A Candidate Tumor Suppressor in Lung Cancer

FS Falvella1, F Colombo1, M Spinola1, M Campiglio1, U Pastorino2 and TA Dragani1

1Department of Experimental Oncology and Laboratories, Fondazione IRCCS Istituto Nazionale Tumori, Milan, Italy and 2Unit of Thoracic Surgery, Fondazione IRCCS Istituto Nazionale Tumori, Milan, Italy

BHLHB3 is a basic helix-loop-helix (bHLH) domain- (Fujimoto et al., 2001). Both genes encode bHLH containing protein that acts as a transcriptional repressor. transcriptional repressors involved in the regulation of We found that BHLHB3 transcript levels were low in the mammalian molecular clock (Honma et al., 2002). three human lung cancer cell lines and downregulated Retroviral insertional mutagenesis analysis has impli- in human lung adenocarcinomas as compared to normal cated the mouse Bhlhb3 gene in the development of lung tissue. BHLHB3 gene overexpression inhibited mouse leukemia/lymphoma (Suzuki et al., 2002). In- colony formation of A549, NCI-H520 and NCI-H596 deed, in two cases of leukemia/lymphoma, the retrovirus lung cancer cells. The reduced colony growth was likely integration disrupted the gene coding sequence (Suzuki due to inhibition of cell proliferation as suggested by the et al., 2002) (http://genome2.ncifcrf.gov/RTCGD), sug- downregulation of cyclin D1 (CCND1) expression in gesting that Bhlhb3 loss of function is required for NCI-H520 cells transfected to overexpress the BHLHB3 leukemia/lymphoma development. No information is gene; no evidence of apoptosis was observed. These results currently available on the role of Bhlhb3 in other tumor point to the potential role of the BHLHB3 protein as a types in mice. However, on the basis of our previous tumor suppressor for lung cancer. report suggesting an association between the BHLHB3- Oncogene (2008) 27, 3761–3764; doi:10.1038/sj.onc.1211038; Ala298Val polymorphism and survival of lung adeno- published online 28 January 2008 carcinoma (ADCA) patients (Falvella et al., 2006), we tested BHLHB3for its putative tumor suppressor Keywords: BHLHB2; DEC1; DEC2; SHARP-1; activity in human lung cancer. STRA13; lung cancer BHLHB3transcript levels in human lung cancer cell lines were B100- (A549, NCI-H596) to B1000-fold (NCI-H520) lower than those in a pool of 20 RNA samples of human normal lung tissue, as determined The basic helix-loop-helix (bHLH) proteins form by kinetic reverse transcriptase–PCR. Analysis of heterodimers or homodimers via their HLH domains BHLHB3gene expression in 18 pairs of matched and bind to DNA through a region of basic amino acids normal tumor lung tissue obtained from lung ADCA immediately N terminal to the HLH dimerization patients revealed downregulation of the gene in nine domain (Davis and Turner, 2001). The bHLH transcrip- tumors (range: 2- to 11-folds) and upregulation in five tion factors are involved in coordinated regulation tumors (range: 2- to 3.5-folds); four pairs were excluded of gene expression and cell differentiation in most from the analysis since BHLHB3expression was mammalian tissues (Honma et al., 2002), affecting undetectable in either normal or tumor tissue. Overall, neurogenesis, vasculogenesis, mesoderm segmentation, there was a slight downregulation of BHLHB3expres- myogenesis and T lymphocyte development (Davis and sion in lung ADCA as compared to normal lung tissue, Turner, 2001). bHLH are targets of the Notch signaling but the difference was not statistically significant pathway (Davis and Turner, 2001; Iso et al., 2003), (Figure 1, top panel). We also measured BHLHB2 which can be dysregulated in a variety of human mRNA levels in our matched samples; all 14 pairs had neoplasms (Davis and Turner, 2001; Allenspach et al., detectable levels of BHLHB2 mRNA and the gene was 2002). Proteins encoded by the Class B bHLH protein 3 overexpressed in 2 tumor samples, downregulated in 11 (BHLHB3: also known as DEC2, SHARP-1) gene and tumors and unchanged in 1 pair (Figure 1, bottom its homolog Class B bHLH protein 2 (BHLHB2: also panel; P ¼ 0.01, Kruskal–Wallis test). The finding of known as DEC1, STRA13, SHARP-2) share 42% total downregulated BHLHB2 expression in lung cancer sequence identity and 97% identity in the bHLH region contrasts with that of (Li et al., 2003), who reported a slight upregulation of BHLHB2 (DEC1) expression in Correspondence: Dr TA Dragani, Department of Experimental lung cancer in a small series (n ¼ 4) of lung cancer/ Oncology and Laboratories, Fondazione IRCCS Istituto Nazionale normal lung tissue pairs (Li et al., 2003). In addition, we Tumori, Via G Venezian 1, Milan 20133, Italy. found a direct correlation between BHLHB3and E-mail: [email protected] GenBank acc. number: AY665467. BHLHB2 transcript levels (Spearman’s r ¼ 0.508; Received 22 August 2007; revised 6 December 2007; accepted 11 P ¼ 0.006), whereas (Li et al., 2003) described an December 2007; published online 28 January 2008 inverse relationship between BHLHB3and BHLHB2 BHLHB3 in lung cancer FS Falvella et al 3762 cancer cells stably transfected to overexpress