Materials and methods
G- actin/F-actin activity assay
14. Corosolic acid does not exhibit significant inhibitory effects on Huh7 cell invasion
The matrix metalloproteinases (MMPs) are very important factors on cancer migration or metastasis [67]. To examine whether corosolic acid (CA) could inhibit
44
invasion activity of Huh7 cells, studies on the effects of CA for MMPs and NF- kappa B pathway, which is an important event to regulate the MMPs activity were evaluated.
However, in our model, corosolic acid (CA) had no significant inhibitory effect on Huh7 cell invasion (Figure 46A). The expression level of MMP2 and MMP9 and the
activity of MMP1, MMP2, and MMP9 were not affected by CA treatment (Figure 46B).
The level of phosphorylated IκB, IκB, and NFκB was maintained at a stable level (Figure 47) which suggested that NFκB pathway may not participate in CA effect.
45
Discussion
§Part I
Role of L-FABP in hepatocellular carcinoma
HCC is characterized by the high aggressive and angiogenic capacities, and the angiogenic factor, VEGF, has been considered as one of investigated targets for cancer therapy in HCC [1, 6]. We reported here for the first time that L-FABP overexpression plays an important role in VEGF-A expression and cell migration in HCC, and
demonstrates that L-FABP associates with VEGFR2 in cell membrane, following by the activation of VEGFR2 related signaling, including Src/ FAK/cdc42 and
Akt/mTOR/HIF-1α. T94A mutation of L-FABP, which was related to the cholesterol binding activity, significantly decreased the angiogenic potential and migration activity of L-FABP overexpressed cells.
It has been suggested that L-FABP promotes growth of hepatocyte and protects cells from ROS by its anti-oxidative activity, which was related to the methionine and cysteine [68, 69]. Other studies also found the several lines of evidences in correlation of L-FABP and VEGF [12, 19]. However, the link of L-FABP and tumor malignance still remains unclear. In the present study, we found a significant increase of L-FABP expression in tumor part versus their NAT part in 90 HCC patients (p=0.012) by IHC staining. The well correlation between the expression level of L-FABP and VEGF-A in
46
90 clinical tissue pairs of HCC patients was also demonstrated (r=0.737, p<0.01). By screening of liver cell lines of, we also found that L-FABP expressed higher in malignant HCC cell lines, HepG2 and Huh7, but lower expressed in immortalized normal hepatocyte, Hus cells, and the tendency was consistent with that of VEGF-A expression level. Thus, it strongly suggests that L-FABP may regulate VEGF expression in HCC. Furthermore, it was suggested that we generated stable clones of Hus/L-FABP cells, and that the up-regulated VEGF-A expression level and angiogenic potential of Hus/L-FABP cells were observed by in vitro and in vivo studies. These observations were also further proofed by L-FABP knockdown in Huh7 cells and Hus/L-FABP cells (Figure, 24C and 23B). Previous study has suggested that VEGF is essential for HCC cell migration [63], therefore, we have observed that the migration activity of
Hus/L-FABP cells increased significantly than that of control cells. Knockdown of L-FABP in Huh7 cells or L-FABP stably expressed Hus cells also showed a decreased migration activity compared with that of control group (Figure, 24D and 23C). Taken together, these results suggested that L-FABP overexpression plays a critical roles in the angiogenic potential and migration activity of HCC cells, which could be reversely regulated by RNA knockdown technology.
In previous study, L-FABP has been suggested to be interacted with cell membrane
47
[11], however, most studies focused on its biological function in transport fatty acids and the regulation to lipid metabolism [70]. However, in this study, we found that L-FABP co-localized with VEGFR2 on membrane rafts of L-FABP overexpressed cells.
Previous study has reported that L-FABP co-expressed with VEGF in cell membrane [19]. Other studies also suggested that lipid rafts seemed to be capable of acting in signaling platform [71-73]. Comply with this, our confocal microscopy analysis suggested the co-localization of L-FABP and VEGFR2 on apical membrane of
Hus/L-FABP and Huh7 cells (Figure 7 and Figure 25). The downstream signal proteins including Src/ FAK and PI3K/Akt showed an increased level in membrane fraction.
Knockdown of VEGFR2 in Hus/L-FABP cells decreased the phosphorylation level of these downstream signal molecules (Supplementary, Figure 1). Moreover, by protein docking software, we predicted two possible interacting model of L-FABP and VEGFR2 kinase domain (Supplementary, Figure 2). As a result, our observation provides a possible mechanism of how L-FABP activates VEGFR2 signaling.
