Up-regulation of hypoxia-inducible factors HIF-1α and HIF-2α under normoxic conditions in renal carcinoma cells by von Hippel-Lindau tumor suppressor gene loss of function

Abstract

Hypoxia induces transcription of a range of physiologically important genes including erythropoietin and vascular endothelial growth factor. The transcriptional activation is mediated by the hypoxia-inducible factor-1 (HIF-1), a heterodimeric member of the basic helix–loop–helix PAS family, composed of α and β subunits. HIF-1α shares 48 per cent identity with the recently identified HIF-2α protein that is also stimulated by hypoxia. In a previous study of hemangioblastomas, the most frequent manifestation of hereditary von Hippel-Lindau disease (VHL), we found elevated levels of vascular endothelial growth factor and HIF-2α mRNA in stromal cells of the tumors. Mutations of the VHL tumor suppressor gene are associated with a variety of tumors such as renal clear cell carcinomas (RCC). In this study, we analysed the expression of the hypoxia-inducible factors HIF-1α and HIF-2α in a range of VHL wildtype and VHL deficient RCC cell lines. In the presence of functional VHL protein, HIF-1α mRNA levels are elevated, whereas HIF-2α mRNA expression is increased only in cells lacking a functional VHL gene product. On the protein levels, however, in VHL deficient cell lines, both HIF-α subunits are constitutively expressed, whereas re-introduction of a functional VHL gene restores the instability of HIF-1α and HIF-2α proteins under normoxic conditions. Moreover, immunohistochemical analyses of RCCs and hemangioblastomas demonstrate up-regulation of HIF-1α and HIF-2α in the tumor cells. The data presented here provide evidence for a role of the VHL protein in regulation of angiogenesis and erythropoiesis mediated by the HIF-1α and HIF-2α proteins.

Introduction

The hypoxia-inducible factor-1 (HIF-1) was identified as a nuclear factor that activates gene transcription in response to reduced cellular O2 tension. It was initially analysed in studies of the oxygen-regulated expression of the hematopoietic growth factor erythropoietin (Epo) (Semenza and Wang, 1992). Further studies revealed that HIF-1 is responsible for hypoxia-inducible expression of genes involved in regulation of vascular growth and cellular metabolism such as vascular endothelial growth factor (VEGF), glucose transporter 1 (Glut1), glycolytic enzymes and nitric oxide synthetase (Wenger and Gassmann, 1997; Semenza, 1998). Biochemical studies indicated that HIF-1 is a heterodimer of two basic helix–loop–helix PAS domain (bHLH–PAS) proteins, namely HIF-1α and the previously identified aryl hydrocarbon nuclear receptor translocator ARNT (also termed HIF-1β) (Wang et al., 1995). The activation mechanisms of the HIF-1 complex appears to depend primarily on the hypoxia-induced accumulation of the HIF-1α subunit, while the HIF-1β protein level is expressed constitutively and not significantly affected by oxygen (Wang et al., 1995; Huang et al., 1996). Biochemical studies present evidence that the HIF-1α protein is rapidly degraded under normoxic conditions by the ubiquitin-proteasome system, controlled by an oxygen dependent degradation domain within HIF-1α, and that the accumulation of HIF-1α under hypoxia involves stabilization of the protein (Pugh et al., 1997; Salceda and Caro, 1997; Huang et al., 1998).

Recently, a novel bHLH–PAS protein termed endothelial PAS domain protein 1 (EPAS1) was identified, which shares 48 per cent sequence identity with HIF-1α (Tian et al., 1997). The factor was independently cloned by two other groups and reported as HIF-like factor (HLF) (Ema et al., 1997) and HIF-related factor (HRF) (Flamme et al., 1997) and is now termed HIF-2α (Wenger and Gassmann, 1997). Current evidence indicates that HIF-2α forms a functional heterodimer with HIF-1β, resulting in the HIF-2 complex, activating transcription from the same DNA recognition sites as HIF-1, the hypoxia-response element (HRE) of the Epo and VEGF genes. This activation by HIF-1α and HIF-2α is stimulated under hypoxic conditions (Tian et al., 1997; Ema et al., 1997). Although HIF-2α shows these very similar characteristics to HIF-1α, their levels of expression differ widely. HIF-2α mRNA is abundantly expressed in lung, heart, liver and other organs under normoxic conditions, whereas HIF-1α mRNA is ubiquitously expressed albeit at a much lower level (Ema et al., 1997).

