This study was approved by Institutional Review Committee of Beijing Cancer Hospital. As summarized in Table I, a total of 154 NSCLC specimens were obtained from patients who underwent surgical resection from 1995 to 2000 in the Department of Thoracic Surgery II of Beijing Cancer Hospital. The median overall survival time was obtained from all the patients examined and the follow-up was from September 1995 till September 2005. The median follow-up time was 38.9 months (from 1.5 to 109 months). Lung tissues from 12 benign lung disease patients and 2 thoracic injury patients were collected as controls. All the patients and controls gave informed consent for the study.

NSCLC cell types included squamous carcinoma, adenocarcinoma (including bronchioloalveolar carcinoma), large cell carcinoma, adenosqumous carcinoma, and salivary adenocarcinoma. Benign diseases included inflammatory pseudotumor, tuberculosis, sclerotic hemangioma, and harmatoma. The tissue samples were histologically confirmed by two independent pathologists. Monkey kidney cell line COS-1 was obtained from the American Type Culture Collection and maintained in DMEM with 10% fetal bovine serum (FBS) (Gibco).

Tissue microarray was constructed as described previously using a manual tissue arrayer (Beecher Instruments) (21). Briefly, H&E slides of each tissue sample were reviewed by two pathologists. A tumor tissue block was chosen for the tissue microarray. Scarred, myxoid, and hypocellular areas were avoided. Two tissue cores with a diameter of 0.6-mm were taken from each block and inserted into the blank tissue microarray block. The whole set of tissue array contained two chips (A and B), in which chip A had 78 NSCLCs and 3 benign controls and chip B contained 76 NSCLCs and 11 benign controls.

The tissue microarrays were deparaffinized in xylene and ethanol. Antigen retrieval was done by heating the slides in an autoclave at 120°C for 3 min in citric acid buffer. The primary antibody was mouse monoclonal anti-TRA16 (22) (1:200 dilution in PBS) or rabbit monoclonal anti-ERβ (1:100 dilution in PBS, Dako). The immunostaining was carried out using the EnVision method (Dako) according to the manufacturer’s instructions. Brown staining was considered positive. Normal mouse IgG was used instead of the primary antibody as a negative control.

Cos-1 cells, which are TR2-negative and express very low level of TRA16, were transfected with pBig or pBig-TRA16 using SuperFect (Qiagen). The cells were then selected using 100 μg/ml hygromycin B to generate stable clones, which were confirmed by reporter gene assay. The established stable pBig-TRA16-Cos-1 cells (5×10⁴) per well were seeded in 12-well plates. After 12 h, the medium was changed to DMEM medium with 10% FBS for another 4 days and then treated with doxycycline (6 μg/ml) or DMSO for induction of TRA16 expression for 24 h.

For the MTT assay, 2×10³ of Cos-1 cells per well were plated in 96-well plates with RPMI-1640 medium containing 10% FBS and incubated at 37°C overnight. After 1, 2, 3, or 4 days, 200 μl of MTT (5 mg/ml, Sigma) was added to each well for 3-h incubation, then 2 ml of 0.04 M HCl in isopropyl alcohol was added to each well to stop the reaction. After 5 min of incubation at room temperature, the absorbance was read at 570 nm.

Cos-1 cells (2.5×10⁴) per well were plated in 12-well plates. The next day, the medium was changed to a medium containing 10% charcoal-stripped serum to deprive the cells of steroid hormones. The cells were then transfected the following day. Transfections were performed using SuperFect (Qiagen) according to the manufacturer's instructions. The cells were transiently cotransfected with reporter plasmids pERE-TK-Luc (a gift from Dr M. Nichols at University of Pittsburgh, Pittsburgh, PA) and ERβ with or without TRA16. After 24 h, the cells were incubated with 10 nM estrogen or androgen if necessary for 24 h and then harvested using lysis buffer (Promega) to detect the interaction between TRA16 and ERβ.

To further prove the interaction between TRA16 and ERβ, the Cos-1 cells were transiently co-transfected with reporter plasmid pG5-Luc and each of GAL4DBD, VP16, VP16-TRA16, and GAL4-ERβ in mammalian two hybrid system. As negative controls, the Cos-1 cells were transiently cotransfected with PPRE-Luc, MMTV-Luc, PSA-Luc, ARE-Luc, and the pRL-CMV (Promega, Madison, WI) to confirm the specifity of the interaction between TRA16 and ERβ.

