Homeobox Transcription Factor NKX2-1 Promotes Cyclin D1 Transcription in Lung Adenocarcinomas Free

Lung cancer is the leading cause of cancer-related death worldwide. The 5-year survival rate of non–small cell lung cancer (NSCLC) patients remains at 15% despite chemotherapy, radiotherapy, and surgery (1). The discovery of efficient molecular targets for NSCLC treatment and reliable prognostic markers are highly desired. Overexpression of cyclin D1 is frequently found in human breast cancers and is associated with the early steps of breast tumorigenesis (2). Overexpression of cyclin D1 has also been found in NSCLC. The association between cyclin D1 expression and prognosis is controversial (3, 4). It has been reported that overexpression of cyclin D1 is not only positively (5–9), but also negatively (10–17) correlated with poor prognoses. In addition, some reports have shown no correlation between them (18–19). The reason for the positive correlation between overexpression of cyclin D1 and a good prognosis of NSCLC has been unclear.

Cyclin D1, a G1 cyclin, binds to and activates cyclin-dependent kinase 4/6 (CDK4/6) at G₁ phase (20, 21). Cell-cycle progression from G₁ to S-phase is negatively regulated by retinoblastoma protein (pRB) as a gatekeeper of the restriction point (22). In the mid-late G₁ phase, the cyclin D1–CDK4/6 complex phosphorylates several sites, such as Ser780 in pRB, and inactivates pRB (23). Then, pRB-binding transcription factors, such as E2F, dissociate from pRB. Activated E2F promotes transcription of growth-associated target genes and contributes to G₁–S progression (22). Overexpression of cyclin D1 is often observed in human cancers, contributes to unregulated cell-cycle progression by skipping the restriction point, and participates in cancer cell growth (21). There are many studies about the mechanisms of transcriptional regulation of cyclin D1 (24, 25). Jumonji and Tob1 participate in negative regulation of cyclin D1 transcription. Several growth factors, such as EGF and IGF, estrogen, and angiotensin II promote cyclin D1 expression via specific signal transduction pathways. Transcriptional induction of the cyclin D1 gene is mediated by transcription factors such as TCF/LEF, CREB, NF-κB, AP-1, c-myc, and SP1 (26). Deregulated activation or expression of these transcription factors can contribute to enhanced expression of cyclin D1 during tumorigenesis. In colon cancers, β-catenin stabilization by abnormalities in Wnt signaling, such as deletion of APC, enhances cyclin D1 transcription cooperating with TCF/LEF. In lung cancers, we recently reported that YB-1 binds to and activates the cyclin D1 promoter to promote expression of cyclin D1 (26). Moreover, overexpression of cyclin D1 is associated with YB-1 expression in human NSCLC clinical specimens.

NKX2-1, also known as thyroid transcription factor 1 (TTF-1), is a homeobox transcription factor involved in lung and thyroid cell development as a differentiation-promoting factor (27, 28). During lung development, NKX2-1 positively regulates transcription of genes involved in pulmonary development and functions, such as surfactant protein (SP)-A, SP-B, and SP-C, and secretoglobin. NKX2-1 recognizes the CTTG/GAAC sequence motif in regulatory regions of its target genes (29, 30). It also promotes transcription of cell-cycle–related genes and prosurvival genes such as ROR1 (31) and LMO3 (32). In cancers, expression of NKX2-1 correlates with good prognoses of NSCLC (33–35). NKX2-1 negatively regulates metastasis via suppression of Snail and Slug transcription (36). Moreover, it promotes expression of tight junction proteins, including claudins and occludin, which promote cell–cell interactions (37). NKX2-1 induces MYBPH that inhibits Rock1 to reduce cell motility (38). Therefore, NKX2-1 is involved in pulmonary development as a differentiation-promoting factor in normal lung cells. In contrast, NKX2-1 negatively regulates metastasis-associated genes as a metastasis suppressor in cancer cells. It has been reported that epigenetic gene suppression and allelic loss of NKX2-1 gene are involved in the defects of NKX2-1 expression in human lung cancers (39, 40). Depletion of NKX2-1 is suggested to promote the metastatic potential of cancer cells (41, 42).