We have recently shown that up-regulation of HIF-2α mRNA in hemangioblastomas, highly vascularized tumors of the central nervous system, is correlated with high expression of VEGF in the stromal cells of these tumors (Flamme et al., 1998). Hemangioblastomas are the most frequent manifestation of the autosomal dominantly inherited von Hippel-Lindau (VHL) disease. Mutations of the VHL tumor suppressor gene are responsible for the development of a variety of further tumors including renal clear cell carcinomas (RCC), pheochromocytomas, endolymphatic sac tumors and pancreatic cysts (Neumann et al., 1995). Several groups have shown that the VHL gene product is a negative regulator of hypoxia-inducible genes, such as VEGF, Glut1 and platelet-derived growth factor β. VHL presumably acts via destabilization of the respective mRNAs under normoxic conditions (Gnarra et al., 1996; Iliopoulos et al., 1996; Siemeister et al., 1996). The VHL protein can also act as a transcription elongation inhibitor in vitro by stable association with the two regulatory subunits B and C of the transcription elongation factor elongin (Duan et al., 1995; Kibel et al., 1995). The correlated up-regulation of VEGF and HIF-2α mRNAs in VHL-associated hemangioblastomas indicates an alternative mechanism for transcriptional activation of hypoxia-inducible genes in VHL deficient cells under normal cellular O2 tension mediated by the hypoxia-inducible factor-2. We therefore investigated a series of VHL wildtype and VHL deficient renal clear cell carcinoma cell lines for the presence of the HIF-1α and HIF-2α mRNAs and expression of the corresponding proteins. In all cells lines analysed which lacked the VHL wildtype protein, HIF-1α and HIF-2α protein expression was observed already under normoxic conditions. In contrast, VHL wildtype cells as well as VHL deficient cells with an re-introduced wildtype VHL gene contained the stabilized HIF-α proteins only under hypoxic conditions, providing evidence for regulation of HIF-1α and HIF-2α protein expression by the VHL gene product. In addition, hypoxia-induced up-regulation of HIF-2α mRNA in VHL wildtype expressing cells, indicates a transcriptional regulation of HIF-2α by the VHL gene product.

Maxwell et al., (1999) also showed that the HIF-α subunits are constitutively stabilized in VHL defective cells and that VHL and HIF-α subunits co-immunoprecipitate, suggesting VHL interaction with the two transcription factors.

Results

Characterization of mutations of the VHL gene in human RCCs

We analysed seven human renal clear cell carcinoma cell lines for mutations of the von Hippel-Lindau tumor suppressor gene using the polymerase chain reaction (PCR) and the single-strand conformational polymorphism analyses (SSCP) of DNA. PCR products in cases that showed abnormalities in SSCP analyses were sequenced. The sequence analyses revealed that tumor DNA of RCC6 had a mutation at nucleotide 388 (G to C) resulting in amino acid substitution of codon 130 valine to leucine, while tumor DNA of RCC10 had a deletion of nucleotides 474–477 (deletion of amino acid 159 lysine), creating a stop codon at codon 169. In the remaining lines RCC2, RCC7, RCC8, RCC9 and RCC11 no mutations could be found (data not shown). In cell lines RCC2 and RCC11 no VHL mRNA could be detected by Northern blot analysis (Figure 1), classifying these cell lines as VHL deficient. These reductions of VHL mRNA amounts was supported by Western blot analyses detecting a VHL protein signal only in lines RCC6 to RCC10 (data not shown).

Figure 1
figure1

Northern blot analysis of total RNA from seven different human renal clear cell carcinoma cell lines. Lanes 1–4: VHL deficient cell lines RCC2, RCC6, RCC10 and RCC11; lanes 5–7: VHL wildtype cell lines RCC7, RCC8 and RCC9. Ten μg RNA was loaded per lane. The blot was hybridized with 32P-labeled human VHL, VEGF, HIF-1α, HIF-2α and Glut1 probes. Hybridization with β-actin served as loading control

Distribution of VHL, VEGF, HIF-1α, HIF-2α and Glut1 mRNAs in RCC lines

We investigated the RCC lines for expression of the VHL gene by Northern blot analysis (Figure 1). A strong VHL mRNA signal was observed for the RCC lines 7, 8 and 9, expressing the VHL wildtype gene (lanes 5–7), whereas RCC lines 6 and 10 showed a weak level of VHL mRNA, carrying the above described mutations. No VHL mRNA could be detected in RCC2 and 11. Cells lacking the VHL wildtype protein were found to overexpress the mRNAs encoding the hypoxia-regulated genes VEGF and Glut1 (Gnarra et al., 1996; Iliopoulos et al., 1996; Siemeister et al., 1996). Consistent with these previous reports, we observed a strong up-regulation of VEGF and Glut1 expression in the VHL minus RCC lines 2, 6, 10 and 11 (lanes 1–4), whereas RCC lines 7, 8 and 9 expressing the VHL wildtype gene showed only weak VEGF and Glut1 mRNA levels (lanes 5–7). As the transcription factors HIF-1α and HIF-2α were shown to be responsible for the hypoxia-inducible expression of VEGF and Glut1, we examined the expression pattern of HIF-1α and HIF-2α mRNAs in the RCC lines. RCC lines expressing the VHL wildtype gene (Figure 1: lanes 5–7) seemed to express a higher level of HIF-1α mRNAs compared to VHL minus cell lines (lanes 1–4). However, VHL minus RCCs showed a strong up-regulation of HIF-2α mRNAs in contrast to the presence of the VHL wildtype protein. To support the observed differences in VEGF, Glut1, HIF-1α and HIF-2α mRNA abundance in the presence or absence of a functional VHL protein, VHL minus RCC10 cell line was stably transfected with a plasmid expressing the VHL wildtype gene. The re-introduction of the wildtype VHL protein largely restored the expressing pattern of these mRNAs (Figure 2). Eleven independent RCC10 subclones, expressing the VHL wildtype gene, revealed high levels of exogenous VHL mRNA (lanes 2–12). In the presence of a functional VHL gene product the abundance of VEGF and Glut1 mRNAs were strongly reduced. The reciprocal expression pattern of HIF-1α and HIF-2α mRNAs could be confirmed as well. While the introduction of the VHL wildtype gene led to up-regulation of HIF-1α mRNAs in all analysed cell clones, the abundance of the HIF-2α mRNA exhibited reduced levels in eight of 11 stable cell clones (lanes 2–5, 8–10, 12) compared to the VHL minus RCC10 line (lane 1). To get more information about the VHL dependent regulation of HIF-α mRNAs, we performed Northern blot analyses of RCC cell lines, cultured under hypoxic conditions (Figure 3). As expected, RCC lines 7, 8 and 9, expressing the VHL wildtype gene, showed a strong up-regulation of VEGF and Glut1 mRNAs, when the cells were cultured under hypoxia. Surprisingly, also HIF-2α mRNA levels were increased in hypoxic cells in the presence of a functional VHL protein. However, expression of HIF-1α mRNA seemed not to be affected by reduced cellular O2 tension.