To assess the interaction between TRA16 and TR2, we adopted three approaches. First, we transiently co-transfected TR2 and TRA16 or TR2 alone into Cos-1 cells to measure the inhibition of TR2 transcriptional activity by TRA16. Second, pBig-TRA16-Cos-1 and pBig-Cos-1 stable cells were transiently co-transfected with HCR-1-Luc reporter and TR2 for 24 h, then transfected cells were treated with doxycycline to induce TRA16 expression, and measure the TR2 activity. Third, we transiently co-transfected the reporter plasmid pG5-Luc and each of GAL4DBD, VP16, VP16-TR2, and GAL4-TRA16 in mammalian two-hybrid system to further prove the interaction between TRA16 and TR2. The luciferase activity was measured using the Dual-Luciferase System (Promega) by TD 20/20 luminometer. Values of the luciferase activity were corrected for protein concentration and presented as the mean ± SD of three independent experiments.

Cos-1 cells were transiently transfected with TR2 and either TRA16 or pcDNA4 control for 48 h and then harvested and dissolved in lysis buffer (1% Nonidet P-40, 10% glycerol, 135 mM NaCl, 40 mM Tris, pH 7.4, 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, and 1X protease inhibitor cocktail Roche). Cell lysates containing 500 μg of proteins were precleared with 20 μl of protein A/G Plus-agarose and 1.0 μg of normal mouse IgG (Santa Cruz Biotechnology) for 30 min. The supernatant was then mixed with a 1:100 dilution of mouse TR2 antibody at 4°C for 2 h, followed by adding protein A/G Plus-agarose and incubated for another 2 h. Immunoprecipitates obtained by spinning down protein A/G Plus-agarose were washed with PBS for three times and separated on SDS-8% PAGE. After transferring to the membrane, the protein was detected by anti-His-tag or anti-TR2 antibodies.

Cos-1 cells transiently co-transfected with TR2 and TRA16 were seeded on two-well Lab Tek Chamber slides (Nalge) for 48 h. Immunostaining was performed as described previously (18) by incubating with mouse anti-TRA16 monoclonal antibody and/or rabbit anti-TR2 monoclonal antibody, and then followed by incubating with either fluorescein-conjugated goat anti-mouse or anti-rabbit antibodies. Coverslips were fixed on the glass slides with a drop of DAPI to stain the nucleus. The slides were observed under 400-fold magnification of a fluorescence microscope or confocal fluorescence microscope.

SPSS v11.0 software was used to perform statistical analysis. Chi-square test was used to analyze the assocation between TRA16 expression and different clinical factors in NSCLC patients. Univariate and multivariate Cox regression hazards models were used to analyze the correlation of individual factors with overall survival of NSCLC patients. Log-rank test and Kaplan-Meier survival analysis were used to examine the effect of individual factors on prognosis of NSCLC patients. Values of P<0.05 were considered significant.

We previously found that TRA16 was highly expressed in lung adenocarcinoma cell line H1299 (18). To investigate the expression of TRA16 in NSCLC tissues, we performed tissue microarrays with 154 NSCLCs, in which 134 NSCLC samples were qualified for data analysis (Table II). The results of TRA16 tissue microarray and its expression level and distribution in different NSCLC cell types are exemplified in Fig. 1A-D. We found that all the control samples were TRA16-negative, but 120 lung cancer samples were TRA16-positive (+, ++, or +++) with the positive rate of 89.55% (120/134) (P<0.001). When categorizing the samples with the staining intensity of ++ or +++ as strong positive, there were 105 strong positive samples accounting for 78.36% (105/143) of the NSCLCs. The 105 samples included 55 adenocarcinoma (83.3%, 55/66), 43 squamous carcinoma (71.7%, 43/60). No correlation of TRA16 expression with NSCLC cell types was observed (P=0.230). The Chi-square test indicated that the TRA16 expression was correlated with TNM stage (P=0.016).

An univariate Cox regression analysis of clinicopathologic parameters and TRA16 expression showed that TNM stage, tumor size, lymph node metastasis, tumor thrombus, cell differentiation, and TRA16 expression were significantly associated with overall survival of NSCLC patients (Table III). The multivariate Cox regression analysis showed that lymph node metastasis, tumor thrombus formation, poor cell differentiation, and high expression of TRA16 were independent prognostic factors in NSCLC (Table III). The Kaplan-Meier analysis showed that the patients with higher TRA16 expression had significantly shorter overall survival than those with lower TRA16 expression (P=0.0348) (Fig. 1E), similarly to the standard clinical prognostic factors of TNM stage, lymph node metastasis, and tumor differentiation.

