|Home | About | Journals | Submit | Contact Us | Français|
Lung cancer is a leading cause of cancer death, where the amplification of oncogenes contributes to tumorigenesis. Genomic profiling of 128 lung cancer cell lines and tumors revealed frequent focal DNA amplification at cytoband 14q13.3, a locus not amplified in other tumor types. The smallest region of recurrent amplification spanned the homeobox transcription factor TITF1 (thyroid transcription factor 1; also called NKX2-1), previously linked to normal lung development and function. When amplified, TITF1 exhibited increased expression at both the RNA and protein levels. Small interfering RNA (siRNA)- mediated knockdown of TITF1 in lung cancer cell lines with amplification led to reduced cell proliferation, manifested by both decreased cell-cycle progression and increased apoptosis. Our findings indicate that TITF1 amplification and overexpression contribute to lung cancer cell proliferation rates and survival and implicate TITF1 as a lineage-specific oncogene in lung cancer.
Lung cancer is the leading cause of cancer death in the United States (Jemal et al., 2007). In lung cancers, the amplification of oncogenes such as MYC, KRAS, MET, EGFR, ERBB2, CCND1 and CDK4 contributes to tumor development and progression, and amplified genes have become important targets for molecularly-directed therapies (Sato et al., 2007). To discover novel amplicons, we profiled 52 non-small cell lung cancer (NSCLC) cell lines and 76 NSCLC tumors (36 adenocarcinomas including 2 metastases, and 40 squamous cell carcinomas including 1 metastasis), by array-based comparative genomic hybridization (aCGH) on cDNA microarrays (Pollack et al., 1999) covering ~22 000 human genes with a median inter-probe distance of ~30 kb. The most frequent focal DNA amplification not associated with a previously known oncogene occurred at cytoband 14q13.3 (Figure 1a).
Gain at 14q13.3 was present in 17 of 52 (33%) lung cancer cell lines, where it was more often observed in cell lines derived from adenocarcinomas (including bronchioloalveolar carcinomas) compared to other histologies (P = 0.04, Fisher’s exact test; unspecified NSCLCs excluded from the analysis) (Figure 1b). 14q13.3 gain was also detectable in 4 of 36 (11%) lung adenocarcinomas and in 1 of 40 (3%) squamous cell carcinomas (all samples with gain were primary tumors). The lower frequencies observed in patient tumors may reflect an under-calling of gains due to contaminating non-cancerous cells in tumor samples or alternatively to a bias in the tumors attempted or successfully established as cultures or to selective pressures on cultured cells. Gain of 14q13.3 was significantly associated with the presence of EGFR-activating mutations (P = 0.03, Mann–Whitney U-test) (but not KRAS or TP53 mutations), as well as the presence of specific DNA gains/losses elsewhere in the genome, including gain at 5p15.33 (TERT) (Supplementary Table 1). Notably, we have not observed the 14q13.3 locus to be amplified in other tumor types that we have profiled on the same platform, including cancers (totaling 385 specimens) of the breast, prostate, colon and pancreas (Bashyam et al., 2005; Bergamaschi et al., 2006; Lapointe et al., 2007; unpublished data), suggesting that the putative driver oncogene(s) within this locus is lung cancer specific.
The smallest region of recurrent amplification, corroborated by CGH on a custom high-definition oligonucleotide microarray with probes spanning 14q13.2–q13.3 at 300 bp intervals (Figure 1c), included just eight named genes: NFKBIA, INSM2, GARNL1, BRMS1L, MBIP, TITF1, NKX2-8 and PAX9. Because TITF1 (thyroid transcription factor 1; also called TTF-1 and NKX2-1) was known to participate in normal lung development (Kimura et al., 1996) and had been characterized as a histological marker of lung adenocarcinoma (Travis et al., 2004), we sought to explore a possible functional connection of TITF1 gene amplification with lung cancer.
Consistent with an oncogenic role, TITF1 exhibited increased expression at both the RNA (P = 0.046, Mann–Whitney U-test) (Figure 2b) and protein level (Figure 2c) in NSCLC cell lines with amplification. Notably, while other genes at 14q13.3 also exhibited increased expression when amplified, TITF1 was the only well-measured gene that also exhibited significantly increased expression in tumorigenic compared to non-tumorigenic cell lines (Figure 2b and Supplementary Table 2). Sequencing of the TITF1 open reading frame (and splice sites) from four NSCLC cell lines (HCC1195, HCC1833, H2009 and H661) with amplification revealed no DNA mutations, indicating that amplification- driven overexpression of the wild-type gene product would be of relevance.
To assess the functional significance of TITF1 amplification and overexpression, we used RNA interference to target TITF1 knockdown in two lung cancer cell lines, HCC1833 and HCC1195, with TITF1 amplification validated by fluorescence in situ hybridization (Figure 2a). Transfection of a Dharmacon On- TARGETplus pool of four different short interfering RNAs (siRNAs), designed and chemically modified to minimize off-target effects (Birmingham et al., 2006; Jackson et al., 2006), led to decreased TITF1 protein (Figure 3a) and decreased cell proliferation compared to a negative control siRNA pool (Figure 3b). Transfection of individual siRNAs from the same pool showed comparable effects (data not shown). In contrast, siRNA transfection of a lung cancer cell line (H1155) without TITF1 amplification but with detectable expression did not diminish cell proliferation, indicating the functional importance of amplification-driven overexpression. Transfection of a cell line (H1299) without amplification or detectable expression also did not alter cell proliferation, supporting the specificity of TITF1 targeting (Figure 3b). Similar negative results were observed upon transfection of a non-lung cancer cell line (colorectal cancer line SW48; data not shown). In the lung cancer cell lines with 14q13.3 amplification, the TITF1-targeted reduction in cell proliferation was attributable to both decreased cell-cycle progression (as evidenced by decreased S-phase fraction with G1 block; Figure 3c) and increased apoptosis (Figure 3d). The effects were more pronounced in HCC1833 compared to HCC1195.