The regulation of VEGF in HCC has been highlighted since its related pathway plays an important role in cancer progression [2]. In fact, only the anti-VEGFR2 therapy revealed a significant benefit on clinical HCC patients, and was approved by FDA [9].
In our experiment, we found that the increased VEGF-A expression was via translation
48
regulation of PI3K/Akt and its downstream mTOR/P70S6K/4EBP1 pathway. Since VEGF-A could be regulated by HIF-1α in transcription level of cancer cells, in our model, both mRNA level and transcriptional activity of VEGF-A showed a significant up-regulation by L-FABP overexpression. Interestingly, previous study also showed that L-FABP revealed a positive correlation with VEGF-A in mRNA level [12]. Taken together, these data suggested the possible mechanism which regulates VEGF-A expression in HCC cells.
L-FABP is the only member of mammalian FABP family to transfer fatty acids to membranes by aqueous diffusion [74], furthermore, direct interaction of L-FABP and PPARα has also been reported for ligand trafficking to nucleus [75]. Therefore, the studies on ablation or mutation of L-FABP protein in normal hepatocyte has been studied for a long time. In L-FABP knockout mice, it showed decreased lipid metabolism and exacerbated obese phenotype with high-fat diet [70, 76]. For the mutation studies, L-FABP (F3W) and (K31E) mutants showed a significance decreased binding ability to phospholipid [57, 58]. Moreover, L-FABP (T94A) mutant altered structure and stability of L-FABP and caused a loss-of-function [59-61]. In present studies, we have mutated four amino acids which located in different domains of L-FABP protein (Supplementary, Figure 3): F3 to W (β sheet A, N-terminal), K20 to E
49
(α-helix I), K30 to E (α-helix II), T94 to A (β sheet G, C-terminal) to examine the functional amino acids of L-FABP in L-FABP induced angiogenesis and migration. The results demonstrated that a decreased level of VEGF in L-FABP K31E and T94A
mutants, and a significant down-regulation in migration activity. The treatment of MβCD, a membrane cholesterol depletion agent [77] with Hus/L-FABP cells also
support the above-mentioned observation, and it suggests that the cholesterol-binding activity of L-FABP is indispensable to its function. Taken together, the function of L-FABP in cell membrane not only for metabolism, but also for its oncogenic role in HCC tumorigenesis.
Previous studies reported that L-FABP promoted diet induced fatty liver disease and hepatic steatosis [78]. It also suggested that VEGF level was correlated with HCC malignance and poor prognosis [79]. In the present study of clinical sample data, we found that L-FABP up-regulated significantly in HCC patients, with and without cirrhosis. Moreover, in the cirrhosis patients, high L-FABP expression indicated high risk and poor survival time (Figure 26). Previous study suggested that it needs
“angiogenic switch” to become a solid HCC tumor [80], and VEGF showed an autocrine feed-forward loop to trigger angiogenesis [55, 81], Since the correlation of L-FABP expression and HCC progression remains unclear, and there was no appropriate
50
prognosis marker in HCC with cirrhosis [82], L-FABP may serve as a potential research target for further studies.
§Part II
Effects of corosolic acid on hepatocellular carcinoma
CA is an ursane-type triterpenoid, and is known to be a STAT3 inhibitor in
macrophages, myeloid cells, and ovarian cancer cells [30-32]. CA also has a significant inhibitory effect on endothelial angiogenic tube formation [29], and tumor growth in lung and ovarian cancer cells [31, 33]. In this study, we found that CA significantly reduced the migration activity of HCC cells, including Huh7, HepG2 and Hep3B at a low-cytotoxicity dosage. When combined with sorafenib, CA showed synergistic effects on HCC cell growth and migration. An in vivo xenograft mouse model was used to verify the anti-HCC activity of CA, which showed significant inhibitory effects on Huh7 cells at 5 mg/kg/day.
VEGFR2 is the major receptor in the VEGF signaling pathway that regulates cell migration, proliferation, and angiogenesis. This study revealed that CA reduces the tyrosine phosphorylation level of VEGFR2, with an IC50of kinase activity of 0.95 μM.
Further studies also found that CA suppressed the activation of Src, FAK, and cdc42.
51
These results provide a potential mechanism for the anti-migration effects of CA on Huh7 cells in HCC.
