Cell-Cell communication via gap junctions plays a key role in carcinogenesis and in growth control. One of the gap junction proteins, Connexin 26 (Cx26) was considered as tumor suppressor in various cancers. In our study, the expression of Cx26 was analyzed in human lung cancer. The reduced mRNA expression was observed in 17 lung cancer cell lines examined by Northern blot analysis and RT-PCR. In 138 primary carcinomas comprising all subtypes analyzed by immunohistochemistry, 85 cases (62%) exhibited no expression of Cx26, whereas in other 53 cases the Cx26 staining was positive (38%). Additionally, an association between Cx26 protein expression and higher grading of tumors was found in whole tumor samples (p =0.028) but no statistically significant correlations could be observed with tumor stage, tumor size and node status. In squamous cell carcinoma, tumors with higher stage and grading were linked to higher expression of Cx26 (p = 0.015 and 0.017, respectively). To explore the mechanism responsible for the downregulation of Cx26, we treated 2 lung cancer cell lines H2170 and H226 with the demethylation agent 5-aza-2′-deoxycytidine and found the reexpression of Cx26 mRNA. Methylation status of these 2 cell lines was further analyzed by PCR amplification of bisulfite modified DNA and sequencing. A heterogeneous methylation pattern turned out. Our results suggest the inactivation of Cx26 in lung cancer may be explained by promoter methylation.
Lung cancer is one of the most common malignancies and is the leading cause of cancer-related death in the world. Despite the major progress made in cancer treatment during the last decades, the prognosis of lung cancer has not greatly improved due to its high frequency of recurrence. To achieve new therapeutic approaches for this fatal disease, a better understanding of the molecular mechanisms underlying the complex process of tumorigenesis in lung cancer is therefore required.
In order to identify new candidate genes in lung carcinogenesis, we performed suppression subtractive hybridization (SSH) to reveal lung cancer associated genes in previous studies.1, 2 Comparing the gene expression between normal human bronchial epithelial cells (HBEC) and a lung squamous carcinoma cell line, we cloned 2 cDNA libraries that represented mainly the genes that are overexpressed and underexpressed in HBEC and the tumor cell line, respectively. The clone HBEC-15 with high similarity to the human Connexin 26 gene (gap junction protein, beta 2) was found in the library enriched for the genes that were downregulated in tumor cell lines.
Connexins (Cxs) are member of a multigene family of at least 20 highly conserved proteins that compose a hexameric transmembrane functional channel called a connexon.3 Gap junctions are formed by the interaction of connexons or hemichannels on adjacent cells. Intercellular communication mediated by gap junctions plays an important role in a variety of cellular processes including homeostasis, morphogenesis, cell differentiation and growth control. Modulation of gap junction communication can be achieved by multiple mechanisms such as alteration in transcription, translation, stability and posttranslational processing. It has been reported that Cx is expressed in a tissue-specific manner during development and adult life, and the reduction or alteration in the level or types of connexin expressed in a given cell type is correlated with tumor progression and metastasis.4
One of the Cx family members, Cx26, has been recently considered as a potential tumor suppressor in various kind of tumors with the evidence that the Cx26 gene confers a growth suppression when transfected into these tumor cells.5, 6, 7, 8 However, the mechanism by which Cx26 acts as a potent tumor suppressor however still needs to be elucidated.
DNA methylation is known to play a key role in regulating gene expression during cell development. Many genes contain CpG islands in their promoter regions and the methylation of these regions leads to transcriptional silencing in cancer.9, 10 Recently, Tan et al.11 have reported that methylation of the promoter region of the Cx26 gene is likely to be an important mechanism in modulating the expression of Cx26 in breast cancer by using methylation-specific single-stranded conformational analysis and genomic sequencing.
Cx26 expression persists throughout life in the lung, suggesting that gap junctions serve more perennial intercellular communication functions in the peripheral lung.12 However, so far limited attention has been given to the involvement of Cx26 in lung carcinogenesis. In our study, we analyzed the expression of Cx26 in lung cancer cells as well as in primary lung tumors and also explored the mechanism that is possibly responsible for the downregulation of this gene in lung cancer.
