The bone morphogenetic protein antagonist gremlin 1 is overexpressed in human cancers and interacts with YWHAH protein

Background Basic studies of oncogenesis have demonstrated that either the elevated production of particular oncogene proteins or the occurrence of qualitative abnormalities in oncogenes can contribute to neoplastic cellular transformation. The purpose of our study was to identify an unique gene that shows cancer-associated expression, and characterizes its function related to human carcinogenesis. Methods We used the differential display (DD) RT-PCR method using normal cervical, cervical cancer, metastatic cervical tissues, and cervical cancer cell lines to identify genes overexpressed in cervical cancers and identified gremlin 1 which was overexpressed in cervical cancers. We determined expression levels of gremlin 1 using Northern blot analysis and immunohistochemical study in various types of human normal and cancer tissues. To understand the tumorigenesis pathway of identified gremlin 1 protein, we performed a yeast two-hybrid screen, GST pull down assay, and immunoprecipitation to identify gremlin 1 interacting proteins. Results DDRT-PCR analysis revealed that gremlin 1 was overexpressed in uterine cervical cancer. We also identified a human gremlin 1 that was overexpressed in various human tumors including carcinomas of the lung, ovary, kidney, breast, colon, pancreas, and sarcoma. PIG-2-transfected HEK 293 cells exhibited growth stimulation and increased telomerase activity. Gremlin 1 interacted with homo sapiens tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein, eta polypeptide (14-3-3 eta; YWHAH). YWHAH protein binding site for gremlin 1 was located between residues 61–80 and gremlin 1 binding site for YWHAH was found to be located between residues 1 to 67. Conclusion Gremlin 1 may play an oncogenic role especially in carcinomas of the uterine cervix, lung, ovary, kidney, breast, colon, pancreas, and sarcoma. Over-expressed gremlin 1 functions by interaction with YWHAH. Therefore, Gremlin 1 and its binding protein YWHAH could be good targets for developing diagnostic and therapeutic strategies against human cancers.


Background
The identification of molecular alterations in cancerous and pre-cancerous cells has provided insight into the role of oncogenes and tumor suppressor genes in tumor initiation and progression [1]. Oncogenes are derived from highly conserved proto-oncogenes that are altered by chromosomal point mutations, gene amplifications, or gene rearrangements [2]. Structural alteration of protooncogenes leads to a quantitative or qualitative change in the expression of the corresponding protein product. The signal transduction pathways subverted by oncoproteins govern fundamental cell functions, including proliferation, cell cycle regulation, and apoptosis [3].
Although genetic characterization of tumor tissues demonstrates that mutation of the p53 gene is the most common genetic alteration in human cancers, the mutation ratio of the p53 gene in uterine cervical cancer is relatively low [4,5]. It suggests that there are other oncogenes involved in cervical carcinogenesis. We applied the DDRT-PCR method to discover genes involved in tumorigenesis of human cervical tissue, and identified the new human cervical cancer-related gene, proliferation-inducing gene 2 (PIG-2) (GenBank accession number AY232290), which exhibits close similarity to gremlin 1 cDNA (GenBank accession number NM_013372) in the database.
Gremlin gene encodes a member of the bone morphogenic protein (BMP) antagonist family. Like BMPs, BMP antagonists contain cystine knots and typically form homo-and heterodimers. The cerberus and dan subfamily of BMP antagonists, to which this gene belongs, is characterized by a C-terminal cystine knot with an eight-membered ring. The antagonistic effect of the secreted glycosylated protein encoded by this gene is likely due to its direct binding to BMP proteins. As an antagonist of BMP, this gene may play a role in regulating organogenesis, body patterning, and tissue differentiation. In mouse, this protein has been shown to relay the sonic hedgehog signal from the polarizing region to the apical ectodermal ridge during limb bud outgrowth [17].
The action of Drm and Dan on development and possibly diabetic nephropathy is mediated by heterodimerizing with certain BMPs [9], in particular BMP2, 4, and 7 [7,8,16,18] to subsequently block the ability of BMPs to bind their receptors [7,18,19]. Chen et al. have previously shown that the capacity of Drm to suppress transformation and tumorigenesis [6,14,15] is mediated by a mechanism that is independent of BMPs and involves both upregulation of p21 Cip1 and down-regulation of p42/44 MAPK [14], suggesting additional target(s) for Drm and other Dan family members.