Our previous studies demonstrated that TRA16 inhibited the function of TR4 in lung cancer cell line H1299 and resultantly relieved TR4-repressed ERβ activity (18). To further determine the relationship between TRA16 and ERβ in NSCLC, we analyzed the expressions of TRA16 and ERβ in 71 NSCLC tissues and 2 benign samples using the tissue microarrays. We found that among the 71 NSCLC tumors, 35 cases were ERβ-positive with 22 cases (+), 9 cases (++), and 4 cases (+++), while 36 cases and two benign controls were ERβ-negative. When using the TRA16 expression to stratify the 71 cases into no or low (+) level group (n=15) and high level (++ − +++) group (n=56) and then analyzing the difference of the ERβ expression between the two groups, we found that the ERβ expression in the high level group of TRA16 was significantly higher than that in the no or low level group, while only one of 15 samples in the no or low level group of TRA16 showed ERβ-positive (P<0.001) (Table II), suggesting their functional association.

The increased expression of TRA16 in NSCLC indicated that TRA16 could play an important role in lung cancer development. Thus, we performed the MTT assay to study the effect of TRA16 on cell growth by transfecting doxycycline-inducible TRA16 in pBig vector into Cos-1 cells. We found that the doxycycline-treated group had a significantly higher cell growth than the DMSO-treated control group.

To further explore the functional association between TRA16 and ERβ, we performed several reporter gene assays by transient transfection of different combinations of plasmids into Cos-1 cells (Fig. 2). We found that when transiently co-transfecting ERβ response element (ERE)-Luc reporter and ERβ with or without TRA16 into Cos-1 cells, after adding 10 nM of estradiol (E₂) to stimulate ERβ activity, TRA16 significantly enhanced the ERβ activity in Cos-1 cells compared with the ERβ alone (Fig. 2A). Further, in the mammalian two-hybrid assay by transfecting the reporter plasmid pG5-Luc and each of GAL4DBD, VP16, VP16-TRA16, and GAL4-ERβ in various combinations as indicated into Cos-1 cells, we found that after the stimulation of E₂, the co-presence of TRA16 and ERβ significantly enhanced the ERβ activity in Cos-1 cells (Fig. 2B), indicating that TRA16 interacted with ERβ.

To clarify the specificity of the association of TRA16 and ERβ, similar but negative control experiments were performed by co-transfecting TRA16 and PPAR or androgen receptor (AR) in Cos-1 cells. As shown in Fig. 2C and D, TRA16 had little effect on activation of both PPAR and AR.

We previously demonstrated that TRA16 bound to TR4 and blocked TR4 dimerization and thus inhibited TR4 activity in lung cancer cells (18). TR2 and TR4 are highly homologous and TR2 can suppress ERβ-mediated cell growth in lung cancer cells in vitro(19,20). To assess a potential relationship between TRA16 and TR2, we performed similar reporter gene assays as described above. We found that when HCR-1-Luc reporter gene and TR2 with or without TRA16 were co-transfected into Cos-1 cells, TRA16 strongly inhibited TR2 transcriptional activity in Cos-1 cells (Fig. 3A). Similarly, the doxycycline-inducible assay using the pBig-Cos-1 and pBig-TRA16-Cos-1 cell system showed that after treatment of 6 μg/ml doxycycline, TR2 activity was significantly lower in pBig-TRA16-Cos-1 cells than that in pBig-Cos-1 cells (Fig. 3B). Moreover, the mammalian two-hybrid assay demonstrated that the co-expression of TRA16 and TR2 significantly enhanced the reporter Luc activity compared with TR2 or TRA16 alone (Fig. 3C), suggesting the association of TRA16 and TR2.

Furthermore, the immunoprecipitation assay showed that TRA16 directly interacted with TR2 (Fig. 3D), and TRA16 and TR2 were co-localized in the nuclear membrane by immunocytofluorescence assay (Fig. 3E). Taken together, the findings suggest that TRA16 interact with TR2 to release TR2-inhbitory role in ERβ activity in addition to its direct stimulation of ERβ signaling.