Introduction

Lung cancer is one of the most prevalently diagnosed cancer types with high mortality worldwide1. From pathological standpoints, lung cancer can be roughly classified into two major subtypes: non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC). Featured as the most common type, NSCLC that accounts for 80% of all cases can be further sub-divided into adenocarcinoma (ADC, ∼48%), squamous cell carcinoma (SCC, ∼28%) and large cell carcinoma (LCC, ∼24%)2,3. Despite the great advances achieved in recent targeted therapies, lung cancer still holds a quite poor prognosis and its 5 year survival rate remains as low as 10%-15%4.

Hippo (Hpo) pathway is a newly discovered signaling cascade, which is initially considered as a key regulator of tissue growth and controls organ size via regulating the YAP-TEAD complex5,6,7. In recent years, accumulating evidence has suggested that Hpo pathway also plays an important role in cancer development. Inactivation of upstream members in Hpo pathway either by mutations or epigenetic silencing have been reported in human epithelial cancers8. Moreover, YAP, the essential downstream effector of Hpo pathway, is frequently highly expressed in a wide spectrum of human solid tumors and significantly associated with poor clinical outcomes9,10,11,12. Functional studies show that ectopic YAP expression is sufficient to drive cell proliferation, transformation, invasion and epithelial-mesenchymal transition (EMT)13. Activation of YAP can drive the expansion of epidermal stem cell compartment as well as the formation of SCC-like tumors in mouse skin14. Liver-specific activation of YAP either by transgene overexpression or Mst1/2 deletion results in a dramatic increase in liver size and the development of liver tumor15,16,17.

Deregulation of Hpo pathway leads to the stabilization and nuclear sequestration of YAP. Nuclear YAP binds to and activates the transcription factors TEADs (the TEAD/TEF family transcription factors) and eventually turns on the expression of downstream genes including CTGF and CYR615. The mammalian genome contains four highly homologous TEAD/TEF family members (TEAD1-4), which are expressed in diverse tissues from pre-implantation embryos to adult tissues18. Previous studies have shown that TEADs are also involved in human cancers. High TEAD1 expression correlates with poor clinical outcome in prostate cancer19. Amplification and high expression of TEAD4 have been found in serous fallopian tube carcinoma, basal breast cancer and testicular germ cell tumors20,21,22. Experimentally, knockdown of TEAD1 reduced cell proliferation and impaired acinar formation ability19. TEAD4 alone also promoted anchorage-independent growth in MCF10A cells23. These observations indicate that TEADs may also play an important role in human cancers.

Vestigial-like (VGLL) proteins have recently emerged as a new group of TEAD-interacting partners participating in tumorigenesis. VGLL proteins are transcriptional cofactors that are named after Drosophila transcriptional co-activator Vestigial (Vg), the master regulator of wing development. There are 4 VGLL proteins in mammals, named VGLL1-4. These proteins bear no sequence similarity except for the TDU domain (the TEAD-interacting domain)24,25,26,27,28. Previous studies showed that VGLL1 promotes cell proliferation and exhibits high expression in basal-like breast cancer29,30. Similarly, VGLL3 is amplified in soft tissue sarcoma and inhibition of VGLL3 results in decreased cell proliferation and migration31. Different from other members in VGLL family, VGLL4 contains an extra TDU domain and is considered to be functionally different. For example, VGLL4 can promote apoptosis via negatively regulating inhibitor of apoptosis proteins (IAPs)32. However, the exact role of VGLL4 in cancers, especially in lung cancer, and whether and how VGLL4 is involved in the Hpo pathway are not clear.

Here, we found that VGLL4 is consistently downregulated in both murine and human lung ADC specimens. Our data further proved that VGLL4 functions as a suppressor of lung cancer growth and progression via direct competition with YAP in forming the complex with TEADs through two TDU domains.

Results

VGLL4 is downregulated in both mouse and human lung ADC specimens

To study the potential role of VGLL4 in lung cancer, we first examined the expression level of VGLL4 in KrasG12D-based lung cancer mouse model. Through quantitative RT-PCR analysis, we found that murine lung ADCs derived from KrasG12D model expressed lower level of Vgll4 mRNA relative to mouse normal lungs (Figure 1A). Through immunohistochemistry (IHC) study, we further showed that VGLL4 protein levels decreased in mouse lung ADCs (Figure 1B). Moreover, VGLL4 displayed a more diffused cytoplasmic staining in lung ADCs, but a predominant nuclear staining was seen in normal lungs (Figure 1B).