Figure 2
figure2

Northern blot analysis of total RNA from VHL deficient cell line RCC10 and 11 stably transfected, VHL wildtype expressing cell clones. Lane 1: RCC10 (VHL minus), lanes 2–12: VHL wildtype cell clones RCC10wt53, RCC10wt63, RCC10wt64, RCC10wt69, RCC10wt71, RCC10wt84, RCC10wt90, RCC10wt92, RCC10wt106, RCC10wt115 and RCC10wt119. Ten μg RNA were loaded per lane. The blot was hybridized with 32P-labeled human VHL, VEGF, HIF-1α, HIF-2α and Glut1 probes. Hybridization with β-actin served as loading control

Figure 3
figure3

Northern blot analysis of total RNA from seven different human renal clear cell carcinoma cell lines: VHL deficient cell lines (mut) RCC2, RCC6, RCC10 and RCC11; VHL wildtype cell lines (wt) RCC7, RCC8 and RCC9. Hypoxic stimulation of the cells was carried out for 5 h. Ten μg RNA was loaded per lane. The blot was hybridized with 32P-labeled human VHL, VEGF, HIF-1α, HIF-2α and Glut1 probes. Hybridization with β-actin served as loading control

Tumorigenicity analyses

To analyse the function of the VHL wildtype gene in tumor growth, tumorigenicity of RCC10 cells (VHL minus) and of three stably transfected VHL wildtype cell lines (RCC10wt53, RCC10wt63, RCC10wt64) was characterized by subcutaneous injection of tumor cells into nude mice. Sixteen days after transplantation tumor growth was only detectable in mice treated with VHL negative RCC10 cells, whereas stably transfected RCC10 cells expressing wildtype VHL were non-tumorigenic (data not shown).

Expression of the HIF-1α and HIF-2α proteins in RCC cell lines

To examine the HIF-1α and HIF-2α regulation by the VHL gene product on the protein level, we carried out Western blot analyses using monoclonal HIF-1α and HIF-2α antibodies. Previous studies of nuclear and whole cell extracts showed increased HIF-1α (Wang et al., 1995; Huang et al., 1996) and HIF-2α (Wiesener et al., 1998) protein levels under hypoxic stimulation. To determine the HIF-1α and HIF-2α protein levels in VHL deficient and VHL wildtype expressing cell lines, whole cell extracts were prepared following normoxic and hypoxic incubation for 5 h. In RCC lines 7, 8, 9 expressing the VHL wildtype gene, the HIF-1α protein was detectable only in extracts of hypoxic cells (Figure 4a: lanes 3–8), whereas RCC lines 6 and 10, expressing a mutant VHL gene, showed similar levels of HIF-1α protein expression under normoxic and hypoxic conditions (lanes 1, 2, 9–10). This expression pattern could be confirmed by analyses of the VHL minus RCC10 line and three independent, stably transfected, VHL wildtype expressing cell clones (Figure 4b: RCC10wt69, RCC10wt90, RCC10wt63). The re-introduction of the VHL wildtype gene inhibited the expression of HIF-1α protein under normoxic condition and regained the hypoxic stimulation of HIF-1α expression (lanes 3–8), whereas HIF-1α protein is constitutively expressed in the VHL deficient RCC10 line (lanes 1, 2).

Figure 4
figure4

Expression of HIF-1α protein in different VHL deficient and VHL wildtype expressing RCC cell lines. Cells were cultured in parallel in normoxia (N) or hypoxia (H) for 5 h. Immunoblots of RCC whole cell extracts (100 μg) with HIF-1α antibody (Novus). (a) Lanes 1, 2, 9, 10: VHL deficient cell lines RCC6 and RCC10, lanes 3–8: VHL wildtype cell lines RCC7, RCC8 and RCC9. (b) Lanes 1, 2: VHL minus cell line RCC10, lanes 3–8: stably transfected, VHL wildtype expressing cell clones RCC10wt69, RCC10wt90 and RCC10wt63

Analysis of the expression of the HIF-2α protein revealed that in VHL wildtype expressing RCC lines 7, 8, 9, HIF-2α accumulated only under hypoxic conditions (Figure 5a: lanes 5–10). In contrast, the loss of a functional VHL gene product in RCC lines 2, 6, 10 and 11 led to constitutive expression of HIF-2α independently of the oxygen tension of the cells (lanes 1–4, 10–14). The correlation between up-regulation of HIF-2α protein and loss of function of the von-Hippel-Lindau tumor suppressor gene was confirmed by Western blot analyses of the VHL minus cell line RCC10 and stably transfected, VHL wildtype expressing subclones (Figure 5b). RCC10 (VHL minus) cell extracts revealed constitutive expression of the HIF-2α protein (lanes 1, 2, 9, 10). However, re-introduction of the VHL wildtype gene in six stable cell clones (Figure 5b: RCC10wt53, RCC10wt63, RCC10wt64, RCC10wt69, RCC10wt90, RCC10wt71) inhibited HIF-2α protein expression under normoxic conditions, leading to an expression of HIF-2α that is restricted to hypoxic conditions (lanes 3–8, 11–16).

