Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Lung cancer is the leading cause of cancer death in the United States (1). At least 90% of the human lung cancer is associated with cigarette smoking (2). However, the mechanisms of pathogenesis of lung cancer are poorly understood. It is believed that tumorigenesis of lung cancer, similar to that of colorectal cancer, occurs through multiple and sequential morphologic and molecular changes (3, 4).

In the respiratory tract, squamous metaplasia of the tracheobronchial epithelium has been considered a preneoplastic change. Squamous metaplasia of the respiratory epithelium occurs in vitamin-A deficiency (5), or by exposure to carcinogen (6) or chronic tobacco smoke (7, 8). Concomitant with squamous differentiation as evidenced by the appearance of cellular cornification, our laboratory has identified a small proline-rich protein (SPR1) that is overexpressed in the primate airway epithelium (9). SPR1 is downregulated by vitamin A and upregulated by a tumor promoter, phorbol ester (10, 11). However, SPR1 expression is markedly diminished or lost in lung cancer (12, 13). In a deletion study of the 5'-flanking region of the SPR1 gene, we demonstrated that the first 98 base pairs (bp) relative to the transcription start site contained basal promoter activity, and that the region between 96 and 162 bp contained phorbol ester-responsive element (TRE) and TRE-like enhancer elements (14).

We postulate that expression of SPR1 is closely linked to multistep carcinogenesis of lung cancer. The progressive loss of SPR1 expression through the process of carcinogenesis is likely a consequence of dysregulation of the SPR1 promoter. In this report we studied the expression and regulation of SPR1 in a series of human bronchial epithelial cell lines with graded levels of tumorigenic potential representing different stages of malignant transformation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Human Bronchial Epithelial Cell Lines

A series of human bronchial epithelial cell lines with graded levels of tumorigenicity were cultured. These cell lines included: primary human tracheobronchial epithelial cells (HTBE) isolated from tracheobronchial tissue of a trauma patient without a history of smoking (15); a papilloma virus- immortalized human tracheobronchial epithelial cell line (HBE1) (16); an HBE1-C line derived from HBE1 after exposure for 24 passages to a tobacco carcinogen, N-methylacetoxynitrosamine (2 µm) (NCI Chemical Carcinogen Repository, Kansas City, MO) and phorbol 12-myristate 13-acetate (PMA) (100 ng/ml); a simian virus-40-immortalized human bronchial epithelial cell line (BEAS-2B) (17, 18); and a bronchogenic carcinoma cell line, H460, derived from a human large-cell lung carcinoma (19). The H460 cell line was cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and penicillin/streptomycin. The other cell lines were maintained in serum-free Ham's F12 medium supplemented with insulin (5 µg/ml), transferrin (5 µg/ml), epidermal growth factor (20 ng/ml), cholera toxin (20 ng/ml), dexamethasone (0.1 µm), and bovine hypothalamus extract (30 µg/ml) as previously described (15).

Anchorage-Independent Assay

For in vitro testing of tumorigenicity of each cell line, the anchorage-independent assay was performed using a modified soft-agar clonogenic method as previously described (20). Relative colony-forming efficiency was determined by counting the number of colonies (defined as aggregates of 10 or more cells) in a 1-cm² agar well.

Transplantation of Cell Lines in Nude Mice

Each bronchial cell line was tested for its in vivo tumorigenicity in nude mice as described by Reddel and colleagues (21). Cells were scraped off culture flasks and ~ 5 × 10⁶ cells in 0.1 ml of culture medium were inoculated subcutaneously in the lateral thoracic regions of female athymic nude mice. Tumor formation was followed for a period of 3 mo.