The inhibition of VEGFR2 has been proposed as a novel therapeutic strategy for HCC patients. Various VEGFR2 kinase inhibitors such as sorafenib, sunitinib, and linifanib were developed and used in clinical trials. Recently, anti-HCC therapy with sorafenib has been approved by FDA [9, 64]. To further investigate how CA inhibits VEGFR2, a structure-based interaction model between CA and VEGFR2 was developed by molecular docking analysis. The results suggest that the ATP binding pocket in the VEGFR2 catalytic domain binds CA with lower binding energy than ATP (-15.2 kcal/mol versus -12.3 kcal/mol). Moreover, the surface charge distribution of VEGFR2 demonstrated that the OH groups of CA showed stable interactions with the ATP binding pocket. It also revealed that most uncharged areas of CA could generate
hydrophobic forces with valine and cysteine resulting in stabilizing the binding affinity.
This strongly suggests that the binding of CA to the ATP-binding pocket of VEGFR2 mediates the down-regulation of VEGFR2 phosphorylation and subsequent signals.
Furthermore, the combination of CA and sorafenib had significant synergistic effects on Huh7 cell migration and VEGFR2 phosphorylation. The in vivo combinatorial
experiment further verified that CA combined with sorafenib shows potential for HCC
52
treatment without toxic effects to mice (data not shown). We also observed that CA down-regulated the phosphorylation level of Src and FAK kinases when combined with sorafenib, since sorafenib alone did not show any inhibitory effect to the activation of FAK kinase in the xenograft model. Collectively, these results indicate that CA shows potential as a novel VEGFR2 inhibitor or an adjuvant therapy to be used with existing anti-cancer drugs.
Previous studies have discussed the pharmacophore modeling of different VEGFR2 inhibitors [66]. These inhibitors could be divided in two types, sunitinib-like or
sorafenib-like, depending on the interacting hydrogen bonds. The binding of type I inhibitor (sunitinib) formed hydrogen bonds with Asp1044, Cys917, and Asn921 near the protein surface. On the other hand, the type II inhibitor (sorafenib) could interact with Asp1044, Cys917, and Glu883. By docking analysis, we found that CA formed hydrogen bond and relatively closed to Glu883 than Asn921 (2.67 Å versus 9.2 Å, Figure 42). Although the interaction model of CA with VEGFR2 are likely to sorafenib, however, the chemical structure of CA varied widely with both two types of VEGFR2 inhibitor. Thus, it could be interesting to explore and design novel VEGFR2 inhibitors based on present findings.
53
Summary
We have been focusing on finding out novel oncogenic mechanisms and therapeutic agents of HCC, and the first study reveals for the first time that L-FABP potently induces the up-regulation of VEGF-A and increases angiogenic potential and migration activity in HCC cells. The results also suggest that the function of L-FABP in HCC could be influenced by mutations in its cholesterol interaction sites. When considered alongside previous studies, our findings indicate that L-FABP is a potential therapeutic target in HCC therapy. Next, we demonstrates that corosolic acid could be a potential anti-HCC agent. We provide evidence that CA’s anti-cancer effects stem from its anti-migratory effect, by blocking the VEGFR2 ATP binding pocket and
down-regulating the downstream Src/FAK/cdc42 signaling axis. This study further demonstrates that the combination of CA and sorafenib may have potential as a chemotherapy for HCC.
54
Figure 1. Correlation between the expression levels of L-FABP and VEGF-A
L-FABP and VEGF-A expression in 90 cases of HCC patients (normal and tumor paired tissue) was examined by IHC staining. (A) Representative images of different
expression levels of HCC tissue pairs. a-d: Staining of L-FABP was observed in tumor parts (a and c) and their normal adjacent tissues (b and d). a: Strong staining; b and c:
moderate staining; d: weak staining of L-FABP IHC results. e-h: Staining of VEGF-A was observed in tumor parts (e and g) and their normal adjacent tissues (f and h). e:
Strong staining; f and g: moderate staining; h: weak staining of VEGF-A IHC results. (B) Correlation between L-FABP and VEGF-A expression in 90 HCC tissues (with and without cirrhosis). L-FABP exhibited a positive correlation with VEGF-A by the Pearson correlation coefficient (r = 0.737, **p < 0.01).
(A)
(B)
55
Figure 2. L-FABP expression is associated with VEGF-A expression of HCC cells.
(A) Western blot analysis for L-FABP expression in normal immortalized hepatocyte (Hus) and hepatocellular carcinoma cell lines (HepG2, Hep3B, Huh7 and PLC/PRF/5).
(B) Angiogeic potential of Hus, HepG2, Hep3B, Huh7 and PLC/PRF/5 cells was
assessed by HUVEC endothelial cell tube formation assay. *p < 0.05, **p < 0.01 versus control group (Hus cells).
(A)
(B)
56
Figure 3. Expression level of VEGF-A is up-regulated Hus cells stably expressed L-FABP.