Figure 5
figure5

Expression of HIF-2α protein in different VHL deficient and VHL wildtype expressing RCC cell lines. Cells were cultured in parallel in normoxia (N) or hypoxia (H) for 5 h. Immunoblots of RCC whole cell extracts (100 μg) with HIF-2α antibody (190b). (a) Lanes 1–4, 11–14: VHL deficient cell lines RCC2, RCC6, RCC10 and RCC11, lanes 5–10: VHL wildtype cell lines RCC7, RCC8 and RCC9. (b) Lanes 1, 2, 9, 10: VHL minus cell line RCC10, lanes 3–8, 11–16: stably transfected, VHL wildtype expressing cell clones RCC10wt53, RCC10wt63, RCC10wt64, RCC10wt69, RCC10wt90 and RCC10wt71

Analyses of HIF-2α protein expression in hemangioblastomas and normal cerebellum

Hemangioblastoma formation depends on mutational inactivation of the VHL tumor suppressor gene. Therefore, we investigated six capillary hemangioblastomas by Western blot analysis for HIF-2α protein expression using a monoclonal HIF-2α antibody. The HIF-2α protein was detected in all hemangioblastomas tested (Figure 6: lanes 2–7), while in normal cerebellum HIF-2α was expressed at lower level (lane 1). Detection of HIF-2α protein in extracts of hypoxic RCC8 cells, characterized as VHL wildtype expressing cell line, served as positive control (lane 8).

Figure 6
figure6

Expression of HIF-2α protein in normal cerebellum and hemangioblastomas. Immunoblot of total protein extracts (100 μg) of tumor and control tissues with HIF-2α antibody (190b). Lane 1: normal cerebellum, lanes 2–7: six different hemangioblastomas, lane 8: whole cell extract (100 μg) of VHL wildtype cell line RCC8 under hypoxic conditions

Immunohistochemistry for HIF-1/2α expression in RCCs and hemangioblastomas

We previously showed up-regulation of HIF-2α mRNA in hemangioblastomas by Northern blot analysis and in situ hybridization (Flamme et al., 1998). Here, we investigated HIF-1α and HIF-2α protein expression in 12 VHL-associated tumors, namely five RCCs and seven hemangioblastomas by immunohistochemistry using a commercially available monoclonal HIF-1α antibody and a polyclonal HIF-2α antibody. In all tumors examined, both HIF-1α and HIF-2α were up-regulated in tumor cell nuclei. In clear cell RCCs, immunolabeling was primarily detected in tumor cells (Figure 7), with few immunoreactive cells visible in the tumor stroma. One papillary renal cell carcinoma, known not to be associated with VHL disease, was used for control purposes and did not show significant expression of HIF-1α or HIF-2α (data not shown). In hemangioblastomas, immunolabeling revealed HIF-1α and HIF-2α expression in stromal cells (Figure 7a,c), the presumed neoplastic component of this tumor. In addition, HIF-2α was expressed at lower levels in the endothelial cell layer of smaller vessels (Figure 7).

Figure 7
figure7

Immunohistochemistry for HIF-1α and HIF-2α in hemangioblastoma (left column) and renal cell carcinoma (right column) with HIF-1α antibody (Novus) and polyclonal HIF-2α antibody. HIF-1α and HIF-2α in tumor cell nuclei of hemangioblastoma (stromal cell component) and RCC, note low cytoplasmic expression of HIF-2α. Moderate HIF-2α expression was also observed in endothelial cells (see 1000× magnification of hemangioblastoma); upper rows: 400×, lower rows: 1000× magnification; insets show negative controls without first antibody

Discussion

In this study we have analysed whether the expression of the hypoxic inducible factors HIF-1α and HIF-2α in seven different human renal clear cell carcinoma cell lines is correlated with the VHL genotype. The RCC lines were investigated for mutations of the VHL tumor suppressor gene by PCR and SSCP analyses of DNA. Sequence analyses revealed mutations in the coding sequence of the VHL gene for RCC6 and RCC10. No mutations could be found for RCC7, RCC8 and RCC9, classifying these cells as VHL wildtype. RCC2 and RCC11 were classified as VHL loss of function cell lines as a result of gene silencing, since neither expression of VHL mRNA (Figure 1) nor of protein could be detected. VHL inactivation due to gene silencing is a common phenomenon, since hypermethylation of CpG islands in the 5′ region leading to inactivation of the VHL gene was reported in 19 per cent of spontaneous renal clear cell carcinomas (Herman et al., 1994).

Northern blot analysis demonstrated that RCC2, RCC6, RCC10 and RCC11, lacking the wildtype VHL protein, contained elevated mRNA levels of the hypoxia inducible genes encoding VEGF and Glut1 (Figure 1). Several studies showed that the VHL wildtype protein negatively regulates the expression of hypoxia inducible mRNAs (Gnarra et al., 1996; Iliopoulos et al., 1996; Siemeister et al., 1996). Accordingly, re-introduction of the wildtype VHL gene to RCC10 cells, where the VHL gene seems to be inactivated by silencing, reduced the levels of VEGF and Glut1 mRNAs (Figure 2).