BHLHB3(see Figure 1 legend for details of kinetic reverse transcriptase) revealed only a slight decrease (0.56±0.10-fold, mean±s.e., n ¼ 12) as compared to vector-transfected cells, arguing against any significant, if any, direct modulation of BHLHB2 expression by BHLHB3and instead, suggesting that the observed correlation between BHLHB2 and BHLHB3expression in the normal lung and lung ADCA tissues rests in common transcription regulation. The discrepancy between the experiments carried out in mice and our findings may rest in the differences in the cell lines or species analysed. To analyse the potential functional role of BHLHB3 in human lung cancer, we assessed the clonogenicity of human lung cancer cell lines A549, NCI-H520 and NCI- H596, which expressed low levels of BHLHB3mRNA compared to normal lung tissue, after transfection with mammalian expression vectors containing full-length BHLHB3cDNAs corresponding to either the Ala298 or Val298 variant (Figure 2a). We also transfected a non- lung cancer cell line (HepG2) expressing B1000-fold lower levels of BHLHB3transcript than in normal lung tissue. Clonogenic activity decreased B3- (A549) to B15-fold (NCI-H520, NCI-H596) in the BHLHB3- overexpressing lung cancer cells (Po0.0001), with no significant difference between the two variants (Figures 2b and c). HepG2 transfectants showed no Figure 1 mRNA expression levels in normal and tumor lung difference from controls in clonogenic activity, which tissue of lung adenocarcinoma (ADCA) patients. Lung ADCA surgical specimens and corresponding normal lung tissue were was uniformly high in both cases (not shown). obtained from pathologically documented patients enrolled at To determine whether the decreased clonogenic Istituto Nazionale Tumori, Milan, Italy. Total RNAs were activity of BHLHB3-transfected NCI-H520 cells might extracted with RNeasy Midi kit (Qiagen, Valencia, CA, USA). reflect induction of apoptosis, NCI-H520 cells were cDNAs were reverse transcribed using oligo(dT) primers and transiently transfected and analysed after 48 and 72 h by ThermoScript RT–PCR System (Invitrogen, Carlsbad, CA, USA). mRNA levels were assayed by kinetic reverse transcriptase–PCR indirect immunofluorescence assay in live cells using the using Hs00229146_m1 (BHLHB3) and Hs00186419_m1 Annexin V-FITC kit (Bender MedSystems, Vienna, (BHLHB2) TaqMan Gene Expression Assays (Applied Biosys- Austria); no evidence of apoptosis was obtained at tems). The real-time PCR amplification mixture contained cDNA either time point (data not shown). As an additional template diluted in RNase-free water, 10 ml 2X TaqMan Universal PCR Master Mix No AmpErase UNG and 1 ml of 20X TaqMan assay of apoptosis, we used an anti-PARP antibody Gene Expression Assay Mix (Applied Biosystems). Reactions were (Cell Signaling Tech Inc., Danvers, MA, USA) by run in duplicate on the 7900HT System (Applied Biosystems) for western blot in stable transfected cells; again, we found 40 cycles (95 1C for 15 s, 60 1C for 1 min) after enzyme activation no evidence of PARP protein cleavage (not shown). (95 1C for 10 min). The human hypoxanthine phosphoribosyltrans- Instead, quantitative PCR of NCI-H520 cells stably ferase 1 (HPRT1) gene (Hs99999909_m1 TaqMan Gene Expres- sion Assays, Applied Biosystems) was used as control for possible transfected to overexpress BHLHB3(Figure 3)and differences in cDNA amounts. Values were calculated using the using cyclin D1 as a marker of proliferation (Berthet comparative Ct method, and relative expression levels are directly and Kaldis, 2007) revealed a decrease (0.25±0.04-fold, proportional to the ÀDCt values. kRT–PCRs for BHLHB3(top) n ¼ 8) of cyclin D1 mRNA expression (measured using or BHLHB2 (bottom) were carried out in 14 paired normal and tumor samples. The line within each box represents the median Hs00277039_m1 (CCND1) TaqMan Gene Expression ÀDCt value; upper and lower edges of each box represent the Assays; Applied Biosystems, Foster City, CA, USA) in seventy-fifth and twenty-fifth percentile, respectively; upper and BHLHB3-transfected versus vector-transfected cells, lower bars indicate the highest and lowest values determined, consistent with western blot analysis (Figure 3) and respectively. with a previous study (Azmi et al., 2004) showing that BHLHB3overexpression in mouse myoblast cell line C2C12 promotes cell cycle exit in association with decreased levels of cyclin D1 protein expression. Indeed, expression. On the basis of a previous study indicating inhibition of cyclin D1 protein expression is associated repression of mouse Bhlhb2 promoter activity by Bhlhb3 with cell cycle arrest in cancer cell lines (Berthet and transfection in mouse fibroblast cell line 10T1/2 (Azmi Kaldis, 2007). et al., 2003), we tested the hypothesis that BHLHB3 BHLHB3has been reported to suppress transcription directly affects mRNA levels of BHLHB2. Measure- of the CYP7A gene in a hepatocarcinoma cell line ment of BHLHB2 mRNA levels in NCI-H520 lung (Noshiro et al., 2004). Consistent with that study,