Human bronchial epithelial primary cultures (HBEC) were obtained from Clonetics (Belgium) and cultured in BEG media (Clonetics) until population doubling of maximal 10. Human lung carcinoma cell lines containing small cell lung carcinomas (SCLCs: SHP77, CPC-N, COLO677, COLO668, DMS79, H526 and H82) and nonsmall cell lung carcinoma (NSCLCs: H2170, H125, H123, H2030, H23 and H226) were purchased from the American Type Culture Collection (ATCC, Rockville, MD) and from the German Collection of Microorganisms and Cell Culture (DSMZ, Braunschweig, Germany). In addition, 4 NSCLC cell lines were established in our lab including D51, D54, D97 and D117. They were derived from primary tumors of patients who were operated at the Charité University Hospital. These cells were grown in Leibovitz 15 media supplemented with 10% FCS and 1% glutamine.
To study the effect of the pharmacological agent 5-aza-2′-deoxycytidine (Sigma Chemical Co., St. Louis, MO), 2 lung cancer cell lines, H226 and H2170, were plated and cultured in 10 cm Petri dishes until 50% confluency was achieved. Afterwards, 5-aza-2′-deoxycytide was added into the medium and incubated in the presence of various concentrations of 0, 10, 20 and 25 μM respectively for 48 hr. Then the cells were washed with phosphated-buffered saline (PBS) and incubated further in fresh medium containing the same concentrations of 5-aza-2′-deoxycytidine. After 48 hr incubation, the treated cells were harvested for total RNA isolation and reverse transcription polymerase chain reaction (RT-PCR).
Total RNA was prepared from the HBEC and lung tumor cell lines by using Trizol (Life Technologies, Gaithersburg, MD) and then incubated with RQDNaseI (Progema, Madison, WI) for 40 min at 37°C.
First-strand cDNA was reverse-transcribed by 200 U superscript II transcriptase (Gibco, Eggenstein, Germany) from 3 μg of total RNA, in the presence of 1× RT buffer, 100 mM DTT, 20 U RNase inhibitor and 5 mM dNTPs, using the random hexamer primers supplied in the kit. The reaction took place at 42°C for 1 hr followed by 15 min at 70°C. Reactions lacking reverse transcriptase were used to verify the absence of amplification from genomic DNA contamination (data not shown). PCR was performed in a 50 μl volume. A 2 μl aliquot of the cDNA was used for PCR amplification of Cx26 with the following primers (Genbank Accession Number NM_004004): forward: 5′-GCA, TTC, GCT, TTT, TCC, AGA, GC -3′ and reverse: 5′- TGG, GAG, ATG, GGG, AAG, TAG, TG -3′; forward: 5′- CAT, CCG, GCT, ATG, GGC, CCT -3′ and reverse 5′- CAG, TGA, CAT, TCA, GCA, GGA, TG -3′ resulting in a DNA product of 254 bp and 397 bp, respectively. PCR conditions were 95°C 1 min, 58°C 30 sec and 72°C 30 sec.
Northern blot analysis
Ten micrograms of total RNA extracted from HBEC and lung tumor cell lines (H2170, D51 and H526) were separated in 1.2% agarose/2.2 M formaldehyde gel and run for 4 hr at 60 V followed by capillary transfer onto nylon membranes (Amersham, Braunschweig, Germany) overnight. The cDNA clone of HBEC-15 was amplified by nested PCR as described previously and used as probe for labeling by random priming with [32P]-dCTP (Megaprime labeling, Amersham).1 Following at least 30 min of prehybridization with Expresshyb solution (Clontech), the denatured probe was added and hybridization was carried out at 58°C overnight. Blots were washed with 2× SSC/0.1% SDS at room temperature and 0.1× SSCP/ 0.1% SDS at 60°C and then exposed to X-ray films at −80°C.
Southern blot analysis
High-molecular-weight genomic DNA was isolated from lung cell lines including HBEC, H2170, H226, D51, COLO668, H23, SHP77 and COLO677. Ten micrograms of DNA were digested with restriction enzyme Hind III (60 U per sample) overnight at 37°C, size-fractionated on a 0.7% agarose gel, run for 4 hr at 70 V and vacuum transferred to Hybond-N+ membrane for 90 min. Filters were hybridized and washed as described for Northern blot analysis.