Tissues and cell lines
For Differential display (DD) of mRNA, normal exocervical tissue specimen was obtained from uterine myoma patients during hysterectomy and untreated primary cervical cancer tissues and metastatic lymph node tissues were obtained during radical hysterectomy. Patient consent was obtained from each individual and the use of tissue samples was approved by the ethics committee of our institution. The cervical caner cell line used in DD was CasKi and CUMC-6 which was isolated in our laboratory and maintained as previously described [29]. Mammalian cell lines described below were all obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA): CasKi is a human cervical cancer cell line. NCI-H441, NCI-H157, and NCI-H2009 are human lung cancer cell lines.

DDRT-PCR
Total RNA was extracted from tissues and cells using RNA extraction kit (RNeasy total RNA kit; Qiagen Inc., Valencia, CA) and 0.2 µg of total RNA was used to generate cDNA in reverse transcription reaction (RNAimage™ kit, GenHunter, MA). With the use of the differential display kit (RNAimage™ kit), we performed PCR using oligo-dT primers and arbitrary sequences, each 13 bases in length according to the manufacturer's recommendations [30]. After cDNAs of 3' termini of mRNAs were generated, the PCR products were separated by electrophoreses on a 6% denaturing polyacrylamide gel. Bands representing cDNAs of interest were excised from dried sequencing gel.
The cDNAs were eluted in distilled water by boiling for 15 minutes and then were reamplified without [α-35 S]dATP, and with 20 µM dNTPs instead of 2 µM dNTPs. From the films, a 282 bp cDNA (referred to as CC282) was identified that was expressed in primary cervical cancer tissue, metastatic lymph node and cervical cancer cell lines but not in normal cervical tissue. CC282 was identified by the use of 5' arbitrary primer H-AP28 (5' -AAGCTTACGATGC-3') and 3' H-T 11 C anchored primer (5' -AAGCTTTTTTTTTTTC-3') (GenHunter). CC282 was then subcloned into pGEM-T easy vector with the use of the TA cloning system and sequenced with the use of the Sequenase Version 2.0 DNA Sequencing System (United States Biochemical Co., Cleveland, OH).

Northern blot analysis
Total RNA was extracted from fresh human tissues and cell lines using RNeasy total RNA kit (Qiagen). Northern blot analysis was carried out, in which 20 µg of denatured total RNA was electrophoresed on a 1.0% formaldehyde agarose gel and transferred to nylon membrane (Roche Diagnostics GmbH, Mannheim, Germany). The mRNA expression of PIG-2 was also assessed in normal human tissues and a variety of human cancer cell lines with the use of prepared membranes obtained from Clontech (Palo Alto, CA) and processed as recommended by the supplier. Human β-actin cDNA control probe provided by Clontech was used as a loading control. All blots were hybridized with the randomly primed [ 32 P]-labeled PIG-2 partial cDNA probe (the CC282 fragment).

Growth curve
To test the effect of PIG-2 gene on HEK 293 cell growth, 1 × 10 5 wild-type HEK 293 cells, PIG-2 gene transfected HEK 293 cells, and HEK 293 cells transfected with pcDNA3.1 alone were cultured for 13 days. In three independent experiments, cells in triplicate flasks were detached and viable cells counted every other day using trypan blue dye exclusion.

Transformation and morphology
PIG-2-transfected HEK 293 cells were maintained in culture for 4-5 weeks with the corresponding media replaced every 3 days and monitored microscopically. HEK 293 nontransformed cells were seeded in parallel. To examine cell morphologies, clones of HEK 293 cells stably transfected with the PIG-2 gene were grown to approximately 70% of confluency in culture flasks, and photographed by Olympus (Inha, Japan) phase-contrast microscopy (magnification, × 100).

