The von Hippel-Lindau Gene Product Inhibits Vascular Permeability Factor/Vascular Endothelial Growth Factor Expression in Renal Cell Carcinoma by Blocking Protein Kinase C Pathways*

Mutation or loss of function of the von Hippel-Lindau (VHL) tumor suppressor gene is regularly found in sporadic renal cell carcinomas (RCC), well vascularized malignant tumors that characteristically overexpress vascular permeability factor/vascular endothelial growth factor (VPF/VEGF). The wild-type VHL (wt-VHL) gene product acts to suppress VPF/VEGF expression, which is overexpressed when wt-VHL is inactive. The present study investigated the pathways by which VHL regulates VPF/VEGF expression. We found that inhibition of protein kinase C (PKC) represses VPF/VEGF expression in RCC cells that regularly overexpress VPF/VEGF. The wt-VHL expressed by stably transfected RCC cells forms cytoplasmic complexes with two specific PKC isoforms, ζ and δ, and prevents their translocation to the cell membrane where they otherwise would engage in signaling steps that lead to VPF/VEGF overexpression. Other experiments implicated mitogen-activated protein kinase (MAPK) phosphorylation as a downstream step in PKC regulation of VPF/VEGF expression. Taken together, these data demonstrate that wt-VHL, by neutralizing PKC isoforms ζ and δ and thereby inhibiting MAPK activation, plays an important role in preventing aberrant VPF/VEGF overexpression and the angiogenesis that results from such overexpression.

Sporadic RCC as well as the RCC and other tumors that arise in patients with the VHL syndrome are characteristically well vascularized, a property that has been attributed to their consistent overexpression of the potent angiogenic factor, vascular permeability factor, also known as vascular endothelial growth factor (VPF/VEGF) (12)(13)(14)(15)(16). Several different mechanisms have been shown to up-regulate VPF/VEGF expression in different tumors and also in non-tumorigenic cells; these include hypoxia, cytokines such as transforming growth factor-␣ and platelet-derived growth factor, hormones, and more recently, certain oncogenes including activated forms of Src and Ras (17)(18)(19)(20)(21)(22).
The mechanisms responsible for up-regulating VPF/VEGF expression in RCC are not yet fully understood. Recent reports have shown that the VHL tumor suppressor gene down-regulates VPF/VEGF gene expression at both the transcriptional and post-transcriptional levels (23)(24)(25). In addition, the wildtype VHL (wt-VHL) gene product represses VPF/VEGF expression by inhibiting Sp1/transcriptional (23). Several reports suggest that steps in the Ras pathway may also play an important role in up-regulating VPF/VEGF expression in RCC (26,27).
The present experiments were undertaken to elucidate more fully the role of VHL in regulating VPF/VEGF expression in RCC. We hypothesized that the wt-VHL gene product acts to inhibit steps in a signaling cascade that involves several protein kinases, including PKC. Here we present evidence in favor of this hypothesis, demonstrating that 1) PKC inhibitors repress the overexpression of VPF/VEGF characteristic of RCC cells, 2) wt-VHL protein complexes selectively with PKC and ␦, preventing the translocation of these PKC isoforms to the cell membrane and thereby interrupting a signaling cascade that involves MAP kinase, and 3) PKC ␦, but not PKC , complexes directly with wt-VHL protein.
RNA Analysis by Northern Blot Hybridization-Total RNA was isolated and Northern blots were performed as described previously (23). Bolts were washed at high stringency and quantitated by laser densitometry.
Immunoprecipitations and Western Blot Analyses-Cell lysis and immunoprecipitations were performed as described previously (23). Briefly, immunoprecipitations were carried out in antibody excess, using 0.5 mg of total protein with either affinity-purified rabbit polyclonal antibody (1 g of IgG) against PKC isoforms (Chemicon International Inc., CA) or with a mouse monoclonal antibody (1 g) directed against VHL (gift from O. Iliopoulos and W. G. Kaelin). Proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE). Western blot analysis was carried out as described earlier (23).
PKC Assay-Cell samples were prepared and insoluble and soluble components were isolated as manufacturer's protocol (Pierce). Afterward, extracts were immunoprecipitated with specific antibodies to PKC and ␦, and kinase assays were performed using the Pierce colorimetric PKC assay kit according to the manufacturer's protocol. We used two different synthetic peptides (glycogen synthase, product 29510, and neurogranin peptide, product 29580) as PKC substrates. A standard curve was generated for each experiment using an appropriate PKC standard with known activity (supplied by Pierce), and PKC activity of each sample was measured using that standard curve. Each experiment was performed in triplicate.