Human lung cancer cell lines H441, PC3, LC2/Ad, A549, H1299, ABC1, and EBC1, and human embryonic kidney cell line HEK293 were grown in DMEM supplemented with 10% FBS at 37°C in an atmosphere containing 5% CO₂.

The antibodies used in this study were as follows: anti-NKX2-1 (Santa Cruz Biotechnology, #sc-13040), anti-cyclin D1 (Santa Cruz Biotechnology, #sc-20044), anti-cyclin B1 (Santa Cruz Biotechnology, #53236), anti-cyclin E (Santa Cruz Biotechnology, # sc-247), anti-SP-A (DAKO, # 5B001A), and anti-α-tubulin DM1A (Sigma, #sc-32293).

A reporter construct containing -1745 bp of the human cyclin D1 promoter linked to a luciferase reporter gene (-1745-CD1-Luc) was kindly provided by Dr. Suzuki (Department of Oncology, Institute of Medical Science, The University of Tokyo, Tokyo, Japan). The Myc-DDK-human NKX2-1 expression plasmid was purchased from Origene (#RC202477).

Cells were transfected with human NKX2-1 siRNA or control siRNA oligonucleotides using Lipofectamine RNAiMAX (Invitrogen, #13778-150), according to the manufacturer's protocol. The nucleotide sequences shown in Supplementary Table S1.

Cells were lysed in lysis buffer (0.3% Triton X-100, 300 mmol/L NaCl, and 50 mmol/L Tris-HCl, pH 7.5). Cell lysates were denatured by treatment with SDS sample buffer at 95°C for 8 minutes. The lysates were subjected to SDS-PAGE and then transferred onto a polyvinylidene difluoride membrane (Millipore), followed by immunoblotting with the indicated antibodies. Proteins were visualized using an enhanced chemiluminescence system (Perkin Elmer).

Total RNA was isolated using an RNeasy Mini Kit (Qiagen, #74104) for cultured cells and an Isogen Kit (Wako, #311-02501) for human clinical frozen samples according to the manufacturers' instructions. Reverse transcription was performed with random hexanucleotide primers and reverse transcriptase SuperScript II (Invitrogen, #18064014). The resulting cDNA was subjected to real-time PCR using the Rotor-Gene 3000 System (Corbett Research) and a QuantiTect SYBR Green PCR Kit (Qiagen, #204143) or SYBR Green Realtime PCR Master Mix (Toyobo Co., #QPK-201). Primer sequences of NKX2-1 and cyclin D1 were shown in Supplementary Table S1. Transcript levels were normalized to 18S rRNA mRNA.

To analyze NKX2-1 and cyclin D1 mRNA expression in clinical specimens of lung adenocarcinomas, quantitative reverse transcription-PCR (qRT-PCR) was performed using 72 adenocarcinomas in Lung Cancer cDNA Array I, II, III, and V (OriGene). Transcript levels were normalized to β-actin mRNA. The correlation between NKX2-1 and cyclin D1 expression was evaluated by linear regression analysis.

Cells (1 × 10⁵ per well) cultured in 6-well plates were transfected with 0.5 μg cyclin D1-luciferase reporter plasmid -1745CD-Luc, 100 ng CMV-β-gal plasmid, and 1 μg human NKX2-1 expression vector or 1 μg empty vector using Fugene6 reagent (Promega, #E2692) in Opti-MEM (Invitrogen). Cells were lysed at 48 hours after transfection and assayed for luciferase and β-galactosidase activities with the former being normalized to the latter.