Immunohistochemistry
For immunohistochemistry, cryosections (5 µm thick) of human normal and cancer tissues were used. The sections were deparaffinized with xylene and ethanol. After wash-ing with tap water, the sections were treated with methanolic H 2 O 2 for 30 minutes. Before incubation with primary antibody, the sections were permeabilized by incubation in 0.5% Triton X-100 in phosphate-buffered saline (PBS) for 15 minutes and then blocked with normal goat serum for 15 minutes. The sections were incubated with polyclonal anti-Gremlin antibody (IMGENEX, San Diego, CA) for 2 hours at room temperature. After three washes with PBS, the sections were sequentially incubated with biotinylated species-specific secondary antibodies (Vector Laboratories, Burlington, CA) for 1 hour at room temperature, and then avidin and biotinylated horseradish peroxidase according to the manufacturer's recommendations. Aminoethyl carbozole (AEC) was used as the chromogen. After immunostaining, sections were counterstained with hematoxylin. Sections were photographed on an Olympus photomicroscope (Inha, Japan).

Telomerase activity assay
Telomerase activity was measured with the Telo TAGGG Telomerase PCR-ELISA kit (Roche, Germany). The kit provides an immortalized human 293 kidney cell extract as a positive control and 293 cell extract pretreated with RNase as a negative control. All the experiments were performed in triplicate.

Yeast two-hybrid screening and β-galactosidase assay
The MATCHMAKER LexA two-hybrid system was used to identify proteins from the human fetal brain MATCH-MAKER cDNA library that could bind a PIG-2 fusion protein (Clontech, Palo Alto, CA). All experiments were performed in the yeast strain EGY48 transformed with p8op-lacZ, which expresses lacZ and leu genes as reporters (Clontech). We inserted a PIG-2 cDNA fragment into a yeast two-hybrid vector (pLexA) (Clontech) containing the LexA DNA-binding domain. Yeast cells expressing the LexA-PIG-2 were transformed with a human fetal brain cDNA library (Invitrogen) that expresses B42AD fusion proteins. After library transformation, cells are plated on minimal synthetic dropout non-induction medium (Sigma) that selects for both the bait (PIG-2) and the AD/ library plasmids to improve the chances of detecting AD fusion proteins. To confirm the interaction between PIG-2 and binding protein YWHAH, plasmids expressing PIG-2 and YWHAH were co-transformed into yeast cells. βgalactosidase filter lift assays were performed by replicaplating the co-transformants expressing PIG-2 and YWHAH on Trp-, Leu-, His-selection plates. We used a yeast mating assay to eliminate false positive interactions.

GST-tagged YWHAH protein expression and pull down experiments
GST-tagged proteins were expressed and extracts were prepared as recommended by the manufacturer (Amersham Identification and expression level of PIG-2 Figure 1 Identification and expression level of PIG-2. A) Comparison gene expression profiles by DDRT-PCR from total RNA isolated from normal cervical tissue, primary cervical cancer, cervical cancer tissue metastatic to lymph node and from CasKi and CUMC-6 cervical cancer cell lines. Differential display was carried out 5' arbitrary primer H-AP28 (5' -AAGCTTACGATGC-3') and 3' H-T 11 C (5' -AAGCTTTTTTTTTTTC-3'). The PCR products were resolved by electrophoresis. CC282 is the name of the partial PIG-2 gene product. The arrow identifies the location relative to other PCR products. (B) Total RNAs were isolated from normal cervical tissue, primary cervical cancer, cervical cancer tissue metastatic to lymph node and from CasKi and-CUMC-6 cervical cancer cell lines. Blot was hybridized with the randomly primed [ 32 P]-labeled PIG-2 partial cDNA probe (the CC282 fragment). Human β-actin cDNA was used as a control probe (lower panel).

