung cancer specimens (0.032 versus 0.026, respectively; P > 0.05). Eight of the 12 lung cancer cell lines and tumor comparison cases demonstrated MAs either in the cell line or tumor tissue, and in six cases they were present in both specimens (75% correlation). Sixteen of 29 (55%) MAs detected either in cell lines or tumor tissues were detected in the same microsatellite marker in both corresponding specimens. Of those, ten (66%) were identical (Fig. 4). Because artifacts resulting from PCR amplification may be mistaken for MAs, all examples of MAs were confirmed by a second PCR reaction.

Comparison of LOH and MA Analyses between Successfully Cultured and Sporadic Lung Tumors.

We compared and contrasted the molecular changes in successfully cultured lung cancers with those present in uncultured (sporadic) NSCLCs. Of the 37 markers covering 13 chromosomal regions, we found no significant differences between the cultured and the sporadic lung cancers in either LOH or MA frequency (Table 5), and almost identical high frequencies of total LOH were present in the two groups, as represented by the FAL indices (means, 0.45 and 0.49, respectively; P = 0.86; Fig. 5). In addition, similar relatively low frequencies of total MAs were detected in successfully cultured and sporadic lung tumors (Table 5). There were more advanced clinical stages (5 of 12 cases, stages III and IV) in the cultured series compared with the sporadic set (4 of 40 cases, stages III and IV). Nevertheless, the amount of genetic changes were not significantly different between the two groups.

Effect of Culture Time.

The median time in continuous culture of the cell lines prior to testing was 39 months (range, 12–69; Table 1). No correlation between the culture period and concordance of allelic loss was demonstrated. The cell lines with the greatest discordance of allelic loss when compared with their parental tumors were HCC1359 (concordance of 71%) and HCC461 (concordance of 75%), with 40 and 57 months in culture, respectively. However, three other cell lines, HCC95, HCC1171, and HCC1195, that had been cultured for periods of 42–69 months demonstrated 86, 92, and 100% concordance, respectively (Table 4).

Using a panel of 12 NSCLC cell lines and their corresponding archival lung tumor tissues, we determined whether the morphological, phenotypic, and genetic changes detected in the cell lines represent those present in their corresponding tumor tissues. There was an excellent correlation between the morphological, ploidy, and immunohistochemical features of the primary tumors and their corresponding cell lines. Aneuploidy (defined as an abnormal nuclear content of DNA) has been considered evidence of widespread genetic damage and DNA instability (17). In our series, all of the tumors and their cell lines demonstrated one or more aneuploid populations. However, the correlation between the specific degree of aneuploidy of the tumors and their corresponding cell lines was variable, suggesting that only one or more subpopulations of cells present in the original tumor were successfully cultured.

HER2/neu is overexpressed in over a one-third of uncultured NSCLCs, especially adenocarcinomas (18, 19, 20) and in NSCLC cell lines (21). Likewise, HER2/neu immunohistochemical expression was detected in 3 of 12 (25%) of the NSCLC cell lines and their corresponding tumors, showing a complete correlation between tumor samples and their corresponding cell lines. HER2/neu overexpression correlates with a shorter survival as demonstrated in some, but not all, studies of lung cancer (22, 21, 22, 23, 24).

Point mutations in the TP53 gene may result in variant p53 proteins that have an increased half-life and thus can be detected by immunohistochemistry, which fails to immunostain the low amounts of wild-type p53 protein in cells without TP53 mutations (25). TP53 gene mutations have been detected in ∼50% of NSCLCs (4, 26) and 74% of NSCLC cell lines (27). p53 protein expression, as demonstrated by immunohistochemistry, was detected in all of the lung tumor cell lines and their corresponding tumor tissues, indicating complete correlation (100%). However, TP53 gene mutations in exons 5–8 were detected in only 10 of 12 (83%) lung tumor cell lines. Six of those corresponding tumor tissues exhibited the identical TP53 gene mutation, whereas four of these tumors lacked TP53 mutations. These findings suggest either TP53 mutations occurred during in vitro culture or that mutant subpopulations in heterogeneous tumors were selected during in vitro culturing. These differences in TP53 mutations represented the most significant discrepancy between tumor cell lines and their corresponding primary tumors.

ras gene mutations occur in ∼15–20% of NSCLC, mainly in adenocarcinomas (20–30%; Ref. 4) and in 36% of NSCLC cell lines (27). Mutations in K-ras account for ∼90% of ras mutations in lung adenocarcinomas, with 85% of the K-ras mutations affecting codon 12 (4). K-ras gene mutations at codon 12 were detected in two adenocarcinoma cell lines (17% of the NSCLCs and 33% of adenocarcinoma cases), and identical K-ras mutations were identified in their corresponding tumors.

We determined allelic loss at 13 chromosomal regions deleted frequently in lung cancers using 37 polymorphic microsatellite markers. Nearly identical high LOH frequencies at all chromosomal regions analyzed were detected between tumors and their corresponding cell lines (97% concordance). For all of the individual markers, there was an excellent correlation between tumors and cell lines (89% concordance). In all of the 115 (100%) comparisons, when allelic loss of a particular informative microsatellite was present in both the tumor and corresponding cell line, the identical parental allele was lost in both, confirming that the allelic loss originated in the original tumor tissue.

