Protein Kinase C Lies on the Signaling Pathway for Vascular Endothelial Growth Factor-mediated Tumor Development and Angiogenesis

The growth of a solid tumor is now widely recognized to depend on the process of neovascularization. Without angiogenesis, tumors cease to grow beyond even a few millimeters in volume (1, 2, 3, 4, 5). It has been shown that tumor vascular density is an independent prognostic marker in several types of human tumors and is correlated with a poor prognosis (6, 7, 8, 9, 10, 11, 12, 13, 14). To date, many positive and negative angiogenic factors have been identified (1, 3, 15). Among the positive factors, VEGF,² originally identified as a vascular permeability factor, is the most infringing factor with regard to tumor angiogenesis. VEGF is a specific mitogen for vascular ECs in vitro and can be an angiogenic factor for neovascularization in vivo (1, 2, 3, 4, 5). In contrast to basic fibroblast growth factor, which is also a representative positive angiogenic factor, VEGF has a typical signal peptide composed of 26 amino acids. Accordingly, it has been shown that VEGF is secreted abundantly in several human tumors and animal experimental models. An increase in VEGF expression in human surgical specimens has been shown to be correlated with aggressive behavior and prognosis (5, 6, 7, 8, 9, 10, 11, 12, 13, 14). In animal experimental models, overexpression of VEGF enhanced tumor growth, angiogenesis, and dissemination, whereas suppression of VEGF inhibited tumor growth in several tumor cell types (16, 17, 18, 19, 20, 21, 22). VEGF has been shown to act mainly through two distinct type-III tyrosine-kinase receptors, flt-1 and Flk-1/KDR, which are specifically expressed in the vascular ECs. Recent studies revealed that VEGF and its receptor interaction play an important role in a number of pathological events such as tumor angiogenesis, and these two receptors have different roles and signaling pathways (2, 23, 24, 25). Flk-1/KDR appears to play a predominant role in angiogenesis process both in vitro and in vivo (23, 26, 27, 28, 29, 30). However, the biological process that underlines these signaling pathways remains largely unknown.

The Tet system is a novel tetracycline-regulated gene expression system (31). This system allows manipulation of the interest gene in an “on and off” manner in vivo, which the conventional expression system can’t do for their constitutive gene expression level either overexpression or suppression. In this study, we used a retrovirus-mediated modified vector in which the two components of the Tet system have been organized within the same vector (32, 33). With this Retro-Tet System, it is possible to observe tumor kinetics that can be achieved only by VEGF gene expression (21, 34). PKC is composed of a family of serine-threonine kinases. It has been postulated that activation of PKC is an important intracellular signaling pathway for cellular growth (35, 36). Several reports have shown that PKC plays an important role in the VEGF-initiated angiogenesis process (37, 38, 39, 40, 41). In KDR-transfected NIH3T3 cells and ECs, MAP kinase was activated upon VEGF stimulation. This MAP kinase activation was mediated mainly by the PKC pathway, especially the β-isoform (42). Other reports showed that the PKC-β inhibitor strongly suppressed VEGF-dependent EC growth in vitro (37). In a rat diabetic model, VEGF-induced retinal permeability was suppressed by the PKC-β inhibitor (39, 40). However, the role of PKC in tumor angiogenesis has not yet been clarified.

In this study, in both a xenograft model and an orthotopical experimental model, we examined the role of PKC, especially the β -isoform, in tumor development and angiogenesis at different stages, which was due exclusively to VEGF gene expression by means of a combination of the PKC-β inhibitor and the Tet system. We report here for the first time that PKC, especially the β-isoform, is a major regulator of VEGF-mediated tumor development and angiogenesis.

The tetracycline aspect of the retroviral vector is conferred by the response and regulator units of the previously described Tet-inducible promoter system (31). A complete description of the construction of this Retro-Tet system vector has been reported previously (32, 33). The parent PBSTR-1 vector was generously provided by Dr. S. A. Reeves (Massachusetts General Hospital, Boston, MA). Human VEGF cDNA was cloned into the multicloning site of the PBSTR-1 vector at a BamHI site, forming the Tet-VEGF vector.