The function of the VHL wildtype gene in tumor growth in vivo was analysed by determining the tumorigenicity of RCC10 cells (VHL minus) and three stably transfected VHL wildtype subclones in nude mice. Tumor growth was only detectable in mice treated with VHL negative RCC10 cells, whereas stably transfected RCC10 cells expressing wildtype VHL were non-tumorigenic. These findings support a previous report in which the VHL negative 786-0 RCC cell line was reported to be tumorigenic in nude mice. Tumor growth, however, was completely repressed by re-introduction of VHL wildtype (Illiopoulos et al., 1995).

In a previous study of hemangioblastomas, the most frequent tumor manifestations of the hereditary VHL disease, we found that the elevated levels of HIF-2α mRNA expression in stromal cells of the tumor is highly correlated with overexpression of VEGF mRNA (Flamme et al., 1998). This led to the conclusion that loss of the von Hippel-Lindau gene product in these tumors might be related to the accumulation of HIF-2α mRNA and tumorigenesis. We therefore examined here the expression of HIF-1α and HIF-2α in the VHL deficient and VHL wildtype RCC cell lines at the RNA and protein levels.

The expression of endogenous VHL wildtype protein or re-introduction of the VHL wildtype gene by stable transfection into the VHL deficient cell line RCC10 results in increased HIF-1α mRNA expression, whereas HIF-2α mRNA levels are elevated in cells lacking a functional VHL gene product (Figures 1 and 2). Therefore, the expression of the HIF-1α and HIF-2α mRNAs might be regulated by the presence of a functional VHL protein. The differential accumulation of the HIF-α mRNAs implies that HIF-1α and HIF-2α expression could be regulated by a reciprocal mechanism.

Induction of HIF-1α protein by hypoxia occurs primarily by protein stabilization and not by induction of HIF-1α mRNA (Huang et al., 1998). However, prolonged exposure to hypoxia in tissue culture cells and in the whole organism leads to decreased HIF-1α MRNA (Wenger et al., 1997, 1998). Recently an endogenous occurring HIF-1α antisense transcript (termed aHIF) arising from the 3′UTR of HIF-1α mRNA was identified in nonpapillary renal clear cell carcinoma (Trash-Bingham and Tartof, 1999). The aHIF transcript is not overexpressed in papillary renal carcinomas and only nonpapillary tumors are characterized by dysfunction of the VHL gene (Gnarra et al., 1994). The authors speculate that aHIF induction is responsible for the observed decrease in HIF-1α mRNA under prolonged exposure to hypoxia. An accumulation of aHIF in VHL minus cells might represent an auto-feedback mechanism to suppress the HIF-1α mRNA.

The HIF-1α protein is highly unstable under normoxic conditions and hypoxia can prolong its half-life, leading to its accumulation and formation of the HIF-1 complex (Wang et al., 1995; Huang et al., 1996). HIF-2α is regulated in a similar way and showed also high-level induction by hypoxia (Wiesener et al., 1998). Recent reports indicated that the rapid degradation of HIF-1α under normoxic conditions is mediated by the ubiquitin-proteasome system (Salceda and Caro, 1997; Huang et al., 1998). To examine whether the lack of functional VHL protein regulates the hypoxia-inducible factors via protein stability we analysed different RCC lines for HIF-1α and HIF-2α protein expression. Western blotting of whole cell extracts showed that in VHL deficient cells HIF-1α and HIF-2α subunits are constitutively expressed, whereas in cells expressing the VHL wildtype protein or re-introduction of VHL wildtype by stable transfection of RCC10 cells restored the regulation of both HIF-α subunits (Figures 4 and 5).

These Northern and Western blot analyses revealed that under normoxic conditions the levels of both HIF-2α mRNA and protein are elevated in VHL deficient cell lines. It is therefore possible that HIF-2α is regulated at the transcriptional level and protein stability by the VHL gene product. Transcriptional up-regulation of HIF-2α mRNA could also be detected in VHL wildtype expressing cells after hypoxic stimulation (Figure 3), supporting a transcriptional regulation of HIF-2α by VHL.

Mutations of the VHL gene are responsible for the development of hemangioblastomas. To support our data about the up-regulation of HIF-2α mRNA in hemangioblastomas (Flamme et al., 1998), we investigated total protein extracts of different hemangioblastomas for the presence of HIF-2α protein. HIF-2α could be detected at higher levels in all hemangioblastomas analysed when compared with normal cerebellum (Figure 6). In addition, we analysed HIF-1α and HIF-2α protein expression in five RCCs and seven hemangioblastomas by immunohistochemistry (Figure 7). In line with the results of our Western blot analyses, a specific up-regulation of HIF-1α and HIF-2α in the tumor cell component of all tumors examined was found.