Genomic DNA extraction and PCR amplification of bisulfite modified DNA
Genomic DNA was extracted from resuspended 10 × 106 cells by proteinase K digestion and phenol/chloroform extraction. Bisulfite treatment was modified from the method of Frommer et al.13 Briefly, 900 μg genomic DNA isolated from HBEC, H2170 and H226, respectively, was first linearized with restriction enzyme Hind III and denatured with 0.3 M NaOH for 15 min at 42°C in a final volume of 20 μl. After the addition of 20 μl of 4% low melting preheated agarose, the DNA/agarose mixture was immediately transferred into chilled mineral oil to form a single agarose bead, which was then incubated in a total volume of 400 μl with freshly prepared 5 M sodium bisulfite/1 M hydroquinone, pH 5.0, at 50°C for 12–15 hr in the dark under mineral oil, followed by equilibration with 1 ml TE, desulfurization in 500 μl of 0.2 M NaOH and neutralization in 1 M HCL. The bisulfite-treated DNA (the agarose bead) was heated at 95°C for 5 min before PCR amplification.
A pair of primer was designed to amplify both bisulfite modified methylated and unmethylated DNA but not unmodified DNA. The target fragment of 237 bp (−175 to +62) spanning 31 CpG dinucleotides was amplified by using the primer pair: 5′- ATT, CGG, GAA, GTT, TTG, AGG, A –3′ (sense) and 5′- TCT, ACG, CTA, AAA, CTC, CTA, C –3′ (antisense), which were designed from the promoter sequence of Cx26 (Ensembl Genome Browser with the accession number of ENSG 00000165474). PCR was performed in a final volume of 50 μl using 20 μl of sodium bisulfite modified DNA, 0.5 μM of each primer, 200 μM of each dNTP, 2 mM MgCl2 and 2.5 U of Hotstar Taq DNA polymerase supplied in the kit (Qiagen, Germany). The PCR reactions consisted of 38 cycles at 95°C (45 sec), 53°C (60 sec), 72°C (60 sec) with an initial denaturation at 95°C for 3 min and a final elongation at 72°C for 10 min. Ten microliters of the PCR products were electrophoresed at 100 V for 1 hr on a 1.5% agarose gel to verify amplification.
Unmodified genomic DNA from the same cell lines were applied for PCR amplification under the same conditions. Each of the PCR amplifications was repeated at least once to confirm the result.
PCR products cloning and sequencing
The amplified PCR fragments of 237 bp were excised from the agarose gel, purified and cloned in to a TA cloning vector (Invitrogen, San Diego, CA). After plasmid DNA preparation and digestion with restriction enzyme (EcoRI), clones containing inserts were confirmed and subsequently sequenced (Amersham).
Tumor samples, tissue microarray construction and immunohistochemical analysis
Tumor specimens used in this study consisted of 138 patients between ages 42 and 80 years (mean 62 years) who underwent surgical operation of lung cancer at the Department of Surgery of Charité University Hospital from 1995 to 2000. No adjuvant radiotherapy or chemotherapy was administered before surgery. The clinicopathological characteristics were summarized in Table I according to TNM criteria of the UICC.14 Survival data were present for 65 out of 138 patients.
ADC = adenocarcinomas; SCC = squamous cell carcinoma; LCLC = large cell lung carcinoma; SCLC = small cell lung carcinoma.
Total no. (n = 138)
Tissue microarrays (TMAs) were constructed with assistance of Oligene (Oligene GmbH, Berlin, FRG) using a manual tissue arrayer (Beecher Instruments, Woodland, USA). All sections were reviewed by 2 pathologists (T. Knösel and YW. Yu). Suitable areas for tissue retrieval were marked on standard haematoxylin and eosin (H&E) sections. Tissue cylinders with a diameter of 0.6 mm were punched out of the paraffin block and transferred into a recipient array block. After construction, 4 μm sections containing 276 tumors presenting 138 patients were cut from the “donor” blocks and transferred to glass slides without any sectioning aids like adhesive tapes or additionally coated slides.