Figure 1
figure 1

VGLL4 is lowly expressed in both human and mouse lung adenocarcinomas. (A) Real-time PCR quantification of Vgll4 mRNA levels in mouse KrasG12D lung adenocarcinoma and normal lungs. Data are shown as means ± SEM. *P < 0.05. (B) H&E and Vgll4 immunohistochemical staining on mouse KrasG12D lung adenocarcinoma and normal lungs. Scale bar, 50 μm. (C) Quantification of VGLL4 mRNA levels in human lung adenocarcinoma and paired pathologically normal lungs by real-time PCR. The values were presented as log10 ratio of the VGLL4 expression of human lung adenocarcinomas vs normal lung specimens. (D) Immunohistochemical staining of VGLL4 on human lung adenocarcinoma and normal lungs. Scale bar, 500 μm (top) and 50 μm (bottom). (E) Statistical analysis of nuclear VGLL4 staining in human lung adenocarcinoma and normal lung specimens. ADC, adenocarcinoma; NL, normal lung.

Full size image

We further examined the expression status of VGLL4 in human lung ADC specimens. Interestingly, we found that almost all the human lung ADC samples (29 of 30) had a relatively lower VGLL4 expression in comparison with paired pathologically normal lungs (Figure 1C). IHC studies in a cohort containing 27 normal lungs and 77 lung ADCs showed that 92.6% of patients (25 out of 27) exhibited high nuclear VGLL4 expression in their normal lungs, whereas only 22.1% of patients (17 out of 77) had high nuclear expression of VGLL4 in their lung ADCs. The pattern of VGLL4 nuclear expression was statistically different between normal lungs and lung ADCs (P < 0.01) (Figure 1D and 1E). Taken together, these results indicated that VGLL4 is consistently expressed at lower level in lung ADC compared with normal lung tissues.

VGLL4 suppresses lung tumor cell growth in vitro

We next investigated the potential biological function of VGLL4 in lung cancer. We first examined the expression of VGLL4 in a series of lung cancer cell lines and found that A549 and CRL-5872 had relatively low but detectable level of VGLL4 expression (Supplementary information, Figure S1). We then overexpressed VGLL4 in these two cell lines. Our data showed that ectopic VGLL4 expression significantly inhibited the proliferation of these two cell lines (Figure 2A and 2B and Supplementary information, Figure S2). Cell cycle analysis showed that both A549 and CRL-5872 cells with VGLL4 overexpression displayed a reduced cell ratio in G2/M phase (Figure 2C and 2D). We further showed that VGLL4 overexpression inhibited both the anchorage-independent cell growth in soft agar (Figure 2E and 2F) and colony proliferation in Matrigel culture of A549 and CRL-5872 cells (Figure 2G and 2H). These data clearly demonstrated that VGLL4 negatively regulates lung cancer cell proliferation and growth in vitro.

Figure 2
figure 2

VGLL4 inhibits lung cancer cell growth in vitro. (A and B) The proliferations of A549 (A) and CRL-5872 (B) cells with or without ectopic VGLL4 expression were detected by MTT assay. Data are shown as means ± SEM. ***P < 0.005. (C and D) Cell cycle distributions of A549 (C) and CRL-5872 (D) cells with or without VGLL4 overexpression were assessed by FACS assay. (E and F) Colony formation abilities of A549 (E) and CRL-5872 (F) cells with or without ectopic VGLL4 expression were assessed by soft agar assay. Data are shown as means ± SEM. **P < 0.01; ***P < 0.005. (G and H) A549 (G) and CRL-5872 (H) cells with or without VGLL4 overexpression were cultured in Matrigel for 7 days. Colony size was shown by photos. Scale bar, 0.01 mm.

Full size image

VGLL4 suppresses lung cancer progression in de novo mouse model

We next asked whether VGLL4 plays a tumor suppressive function in vivo. To address this, we employed a KrasG12D-based de novo lung cancer mouse model. We first examined the Vgll4 expression in lung tumors derived from either KrasG12D or p53L/L/KrasG12D or Lkb1L/L/KrasG12D mice. Our data showed that the lung tumors from Lkb1L/L/KrasG12D mice exhibited the lowest Vgll4 expression (Supplementary information, Figure S3). We then decided to use Lkb1L/L/KrasG12D mice to verify the in vivo function of VGLL4. We delivered the lentivirus carrying either Cre (Ctrl-Cre) or VGLL4-Cre to Lkb1L/L/KrasG12D mice through nasal inhalation as previously described33. We then analyzed the mice at 13 weeks post viral administration for gross inspection and pathological study (Figure 3A). We confirmed the VGLL4 expression in lung tumors via immunostainning (Figure 3B). Our data showed that the ectopic expression of VGLL4 did not cause a significant change in the total tumor number and tumor area despite a detectable trend that mice infected with VGLL4-Cre harbored less tumor area (Supplementary information, Figure S4). However, detailed pathological analysis revealed that the percentage of large tumors (≥ 0.5 mm2) was significantly lower in VGLL4-Cre group (Figure 3C and 3D). Consistently, a lower proliferation rate indicated by Ki-67-positive staining was observed in lung tumors with VGLL4 overexpression (Figure 3E and 3F). Moreover, the percentage of high-grade tumors was slightly decreased in VGLL4-Cre group in comparison with control group (Supplementary information, Figure S5). Taken together, these data suggested that VGLL4 may function as a negative regulator of lung tumor progression in vivo.