Biochemical studies revealed that the VHL protein interacts with elongin B and elongin C (Duan et al., 1995; Kibel et al., 1995). These two proteins, when bound to elongin A, generate the transcriptional elongation complex termed elongin (Aso et al., 1996). Recent reports revealed binding of the VHL-elongin B/C complex to Cullin2 (Cul2). This complex has been shown to be required for the VHL-mediated regulation of the VEGF mRNA (Pause et al., 1997; Lonergan et al., 1998). VHL does not resemble any other known protein but the complex consisting of VHL, elongin B/C, Cul2 and the recently identified Rbx1 protein share homology to the yeast SCF multi protein complex that targets many cell cycle regulatory proteins for ubiquitin-mediated proteolysis (Kamura et al., 1999; Skowyra et al., 1999). The resolution of the crystal structure of the VHL-elongin B/C complex supports the similarities to the SCF complex (Stebbins et al., 1999). Taken together these findings suggest that the VHL tumor suppressor protein is linked to ubiquitin-mediated proteolysis.

The data presented here show that cells lacking the wildtype VHL do not degrade HIF-1α and HIF-2α proteins under normoxic conditions. It is tempting to speculate that VHL in association with the other proteins is involved in the regulation of HIF-1α and HIF-2α protein stability. Maxwell et al. (1999) proposed a model in which VHL forms complexes with the HIF-α subunits under normal cellular O2 tension, thereby targeting the HIF-α subunits for destruction. In addition, ubiquitination of cellular proteins in VHL deficient cells was elevated after glucose deprivation supporting an involvement of VHL in ubiquitin dependent degradation of proteins (Gorospe et al., 1999). Recent findings point to an involvement of the HIF-1 complex in tumor angiogenesis (Maxwell et al., 1997; Carmeliet et al., 1998). Thus the constitutive HIF-1α and HIF-2α activation might be responsible for the prominent angiogenic phenotype observed in VHL associated tumors.

Materials and methods

Cell culture

Human renal clear cell carcinoma cell lines (designated RCC2, 6, 7, 8, 9, 10, 11) were obtained from Dr C Bauer, Zurich, Switzerland. All cells were grown in Dulbecco's modified Eagle's medium (DMEM, Gibco) containing 10% fetal bovine serum at 37°C in a humidified, 10% CO2-containing atmosphere. Transfection was performed using the calcium-phosphate method (Gorman et al., 1983). RCC 10 clonal cell lines stably transfected with pCMV-derived plasmid expressing the VHL wildtype gene (PCR fragment of VHL coding region amplified of g7 DNA was cloned into pcDNA3 (Invitrogen)) were grown in media supplemented with G418 (1 mg/ml). Hypoxic exposure was in a gaspack system (BBL, Microbiology Systems) for 5 h as described previously (Plate et al., 1993).

Specimens

For Western blot analyses, hemangioblastomas were taken from the brain tumor bank of the Department of Neuropathologie, Freiburg University Medical School, Germany. For control purposes normal cerebellum from a patient without neurological disease was obtained postmortem and snap-frozen in liquid nitrogen. Tissue blocks for immunohistochemistry were retrieved from the tissue bank of the Department of Neuropathology, Freiburg University Medical School, and the Institute of Pathology, Erlangen-Nürnberg University Medical School, Germany.

RCC VHL mutation analyses

High molecular weight genomic DNA was extracted from cell cultures by standard procedures. DNA of RCC lines were tested for mutation within the VHL coding sequence by PCR and SSCP analysis as described by Glavac et al. (1996). For PCR exon specific primers were used for exon 1 VHL28/VHL22, for exon 2 I5/I6 and for exon 3 YH1A/6b (Glavac et al., 1996). PCR was performed in a PT 200 MJ Research Thermocycler in 25 μl volumes. SSCP analysis of PCR products were performed non-radioactively using polyacrylamide gels from ETC, Kirchentellinsfurth, Germany and a Pharmacia Multiphor II electrophoresis device with adjustable temperature and silver staining. ETC-Clean Gels were pretreated according to the manufacturer protocol and electrophoresis was run in ETC Disc Buffer at 200 V for 10 min and 600 V for additional 45–60 min at 4°C, 15°C and 25°C. Sequencing of the PCR products was performed with the same primers by the dye terminator method on a 373 DNA sequencer (Sequiserve, Vaterstetten). One hundred μl PCR product was generated and purified by electrophoresis in a 1.9% Nusieve agarose gel (1.5% Nusieve, 0.4% Seakem agarose) in order to remove excess primer. Bands were excised under UV light and DNA was recovered with a Qiaquick kit (Qiagen). Sequencing of both strands was performed to confirm a mutation.

RNA extraction and Northern blot analyses

Total RNA extraction and Northern blot analyses were carried out as described previously (Flamme et al., 1998). Probes for human VHL cDNA (kindly provided by Dr B Zbar, Frederick, MD, USA), human VEGF cDNA (kindly provided by Dr H Weich, Braunschweig, Germany), human HIF-1α cDNA (kindly provided by Dr H Marti, Bad Nauheim, Germany), human HIF-2α cDNA (nt 1680–2344), human Glut1 cDNA (nt 183–683) and β-actin were labeled with 32P-dCTP using a random priming labeling kit (Stratagene).

Protein extraction and Western blot analyses

The total protein extracts and Western blot analyses were carried out as described in Wiesener et al. (1998). Extracts were quantified using the BioRad DC protein assay (BioRad). For immunoblotting, proteins were resolved in 6% SDS–PAA gels and transferred to Immobilon P (Millipore) overnight in 10 mM Tris base, 100 mM glycine, 10% methanol and 0.005% SDS. Membranes were blocked with PBS, 5% fat-free dried milk and 0.1% Tween 20. For HIF-1α detection a monoclonal HIF-1α antibody (Novus) was diluted 1 : 500. For HIF-2α detection monoclonal antibody 190b supernatant (kindly provided by Dr P Maxwell, Oxford, UK) was diluted 1 : 4. Detection was carried out with horseradish peroxidase (HRP)-conjugated goat anti-mouse Igs (Dianova) at 1 : 1000 and enhanced chemiluminescence (SuperSignal Ultra, Pierce).