In addition to allelic losses, alterations in microsatellite size (MAs) are another genetic change associated with several human cancers (28). MAs represent changes in the size of simple nucleotide repeat polymorphic microsatellite markers, resulting in altered electrophoretic mobility of one or both alleles. In lung cancers, they have been reported to occur at frequencies ranging from 0 to 45% (29, 30). Although the mechanisms underlying MAs are presently unknown, they may represent a manifestation of genomic instability (28). There was a good correlation in the MA frequency between lung cancer cell lines and their corresponding tumor tissues. Genomic instability is a characteristic feature of many tumors, and it may develop early during pathogenesis (17). Genomic instability persists after tumor development, resulting in the frequent appearance of multiple subclonal populations (31), and presumably, the instability continues during culture life. Because cell cultures frequently have population doubling times considerably shorter than those of in vivo tumors, the frequency of mutational change in cultures may be more rapid than their corresponding tumors. However, the tumor cells did not exhibit a greater frequency of this phenomenon in culture and after lengthy culture periods. Identical MAs in cell lines and tumor tissues were detected in cases that had been cultured for periods of time ranging from 12 to 69 months.

Other than differences in the degree of aneuploidy, the incidence of TP53 mutations, and MAs, the other properties studied demonstrated a remarkable degree of concordance between tumors and their corresponding cancer cell lines. These features included morphological characteristics, presence of aneuploidy, immunohistochemical expression profile for HER2/neu and p53 protein, K-ras and types of TP53 gene mutations, and a similar allelic loss pattern for multiple loci deleted frequently in lung carcinoma. Our studies were performed on cell lines cultured for a median period of 39 months. The concordance between tumors and cell lines for all of the comparisons was independent of the time on culture, indicating that the properties of cell lines usually closely resemble those of their parental tumors for culture periods up to 69 months.

The overall LOH and MAs incidences and the frequencies of allelic loss by chromosomal regions analyzed detected in our cultured NSCLC tumors are similar to the LOH frequencies found in uncultured (sporadic) lung cancers. In addition, our successfully cultured NSCLCs demonstrated similar frequencies of K-ras (17% versus 20–30%) and TP53 (50%) mutations compared with the literature (4, 26). These findings indicate that successfully cultured NSCLCs present similar genetic changes to those found in typical NSCLCs. Recently, we have reported that human breast cancer cell lines retain the phenotypic and genotypic properties for lengthy culture periods (up to 60 months; Ref. 8). However, our analyses indicated that only a subset of primary breast carcinomas that have several features indicative of advanced tumors with poor prognosis, including more advanced genetic changes, can be cultured (7, 8). Our present findings indicate that successfully cultured NSCLCs are representative of the general population of lung cancers and that cell lines are useful models for studying this important type of lung neoplasm. However, because the successful culture of a primary NSCLC tumor is an independent negative prognostic factor (32), we presume that successfully cultured tumors have certain, as yet unknown, changes not present in early-stage sporadic tumors.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

        
1

Supported by the Specialized Program of Research Excellence Grant P50-CA70907 from the National Cancer Institute, NIH, Bethesda, MD.

                
3

The abbreviations used are: SCLC, small cell lung carcinoma; NSCLC, non-SCLC; PI, ploidy index; SSCP, single-strand conformation polymorphism; LOH, loss of heterozygosity; MA, microsatellite alteration; FAL, fractional allelic loss.

Fig. 1.

Comparison of morphology (H&E stained) between lung cancer cell lines and their corresponding tumor tissues. Tumor tissue (a) and the corresponding HCC95 cell line (b) showing squamous cell differentiation with keratinization features are shown. Tumor tissue (c) and the corresponding cell line HCC1833 (d) showing poorly differentiated adenocarcinoma features with gland-like structures. Tumor tissue from a spindle and giant cell carcinoma (e) and corresponding HCC1359 cell line (f and g) shows both giant cell (f) and spindle (g) components.

Fig. 1.

Comparison of morphology (H&E stained) between lung cancer cell lines and their corresponding tumor tissues. Tumor tissue (a) and the corresponding HCC95 cell line (b) showing squamous cell differentiation with keratinization features are shown. Tumor tissue (c) and the corresponding cell line HCC1833 (d) showing poorly differentiated adenocarcinoma features with gland-like structures. Tumor tissue from a spindle and giant cell carcinoma (e) and corresponding HCC1359 cell line (f and g) shows both giant cell (f) and spindle (g) components.

Close modal
Fig. 2.

Comparison of immunostaining between lung cancer cell lines and their corresponding tumor tissues. Tumor tissue (a) and the corresponding HCC1833 cell line (b) showing HER2/neu immunostaining. Note that HER2/neu immunostaining is focal and predominantly located in the outer cell membrane in both tumor tissue and cell line. Tumor tissue (c) and the corresponding HCC2279 cell line (d) shows p53 nuclear immunostaining.

Fig. 2.