Oncogene BHLHB3 in lung cancer FS Falvella et al 3763 HepG2 cells transfected with a luciferase construct (Ala298Val) of the BHLHB3gene were observed in the containing the 50 flanking region of rat CYP7A, an suppression of CYP7A transcriptional activity in either expression plasmid containing the rat D site of the HepG2 or A549 cells. Transrepression of CYP7A albumin promoter (albumin D-box) binding protein (DBP) to enhance transcriptional activity of the CYP7A promoter, a mammalian expression vector containing full-length BHLHB3cDNA (Val or Ala variant), and an internal control plasmid showed a B5-fold reduction in luciferase activity (Po0.0001) as compared to controls (Figure 4), indicating the suppression of DBP-induced CYP7A promoter activity by BHLHB3. This suppression was also apparent in similarly transfected A549 cells, although DBP-induced expression levels were Figure 3 Cyclin D1 expression in NCI-H520 cells transfected with lower and the reduction in luciferase activity was slight BHLHB3recombinant or control empty vector. Clones transfected using FuGENE HD Transfection Reagent (Roche Applied (1.4- to 1.7-fold; Po0.01) as compared to those in HepG2 Science) were selected with 2 mgmlÀ1 blasticidin (Invitrogen) for 2 cells (Figure 4). No significant differences (ANOVA and weeks. Nuclear protein extraction and western blot analysis were Sheffe’s test) between the two nonsynonymous alleles carried out as detailed in Figure 2 legend; anti-V5 (Invitrogen), anti-cyclin D1 (DCS-6, Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) or anti-Histone H1 (FL-219, Santa Cruz) antibodies were used for immunoblotting. A pool of two stable empty vector transfectants (lane 1) and a pool of stable BHLHB3 transfectants (two Ala and two Val variants, lane 2) showed undetectable or clearly detectable levels of BHLHB3recombinant protein, respectively (top panel). Middle panel: the cyclin D1 protein-specific band (arrow) was apparent in control (lane 1) and barely detectable in BHLHB3transfectants (lane 2). Molecular size markers are indicated on the right. Histone H1 was used as protein loading control (bottom panel).

