Published OnlineFirst August 31, 2010; DOI: 10.1158/0008-5472.CAN-10-0860 Published OnlineFirst on August 31, 2010 as 10.1158/0008-5472.CAN-10-0860 Molecular and Cellular Pathobiology Cancer Research hnRNP A2/B1 Modulates Epithelial-Mesenchymal Transition in Lung Cancer Cell Lines Jordi Tauler1, Enrique Zudaire2, Huaitian Liu3, Joanna Shih4, and James L. Mulshine1 Abstract Heterogeneous nuclear ribonucleoprotein A2/B1 (hnRNP A2/B1) has been reported to be overexpressed in lung cancer and in other cancers such as breast, pancreas, and liver. However, a mechanism linking hnRNP A2/ B1 overexpression and progression to cancer has not yet been definitively established. To elucidate this mech- anism, we have silenced hnRNPA2/B1 mRNA in non–small-cell lung cancer cell lines A549, H1703, and H358. These cell lines present different levels of expression of epithelial-to-mesenchymal transition (EMT) markers such as E-cadherin, fibronectin, and vimentin. Microarray expression analysis was performed to evaluate the effect of silencing hnRNP A2/B1 in A549 cells. We identified a list of target genes, affected by silencing of hnRNP A2/B1, that are involved in regulation of migration, proliferation, survival, and apoptosis. Silencing hnRNP A2/ B1 induced formation of cell clusters and increased proliferation. In the anchorage-independent assay, silencing hnRNP A2/B1 increased colony formation by 794% in A549 and 174% in H1703 compared with a 25% increase in proliferation, in both cell lines, in a two-dimensional proliferation assay. Silencing hnRNP A2/B1 decreased migration in intermediate cell line A549 and mesenchymal cell line H1703; however, no changes in proliferation were observed in epithelial cell line H358. Silencing hnRNP A2/B1 in A549 and H1703 cells correlated with an increase of E-cadherin expression and downregulation of the E-cadherin inhibitors Twist1 and Snai1. These data suggest that expression of hnRNP A2/B1 may play a role in EMT, in nonepithelial lung cancer cell lines A549 and H1703, through the regulation of E-cadherin expression. Cancer Res; 70(18); OF1–11. ©2010 AACR. Introduction development of lung cancer (16). High levels of hnRNP A2/B1 expression also colocalized with areas of genetic in- Heteregeneous nuclear ribonucleoprotein A2/B1 (hnRNP jury, as assessed by microsatellite instability and loss of A2/B1) is a member of the hnRNP family of proteins. The heterozygosity (17). hnRNP A2/B1 has been consistently re- hnRNPs are a complex group of RNA-binding proteins that ported as an informative biomarker for lung cancer screen- play a key role in mRNA processing and telomere biogenesis ing using a variety of analytic approaches, including (1–4). hnRNP A2/B1 overexpression has been described in expression studies and proteomic analysis (7, 8, 18, 19). many cancers, including breast, pancreas, liver, and gastroin- We have also previously reported that hnRNP A2/B1 expres- testinal cell lines (5–8). Recent reports have described a crit- sion is tightly controlled during lung development (20). A ical role of hnRNP A2/B1 in the regulation of fundamental recent article has shown the impact of cell density on biology especially related to cancer, such as migration (9) hnRNP A2/B1 expression, reporting that mouse lung cell and aerobic glycolysis (10). We have previously reported that lines E9 and E10 both showed increase of hnRNP A2/B1 pro- hnRNP A2/B1 is a marker for early lung cancer (11–15). Over- tein related to proliferation. At confluency, the nontumori- expression of hnRNP A2/B1 in exfoliated bronchial epithelial genic E10 cell line downregulated hnRNP A2/B1 mRNA cells of archival sputum samples correlated with eventual production, whereas the tumorigenic E9 cell line showed an increase of total hnRNP A2/B1 production (21). Authors' Affiliations: 1Laboratory of Lung Cancer Biology, Section of However, the contribution of hnRNP A2/B1 to the carcino- Medical Oncology, Rush University Medical Center, Chicago, Illinois; genic process is still poorly understood. Some reports have 2NCI Angiogenesis Core Facility, National Cancer Institute, NIH, Advanced Technology Center, Gaithersburg, Maryland; 3Science used silencing tools in different cancer cell lines to study Applications International Corporation, Rockville, Maryland; and the function of hnRNP A2/B1 (22–24). Silencing hnRNP A2/ 4Biometric Research Branch, Division on Cancer Treatment and B1 in combination with hnRNP A1 reduced cell growth by Diagnosis, National Cancer Institute, NIH, Bethesda, Maryland inducing cell death in cancer cell lines (22). Silencing hnRNP Note: Supplementary data for this article are available at Cancer A2/B1 alone in squamous carcinoma cell line, Colo16, regu- Research Online (http://cancerres.aacrjournals.org/). lated genes involved in cell cycle and proliferation; however, Corresponding Author: Jordi Tauler, Rush University Medical Center, Room 1413, Jelke Building, 1750 West Harrison Street, Chicago, IL it had a modest effect in Colo16 proliferation (23, 24). 