Pregnancy-specific glycoprotein 9 (PSG9), a driver for colorectal cancer, enhances angiogenesis via activation of SMAD4

Abstract

PSG9 is a member of the pregnancy-specific glycoprotein (PSG) family and has been shown to contribute to the progression of colorectal cancer (CRC) and cancer-related angiogenesis. Here, we aim to investigate abnormal PSG9 levels in patients with CRC and to emphasize the role of PSG9 in driving tumorigenesis. Serum from 140 patients with CRC and 125 healthy controls as well as 74 paired tumors and adjacent normal tissue were used to determine PSG9 levels. We discovered that PSG9 was significantly increased in serum (P<0.001) and in tumor tissues (P<0.001) from patients with CRC. Interestingly, the increased PSG9 levels correlated with poor survival (P=0.009) and microvessel density (MVD) (P=0.034). The overexpression of PSG9 strongly promoted the proliferation and migration of HCT-116 and HT-29 cells. However, PSG9 depletion inhibited the proliferation of SW-480 cells. Using a human umbilical vein endothelial cell tube-forming assay, we found that PSG9 promoted angiogenesis. The overexpression of PSG9 also increased the growth of tumor xenografts in nude mice. Co-immunoprecipitation experiments revealed that PSG9 was bound to SMAD4. The PSG9/SMAD4 complex recruited cytoplasmic SMAD2/3 to form a complex, which enhanced SMAD4 nuclear retention. The PSG9 and SMAD4 complex activated the expression of multiple angiogenesis-related genes (included IGFBP-3, PDGF-AA, GM-CSF, and VEGFA). Together, our findings illustrate the innovative mechanism by which PSG9 drives the progression of CRC and tumor angiogenesis. This occurs via nuclear translocation of PSG9/SMAD4, which activates angiogenic cytokines. Therefore, our study may provide evidence for novel treatment strategies by targeting PSG9 in antiangiogenic cancer therapy.

Keywords: PSG9; SMAD4; angiogenesis; colorectal cancer (CRC).

Figure 1. High levels of PSG9 are associated with micro-vessel density (MVD) and poor outcomes in patients with colorectal cancer (CRC)

A. PSG9 levels in serum from healthy controls and patients with CRC. B. PSG9 levels in CRC patients with a tumor diameter ≤ 5 cm and > 5 cm. C. The levels of PSG9 range from negative (−) (score of 0~2 score), weak (+) (score of 3~4), moderate (++) (score of 4) to strong (+++) (score ≥ 5) in CRC tumors. D. CD31 and PSG9 immunohistochemical staining in a CRC tissue section illustrates the location of PSG9 and the MVD. Kaplan-Meier curves show the differences between the high and low expression of PSG9 (E) and MVD (F).

Figure 2. PSG9 promotes the proliferation and migration of CRC cells

A. Quantitative real-time PCR detected the expression of PSG9 in HT-29/HCT-116-pcDNA3.1/pcDNA-3.1-PSG9 (pcPSG9) and SW-480-shControl (shCon) and shPSG9 cells. The data represent the mean±S.D. values of three replicates. *P<0.05. B. Western blot determined the expression of PSG9 in HT-29/HCT-116-pcDNA3.1/pcPSG9 cells and SW-480-shCon/shPSG9 cells. Cells transfected with the pcPSG9 plasmid displayed significantly higher PSG9 expression compared with cells that were transfected with the pcDNA3.1 plasmid. Conversely, PSG9 shRNA decreased the expression of PSG9 in SW-480 cells. C. A CCK-8 assay was used to analyze the proliferation of HT-29 and HCT-116 cells stably transfected with pcDNA3.1 and pcPSG9 and in SW-480 cells stably transfected with shCon and shPSG9. D. EdU incorporation assays compared the percentage of proliferating cells in HT-29/HCT-116-pcPSG9 and HT-29/HCT-116-pcDNA3.1 cells as well as in SW480 cells stably transfected with shPSG9 and shCon. E. Colony formation assays revealed the colony formation rates of HT-29/HCT-116-pcDNA3.1 and HT-29/HCT-116-pcPSG9 cells as well as SW-480-shCon and SW-480-shPSG9 cells. F. Transwell migration assays determined the migration abilities of HT-29/HCT-116-pcDNA3.1 and HT-29/HCT-116-pcPSG9 cells. The data represent the mean±S.D. values of three replicates. *P<0.05.