Comparison of immunostaining between lung cancer cell lines and their corresponding tumor tissues. Tumor tissue (a) and the corresponding HCC1833 cell line (b) showing HER2/neu immunostaining. Note that HER2/neu immunostaining is focal and predominantly located in the outer cell membrane in both tumor tissue and cell line. Tumor tissue (c) and the corresponding HCC2279 cell line (d) shows p53 nuclear immunostaining.

Close modal
Fig. 3.

Comparison of TP53 and K-ras gene mutations between lung cancer cell lines and their corresponding tumor tissues. A, SSCP analysis (left) of tumor tissue (T), corresponding HCC1833 cell line (CL) and normal tissue (L, lymphocytes) for TP53 mutation at exon 8, followed by sequencing (right). Both cell line and corresponding tumor tissue demonstrated the same bandshift in the SSCP analysis (arrowhead). In the sequencing analysis, a 9-bp deletion [indicated by dotted line in the L (wild-type) sequencing panel] was detected in both tumor (T) and cell line (CL) paired specimens. B, designed RFLP analysis of K-ras mutations at codon 12 of tumor tissue (T), corresponding HCC461 cell line (CL) and normal tissue (L, lymphocytes) (left), followed by sequencing (right). The nondigested, 97-bp band represents the mutant type, and the digested 77-bp band represents the wild-type codon 12 of the K-ras gene. Although the tumor sample had both wild (GGT) and mutant (GAT) types, the CL specimen had only the mutant form (GAT). SM, 100-bp size marker; PC, positive control; NG negative control.

Fig. 3.

Comparison of TP53 and K-ras gene mutations between lung cancer cell lines and their corresponding tumor tissues. A, SSCP analysis (left) of tumor tissue (T), corresponding HCC1833 cell line (CL) and normal tissue (L, lymphocytes) for TP53 mutation at exon 8, followed by sequencing (right). Both cell line and corresponding tumor tissue demonstrated the same bandshift in the SSCP analysis (arrowhead). In the sequencing analysis, a 9-bp deletion [indicated by dotted line in the L (wild-type) sequencing panel] was detected in both tumor (T) and cell line (CL) paired specimens. B, designed RFLP analysis of K-ras mutations at codon 12 of tumor tissue (T), corresponding HCC461 cell line (CL) and normal tissue (L, lymphocytes) (left), followed by sequencing (right). The nondigested, 97-bp band represents the mutant type, and the digested 77-bp band represents the wild-type codon 12 of the K-ras gene. Although the tumor sample had both wild (GGT) and mutant (GAT) types, the CL specimen had only the mutant form (GAT). SM, 100-bp size marker; PC, positive control; NG negative control.

Close modal
Fig. 4.

Comparison of allelic loss among lung cancer cell lines and their corresponding tumor tissues. Ten representative examples of LOH and MAs in paired lung cancer cell lines and their corresponding invasive tumors are shown. LOH in both sample types was detected by complete absence of one of the two alleles present in constitutional DNA. Bars, the positions of the major allelic bands. Arrowheads, either alleles lost (LOH) or shifted (MAs). L, lymphocytes or stromal tissue (as source of constitutional DNA); T, microdissected tumor tissue; CL, cell line.

Fig. 4.

Comparison of allelic loss among lung cancer cell lines and their corresponding tumor tissues. Ten representative examples of LOH and MAs in paired lung cancer cell lines and their corresponding invasive tumors are shown. LOH in both sample types was detected by complete absence of one of the two alleles present in constitutional DNA. Bars, the positions of the major allelic bands. Arrowheads, either alleles lost (LOH) or shifted (MAs). L, lymphocytes or stromal tissue (as source of constitutional DNA); T, microdissected tumor tissue; CL, cell line.

Close modal
Fig. 5.

Comparison of LOH in successfully established cultured and uncultured (sporadic) lung carcinomas using the FAL Index (an indicator of LOH at all informative loci analyzed per case).

Fig. 5.

Comparison of LOH in successfully established cultured and uncultured (sporadic) lung carcinomas using the FAL Index (an indicator of LOH at all informative loci analyzed per case).

Close modal
Table 1

Clinicopathological data of the tumors from which the lung cancer cell lines were established and phenotypic properties of lung cancer cell lines and their corresponding tumor tissues

