Hepatocellular carcinoma (HCC) is a typical hypervascular tumor, and increased levels of vascular endothelial growth factor (VEGF) are associated with progression of HCC. Tumor suppression gene PTEN (phosphatase and tensin homolog deleted on chromosome 10), an important antagonist of the phosphoinositide-3-kinase (PI3K)/adenosine triphosphate-dependent tyrosine kinase (Akt) pathway, is also commonly lost or mutated in HCC. However, the effect of PTEN on VEGF-mediated angiogenesis in HCC remains unknown. To explore this relationship, we expressed a panel of PTEN mutants in human HCC cells with low expression of PTEN (HepG2 cells). Overexpression of PTEN in HepG2 cells resulted in the downregulation of proliferation and migration of cocultured endothelial cells and decreased expression of hypoxia-inducible factor 1 (HIF-1) and VEGF. Similarly, using a nude mouse model, we demonstrated that PTEN decreased expression of HIF-1 and VEGF and suppressed HepG2-induced angiogenesis. This inhibitory effect was not observed in cells expressing a phosphatase-deficient PTEN mutant, suggesting that PTEN inhibits angiogenesis and VEGF through a phosphatase-dependent pathway. Strikingly, reintroducing the C2 domain of PTEN also resulted in a significant decrease in angiogenesis and VEGF expression, although it did not affect Akt phosphorylation or HIF-1 expression. In summary, this study suggests the novel viewpoint that PTEN suppresses angiogenesis and VEGF expression in HCC through both phosphatase-dependent and -independent mechanisms.
Hepatocellular carcinoma (HCC) is the most common liver malignancy and rates fifth in incidence and third in mortality in the world ( 1 ). Despite advances in HCC treatment, most patients with HCC die within 1 year after diagnosis, largely due to recurrence and metastases. HCC is a typical hypervascular tumor, and tumor angiogenesis is required for both growth and metastasis of HCC ( 2 , 3 ). Tumor angiogenesis begins with the activation of endothelial cells by a few specific angiogenic factors, among which vascular endothelial growth factor (VEGF) is a fundamental player ( 4 ). VEGF is a general activator of endothelial cell proliferation and mobility. Enhanced levels of VEGF are often observed in the liver and serum of HCC patients, and VEGF levels are useful for predicting overall survival and defining prognosis ( 5 ), suggesting that it plays an important role in HCC.
Tumor suppression gene PTEN (phosphatase and tensin homolog deleted on chromosome 10) is one of the most frequently lost or mutated genes in a variety of human cancers, including HCC ( 6–8 ). PTEN has been shown to exhibit many negative effects on cellular processes, such as growth, self-renewal, metastasis, etc. all of which can initiate and sustain a malignant phenotype ( 9–11 ). PTEN is composed of two major structural domains: the phosphatase domain and the C2 domain. The crystal structure of PTEN reveals an extensive and tight link between the phosphatase and C2 domains ( 12 ). As a dual specificity phosphatase, PTEN can dephosphorylate phosphatidylinositol 3,4,5-trisphosphate to phosphatidylinositol-4,5-bisphosphate, thus opposing the activity of phosphoinositide-3-kinase (PI3K)/adenosine triphosphate-dependent tyrosine kinase (Akt) signaling. Activation of PI3K/Akt modulates the activities of several important proteins involved in cell survival, proliferation and migration ( 13–15 ). Recently, some studies have implied that PTEN decreases secretion of VEGF by inhibiting phosphorylation of Akt and its downstream target, hypoxia-inducible factor 1 (HIF-1). Thus, it seems clear that PTEN exerts its tumor-suppressing effect, including its anti-angiogenic effect, by antagonizing the PI3K/Akt pathway. However, other potential mechanisms of action of PTEN have been identified. Growing evidence has revealed that some anti-oncogenic functions of PTEN, including regulation of cell migration and p53 expression, are independent of PTEN’s phosphatase activity ( 16–18 ). Furthermore, the C2 domain plays an essential role in PTEN nuclear localization and maintaining chromosomal integrity ( 19–21 ). Indeed, >40% of the mutations of PTEN have been mapped to the C2 domain ( 22 ). This suggests that PTEN may exert tumor suppression through its C2 domain via a mechanism independent of the PI3K/Akt pathway.