Tumorigenicity analyses

RCC10, RCC10wt53, RCC10wt63 and RCC10wt64, respectively, were harvested by trypsinization and washed twice in phosphate-buffered saline (PBS). The resulting cell pellets were resuspended at a density of 2.5×107 cells per 300 μl PBS and were injected subcutaneously into two female NCR Balb/c mice per cell line. Growth was observed for 16 days after tumor cell injection. Subcutaneous tumors were excised from the mice and frozen in liquid nitrogen.

Production of polyclonal HIF-2α antibody

A 2.315 bp SacI fragment from the open reading frame of the murine HIF-2α cDNA was cloned in frame into the SacI site of the procaryotic HIS-tag expression vector pQE31 (Qiagen). By Ni-affinity chromatography a 87-kDa 6×HIS-HIF-2α fusion protein was purified from E. coli cell lysates, after induction with IPTG according to the instructions of the manufacturer (Qiaexpressionist, Qiagen). The protein was used for immunization of a rabbit according to standard protocols (Eurogentech) and immunoglobulin from sera was affinity purified over protein-G sepharose (Pharmacia). The antiserum readily detected HIF-2α in the nuclei and cell lysates of transfected cells and did not cross react with HIF-1α in Western blot or immunohistochemistry of cells transiently transfected with HIF-1α and HIF-2α, respectively (I Flamme, unpublished observations).

Immunohistochemistry

Paraffin sections of seven hemangioblastomas and five renal clear cell carcinomas (RCC) were analysed for HIF-1α and HIF-2α expression using a monoclonal HIF-1α antibody (Novus) and the above described polyclonal HIF-2α antibody. Immunohistochemistry was performed according to manufacturer instructions using the DAKO antigen retrieval system (DAKO, Glostrup, Denmark). Endogenous peroxidase was blocked by incubating slides in 100% methanol/1% H2O2 at 4°C for 20 min. Antigen retrieval was performed by microwave cooking in Tris/EDTA. Slides were rinsed in PBS and incubated at room temperature for 30 min in 5% BSA/PBS (Bovine Albumine Serum, Sigma Fraction 5) and for 60 min in 20% NGS/PBS (Normal Goat Serum, Sigma) to block non-specific binding of antibodies. Slides were then incubated overnight in 4°C in 10% NGS/PBS with the primary antibody in a 1 : 200 dilution. Omission of the first antibody served as negative control. Slides were rinsed in PBS/0.1% Triton and incubated for 1 h with 20% NGS/PBS/0.1% Triton. Incubation with biotinylated rat anti-mouse secondary antibodies (Dianova, diluted 1 : 300 in 20% NGS/PBS/0.1% Triton) for 1 h was followed by application of alkaline phosphatase conjugated streptavidin (Vector, Vectastain Kit ABC-AP) for 1 h at room temperature. Binding was visualized with Fast Red (Sigma, Germany). The slides were counterstained with hematoxylin and mounted in ethanol.

References

  1. Aso T, Haque D, Barstead RJ, Conaway RC and Conaway JW. . 1996 EMBO 15: 5557–5566.

  2. Carmeliet P, Dor Y, Herbert J-M, Fukumura D, Brusselmans K, Dewerchin M, Neeman M, Bono F, Abramovitch R, Maxwell P, Koch CJ, Ratcliffe P, Moons L, Jain RK, Collen D and Keshet E. . 1998 Nature 394: 485–490.

  3. Duan DR, Pause A, Burgess WH, Aso T, Chen DYT, Garrett KP, Conaway RC, Conaway JW, Linehan WM and Klausner RD. . 1995 Science 269: 1402–1406.

  4. Ema M, Taya S, Yokotani N, Sogawa K, Matsuda Y and Fujii-Kuriyama Y. . 1997 Proc. Natl. Acad. Sci. USA 94: 4273–4278.

  5. Flamme I, Fröhlich T, von Reutern M, Kappel A, Damert A and Risau W. . 1997 Mech. Dev. 63: 51–60.

  6. Flamme I, Krieg M and Plate KH. . 1998 Am. J. Pathol. 153: 25–29.

  7. Glavac D, Neumann HPH, Wittke C, Jaenig H, Masek O, Streicher T, Pausch F, Engelhardt D, Plate KH, Höfler H, Chen F, Zbar B and Brauch H. . 1996 Hum. Genet. 98: 271–280.

  8. Gnarra JR, Tory K, Weng Y, Schmidt L, Wei MH, Latif F, Li H, Liu S, Chen F, Duh F-M, Lubensky I, Duan DR, Florence C, Pozzatti R, Walther MM, Bander NH, Grossman HB, Brauch H, Pomer S, Brooks JD, Isaacs WB, Lerman MI, Zbar B and Linehan WM. . 1994 Nat. Genet. 7: 85–90.

  9. Gnarra JR, Zhou S, Merrill MJ, Wagner JR, Krumm A, Papavassiliou E, Oldfield EH, Klausner R and Linehan WM. . 1996 Proc. Natl. Acad. Sci. USA 93: 10589–10594.