Telomerase Assay

Telomerase activity of each cell line was quantified by using a polymerase chain reaction (PCR) enzyme-linked immunosorbent assay (ELISA) kit (Boehringer Mannheim, Indianapolis, IN) based on the principles of telomeric repeat amplification protocol (TRAP) described by Kim and associates (22). Extracts were obtained from cells grown to approximately 80% confluency. Protein concentration was determined by the Bio-Rad DC Protein Microassay (Bio-Rad, Hercules, CA), and an equal amount of protein from each cell extract was used for the ELISA assay. To prove that the assay was an RNA-dependent enzymatic reaction, cell extract of H460 was preincubated with ribonuclease (RNase) (1 µg/µl) for 20 min before the ELISA assay. Relative telomerase activity of each cell line was expressed as percent of absorbance at 450 nm of the cell extract to that of H460.

Analysis of K-ras and p53 Mutations

Mutations of p53 (exons 5 to 8) and K-ras (exon 1) are commonly detected in bronchial preneoplasia and frank carcinoma (4, 23). Genomic DNA was isolated from each cell line. Genomic mutations of p53 (exons 5 to 8) and K-ras (exon 1) were analyzed using the technique of nonradioisotopic single-strand conformational polymorphism (SSCP) and DNA sequencing as described by Lau and coworkers (24).

Expression of SPR1: Reverse Transcription and PCR

A semi-quantitative reverse transcription (RT)/PCR method was used to measure relative expression of SPR1 messenger RNA (mRNA) at basal level and under the influence of vitamin A (1 µm) and PMA (100 ng/ml) (25). This method was more sensitive than Northern blotting in detecting very low levels of SPR1 transcripts in some of the cell lines. Total cell RNA was prepared according to the single-step method of Chomczynski and Sacchi (26). The downstream primer, 5'CGTTTGCAGCATGAGTTC3', and the upstream primer, 5'TTCAGAGACTCAGAGTG3', amplified the SPR1 complementary DNA extending from nucleotides #36 to #475 and yielded a PCR product of 440 bp. For internal control, a primer pair of 5'GAGAAAATCTGGCACCACAC3' and 5'TACCCCTCGTAGATGGGCAC3' was used to amplify a 259-bp product of -actin. Conditions for PCR were: 95°C for 30 s, 55°C for 1 min, and 72°C for 30 s for 28 cycles, with a 10-min extension at 72°C. PCR cycle range-finding study revealed that optimal PCR product signal was observed between 25 and 30 cycles without reaching a plateau. The relative level of SPR1 mRNA was expressed as a ratio of optical density of the SPR1 band to that of the -actin.

Expression of SPR1: Western Blotting

Whole-cell protein was extracted with keratin extraction buffer and protein concentration was determined by the Bio-Rad DC Protein Microassay (Bio-Rad). An equal amount of protein was loaded to each lane of a 15% polyacrylamide gel for electrophoresis. Western blotting for SPR1 was performed as previously described (12, 27). A primary rabbit polyclonal antiserum to a polypeptide, corresponding to the first 15 amino acids of the C-terminal region of the human SPR1 peptide, was used at a concentration of 1:1,000. A secondary biotinylated goat-antirabbit immunoglobulin G (Bio-Rad) was used at a concentration of 1:2,000. Detection of SPR1 signals was accomplished by using streptavidin-peroxidase and ECL Reagents (Amersham Life Sciences, Arlington Heights, IL).

Transfection of Chimeric SPR1 Promoter and Chloramphenicol Acetyltransferase Reporter Plasmids

Constructs of the 5'-flanking region of the SPR1 gene were inserted into a pBL-CAT3 plasmid carrying a reporter gene, chloramphenicol acetyltransferase (CAT), as previously described by Reddy and colleagues (14). Thus, CAT3 plasmids containing the SPR1 5'-flanking nucleotides of 96 to +9 and 162 to +9 relative to the transcription start site were designated, respectively, as 96-CAT3 and 162-CAT3. Each of these plasmids was transfected into each cell line with Lipofectin (GIBCO BRL, Gaithersburg, MD) (14). A pSV -galactosidase (-Gal) reporter plasmid (Clontech, Palo Alto, CA) was used as an internal control for transfection efficiency. Each cell culture was exposed to 2 µg of a plasmid DNA and 2 to 8 µl of Lipofectin in 1 ml of serum-free medium for 4 h, followed by incubation in fresh culture medium for 48 h. After transfection, each cell line was also exposed to PMA (50 ng/ml) or vitamin A (1 µm) for 24 h before preparation of cell extract. Cell extracts were assayed for CAT and -Gal levels using ELISA kits (Boehringer Mannheim). Relative CAT level was expressed as a ratio of CAT level to that of -Gal.