Figure 3. PSG9 promotes tumor proliferation in vivo and enhances angiogenesis in vivo and in vitro

A. Tumor xenografts from nude mice that were injected with HT-29-pcPSG9 cells (n=5) exhibited significantly higher growth rates compared with tumors from mice that were injected with HT-29-pcDNA3.1 cells (n=4). Tumors from mice that were injected with HCT-116-pcPSG9 cells (n=5) also exhibited increased growth rates compared with tumors from mice injected with HCT-116-pcDNA3.1 cells (n=5). *P<0.05. B. Representative immunohistochemical staining for CD31 imaged from HT-29-pcPSG9 cells that formed tumor xenografts. C. IHC staining for CD31 in the blood vessels of tumor xenografts derived from HT-29/HCT-116-pcPSG9 cells was significantly higher than staining in tumors derived from HT-29/HCT-116-pcDNA3.1 cells. *P<0.05. D. Endothelial cell medium (ECM) and conditioned medium (CM) collected from HCT-116-pcDNA3.1 and HCT-116-pcPSG9 cells, along with 5 μg/ml purified human recombinant PSG9 proteins, PSG9 antibody (ab), and VEGFA ab were used to induce the formation of vessel-like tubes in HUVECs. E. pcPSG9 CM and PSG9 proteins increased the number (Nb.) of HUVECs that formed branches compared with treatment with HCT-116-pcDNA3.1 CM, ECM, PSG9 ab, and VEGFA ab. F. pcPSG9 CM and PSG9 proteins led to a greater total length of vessel-like tubes that formed in HUVECs. The data (Nb. branches and length) represent the mean±S.D. values of three replicates.

Figure 4. PSG9 binds directly to SMAD4

A. BioGRID^3.4 identified 16 proteins bound to PSG9. SMAD4 is the closest molecule to PSG9 in the protein-protein interaction network. B. A co-immunoprecipitation (Co-IP) assay was performed to determine the interaction between PSG9 and SMAD4. Left, anti-SMAD4 antibody was used for PSG9 immunoprecipitation (IP), and a mouse IgG (IgG M) antibody was used as the negative control. Immunoblotting (IB) with a PSG9 antibody was used to determine the PSG9 levels in SW-480, HT-29, and HCT-116 cells. Right, anti-PSG9 antibody was used for SMAD4 immunoprecipitation, and a rabbit IgG (IgG R) antibody was used as the negative control. C. Exogenous expression of PSG9 and SMAD2/3/4 in HCT-116 cells. Vectors of c-myc-pcDNA3.1 PSG9 combined with Flag-SMAD2, Flag-SMAD3 or Flag-SMAD4 were transfected into HCT-116 cells for 24 hours. Immunoblotting with an anti-Flag antibody detected higher levels of Flag-SMAD4 than Flag-SMAD2/3 when an anti-c-myc antibody was subjected to Flag-SMAD2/3/4 immunoprecipitation (Top). An anti-Flag antibody subjected to c-myc-PSG9 and anti-c-myc antibody was used to immunoblot for PSG9 (Bottom). D. IF shows the location of the PSG9 and SMAD4 proteins in HCT-116 cells. DAPI was used to stain the nucleus.

Figure 5. PSG9 increases the nuclear retention of SMAD4

A. PSG9 increased the nuclear retention of SMAD4 in stable transfected pcPSG9 HT-29 cells compared with stable transfected pcDNA3.1 cells. B. SMAD4 nuclear expression was determined by IF in HCT-116-pcDNA3.1 and HCT-116-pcPSG9 cells after treatment with 10 ng/ml TGF-β for 4 hours. C. PSG9 increased the nuclear retention of SMAD4 in HCT-116 cells. HCT-116-pcDNA3.1 and HCT-116-pcPSG9 cells were treated with 10 ng/ml TGF-β for 4 hours; cells were washed 3 times to remove TGF-β and were then treated with SB431542 for up 4 hours. The cells were harvested at the indicated times, and both the nuclear and cytoplasmic fractions were collected. SMAD4 nuclear and cytoplasmic ratios were analyzed after treatment with SB431542. Data are expressed as the mean±S.D. values from three replicates. D. Depletion of PSG9 reduced the levels of nuclear SMAD4 compared with shCon in SW-480 cells. SW-480 cells were treated with 10 ng/ml TGF-β for 4 hours. Data are expressed as the mean±S.D. values from 2 experiments.

Figure 6. PSG9/SMAD4 induces the expression of angiogenesis-related genes

A. An angiogenesis membrane-based antibody array was used to determine human angiogenesis-related proteins in HCT-116-pcDNA3.1 and HCT-116-pcPSG9 cell lysates. B. An antibody array was used to test the levels of angiogenesis-related proteins in SW-480-shControl (Con) and SW-480-shPSG9 cell lysates. Bottom, semi-quantitative data from dot density relative reference (Ref). C. Quantitative PCR determined the mRNA levels of VEGFA, PDGF-AA, GM-CSF, and IGFBP-3. The data represent the mean±S.D. values of at least four replicates. *P<0.05. D. Chromatin immunoprecipitation (ChIP) revealed that the PSG9/SMAD4 complex was bound to the promoters of IGFBP-3, PDGF-AA, GM-CSF, and VEGFA in HCT-116-pcDNA3.1 and HCT-116-pcPSG9 cells.