Standard indirect biotin-avidin immunohistochemical analysis was used to evaluate the Cx26 protein expression. Briefly, a section from the tissue microarrays was dewaxed with xylene and gradually hydrated. Antigen retrieval was performed by treatment in a pressure cooker for 6 min. The mouse anti-Connexin 26 monoclonal antibody with working concentration of 1:250 (Zymed laboratories, Inc.) was incubated at room temperature for 1 hr. Detection took place according to the manufacturer's instructions (LSAAB-kit, DAKO). Three large sections from normal lung tissue were used as control. All slides were manually read by one pathologist (T. Knoesel) who was blinded to the clinical information. Immunohistochemistry was scored semiquantitatively as negative (score 0), weak (score 1), moderate (score 2) or strong (score 3) as previously described15, 16 and exemplified in Figure 5. Weak staining corresponded to faint signal intensities of the tumor cells that were hardly distinguishable from background or unspecific staining. For statistical evaluation, scores 0–1 were therefore considered as negative and scores 2–3 as positive.
To compare the protein expression of Cx26 with clinicopathological parameters, 2 × 2 contingency tables (e.g., Cx26 negative and Cx26 positive and G1 and G2 vs. G3) were set up, and chi-square (χ2) was applied by using the statistical software package SPSS 11.5. All presented p values were calculated 2-sided; p values < 0.05 were considered statistically significant.
Expression analysis of Cx26 mRNA in lung cancer cell lines
Northern blot analysis of cultured lung cell lines showed that Cx26 mRNA was strongly expressed in HBEC, only slightly expressed in D51, while in H2170 and H526, there was no expression detectable (Fig. 1).
We extended the expression analysis of Cx26 mRNA to 17 lung tumor cell lines including H2170, D51, H526, H125, H123, SHP77, H2030, D117, D54, CPC-N, COLO677, COLO688, H23, H82, DMS79, H226 and D97 by RT-PCR and found the markedly increased expression of Cx26 in HBEC and reduced expression in D51, CPC-N and H82, whereas in other cell lines no mRNA expression of Cx26 could be seen (Fig. 2). This result is in good agreement with that observed in Northern blot analysis.
Demethylation of H226 and H2170 by 5-aza-2′-deoxycytidine
RT-PCR showed that the lung cancer cell lines H226 and H2170 lack Cx26 mRNA expression, making them suitable candidates for the demethylation test. Therefore, these 2 cell lines were selected for the treatment by the demethylating agent 5-aza-2′-deoxycytidine.
The expression of Cx26 in H226 and H2170 was restored after treatment with 5-aza-2′-deoxycytidine (Fig. 3). In H2170, reexpression increased with rising concentration of 5-aza-2′-deoxycytidine. Compared to H2170, the reexpression of Cx26 in H226 was however relatively low and not concentration-dependent. Especially, when H226 was treated with the concentration of 25 μM 5-aza-2′-deoxycxtidine, the reexpression of Cx26 disappeared. The demethylation test suggested a role of CpG methylation in regulating Cx26 mRNA expression in lung cancer.
Southern blot analysis
No deletions or rearrangements of connexin 26 could be detected by Southern blot analysis (data not shown).
Analysis of the methylation status in lung cancer cell lines
The methylation status of the Cx26 gene in lung cancer cell lines was determined by PCR after bisulfite treatment of DNA followed by sequencing of the PCR products. We amplified a 237 bp fragment in the promoter region spanning 31 CpG sites and analyzed the methylation status in more than 20 colonies from the lung cancer cell lines H226 and H2170, respectively. The target fragment of 237 bp was only observed in bisulfite-treated genomic DNA but not in the unmodified DNA, indicating a successful modification of the DNA by the bisulfite treatment (Fig. 4). Sequencing analysis showed that out of 22 colonies from H2170, methylation occurred in 16 colonies (73%) among which a heterogeneous methylation pattern was found. The hotspots sites were concentrated on −60, −55, −36, −30 and −25, while on the sites at −172, −143, −141, −5, +33, +43 and +57, there was no methylation detectable. In H226, only 4 colonies (20%) were positive indicating a less pronounced methylation compared to H2170 (Table II). This result was in agreement with the 5-aza-2′-deoxycxtidine-demethylation test, where Cx26 reexpression of H226 was less prominent than for H2170. For HBEC, we did not find any methylation within the CpG dinucleotides of the Cx26 gene from 10 randomly picked colonies (data not shown).