Figure 3
figure 3

VGLL4 significantly suppresses lung cancer progression in de novo mouse model. (A) Scheme of viral infection strategy for ectopic VGLL4 expression in Lkb1 L/L/KrasG12D mouse model. (B) VGLL4 immunohistochemical staining in lung tumors of Lkb1L/L/KrasG12D mouse at 13 weeks post viral infection with Lenti-Cre or Lenti-VGLL4-Cre. Scale bar, 50 μm. (C) Representative H&E photos of Lkb1L/L/KrasG12D mice lung at 13 weeks post viral infection with Lenti-Ctrl-Cre or Lenti-VGLL4-Cre. Scale bar, 500 μm (top) and 50 μm (bottom). (D) Statistical analysis of lung tumors with the indicated size from Lkb1L/L/KrasG12D mice virally infected with Lenti-Ctrl-Cre or Lenti-VGLL4-Cre. Data are shown as means ± SEM. *P < 0.05. (E) Ki-67 immunohistochemical staining of lung sections from Lkb1L/L/KrasG12D mice virally infected with Lenti-Ctrl-Cre or Lenti-VGLL4-Cre. Scale bar, 50 μm. (F) Quantitative analyses of proliferative index in lung tumors from Lkb1L/L/KrasG12D mice virally infected with Lenti-Ctrl-Cre or Lenti-VGLL4-Cre. Data are shown as means ± SEM. *P< 0.05.

Full size image

TEADs mediate VGLL4's suppression on lung tumor cell growth

Previous study showed that VGLL4 physically interacts with TEAD124. As mentioned above, TEAD family contains four highly homologous proteins TEAD1-4. Among them, TEAD4 is highly expressed in the lung. We therefore examined the potential interaction of VGLL4 and TEAD4. Our data showed that VGLL4 and TEAD4 colocalized in the nucleus and had strong binding with each other (Supplementary information, Figure S6). We then asked whether TEADs are important for VGLL4's suppressive role in lung cancer cell growth. We found that VGLL4 no longer exhibited a suppressive function when TEADs were inactivated through TEAD1/3/4 RNAi (Supplementary information, Figure S7). Previous studies using mammalian two-hybrid system have indicated that deletion of both TDU domains in VGLL4 prevents its interaction with TEAD124. To confirm the importance of these TDU domains in VGLL4-TEAD4 interaction, we performed co-immunoprecipitation experiments and found that deletion of both TDU domains abolished the interaction between VGLL4 and TEAD4 (Figure 4A). To further prove our hypothesis, we used a VGLL4 mutant form (VGLL4 with the deletion of both TDU domains) to see whether VGLL4-TEAD interaction is required for VGLL4's inhibitory role in tumor cell growth. Our data revealed that deletion of both TDU domains significantly attenuated the inhibitory function of VGLL4 in A549 cell growth in 2D culture and colony formation in soft agar (Figure 4B and 4D and Supplementary information, Figure S8A). Similar results were obtained in another lung cancer cell line CRL-5872 (Figure 4C and 4E and Supplementary information, Figure S8B). Thus, these findings suggested that TEADs mediates the inhibitory function of VGLL4 in lung cancer cell growth.

Figure 4
figure 4

TEADs mediate VGLL4's inhibitory role in lung tumor cell growth. (A) Co-immunoprecipitation experiments of TEAD4 with wild-type (WT) VGLL4 or VGLL4 mutant with deletion of two TDU domains in HEK-293T cells. TEAD4 is HA-tagged and VGLL4 is FLAG-tagged. (B and C) Deletion of two TDU domains abolishes the inhibitory role of VGLL4 in the cell proliferation of human lung cancer cell lines A549 (B) and CRL-5872 (C). Data are shown as means ± SEM. **P < 0.005. (D and E) Deletion of two TDU domains abolishes the inhibitory role of VGLL4 in the colony formation of human lung cancer cell lines A549 (D) and CRL-5872 (E) in soft agar. *P< 0.05; **P< 0.01.