Figure 2 Colony formation in human lung cancer cell transfectants. Human cell lines A549, NCI-H520 and NCI-H596 (American Type Culture Collection, Rockville, MD, USA), derived from lung carcinoma, lung squamous cell carcinoma and lung adenosquamous carcinoma, respectively, were used for in vitro functional assays. The BHLHB3gene coding region was amplified from the cDNA clone isolated from a full-length human lung plasmid cDNA library (OriGene Technologies, Rockville, MD, USA; Val variant) or from lung tissue (Ala variant) using primers 50-agaacatggacgaaggaatt-30 and 50-gggagcttcctttcctgg-30. PCR aliquots were subcloned in the pEF6/V5-His-TOPO vector (Invitrogen) in-frame with the V5 epitope and polyhistidine tag. Clones were screened and sequenced using vector-specific primers. (a) Immunodetection of recombinant BHLHB3variant proteins in transfected A549, NCI-H520 and NCI-H596 cells using anti-V5 antibody (Invitrogen). Cells were transfected using FuGENE HD Transfection Reagent (Roche Applied Science, Basel, Switzerland). Nuclear proteins were extracted with CelLytic NuCLEAR Extraction Kit (Sigma, St Louis, MO, USA). Equal amounts of protein were resolved by SDS–polyacrylamide gel electrophoresis, transferred onto Hybond- C nitrocellulose membranes (GE Healthcare Life Sciences, Buck- inghamshire, UK) and immunoblotted with anti-V5 (Invitrogen). The ECL Plus Western Blotting Detection System (GE Healthcare Life Sciences) was used for detection. Lane 1, Ala variant, lane 2, Val variant, lane 3, empty vector. (b) Inhibition of in vitro colony formation in BHLHB3-transfected human lung carcinoma A549, NCI-H520 and NCI-H596 cells. Cell lines were transfected with recombinant or control empty vector (pEF6/V5-His) using Super- Fect Transfection Reagent (Qiagen) (A549 and NCI-H596) or Effectene Transfection Reagent (Qiagen) (NCI-H520). Transfected clones were selected with 5 mgmlÀ1 (A549) or 2 mgmlÀ1 (NCI-H520 and NCI-H596) blasticidin (Invitrogen) for 2 weeks, methanol-fixed and stained with 10% Giemsa. (c) Numbers (mean±s.e.) of Giemsa-stained colonies produced by empty vector (empty bars) or BHLHB3-transfected cells. Colonies were automatically counted with either Quantity One software (Bio-Rad Laboratories, Her- cules, CA, USA) (NCI-H520 or NCI-H596 cells) or ImageQuant TL software (GE Healthcare Life Sciences) (A549 cells). All experiments were carried out in sextuplicate (Ala298, diagonal- filled bars; Val298, black-filled bars; asterisk, Po0.0001).