H441 cells (4.4 × 10⁶) were treated with 1% formaldehyde. Crosslinking was terminated by addition of 0.125 mol/L glycine. The cells were lysed with cell lysis buffer (50 mmol/L HEPES, pH 7.5, 0.5% NP-40, 140 mmol/L NaCl, 1 mmol/L EDTA, 10% glycerol, 0.25% Triton X-100, and a protease inhibitor cocktail) on ice. After centrifugation, the cell pellets were lysed by nuclear lysis buffer (10 mmol/L Tris, pH 8.0, 200 mmol/L NaCl, 1 mmol/L EDTA, 0.5 mmol/L EGTA, and a protease inhibitor cocktail) and then sonicated with buffer (1% SDS, 10 mmol/L EDTA, 50 mmol/L Tris, pH 8.0, and a protease inhibitor cocktail). After centrifugation, the lysates were diluted with an equal volume of dilution buffer (1.1% Triton X-100, 50 mmol/L Tris, pH 8.0, 167 mmol/L NaCl, 0.11% sodium deoxycholate, and a protease inhibitor cocktail). Immunoprecipitation was performed with an anti-NKX2-1 antibody (Santa Cruz Biotechnology, #sc-13040 X) and normal rabbit IgG as a control. After immunoprecipitation, 20 μL of Dynabeads protein G (Thermo Fisher, #10004D) was added, followed by 1 hour of incubation. The elutes were incubated at 65°C for de-crosslinking. DNA fragments were purified using UltraPure Phenol*Chloroform*Isoamyl Alcohol (Thermo Fisher, #15593031). PCR was performed using Platinum Taq polymerase (Invitrogen). Primer pairs are shown in Supplementary Table S1. The PCR program was initial denaturation at 94°C for 2 minutes, and then 34 cycles of 94°C for 30 seconds, 58°C for 30 seconds, and then 72°C for 30 seconds. PCR products were analyzed on an agarose gel by electrophoresis.

Lung cancer specimens were obtained from 307 Japanese patients with primary NSCLC, whose tumors had been completely removed by surgery at the Department of Surgery of Hamamatsu University School of Medicine between 1988 and 2007. Among the 307 patients, 197 patients were diagnosed histologically as having adenocarcinoma and 110 patients were diagnosed as having squamous cell carcinoma. NSCLC patient ages ranged from 33 to 86 years (median, 66 years). Of the total patient numbers, 208 were men and 99 were women (Supplementary Table S2). Although there were no significant differences in each parameter in the YB-1–positive population, in the case of NKX2-1, EGFR protein expression in female population was increased. These results might reflect the fact that NKX2-1 is a lineage survival oncogene of lung adenocarcinoma. Furthermore, NKX2-1–positive patients tended to be in an earlier T or N stage, and had a smaller tumor volume at surgery.

This study was performed in accordance with the guidelines of the Declaration of Helsinki. The study protocol was approved by the Research Ethics Committee of Hamamatsu University School of Medicine (Approved No. 23-91). A written letter of consent was processed after obtaining informed consent from patients to participate in this study.

For IHC analysis, tissue microarray blocks were prepared, deparaffinized, rehydrated, and then boiled to retrieve antigens for 30 minutes in Tris-EDTA buffer (pH 9.0) for SP-A detection or 10 mmol/L sodium citrate buffer (pH 6.0) for cyclin D1 and NKX2-1 detection. Endogenous peroxidase activity was blocked by incubation in a hydrogen peroxide solution for 30 minutes. The sections were then incubated with a rabbit anti-NKX2-1 polyclonal antibody (Abcam, 1:7,500), rabbit anti-cyclin D1 mAb (Nichirei, Tokyo, #413521, 1:25), and a rabbit anti-SP-A polyclonal antibody (DAKO, #5B001 PE-10). The antigen–antibody complex was visualized using Histofine Simple Stain Max-Po (Multi; Nichirei, #424131) and 3,3′-diaminobenzidine tetrahydrochloride. Counterstaining was performed using hematoxylin. To evaluate NKX2-1, cyclin D1, and SP-A expression, the number of stained cells was counted, and at least five high power fields were chosen randomly for scoring the percentage of cells with positive staining among 1,000 cells examined per section.