Full size image

VGLL4 inhibits the transcriptional activity of TEADs

Increased TEAD expression has been observed in human cancers27. Also, TEADs have been reported to facilitate the growth of certain cell types19,23. As transcriptional factors, TEADs regulate a number of growth-promoting genes. In consideration of this, we asked whether VGLL4 modulates the transcriptional activity of TEADs. We took advantage of an established TEAD4-responsive element-driven luciferase reporter system34 and found that ectopic VGLL4 expression significantly reduced the TEAD4-dependent luciferase activity (Figure 5A). In contrast, the VGLL4 mutant without two TDU domains obviously lost its inhibitory function. We further examined whether VGLL4 altered the TEAD transcription downstream genes13,35. Our data showed that ectopic expression of VGLL4, but not VGLL4 lacking two TDU domains decreased the mRNA levels of two well-known downstream target genes CTGF and CYR61 in HEK-293T and A549 cells, respectively (Figure 5B and 5C). Consistently, VGLL4 knockdown significantly increased TEADs' transcriptional activity in HEK-293T (Figure 5D and Supplementary information, Figure S9) as well as the CTGF and CYR61 mRNA levels in human bronchial epithelial cells (Beas2B) and CRL-5807 cells (Figure 5E and 5F and Supplementary information, Figure S10). Together, these data identified VGLL4 as a negative regulator of TEADs' transcriptional activity.

Figure 5
figure 5

VGLL4 inhibits the transcriptional activity of TEADs and the expression of their downstream target genes. (A) Ectopic expression of wild-type (WT) VGLL4 but not VGLL4 mutant with deletion of two TDU domains decreases TEAD4-dependent luciferase reporter activity in HEK-293T cells. Data are shown as means ± SEM. ***P< 0.005. (B and C) Ectopic expression of WT VGLL4 but not VGLL4 mutant with deletion of two TDU domains significantly downregulates the expression of TEAD downstream target genes (CTGF and CYR61) in HEK-293T cells (B) and A549 cells (C). Data are shown as means ± SEM. *P< 0.05; **P < 0.01; ***P< 0.005. (D) Knockdown of VGLL4 increases TEAD4-dependent luciferase reporter activity in HEK-293T cells. Data are shown as means ± SEM. **P < 0.01; ***P < 0.005. (E and F) Knockdown of VGLL4 upregulates the transcripts of CTGF and CYR61 in Beas2B (E) and CRL-5807 (F) cells. Data are shown as means ± SEM. *P < 0.05; **P < 0.01; ***P < 0.005.

Full size image

VGLL4 inhibits the function of YAP-TEAD transcri-ptional complex through direct competition

TEADs require transcriptional cofactors to fully activate the downstream gene transcription. YAP has been well documented as a major transcriptional co-activator of TEADs through direct binding to TEADs and facilitates downstream target gene expression13. We therefore asked whether VGLL4 suppresses TEADs' transcriptional activity by attenuating YAP-induced activation of TEADs. Through luciferase reporter assay, we found that overexpression of YAP enhanced the activity of TEAD4-responsive luciferase reporter, whereas VGLL4 and YAP co-expression led to a dramatic decrease of TEAD4-responsive luciferase reporter activity (Figure 6A). Similar results were observed using another luciferase reporter driven by the promoter of CTGF (Supplementary information, Figure S11). We further analyzed TEAD downstream target gene expression by quantitative RT-PCR. Overexpression of YAP activated the transcription of CTGF and CYR61. However, the mRNA levels of these two genes were downregulated sharply when YAP and VGLL4 were co-expressed in HEK-293T and A549 cells (Figure 6B and 6C).

Figure 6
figure 6

VGLL4 inhibits the growth-promotive function of YAP through the direct competition for its binding to TEAD4. (A) Ectopic VGLL4 expression decreases the TEAD4-dependent luciferase reporter activity promoted by YAP overexpression in HEK-293T cells. Data are shown as means ± SEM. ***P < 0.005. (B and C) Ectopic VGLL4 expression decreases YAP-induced CTGF and CYR61 expression in HEK-293T cells (B) and A549 cells (C). Data are shown as means ± SEM. **P < 0.01; ***P < 0.005. (D) VGLL4 directly competes with YAP for binding to TEAD4 in vitro. YAP, amino acids 50-504 containing TEAD-binding domain. TEAD4, amino acids 217-434; VGLL4, amino acids 206-259 containing two TDU domains. (E) Ectopic VGLL4 expression significantly suppresses the growth-promotive function of YAP in A549 cells. Data are shown as means ± SEM. ***P < 0.005. (F) Ectopic expression of VGLL4 significantly inhibits the colony formation of A549 cells promoted by YAP overexpression. *P < 0.05.

