1. ¹Human Genome Research Institute and Cancer Research Institute, Seoul National University College of Medicine, 28 Yongon-dong, Chongno-gu, Seoul 110-799, Korea
2. ²Department of Pharmacology, Chosun University College of Medicine, 375 Seusuk-dong, Kwangju 501-759, Korea
3. ³Department of Pharmacology, Seoul National University College of Medicine, 28 Yongon-dong, Chongno-gu, Seoul 110-799, Korea

Correspondence: YS Chun, E-mail: chunys@snu.ac.kr

⁴These two authors contributed equally to this work

Received 9 February 2004; Revised 18 June 2004; Accepted 29 June 2004; Published online 15 November 2004.

Abstract

Hypoxia-inducible factor-1alpha (HIF-1) plays crucial roles in tumor promotion by transactivating approximately 60 kinds of its target genes. Recently, we reported a novel splice variant HIF-1⁷⁸⁵, which is regulated primarily by phorbol ester. This variant can be stabilized under normoxic conditions because it loses an acetylation site Lys532. Its expression was found to promote xenografted tumor growth in nude mice. We here found that the Ras oncogene regulates HIF-1⁷⁸⁵ expression via the Raf/MEK/ERK pathway, and that both phorbol ester and epidermal growth factor also induced HIF-1⁷⁸⁵ via the same pathway. We also identified the nonhypoxic regulatory domain responsible for phorbol ester-induced HIF-1⁷⁸⁵ expression. These results imply that HIF-1⁷⁸⁵ may play an important role in tumor promotion mediated by the Ras oncogene, phorbol ester or tumor growth factors.

HIF-1 is composed of two subunits, HIF-1 and HIF-1, and its transcriptional activity depends mainly on the amount of HIF-1 (Semenza, 1998). In human tumors, HIF-1 is overexpressed as a result of intratumoral hypoxia, or when genetic alterations upregulate oncogenes (HRAS, HER2, FRAP, or CSRC), or downregulate tumor suppressor genes (VHL, PTEN, or p53) (Semenza, 2002). The expression level of HIF-1, in biopsies of various solid tumors, is positively related with tumor aggressiveness, poor vascularity, treatment failure, and mortality (Zhong et al., 1999; Birner et al., 2000). A recent study demonstrating that the growth of grafted tumors is halted by HIF-1 inhibition supports the causative role of HIF-1 in tumor promotion (Yeo et al., 2003).

To date, several isoforms of human HIF-1 have been reported (Gothie et al., 2000; Chun et al., 2001). Recently, we reported a novel HIF-1 isoform, designated HIF-1⁷⁸⁵, which produces a shorter protein by 41 amino acids due to the deletion of 123 nucleotides, representing exon 11 (Chun et al., 2003). Since HIF-1⁷⁸⁵ possesses two transcription activation domains, it retains its normal transactivation activity to hypoxia-response element-containing genes. Importantly, HIF-1⁷⁸⁵ lacks the Lys532 residue, which causes a unique regulation difference between it and wild-type HIF-1, because the acetylation of Lys532 is one of the key requirements for HIF-1 destruction in the presence of oxygen (Jeong et al., 2002). HIF-1⁷⁸⁵, which escapes lysine acetylation, can be stabilized even under normoxic conditions by PMA treatment. Furthermore, PMA was found to strikingly prolong the half-life of HIF-1⁷⁸⁵ to 15.5 min, which is three times longer than that of HIF-1⁷⁸⁵ in the absence of PMA. In addition, the ectopic expression of HIF-1⁷⁸⁵ markedly enhanced the growth of xenografted tumors in immune-deficient mice (Chun et al., 2003). Therefore, we believe that HIF-1⁷⁸⁵ plays an important role in tumor promotion by phorbol ester. However, the mechanism of HIF-1⁷⁸⁵ regulation and the physiological inducer(s) of HIF-1⁷⁸⁵ have not been determined. Therefore, in the present study, we addressed in detail the signal transduction cascade regulating HIF-1⁷⁸⁵, and tried to identify those endogenous molecules that stimulate HIF-1⁷⁸⁵ expression.