60612. Phone: 312-942-3595; Fax: 312-563-3377; E-mail: Jordi_Tauler@ Recent evidence is showing a prominent role of hnRNP rush.edu. A2/B1 in fundamental biological functions such as regulation doi: 10.1158/0008-5472.CAN-10-0860 of cell metabolism (10), migration and invasion (9, 25, 26), ©2010 American Association for Cancer Research. proliferation (22–24), and cellular response to mitochondrial www.aacrjournals.org OF1 Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst August 31, 2010; DOI: 10.1158/0008-5472.CAN-10-0860 Tauler et al. stress (26). Moreover, hnRNP A2/B1 activation of invasion cDNA expression vectors phenotype involves different mechanisms including activa- cDNA expression vector for hnRNP A2/B1, based on tion of the CXCL-12/CXCR4 axis (25), alternative splicing TrueORF cDNA clones system, was acquired from Origene. of TP53INP2 (9), and activation of invasive behavior, after Cell line A549 was transfected with Fugene 6 (Roche), depletion of mitochondrial DNA, through hnRNP A2/B1– and transfected cells were selected using the protocol de- mediated cooperative effect of NF-κB, NFAT, CREB, scribed above. Cell lines A549 EV-TC (Empty Vector True and C/EBPδ (26). Interestingly, it has been reported that Clone) and A549 A2B1rec (expressing hnRNP A2/B1) were depletion of mitochondrial DNA in cancer cells promotes obtained. epithelial-mesenchymal transition (EMT; ref. 27). Moreover, it has also been proposed that changes in morphology, in- Microarray data generation vasion, and metastasis potential could be the result of A549 EV and A549 siA cells were seeded at 1.5 × 106 in a EMT (28, 29). T75 flask. At 60 hours postseeding, total RNA was extracted To investigate how the overexpression of hnRNP A2/B1 af- using RNeasy mini-kit (Quiagen). First-strand cDNA, double- fects this important functional biology in lung cancer, we strand cDNA, and cRNA were synthesized and processed modified the expression of hnRNP A2/B1. Because hnRNP according to the manufacturer's protocol (Affymetrix). Mi- A2/B1 is playing a role in the regulation of migration activity, croarray study was performed using Affymetrix HG U133 we used lung cancer cell lines with different levels of EMT Plus 2.0. Protocol for data analysis can be found in Supple- phenotype markers such as mesenchymal H1703, intermedi- mentary Materials (Supplementary Microarray Data Analysis ate A549, and epithelial H358 (30). We have, for the first time, and Legends, and Supplementary Figs. S1–3). noted significant changes in proliferation and migration in nonepithelial H1703 and A549 lung cancer cell lines associat- Real-time PCR ed with changes in cell morphology, suggesting that hnRNP For real-time PCR of samples depicted in Figs. 1A,B, 4A,B, A2/B1 may be an important regulator of EMT and could pro- and 5D, cells were seeded at 2 × 106 in a T75 flask. For real- vide a mechanistic link between hnRNP A2/B1 overexpres- time PCR for samples in Fig. 5A to C, cells were seeded at sion and tumor progression. 0.7 × 106 in a T25 flask. Total RNA was extracted after 24 hours postseeding (Figs. 1A,B, 4A,B, and 5D) and at 0, 2, 4, and 6 hours (Figs. 5A–C) with RNeasy Mini-kit (Qiagen). Materials and Methods Total RNA (0.8–2 μg) for each analyzed sample was reverse transcribed in a final volume of 25 μL using SuperScript Cell cultures First-Strand Synthesis kit (Invitrogen). Quantitative real-time Cell lines used in this study include A549, H1703, and PCR reaction was performed in a final volume of 25 μL con- H358 non–small cell lung cancer (NSCLC) cell lines ob- taining 2 μL of cDNA (1:10 dilution) and 400 nmol/L of pri- tained from the American Type Culture Collection. These mers and SybrGreen PCR master mix (Applied Biosystems), cell lines present different EMT phenotypes as determined and run in an Opticon Cycler (MJ Research) and in a 7300 by the expression of EMT markers. H1703 has been classi- Cycler (Applied Biosystems). All samples were amplified in fied as mesenchymal, whereas A549 and H358 display a triplicate using the following cycle scheme: 50°C for 2 min- more epithelial phenotype (30). Cells were maintained in utes; 95°C for 10 minutes; and 45 cycles of 95°C for 30 seconds, RPMI 1640 supplemented with 5% fetal bovine serum 60°C for 30 seconds. Final mRNA levels were normalized to (FBS; Hyclone), in 95% air and 5% CO2 at 37°C. Cells were 18S rRNA levels. All results are plotted as columns (mean) counted and seeded at 1 × 106 per flask in a new T75 flask and bars (SD), calculated with 7300 Real-Time PCR System at every passage. SD Software (Applied Biosystems). Primer sequences can be found in Supplementary Table S1. siRNA expression vectors Oligonucleotides for the siRNA vectors were designed with Western blotting the Web-based software provided by Oligoengine.