To evaluate the anti-angiogenic function and mechanism of PTEN in HCC, we generated a panel HepG2 cells expressing wild-type or mutant PTEN . Our in vitro and in vivo studies demonstrated that PTEN displays an inhibitory function in HCC angiogenesis through suppression of VEGF expression. We report that VEGF expression was dramatically reduced in HepG2 cells expressing PTEN and that PTEN inhibited VEGF by inactivating Akt and HIF-1α. Significantly, ectopic expression of the C2 domain of PTEN also resulted in a significant decrease in angiogenesis. Further, this mutant of PTEN was able to suppress VEGF expression, even in the absence of phosphatase activity, suggesting that PTEN regulates VEGF via a phosphatase-independent mechanism. Together, our data provide the first evidence that PTEN regulates angiogenesis and VEGF through both phosphatase-dependent and -independent mechanisms.
Materials and methods
The wild-type PTEN (PTEN) and the PTEN mutant construct with the cysteine 124 to serine mutation (PTEN.C124S) were gifts from Prof. Yu-Xin Yin (Columbia University, New York, NY). We used the PTEN plasmid as a template to polymerase chain reaction (PCR) amplify the fragment containing the PTEN C2 domain using an upstream primer 5′-TATGCCAGTGGCACTGTTGT-3′ and a downstream primer 5′-GTGTCAAAACCCTGTGGATGTA-3′. The fragment was subcloned into the pMD18-T Vector (TaKaRa, Dalian, China) to construct the PTEN.C2-18T plasmid. The PTEN C2 domain expression plasmid (PTEN.C2) was then generated by cutting the PTEN.C2-18T plasmid with XbaI and HindIII (TaKaRa) and subcloning the C2 domain into pcDNA3.1/Hygro (Invitrogen, Carlsbad, CA) cut with XbaI and HindIII. Plasmid sequences were confirmed by sequencing.
Cell culture and transfection
HepG2 cells, a human hepatoma cell line, expressing little to no PTEN protein ( 23 , 24 ) were cultured in RPMI-1640 medium containing 10% fetal bovine serum (FBS) in 5% CO 2 at 37°C. For pharmacological inhibition assays, HepG2 cells were switched to 2 ml RPMI-1640 containing 1% FBS and PI3K/Akt inhibitor LY294002 (10 or 20 μM, Sigma, Saint Louis, MO) at 60% confluence. After 48 h incubation, cell supernatant was harvested as conditioned medium (CM) and cell extracts were prepared as described in the following. Human umbilical vein endothelial cells (HUVEC) were obtained from the American Type Culture Collection (Rockville, MD) and grown on 1% gelatin-coated plates in M199 medium supplemented with 20% FBS, 100 IU/ml penicillin and streptomycin and 2 mmol/l L -glutamine at 37°C in a 5% CO 2 humidified atmosphere.
Stable transfections of PTEN, PTEN.C124S, PTEN.C2 or pcDNA3.1 empty vector were performed with Lipofectamine™ 2000 (Invitrogen) reagent following the manufacturer's guidelines. After transfection for 24 h, cells were diluted 1:20 and selected at 500 μg/ml of hygromycin (Sigma) for 1–2 months. Stable transfectants were confirmed by western blot and maintained in medium containing 250 μg/ml hygromycin.
Preparation of CM
CM was prepared as described previously with some modifications ( 25 ). Briefly, the various cell lines were seeded at a density of 1 × 10 5 cells/ml in six-well plates. One day later, cells were washed in phosphate-buffered saline two times and were switched to 2 ml RPMI-1640 containing 1% FBS. After 48 h incubation, CM was collected from the six-well plates with cell density of ∼80%. For the different treatment, the CM from transfected cells was preincubated with recombinant human VEGF 165 (20 ng/ml; R&D Systems, Minneapolis, MN) or anti-VEGF blocking antibody (50 ng/ml; Millipore, Danvers, MA) overnight at 4°C.
