Krüppel-like factor 8 regulates VEGFA expression and angiogenesis in hepatocellular carcinoma

Tumor angiogenesis plays a critical role in hepatocellular carcinoma (HCC) development and progression, but its mechanism is unclear. Krüppel-like factor 8 (KLF8) is a transcription factor that plays an important role in HCC progression. Here, we investigated the role of KLF8 in angiogenesis in HCC and its possible mechanism. Immunohistochemistry, quantitative RT-PCR, western blotting, promoter reporter assays, chromatin immunoprecipitation (ChIP), and chicken chorioallantoic membrane (CAM) and nude mouse tumor models were used to show that the mRNA and protein expression levels of KLF8 and VEGFA are highly correlated in HCC tissue samples. The up-regulation of KLF8 increased VEGFA protein levels and induced VEGFA promoter activity by binding to the CACCC region of the VEGFA promoter. In addition, KLF8 regulated HIF-1α and Focal adhesion kinase (FAK) expression. The PI3K/AKT inhibitor LY294002 inhibited KLF8-induced VEGFA expression, whereas PI3K/AKT signaling pathway proteins, such as P-PDK1(Ser241) and P-AKT(Thr308), were decreased significantly. KLF8-overexpressing HCC cells had a higher potential for inducing angiogenesis. Thus, our results indicate that KLF8 may induce angiogenesis in HCC by binding to the CACCC region of the VEGFA promoter to induce VEGFA promoter activity and through FAK to activate PI3K/AKT signaling to regulate HIF-1α expression levels.

. The mRNA and protein levels of KLF8 and VEGFA are highly correlated in HCC. (a) KLF8 and VEGFA protein expression levels were detected in paraffin-embedded HCC specimens by immunohistochemistry staining. Integrated densities were measured by ImageJ. (b) Pearson's correlation analysis indicated that KLF8 protein expression was highly correlated with VEGFA protein expression, r = 0.622, p < 0.05. (c) mRNA expression levels of KLF8 and VEGFA in 50 fresh HCC tissue samples were detected by qRT-PCR, and the relative expression levels were calculated using the 2 −ΔΔCt method. (d) Pearson's correlation analysis indicated that KLF8 mRNA expression was highly correlated with VEGFA mRNA expression, r = 0.414, p < 0.05. was cloned from human genomic DNA, and this fragment was inserted into pGL3-Basic to construct the pGL3-Basic-VEGFA-P plasmid. Promoter reporter assays were used to detect the effects of KLF8 up-regulation on VEGFA promoter activity. VEGFA promoter activity was induced significantly by KLF8 up-regulation (1.49 ± 0.04 vs. 3.08 ± 0.04, P < 0.0001, n = 3) (Fig. 3a).
KLF8 binds to the CACCC region of the VEGFA promoter. KLF8 can function as either a transcription repressor or activator by binding to the GT-box (CACCC) promoter sequence via its three C-terminal C2H2 zinc fingers that are highly conserved among KLFs 9,14-17 . To confirm whether KLF8 binds to the CACCC region of the VEGFA promoter or not, ChIP assays were performed to identify the KLF8-binding region of the VEGFA promoter. Several primers were designed to amplify the CACCC sites of the VEGFA promoter. The amplification for anti-KLF8 in KLF8-overexpressing SMMC7721 cells and pcDNA3.1 transfected SMMC7721 cells is 715.0 ± 42.23 vs. 2.15 ± 0.16 (p < 0.05, n = 3) (Fig. 3b).This result indicated that KLF8 could bind to the CACCC region of the VEGFA promoter.

Discussion
The growth and metastasis of cancer depend on angiogenesis. Vascular endothelial growth factor (VEGF) has been identified as a key mediator of tumor angiogenesis. The VEGF family includes VEGF-A, VEGF-B, VEGF-C, VEGF-D, VEGF-E, and placental growth factor (PlGF), and VEGF-A appears to be the most important in the growth of blood vessels in a variety of normal and pathological circumstances 18 . The effects of VEGF are mediated by endothelial cells via its receptors VEGFR-1 (Flt-1) and VEGFR-2 (KDR) 19 . Some tumor cells may also express VEGF receptors, and VEGF may act as an autocrine growth factor that stimulates the proliferation of some cancer cells 20 . In addition to the HIF1-α pathway, HBx protein activation, another mechanism that activates oncogenes; tumor suppressor gene loss or inactivation; and multiple signal transduction pathways, including Egr-1 and Sp1, may be involved in VEGF regulation in HCC. However, the mechanism of VEGF expression and its regulation in HCC are mostly unknown.