First, we re-evaluated the induction of HIF-1⁷⁸⁵ by PMA and the inhibition of this induction by MEK-1 inhibitors. Both MEK-1 inhibitors, that is, PD98059 and U0126, abolished HIF-1⁷⁸⁵ expression induced by PMA, but the PI3K inhibitors wortmannin and LY294002 and the p38 MAPK inhibitors SB203580 and SB212080030 did not (Figure 1a). It was also confirmed that PKA, PKC, PKG, and tyrosine kinase inhibitors did not affect the PMA-induced expression of HIF-1⁷⁸⁵ (data not shown), as reported previously (Chun et al., 2003). In contrast, none of the inhibitors tested affected the hypoxia-induced expression of wild-type HIF-1 (Figure 1b). Hyperthermia is also a strong stimulus capable of nonhypoxic HIF-1⁷⁸⁵ expression (Chun et al., 2003). However, this HIF-1⁷⁸⁵ expression was suppressed by PI3K inhibitors instead of MEK-1 inhibitors (Figure 1c), which indicates that hyperthermia regulates HIF-1⁷⁸⁵ via the PI3K pathway, not via the MEK pathway. Although HIF-1⁷⁸⁵ was not induced distinctly by hypoxia, its expression was enhanced by a PI3K inhibitor, LY294002, reversely (Figure 1d). These results suggest that HIF-1⁷⁸⁵ expression is regulated in different ways by different stimuli.

Figure 1.

The Raf-1/MEK-1/ERK-dependent regulation of HIF-1⁷⁸⁵. Hemagglutinin (HA)-tagged HIF-1 and HA-HIF-1⁷⁸⁵ expression plasmids were constructed as described previously (Chun et al., 2003). HEK 293 cells were transfected with the plasmids using the calcium phosphate method and selected by G418. Cell lines were cultured in Dulbecco's modified Eagle's medium, supplemented with 10% fetal calf serum. HIF-1⁷⁸⁵-expressing (a) or wild-type HIF-1-expressing cells (b) were pre-incubated with various kinase inhibitors for 30 min and then treated with 5 nM PMA (Alexis Biochemicals) for 4 h or 1% O₂ hypoxia for 8 h. W, 200 nM wortmannin (Alexis); LY, 20 M LY294002 (Alexis); PD, 100 M PD98059 (Calbiochem); U, 10 M U0126 (Alexis); SB1, 10 M SB203580 (Calbiochem); SB2, 5 M SB212080030 (Alexis). HIF-1⁷⁸⁵-expressing cells were pre-incubated with various kinase inhibitors for 30 min and then exposed to 42°C (c) or to 1% O₂ hypoxia (d) for 8 h. HIF-1⁷⁸⁵-expressing cells were treated with PMA at 0.1–10 nM for 4 h (e) or treated with 1 nM PMA for 2 min–4 h (f). HIF-1 isoforms and active forms of signaling molecules were detected by Western blotting using anti-HA (BIODESIGN International), anti-phosphoRaf-1 (p-Raf-1, Cell Signaling Technology), anti-phosphoMEK-1 (p-MEK-1, Cell Signaling), and anti-phosphoERK (p-ERK, Cell Signaling) antibodies. Horseradish peroxidase-conjugated antiserum was used as a secondary antibody, and the antigen–antibody complexes were visualized using an ECL kit (Amersham Biosciences). -Actin was used as an internal standard and quantified using mouse monoclonal anti--actin antibody (Santa Cruz Biotechnology). The results presented in each panel are representative of three separate experiments

Full figure and legend (127K)

To determine the nature of the signal transduction pathway mediating the PMA induction of HIF-1⁷⁸⁵, we measured active phosphorylated forms of Raf-1, MEK-1, and ERK proteins. PMA activated the Raf-1/MEK-1/ERK pathway in a dose-dependent manner, and this was well matched by HIF-1⁷⁸⁵ levels (Figure 1e). Increased HIF-1⁷⁸⁵ levels were observed as early as 1 h after PMA treatment, and these continued to increase up to 4 h. Moreover, the Raf-1/MEK-1/ERK pathway was activated immediately after PMA treatment, and this activity was maintained at a high level (Figure 1f). To confirm the involvement of this signaling pathway, plasmids expressing active 12V-Ras, MEK1ED, or 12V-Rac1 were transfected into the HIF-1⁷⁸⁵ stable cell line. 12V-Ras markedly stimulated the expression of HIF-1⁷⁸⁵ in a gene-dose-dependent manner. Of MEK-1 and Rac, both of which are known to be downstream of Ras, only MEK-1 induced HIF-1⁷⁸⁵ expression (Figure 2a). PMA treatment and 12V-Ras transfection in the HIF-1⁷⁸⁵ stable cell line activated all components of Ras/Raf-1/MEK-1/ERK signal transduction (Figure 2b) and resulted in the induction of HIF-1⁷⁸⁵ expression (Figure 2a). HIF-1⁷⁸⁵ expressed by 12V-Ras was diminished by the MEK-1 inhibitors PD98059 and U0126 (Figure 2c), as was PMA-induced expression. These results indicate that the activation of the Ras oncogene may induce the expression of tumor-promoting HIF-1⁷⁸⁵ by stimulating the Raf-1/MEK-1/ERK cascade, and that PMA also causes HIF-1⁷⁸⁵ expression via the same pathway.

Figure 2.