NKX2-1, cyclin D1, and SP-A expression was scored as follows: 0, no staining or expression in <10% of cancer cells; 1+, expression in >10% and <50% of cancer cells; 2+, expression in >50% of cancer cells. Scores 1+ and 2+ were regarded as positive, and score 0 was regarded as negative (Supplementary Fig. S1).

Data are presented as means ± SD. Data were analyzed by the Student t test or Fisher exact test, where P < 0.05 was considered to be statistically significant. Overall survival was calculated as the time from surgery to death or last contact. Standard methods for time-to-event data were analyzed by the Kaplan–Meier method and log-rank test. Univariate and multivariate Cox proportional hazards models were fitted to calculate the HRs of death with adjustment for other potential confounding factors. All data analyses were performed using statistical software package JMP (version 10, SAS Institute Inc.).

NKX2-1 is an important transcription factor for lung cell differentiation and lung carcinogenesis, which is often overexpressed in NSCLC. NKX2-1 recognizes the CTTG/GAAC sequence motif in its target genes. We searched for the motif and found 14 CTTG/GAAC motifs in the upstream regulatory region of the human cyclin D1 gene (Fig. 1A). Therefore, we speculated that cyclin D1 might be a novel target gene of NKX2-1. To evaluate the contribution of endogenous NKX2-1 to cyclin D1 protein expression in human NSCLC, we analyzed NKX2-1 expression in various human lung cancer cell lines: A549 (alveolar adenocarcinoma), H1299 (NSCLC), ABC-1 (adenocarcinoma), EBC1 (squamous cell carcinoma), LC2/Ad (adenocarcinoma), PC3 (adenocarcinoma), and H441 (papillary adenocarcinoma). NKX2-1 was highly expressed in H441 and PC3 cells, and weakly expressed in LC2/Ad cells, but it was not detected in A549, H1299, ABC-1, or EBC1 cells (Fig. 1B). This result suggests that NKX2-1 is not essential for cyclin D1 expression.

Next, to evaluate the contribution of NKX2-1 to cyclin D1 expression, we investigated whether expression of cyclin D1 was affected by NKX2-1 depletion in NKX2-1-expressing lung cancer cell lines. Depletion of NKX2-1 by both si-NKX2-1-#1 and -#2 siRNAs strongly suppressed the cyclin D1 protein in H441 cells (Fig. 1C). Expression of cyclin B, which has been reported as one of the target genes of NKX2-1, was also decreased by depletion of NKX2-1. Moreover, in another NKX2-1–expressing NSCLC cell line, PC-3, we confirmed that depletion of NKX2-1 inhibited expression of cyclin D1 protein (Fig. 1D). Cyclin D1 mRNA was also decreased in both H441 and PC-3 cells treated with si-NKX2-1-#1 and -#2 (Fig. 1E and F). We next investigated whether overexpression of NKX2-1 promoted cyclin D1 expression in NKX2-1–negative cell lines such as A549. The Myc-DDK-NKX2-1 plasmid or control vector was transfected into A549 cells. As shown in Fig. 1G and H, cyclin D1 protein expression was significantly increased by forced expression of NKX2-1. Moreover, cyclin D1 mRNA expression was increased by NKX2-1 (Fig. 1I). Therefore, supplementation of NKX2-1 in NKX2-1–negative cells promotes cyclin D1 expression. These results suggest that NKX2-1 is a positive transcriptional regulator of the cyclin D1 gene in NSCLC.