KLF8 is a member of the Krüppel-like C2H2 zinc-finger transcription factor family of proteins 9 . In our previous research, KLF8 up-regulation promoted HCC cell proliferation and invasion and inhibited apoptosis, and the over-expression of KLF8 increased HCC progression and metastasis. In the present study, we examined the expression of and relationship between KLF8 and VEGF in the tumor tissues of HCC patients. We found that the expression levels of KLF8 and VEGFA were highly related in HCC samples. KLF8-overexpressing HCC cells had higher VEGFA mRNA and protein levels. KLF8-overexpressing HCC cells had a higher potential for inducing angiogenesis according to chick chorioallantoic membrane (CAM) assays and a nude mouse HCC model.
Because KLF8 is a transcriptional factor, KLF8 up-regulation induced VEGFA promoter activity by binding to the CACCC region of the VEGFA promoter. PI3K/AKT signaling plays an important role in angiogenesis; 21 once PI3K signaling is activated by PIP3, the pleckstrin homology (PH) domain of PDK1 is recruited to the plasma membrane, which results in the activation of membrane-associated AKT at threonine 308. AKT phosphorylation at serine 473 occurs independently via mammalian target of rapamycin complex 2 or is induced by PIP3. In addition, PIP3 binding activates PDK1 by promoting serine 241 autophosphorylation 22 . The mutation of PDK1 at serine 241 significantly reduces PDK1 activity toward AKT 23,24 . Activation of the EGFR/PI3K/ AKT/mTOR pathway could increase VEGF expression by up-regulating HIF-1α 25 . Here, we showed that KLF8 up-regulation in HCC cells increased HIF1-α expression levels and that KLF8 down-regulation decreased HIF1-α expression levels. The induction of VEGF expression via KLF8 overexpression was blocked by the PI3K/ AKT-specific inhibitor LY294002; in addition, the PI3K/AKT signaling pathway proteins P-PDK1(Ser241) and P-AKT(Thr308) decreased significantly, but the protein expression levels of P-AKT(Ser473) were not different. In pcDNA3.1-transfected SMMC7721 cells treated with LY294002 or DMSO, the protein levels of P-AKT (Thr308) were not different, and KLF8-overexpressing HCC cells had higher levels of P-PDK1(Ser241), P-AKT(Thr308) and P-AKT(Ser473). These results indicated that KLF8 up-regulation may act through the PI3K/AKT signaling pathway to increase P-PDK1(Ser241) levels; then, increased P-AKT(Thr308) or P-AKT(Ser473) protein levels could induce VEGFA protein expression.
Focal adhesion kinase (FAK) is a cytoplasmic protein tyrosine kinase that participates in regulating diverse cellular functions, such as cell spreading, migration, proliferation, and apoptosis 14 .The FAK/PI3K/AKT signaling pathway plays an important role in HCC invasion 26 , and KLF8 overexpression causes the CXCL12/ CXCR4-dependent activation of FAK 27 . Here, we showed that KLF8-overexpressing HCC cells had higher FAK levels (Supplementary Figure 1a), and the protein expression level of p-AKT decreased significantly in FAK down-regulated SMMC7721 cells (Supplementary Figure 1b),so it is possible that KLF8 activates PI3K/AKT signaling through FAK.