Involvement of the Ras oncogene in HIF-1⁷⁸⁵ expression. Dominant-positive 12V-H-Ras and dominant-positive 12V-Rac1 were constructed as described previously (Cho et al., 2002). Active MEK1ED was kindly donated by Dr Natalie Ahn (University of Colorado, USA). (a) The plasmid was transfected into HIF-1⁷⁸⁵-expressing stable cells with 0.6, 1.25 or 2.5 g of each plasmid, respectively. After 48 h, HIF-1⁷⁸⁵ levels were analysed by Western blotting with anti-HA antibody. (b) In HIF-1⁷⁸⁵-expressing cells transfected with 12V-H-Ras plasmid, we analysed for Ras expression and the activation of its downstream signal cascade by Western blotting. (c) 12V-H-Ras-transfected, HIF-1⁷⁸⁵-expressing cells were incubated with various kinase inhibitors for 12 h on the second day after transfection. The abbreviations, treatment conditions, and assay methods are the same as those detailed in Figure 1

Full figure and legend (71K)

To identify which cytokine induces HIF-1⁷⁸⁵ by activating the Ras pathway, we first examined the effect of serum on HIF-1⁷⁸⁵ expression in cells pre-incubated in serum-free media. After adding fetal calf serum, HIF-1⁷⁸⁵ expression increased in a dose-dependent manner (Figure 3a). This implies the possibility that HIF-1⁷⁸⁵ is expressed by some growth factor(s) contained in the serum. Previously, growth factors, that is, insulin and insulin-like growth factor (IGF) (Zelzer et al., 1998; Feldser et al., 1999), epidermal growth factor (EGF) (Zhong et al., 2000), or prostaglandin E2 (PGE2) (Liu et al., 2002), were found to induce HIF-1 expression in several types of cancer cell lines. Thus, we examined whether these factors stimulate the expression of HIF-1⁷⁸⁵ via Ras–Raf–MEK–ERK signaling. Both EGF and PGE2 strongly stimulated the expression of HIF-1⁷⁸⁵ in a dose-dependent manner, whereas insulin did not (Figure 3b). The expression of HIF-1⁷⁸⁵ was further enhanced in the presence of the serum. In serum-deficient conditions, EGF strongly induced HIF-1⁷⁸⁵ in the pico-molar range, but PGE2 affected HIF-1⁷⁸⁵ expression in the micro-molar range, which is much higher than the effective concentration of EGF. In EGF-treated cells, the levels of activated MEK-1 and ERK increased in the same manner as HIF-1⁷⁸⁵ expression (Figure 3c). Moreover, this effect of EGF on HIF-1⁷⁸⁵ expression was abolished by MEK-1 inhibitors (Figure 3d) or by either dominant-negative Ras or Raf-1 (Figure 3e). These results suggest that EGF is a physiological stimulus for HIF-1⁷⁸⁵ expression via the Ras/Raf-1/MEK-1/ERK pathway.

Figure 3.

Growth factor-stimulated expression of HIF-1⁷⁸⁵ via the Ras/Mek-1 pathway. (a) HIF-1⁷⁸⁵-expressing cells were pre-incubated in serum-free medium for 24 h, and then treated with 0.5–10% of the fetal calf serum (GIBCO/BRL). HIF-1⁷⁸⁵ expression was analysed by Western blotting using anti-HA antibody. (b) The cells were treated with insulin (Sigma-Aldrich), PGE2 (Sigma-Aldrich), or EGF (Sigma-Aldrich) at various concentrations for 4 h. The concentration ranges of the growth factor used were 1–50 nM for insulin, 1–50 M for PGE2, or 8–300 pM for EGF. (c) In the cells treated with 8–300 pM EGF for 4 h, phosphoMEK and phosphoERK levels were analysed by Western blotting. (d) The cells were pre-incubated with various kinase inhibitors for 30 min, and then treated with 300 pM EGF for 4 h. The abbreviations, treatment conditions, and assay methods used are the same as those detailed in Figure 1. (e) Dominant-negative 12N-H-Ras and dominant-negative RafR89L plasmids were constructed as described previously (Alahari et al., 2000; Cho et al., 2002). The plasmid was transfected into HIF-1⁷⁸⁵-expressing cells using 1, 2.5, or 5 g of each plasmid, respectively. After stabilizing for 48 h, the cells were treated with 150 pM EGF

Full figure and legend (79K)

How is HIF-1⁷⁸⁵ stabilized by PMA? To answer this question, we firstly examined whether the prolyl hydroxylation process is involved in HIF-1⁷⁸⁵ regulation, because this is the key step for HIF-1 destabilization. Although we deleted each or both of two proline residues (P402 and P564), these mutants were still stabilized by PMA (Figure 4a). This suggests that PMA regulates HIF-1⁷⁸⁵ stability through a novel pathway other than prolyl hydroxylation. We secondly examined whether the oxygen-dependent degradation domain (ODDD) other than two proline hydroxylation sites is involved in HIF-1⁷⁸⁵ regulation. The mutant protein lacking amino acids 406–512 was slightly induced by either hypoxia or PMA treatment (Figure 4a). However, the PMA induction of this mutant was markedly diminished, compared to the extent of PMA induction of HIF-1⁷⁸⁵. Figure 4b also demonstrated that the 406–512 domain is responsible for PMA stabilization of HIF-1⁷⁸⁵ via the MEK/ERK pathway.

