Carcinoembryonic antigen interacts with TGF-{beta} receptor and inhibits TGF-{beta} signaling in colorectal cancers

Abstract

As a tumor marker for colorectal cancers, carcinoembryonic antigen (CEA) enhances the metastatic potential of cancer cells. CEA functions as an intercellular adhesion molecule and is upregulated in a wide variety of human cancers. However, the molecular mechanisms by which CEA mediates metastasis remain to be understood. Transforming growth factor-β (TGF-β) signaling regulates both tumor suppression and metastasis, and also contributes to the stimulation of CEA transcription and secretion in colorectal cancer cells. However, it remains unknown whether CEA, in turn, influences TGF-β functions and if a regulatory cross-talk exists between CEA and the TGF-β signaling pathway. Here, we report that CEA directly interacts with TGF-β receptor and inhibits TGF-β signaling. Targeting CEA with either CEA-specific antibody or siRNA rescues TGF-β response in colorectal cancer cell lines with elevated CEA, thereby restoring the inhibitory effects of TGF-β signaling on proliferation. CEA also enhances the survival of colorectal cancer cells in both local colonization and liver metastasis in animal study. Our study provides novel insights into the interaction between CEA and TGF-β signaling pathway and establishes a negative feedback loop in amplifying the progression of colon cancer cells to more invasive phenotypes. These findings offer new therapeutic opportunities to inhibit colorectal cancer cell proliferation by cotargeting CEA in promoting tumor-inhibitory action of the TGF-β pathway.

Figure 1. Direct association of CEA with TBRI

A) 293T cells were co-transfected with wt CEA and one of the five elements of TGF-β signaling pathway, namely β2SP, Smad3, TBRI, TBRII, and Smad4 as indicated. Coimmunoprecipitation experiments were performed to examine the interaction between CEA and these five elements. B) LS174T and LS180 cells lysate were used for the coimmunoprecipitation assay to determine the association of endogenous CEA with endogenous TBRI. C) In vitro binding assay. Purified CEA protein was incubated with purified HA-TBRI or HA-TBRII. Binding of CEA to TBRI or TBRII was assessed by immunoprecipitation followed by immunoblotting. D) 293T Cells were transfected with wt CEA plasmid or vector pcDNA. Cells were then fixed and stained for CEA and TBRI 24 h after transfection.

Figure 2. Inhibition of TGF-β signaling by CEA

A) 293T cells were co-transfected with HA-TBRI, V5- Smad3 and CEA as indicated. 24 h after transfection, cells were treated with TGF-β (100 pM) for 1 h. Coimmunoprecipitation was carried out to evaluate the association of TBRI with Smad3. B) 293T cells were transfected with CEA or vector plasmid for 24 h, and then stimulated with TGF-β (100 pM) for different time periods as indicated. Cells were harvested and p- Smad3 and Smad3 protein levels were evaluated by Western blotting. Histogram shows the fold increase of p- Smad3 intensity compared to that at the time point 0. C) 293T cells were transfected as in B, then stimulated with TGF-β (100 pM) for I h. Cells were then fixed and stained for Smad3 and CEA. Counterstain, DAPI). D) Cells were transfected and stimulated with TGF-β (100 pM) for I h. Cells were lysed, and cytoplasmic and nuclear fractions were separated. Smad3 levels in different fractions were evaluated by immunoblotting. Actin served as loading control. E) 293T cells were treated as in C. c-myc mRNA levels were assessed by RT-PCR. GAPDH served as loading control.

Figure 3. TGF-β signaling is impaired in colorectal cancer cells with elevated CEA

A) CEA expression and TGF-β induced Smad3 phosphorylation were evaluated in 12 colorectal cancer cell lines by immunoblotting. The extent of Smad3 phosphorylation was measured by the fold increase of p- Smad3 (ratio of p- Smad3 with TGF-β stimulation to p- Smad3 without TGF-β stimulation). According to CEA expression levels, the cell lines were classified into 3 groups. The scatter plot graph demonstrates an inverse correlation between CEA expression levels and the extent of Smad3 phosphorylation. Blue circles: cell lines 1,2,3 and 8; Yellow squares: cell lines 4,6,7, and 12; Red diamonds: cell lines 5,9,10 and 11. Broken lines represent average levels of p-Smad3 fold increase in each group. The bars indicate the standard error. 1. SK-CO-1; 2. LS180; 3. LS174T; 4. Caco-2; 5. HCT-6; 6. Colo205; 7. HT-29; 8. Lovo; 9. HCT116; 10. SW480; 11. HCT-15; 12. DLD-1. B) LS174T cells were treated with anti-CEA antibody (3 μg/ml) or naive IgG for 24 h to block CEA and then treated with or without TGF-β (100 pM) for 1 h. Nuclear translocation of Smad3 was determined as in Figure 2C. C). Five CRC cell lines were treated with or without anti-CEA antibody for 24 h as indicated, then treated with TGF-β (100 pM) for 1 h. Transcription levels of c-myc were assessed as in Figure 2E. The histogram shows the quantification of c-myc mRNA levels in the left graph. D–F) Anti-CEA Ab promotes Smad3-dependent repression of c-Myc expression by TGF-β. D. Effect of IgG or CEA antibody on the c-Myc-luc promoter activity (lower panel) and on the c-Myc protein in the HCT116 cells treated with or without TGF-β. *P<0.05. Western blot analysis was performed with the cell lysates obtained from the luciferase assay samples. E. ChIP analysis showing the recruitment of Smad3 but not β-spectrin onto human c-myc promoter in the HCT116 cells treated with IgG or CEA antibody in the presence or absence of TGF-β treatment. F. EMSA analysis of the Smad3 binding in the human c-Myc promoter using the PCR product encompassing the region −5 to −233 in HCT116 cells treated with TGF-β in presence of either IgG or CEA antibody.

Figure 4. Targeting CEA rescues the inhibitory effects of TGF-β signaling on proliferation of colorectal cancer cells

A) Clone A cells were stably transfected either with pcDNA vector or with plasmids encoding wt CEA. 10,000 cells were seeded into each well of 96 well plates, and treated with TGF-β (100 pM) and/or anti-CEA antibody (3 μg/ml) as indicated for 24 h. Cell proliferation rates were assessed by WST assay. Proliferation rate was calculated as (absorbance with treatment/absorbance without treatment) × 100%. The graph depicts percentages of proliferation with data from two independent experiments (each carried out in triplicate). *, P< 0.05 for the comparison with cells treated with TGF-β alone (Student’s t test). B). Three colorectal cancer cell lines were treated with TGF-β and different doses of anti-CEA antibody as indicated. Cell proliferation assays were performed as in A. C). DLD- cells were infected with lentiviral control siRNA or siRNA targeting CEA. Western blot demonstrated that the siCEA suppressed CEA level efficiently. Stable infected cells were selected by puromycin treatment. Cells were then treated with or without TGF-β (100 pM) for different time periods. Cell proliferation assay was performed as in A. *: p<0.05 for the comparison between the TGF-β treatment and non-treatment groups at the same time point.

Figure 5. CEA enhances liver metastasis of colorectal cancer cells

Clone A cells were stably transfected either with pcDNA vector or with plasmids encoding wt CEA, and then injected intrasplenically into nude mice. 2×10⁶ viable cells were injected for each of 10 mice per group. Mice were autopsied to determine spleen and liver colonization 30 days after injection. A) Representative pictures show the tumor colonies in spleens and livers. Arrows indicate the colonies formed by different cell types. B) The graph shows the percentage of mice with liver colonies.