As described above, there are multiple NKX2-1–recognized CTTG/GAAC sequence motifs in the upstream regulatory region of the human cyclin D1 gene (Fig. 2A). We used -1745CD-Luc, a luciferase reporter plasmid containing the upstream region from −1745 to +1 in the human cyclin D1 gene (Fig. 2A), to investigate whether NKX2-1 promotes transcription of cyclin D1. -1745CD-Luc was transfected with or without the NKX2-1 overexpression plasmid into HEK293 cells. As shown in Fig. 2B, promoter activity of the cyclin D1 gene was significantly enhanced by ectopic expression of NKX2-1. Furthermore, we investigated whether endogenous NKX2-1 contributes to cyclin D1 transcription. siRNA against NKX2-1 or control siRNA were transfected with -1745CD-Luc into the human lung cancer cell line PC3 that expresses endogenous NKX2-1, and then luciferase activities were measured. Cyclin D1 promoter activity was significantly suppressed by depletion of NKX2-1 (Fig. 2C).

To investigate whether NKX2-1 binds to the human cyclin D1 promoter, we performed chromatin immunoprecipitation (ChIP) assays using the anti-NKX2-1 antibody. As shown in Fig. 2A, there were 23 CTTG motifs and 19 CAAG motifs in the upstream region from −6000 to +1 in the human cyclin D1 gene. We designed seven primer sets, “a” (−2341 to −2048 bp), “b” (−1941 to −1583 bp), “c” (−1539 to −1273 bp), “d” (−758 to −470 bp), “e” (−542 to −279 bp), and “f” (−1295 to −891 bp), and conducted ChIP assays. As shown in Fig. 2D and E, NKX2-1 widely bound to the human cyclin D1 promoter containing the CTTG/GAAC sequence motif. These results suggest that NKX2-1 binds to the cyclin D1 promoter to directly activate transcription of cyclin D1.

As described above, YB-1 also promotes cyclin D1 transcription (26). We examined whether YB-1 promotes NKX2-1–dependent transcription of cyclin D1 using the reporter assay. As shown in Fig. 2F, additional expression of YB-1 enhanced NKX2-1–mediated promoter activity of cyclin D1. Therefore, YB-1 and NKX2-1 collaboratively promote transcription of cyclin D1.

To investigate whether NKX2-1 is involved in cyclin D1 expression of human lung cancers, we performed IHC analysis using tissue microarrays of human NSCLC clinical specimens. There was a tendency for NKX2-1–positive adenocarcinomas to coexpress NKX2-1 with cyclin D1 (Supplementary Fig. S1, #1, #2, and #3), and NKX2-1–negative adenocarcinomas were negative for cyclin D1 (Supplementary Fig. S1, #4, #5, and #6). As shown in Table 1, expression of NKX2-1 was detected in 164 (53%) of 307 NSCLC patients. Expression of cyclin D1 was detected in 255 (83%) of 307 NSCLC patients. A total of 148 (48%) of 307 NSCLC patients showed expression of both NKX2-1 and cyclin D1. There was a significant correlation between expression of NKX2-1 and cyclin D1 in total NSCLCs (P = 0.0003). The significant correlation between NKX2-1 and cyclin D1 expression was observed in adenocarcinomas (P = 0.0005), but not in squamous cell carcinomas (P = 0.7147). Interestingly, expression of SP-A, a known target gene of NKX2-1, showed a positive correlation with NKX2-1 in both adenocarcinomas (P = 0.0001) and squamous cell carcinomas (P = 0.0001). In addition, we analyzed whether the protein expression level of NKX2-1 was correlated with cyclin D1 protein expression in the lung adenocarcinomas specimens. NKX2-1 and cyclin D1 levels were scored and then the cyclin D1 level was compared in each score group of NKX2-1. As shown in Fig. 3A, the protein expression level of NKX2-1 was correlated with cyclin D1 protein expression in adenocarcinomas.

Moreover, to determine whether mRNA expression of cyclin D1 correlated with NKX2-1 mRNA expression, 72 adenocarcinomas in Lung Cancer cDNA Array I, II, III, and V were subjected to qRT-PCR analysis. Linear regression analysis indicated that mRNA expression of cyclin D1 was moderately correlated with NKX2-1 mRNA expression (r = 0.420, P = 0.00023, y = 0.027x + 0.5610; Fig. 3B). These results suggest that NKX2-1 is a transcription factor involved in cyclin D1 expression of human lung adenocarcinomas.