BALB/c mice-derived BNL. 1. 7R. 1 HCC (BNL-HCC) and BOSC 23 retrovirus packaging cells were purchased from the Japanese Cancer Research Resources Bank (Tokyo, Japan) and American Type Culture Collection (Manassas, VA), respectively. Cells were grown in a supplier-recommended medium. Stable Tet-VEGF-expressing clone (Tet-VEGF-HCC) was generated as described previously (34). Briefly, after the transduction of Tet-VEGF into BNL-HCC cells, followed by 14 days of selection with puromycin (1.5 μg/ml; Sigma), surviving single cell colonies were picked up and expanded. The VEGF expression level in the condition media for each colony was measured by an ELISA (R&D Systems). For further study, we used the colony that showed the highest and lowest VEGF clone in the absence or presence of tetracycline (tet: 1 μg/ml; Sigma), respectively. Tet-lacZ vector was also transduced into BNL-HCC cells and used as control cells (Tet-lacZ-HCC).

In the xenograft model, 1 × 10⁶ Tet-VEGF or lacZ-HCC cells were syngenetically inoculated into the flank of BALB/c mice. PKC-β inhibitor (LY333531) was administered orally to mice by gavage on the same day as transplantation. For the dose-ranging study, mice were given 10, 50, and 100 mg/kg PKC-β inhibitor, respectively. The tumor was measured twice a week using calipers, and the tumor volume was calculated as described previously (21). The next experiment was conducted to examine the effect of the PKC-β inhibitor on the fully established tumor. In this experiment, PKC-β inhibitor administration started on day 18 (tumor volume was 438 mm³). To suppress VEGF expression in vivo, the mice were given tet-containing drinking water (1 mg/ml). For the orthotopical transplantation experiment, 1 × 10⁶ Tet-VEGF or lacZ-HCC cells were injected directly into the liver, and mice were given 10 mg/kg of PKC-β inhibitor. Fourteen days after injection, mice were sacrificed and examined for HCC tumor development in the liver, macroscopically and microscopically.

To evaluate the expression of platelet/EC adhesion molecule (PECAM/CD31) RNA, which is widely used as a neovascularization marker, we used semiquantitative RT-PCR analysis. The tumor pool was prepared from three of Tet-VEGF, lacZ-transduced mice and snap-frozen into the liquid nitrogen immediately. The mRNA was extracted from the tumor pools, and RT-PCR was performed as described previously (34). PCR products of CD31 were examined by 1.2% agarose gel electrophoresis and stained by ethidium bromide. Densitometric analysis was performed by measuring the absorbance with a Fuji BAS 2000 image analyzer. The level of gene expression was calculated after normalization with the glyceraldehyde-3-phosphate dehydrogenase internal control.

The tumor pool was prepared in the same manner as described for the CD31 expression evaluation. MAP kinase activity was examined with some modification, as described previously (41, 42). Equal concentrations of protein from concentrated tumor lysate were subjected to SDS-PAGE and transferred onto a nitrocellurose membrane, and Western blot analysis was done using a phospho-specific p44/42 MAP kinase antibody (New England BioLabs) diluted 1:1000. Blots were developed using an amplified alkaline phosphates immune-blot assay kit (Bio-Rad).

Immunohistochemistry of CD31 was analyzed as described previously (43). Frozen sections were allowed to thaw at room temperature for 30 min and then washed once in PBS. Nonspecific binding was conducted by incubation with 0.1% BSA in PBS for 15 min and with normal rabbit serum for 30 min. Endogenous biotin was blocked with 0.1% avidin for 15 min, followed by biotin for 15 min. The section was washed once with PBS after each 15-min incubation period, then reacted with a primary rat anti-mouse CD31 antibody (PharMingen, San Diego, CA; 1:100) for 1 h. After washing twice with PBS, the sections were incubated with a second biotin-labeled rabbit antirat IgG (Novocastra). Endogenous peroxidase was blocked with 0.3% hydrogen peroxide in PBS for 30 min, followed by a 5-min washing with PBS. Sections were then incubated with conjugated streptavidin for 30 min, rinsed again in PBS, and incubated with diaminobenzidine for 3 min.

In vivo tumor cell proliferation was evaluated by quantitation of cells positive for PCNA and mitotic activity in tissue sections. Apoptosis was detected with the DNA fragmentation products that were stained by in situ 3′ end labeling (TUNEL). Immunohistochemistry for PCNA and TUNEL was performed as described previously (43, 44). Paraffin-embedded sections were used for the PCNA and TUNEL assay.