In Vitro Binding Experiments-In vitro binding assays were performed as described previously (23). Briefly, the glutathione-Sepharose beads bound with GST-fused VHL proteins were incubated with purified PKC isoforms (50 ng, PanVera Corp., WI). The beads were then incubated for 1 h at 4°C, washed four times with cold binding buffer, and boiled in SDS sample buffer. Bound proteins were resolved on SDS-PAGE, and blots were performed with antibodies to specific PKC isoforms.

Inhibition of VPF/VEGF Expression by Different Protein
Kinase Inhibitors-Earlier work had shown that the wt-VHL gene suppresses VPF/VEGF expression in RCC and that genistein, a potent inhibitor of protein tyrosine kinases, inhibits hypoxia-induced VPF/VEGF overexpression (19). We had also shown that genistein at a level of 100 M suppressed completely the high base-line VPF/VEGF expression characteristic of RCC cells (data not shown). Taken together, these findings suggested that wt-VHL regulated VPF/VEGF expression by interacting with a signaling pathway that involved protein tyrosine kinases. To test this hypothesis further, we studied the effects of several more specific kinase inhibitors on the expression of VPF/VEGF mRNA in a renal carcinoma cell line, 786-0, which constitutively overexpresses VPF/VEGF. Staurosporine, a potent and specific PKC inhibitor, suppressed basal VPF/VEGF expression by 90% at a concentration of 1 M (Fig.  1). However, VPF/VEGF mRNA levels were not significantly altered by H89, a PKA inhibitor, or by GO6976, a selective inhibitor of the ␣ and ␤ PKC isoforms (Fig. 1). These data indicate that activation of VPF/VEGF mRNA in RCC is mediated through PKC but not through its ␣ or ␤ isoforms or by PKA.
Interaction of VHL with Specific PKC Isoforms-To investigate further the interaction between VHL and PKC in regulating VPF/VEGF expression, we looked for a possible physical interaction between the VHL gene product and various PKC isoforms. To that end, we prepared lysates of wt-VHL cells (VHL) and immunoprecipitated with affinity-purified antibodies specific for various PKC isoforms, followed by Western blotting with a monoclonal antibody directed against wt-VHL protein. In the case of immunoprecipitates prepared with antibodies to PKC and PKC ␦, Western blots revealed intense bands ( Ͼ ␦). Very weak VHL-positive bands were also detected when immunoprecipitates were prepared with antibodies against the ␣ or ⑀ isoforms of PKC (Fig. 2, a-d). However, no detectable band was found against the ␤, ␥, and PKC isoforms (data not shown).
In contrast, anti-VHL reactive protein bands were not detected in Western blots of immunoprecipitates prepared with antibodies to PKC isoforms on cell extracts of either RCC cells transfected with an empty neo cassette (neo-786-O cells, Neo) or stably transfected with a deletion mutation of VHL (VHL-(1-115), ⌬VHL) (Fig. 2, a-d). In the case of VHL-(1-115) cells, these findings could not be attributed to a failure of protein expression because VHL-(1-115) cells contained mutant VHL protein in amounts comparable with those of wt-VHL protein present in VHL cells (Fig. 2e). Also, making use of split samples, we determined that the levels of PKC isoforms were similar in all three cell lines (Fig. 2, a-d).
We also performed reciprocal experiments. Extracts were prepared from the same three cell lines and were immunoprecipitated with a monoclonal antibody directed against VHL, followed by Western blotting with antibodies specific for PKC isoforms ␣, ␤, ␦, ⑀, ␥, , and . Only extracts of wt-VHLexpressing cell lines formed complexes that reacted with anti-PKC (Fig. 2f). Surprisingly, we were unable to detect a band that reacted with anti-PKC ␦, in wt-VHL extracts, perhaps because of the low blotting sensitivity of the anti-PKC ␦ antibody. As expected, no bands were detected with antibodies to PKC isoforms ␣, ␤, ⑀, ␥, and .
Subcellular Localization of VHL and Its Complexes with PKC Isoforms-We next determined the subcellular localization of wt-VHL protein and of its complexes with PKC isoforms and ␦. Cytoplasmic and membrane fractions were prepared from extracts of Neo and VHL cells. Western blots with antibodies to VHL revealed that the great bulk of VHL was found in the cytoplasmic fraction of VHL cells; as expected, VHL protein was not detected in cytoplasmic or membrane fractions of the Neo cells (Fig. 3a). We next immunoprecipitated these cytoplasmic extracts with antisera against PKC and ␦. We found VHL-reactive bands in cytoplasmic fractions (Fig. 3, b and c) but not in membrane fractions (data not shown) of VHL cells. Thus, PKC and ␦ complexes were localized primarily to the cytoplasm.