HUVEC proliferation inhibition assay
HUVECs were seeded at a concentration of 5000 cells per well in 1% gelatin-coated 96-well plates. At 80% confluency, the media was removed and various CMs were added to the wells. Relative cell numbers were quantified every day via [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] (MTT) assay. For each well, media was removed and 20 μl of 5 mg/ml MTT (Sigma) was added. After 4 h of incubation at 37°C, 150 μl dimethyl sulfoxide was added to each well and the absorbance was measured at 492 nm on a Multifunction Microplate Reader (POLARstar OPTIMA; BMG, Offenburg, Germany).
HUVEC cell cycle assay
HUVECs were cultured in 1% gelatin-coated six-well plates at 1 × 10 5 cells/ml. At 80% confluency, the media was removed and various CMs were added to the wells. After 48 h of incubation, the HUVECs were harvested and fixed in 70% alcohol for 30 min on ice. Cells were then stained with propidium iodide and RNase A (Sigma) at 37°C in the dark for 30 min. Cell cycle was assessed by flow cytometry (FACSCalibur; BD, Franklin Lakes, NJ) and the data was analyzed using the Cell Quest software.
HUVEC migration assay
Cell migration assays were performed with the modified Transwell system as described previously ( 26 ). Confluent HUVECs were trypsinized and resuspended in M199 medium containing 1% FBS. Cell suspensions (100 μl, 5000 cells) were seeded into Transwell inserts (8 μm pore; Millipore). Simultaneously, 500 μl of the various CMs (or M199 containing 20% FBS as a positive control) was placed in the lower chamber of the Transwell. The cells were allowed to migrate for 24 h at 37°C. The non-migrated cells were removed from the upper surface by scraping with a cotton swab. Cells that had migrated were stained with crystal violet, and the number of migrating cells per field was assessed by counting 10 random fields at ×200 magnification.
Quantification of VEGF in CM
VEGF concentrations in the CM were quantified using a VEGF enzyme-linked immunosorbent assay (ELISA) kit (Shanghai Senxiong Science and Technology Co., Ltd, Shanghai, China) according to the manufacturer's instructions.
Matrigel plug angiogenesis assay
Male BALB/c nude mice (4 weeks old) were purchased from the Shanghai Experimental Animal Center and maintained in the Laboratory Animal Center of Xi'an Jiaotong University, in accordance with the University Institutional Animal Care and Use Committee. Mice received food and water ad libitum .
Cells were cultured in fresh medium for 24 h and harvested, adjusting the cell concentration to 2.5 × 10 7 /ml in 50% mixture of Matrigel (Sigma) in serum-free RPMI-1640 medium. A volume of 0.2 ml of the mixture was injected subcutaneously into the flank of each mouse. Tumor dimensions were recorded every 3 days, starting with the first day of implantation, and tumor volumes were calculated using the equation ( l × w 2 )/2, where l and w refer to the larger and smaller dimensions collected at each measurement. Animals were killed 21 days after injection. The tumor xenografts were removed from the mice and immediately weighed. Next, the tumor xenografts were bisected. Half of each tumor was fixed in 4% paraformaldehyde overnight and analyzed by immunohistochemistry. The other half of each tumor was homogenized to generate lysates for reverse transcription–polymerase chain reaction (RT–PCR) and western blot analysis.
Real-time PCR assay for HIF-1α and VEGF
Real-time RT–PCR was performed to determine messenger RNA (mRNA) levels of HIF-1α and VEGF. Total mRNA was extracted using the TRIzol reagent (Invitrogen) and reverse transcription was performed using an RT–PCR kit (TaKaRa). Real-time experiments were conducted on an iQ5 Multicolor Real-Time PCR Detection System (Bio-Rad, Hercules, CA) using SYBR Green Real-time PCR Master Mix (TOYOBO, ShangHai, China). The PCRs consisted of 1 min at 95°C followed by 40 cycles of denaturation for 15 s at 95°C, annealing for 15 s at 55°C and a primer extension for 45 s at 72°C. The PCR primer sequences of HIF-1α, VEGF and glyceraldehyde 3-phosphate dehydrogenase are shown in supplementary Table S1 (available at Carcinogenesis Online). The comparative CT method was used to quantitate the expression of HIF-1α and VEGF using glyceraldehyde 3-phosphate dehydrogenase as a normalization control ( 27 ).