PTEN (phosphate and tensin homologue deleted on chromosome TEN) acts as a key negative regulator of the ligand-activated PI3K-AKT pathway; 28,29 PTEN dephosphorylates phosphatidylinositol (3,4,5) triphosphate to its diphosphate (4,5) form, thus reducing the activation of AKT 30 . PTEN also has a restrictive role in angiogenesis 31 . The activation of Wnt signaling up-regulates VEGF expression 32 . GSK-3β is a negative regulator of Wnt signaling, and inhibiting GSK-3β increases VEGF promoter activity 33 . GSK-3β down-regulates HIF-1 and VEGF expression, thus inhibiting tumor angiogenesis in vivo 34 . Raf isoforms (ARAF, BRAF and CRAF in humans) initiate Raf/MEK/ERK signaling and can activate serine/threonine kinases; inhibiting the phosphorylation of c-Raf decreases the levels of p-MEK and p-ERK 35 . PI3K/AKT and Raf/MEK/ERK signaling cascades concurrently participate in angiogenesis via HIF-1α-mediated VEGF expression that is stimulated by notoginsenoside Ft1 (Ft1) 36 . In our study, KLF8-overexpressing SMMC7721 cells had higher levels of p-PTEN, P-GSK-3β and P-c-Raf, and these proteins levels decreased after LY294002 treatment. In pcDNA3.1-transfected SMMC7721 cells treated with LY294002 or DMSO, the protein levels of p-PTEN and P-c-Raf were not different, and P-GSK-3β protein expression levels decreased after LY294002 treatment. The roles of p-PTEN, P-GSK-3β and P-GSK-3β in KLF8-regulated angiogenesis in HCC need further investigation. Taken together, our research showed that KLF8 induced angiogenesis in HCC by binding to the CACCC region of the VEGFA promoter to induce VEGFA promoter activity and through FAK to activate the PI3K/AKT signaling pathway to increase P-PDK1(Ser241) levels; then, increased P-AKT(Thr308) or P-AKT(Ser473) and HIF-1α levels induced VEGFA protein expression Figure 8. KLF8 promotes tumor growth and angiogenesis in vivo SMMC7721 cells (5 × 10 6 ) transfected with pcDNA3.1-KLF8 or pcDNA3.1 were inoculated into the liver parenchyma of nude mice under ketamine/ xylazine anesthesia after the abdomen was opened. All mice were monitored once every 3 days and sacrificed 5 weeks later. Tumor tissue sections were prepared, and immunoreactivity was analyzed as above using KLF8, VEGF and CD31 antibodies. (a,b) SMMC7721 cells transfected with pcDNA3.1-KLF8 had greater growth potential than SMMC7721 cells transfected with pcDNA3.1. In the nude mouse livers, the tumor weights were significantly higher in the SMMC7721-pcDNA3.1-KLF8 group than in the SMMC7721-pcDNA3.1 group (3.6 ± 0.6 g vs 1.0 ± 0.3 g, P < 0.01, n = 3). (c,d) VEGF and CD31 expression levels were detected by immunohistochemistry. The integrated density of VEGF staining was higher in the SMMC7721-pcDNA3.1-KLF8 group than in the SMMC7721-pcDNA3.1 group (129.2 ± 1.6 vs 46.3 ± 7.2, P < 0.01), and the tumor vessel density was significantly increased in the SMMC7721-pcDNA3. Cell transfection. Cells were seeded at a density of 5 × 10 5 cells/well in 6-well plates one day before transfection. Four KLF8 shRNA expression plasmids pGPU6/GFP/Neo-KLF8 were constructed by Genepharma, Shanghai, and pGPU6/GFP/Neo-ShNC was used as a control; the most effective pGPU6/GFP/ Neo-KLF8 was screened by real-time PCR and used to down-regulate KLF8 expression. FAK siRNA(sense: GUAUUGGACCUGCGAGGGA, anti-sense: UCCCUCGCAGGUCCAAUAC) was used to down-regulate FAK expression, siRNA (sense:TTCTCCGAACGTGTCACGT, anti-sense: ACGTGACACGTTCGGAGAA) was used as control (GenePharma, Shanghai).The shRNA sequences were as follows: KLF8-shRNA-1: GAAGACCTAGCATGCTACAAGCTCCAATTCAAGAGATTGGAGCTTGTAGC ATGCTAGTTTTTG KLF8-shRNA -2: GATCCAAAAACTAGCATGCTACAAGCTCCAATCTCT TGAAT TGGAG CTTGTAGCATGCTAG.