Cell lineTumor sourceLung cancer histological typeAgeSexTumor size (cm)LNa metastasisTNM stageCulture initiation dateGrowth period(months)Morphological differentiation T/CLPloidy index T/CL
HCC95 Pl. effusion Squamous 65 NA NA IV 6/25/92 69 Poor/Poor 2.4/1.5–2.5 
HCC461 Pr. Lung Adenocarcinoma 69 3.0 1/5 IIIA 5/27/93 57 Poor/Poor 1.6/1.6 
HCC1171 Pr. Lung Adenocarcinoma 58 1.5 0/6 9/7/94 42 Poor/Poor 1.6/1.8–2.2 
HCC1195 Pr. Lung Adeno-squamous 47 4.5 0/10 9/15/94 42 Poor/Poor 2.4–2.9/2–2.4–2.8 
HCC1313 Pr. Lung Squamous 51 2.0 3/22 IIIA 11/2/94 40 Poor/Poor 1.8/1.5–2.4 
HCC1359 Pr. Lung Spindle-giant cell 55 5.5 0/9 IIIA 11/22/94 40 Poor/Poor 2.2/1–1.6 
HCC1438 Pr. Lung Large cell 43 1.5 0/8 1/9/95 38 Poor/Poor 1.7/1.4 
HCC1588 Pr. Lung Squamous 63 7.0 0/6 3/20/95 36 Poor/Poor 3/1.7 
HCC1833 Pr. Lung Adenocarcinoma 69 3.5 0/5 8/9/95 31 Poor/Poor 1.8/1.6 
HCC2108 Pr. Lung Adenocarcinomas 59 4.0 1/8 II 2/6/96 25 Poor/Poor 2/1–1.4–1.8 
HCC2279 Pr. Lung Adeno-squamous 52 4.3 2/4 IIIA 5/30/96 21 Poor/Poor 1.6–2.6/1.6–2 
HCC2450 Pr. Lung Adenocarcinoma 52 1.6 0/4 3/13/97 12 Poor/Poor 2.8/2.9 
Cell lineTumor sourceLung cancer histological typeAgeSexTumor size (cm)LNa metastasisTNM stageCulture initiation dateGrowth period(months)Morphological differentiation T/CLPloidy index T/CL
HCC95 Pl. effusion Squamous 65 NA NA IV 6/25/92 69 Poor/Poor 2.4/1.5–2.5 
HCC461 Pr. Lung Adenocarcinoma 69 3.0 1/5 IIIA 5/27/93 57 Poor/Poor 1.6/1.6 
HCC1171 Pr. Lung Adenocarcinoma 58 1.5 0/6 9/7/94 42 Poor/Poor 1.6/1.8–2.2 
HCC1195 Pr. Lung Adeno-squamous 47 4.5 0/10 9/15/94 42 Poor/Poor 2.4–2.9/2–2.4–2.8 
HCC1313 Pr. Lung Squamous 51 2.0 3/22 IIIA 11/2/94 40 Poor/Poor 1.8/1.5–2.4 
HCC1359 Pr. Lung Spindle-giant cell 55 5.5 0/9 IIIA 11/22/94 40 Poor/Poor 2.2/1–1.6 
HCC1438 Pr. Lung Large cell 43 1.5 0/8 1/9/95 38 Poor/Poor 1.7/1.4 
HCC1588 Pr. Lung Squamous 63 7.0 0/6 3/20/95 36 Poor/Poor 3/1.7 
HCC1833 Pr. Lung Adenocarcinoma 69 3.5 0/5 8/9/95 31 Poor/Poor 1.8/1.6 
HCC2108 Pr. Lung Adenocarcinomas 59 4.0 1/8 II 2/6/96 25 Poor/Poor 2/1–1.4–1.8 
HCC2279 Pr. Lung Adeno-squamous 52 4.3 2/4 IIIA 5/30/96 21 Poor/Poor 1.6–2.6/1.6–2 
HCC2450 Pr. Lung Adenocarcinoma 52 1.6 0/4 3/13/97 12 Poor/Poor 2.8/2.9 
a

LN, mediastinal lymph node status (number positive/number examined); TNM, Tumor-Node-Metastasis; T, tumor; CL, cell line; Pl., pleural; Pr., primary; NA, not applicable.

Table 2

Comparison of ploidy, immunohistochemistry, and LOH frequencies among 12 lung tumor cell lines and their corresponding tumor tissues

Comparison of tumor tissue/cell line
FeatureFrequency Tumor tissue/Cell line+/++/−−/+−/−Concordance
Protein expressiona       
 HER2/neu 25%/25% 12/12 (100%) 
 p53 protein 100%/100% 12 12/12 (100%) 
LOH frequencies by chromosomal regionsb       
 3p25 38%/38% 8/8 (100%) 
 3p22–24 55%/55% 11/11 (100%) 
 3p21 58%/58% 12/12 (100%) 
 3p14–21 22%/22% 9/9 (100%) 
 3p14.2 (FHIT gene) 50%/50% 10/10 (100%) 
 3p12 25%/25% 8/8 (100%) 
 Any 3p 67%/67% 12/12 (100%) 
 5q22 (APC-MCC region) 44%/44% 9/9 (100%) 
 8p23 91%/91% 10 11/11 (100%) 
 8p22 83%/92% 10 11/12 (92%) 
 8p21 58%/75% 10/12 (83%) 
 Any 8p 100%/100% 12 12/12 (100%) 
 9p21 78%/89% 8/9 (89%) 
 13q (RB gene) 33%/33% 6/6 (100%) 
 17p (TP53 gene) 78%/89% 8/9 (89%) 
 Total  92 53 145/150 (97%) 
MA 42%/58% 9/12 (75%) 
Gene mutations       
TP53 gene (exons 5–8) 50%/83% 8/12 (67%) 
 K-ras gene (codons 12–13) 17%/17% 12/12 (100%) 
Comparison of tumor tissue/cell line
FeatureFrequency Tumor tissue/Cell line+/++/−−/+−/−Concordance
Protein expressiona       
 HER2/neu 25%/25% 12/12 (100%) 
 p53 protein 100%/100% 12 12/12 (100%) 
LOH frequencies by chromosomal regionsb       
 3p25 38%/38% 8/8 (100%) 
 3p22–24 55%/55% 11/11 (100%) 
 3p21 58%/58% 12/12 (100%) 
 3p14–21 22%/22% 9/9 (100%) 
 3p14.2 (FHIT gene) 50%/50% 10/10 (100%) 
 3p12 25%/25% 8/8 (100%) 
 Any 3p 67%/67% 12/12 (100%) 
 5q22 (APC-MCC region) 44%/44% 9/9 (100%) 
 8p23 91%/91% 10 11/11 (100%) 
 8p22 83%/92% 10 11/12 (92%) 
 8p21 58%/75% 10/12 (83%) 
 Any 8p 100%/100% 12 12/12 (100%) 
 9p21 78%/89% 8/9 (89%) 
 13q (RB gene) 33%/33% 6/6 (100%) 
 17p (TP53 gene) 78%/89% 8/9 (89%) 
 Total  92 53 145/150 (97%) 
MA 42%/58% 9/12 (75%) 
Gene mutations       
TP53 gene (exons 5–8) 50%/83% 8/12 (67%) 
 K-ras gene (codons 12–13) 17%/17% 12/12 (100%) 
a