The statistical significance was analyzed by Tukey’s multiple comparison test.

In vitro VEGF expression in Tet-VEGF-HCC was tightly regulated by tetracycline (1 μg/ml), as described previously (34). In vivo, Tet-VEGF-HCC cells showed a marked increase in tumor development compared to the control group (lacZ-HCC cells). When VEGF expression was inhibited by the addition of tetracycline (1 mg/ml) to the drinking water, tumor growth was significantly decreased, almost to the same level as the lacZ-HCC cells (Fig. 1). This indicates that the tumor growth kinetics of this study should be attributed exclusively to VEGF expression. Detectable endogenous VEGF was not seen in the wild-type BNL-HCC cells (data not shown).

First, we performed the syngeneic xenograft experiment to examine the role of PKC-β in VEGF-induced tumor development. To maintain VEGF overexpression, Tet-VEGF-HCC-bearing mice (the Tet-VEGF group) were given tetracycline-free normal water throughout the experiment. Three different concentrations of PKC-β inhibitor were given orally on a daily basis to the Tet-VEGF group at doses of 100, 50, and 10 mg/kg. As shown in Fig. 2,A, even 10 mg/kg PKC-β inhibitor significantly inhibited tumor growth compared to the nontreated Tet-VEGF group (P < 0.001). Next we examined the effect of the PKC-β inhibitor on tumor development in fully established tumors. Because 10 mg/kg PKC-β showed a sufficient inhibitory effect, a 10 mg/kg dose was used for additional experiments. All mice of the Tet-VEGF group were allowed to grow in the absence of tetracycline (switch “ on”) until day 18 (tumor volume was 438 mm³). Next, half of the mice were given the PKC-β inhibitor starting on day 18. The tumor growth rate of the PKC-β inhibitor-treated mice was markedly reduced to almost the same growth rate observed in mice receiving the PKC-β inhibitor from day 0 (Fig. 2 B). These results indicated that tumor augmentation induced exclusively by VEGF was significantly reduced by LY333531 at a dose of 10 mg/kg. Furthermore, this inhibitory effect appeared even after the tumor was established. In this xenograft experimental model, PKC activity in the tumor was significantly inhibited by the 10-mg/kg dose of the PKC-β inhibitor (data not shown).

Because an orthotopical transplantation experiment has been shown to be much closer to the real biological situation (45, 46), we also examined the effect of PKC-β inhibitor on HCC development in the liver. Table 1 indicates that the PKC-β inhibitor significantly reduced the number of tumors in the liver (P < 0.001). In addition, most mice of Tet-VEGF group had a large tumor (over 5 mm in diameter; seven of seven mice; 100%), whereas such a large tumor was seldom found in the PKC-β inhibitor-treated mice (one of seven mice: 14.2%; Table 1). Neither tumor invasion into the other organ nor metastasis was observed at the sacrifice period (day 14). Systemic administration of PKC-β inhibitor had no obvious effect on the state of health of the mice in both the xenograft and orthotopical experiments (data not shown).

To determine whether the inhibitory effect of the PKC-β inhibitor on tumor development was accompanied by the suppression of tumor angiogenesis, we examined CD31 expression in the tumor. First we performed an immunohistochemical analysis for CD31. As shown in Fig. 3, CD31-positive vessels were significantly decreased in PKC-β inhibitor-treated mice. However, it seemed difficult to obtain reliable vascular counts because there were only small slit-positive vessels in PKC-β inhibitor-treated mice. Accordingly, we next performed the semiquantitative RT-PCR analysis for CD31 gene expression. CD31 RNA expression in PKC-β inhibitor-treated mice was five times lower than that in Tet-VEGF mice without PKC-β inhibitor treatment (Fig. 4 and Table 2).

The p44/42 MAP kinases are shown to be associated with cell proliferation induced by a number of mitogens, including VEGF and basic fibroblast growth factor (42, 47, 48). In KDR-transfected NIH3T3 cells and ECs, the p44/42 MAP kinase was activated upon VEGF stimulation, and signals from VEGF receptors to the MAP kinase were mediated mainly by PKC, especially the β isoform-dependent pathway (41, 42). To assess the role of PKC-β on the activation of MAP kinase in VEGF-mediated tumor development, we examined the MAP kinase with the p44/42 MAP kinase antibody, which recognized only the catalytically active, phosphorylated active species. Fig. 5 revealed that the phosphorylated activation of p44/42 MAP kinase was significantly suppressed upon administration of the PKC-β inhibitor at a dose of 10 mg/kg.