Identification of the PIG-2 gene and different expressions of PIG-2 in human cervical tissues
To identify oncogenes that may be responsible for the formation of cervical and other cancers, we used differential display RT-PCR and identified the 200 bp partial cDNA fragment CC282. As shown in Figure 1A, among the multiple bands, CC282 was strongly expressed in cervical cancer tissue, metastatic lymph node tissue, and CasKi and CUMC-6 cervical cancer cells, but not normal cervical tissue. Using the partial cDNA CC282 as a probe, we screened normal human lung, normal uterine cervical tissue, and cervical cancer cDNA library to identify the fulllength cDNA. One clone with size of 4151 bp was isolated from the human lung cDNA library. This clone, named PIG-2 for human proliferation-inducing gene 2 (GenBank accession number AY 232290), encoded a polypeptide of 184 amino acids with a predicted relative molecular mass of 21 kDa. PIG-2 exhibited 99% homology with the recorded human gremlin 1, cysteine knot superfamily, homolog (Xenopus laevis) (GREM1), mRNA sequence (GenBank accession number NM_013372). The CC282 partial cDNA identified in the differential display corresponded to nucleotides 3901-4100 of PIG-2. Next, we performed Northern blot analysis to examine the expression of PIG-2 in normal cervical tissues, cervical cancer tissues, metastatic lymph node tissues, and cervical cancer cell lines human tissue using the CC282 partial cDNA fragment. PIG-2 (~4.5 kb) mRNA was strongly expressed in cervical cancer tissue, metastatic lymph node tissue, and cervical cancer cells. I normal cervical tissues, the expression of PIG-2 was very low ( Figure 1B).

Differential expression of PIG-2 in various types of human normal and cancer tissues
Northern blot analysis was performed to examine the expression of PIG-2 in human tissue using the CC282 partial cDNA fragment. A transcript of approximately 4.5 kb is present in skeletal muscle, colon, and small intestine tissues examined, as well as a weaker band at 3.7 kb in colon and small intestine (Figure 2A). The levels of mRNA expression were quantified by comparison with the levels of expression of β-actin. We also examined the expression of PIG-2 in several cancer cell lines. PIG-2 mRNA was overexpressed only in human lung cancer cell line A549. Protein-interaction of PIG-2. (A)PIG-2 binds to YWHAH. Co-immunoprecipitation from transfected HEK 293 cells producing PIG-2 and YWHAH proteins was done. Immunoprecipitation was performed with anti-Myc mAb. The protein in the pellets was detected with anti-FLAG mAb. (B) YWHAH binds to PIG-2. Co-immunoprecipitation from transfected HEK 293 cell lines producing YWHAH and PIG-2 proteins was done. Immunoprecipitation was performed with anti-FLAG mAb. The protein in the pellets was detected with anti-Myc mAb.

Expression of PIG-2 in human tissues by Northern analyses and immunohistochemical studies
lung cancer cells among 8 cancer cell lines tested, we studied the expression levels of PIG-2 in normal lung tissues, fresh primary lung cancer tissues, and lung cancer cell lines (NCI-H441, NCI-H157 and NCI-H2009). Northern blot analysis revealed that fresh primary human lung cancer tissues and lung cancer cell lines showed increased expression of PIG-2 levels compared to normal lung tissues ( Figure 2C). Northern blot analysis also revealed that fresh primary human tumor tissues, including carcinomas of the ovary, kidney and breast, showed increased expression of PIG-2 when compared with their normal counterparts ( Figure 2D).
Using anti-gremlin polyclonal antibody, normal or cancer tissue from muscle, colon and pancreas were subjected to immunohistochemical experiments. As a result, PIG-2 was over-expressed in all muscle, colon and pancreas cancer tissues. The immunoreactivity was observed mainly in cancer cells with a cytoplasm dominant manner ( Figures  2F, 2H and 2I). In corresponding normal muscle and colon tissues, there were very weak expressions of PIG-2 ( Figures 2E and 2G). Although further investigation with larger number of samples will be needed, these results indicate that increased expression of the PIG-2 may be associated with human carcinogenesis.