To clarify the clinical relevance of NKX2-1 and cyclin D1 expression, we analyzed the relationship between their expression and prognosis in the NSCLC patients. As shown in Fig. 4A, expression of cyclin D1 was significantly correlated with a good prognosis in patients with adenocarcinomas (P = 0.0047). Expression of NKX2-1 was also correlated with a good prognosis in patients with adenocarcinomas (P = 0.0004; Fig. 4C). Interestingly, cyclin D1–positive [cyclin D1(+)]/NKX2-1–negative [NKX2-1(−)] patients showed a poorer prognosis than cyclin D1(+)/NKX2-1–positive [NKX2-1(+)] patients with adenocarcinomas (P = 0.0157; Fig. 4E). Conversely, neither NKX2-1 nor cyclin D1 expression was correlated with the prognosis of patients with squamous cell carcinomas (Fig. 4B, D, and F). To determine the reason, we investigated whether metastatic states were correlated with the expression of NKX2-1 or cyclin D1. Metastatic states were inversely correlated with the expression of NKX2-1, but not cyclin D1 (Table 2). Furthermore, our multivariate analysis revealed that lymph node metastasis was an independent poor prognostic factor for adenocarcinoma (HR, 1.735; 95% CI, 1.010–2.962, P = 0.046; Table 3). Multivariate analysis also revealed that NKX2-1 protein expression was an independent good prognostic factor (HR, 0.486; 95% CI, 0.264–0.901, P = 0.022) for adenocarcinoma but not squamous cell carcinoma (HR, 1.449; 95% CI, 0.542–3.438, P = 0.439). However, for cyclin D1, the positive correlation with prognosis in univariate analysis was not found in the multivariate analysis. There were no significant differences in both adenocarcinoma (HR, 0.537; 95% CI, 0.279–1.078, P = 0.079) and squamous cell carcinoma (HR, 1.410; 95% CI, 0.642–3.501, P = 0.408; Table 3; Supplementary Table S3). These results suggest that the prognosis of cyclin D1(+) patients depends on NKX2-1 expression, and NKX2-1, but not cyclin D1, contributes to a good prognosis of patients with adenocarcinomas.

Because it has been reported that SP-A negatively correlates with tumor progression (43), we evaluated this possibility. In Kaplan–Meier analysis, there was no significant survival advantage between cyclinD1(+)/NKX2-1(+)/SP-A(+) and cyclinD1(+)/NKX2-1(+)/SP-A(−), whereas NXK2-1/cyclin D1(+) adenocarcinomas showed a tendency for SP-A positivity (Table 1 and data not shown). This is consistent with the results of univariate and multivariate Cox proportional hazards model analyses of overall survival of patients with adenocarcinoma (Table 3). Therefore, other NKX2-1 target genes may participate in a good prognosis of NKX2-1–positive lung adenocarcinomas.

This is the first report to show that NKX2-1/TTF-1 contributes to the transcription of cyclin D1 because the functional association has not been elucidated between NKX2-1 and cyclin D1. In this study, we found many NKX2-1–binding motifs in the human cyclin D1 promoter, and demonstrated that NKX2-1 bound to them and directly activated cyclin D1 expression in NSCLC cell lines. Moreover, depletion of NKX2-1 decreased the expression of cyclin D1 at both protein and mRNA levels. Supplementation of NKX2-1 in NKX2-1–negative cells promoted cyclin D1 expression. Furthermore, cyclin D1 expression was positively correlated (P = 0.0005) with NKX2-1 in specimens of lung adenocarcinomas. These findings suggest that NKX2-1 is one of the positive transcriptional regulators of the cyclin D1 gene in NSCLC.