L-FABP was stably expressed in Hus cell using pcDNA3.1 expression system. The vector only cells were used as control group. (A)The protein expression level of L-FABP and VEGF-A was analyzed by western blotting. (B)The mRNA expression level of VEGF-A was determined by qRT-PCR. *p < 0.05, versus control group (Hus/Vector cells).
(A) (B)
57
(A)
(C)
(B)
Figure 4. L-FABP promotes in vitro and in vivo angiogenic activity of Hus cells.
(A) The in vitro angiogenic activity was studied by tube formation assay which performed by HUVEC endothelial cells to determine angiogenesis activities of Hus/Vector and Hus/L-FABP cells. Angiogenic vascular tube was imaged at 8 h. The quantification of S.CORE tube formation was shown as panel bar. ***p < 0.001 versus control group (Hus/Vector cells). (B) The in vivo angiogenic activity was studied by matrigel plug in assay. Left: Macroscopic view of matrigel plugs recovered from mice injected with Hus/Vector and Hus/L-FABP cells, and the infiltration of blood vessels were indicated by arrows. (C) Immunohistochemical staining of CD31 (angiogenesis marker) in matrigel plugs were presented and quantified. n=3, *p < 0.05 versus control group (Hus/Vector cells).
58
Figure 5. Sequence aliment of L-FABP interacting domains
Amino acid sequence alignment of L-FABP interacting domains, including: CD36 TSP binding domain, DSLR cytoplasmic domain, integrin α1 cytoplasmic domain, and integrin α2 cytoplasmic domain. Strictly conserved residues are highlighted in blue;
residues with similar property are highlighted in green, respectively.
59
(A) (B)
Figure 6. Co- immunoprecipitation of L-FABP and VEGFR2 in Hus/L-FABP cells (A) The cell lysates of Hus/Vector and Hus/L-FABP cells were subjected to
immunoprecipitation (IP) with VEGFR2 antibody, followed by blotting with L-FABP;
or L-FABP antibody, followed by blotting with VEGFR2. (B) Cell lysates (50 μg) were immunoblotted as input control.
60
(A)
(B)
Figure 7. L-FABP associates with VEGFR2 in apical membrane of Hus/L-FABP cells
(A) Cells were fixed and stained with antibodies against to VEGFR2 and L-FABP.
Three-color confocal images were acquired on a ZEISS, LSM 510 META Confocal Microscope (Magnification, 63 ×). (B) Red or green lines showed the X-Z or Y-Z optical section of Hus/Vector and Hus/L-FABP cells, respectively. The co-localization of VEGFR2 and L-FABP on the upside of cells was indicated by red arrows.
61
Figure 8. Localization of L-FABP and signaling molecules in lipid rafts
Membrane localization of L-FABP, VEGFR2, PI3K (p85), phospho-Akt (Ser473), Akt, phosho-Src (Tyr416), Src, FAK and phosho-FAK (Tyr397) in Hus/L-FABP or control cells. Membrane rafts were obtained by sucrose gradient based ultra-centrifugation and analyzed by western blot analysis (Fraction #3~#5).
62
Figure 9. L-FABP increases the phosphorylation level of VEGFR2 in Hus cells Phosphorylated levels of VEGFR2 in Hus/Vector and Hus/L-FABP cells were analyzed by immunoprecipitation (IP) of VEGFR2 antibody and blotted with phospho-tyrosine antibody. *p < 0.05 versus control group (Hus/Vector cells).
63
Figure 10. L-FABP increases the phosphorylation level of Src and FAK kinases in Hus cells
Phosphorylated level of Src (Tyr416) and FAK (Tyr397) in Hus/Vector and
Hus/L-FABP cells were analyzed by western blot analysis. **p < 0.01 versus control group (Hus/Vector cells).
64
Figure 11. L-FABP promotes cdc42 activity of Hus cells
Small GTPase binding assay was carried out to Hus/L-FABP or control cells. Active cdc42 and Rac1 were detected by western blot analysis, however, active RhoA was not detectable in this study. For cdc42 activity, ***p < 0.001 versus control group
(Hus/Vector cells).
65
(A)
(B)
Figure 12. Analysis of migration activity of L-FABP stably expressed Hus cells (A) Wound-healing migration assay of Hus/Vector and Hus/L-FABP were performed to examine two-dimensional migration activity and the migrated distance during the designated period was quantified. ***p < 0.001 versus control group (Hus/Vector cells).
(B) For studying three-dimensional migration activity, Hus/Vector and Hus/L-FABP were seeded onto Boyden chambers and allowed to migrate toward 10% serum containing medium for 16 h. ***p < 0.001 versus control group (Hus/Vector cells).