Statistical Analysis

Significant differences (P < 0.05) were determined by analysis of variance (ANOVA) using the StatView Version 4.02 (Abacus Concepts, Inc., Berkeley, CA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Tumorigenicity: Anchorage-Independent Assay and Transplantation in Nude Mice

Results of the anchorage-independent assay and transplantation in nude mice are shown in Table 1. For the anchorage-independent assay, no colony was detected with the primary culture of HTBE and the HBE1 cell line. Relative colony-forming efficiency was very low for HBE1-C and BEAS-2B, with colonies of 1 ± 1 and 2 ± 1 detected per well, respectively. For H460, more than 50 colonies were observed in each agar well. For the in vivo tumorigenic assay, only the H460 cells formed detectable tumors in nude mice.

Telomerase Assay

Relative telomerase activities of the cell lines measured by TRAP ELISA are shown in Table 1. Among the cell lines tested, H460 exhibited the highest telomerase activity comparable with that of a positive-control cell extract provided with the ELISA kit. Hence, the relative activity of H460 was expressed as 100%. No telomerase activity was detected with the HTBE. There was a progressive increase of telomerase activity from HBE1 to HBE1-C, BEAS-2B, H460, as shown in Table 1. The relative telomerase activities among the cell lines were significantly different (P < 0.05) as determined by ANOVA. The telomerase activity of H460 could be totally abolished by RNase, which verified that we were indeed measuring the activity of an RNA-dependent enzyme.

p53 and K-ras Mutations

No genomic mutation of exons 5 to 8 of p53 and exon 1 of K-ras was found in any of the cell lines analyzed by SSCP and DNA sequencing (data not shown).

SPR1 mRNA Expression

Expressions of SPR1 and -actin mRNA Detected by RT-PCR are shown in Figure 1. The relative levels of basal SPR1 expression among the cell lines were 1, 1, 0.7, 0.3, and 0.0 (mean of two determinations) for HTBE, HBE1, HBE1-C, BEAS-2B, and H460, respectively. For HTBE, HBE1, and HBE1-C, SPR1 level was upregulated by 30 to 50% after treatment with 100 ng/ml of PMA for 24 h, and was downregulated by 15 to 50% after exposure to 1 µm of vitamin A for 24 h. For BEAS-2B or H460, no effect on SPR1 expression was observed with PMA or vitamin-A treatment.

View larger version (22K): [in this window] [in a new window]   Figure 1.   Expression of SPR1 and -actin mRNA detected by RT-PCR. Of each cell line: lane 1, control; lane 2, cells pretreated with the phorbol ester PMA for 24 h; lane 3, cells pretreated with vitamin A for 24 h.

SPR1 Protein Expression: Western Blotting

Expression of SPR1 protein detected by Western blot analysis is shown in Figure 2. The relative levels of basal SPR1 expression among the cell lines were 1, 0.6, 0.6, 0.0, and 0.0 (mean of two determinations) for HTBE, HBE1, HBE1-C, BEAS-2B, and H460, respectively. For HTBE, HBE1, and HBE1-C, SPR1 was upregulated by 100% with PMA pretreatment. After exposure to vitamin A, SPR1 expression was completely abolished in HTBE, whereas its expression was suppressed by approximately 30% in HBE1 and HBE1-C. For BEAS-2B or H460, no effect was observed on SPR1 expression after treatment with PMA or vitamin A. 