Morphological changes of HEK 293 cells after transfection with PIG-2 gene
Human embryonic kidney (HEK) 293 cells, a differentiated fibroblast cell line, is a spindle shaped fibroblast cell having a long slender nucleus and a scanty amount of cytoplasm ( Figure 3A; left panel). There were no discernable differences in cell morphology between wild-type HEK 293 cells and cells transfected with vector alone ( Figure  3A; middle panel). But HEK 293 cells expressing PIG-2 are altered to polygonal cells with an ovoid nucleus and plump cytoplasm ( Figure 3A; right panel).

Growth stimulation of HEK 293 cells by PIG-2
The rate of PIG-2-transfected HEK 293 cell growth was increased compared to those of cells transfected with vector alone or wild-type HEK 293 cells. About 150% of PIG-2-transfected HEK 293 cells remained viable at 13 days when compared with wild-type HEK 293 cells ( Figure 3B).

Telomerase activity in the PIG-2-transfected HEK 293 cells
Alterations in telomere biology both suppress and facilitate malignant transformation by regulating genomic stability and cellular life span [31]. Telomerase is an enzymatic ribonucleoprotein complex that acts as a reverse transcriptase in the elongation of telomeres. Telomerase activity is almost absent in somatic cells, but it is detected in embryonic stem cells and in the vast majority of tumor cells [32]. To explain the possible oncogenic role of PIG-2-transfected cells, we determined telomerase activity in PIG-2-transfected HEK 293 cells. Wild-type HEK 293 cells showed detectable telomerase activity (Figure 3C). PIG-2 gene transfection increased telomerase activity up to about 2-fold when compared with HEK 293 wild-type cells ( Figure 3C).

PIG-2 interacts with YWHAH in vivo
To understand the PIG-2 tumorigenesis pathway, we performed a yeast two-hybrid screen and identified the YWHAH protein, encoded by the gene homo sapiens tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein, eta polypeptide (GenBank accession number NM_003405), was interacted with PIG-2. The interaction of PIG-2 and YWHAH identified by yeast two hybrid screen was confirmed by glutathione-S-transferase (GST) pull-down experiments and immunoprecipitation studies in vitro. The physical interaction between PIG-2 and YWHAH was examined in vivo by co-immunoprecipitation in HEK 293 cells. The FLAG epitope-tagged PIG-2 was transiently overexpressed together with the Myc-tagged YWHAH in HEK 293 cells, and the cell lysates were immunoprecipitated with the anti-Myc antibody and detected by western blotting with anti-FLAG antibody ( Figure 4A). We also performed the immunoprecipitation of reverse form of co-expression, YWHAH-Myc and PIG-2-FLAG proteins, in vivo ( Figure 4B). As shown in Figure 4A, PIG-2 could be co-immunoprecipitated with YWHAH, indicating that YWHAH is likely to be a binding partner. Taken together, co-immunoprecipitation showed that PIG-2 binds YWHAH protein (Figure 4).
analysis indicates that the YWHAH binding site for PIG-2 is located between residues 61-80, which constructively include in the α helix C-α helix D linker, may play an important role in this intracellular interaction ( Figure  5C).