66
Figure 13. L-FABP up-regulates migration activity through VEGFR2/ Src pathway Hus/L-FABP cells were treated with PP1 (Src inhibitor; 5, 10 μM, respectively) or Sorafenib (VEGFR2 inhibitor; 1, 2, 4 μM, respectively) for 16 h and analyzed by transwell assay. ***p < 0.001 versus control group (DMSO only treated cells).
67
Figure 14. L-FABP activates Akt/ mTOR/ P70S6K/ 4EBP1 signaling
The phosphorylation level of Akt (Ser473), mTOR (Ser2448), P70S6K (Thr421/Ser424) and 4EBP1 (Thr37/46) in Hus/Vector and Hus/L-FABP cells were studied by western blot analysis. *p < 0.05, **p < 0.01 versus control group (Hus/Vector cells).
68
Figure 15. HIF-1α significantly increases in the nucleus of L-FABP overexpressed cells
Nucleus and cytoplasmic localization of HIF-1α in Hus/L-FABP cells was studied, and α-tubulin and lamin A/C were represented as loading controls for cytoplasmic and nucleus, respectively. Note that HIF-1α level was increased in Hus/L-FABP cells 1.7 fold higher than that of control group as the bar graph. *p < 0.05 versus control group (Hus/Vector cells).
69
Figure 16. Role of HIF-1α in VEGF-A transcriptional activity of L-FABP overexpressed cells
The diagrams of the receptor constructs of full length and various deletion mutants of VEGF-A promoter (D1-D3) were showed as the graph. The luciferase activity of cell extracts was analyzed by luciferase reporter assay, and the data were presented as bar graph. For comparing full length activity, ***p < 0.001 versus control group
(Hus/Vector cells); for deletion experiments, ***p < 0.001 versus control group (Hus/L-FABP cells); for adding HIF-1α inhibitor, ***p < 0.001 versus control group (Hus/L-FABP cells).
70
(A) (B)
(C)
Figure 17. Post-transcriptional regulation of VEGF-A in L-FABP stably expressed Hus cells
(A) Hus/L-FABP cells were treated with Rapamycin (mTOR inhibitor) or
cyclohexamide for 12 h and analyzed by western blot analysis. (B) On the other hand, Hus/Vector cells were treated with MG132 (proteasome inhibitor) for 24 h and analyzed by western blot analysis. (C) Cells were treated with Rapamycin or cyclohexamide for 12 h and the conditioned medium were subjected to tube formation assay to measure the in vitro angiogenic activity. Angiogenic vascular tube was imaged at 12 h and the quantification of S.CORE tube formation was shown as panel bar. ***p < 0.001 versus control group (DMSO only treated Hus/L-FABP cells).
71
(A)
(B) (C)
Figure 18. L-FABP promotes tumor growth in vivo
To study the effect of L-FABP on tumor growth, 2 × 106 of Hus/L-FABP or control cells were subcutaneously injected into the hind limb of NOD/SCID mice, and the resulting in situ tumors were removed 8 weeks later for analysis. (A) Representative photograph and average weight of tumors are presented. (n=5 per group). (B) VEGF-A contents in serum of the above mice were measured and presented in the bar graph. (C) The tumor sections analyzed by H&E staining (a & b) or anti-CD31 antibody IHC staining (c & d) indicated the strong angiogenesis activity in Hus/L-FABP mice group. Image a & c indicated Hus/Vector group; b & d indicated Hus/L-FABP group.
72
(A)
(B)
Figure 19. L-FABP promotes in vivo metastasis by lung metastasis model
The metastatic activity of Hus/L-FABP cells was carried out by lung metastasis model.
5 x 106 of Hus/L-FABP or control cells were intravenously injected into the lateral tail vein of NOD/SCID mice. After 10 weeks, the lungs were excised from each mice for analysis. (A) Metastatic nodules were presented and counted (n=5 per group). (B) The immunohistochemistry analysis by H&E staining (a, b, d and e) or anti-CD31 antibody IHC staining (c and f) were also studied. Image a – c indicated Hus/Vector group; d – f indicated Hus/L-FABP group. **P < 0.01 versus control group (mice injected with
5 x 106 of Hus/L-FABP or control cells were intravenously injected into the lateral tail vein of NOD/SCID mice. After 10 weeks, the lungs were excised from each mice for analysis. (A) Metastatic nodules were presented and counted (n=5 per group). (B) The immunohistochemistry analysis by H&E staining (a, b, d and e) or anti-CD31 antibody IHC staining (c and f) were also studied. Image a – c indicated Hus/Vector group; d – f indicated Hus/L-FABP group. **P < 0.01 versus control group (mice injected with