Figure 4.

Identification of PMA-regulated domain of HIF-1⁷⁸⁵. (a) To rule out the involvement of proline hydroxylation in PMA induction of HIF-1⁷⁸⁵, each proline hydroxylation site in HIF-1⁷⁸⁵ was blocked by deleting the amino-acid motif (397–405, NP; 513–595, CP), or both sites were blocked by deleting both motifs (NPCP). The ODDD other than proline hydroxylation sites was deleted from HIF-1⁷⁸⁵ (406–512). HEK293 cells stably expressing the mutated plasmids were incubated under normoxic (C), hypoxic (H), or were treated with 2 nM PMA for 4 h. (b) To identify the domain regulated by PMA, cells were transfected with pHIF-1⁵¹² or pHIF-1³⁹⁶, and treated with various concentrations of PMA for 4 h. The cells were preincubated with 100 M PD98059 (PD) or 10 M U0126 (U) for 30 min and then treated with 2 nM PMA. The total cell lysates were analysed using Western blotting with anti-HA antibody. All plasmids expressing mutant proteins were constructed using a PCR-based mutagenesis kit (Stratagene). (c) Summary of HIF-1⁷⁸⁵ regulation. EGF, PMA, and gain-of-function of the Ras gene stimulate Ras/Raf/MEK/ERK pathway, which in turn targets a specific domain (406–512) and stabilizes HIF-1⁷⁸⁵ protein. HIF-1⁷⁸⁵ translocates to the nucleus and then dimerizes with ARNT and transactivates its downstream genes, which may promote tumor growth and angiogenesis

Full figure and legend (84K)

In cancer cells, PI3K is regarded a main signaling molecule that mediates wild-type HIF-1 induction by growth factors (Zhong et al., 2000; Chun et al., 2002). This mechanism was first demonstrated in prostate cancer cells, in which HIF-1 is constitutively expressed even under aerobic conditions (Zhong et al., 2000). Growth factors, including EGF and insulin/IGF, are known to bind to their receptors and to activate receptor tyrosine kinases, which in turn activate the PI3K/AKT/mTOR pathway. Finally, mTOR stimulates the expression of HIF-1 (Zhong et al., 2000; Hudson et al., 2002). In addition to the PI3K pathway, the MEK/ERK pathway can also enhance HIF-1 activity, but this is not due to the stimulation of HIF-1 expression. The MEK/ERK pathway promotes the nuclear translocation of HIF-1 in PGE2-treated cells (Liu et al., 2002), or further enhances the transcriptional activity of HIF-1 induced by hypoxia (Hur et al., 2001). In contrast to wild-type HIF-1, HIF-1⁷⁸⁵ expression was attributed to the activation of the MEK/ERK pathway, rather than PI3K activation. Moreover, a domain containing 406–512 aa is likely to be a target site of the MEK/ERK pathway, as illustrated in Figure 4c. However, the detailed interaction between this domain and the MEK/ERK pathway remains to be investigated.

The Ras oncogene has been shown to be important in both the genesis and maintenance of solid tumors (Shapiro, 2002). In about 30% of human tumors, Ras proteins are present in structurally altered forms that enable them to continuously provide cells with mitogenic signals, even without ongoing stimulation by growth factors. In addition, the angiogenic switch in Ras-transformed cells is another critical control point for tumor expansion, that is, rapid tumor growth and metastases (Rak et al., 1995). Tumor angiogenesis is prompted by the hypoxic microenvironment through VEGF induction in both tumor and surrounding stromal cells. VEGF induction is potentiated by activated Ras (Rak et al., 1995). The present and our previous (Chun et al., 2003) works support the possibility that HIF-1⁷⁸⁵ is an effector molecule that enables VEGF expression by Ras in Ras-transformed cells or in growth factor-stimulated cancer cells.

In conclusion, HIF-1⁷⁸⁵ expression was attributed to the activation of Ras oncogene and signal transduction via Raf-1/MEK-1/ERK. Both PMA and EGF induced HIF-1⁷⁸⁵ expression via the same pathway. Taken together with our previous results, these results suggest that HIF-1⁷⁸⁵ is an important HIF-1 isoform that is involved in tumor promotion in Ras-transformed or PMA/EGF-stimulated cells.