View larger version (16K): [in this window] [in a new window]   Figure 2.   Expression of SPR1 detected by Western blotting. Of each cell line: lane 1, control; lane 2, cells pretreated with the phorbol ester PMA for 24 h; lane 3, cells pretreated with vitamin A for 24 h.

Promoter Study

Figures 3A and 3B illustrate the relative CAT levels in HBE1, HBE1-C, BEAS-2B, and H460 transfected with the chimeric 96-CAT3 and 162-CAT3 plasmids. For HTBE, similar experiments were previously conducted and are reported elsewhere (14). For the 96-CAT3 plasmid, low levels of control CAT were detected in descending order in HBE1, HBE1-C, and BEAS-2B, whereas no CAT was detected in H460 (Figure 3A). The control CAT levels were significantly different among the cell lines (P < 0.05). No effect on CAT expression was observed with PMA or vitamin A. For the 162-CAT3 plasmid, CAT level was readily detectable for HBE1, HBE1-C, and BEAS-2B cell lines, but barely detectable for H460 (Figure 3B), as shown by the respective CAT levels of 0.49, 0.31, 0.21, and 0.04. For this plasmid, CAT level was significantly (P < 0.05) upregulated from the control level by PMA for HBE1, HBE1-C, and BEAS-2B, but not for H460. However, no significant suppression of CAT by vitamin A was observed for all the cell lines.

View larger version (31K): [in this window] [in a new window]   Figure 3.   Relative levels of the reporter gene CAT, ligated to two constructs of the 5'-flanking promoter region of SPR1 gene. (A) Chimeric construct of the first 96 nucleotides. (B) Chimeric construct of the first 162 nucleotides. Control, cells without pretreatment; + PMA, cells pretreated with the phorbol ester PMA for 24 h; + Vit A, cells pretreated with vitamin A for 24 h. # Significant differences (P < 0.05) in CAT levels among the cell lines. * Significant increase (P < 0.05) in CAT level after PMA pretreatment. Mean ± standard deviation (SD) (n = 4).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Tumorigenesis is believed to occur through multiple-step genetic changes as elucidated for colorectal cancer by Fearon and Volgelstein (3). Such a multistep pathway is difficult to demonstrate in the pathogenesis of lung cancer because the human lung is not easily accessible for close observation and tissue biopsy. Although experimental proof is incomplete, it is believed that morphologic and molecular changes leading to lung cancer also occur in a stepwise fashion (4). There are few experimental models suitable for studying multistep carcinogenesis of lung cancer. Animal models have been employed for studying chemical-induced bronchial carcinogenesis (28). However, animal models are expensive to maintain and the tracheobronchial tree, similar to that of the human, is not readily accessible for sequential tissue sampling.

In this study, we have reported the identification of a series of cell lines that appears to be an appropriate model for multistep bronchial carcinogenesis. This series of cell lines demonstrates increasing levels of telomerase activity indicative of progressive potential of tumorigenicity. It has been shown by other investigators that an elevated telomerase activity is associated with an increase in likelihood of malignant transformation (29). This is further corroborated by the results of the anchorage-independent assay and in vivo transplant study in mice. In addition, we have observed squamous differentiation of HBE and HBE1-C, but no discernable histology for BEAS-2B and H460, in suspended cultures (unpublished observations).

In the respiratory tract, squamous metaplasia occurs in vitamin-A deficiency (5) or by exposure to carcinogen (6) and chronic tobacco smoke (7, 8). A previous study from our laboratory revealed that SPR1 was overexpressed in association with squamous differentiation in the bronchial epithelium (9). Further, SPR1 could be downregulated by vitamin A and upregulated by the tumor promoter phorbol ester (10, 11) and by tobacco smoke (unpublished data). These observations indicate that regulation of SPR1 expression is closely linked to early neoplastic transformation of the tracheobronchial epithelium. Using the in vitro model representing various steps of bronchial malignant transformation, we were able to characterize the expression and regulation of SPR1 as a molecular marker for multistep bronchial carcinogenesis.