We previously reported that YB-1 is a transcription factor contributing to cyclin D1 expression. YB-1 was widely expressed in lung cancer cell lines (Fig. 1B) and expression of cyclin D1 correlates significantly with YB-1 in human NSCLC clinical specimens (26). Therefore, enhanced expression of cyclin D1 may be due to overexpression of YB-1 in NSCLC. In the sample set, YB-1 was overexpressed in 91.8% of squamous cell carcinomas and 92.4% of adenocarcinomas (Supplementary Tables S1 and S2). Because overexpression of both YB-1 and cyclin D1 was found in 72.7% of squamous cell carcinomas and 80.7% of adenocarcinomas, it strongly suggested that YB-1 widely promotes cyclin D1 expression in both adenocarcinomas and squamous cell carcinomas. In this study, most NSCLC tissue microarray samples were the same as those in our previous report (26). In this study, NKX2-1 was frequently overexpressed in adenocarcinomas (71%) but not in squamous cell carcinomas (22%), whereas cyclin D1 was highly overexpressed in both adenocarcinomas (87%) and squamous cell carcinomas (76%). Overexpression of both NKX2-1 and cyclin D1 was found in 65.5% of adenocarcinomas but in only 17.3% of squamous cell carcinomas. Cyclin D1 expression was significantly correlated with NKX2-1 expression in adenocarcinomas (P = 0.0005) but not squamous cell carcinomas (P = 0.7147), whereas SP-A, one of the known targets of NKX2-1, was significantly associated with NKX2-1 expression in both adenocarcinomas (P = 0.0001) and squamous cell carcinomas (P = 0.0001). These results suggest that NKX2-1 participates in cyclin D1 expression of adenocarcinomas but not squamous cell carcinomas. In squamous cell carcinomas, other transcription factors including YB-1 rather than NKX2-1 may contribute to cyclin D1 expression. In NSCLCs, overexpression of YB-1 or NKX2-1 was observed frequently and significantly correlated with cyclin D1 expression. Although we investigated the relationship between expression of YB-1 and it of NKX2-1 in NSCLC, there was no significant correlation between them (Supplementary Table S4). As shown in Fig. 1B, YB-1 was widely expressed in lung cancer cell lines, whereas NKX2-1 was restrictedly expressed in PC3 and H441 cells. In addition, we found that YB-1 and NKX2-1 collaboratively activated the cyclin D1 promoter (Fig. 2F). Therefore, we speculate that YB-1 and NKX2-1 collaboratively contribute to cyclin D1 expression in NSCLC. Frequent defects of NKX2-1 expression are found in human lung cancer cell lines (Fig. 1B). On the basis of a search result using the GEO database in NCBI, deficiencies of NKX2-1 are found in 16 of 29 lung cancer cell lines (55%). It has been reported that epigenetic gene suppression and allelic loss of the NKX2-1 gene are involved in the defects of NKX2-1 expression in human lung cancers (39, 40). These defects of NKX2-1 expression may contribute to the malignancy of these cancer cells during late onset of tumor progression processes. In NKX2-1–negative malignant cancer cells, other transcription factors such as TCF/LEF, CREB, NF-κB, SP-1, AP-1, and YB-1 may participate in cyclin D1 expression in a context-dependent manner to promote cancer cell proliferation on behalf of NKX2-1. We found that forced expression of NKX2-1 upregulated cyclin D1 expression in A549 cells, an NKX2-1–negative cell line (Fig. 1G–I). Therefore, NKX2-1 has the potential to be a positive transcriptional regulator of cyclin D1. Alternatively, depletion of NKX2-1 inhibited cyclin D1 expression by about 70% in NKX2-1–positive cell lines such as H441 and PC3 (Fig. 1E and F). These results suggest that NKX2-1 contributes to cyclin D1 transcription in NKX2-1–positive cells.