Cells and tumor tissues were lysed using cell lysis buffer [40 mmol/l Tris–HCl (pH 7.4), 10% glycerol, 50 mmol/l β-glycerophosphatase, 5 mmol/l ethyleneglycol-bis(aminoethylether)-tetraacetic acid, 2 mmol/l ethylenediaminetetraacetic acid, 0.35 mmol/l vanadate, 10 mmol/l NaF and 0.3% Triton X-100] containing protease inhibitors (Roche, Penzberg, Germany). Equivalent amounts of protein were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to Immobilon membranes (Millipore). The membranes were incubated with primary antibodies as follows: anti-PTEN antibody (28H6, 1:250), anti-Akt1/2 antibody (N-19, 1:500), anti-VEGF antibody (C-1, 1:200), anti-HIF-1α antibody (H1α 67, 1:200) and anti-β-actin antibody (C4, 1:800) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) and the anti-phospho-Akt (Ser473) antibody (193H12, 1:500) was obtained from Cell Signaling Technology (Danvers, MA). Blots were visualized with a secondary antibody coupled to horseradish peroxidase (Santa Cruz Biotechnology) and an ECL detection system (Millipore).
Immunohistochemistry and microvascular density assay
Paraffin embedded sections were deparaffinized, rehydrated, immersed in retrieval solution (citrate buffer) and blocked by 3% hydrogen peroxide. Sections were then incubated with primary antibodies against PTEN (28H6, 1:100), phosphorylated Akt (193H12, 1:50), HIF-1α (H1α 67, 1:100), VEGF (C-1, 1:100) and CD31 (M-20, 1:50, Santa Cruz Biotechnology) overnight at 4°C. Control sections were incubated with isotype-matched control antibody. The sections were stained with biotin-conjugated secondary antibodies and tertiary antibodies conjugated to streptavidin peroxidase. Immunoreactive products were stained with 3,3′-diaminobenzidine and subsequently counterstained with hematoxylin. The sections were then examined on a microscope (Q550CW; Leica, Manheim, Germany).
Microvascular density (MVD) was determined according to the criteria introduced by Weidner et al. ( 28 ). Briefly, the stained sections were screened at ×40 magnification under a light microscope (Leica; Q550CW) to identify the areas of highest CD31-positive vessel density. These areas were then counted at ×200 magnification in 10 random fields. Data was collected by two independent observers without knowledge of which tumors were viewed. The number of microvessels in each field was determined and expressed as the MVD.
All data are shown as mean ± standard deviation and were analyzed using the SPSS 13.0 software (Chicago, IL). Analysis of data was performed using One-way analysis of variance test and Dunnett's test. P < 0.05 was considered statistically significant.
PTEN suppressed the angiogenic ability of HepG2 cells in vitro
The growth and migration of endothelial cells are essential to physiological and pathological angiogenesis ( 29 ). To determine whether PTEN reduced HUVEC growth, we used CM from HepG2 cells expressing wild-type PTEN (PTEN), a phosphatase-deficient PTEN mutant (PTEN.C124S), the C2 domain of PTEN (PTEN.C2) or an empty vector (pcDNA3.1) to treat HUVEC cells and compared the effects of the various CMs on HUVEC growth. The M199 medium containing 20% FBS were used as positive control. No significant differences in HUVEC growth were observed after 1 day of exposure to the various CMs. However, the growth of HUVECs treated with CM from HepG2 cells expressing PTEN or PTEN.C2 was reduced by 57 and 45% after 7 days of culture, respectively. In contrast, treatment with CM from HepG2 cells expressing PTEN.C124S had no inhibitory effect on HUVEC growth after 7 days of culture ( Figure 1A ). Next, we analyzed the effect of the various CMs on HUVEC cell cycle progression by flow cytometry. The results indicate that HUVECs treated with CM from HepG2 cells expressing PTEN or PTEN.C2 accumulated significantly in the G 1 /G 0 phase. Specifically, within 48 h after treatment, 78.54 ± 1.96% and 75.77 ± 1.31% of the HUVECs treated with CM from HepG2 cells expressing PTEN and PTEN.C2 were in G 1 /G 0 phase, respectively. This was in contrast to the 67.73 ± 3.02% of the HUVEC cells treated with CM from HepG2 cells expressing pcDNA3.1 ( P < 0.01). Consistent with these data, there was a substantial reduction in the number of cells in S phase. Specifically, 11.18 ± 0.98% and 12.12 ± 1.15% of the HUVECs treated with CM from HepG2 cells expressing PTEN and PTEN.C2, respectively versus 18.96 ± 4.38% of HUVECs cultured with CM from HepG2 cells expressing pcDNA3.1. However, we observed no difference between the proportions of cells in G 2 /M phase between the treatments ( Figure 1B and C ). The observed increase of G 1 /G 0 phase cells and the decrease of cells in S phase might be the result of inhibition of HUVEC proliferation.