As determined by immunostaining.

b

For LOH studies: +, LOH; −, retention of heterozygosity.

Table 3

Concordance for TP53 mutations and LOH and K-ras mutations among lung cancer cell lines and corresponding tumor tissues

GeneCell line17p13 LOH T/CLaExon/CodonTumorBase substitution/Amino acid change cell line
TP53 HCC461 LOH/LOH Exon 5/Codon 148 None detected GAT to GGb 
TP53 HCC1171 LOH/LOH Exon 7/Codon 247 AAC to ATC/Asn to Ile AAC to ATC/Asn to Ile 
TP53 HCC1195 LOH/LOH Exon 8/Codons 262, 263 GGT AAT to CCT GAT/Gly Asn to Pro Asp GGT AAT to CCT GAT/Gly Asn to Pro Asp 
TP53 HCC1359 LOH/LOH Exon 5/Codon 130 CTC to TTC/Leu to Phe CTC to TTC/Leu to Phe 
TP53 HCC1438 NI/NI Exon 5/Codon 139 None detected 1 base pair deletion (G)b 
TP53 HCC1588 NI/NI Exon 8/Codons 296 to 304 Deletion CAC CAC GAG CTG CCC CCA GGG AGC Ab Deletion CAC CAC GAG CTG CCC CCA GGG AGC Ab 
TP53 HCC1833 No LOH/No LOH Exon 8/Codons 284 to 286 Deletion ACA GAG GAAb Deletion ACA GAG GAAb 
TP53 HCC2108 LOH/LOH Exon 5/Codon 154 None detected GGC to GTC/Gly to Val 
TP53 HCC2279 No LOH/LOH Exon 7/Codon 234 None detected TAC to TGC/Tyr to Cys 
TP53 HCC2450 LOH/LOH Exon 7/Codon 249 AGG to ACG/Arg to Thr AGG to ACG/Arg to Thr 
K-ras HCC461 NA Exon 1/Codon 12 GGT to GAT/Gly to Asp GGT to GAT/Gly to Asp 
K-ras HCC1171 NA Exon 1/Codon 12 GGT to TGT/Gly to Cys GGT to TGT/Gly to Cys 
GeneCell line17p13 LOH T/CLaExon/CodonTumorBase substitution/Amino acid change cell line
TP53 HCC461 LOH/LOH Exon 5/Codon 148 None detected GAT to GGb 
TP53 HCC1171 LOH/LOH Exon 7/Codon 247 AAC to ATC/Asn to Ile AAC to ATC/Asn to Ile 
TP53 HCC1195 LOH/LOH Exon 8/Codons 262, 263 GGT AAT to CCT GAT/Gly Asn to Pro Asp GGT AAT to CCT GAT/Gly Asn to Pro Asp 
TP53 HCC1359 LOH/LOH Exon 5/Codon 130 CTC to TTC/Leu to Phe CTC to TTC/Leu to Phe 
TP53 HCC1438 NI/NI Exon 5/Codon 139 None detected 1 base pair deletion (G)b 
TP53 HCC1588 NI/NI Exon 8/Codons 296 to 304 Deletion CAC CAC GAG CTG CCC CCA GGG AGC Ab Deletion CAC CAC GAG CTG CCC CCA GGG AGC Ab 
TP53 HCC1833 No LOH/No LOH Exon 8/Codons 284 to 286 Deletion ACA GAG GAAb Deletion ACA GAG GAAb 
TP53 HCC2108 LOH/LOH Exon 5/Codon 154 None detected GGC to GTC/Gly to Val 
TP53 HCC2279 No LOH/LOH Exon 7/Codon 234 None detected TAC to TGC/Tyr to Cys 
TP53 HCC2450 LOH/LOH Exon 7/Codon 249 AGG to ACG/Arg to Thr AGG to ACG/Arg to Thr 
K-ras HCC461 NA Exon 1/Codon 12 GGT to GAT/Gly to Asp GGT to GAT/Gly to Asp 
K-ras HCC1171 NA Exon 1/Codon 12 GGT to TGT/Gly to Cys GGT to TGT/Gly to Cys 
a

T, tumor; CL, cell line; NI, not informative; NA, not applicable.

b

Resulting in a frame shift mutation.