To examine whether the tumor inhibition by the PKC-β inhibitor was reflected by tumor cell proliferation and apoptosis, we performed an immunohistochemical analysis of PCNA and TUNEL. We also counted the mitotic index with routine H&E-stained tumor sections. As shown in Table 3, PCNA-positive cells and the mitotic index did not show a remarkable difference between PKC-β inhibitor-treated mice and nontreated mice. In vitro, even 500 nm PKC-β inhibitor, which corresponds to the expected blood concentrations of LY333531 in mice treated with a dose of 100 mg/kg in vivo, did not affect the proliferation of Tet-VEGF-HCC cells (data not shown). Our in vivo result was consistent with this in vitro data. On the contrary, the number of TUNEL-positive cells was significantly increased by treatment with 10 mg/kg PKC-β inhibitor (P < 0.001; Table 3). Histological examination of routine H&E-stained sections showed an increase of inflammatory cells consisting mainly of macrophages and extensive necrosis in the PKC-β inhibitor-treated mice (data not shown).

The present study revealed that PKC-β played an important role in VEGF-mediated tumor development not only at the initial stage of tumor development, but also after the tumor was fully established. The role of VEGF as a major angiogenic factor is well established from a number of reports, including those regarding HCCs (1, 2, 3, 4, 5). In human malignancies, increased expression of VEGF correlated with a more aggressive neoplastic phenotype and a poor prognosis (6, 7, 8, 9, 10, 11, 12, 13, 14). Preclinical in vivo experiments have shown that overexpression of VEGF augmented tumor growth, whereas suppression of VEGF inhibited tumor growth in several types of tumors (16, 17, 18, 19, 20, 21, 22). However, most of these studies used the conventional gene modification system, the expression level of which was constant throughout the experiment (either overexpression or suppression).

The Tet system is a novel gene expression system that allows the gene of interest to be modified in an on/off manner (31). We used the retroviral modified (Retro-Tet) vector in this study (32, 33). The major advantage of using this vector in vivo is that gene expression is tightly regulated by tetracycline in the drinking water and can be efficiently switched on and off. To decrease leaky expression in the target gene at the off state, the two components of the Tet system were inserted in opposite directions within the same vector. The apparent antisense inhibitory effect minimized the leaky basal gene expression compared to the original two-plasmid Tet system. With this system, tumor growth kinetics that are mediated exclusively by VEGF gene modification can be ascertained.

The biological actions of VEGF are mediated mainly by two distinct type-III tyrosine kinase receptors: flt-1 and KDR/Flk-1, which are almost exclusively expressed on ECs (2). A third VEGF receptor that binds VEGF₁₆₅ was identified recently and is named neurophilin. Neurophilin modulated VEGF binding to KDR/Flk-1 and subsequently bioactivity (48). The first two receptors, flt-1 and KDR/Flk-1, have been shown to possess a different role and signal transduction pathway (2, 23, 24, 25). Each gene knockout model suggested that KDR/Flk-1 stimulates the proliferation of ECs, and flt-1 is required more for differentiation and vascular organization (49, 50). In the angiogenesis process, KDR/Flk-1 has been shown to be a major regulator, both in vitro and in vivo (23, 26, 27, 28, 29, 30). However, the intracellular signal transduction pathway initiated by VEGF is not fully understood. It has been suggested that a major target molecule of KDR/Flk-1 was PLC-γ, which binds to the KDR/Flk-1 and is tyrosine-phosphorylated in response to VEGF stimulation. It also seems likely that the activated PLC-γ leads to PKC activation (42). The PKC family consists of at least 12 isozymes, which are divided into three subfamilies: (a) classical PKCs; (b) novel PKCs; and (c) atypical PKCs. This kinase family has been implicated in regulating a variety of biological actions, such as cellular differentiation, gene regulation, and proliferation (35, 36). It has been shown that activation of PKC is important for EC growth during angiogenesis (51, 52, 53, 54). Selective inhibition of the PKC-β isoform strongly suppressed VEGF-dependent EC growth in a concentration-dependent manner. In vitro, VEGF appears to mediate its mitogenic effects through activation of the PKC pathway, predominantly involving PKC-β isoform activation in EC (37). In vivo, it has been shown that oral administration of the PKC-β inhibitor normalized many of the early retinal and renal hemodynamic abnormalities in rat models of diabetes (39, 40). However, the role of PKC in tumor development and angiogenesis has not yet been clarified. In this study, using a combination of the Retro-Tet system and the PKC-β inhibitor, we elucidated the role of PKC-β in VEGF-mediated tumor development and angiogenesis. We found that PKC-β inhibition markedly inhibited VEGF-mediated tumor development, which was accompanied by the suppression of neovascularization. This inhibitory effect was achieved even after the tumor was established. We administered this oral PKC-β inhibitor in three different concentrations: 100; 50; and 10 mg/kg. Even the 10 mg/kg dose of PKC-β inhibitor showed a sufficient inhibitory effect on tumor development. At the 100 mg/kg dose, this compound would inhibit not only the β-isoform but also many other types of PKC. At the oral dose of 10 mg/kg, the steady-state plasma level of this compound active metabolite is approximately ∼40 nm,³ which would be a 10–100-fold less specificity for any other PKC isoform in vitro, including PKC-α (IC₅₀ = 360 nm) (40, 55). Consequently, only the PKC-β isoform would be inhibited at a dose of 10 mg/kg.