To investigate whether PTEN modulated endothelial cell migration, we assessed the ability of HUVECs to migrate toward various CMs in a lower chamber. The M199 medium containing 20% FBS were used as positive control. HUVECs migration toward CMs from HepG2 cells expressing PTEN and PTEN.C2 was markedly inhibited ( P < 0.01). In contrast, CM from HepG2 cells expressing PTEN.C124S had no effect on HUVEC migration ( Figure 1D and E ). Importantly, treatment of HUVEC with various CMs up to 24 h did not exhibit any significant change in cell viability ( Figure 1A ), suggesting that the inhibition of HUVECs migration mediated by full-length PTEN and the C2 domain of PTEN was due to decreased motility and not a growth inhibitory effect. Together, our findings showed that PTEN and the C2 domain of PTEN can inhibit the angiogenic response of HepG2 cells in vitro , as indicated by reduced endothelial cell growth and migration.
Full-length PTEN and the C2 domain of PTEN suppress the expression of VEGF
VEGF, a secreted protein, is a crucial angiogenic factor involved in blood vessel formation during tumor development responsible for regulation of endothelial cell survival, proliferation and migration ( 30 , 31 ). We sought to determine whether PTEN inhibited endothelial cell growth and migration by suppressing VEGF secretion. We used ELISA assays to assess VEGF expression in various CMs. The ELISA data showed that HepG2 cells and HepG2 cells containing the empty vector secreted 2138.03 ± 75.31 pg/ml/10 5 cells and 1728.72 ± 220.89 pg/ml/10 5 cells of VEGF, respectively. Expression of PTEN significantly reduced VEGF levels to 925.59 ± 73.00 pg/ml/10 5 cells ( P < 0.01). Similarly, the levels of VEGF were lower in the CM from HepG2 cells expressing PTEN.C2 (1189.59 ± 61.73 pg/ml/10 5 cells, P < 0.05). Consistent with previous results, VEGF levels were not inhibited in the CM from HepG2 cells expressing PTEN.C124S (1670.86 ± 78.83 pg/ml/10 5 cells) ( Figure 2D ). Western blot analysis of VEGF expression in the various cells confirmed these results ( Figure 2C ).
RT–PCR and real-time PCR assays were performed to determine whether PTEN suppressed VEGF secretion by inhibiting the VEGF mRNA. Expression of PTEN or the C2 domain of PTEN in HepG2 cells reduced VEGF expression, compared with the expression of parental HepG2 cells and HepG2 cells expressing PTEN.C124S (shown as 409 bp and 541 bp bands, indicating the VEGF 121 isoform and VEGF 165 isoform, respectively, Figure 2A ). Real-time PCR analysis was used to quantitate the expression of target genes and the results showed that VEGF mRNA in HepG2 cells expressing PTEN or PTEN.C2 decreased by 62 and 51%, respectively, compared with parental HepG2 cells ( Figure 2B ).