Full size image

As VGLL4 inhibited TEADs' transcriptional activity enhanced by YAP, we then investigated the underlying mechanism. As our data showed, unlike currently known mechanisms, ectopic expression of VGLL4 neither decreased the YAP protein level nor led to YAP cytoplasmic sequestration (Supplementary information, Figure S12). Because of the predominant nuclear location of VGLL4, we wondered whether it is possible that VGLL4 competes with YAP for binding to TEADs. To answer this, MBP-fused YAP coupled on amylose resin was mixed with TEAD4 in the presence of different amounts of VGLL4. The dosage-dependent pull-down assay clearly showed a decreasing binding of TEAD4 to YAP when increasing amounts of VGLL4 protein were added (Figure 6D). Therefore, our observations supported the notion that VGLL4 suppresses TEADs' activity through competing with YAP for TEAD binding. YAP had been reported to promote cell growth in several types of tumor cells10,12,36. Therefore, we then investigated whether VGLL4 can abrogate YAP's promotive effect on tumor cell growth. We found that ectopic expression of YAP facilitated A549 cell proliferation and enhanced its colony formation ability, whereas VGLL4 co-expression abolished this effect (Figure 6E and 6F and Supplementary information, Figure S13). Thus, the above data indicated that VGLL4 can diminish YAP-enhanced TEADs' transcriptional activity and YAP-promoted tumor cell growth by impairing YAP-TEAD interaction.

Discussion

In this study, we have identified VGLL4 as a potential tumor suppressor in lung cancer via negatively regulating YAP-TEAD transcriptional complex activity. Through studies of murine and human lung cancer specimens, we find that VGLL4 is significantly decreased in lung cancer. Taking advantage of lung cancer cells culture system and de novo lung cancer mouse model, we clearly showed that VGLL4 exhibits a strong inhibitory role in lung cancer growth and progression. Moreover, our data show that VGLL4 acts as a novel negative regulator of YAP-TEAD transcriptional complex through direct competition with YAP for binding to TEADs. Thus, these data have provided a novel mechanism in lung carcinogenesis through VGLL4 involving the regulation of YAP-TEAD transcription activation and Hpo pathway.

The biological function of VGLL4 was initially characterized in heart. Ectopic expression of VGLL4 in cardiac myocytes suppresses a TEF-1-dependent skeletal α-actin promoter and interfered with its α1-adenergic induction24. VGLL4 is also shown to contribute to apoptosis signaling via forcing the nuclear relocation of IAPs32. VGLL4 was also identified as one of the mutated genes in pancreatic ADC through mutagenic screen37. However, the potential contribution of VGLL4 to cancer development, especially lung cancer, remains to be elucidated. Our study here provides strong evidence that VGLL4 is downregulated in a large proportion of lung cancer specimens and plays a suppressive role in lung cancer. Moreover, our data show that ectopic VGLL4 expression also significantly impairs the growth of other types of cancer cells including liver cancer cell line HepG2 as well as cervical cancer cell line Hela (Supplementary information, Figure S14). Coincidently, a recent study has convincingly shown that VGLL4 plays an inhibitory role in hepatocellular carcinoma (HCC) formation38. Collectively, these data together suggest that VGLL4 may act as a common tumor suppressor in a variety of human cancers including lung cancer and liver cancer.

VGLL4 is the only member of VGLL family that carries two TDU domains. An interesting question remains here is that which TDU domain is more important for VGLL4's suppression on lung cancer cell growth. Deletion of the second TDU domain (aa 242-252) completely abolishes the inhibitory role of VGLL4 in lung cancer cell growth while deletion of the first one (aa 214-224) does not affect VGLL4's function, indicating that the second TDU domain is more important for VGLL4's inhibitory function (Supplementary information, Figure S15). VGLL1 is another VGLL family member that has only one TDU domain. Previous studies and our own data showed that VGLL1 plays a promotive role in tumor cell growth (Supplementary information, Figure S16). Therefore, it will be interesting to further investigate how VGLL4 functions differently from other family members in relation to the TDU domain.