Table II. Methylation Patterns of Cx26 from Aberrantly Methylated Lung Tumor Cell Lines. Sequencing of Colonies from H2170 and H226 Revealed a Heterogeneous Methylation status. Open Circles Represent the Unmethylated CpG Dinucleotides; Methylated CpG Dinucleotides are Indicated by Closed Circles
Cx26 protein expression in primary lung tumor
A tissue microarray was constructed to explore the Cx26 protein expression in 276 primary lung cancer samples representing 138 patients. One section is shown in Figure 5a. Normal human bronchial epithelia cells (HBEC) were staining positive (Fig. 5b). As shown in Table I, 85 cases (62%) exhibited no expression of Cx26 (Fig. 5c), while in other 53 cases, the Cx26 staining was positive (38%). The pattern of positive staining for Cx26 in lung cancer was intracytoplasmic (Fig. 5d–f). The expression of Cx26 did not differ significantly by age, sex, tumor stage, tumor size and node status; however, an association between Cx26 protein expression and higher grading of tumors was found (p = 0.028, Table III). Especially in squamous cell carcinoma (SCC), higher stage and higher grading tumors were linked to increased expression of Cx26 (p = 0.015 and 0.017, respectively), while in adenocarcinoma (ADC), the Cx26 expression was not significantly related to any of the clinicopathological parameters that we investigated. Interestingly, the difference of Cx26 expression between SCC and ADC reached statistical significance with SCC showing more frequently positive expression (27% in SCC and 10% in ADC, Table I).
Table III. Correlation Between Connexin 26 Expression and Clinicopathological Data (p Values)1
p values were calculated by chi-square test. ADC = adenocarcinoma; SCC = squamous cell carcinoma.
On analyzing 65 patients with primary lung cancer and a median 24 months survival (range 2–66 months), we found no significant effect of Cx26 protein expression on overall survival (p= 0.32; Kaplan-Meier statistical analysis).
Gap junctional intercellular communication (GJIC) is the only means by which multicellular organisms exchange low molecular weight signals directly from within 1 cell to the interior of neighboring cells.17 Cell contact-mediated communication has been found to have an important role in carcinogenesis with decreased or absent communication being frequently observed in various tumor cells, and the gap junction proteins, the connexins (Cx), involved in this type of communication appear to form a family of tumor suppressor.18
One of the Cx family members, the Cx26 gene, was initially identified by using subtractive hybridization technique with the mRNA expression only in normal but not in malignant mammary epithelial cell lines.19 Later, diminished expression of Cx26 has also been found in other kind of cancer cell lines and in primary tumors.7, 8, 20 Transfection of the Cx26 gene into gap junction-deficient human breast cancer cell lines, hepatocellular carcinoma cell lines and choriocarcinoma cells reduced their growth rate, tumorigenicity, and resorted their potential to differentiate.5, 6, 7, 8 Growing evidence supports the function of Cx26 as a tumor suppressor.
Connexin 26 was found to be expressed in human and ferret airways during development and the expression persists throughout the adult life.12 Besides, the Cx26-null mice exhibited an abrogated lobulo-alveolar development and increased cell death during pregnancy, suggesting an essential role in the lung.21 However, to our knowledge, except for one report about the reduced expression of Cx26 protein in lung cancer, suggesting its potential tumor suppressor function, no other literature provide information about the alteration of Cx26 expression in this kind of tumor.22 Furthermore, the potential importance of the Cx26 gene in lung carcinogenesis was largely enhanced, as this gene was isolated in our SSH library when we compared human bronchial epithelial cells (HBEC) with the lung cancer cell line H2170. Thus, we turned our attention to investigate the potential role of Cx26 in lung cancer.