  10. Gorman CM, Padmanabliam R and Howard BH. . 1983 Science 221: 551–553.

  11. Gorospe M, Egan JM, Zbar B, Lerman M, Geil L, Kuzmin I and Holbrook NJ. . 1999 Mol. Cell. Biol. 19: 1289–1300.

  12. Herman JG, Latif F, Wenig Y, Lerman MI, Zbar B, Liu S, Samid D, Duan DR, Gnarra JR, Linehan WM and Baylin SB. . 1994 Proc. Natl. Acad. Sci. USA 91: 9700–9704.

  13. Huang LE, Arany Z, Livingston DM and Bunn HF. . 1996 J. Biol. Chem. 271: 32253–32259.

  14. Huang LE, Gu J, Schau M and Bunn HF. . 1998 Proc. Natl. Acad. Sci. USA 95: 7987–7992.

  15. Iliopoulos O, Levy AP, Jiang C, Kaelin WG and Goldberg MA. . 1996 Proc. Natl. Acad. Sci. USA 93: 10595–10599.

  16. Iliopoulos O, Kibel A, Gray S and Kaelin WG. . 1995 Nature Med. 1: 822–826.

  17. Kamura T, Koepp DM, Conrad MN, Skowyra D, Moreland RJ, Iliopoulos O, Lane WS, Kaelin WG, Elledge SJ, Conaway RC, Harper JW and Conaway JW. . 1999 Science 284: 657–661.

  18. Kibel A, Iliopoulos O, DeCaprio JA and Kaelin WG. . 1995 Science 269: 1444–1446.

  19. Lonergan KM, Iliopoulos O, Ohh M, Kamura T, Conaway RC, Conaway JW and Kaelin WG. . 1998 Mol. Cell. Biol. 18: 732–741.

  20. Maxwell PH, Dachs GU, Gleadle JM, Nicholls LG, Harris AL, Stratford IJ, Hankinson O, Pugh CW and Ratcliffe PJ. . 1997 Proc. Natl. Acad. Sci. USA 94: 8104–8109.

  21. Maxwell PH, Wiesener MS, Chang G-W, Clifford SC, Vaux EC, Cockman ME, Wykoff CC, Pugh CW, Maher ER and Ratcliffe PJ. . 1999 Nature 399: 271–275.

  22. Neumann HPH, Lips CJM, Hsia YE and Zbar B. . 1995 Brain Pathol. 5: 181–193.

  23. Pause A, Lee S, Worrell RA, Chen DYT, Burgess WH, Linehan WM and Klausner RD. . 1997 Proc. Natl. Acad. Sci. USA 94: 2156–2161.

  24. Plate KH, Breier G, Millauer B, Ullrich A and Risau W. . 1993 Cancer Res. 53: 5822–5827.

  25. Pugh CW, O'Rourke JF, Nagao M, Gleadle JM and Ratcliffe PJ. . 1997 J. Biol. Chem. 272: 11205–11211.

  26. Salceda S and Caro J. . 1999 J. Biol. Chem. 272: 22642–22647.

  27. Semenza GL. . 1998 J. Lab. Clin Med. 131: 207–214.

  28. Semenza GL and Wang GL. . 1992 Mol. Cell. Biol. 12: 5447–5454.

  29. Siemeister G, Weindel K, Mohrs K, Barleon B, Martiny-Baron G and Marmé D. . 1996 Cancer Res. 56: 2299–2301.

  30. Skowyra D, Koepp DM, Kamura T, Conrad MN, Conaway RC, Conaway JW, Elledge SJ and Harper JW. . 1999 Science 284: 662–665.

  31. Stebbins CE, Kaelin WG and Pavletich NP. . 1999 Science 284: 455–461.

  32. Tian H, McKnight SL and Russell DW. . 1997 Genes Dev. 11: 72–82.

  33. Trash-Bingham CA and Tartof KD. . 1999 J. Nat. Cancer Institute 91: 143–151.

  34. Wang GL, Jiang B-H, Rue AE and Semenza GL. . 1995 Proc. Natl. Acad. Sci. USA 92: 5510–5514.

  35. Wenger RH and Gassmann M. . 1997 Biol. Chem. 378: 609–616.

  36. Wenger RH, Kvietikova I, Rolfs A, Gassmann M and Marti HH. . 1997 Kidney Int. 51: 560–563.

  37. Wenger RH, Rolfs A, Spielmann P, Zimmermann DR and Gassmann M. . 1998 Blood 91: 3471–3480.

  38. Wiesener MS, Turley H, Allen WE, William C, Eckardt K-U, Talks KL, Wood SM, Gatter KC, Harris AL, Pugh CW, Ratcliffe PJ and Maxwell PH. . 1998 Blood 92: 2260–2268.

Download references

Acknowledgements

Support was given by DFG grants Pl 158/3-1, 3-2 and Pl 158/4-1 (SPP 1069) to KH Plate, and Fl 223/3-1 to I Flamme, and by grant C4 from the Center for Clinical Research I from Freiburg University Medical School to KH Plate. We thank Manuel Rauter for technical assistance, and Angelika Burger for help with animal experiments.

Author information

Correspondence to Karl H Plate.

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • von Hippel-Lindau (VHL) tumor suppressor gene
  • hypoxia-inducible factors HIF-1α and HIF-2α
  • protein stability
  • renal clear cell carcinoma cell lines

Further reading