To investigate whether the reduction in VEGF expression by PTEN or C2 domain of PTEN contributed to the decreased angiogenic ability of HepG2 cells, we added VEGF 165 or anti-VEGF blocking antibody into CM. As expected, VEGF 165 enhanced cell proliferation compared with controls ( Figure 3A ); however, the proliferative response was markedly blunted when cells were cultured in CMs by addition of anti-VEGF blocking antibody ( Figure 3B ). Similarly, VEGF 165 enhanced the migration of HUVECs ( Figure 3C ); however, the HUVECs migration-stimulating activity of CMs was significantly reduced by addition of anti-VEGF blocking antibody ( Figure 3D ). It is notable that the increase of HUVECs proliferation and migration by addition of VEGF 165 were more significant in cells expressing PTEN or PTEN.C2 and reduction of HUVECs proliferation and migration by anti-VEGF blocking antibody were more dramatic in cells expressing empty vector or PTEN.C124S. These data support the notion that the inhibition of HUVECs growth and migration mediated by PTEN or the C2 domain of PTEN is attributable to suppression of VEGF expression.
PTEN regulates VEGF expression through phosphatase-dependent and -independent mechanisms
Since several studies have reported that PTEN-mediated inhibition of VEGF expression was due to effects on the PI3K/Akt/HIF-1/VEGF signaling pathway ( 32–35 ), we examined the activation of PI3K/Akt signaling and its downstream targets. Decreased Akt phosphorylation and HIF-1α expression were only observed in the HepG2 cells expressing full-length PTEN, but not in the other cell treatments ( Figure 2 ). The data showed that exogenous PTEN expression resulted in decreased Akt phosphorylation and downregulation of expression of HIF-1α and VEGF. However, this effect was clearly decreased in HepG2 cells expressing PTEN.C124S. Furthermore, we used the highly specific PI3K inhibitor LY294002 to assess the role of the PI3K/Akt pathway in angiogenesis. As expected, cell proliferation and migration were significantly inhibited when HUVECs were cultured in CM by addition of LY294002 ( Figure 4A and B ). Meanwhile, RT–PCR, western blot and ELISA analysis indicated that LY294002 was able to decreased Akt phosphorylation, HIF-1α and VEGF expression in a concentration-dependent manner ( Figure 4C–E ). Thus, it is not surprising that PTEN inhibits VEGF expression through the PI3K/Akt/HIF-1 pathway.
To explore whether other regions of PTEN could have a similar inhibitory effect on VEGF expression, the C2 domain mutant, which lacked the phosphatase domain, was tested in the same assay. Surprisingly, the results of this assay showed that the C2 domain of PTEN could also suppress VEGF mRNA and protein expression, although it had no effect on Akt phosphorylation or HIF-1α expression ( Figure 2 ). This data indicates that PTEN inhibits VEGF, at least in part, by a mechanism independent of its phosphatase activity.
PTEN inhibits tumor xenograft growth and tumor-induced angiogenesis in vivo
We subsequently conducted in vivo experiments to further confirm the findings of our in vitro experiments. A Matrigel plug assay, in which transfected cells were subcutaneously injected into the flank site of BALB/c nude mice, was conducted to monitor tumor growth in an in vivo system. We observed a dramatic difference in the growth of tumors derived from cells expressing full-length PTEN or the C2 domain of PTEN, compared with tumors derived from cells transfected with the empty vector ( Figure 5A ). The average volume of the tumor xenografts derived from HepG2 cells transfected with empty vector 21 days after implantation was 864.91 ± 669.55 mm 3 compared with 162.00 ± 72.56 mm 3 and 210.83 ± 159.82 mm 3 for tumor xenografts derived from HepG2 cells expressing PTEN and PTEN.C2, respectively ( P < 0.05). Similarly, the weights of the tumor xenografts derived from HepG2 cells expressing PTEN and PTEN.C2 21 days after inoculation were markedly decreased ( Figure 5B ). However, the volume and weight of tumors derived from HepG2 cells expressing PTEN.C124S were not decreased 21 days after implantation. Together, these data support the hypothesis that PTEN and the C2 domain of PTEN can inhibit tumor xenograft growth in vivo .