Table 4

Concordance for retention or LOH

Cell lineCulture time (months)aRetention of heterozygositybLOH tumor/Cell lineConcordancecLoss of the same alleled
HCC95 69 2/2 10/12 12/14 (86%) 10/10 (100%) 
HCC461 57 13/13 5/11 18/24 (75%) 5/5 (100%) 
HCC1171 42 15/15 8/10 23/25 (92%) 8/8 (100%) 
HCC1195 42 11/11 8/8 19/19 (100%) 8/8 (100%) 
HCC1313 40 3/3 18/22 21/25 (84%) 18/18 (100%) 
HCC1359 40 15/15 5/13 20/28 (71%) 5/5 (100%) 
HCC1438 38 13/13 8/10 21/23 (91%) 8/8 (100%) 
HCC1588 36 2/2 15/18 17/20 (85%) 15/15 (100%) 
HCC1833 31 10/10 11/11 21/21 (100%) 11/11 (100%) 
HCC2108 25 12/12 13/13 25/25 (100%) 13/13 (100%) 
HCC2279 21 14/14 5/8 19/22 (86%) 5/5 (100%) 
HCC2450 12 14/14 9/9 23/23 (100%) 9/9 (100%) 
Total  124/124 (100%) 115/145 (79%) 239/269 (89%) 115/115 (100%) 
Cell lineCulture time (months)aRetention of heterozygositybLOH tumor/Cell lineConcordancecLoss of the same alleled
HCC95 69 2/2 10/12 12/14 (86%) 10/10 (100%) 
HCC461 57 13/13 5/11 18/24 (75%) 5/5 (100%) 
HCC1171 42 15/15 8/10 23/25 (92%) 8/8 (100%) 
HCC1195 42 11/11 8/8 19/19 (100%) 8/8 (100%) 
HCC1313 40 3/3 18/22 21/25 (84%) 18/18 (100%) 
HCC1359 40 15/15 5/13 20/28 (71%) 5/5 (100%) 
HCC1438 38 13/13 8/10 21/23 (91%) 8/8 (100%) 
HCC1588 36 2/2 15/18 17/20 (85%) 15/15 (100%) 
HCC1833 31 10/10 11/11 21/21 (100%) 11/11 (100%) 
HCC2108 25 12/12 13/13 25/25 (100%) 13/13 (100%) 
HCC2279 21 14/14 5/8 19/22 (86%) 5/5 (100%) 
HCC2450 12 14/14 9/9 23/23 (100%) 9/9 (100%) 
Total  124/124 (100%) 115/145 (79%) 239/269 (89%) 115/115 (100%) 
a

Ranked by cell line time period of culture.

b

Number of informative loci with retention of heterozygosity in both tumors and cell lines.

c

Concordance for both retention and LOH.

d

For comparisons when both tumor and cell line demonstrated LOH for an individual marker.

Table 5

Comparison of LOH and MA frequencies between successfully cultured and uncultured (sporadic) lung carcinomas

Molecular abnormalityCultured tumors(n = 12)Uncultured tumors(n = 40)P
LOH    
 3p25 3/8 (38%) 17/30 (57%) 0.28 
 3p22–24 6/11 (55%) 27/40 (68%) 0.33 
 3p21 7/12 (58%) 25/40 (63%) 0.53 
 3p14–21 2/9 (22%) 15/27 (56%) 0.08 
 3p14.2 (FHIT gene) 4/10 (40%) 26/39 (67%) 0.12 
 3p12 2/8 (25%) 23/37 (62%) 0.06 
 Any 3p 8/12 (67%) 33/40 (83%) 0.22 
 5q22 (APC-MCC region) 4/9 (44%) 3/26 (12%) 0.50 
 8p23 10/11 (91%) 27/40 (68%) 0.12 
 8p22 10/12 (83%) 21/38 (55%) 0.80 
 8p21 7/12 (58%) 27/40 (68%) 0.41 
 Any 8p 12/12 (100%) 36/40 (90%) 0.41 
 9p21 7/9 (78%) 20/31 (65%) 0.30 
 13q (RB gene) 2/6 (33%) 8/21 (38%) 0.36 
 17p (TP53 gene) 7/9 (78%) 13/19 (68%) 0.47 
FAL indexa mean  (range) 0.45 (0.19–0.75) 0.49 (0.05–1.0) 0.86 
MA frequency 5/12 (39%) 18/40 (45%) 0.55 
MA indexa mean (range) 0.026 (0–0.13) 0.030 (0–0.17) 0.40 
Molecular abnormalityCultured tumors(n = 12)Uncultured tumors(n = 40)P
LOH    
 3p25 3/8 (38%) 17/30 (57%) 0.28 
 3p22–24 6/11 (55%) 27/40 (68%) 0.33 
 3p21 7/12 (58%) 25/40 (63%) 0.53 
 3p14–21 2/9 (22%) 15/27 (56%) 0.08 
 3p14.2 (FHIT gene) 4/10 (40%) 26/39 (67%) 0.12 
 3p12 2/8 (25%) 23/37 (62%) 0.06 
 Any 3p 8/12 (67%) 33/40 (83%) 0.22 
 5q22 (APC-MCC region) 4/9 (44%) 3/26 (12%) 0.50 
 8p23 10/11 (91%) 27/40 (68%) 0.12 
 8p22 10/12 (83%) 21/38 (55%) 0.80 
 8p21 7/12 (58%) 27/40 (68%) 0.41 
 Any 8p 12/12 (100%) 36/40 (90%) 0.41 
 9p21 7/9 (78%) 20/31 (65%) 0.30 
 13q (RB gene) 2/6 (33%) 8/21 (38%) 0.36 
 17p (TP53 gene) 7/9 (78%) 13/19 (68%) 0.47 
FAL indexa mean  (range) 0.45 (0.19–0.75) 0.49 (0.05–1.0) 0.86 
MA frequency 5/12 (39%) 18/40 (45%) 0.55 
MA indexa mean (range) 0.026 (0–0.13) 0.030 (0–0.17) 0.40 
a