Construction of the VEGFA promoter-luciferase plasmids and co-transfection and luciferase assays. To assess the activity of the VEGFA promoter induced by KLF8, we constructed the pGL3-Basic-VEGFA-P plasmid. The promoter region of VEGF-A (−2068/50 bp) was cloned from human genomic DNA, and the fragment was inserted into pGL3-Basic. The constructs were verified by restriction endonuclease digestion and sequencing. SMMC7721 cells were seeded in 24-well plates at 0.5 × 10 5 cells/well one day before transfection. SMMC7721 cells were co-transfected with 0.4 μg of the VEGF-A promoter luciferase reporter constructs, 0.2 μg of pGL3-basic-VEGFA promoter or pGL3-Basic reporter plasmids (Promega), and 0.2 μg of pcDNA3.1 (blank vector as a control) or pcDNA3.1-KLF8. Luciferase activity analyses were performed two days after transfection using a luciferase assay kit (Dual-Luciferase Reporter Assay System, Promega), and the data were normalized to Renilla luciferase activity.
Chromatin Immunoprecipitation (ChIP) Assay. Nuclei for the ChIP assays were sonicated in shearing buffer, and the shearing effectiveness was confirmed by electrophoresis in ethidium bromide-stained agarose gels. The samples were then processed for immunoprecipitation using a kit (EZ-ChIP ™ Chromatin Immunoprecipitation Kit, Millipore) and antibodies to KLF8 (Santa) according to the manufacturer's instructions. After precipitation, the cross-linking was reversed, and PCR was carried out using 1 μL of each sample (input DNA dilution 1:10; immunoprecipitated fractions were undiluted) in PCR buffer (Qiagen, Valencia, CA) containing dNTPs (Invitrogen) and TAQ DNA polymerase (Qiagen) with the primers. Three sets of primers were used to amplify three "CACCC" sites of the VEGFA promoter region. ChIP assay real-time PCR results indicated that KLF8 binds to the "CACCC" site 637 nucleotides upstream of the VEGFA promoter region. Therefore, we used the 1386-5′GCTGTTTGGGAGGTCAGAAATAGG 3′-1409 and 1545-5′ ACGCTGCTCGCTCCATTCAC 3′-1526 primers; in addition, we used normal rabbit IgG as a negative control. pcDNA3.1-transfected SMMC7721 cells were used as a control group. Western blotting and immunohistochemistry staining. Total protein was prepared from the cell lines. Immunoblot experiments were performed according to standard procedures, and the following antibodies were used were used for the immunocytochemistry analysis: mouse anti-human monoclonal KLF8 (1:1000; Abnova); rabbit anti-human multiclonal P-c-Raf(Ser259), P-GSK-3β(Ser9), P-PTEN(Ser380), P-PDK1(Ser241), P-AKT(Thr308), P-AKT(Ser473), and AKT(pan) (Cell Signaling Technology); rabbit monoclonal anti-human VEGFA, mouse monoclonal anti-human GAPDH and mouse anti-human monoclonal KLF8 (1:1000; Abnova); anti-human focal adhesion kinase (FAK) (1:3000); and rabbit monoclonal anti-human VEGFA.
Chick Chorioallantoic Membrane (CAM) and Nude Mice Tumor Growth Models. All animal procedures were conducted in accordance with international standards and were approved by the Commission for Ethical Experimentation on Animals of Chongqing Medical University. Fresh fertile eggs were cleaned using a 1% solution of geramine and then incubated at 37.8 °C and 60%-80% relative humidity for 7 days. The shell was cut to create a small window (10 × 10 mm2), and the shell membrane was removed with sterile forceps to create an air chamber. The eggs were then sealed with sterile medical tape and incubated again. After 24 h, SMMC7721 cells were mixed with 100 µl of RPMI 1640 and 100 µl of Matrigel. When the mixture was nearly frozen, the cells were inoculated directly into the air chambers of the CAMs. The eggs were sealed with sterile medical tape and incubated at 37.8 °C and 60% humidity. After 120 h of incubation, the CAMs were fixed in methanol and acetone (1:1 volume) for 15 min. Then, they were cut and spread in distilled water. The status of the CAMs and the tumors (a 2-mm diameter was positive) could then be observed. The CAM vasculature was photographed using a scanner. The images were analyzed using an image analysis program, IPP, and the blood vessel density was calculated. A total of 5 × 10 6 SMMC7721 cells transfected with pcDNA3.1-KLF8 or pcDNA3.1 were inoculated into the liver parenchyma of nude mice under ketamine/xylazine anesthesia after the abdomen was opened. All mice were monitored once every 3 days and sacrificed 5 weeks later. Tumor tissue sections were prepared, and immunoreactivity was analyzed as above using KLF8, VEGF and CD31 antibodies (BD PharMingen).