Induction of Hypoxia-inducible Factor 1 Activity by Muscarinic Acetylcholine Receptor Signaling*

From the ‡Department of Anesthesia, The Tazuke Kofukai Medical Research Institute Kitano Hospital, 2-4-20, Ohgimachi, Kita-ku, Osaka 530-8480, Japan, §Program in Vascular Cell Engineering, The Johns Hopkins University School of Medicine, Baltimore, MD 21205, the Department of Anesthesia, Kyoto University Hospital, Kyoto University, Sakyo-ku, Kyoto 606-8507, Japan, and the **COE Formation for Genomic Analysis of Disease Model Animals with Multiple Genetic Alterations, Kyoto University Graduate School of Medicine, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8507, Japan

Hypoxia activates a number of genes that are important in cellular and tissue adaptation to low oxygen conditions (1). These genes include erythropoietin, glucose transporters, glycolytic enzymes, and vascular endothelial growth factor (VEGF). 1 The hypoxic expression of these different genes is controlled at the transcriptional level by the ubiquitously expressed transcription factor hypoxia-inducible factor 1 (HIF-1) (2). HIF-1 is a heterodimer composed of a constitutively expressed HIF-1␤ subunit and an inducibly expressed HIF-1␣ subunit (3). The regulation of HIF-1 activity occurs at multiple levels in vivo. Among these, the mechanisms regulating HIF-1␣ protein expression and transcriptional activity have been most extensively analyzed (4). The von Hippel-Lindau tumor-suppressor protein (VHL) has been identified as the HIF-1␣-binding component of a ubiquitin-protein ligase that targets HIF-1␣ for proteasomal degradation in non-hypoxic cells (5). Under hypoxic conditions, the hydroxylation of specific proline and asparagine residues in HIF-1␣ is inhibited due to substrate (O 2 ) limitation, resulting in HIF-1␣ protein stabilization and transcriptional activation (6,7). The iron chelator deferrioxamine (DFX) inhibits the prolyl and asparaginyl hydroxylases, which contain Fe 2ϩ at their catalytic sites, causing HIF-1␣ stabilization and transactivation under normoxic conditions (6,8).
In this study, we demonstrate that muscarinic acetylcholine receptor (mAchR)-mediated signals induce HIF-1 activation in a receptor-subtype specific manner using HEK293 cells transfected with various types of mAchR. Signaling via M1-and M3but not M2-or M4-AchRs promote accumulation and transcriptional activation of HIF-1␣. We also provide evidence that the activation is dependent on G␣-or G␤␥-dependent and tyrosine kinase, MAPK, and PI3K activity.
Inhibitor Treatments-PD98059, SB203580, LY294002, GF109203X, or genistein was added 1 h before exposure to CCH or 1% O 2 . CHX was added to the medium of HEK293 cells that were treated with CCH, or DFX for 4 h, and whole cell extracts were prepared at 15, 30, and 60 min (12,13).
RNA Blot Hybridization-Total RNA was extracted from HEK293 cells using TRIzol reagent (Invitrogen) 24 h after CCH stimulation and 48 h after transfection of expression plasmid encoding AchR. 10-g aliquots of RNA were fractionated by electrophoresis in 1.5% agarose, 2.2 M formaldehyde gels, transferred to Hybond N ϩmembranes (Amersham Biosciences), and hybridized with a 32 P-labeled human HIF-1␣ or VEGF cDNA probe as described previously (13). Reporter Gene Assays-Reporter assays were performed in HEK293 cells as described previously (29 -31). Cells were transferred to 24-well plates at a density of 5 ϫ 10 4 cells per well on the day before transfection. FuGENE 6 reagent (Roche Applied Science) was used for transfection. For luciferase assay, each mAchR, the reporter gene plasmid, and the control plasmid pTK-RL (Promega), containing a thymidine kinase promoter upstream of Renilla reniformis luciferase coding sequences, were pre-mixed with the transfection reagent. In each assay the total amount of DNA was held constant by addition of empty vector. After treatment, the cells were harvested and the luciferase activity was determined using the Dual-Luciferase Reporter Assay System (Promega). The ratio of firefly to Renilla luciferase activity was determined. For each experiment, at least two independent transfections were performed in triplicate. ␤-galactosidase (␤-gal) activity was determined using a commercial assay system (Roche Applied Science) (32). To normalize ␤-gal activity of each sample, pGL-Control plasmid was co-transfected with a ␤-gal-coding plasmid. Net ␤-gal count of each sample was divided by its luciferase count and normalized mean count Ϯ S.D. of three independent transfections is shown as relative activity.
Metabolic Labeling Assay-The protocol is described elsewhere (12). Briefly, HEK293 cells were plated in a 10-cm dish, and transfected with M1-AchR plasmid, and 18 h later the cells were serum-starved for 20 h. The cells were pretreated with 100 M CCH or IGF-1 for 30 min in methionine-free Dulbecco's modified Eagle's medium. [ 35 S]Met-Cys was added to a final concentration of 0.3 mCi/ml, and the cells were pulselabeled for 20 -40 min and then harvested. Whole cell extracts were prepared, and 1 mg of extract was precleared with 60 l of protein A-Sepharose for 1 h. Twenty microliters of anti-HIF-1␣ antibody H1␣67 was added to the supernatant and rotated overnight at 4°C. Forty microliters of protein A-Sepharose was added, rotated for 2 h at 4°C, pelleted, and washed five times with 1 ml of radioimmune precipitation assay buffer. The samples were analyzed by SDS-polyacrylamide gel electrophoresis. The gel was dried and exposed to x-ray film.