We observed that Cx26 mRNA was not only diminished in the cell line used for the SSH library construction but also in other 17 lung cancer cell lines representing 4 major types of lung cancer. Alterations of Cx26 expression at the transcriptional level might contribute to the migratory ability of lung cancer cells, a necessity for invasive tumor growth, since the loss of communication between tumor cells and the surrounding normal cells is a critical factor for malignancy.23 Intriguingly, we found that the Cx26 gene was among the subset of 918 clones that we identified to be most valuable for the clustering/classification of primary lung carcinomas using cDNA microarrays revealing its important role in lung cancer.2, 24
Using a monoclonal antibody directed against Cx26, we observed that the majority of the primary lung tumor samples (62%) showed no protein expression by immunohistochemistry, supporting the fact that alterations of Cx26 expression exist not only in cultured lung cancer cells but also in primary tumors. Since the sample sizes of small cell and large cell lung carcinomas were too small to be further evaluated, we only divided the primary tumors into the 2 major subgroups of adenocarcinoma and squamous cell carcinoma for statistical analysis. Interestingly in SCC, the higher expression of Cx26 was related to higher grade and pathologic stage. This pattern is similar to that of the widely known tumor suppressor gene p53, whose immunopositivity increases with tumor stage and grade in various kinds of cancer including lung carcinoma, suggesting a role during the early stage of lung carcinogenesis.25, 26, 27
We investigated the mechanism that might lead to the inactivation of cx26 in lung cancer. No deletion or gene rearrangement was found. Indeed, deletion of the connexin family is a rare event in human cancer. So far, deletion of the Cx26 gene was only found in the patients with hereditary deafness.28 On the contrary, recent studies showed that epigenetic inactivation of connexins through hypermethylation of the promoter region could lead to gene silencing expression. Aberrant methylation patterns have been identified in the promoter regions of the Cx43 and the Cx32 genes.29, 30 For the Cx26 gene, variable promoter region CpG island methylation has been reported in breast cancer,11 whereas in esophageal cancer cells, methylation of Cx26 seems not to be involved as a primary mechanism responsible for the downregulation of this gene.31 In view of increasing literature that implicates the inactivation of tumor-suppressor genes due to methylation in lung cancer,9, 10, 32 we analyzed the methylation status of this tumor suppressor candidate. First, we treated the 2 lung tumor cells of H2170 and H226 lacking the mRNA expression of Cx26 with demethylating agent 5-aza-2′-deoxycytidine. The restoration of the mRNA expression in both cell lines is a powerful support of the importance of Cx26 methylation in inactivating this gene. Furthermore, PCR was performed to amplify the bisulfite modified DNA isolated from these 2 cell lines together with that from HBEC. It turned out that Cx26 promoter methylation was in 16 out of 22 colonies (73%), and 4 out of 20 colonies (20%) from the examined cell lines of H2170 and H226, respectively, further supporting the evidence that hypermethylation of Cx26 might be a fundamental mechanism for silencing this gene in lung cancer.
Sequencing of bisulfite-treated tumor DNA indicated a heterogeneous methylation pattern in Cx26 with some colonies exhibiting hypermethylation in the same CpG sites, whereas others exhibiting hypermethylation in various sites within the promoter region. It is noteworthy that heterogeneous methylation patterns have also been observed in other genes, for example, the RB promoter in retinoblastomas and the p15 promoter in acute myeloid leukaemia.33, 34 One possible explanation is that cells sampled at various stages of tumor progression, in different cycles or cells undergoing clonal evolution could give rise to heterogeneous methylation patterns.33 Additionally, methylation only occurred in 20% colonies from H226, which was in line with the demethylation test by using 5-aza-2′-deoxycytidine. However, it occurred much less frequently than from H2170 (73%), suggesting that besides hypermethylation in the promoter region leading to inactivation of Cx26, other mechanisms might be involved, e.g., histone deacetylation or absence of the recruitment of a regulatory gene to the upstream region.35, 36
In summary, the downregulation of Cx26 in lung cancer cell lines as well as in primary lung tumors as shown in this study is consistent with the potential role of Cx26 as a tumor suppressor gene in lung cancer. Methylation of Cx26 in the promoter region is obviously one mechanism for transcriptional silencing of the gene. The elucidation of the mechanism that mediates the loss of Cx26 has an important clinical implication. Since demethylation agents are under clinical evaluation, our finding might provide an advantage in therapeutic application in lung cancer. Further studies will be needed to elucidate the functional consequences of Cx26 deregulation in lung cancer.
We thank Dr. L. Tan for the idea about PCR amplification of bisulfite modified DNA.