Regulation of PKC Activity by wt-VHL-We immunoprecipitated cytoplasmic and membrane fractions of the RCC cell lines with antibodies to PKC and blotted these with the same antibody. Fig. 3d shows that the amount of PKC present in the membrane fraction of the wt-VHL-expressing cell line is significantly lower than that present in its cytoplasmic fraction. In contrast, relatively more PKC was found in the membrane fractions (as compared with the cytoplasmic fractions) of Neo and VHL-(1-115) (⌬VHL) cells. We measured PKC activity in membrane and cytoplasmic fractions of these same immunocomplexes using a colorimetric assay. Levels of PKC activity in the cytoplasmic fraction of wt-VHL-expressing cells were higher than those in cytoplasmic fractions of either Neo or VHL-(1-115) cells but did not quite achieve statistical significance (analysis of variance, p ϭ 0.06, Fig. 3e). However, PKC activity was significantly reduced in membrane preparations of VHL cells compared with membrane preparations of Neo or VHL-(1-115) cells (p Ͻ 0.005). Also, the amount of PKC activity found in VHL cell cytoplasmic fractions was significantly greater than that found in membrane prepa-FIG. 1. Effect of protein kinase inhibitors on VPF/VEGF mRNA expression. Total RNA (5 g) was extracted from 786-O RCC cells that had been cultured with different protein kinase inhibitors and was subjected to Northern blotting with 32 P-labeled VPF/VEGF cDNA. Fold expression was calculated by densitometry using 36B4 expression as a normalization control. Stauros., staurosporine.

wt-VHL Blocks PKC Pathways in Renal Cancers
rations of these same cells (p Ͻ 0.005, Student's t test), whereas no significant differences in activity were found in cytoplasmic versus membrane fractions of Neo or VHL-(1-115) cells. We obtained similar results with anti-PKC ␦ antibody. Also similar results were obtained when we used neurogranin peptide as PKC substrate (data not shown). Together these data are consistent with the hypothesis that wt-VHL binds PKC and ␦ in the cytoplasm, preventing their translocation to the cell membrane.
Interaction of VHL with PKC and ␦ in Vitro-The experiments just described did not determine whether the cytoplasmic complexes of wt-VHL with PKC or ␦ resulted from a direct interaction between these proteins or required the presence of additional components. To distinguish between these possibilities, we performed in vitro association experiments using GST-VHL fusion proteins and recombinant PKC and ␦ isoforms. Bacterially expressed GST protein alone or GST protein fused to full-length wt-VHL or VHL-(1-115) was bound to glutathione-Sepharose beads, and these were mixed with purified recombinant PKC or ␦ in a buffer designed to approximate intracellular ionic concentrations (see "Experimental Procedures" for details). After suitable incubation and extensive washing with the same buffer, the bound proteins were separated by SDS-PAGE and subjected to Western blotting with antibodies to PKC and ␦. We observed a strong association of PKC ␦, but not PKC , with immobilized wt-VHL (Fig. 3f). These results indicate that PKC ␦ interacted directly with wt-VHL protein; in contrast, wt-VHL did not interact directly with PKC , and therefore the wt-VHL-PKC complex that forms in living cells must involve additional as yet unidentified component(s).
wt-VHL Inhibits Phosphorylation of MAP Kinase-MAPK is a downstream target of both PKC and PKC ␦ (30 -32). To determine whether wt-VHL interferes with downstream signaling of these PKC isoforms, we examined the status of the activated form of MAPK (phospho-MAPK) in VHL cells as well as in the Neo and VHL-(1-115) cells. Total levels of MAP kinase were similar in all three cell lines. However, we found significantly lower levels of phospho-MAPK in VHL cells than in the two, non-wt-VHL-expressing cell lines (Fig. 4a). These results provide further evidence that formation of wt-VHL-PKC complexes in VHL cells prevents PKC translocation to the cell membrane and thus subsequent activation of MAPK.
Inhibition of MAPK Phosphorylation Down-regulates VPF/ VEGF Expression-To determine whether phosphorylation of MAPK has an effect on VPF/VEGF expression, we treated RCC cells with a specific inhibitor of MAPK kinase (MEK), PD098059, and examined VPF/VEGF mRNA levels by Northern blot analysis (33). Fig. 4b shows that PD098059 reduced VPF/VEGF mRNA levels by 80% as compared with the untreated control. These data suggest that MEK activation and MAPK phosphorylation have a role in the overexpression of VPF/VEGF found in RCC.