Studies over past decade have uncovered the vital role of Hpo pathway in organ size control and tumorigenesis8. Deregulation of Hpo pathway eventually triggers the transcriptional activation of the YAP-TEAD complex and initiates downstream target gene expression39. Both TEADs and YAP have been reported to be overexpressed or hyper-activated in various human cancers. Previous studies have demonstrated YAP as a growth promoter in both in vitro cell culture system and in vivo mouse models40. Recent works have also shown a critical role of YAP in stem cell self-renewal and tissue regeneration41. Multiple mechanisms have been proposed previously for the regulation of YAP activity. Canonical Hpo kinase cascade phosphorylates YAP at Ser 127 and leads to YAP cytoplasmic sequestration and inactivation42. α-catenin acts as an upstream negative regulator of YAP through modulating its interaction with 14-3-3 and PP2A phosphatase14. The activity of YAP can also be modulated by protein-protein interaction and cytosol retention. For example, angiomotin (AMOT) family proteins bind to YAP and restrain YAP to tight junction and thus inhibit its nuclear translocation34. Moreover, Mask is proved to form complex with YAP in nucleus and is required for the full activity of YAP43,44. However, there is no endogenous negative regulator of YAP activity in the nucleus reported up to date. We here convincingly show that VGLL4, predominantly located in nucleus, inhibits YAP-driven transcriptional activity and growth advance through the competition for YAP-TEAD complex formation. Our data further showed that the two TDU domains of VGLL4 are important for this competition. Interestingly, similar findings are also observed in Drosophila. Sd Binding Protein (SdBP)/Tgi, the Drosophila homologue of VGLL4, interacts with Sd through its TDU domains and competes with Yki in nucleus to restrict tissue growth45. It is worth noting that Tgi is slightly different from VGLL4 with three extra PPxY motifs, which are essential for Tgi-Yki binding. Despite the functional similarity between VGLL4 and Tgi, ectopic Tgi expression results in the formation of an inactivated Yki-Tgi-Sd complex due to the extra PPxY motifs. In contrast, VGLL4 cannot form the complex with YAP-TEAD (data not shown). Interestingly, deletion of the PPxY motifs results in the loss of ability of Tgi in forming the Yki-Tgi-Sd ternary complex but does not affect the competition with Yki for binding to Sd45. These data indicates that the competition of VGLL4 with YAP for TEADs binding is functionally and evolutionally conserved despite the different manner. Recently, Koontz et al.38 have also demonstrated that YAP can compete with VGLL4 for binding to TEAD2. Taken together, these data show that VGLL4 and YAP compete with each other for TEAD binding and modulate downstream gene transcription and functional outcomes. As YAP acts as an oncogene in multiple human cancers and plays a crucial role in stem cell homeostasis and tissue regeneration, manipulating the interaction among YAP, TEADs and VGLL4 may provide a new entry point for human cancer treatment and regenerative medicine.

Materials and Methods

Human lung cancer specimen collection

All the human lung cancer and normal lung specimens were collected in Fudan University Shanghai Cancer Center from January 2008 to December 2009 with written consents of patients and the approval from the Institute Research Ethics Committee. A total of 30 human lung ADC samples with paired pathologically normal lungs were used for real-time PCR analysis and another 77 lung ADCs and 27 pathologically normal lungs were used for IHC analysis.

Mouse cohorts, treatment and histopathological analysis

KrasG12D, Lkb1L/L and p53L/L mice were generously provided by Drs T Jacks and R Depinho. Lung cancer mouse models with KrasG12D, p53L/L/KrasG12D, Lkb1L/L/KrasG12D mice were generated as previously described46. All mice were housed in a specific pathogen-free environment at the Shanghai Institute of Biochemistry and Cell Biology and treated in strict accordance with protocols approved by the Institutional Animal Care and Use Committee of the Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences.

The KrasG12D, p53L/L/KrasG12D, Lkb1L/L/KrasG12D mice at 6-8 weeks of age were treated with either 2 × 106 PFU of Adeno-Cre virus or lentivirus (Lenti-Cre or VGLL4-CRE) via nasal inhalation as described before47. The lentiviruses were prepared as previously described33. All mice were sacrificed for gross inspection and histopathological examination. Lung tumors were dissected for molecular analysis. Histopathological analysis was conducted as previously described. Briefly, mice lung tissues were inflated and fixed in 4% formalin, embedded in paraffin and sectioned for hematoxylin and eosin (H&E) staining. Tumor number was counted under microscope and tumor size was analyzed using Image-J software.

Cell culture, transfection and lentiviral infection

Human lung ADC cell lines A549, CRL-5807, CRL-5872 were cultured in RPMI 1640 supplemented with 10% FBS. HEK-293T, Beas2B, Hela and HepG2 were cultured in DMEM containing 10% FBS. Lentiviral infection was done as follows: HEK-293T cells were co-transfected with pLKO.1 or pCDH constructs and packaging plasmids. The progeny viruses released from HEK-293T cells were filtered, collected and used to infect A549, CRL-5807, CRL-5872, Beas2B, Hela and HepG2 cells.

Cellular functional assay

Virus-infected cells were seeded in 96-well plate and the viability of cells was measured by MTT assay daily for 5 days. Briefly, 20 μl of MTT working solution (5 mg/ml) was added into each well and incubated at 37 °C for 4 h. Then the supernatants were removed and the resultant MTT formazan was dissolved in 100 μl of DMSO. The absorbance was measured at the wavelengths of 570 nm and 630 nm.