As PTEN regulates many aspects of tumor biology, including angiogenesis, experiments were conducted to determine whether PTEN could control over-angiogenesis in this model. MVD of tissue sections is considered an index of neovascularization. We stained paraffin sections from tumor xenografts for CD31 ( Figure 5D ), a specific endothelial cell marker that has commonly been used for microvessel quantification in tumors ( 36 ). Quantification of MVD in our tumor xenograft samples indicated that samples derived from cells expressing PTEN and the C2 domain of PTEN revealed ∼63 and 55% suppression of angiogenesis, respectively ( Figure 5C ; P < 0.05). The MVD of tumors derived from HepG2 cells expressing PTEN.C124S was not decreased in comparison with controls ( Figure 5C and D ). These findings suggest that the inhibition of tumor xenograft growth mediated by PTEN and the C2 domain of PTEN might be due to suppression of angiogenesis.
We assessed the effect of PTEN on VEGF expression and signaling through the PI3K/Akt/HIF-1 cascade in these mice tumor xenografts. RT–PCR analysis demonstrated that expression of PTEN and the C2 domain of PTEN, but not the phosphatase-deficient PTEN, decreased the mRNA level of VEGF in vivo ( Figure 6A ). Western blot analysis of VEGF expression in these tumor xenograft samples confirmed these results ( Figure 6B ). To confirm these results, we performed immunohistochemistry analysis in the tumor xenograft samples ( Figure 6C ). We noted that PTEN and the C2 domain of PTEN could suppress expression of VEGF, as assessed by VEGF staining ( Figure 6C ). These findings indicate that both full-length PTEN and the C2 domain of PTEN inhibit angiogenesis through suppression of VEGF. Next, we examined Akt phosphorylation and HIF-1α expression in the tumor xenografts. Wild-type PTEN, which suppressed the activation of phospho-Akt, suppressed HIF-1α expression, whereas expression of PTEN.C124S had no effect on phospho-Akt or HIF-1α expression ( Figure 6B and C ). Our results suggest that PTEN may regulate angiogenesis through the control of the Akt/HIF-1α pathway. However, it is quite striking that expression of the C2 domain of PTEN also inhibited angiogenesis and VEGF expression in vivo , although the C2 domain did not suppress Akt phosphorylation or HIF-1α expression ( Figure 6 ). Together, our data demonstrate that PTEN could inhibit VEGF through multiple mechanisms, one of which does not require phosphatase activity.
The development of anti-angiogenic therapy as an effective anticancer treatment has greatly increased the value of understanding the mechanisms regulating angiogenic mediators such as VEGF. Recent studies have shown that overexpression of PTEN can inhibit angiogenesis and expression of VEGF ( 32 , 33 ). However, the role and mechanism of PTEN in HCC angiogenesis have remained undetermined. In this study, we explored the role of PTEN in the regulation of angiogenesis in HCC using HepG2 cells stably transfected with wild-type PTEN plasmid and two mutants expressing phosphatase-deficient PTEN and the PTEN C2 domain. We found that expression of wild-type PTEN or the C2 domain of PTEN in HepG2 cells significantly inhibited angiogenesis both in vitro and in vivo . In addition, expression of both forms of PTEN resulted in downregulation of expression of both VEGF mRNA and protein. We also found that PTEN suppressed angiogenesis and VEGF expression through both phosphatase-dependent and -independent pathways.
As an important antagonist of the PI3K/Akt pathway, PTEN leads to activation of Akt and its downstream target genes. Much evidence has shown that activation of Akt correlates with increased expression of VEGF and overexpression of PTEN can inhibit physiological and pathological angiogenesis ( 33 , 34 ). HIF-1, a key transcription factor that is central to oxygen homeostasis, activates the transcription of many angiogenesis-related genes, including VEGF ( 37 , 38 ). HIF-1α, a component of HIF-1, can be induced by hypoxia or loss of tumor suppressors, including PTEN , as well as by the activation of PI3K/Akt ( 39 ). In this study, introduction of PTEN into HepG2 cells inhibited proliferation and migration of endothelial cells and led to decreased Akt phosphorylation and low expression levels of HIF-1α and VEGF. Additionally, we used a Matrigel plug model to verify the PTEN-mediated suppression of VEGF-induced angiogenesis in vivo . Based on CD31-immunostaining of endothelial cells, we found that the number of neomicrovessels decreased significantly in tumor exografts derived from HepG2 cells expressing PTEN. RT–PCR, western blot and immunohistochemical staining of the tumor xenographs indicated that expression of PTEN decreased HIF-1α and VEGF mRNA and protein levels. This finding confirmed our in vitro results that indicate that overexpression of PTEN inhibits angiogenesis through HIF-1α and VEGF expression. To determine whether the phosphatase activity of PTEN is required for inhibition of angiogenesis, we introduced PTEN.C124S into HepG2 cells. Expression of this phosphatase-deficient form of PTEN did not show an inhibitory effect on angiogenesis and did not inhibit HIF-1α or VEGF expression in vitro or in vivo . Furthermore, we observed that PI3K inhibitor LY294002 exerted an inhibitory effect on angiogenesis and inhibit HIF-1α or VEGF expression. Consistent with previous results, our data suggest that PTEN inhibits VEGF expression through the PI3K/Akt/HIF-1α pathway.