Oncogene BHLHB3 in lung cancer FS Falvella et al 3764 activity was instead not observed in NCI-H520 and NCI-H596 cells (not shown). These results indicate that transrepression of the CYP7A promoter and inhibition of clonogenic activity by BHLHB3are not correlated, since BHLHB3was an efficient transrepression factor in HepG2 cells but did not inhibit colony formation in these cells and, on the contrary, BHLHB3did not affect transrepression of CYP7A activity in NCI-H520 and NCI-H596 cells but strongly inhibited their clonogenic activity. The possibility that BHLHB3transrepression activity relevant to tumor suppression is associated with another as yet uncharacterized promoter(s) or is mediated by cell-type-specific factors remains to be examined. Our results point to a possible tumor suppressor role of BHLHB3in lung cancer, with no apparent difference Figure 4 Transrepression activity of BHLHB3. To test specific transrepression activity of the BHLHB3variants, A549 and HepG2 between the Ala298Val variants. Such activity may be cells were co-transfected with a luciferase construct under the mediated by cyclin D1 downregulation. control of the pCYP7A promoter (pCYP7A(1.8)-Luc), the expression plasmid pSCT/DBP, the recombinant mammalian expression vector pEF6/V5-His containing the full-length cDNA Acknowledgements of the BHLHB3gene (Val or Ala variant) and the internal control plasmid pGL4.70[hRluc] (Promega, Madison, WI, USA). Cells We thank Dr Mitsuhide Noshiro (Hiroshima University, were harvested 48 h after transfection and luciferase activities Hiroshima, Japan) for the gift of the plasmid pCYP7A(1.8)- were analysed with the Dual-Luciferase Reporter Assay System Luc, and Professor Ueli Schibler (University of Geneva, (Promega). Signals were normalized to Renilla luciferase activity. Each bar represents mean±s.e. (A549, n ¼ 12; HepG2, n ¼ 3). Geneva, Switzerland) for the gift of the plasmid pSCT/DBP. Empty bars, empty vector; diagonal-filled bars, BHLHB3Ala298 This work was funded in part by grants from Associazione and variant; black-filled bars, BHLHB3Val298 variant (asterisks, Fondazione Italiana Ricerca Cancro (AIRC and FIRC) and Po0.01). from Fondo Investimenti Ricerca di Base (FIRB).

References

Allenspach EJ, Maillard I, Aster JC, Pear WS. (2002). Notch signaling new member of basic helix-loop-helix proteins. Biochem Biophys Res in cancer. Cancer Biol Ther 1: 466–476. Commun 280: 164–171. Azmi S, Ozog A, Taneja R. (2004). Sharp-1/DEC2 inhibits skeletal Honma S, Kawamoto T, Takagi Y, Fujimoto K, Sato F, Noshiro M muscle differentiation through repression of myogenic transcription et al. (2002). Dec1 and Dec2 are regulators of the mammalian factors. J Biol Chem 279: 52643–52652. molecular clock. Nature 419: 841–844. Azmi S, Sun H, Ozog A, Taneja R. (2003). mSharp-1/DEC2, a basic Iso T, Kedes L, Hamamori Y. (2003). HES and HERP families: helix-loop-helix protein functions as a transcriptional repressor of E multiple effectors of the Notch signaling pathway. J Cell Physiol box activity and Stra13expression. J Biol Chem 278: 20098–20109. 194: 237–255. Berthet C, Kaldis P. (2007). Cell-specific responses to loss of cyclin- Li Y, Xie M, Song X, Gragen S, Sachdeva K, Wan Y et al. (2003). dependent kinases. Oncogene 26: 4469–4477. DEC1 negatively regulates the expression of DEC2 through Davis RL, Turner DL. (2001). Vertebrate hairy and enhancer of split binding to the E-box in the proximal promoter. J Biol Chem 278: related proteins: transcriptional repressors regulating cellular 16899–16907. differentiation and embryonic patterning. Oncogene 20: 8342–8357. Noshiro M, Kawamoto T, Furukawa M, Fujimoto K, Yoshida Y, Falvella FS, Spinola M, Manenti G, Conti B, Pastorino U, Skaug V Sasabe E et al. (2004). Rhythmic expression of DEC1 and DEC2 in et al. (2006). Common polymorphisms in D12S1034 flanking genes peripheral tissues: DEC2 is a potent suppressor for hepatic RASSF8 and BHLHB3are not associated with lung adenocarcino- cytochrome P450s opposing DBP. Genes Cells 9: 317–329. ma risk. Lung Cancer 56: 1–7. Suzuki T, Shen H, Akagi K, Morse HC, Malley JD, Naiman DQ et al. Fujimoto K, Shen M, Noshiro M, Matsubara K, Shingu S, Honda K (2002). New genes involved in cancer identified by retroviral et al. (2001). Molecular cloning and characterization of DEC2, a tagging. Nat Genet 32: 166–174.

Oncogene