For cell cycle analysis, virus-infected cells with 80% confluence were harvested and fixed with 75% ethanol. Then cells were taken for PI staining and cell cycle was analyzed using flow cytometry.

For soft agar colony formation assay, virus-infected cells were added to growth medium with 0.2% agar and layered onto 1% agar beds in six-well plates. Cells were fed with 1 ml of medium every three days. The colonies were stained with 0.005% crystal violet and counted in 2-3 weeks.

For 3D cell culture, virus-infected cells were seeded in medium containing 2% Matrigel (BD Biosciences) on the top of another layer of solidified Matrigel. Cells were then cultured at 37 °C for 1 week. Photos were taken using a light microscope (Leica).

Luciferase reporter assay

HEK-293T cells were seeded in 24-well plates. Luciferase reporter and the indicated plasmids were co-transfected. Luciferase activities were measured 48 h after transfection using Dual-Luciferase Assay kit (Promega) on GloMax 20/20 luminometer (Promega) following the manufacturer's instructions. pRL-TK was co-transfected as internal control. Experiments were done in triplicates.

RT-PCR and quantitative real-time PCR

Total RNA was isolated using Trizol reagent (Invitrogen) and retro-transcribed into first-strand cDNA using RevertAidâ„¢ First Strand cDNA Synthesis Kit (Fermentas). cDNAs were subjected to quantitative real-time PCR with gene-specific primers on 7500 Fast Real-Time PCR System (Applied Biosystems) using SYBR-Green Master PCR mix (Invitrogen). GAPDH (human) or actin (mouse) served as internal control.

Primers used were:

Human VGLL4: 5′-AACTGCAACCTCTCGCACTG-3′ (forward); 5′-GCTCGGGCTCCTTGTAATTCT-3′ (reverse).

Human YAP: 5′-CACAGCATGTTCGAGCTCAT-3′ (forward); 5′-GATGCTGAGCTGTGGGTGTA-3′ (reverse).

Human CTGF: 5′-CCTGGTCCAGACCACAGAGT-3′ (forward); 5′-TGGAGATTTTGGGAGTACGG-3′ (reverse).

Human CYR61: 5′-ACCGCTCTGAAGGGGATCT-3′ (forward); 5′-ACTGATGTTTACAGTTGGGCTG-3′ (reverse).

Human GAPDH: 5′-CAGGTGGTCTCCTCTGACTT-3′ (forward); 5′-CCAAATTCGTTGTCATACCA-3′ (reverse).

Human TEAD1: 5′-GGCCGGGAATGATTCAAACAG-3′ (forward); 5′-CAATGGAGCGACCTTGCCA-3′ (reverse).

Human TEAD3: 5′-TCATCCTGTCAGACGAGGG-3′ (forward); 5′-TCTTCCGAGCTAGAACCTGTAT-3′ (reverse).

Human TEAD4: 5′-GAACGGGGACCCTCCAATG-3′ (forward); 5′-GCGAGCATACTCTGTCTCAAC-3′ (reverse).

Mouse Actin: 5′-TGAGCGCAAGTACTCTGTGTGGAT-3′ (forward); 5′-ACTCATCGTACTCCTGCTTGCTGA-3′ (reverse).

Mouse Vgll4: 5′-ATGAACAACAATATCGGCGTTCT-3′ (forward); 5′-GGGCTCCATGCTGAATTTCC-3′ (reverse).

IHC study

IHC was performed as previously described33. The proliferation rate was evaluated by counting Ki-67-positive nuclear staining at high-power field (HPF) for more than 30 fields for each group. The VGLL4 immunostainning was reviewed and scored blindly. Briefly, the scoring system used in grading the level of nuclear VGLL4 expression was as follows: VGLL4Low, nuclear expression is low or nuclear expression is less than 50% of VGLL4-positive staining; VGLL4High, nuclear expression is more than 50% of VGLL4-positive staining.

Western blot and immunofluorescence

Western blot was conducted as described before48. For immunofluorescence, cells cultured on cover slips were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100 in 1 × PBS. Cells were blocked with 3% bovine serum albumin (BSA) and incubated with primary antibody overnight at 4 °C. Then cells were washed with immunofluorescence buffer and incubated with Alexa Fluor 555 or 488 conjugated secondary antibodies. Nuclei were counterstained with Hoechst. After washing with immunofluorescence buffer, cells were mounted and visualized using a laser scanning confocal microscope.

Co-immunoprecipitation

HEK-293T cells were transfected with the indicated plasmids using Lipofectamine according to the manufacturer's instructions. Cells were lysed and the supernatants were incubated with anti-HA, anti-FLAG antibodies and protein A/G agarose (Santa Cruz). The immune complexes were subjected to SDS-PAGE, and analyzed by western blot.