TITF1 is a tissue-specific transcription factor required for branching morphogenesis during normal lung development, as well as for the differentiation of pulmonary epithelial cells, as marked by the expression of surfactant proteins (which are among its transcriptional targets) (Bohinski et al., 1994; Kimura et al., 1996; Minoo et al., 1999). In the developing and adult lung, TITF1 is expressed mainly in peripheral airway cells and small-sized bronchioles (Yatabe et al., 2002). TITF1 has also been found to be expressed in approximately 40–50% of NSCLCs, more frequently in adenocarcinomas compared to squamous cell carcinomas and with expression linked to more favorable prognosis in some but not all studies (Puglisi et al., 1999; Tan et al., 2003). Immunostaining of TITF1 is used as the major lineage-specific marker to distinguish primary lung adenocarcinoma from metastatic adenocarcinoma to the lung (Travis et al., 2004). Our findings indicate that TITF1 amplification and resultant overexpression contribute to increased cell proliferation rates and survival in lung cancer cells, and now implicate TITF1 as a lung cancer-specific oncogene.
Recently, Tanaka et al. (2007) reported that TITF1 knockdown led to decreased colony formation, which they attributed to increased apoptosis in a cell line (NCIH358) with TITF1 expression but (in their hands) no amplification. In our study, we showed that TITF1 amplification led to protein overexpression and sensitivity to siRNA-mediated knockdown, highlighting the role of TITF1 amplification as a critical event in the pathogenesis of a subset of lung cancers. Tanaka et al. also reported increased gene dosage of TITF1, measured by Southern blot and TaqMan PCR, in 2% of primary lung adenocarcinomas. The higher proportion we observed (11%) may reflect a higher sensitivity of aCGH (where multiple probes per locus are considered), or differences between the patient cohorts. Importantly, aCGH analysis also permitted us to define the 14q13 amplicon structure and boundaries, where our studies placed TITF1 squarely within the amplicon core, consistent with a ‘driver’ role.
Given its connection to pulmonary epithelium differentiation, an oncogenic role of TITF1 might seem counterintuitive. However, other tissue-specific transcription factors have been found to be amplified in cancers, including MITF in melanoma (Garraway et al., 2005), AR in prostate cancer (Visakorpi et al., 1995) and ESR1 in breast cancer (Holst et al., 2007). The deregulated expression of such transcription factors, with normal roles in lineage proliferation or survival, may be required for tumor survival and progression in some cellular and genetic contexts, reflecting a state of ‘lineage dependency’ (Garraway and Sellers, 2006). More generally, the deregulated expression of transcription factors with roles in normal development reflects the principle of ‘oncology recapitulates ontogeny’ (Lechner et al., 2001).
Our finding that 14q13.3 amplification occurs mainly in lung adenocarcinomas (and their derived cell lines), the same histology in which TITF1 expression (even when not amplified) is predominantly restricted, is consistent with TITF1 being the primary driver oncogene within 14q13.3. Nonetheless, other genes within the 14q13 amplicon may contribute to tumorigenesis. Notably, also residing within the amplicon core are two other homeodomain-containing genes, NKX2-8 and PAX9; the former of which has been recently implicated in the control of normal lung development (Tian et al., 2006). Our preliminary data (not shown) indicate that PAX9 is overexpressed when amplified in some NSCLC cell lines, and exhibits positive immunostaining in a subset of lung tumors. In another context, TITF1 and the paired-box member PAX8 have been shown to cooperatively activate the expression of thyroid-specific target genes (Miccadei et al., 2002). It is tempting to speculate that co-amplification of NKX2-8 and/or PAX9, together with TITF1, may contribute to lung cancer development or progression. Indeed, very recently Kendall et al. (2007), who discovered the same 14q13.3 amplicon, reported that both NKX2-8 and PAX9 can synergize with TITF1 to promote proliferation of immortalized human lung epithelial cells.
Future investigations are required to more precisely define the transcriptional effectors and pathways through which TITF1 mediates its oncogenic function. Nonetheless, our genomic profiling and functional studies implicate TITF1 as a lineage-specific oncogene in lung cancer, a discovery that may lead to new opportunities for therapeutic intervention.
We thank the SFGF for microarray manufacture, the SMD for database support, Ilana Galperin (Stanford Cytogenetics Laboratory) for assistance with FISH analysis and Eon Rios for assistance with FACS analysis. We also thank the members of the Pollack lab for helpful discussions. This work was supported in part by grants from the NIH: R01 CA97139 (JRP), T32 CA09151 (KAK), SPORE P50CA70907 (JDM), the DOD (JDM), the Vital, Anderson and Longenbaugh Foundations (JDM) and the Deutsche Krebshilfe: 10-2210-Pe4 (IP).