However, the Akt and HIF-1α-dependent pathway is probably not the only mechanism through which PTEN inhibits angiogenesis. Given that somatic PTEN mutations also occur in the C2 domain in many tumor types, including HCC ( 6 , 40 ), we hypothesized that PTEN could also suppress angiogenesis through its C2 domain. Thus, we introduced the PTEN.C2 plasmid into HepG2 cells and performed the same assay to evaluate angiogenesis. We discovered a phosphatase-independent role of PTEN involved in controlling VEGF expression. Our conclusions are based on two points: (i) the C2 domain of PTEN exerted a similar inhibitory effect on angiogenesis in vitro and in vivo , even though it did not affect Akt phosphorylation, a hallmark of PTEN's phosphatase activity and (ii) HIF-1α mRNA and protein levels did not decrease when HepG2 cells were transfected with PTEN.C2. Since HIF-1α is a key mediator of PI3K/Akt regulation of VEGF expression, as suggested by previous studies, our results indicate that PTEN inhibits VEGF, at least in part, through a currently unknown mechanism that is independent of the PI3K/Akt/HIF-1α pathway.
However, a confusing problem emerged, in that phosphatase-deficient PTEN containing an intact C2 domain did not suppress VEGF expression or angiogenesis. Recent work support a hypothesis that dephosphorylation of the C-terminal phosphorylation sites, particularly Thr 383 , was required for the inhibitory effect of the C2 domain to be exerted in the full-length protein ( 16 ). This process might be mediated by an intra-molecular auto-dephosphorylation of PTEN or by an unknown protein phosphatase. The fact that the phenotype of the C124S mutant-expressing cells could be partially recovered by mutation of the phosphorylation sites supports this hypothesis ( 16 , 17 ). We assessed the effect of dephosphorylation of the C-terminal of PTEN on VEGF expression in vitro ; however, we did not get similar results (data not shown). Therefore, the answer to the complicated and interesting puzzle should be investigated further.
Significantly, it is worth noting that the HepG2 cells expressing PTEN.C2 showed an intermediate level of angiogenesis suppression, suggesting that the anti-angiogenic role of PTEN may be the result of synergy between the phosphatase domain and the C2 domain. However, the mechanism by which PTEN modulates VEGF expression through a phosphatase- and HIF-1α-independent pathway is currently unknown and should be investigated further.
Taken together, our study provides both in vitro and in vivo evidence that PTEN inhibits angiogenesis by suppressing VEGF expression. Moreover, our study suggests a novel hypothesis that PTEN controls VEGF expression through both phosphatase- and HIF-1α-dependent and -independent mechanisms.
adenosine triphosphate-dependent tyrosine kinase
enzyme-linked immunosorbent assay
fetal bovine serum
hypoxia-inducible factor 1
human umbilical vein endothelial cells
phosphatase and tensin homolog deleted on chromosome 10
polymerase chain reaction
reverse transcription–polymerase chain reaction
vascular endothelial growth factor
We thank Prof. Yu-Xin Yin (Columbia University, New York, NY) for the PTEN and PTEN.C124S plasmids and his useful comments. We are grateful to Dr Wei-Min Liu (Xi'an Jiaotong University, Xi'an, China) for his help with HUVEC culture.
Conflict of Interest Statement: None declared.