Pull-down assay

MBP-fused YAP (aa 50-504 containing TEAD-binding domain) coupled on amylose resin were mixed with 10 μg TEAD4 (aa 217-434) in the presence of gradually increased amounts of VGLL4 (aa 206-259 containing two TDU domains) from 0 to 40 μg at 4 °C for 1 h in 20 mM HEPES (pH 7.5), 300 mM NaCl, 5% glycerol, and then washed three times. The proteins bound on the resin were eluted by the same buffer with 20 mM maltose. The input and output samples were loaded on 15% SDS-PAGE followed by Coomassie Blue staining.

Antibodies and plasmids

The following antibodies were used: VGLL4, Actin, HA, FLAG (all from Sigma-Aldrich); Ki-67 (Epitomics Inc.); YAP (Santa Cruz); HRP-conjugated rabbit and mouse secondary antibodies (Santa Cruz); Alexa Fluor 555 or 488 conjugated anti-mouse or anti-rabbit secondary antibodies (Invitrogen).

The following plasmids were used: pCI-HA-TEAD4 was generously provided by Dr Zengqiang Yuan (Institute of Biophysics, Chinese Academy of Sciences); TEAD1/3/4 shRNA construct, CTGF promoter and pGal4-TEAD4/5xUAS-Luc reporter system were kindly given by Dr Kunliang Guan (Fudan University).

To construct the human YAP1 expression plasmid, the coding sequence of human YAP1 was amplified with primers:

Forward: 5′-ACCTCTACGAATTCGCCACCATGGGATACCCCTACGACGTCCCCGACTACGCCGATCCCGGGCAGCAGCCGC-3′; reverse: 5′-ACCTCTACGGATCCCTATAACCATGTAAGAAAGCTTTCTTTATCTAGCTTGG-3′.

To construct the human VGLL4 expression plasmid, the coding sequence of human VGLL4 was amplified with primers:

Forward: 5′-CGGAATTCGCCACCATGCTATTTATGAAGATGGACC-3′; reverse: 5′-ATTTGCGGCCGCTCAGGAGACCACAGAGGGGGAG-3′.

To construct the human VGLL1 expression plasmid, the coding sequence of human VGLL1 was amplified with primers:

Forward: 5′-GCTCTAGAGCCACCATGGGATACCCCTACGACGTCCCCGACTACGCCATGGAAGAAATGAAGAAGACT-3′; reverse: 5′-CGGGATCCCTAAAGATGCTGCAGGTATCGATG-3′.

The DNA fragment was inserted into pCDH vector or pCDNA3.1 vector. The plasmids pCDH-Flag-VGLL4-ΔTDU1, pCDH-Flag-VGLL4-ΔTDU2, pCDH-Flag-VGLL4-ΔTDU1&2, pCDNA3.1-Flag-VGLL4-ΔTDU1&2 were constructed using QuikChange®XL Site-Directed Mutagenesis Kit (Agilent Technologies). To generate the human VGLL4 knockdown construct, the following oligonucleotides were cloned into pLKO.1 at the AgeI/EcoRI sites:

shVGLL4-1:

sense: 5′-CCGGGAGCCTGGGCAAGAATTACAACTCGAGTTGTAATTCTTGCCCAGGCTCTTTTTG-3′; antisense: 5′-AATTCAAAAAGAGCCTGGGCAAGAATTACAACTCGAGTTGTAATTCTTGCCCAGGCTC-3′.

shVGLL4-2

sense: 5′-CCGGCATCTGAACAAGACTGCCAATCTCGAGATTGGCAGTCTTGTTCAGATGTTTTTG-3′; antisense: 5′-AATTCAAAAACATCTGAACAAGACTGCCAATCTCGAGATTGGCAGTCTTGTTCAGATG-3′.

shVGLL4-3:

sense: 5′-CCGGCAGGAGCCTGGGCAAGAATTACTCGAGTAATTCTTGCCCAGGCTCCTGTTTTTG-3′; antisense: 5′-AATTCAAAAACAGGAGCCTGGGCAAGAATTACTCGAGTAATTCTTGCCCAGGCTCCTG-3′.

For pull-down assay, YAP (aa 50-504) was inserted into pET28a with MBP tag, and VGLL4 (aa 206-259) and TEAD4 (aa 217-434) were inserted into pET28a with His-SUMO tag.

Statistical analysis

All statistical analyses were carried out using the SPSS 13.0 statistical software package. Data were analyzed by Student's t-test (two tailed). P < 0.05 was considered to be significant. Error bars represent SEM.