Discussion
To discover genes involved in human cervical carcinogenesis, we applied DDRT-PCR and identified the candidate human cervical cancer-related gene, proliferation-inducing gene 2 (PIG-2) (GenBank accession number AY232290). PIG-2 exhibited close similarity (99%) to gremlin 1 cDNA (GenBank accession number NM_013372) in the database.
Drm/Gremlin and Dan, two homologous secreted antagonists of bone morphogenic proteins, have been shown to regulate early development, tumorigenesis, and renal pathophysiology [6,[10][11][12][13][14][15][16]. Topol et al. had previously shown that most tumor-derived cells fail to express Drm [33] and that in fibroblasts Drm expression is inhibited following oncogene-induced transformation [6]. Human Drm maps to chromosome 15q13-q15, within a region whose loss is associated with metastatic breast cancer and other metastatic carcinomas [34]. These properties suggested that Drm might play an inhibitory role in cell transformation or tumorigenesis. They also demonstrated that overexpression of Drm in the tumor-derived cell lines Daoy (primitive neuroectodermal) and Saos-2 (osteoblastic) significantly inhibited tumorigenesis and provided evidence that Drm can function as a novel transformation suppressor and suggested that this may occur through its affect on the levels of p21 Cip1 and phosphorylated p42/44 MAPK [14,35]. Recent publication also demonstrated that gremlin mRNA is expressed in non-malignant epithelial cells and lost in many human cancer cell lines via promoter methylation [36]. Similar finding is also reported by other group [31].
On the contrary, our experiments showed that PIG-2 which is identical with gremlin 1 was overexpressed in various human tumors including carcinomas of the cervix, lung, ovary, kidney, breast, colon, pancreas and sarcoma. However, expression of PIG-2 was generally downregulated in diverse human normal tisues. In our experiments, PIG-2-transfected HEK 293 cells exhibited growth stimulation and increased telomerase activity. Although further investigation with larger number of samples will be needed, it suggests PIG-2 may play a fundamental oncogenic role in multiple body organs.
However, it is unknown how PIG-2 contributes to the cellular and biochemical mechanisms of human tumorigenesis. In this study, we identified an oncogene that is expressed in multiple different human cancers, and inves-tigated whether the oncogene is responsible for the genesis of human cancer. To understand the PIG-2 tumorigenesis pathway, we performed a yeast two-hybrid screen and identified the 14-3-3 eta (YWHAH) protein was interacted with PIG-2 [20][21][22][23][24][25][26][27][28].
This gene product belongs to the 14-3-3 family of proteins which mediate signal transduction by binding to phosphoserine-containing proteins. This highly conserved protein family is found in both plants and mammals, and this protein is 99% identical to the mouse, rat and bovine orthologs. This gene contains a 7 bp repeat sequence in its 5' UTR, and changes in the number of this repeat has been associated with early-onset schizophrenia [26]. The 14-3-3 proteins are a family of conserved regulatory molecules expressed in all eukaryotic cells. A striking feature of the 14-3-3 proteins is their ability to bind a multitude of functionally diverse signaling proteins, including kinases, phosphatases, and transmembrane receptors. This plethora of interacting proteins allows 14-3-3 to play important roles in a wide range of vital regulatory processes, such as mitogenic signal transduction, apoptotic cell death, and cell cycle control [37]. Acronyms14-3-3 family proteins interact with many signaling molecules, such as MAPK kinase kinase, Raf-1, Wee1, Cdc25, cyclin B1, protein kinase C, IGF-I receptor, insulin receptor substrate 1, Bad, and Bcl [38][39][40][41][42], and regulate several signal transduction pathways [43][44][45]. Also, 14-3-3 proteins help two molecules to interact or to interrupt the association between two molecules by functioning as molecular scaffolds [46].
Binding of a protein to a 14-3-3 protein may result in stabilization of the active or inactive phosphorylated form of the protein, to a conformational alteration leading to activation or inhibition, to a different subcellular localization or to the interaction with other proteins. Currently genome-and proteome-wide studies are contributing to a wider knowledge of this important family of proteins [47].
The molecular consequences of 14-3-3 binding are diverse and only partly understood. Disturbance of the interactions with 14-3-3 proteins may lead to disease like cancer. In this study, gremlin 1 binds YWHAH protein in vitro and in vivo. We suspect that this binding may disturb the interactions with 14-3-3 proteins and lead to disease like cancer.

Conclusion
Gremlin 1 was overexpressed in various human tumors and plays a oncogenic role especially in carcinomas of the cervix, lung, ovary, kidney, breast, colon, pancreas and sarcoma. Although further investigation with larger number of samples will be needed, these results indicate that increased expression of the PIG-2 may be associated with human tumorigenesis. Our study suggests that overexpressed gremlin 1 functions by interaction with YWHAH in human tumorigenesis. While further studies are needed to characterize cellular functions and regulatory mechanisms, gremlin 1 is a candidate oncoprotein in the development of many types of human cancers, and gremlin 1 and its binding protein YWHAH could be good targets for developing diagnostic and therapeutic strategies against human cancers.
Publish with Bio Med Central and every scientist can read your work free of charge