Indicator of LOH at all microsatellite markers (n = 37) analyzed by case (range, 0–1).

1
Landis S. H., Murray T., Bolden S., Wingo P. A. Cancer statistics, 1998.
CA Cancer J. Clin.
,
48
:
6
-29,  
1998
.
2
Parkin D. M., Pisani P., Lopez A. D., Masuyer E. At least one in seven cases of cancer is caused by smoking. Global estimates for 1985.
Int. J. Cancer
,
5904
:
494
-504,  
1994
.
3
Colby T. V., Koss M. N., Travis W. D. Tumors of the lower respiratory tract, 3rd series, fascicle 13
1
-554, Armed Forces Institute of Pathology Washington, DC  
1995
.
4
Sekido Y., Fong K. M., Minna J. D. Progress in understanding the molecular pathogenesis of human lung cancer.
Biochim. Biophys. Acta
,
1378
:
F21
-F59,  
1998
.
5
Gazdar A. F., Minna J. D. NCI series of cell lines: an historical perspective.
J. Cell. Biochem. Suppl.
,
24
:
1
-11,  
1996
.
6
Minna J. D., Sekido Y., Fong K., Gazdar A. F. Molecular biology of lung cancer DeVita V. T., Jr. Hellman S. Rosenberg S. A. eds. .
Cancer: Principles and Practice of Oncology
, Ed. 5
849
-857, Lippincott Philadelphia  
1997
.
7
Gazdar A. F., Kurvari V., Virmani A., Gollahon L., Sakaguchi M., Westerfield M., Kodagoda D., Stasny V., Cunningham T., Wistuba I. I., Tomlinson G., Tonk V., Ashfaq R., Leitch M., Minna J. D., Shay J. W. Characterization of paired tumor and non-tumor cell lines established from patients with breast cancer.
Int. J. Cancer
,
78
:
766
-774,  
1998
.
8
Wistuba I. I., Behrens C., Milchgrub S., Syed S., Ahmadian M., Virmani A. K., Cunningham T. H., Ashfaq R., Minna J. D., Gazdar A. F. Comparison of features of human breast cancer cell lines and their corresponding tumors.
Clin. Cancer Res.
,
4
:
2931
-2938,  
1998
.
9
Johnson B. E., Makuch R. W., Simmons A. D., Gazdar A. F., Burch D., Cashell A. W. myc family DNA amplification in small cell lung cancer patients’ tumors and corresponding cell lines.
Cancer Res.
,
48
:
5163
-5166,  
1988
.
10
Oie H. K., Russell E. K., Carney D. N., Gazdar A. F. Cell culture methods for the establishment of the NCI series of lung cancer cell lines.
J. Cell. Biochem. Suppl.
,
24
:
24
-31,  
1996
.
11
Wistuba I. I., Montellano F. D., Milchgrub S., Virmani A. K., Behrens C., Chen H., Ahmadian M., Nowak J. A., Muller C., Minna J. D., Gazdar A. F. Deletions of chromosome 3p are frequent and early events in the pathogenesis of uterine cervical carcinoma.
Cancer Res.
,
57
:
3154
-3158,  
1997
.
12
Sugio K., Molberg K., Albores-Saavedra J., Virmani A. K., Kishimoto Y., Gazdar A. F. K-ras mutations and allelic loss at 5q and 18q in the development of human pancreatic cancers.
Int. J. Pancreatol.
,
21
:
205
-217,  
1997
.
13
Emmert-Buck M. R., Bonner R. F., Smith P. D., Chuaqui R. F., Zhuang Z., Goldstein S. R., Weiss R. A., Liotta L. A. Laser capture microdissection.
Science (Washington DC)
,
274
:
998
-1001,  
1996
.
14
Hung J., Kishimoto Y., Sugio K., Virmani A., McIntire D. D., Minna J. D., Gazdar A. F. Allele-specific chromosome 3p deletions occur at an early stage in the pathogenesis of lung carcinoma.
J. Am. Med. Assoc.
,
273
:
558
-563,  
1995
.
15
Fondon J. W., III, Mele G. M., Brezinschek R. I., Cummings D., Pande A., Wren J., O’Brien K. M., Kupfer K. C., Wei M. H., Lerman M., Minna J. D., Garner H. R. Computerized polymorphic marker identification: experimental validation and a predicted human polymorphism catalog.
Proc. Natl. Acad. Sci. USA
,
95
:
7514
-7519,  
1998
.
16
Wistuba I. I., Behrens C., Milchgrub S., Virmani A. K., Jagirdar J., Thomas B., Ioachim H. L., Litzky L. A., Brambilla E. M., Minna J. D., Gazdar A. F. Comparison of molecular changes in lung cancers in HIV-positive and HIV-indeterminate subjects.