In this study we demonstrated that a relatively high level of SPR1 expression was a rule in HTBE, HBE1, and HBE1-C in the absence of vitamin A. However, there was a gradual decline and loss of SPR1 expression with progression of carcinogenesis as exemplified by BEAS-2B and H460. Concomitant with malignant progression, there was a gradual loss of regulation of SPR1 expression by the retinoid and phorbol ester. We previously reported that SPR1 protein was only sparsely detectable in squamous dysplasia and carcinoma and that no SPR1 was detectable in bronchogenic adenocarcinoma, small-cell carcinoma, and large-cell carcinoma (12). In addition, SPR1 expression was not detected in other human lung-cancer cell lines, including A549 (isolated from a poorly differentiated carcinoma), H226 (from a squamous-cell carcinoma), and H596 (from an adenosquamous carcinoma) (unpublished data). Loss of SPR1 mRNA expression was also reported by DeMuth and associates (13) for 12 human lung-cancer cell lines. Taking these findings together, it is apparent that SPR1 overexpression is a marker for early metaplastic changes and that its loss signifies an irreversible malignant transformation. Thus, SPR1 is a potential intermediate biomarker for initiation and progression of bronchial malignant transformation.

It is believed that SPR1 plays a protective role for cells with squamous differentiation because it participates, similar to loricrin and involucrin, in the formation of the cornified cell envelope (32). It is understandable that SPR1 expression is upregulated in the tracheobronchial epithelium upon exposure to exogenous insults such as carcinogens. On the other hand, its decline associated with progression of carcinogenesis is more difficult to explain. It is not known whether the dysregulation of SPR1 expression is an epiphenomenon associated with bronchial carcinogenesis or if it is one of the essential steps involved in the pathway(s) of tumorigenesis.

It has been shown that loss of SPR1 expression in lung cancer is not a consequence of deletions of the coding regions of the gene (13). Results of transfection studies using chimeric constructs of the SPR1 promoter and a CAT reporter reveal an association of transcriptional dysregulation of SPR1 promoter with bronchial malignant progression. For the 96-CAT3 plasmid, low level of CAT was detected with the HBE1 cell line. Its level diminished progressively with increasing malignant potential, which was in parallel with the trend of SPR1 expression in these cell lines. These findings are in agreement with our previous study that the first 96 nucleotides of the 5'-flanking region of the SPR1 gene contains basal promoter activity (14).

For the 162-CAT3 chimeric plasmid, a relatively high level of CAT was detected in every cell line except H460. Further, the CAT level was enhanced by PMA but was not affected by vitamin A. These observations support our previous report that the SPR1 5'-flanking region between 162 and 96 nucleotides contains enhancer elements of TRE and TRE-like motifs including the binding sites for activator protein-1 (AP-1) transcription factors (14). They are also in agreement with our previous findings that downregulation of SPR1 by retinols occurs at the level of post-transcription (10). Although CAT level was significantly enhanced by PMA in HBE1, HBE1-C, and BEAS-2B, similar enhancement was not observed for H460. Thus, it appears that transcriptional control of the SPR1 promoter is disrupted in H460. It has been shown that expression of SPR1 is dependent on interaction of the transcription factor c-jun and the TRE motifs of SPR1 promoter (14). It is possible that the decline in SPR1 expression in bronchial carcinogenesis is linked to dysregulation of interaction of transcription factors such as c-jun with the TRE motifs.

In summary, we have identified a series of human bronchial epithelial cell lines representing a model of multistep tumorigenesis. Using this model, we showed that SPR1 is a marker for initiation and progression of bronchial carcinogenesis. The loss of SPR1 expression through multistep carcinogenesis is associated with transcriptional dysregulation of the first 162 nucleotides of the 5'-flanking promoter region of the SPR1 gene.