Because the proportion of deficiencies of NKX2-1 expression in human clinical lung adenocarcinomas was 34%, which was not as high as that in lung cancer cell lines (55%–71%), cyclin D1 protein expression was significantly correlated with NKX2-1 protein expression as expected (P = 0.0005; Table 1). We found that the protein expression level of NKX2-1 was correlated with cyclin D1 protein expression in the adenocarcinoma specimens (Fig. 3A). Moreover, mRNA expression of cyclinD1 and NKX2-1 was correlated (Fig. 3B). Interestingly, expression of NKX2-1 was not correlated with cyclin D1 expression in squamous cell carcinomas. Some cell-type–specific factors or pathways may participate in the differential function of NKX2-1. However, future studies are required for clarification. Collectively, our results strongly suggest that NKX2-1 participates in cyclin D1 expression in human lung adenocarcinomas as far as they express NKX2-1, whereas other transcription factors such as YB-1 also contribute to cyclin D1 expression in a context-dependent manner.

Several reports indicate that NKX2-1 is expressed in a cell type–specific manner during the development of normal lung tissues and participates in expression of proliferation-promoting genes as well as differentiation-associated genes (3, 4). In the early step of carcinogenesis, NKX2-1 may participate in expression of proliferation-promoting genes including cyclin D1 for continuous cell growth. Conversely, NKX2-1 negatively regulates metastasis via suppression of Snail and Slug transcription (36). Moreover, it inhibits cell motility to promote expression of tight junction proteins (37) and Rock1 inhibitory protein MYBPH (38). Therefore, NKX2-1 functions as a “double edged sword,” which not only has a cell proliferation–promoting activity, but also a metastasis-suppressive potential (41, 42). In terms of cyclin D1, it has been controversial whether cyclin D1 expression correlates with the prognosis of NSCLC, but there are some studies about such a correlation (3, 4). In this study, we found that NKX2-1–positive adenocarcinomas significantly showed low lymph node metastasis (Table 2) and a better prognosis (Fig. 4C; Tables 3 and 4) regardless of cyclin D1 expression (Fig. 4E) compared with NKX2-1–negative adenocarcinomas. On the basis of our study, we conclude that cyclin D1 expression should not be a reference for the prognosis of lung adenocarcinomas because of its contribution as a good prognosis factor. Furthermore, we speculate about the functions of NKX2-1–mediated cyclin D1 expression as follows. NKX2-1 may contribute to continuous cell proliferation via cyclin D1 expression during the early carcinogenesis step of pulmonary cells. NKX2-1 may also participate in not only cancerous cell proliferation via cyclin D1 expression, but also maintenance of epithelial features to prevent mesenchymal transition via suppression of EMT-promoting transcription factors, such as Snail and Slug in lung adenocarcinoma. Future studies are needed to clarify this possibility. If NKX2-1 expression is epigenetically suppressed or allelic loss of NKX2-1 gene occurs during malignant development of adenocarcinomas, the cancer cells may show a metastatic character and the prognosis of the patient will be poor.

No potential conflicts of interest were disclosed.

Conception and design: M. Harada, M. Kitagawa

Development of methodology: M. Harada

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Harada, S. Sakai, K. Kitagawa, C. Uchida

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Harada, S. Sakai, M. Mikamo, K. Nishimoto, H. Niida, H. Sugimura, T. Suda, M. Kitagawa

Writing, review, and/or revision of the manuscript: M. Harada, S. Sakai, T. Ohhata, K. Kitagawa, C. Uchida, H. Sugimura, M. Kitagawa

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Harada, S. Sakai, T. Ohhata, M. Mikamo, K. Nishimoto, H. Niida, Y. Kotake, H. Sugimura

Study supervision: M. Harada, T. Ohhata, M. Kitagawa

The authors thank Dr. Toru Suzuki for plasmids, Dr. Naoki Inui for useful suggestions, and Mr. Hisaki Igarashi, Mrs. Kiyomi Kinpara, Mrs. Mika Yoshida, and Mrs. Hazuki Yokota for technical support.

This work was supported in part by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to M. Kitagawa and H. Sugimura), the Ministry of Health, Labour and Welfare (to H. Sugimura), the Smoking Research Foundation (to H. Sugimura), The Uehara Memorial Foundation (to M. Kitagawa), and Princess Takamatsu Cancer Research Fund (to M. Kitagawa).

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.