J. Am. Med. Assoc.
,
279
:
1554
-1559,  
1998
.
17
Loeb L. A. Cancer cells exhibit a mutator phenotype.
Adv. Cancer Res.
,
72
:
25
-56,  
1998
.
18
Schneider P. M., Hung M. C., Chiocca S. M., Manning J., Zhao X. Y., Fang K., Roth J. A. Differential expression of the c-erbB-2 gene in human small cell and non-small cell lung cancer.
Cancer Res.
,
49
:
4968
-4971,  
1989
.
19
Shi D., He G., Cao S., Pan W., Zhang H. Z., Yu D., Hung M. C. Overexpression of the c-erbB-2/neu-encoded p185 protein in primary lung cancer.
Mol. Carcinog.
,
5
:
213
-218,  
1992
.
20
Weiner D. B., Nordberg J., Robinson R., Nowell P. C., Gazdar A., Greene M. I., Williams W. V., Cohen J. A., Kern J. A. Expression of the neu gene-encoded protein (P185neu) in human non-small cell carcinomas of the lung.
Cancer Res.
,
50
:
421
-425,  
1990
.
21
Kern J. A., Robinson R. A., Gazdar A., Torney L., Weiner D. B. Mechanisms of p185HER2 expression in human non-small cell lung cancer cell lines.
Am. J. Respir. Cell Mol. Biol.
,
6
:
359
-363,  
1992
.
22
Tateishi M., Ishida T., Mitsudomi T., Kaneko S., Sugimachi K. Prognostic value of c-erbB-2 protein expression in human lung adenocarcinoma and squamous cell carcinoma.
Eur. J. Cancer
,
27
:
1372
-1375,  
1991
.
23
Kern J. A., Slebos R. J., Top B., Rodenhuis S., Lager D., Robinson R. A., Weiner D., Schwartz D. A. c-erbB-2 expression and codon 12 K-ras mutations both predict shortened survival for patients with pulmonary adenocarcinomas.
J. Clin. Invest.
,
93
:
516
-520,  
1994
.
24
Pfeiffer P., Clausen P. P., Andersen K., Rose C. Lack of prognostic significance of epidermal growth factor receptor and the oncoprotein p185HER-2 in patients with systemically untreated non-small cell lung cancer: an immunohistochemical study on cryosections.
Br. J. Cancer
,
74
:
86
-91,  
1996
.
25
Baas I. O., Mulder J. W., Offerhaus G. J., Vogelstein B., Hamilton S. R. An evaluation of six antibodies for immunohistochemistry of mutant p53 gene product in archival colorectal neoplasms.
J. Pathol.
,
172
:
5
-12,  
1994
.
26
Greenblatt M. S., Bennett W. P., Hollatein M., Harris C. C. Mutations in the p53 tumor suppressor gene: clue to cancer etiology and molecular pathogenesis.
Cancer Res.
,
54
:
4855
-4878,  
1994
.
27
Mitsudomi T., Viallet J., Mulshine J. L., Linnoila R. I., Minna J. D., Gazdar A. F. Mutations of ras genes distinguish a subset of non-small-cell lung cancer cell lines from small-cell lung cancer cell lines.
Oncogene
,
6
:
1353
-1362,  
1991
.
28
Fishel R. Genomic instability, mutators, and the development of cancer: is there a role for p53?.
J. Natl. Cancer Inst.
,
88
:
1608
-1609,  
1996
.
29
Fong K. M., Zimmerman P. V., Smith P. J. Microsatellite instability and other molecular abnormalities in non-small cell lung cancer.
Cancer Res.
,
55
:
28
-30,  
1995
.
30
Adachi J. I., Shiseki M., Okazaki T., Ishimuru G., Noguchi M., Hirohashi S., Yokota Y. Microsatellite instability in primary and metastatic lung carcinomas.
Genes Chromosomes Cancer
,
14
:
301
-306,  
1995
.
31
Nowell P. C. The clonal evolution of tumor cell populations.
Science (Washington DC)
,
194
:
23
-28,  
1976
.
32
Stevenson H., Gazdar A. F., Phelps R., Linnoila R. I., Ihde D. C., Ghosh B., Walsh T., Woods E. L., Oie H., O’Connor T., Mulshine J. L. Tumor cell lines established in vitro: an independent prognostic factor for survival in non-small-cell lung cancer.
Ann. Intern. Med.
,
113
:
764
-770,  
1990
.