DISCUSSION
The data presented here investigated the mechanisms by which VHL regulates VPF/VEGF expression in RCC. In accord with earlier work, we found that a PKC inhibitor, staurosporine, repressed VPF/VEGF expression in RCC cells whereas inhibitors of PKC ␣ or ␤ or of PKA did not. Also, the wt-VHL tumor suppressor gene product formed selective complexes with two specific PKC isoforms ( Ͼ ␦) in wt-VHL-expressing cells. The interaction of wt-VHL protein with recombinant PKC ␦ was replicated in vitro; however, wt-VHL did not complex directly with PKC , suggesting that additional, as yet unidentified, components are required for wt-VHL-PKC complex formation. The complexes that formed in cells between wt-VHL and PKC and ␦ were localized primarily in cell cytoplasm. Also, more PKC and ␦ activity was found in the cytoplasm of wt-VHL cells than in either Neo or VHL-(1-115) cells that lacked wt-VHL; conversely, more PKC and ␦ activity was found in cell membrane fractions of Neo and VHL-(1-115) cells than in wt-VHL-expressing RCC. In addition, although total amounts of MAPK were roughly equivalent in cells that lacked or expressed wt-VHL, significantly less activated (phosphorylated) MAPK was found in cells expressing wt-VHL than in cells and c, immunoprecipitates (IP) of cytoplasmic extracts of Neo and VHL cells with antibodies to PKC or ␦, followed by blotting with antibodies to VHL. d, cytoplasmic and membrane fractions of RCC cell lines were immunoprecipitated with antibody to PKC , and Western blots were performed with the same antibody. e, the PKC immunocomplexes described in d were assayed for kinase activity using the Pierce colorimetric PKC assay kit; glycogen synthase peptide was used as a substrate. Units of PKC enzyme activity were determined by a PKC standard supplied by Pierce. One unit of specific activity was defined as the amount of protein that transfers 1 nmol of phosphate to histone H1 per min at 30°C. f, blots of in vitro interactions of bacterially expressed GST, GST wt-VHL, and GST VHL-(1-115) proteins with recombinant PKC and ␦ (50 ng). One-twentieth the amount of PKCs used in the binding buffer was included for comparison (ϫ 0.2). that did not express the VHL protein. Finally, an inhibitor of MEK led to reduced expression of VPF/VEGF mRNA in RCC cells.
Taken together, we propose that wt-VHL, as expressed in VHL cells, forms complexes with PKC and ␦, thereby retaining these PKC isoforms in the cytoplasm and preventing them from translocating to the cell membrane where they could engage in downstream signaling. In contrast, when wt-VHL protein was not available, as in Neo or VHL-(1-115) cells, PKC and ␦ were free to translocate to the cell membrane where they engaged in downstream signaling that led to the VPF/ VEGF overexpression characteristic of cell lines lacking wt-VHL.
PKC is known to be critically important for regulating a number of cell functions, including proliferation and activation of nuclear factor NF-B (34 -37). Previous studies indicated that PKC acts downstream of Ras in the kinase cascade leading to mitogenesis (30,32). Stimulation of Ras initiates a kinase cascade that culminates in the activation of MAP kinase, which is required for cell growth and other cellular responses. We also identified activation of MAPK as a downstream signaling step in the VHL cascade leading to the regulation of VPF/VEGF expression. Thus, although total levels of MAPK were similar in all three cell lines, levels of phospho-MAPK were significantly higher in Neo and VHL-(1-115) cells which lacked wt-VHL than in VHL cells which expressed wt-VHL. Further, selective inhibition of MAPK phosphorylation significantly inhibited VPF/VEGF expression. All of these data, therefore, support a model in which wt-VHL, by forming a cytoplasmic complex with PKC and ␦, prevents their translocation to the cell membrane and where they would otherwise initiate signaling steps leading to MAPK phosphorylation and subsequently VPF/VEGF overexpression. Of relevance, Oka et al. (38) also demonstrated that constitutive activation of MAPKs may be associated with the carcinogenesis of human RCCs.
Earlier studies suggested that VHL inhibits VPF/VEGF expression in RCC by both transcriptional and post-transcriptional mechanisms (23)(24)(25). We also showed that the VHL gene product inhibits VPF/VEGF transcription by inactivating Sp1. The present studies have identified additional, PKC-mediated mechanisms whereby VHL controls VPF/VEGF expression. Current experiments seek to determine whether these mechanisms affect transcription or act to decrease the stability of VPF/VEGF mRNA.