MAP kinases are a family of serine/threonine-specific protein kinases whose activation is commonly induced by various mitogens. The p44/42 MAP kinase pathway consists of a protein kinase cascade linking growth and differentiation signals with transcription in the nucleus. These MAP kinases are known to be associated with cell proliferation induced by a number of mitogens, including VEGF (42, 47, 56). The p44/42 MAP kinase was activated upon VEGF stimulation in ECs, and signals from VEGF receptors to MAP kinase were mediated mainly by PKC, especially the β isoform-dependent pathway. Ras activation was not required for VEGF signaling toward cell proliferation (41). Consistent with these in vitro data, we found that PKC-β inhibition significantly suppressed p44/42 MAP kinase phosphorylation in tumors. Furthermore, this PKC-β inhibitor affected neither the in vitro nor in vivo proliferation of Tet-VEGF-HCC cells. Accordingly, these findings suggest that this PKC-β inhibitor may inhibit p44/42 MAP kinase of ECs in the tumor rather than the tumor cell itself. Additional studies are required to identify the mechanism and the type of cells responsible for this MAP kinase inhibition.

Some reports, on the contrary, showed that VEGF utilized a different signaling pathway than PKC. Human umbilical vein endothelial cell survival depended on the activation of the phosphatidylinositol 3′-kinase/Akt signaling pathway (57). Phosphatidylinositol 3′-kinase/Akt signal was also utilized in Ha-ras-transfected NIH3T3 cells in a Ras-dependent manner (58). It may be possible that VEGF utilizes the cell type-specific signaling cascade to selectively elicit specific cellular responses. VEGF has been shown to act as a survival factor, preventing the apoptotic death of microvascular ECs (24, 59). In glioma cells, VEGF inhibition led to the detachment of ECs from the walls of preformed vessels and to their subsequent death by apoptosis (44). Angiostatin, a potent endogenous angiogenic inhibitor, has been shown to induce apoptosis of tumor cells, whereas tumor cell proliferation itself was not changed in vivo (60). In this experiment, PKC-β inhibition also significantly increased the number of apoptotic cells. On the contrary, tumor cell proliferation and mitotic index were not changed. Taken together, these data suggest that PKC-β inhibition reduced the tumor growth by decreasing VEGF-dependent MAP kinase activation, which in turn may stimulate the induction of apoptosis.

Orthotopic transplantation has been reported to be a preferential site for transplanted tumor cell growth, enhanced tumor growth, and development of metastatic ability in several types of tumors (45, 46). Accordingly, we examined the effect of PKC-β inhibition on HCC development of an orthotopic model in the liver. Similar to the xenograft experimental model, PKC-β inhibition significantly reduced the size and number of HCC tumors. In summary, from the present study, we suggest for the first time that PKC-β lies on the signaling pathway in VEGF-mediated tumor development and angiogenesis not only at the initial stage, but also after the tumor has become established in murine HCC cells.