Muscarinic Acetylcholine Receptors Induce HIF-1 Activity in a Receptor Subtype-specific Manner under Non-hypoxic Condi-
tions-Because HEK293 cells express only low levels of mAchRs (33), it is possible to examine the effect of individual mAchR signaling on HIF-1 activation by overexpression of each mAchR. HEK293 cells were transfected with expression plasmid encoding the M1-, M2-, M3-, or M4-AchR and stimulated by the addition of the cholinergic agonist CCH. As shown in Fig. 1A, p42 ERK2 /p44 ERK1 MAPKs were activated in response to treatment with CCH (100 M) (lanes 4, 6, 8, and 10). CCH simulation also induced transcriptional activation of Elk-1 in mAchR-expressing HEK293 cells (data not shown). Together, these data demonstrate that each mAchR subtype was functionally expressed in HEK293 cells. Furthermore, the levels of  1 and 2) or lysates prepared from M1-expressing HEK293 cells that were untreated (lane 3) or exposed to 100 M CCH (lane 4) or from M2-expressing HEK293 cells that were untreated (lane 5) or exposed to 100 M CCH (lane 6). Glutathione-Sepharose beads were used to capture GST or GST-HIF-1␣ and the presence of bound VHL in the samples was determined by PAGE and autoradiography. One-fifth of the input VHL protein was also analyzed. ERK1/2 phosphorylation in response to CCH stimulation were similar, suggesting that the different mAchRs were expressed at comparable levels.
Using this system we investigated impact of mAchR stimulation on activation of HIF-1. 100 M CCH increased HIF-1␣ protein levels in M1-or M3-AchR-but not in M2-or M4-AchRexpressing HEK293 cells (Fig. 1B). In contrast, HIF-1␤ protein levels were not affected by the binding of CCH to mAchR in HEK293 cells. The induction of HIF-1␣ expression was inhibited by atropine, demonstrating that the effect is receptor agonist-specific (Fig. 1C). The M1-or M3-AchR signal induced the accumulation of HIF-1␣ in a CCH dose-dependent manner to 100 nM (Fig. 1D). In M1-transfected HEK293 cells exposed to CCH (100 M), HIF-1␣ protein levels peaked at 4 h (Fig. 1E). We next investigated involvement of mAchR in HIF-1␣ protein accumulation in SK-N-SH human neuroblastoma cells, which endogenously express M3-AchR. Treatment with CCH (100 M) induced accumulation HIF-1␣ protein in 4 h and this accumulation was inhibited by treatment with atropine (Fig. 1F). Thus mAchR-mediated signals induced HIF-1␣ protein accumulation under non-hypoxic conditions. M1-or M3-AchR Stimulation Activates HIF-1-dependent Gene Expression-We examined the impact of mAchR system on HIF-1-dependent gene expression. As shown in Fig. 2, VEGF gene expression was induced in M1-AchR-expressing HEK293 cells exposed to CCH (100 M) or to the iron chelator DFX. The expression of HIF-1␣ mRNA was not affected by any treatment.
Next, HEK293 cells were transfected with reporter plasmid p2.1, which contains a HIF-1-dependent HRE, or p2.4, which contains a mutated, non-functional HRE. Stimulation of M1-or M3-AchR with CCH induced HRE-dependent expression comparably to 100 M DFX treatment (Fig. 3A). The activation was CCH dose-dependent (Fig. 3B). In contrast, the reporter p2.4 was not activated by the treatment. Expression of a dominant negative form of HIF-1␣ markedly reduced p2.1 reporter gene expression, demonstrating that gene activation was both HREand HIF-1-specific (Fig. 3C). CCH treatment of cells expressing M1-AchR also induced expression of the pVEGF-KpnI reporter, which contains promoter sequences encompassing nucleotides Ϫ2274 to ϩ379 relative to the transcription start site of the VEGF gene. In contrast, CCH had no effect on VEGF promoter activity in cells transfected with expression vector M2-AchR or empty vector (Fig. 3D).
Involvement of G Proteins in M1-or M3-AchR Stimulation-Because mAchRs are G protein-coupled receptors, we examined involvement of G proteins in the activation of HIF-1. Because M1-or M3-AchR mainly couples with G␣ q protein in HEK293 cells (33,34), we first examined involvement of G␣ q in the process. Expression of a constitutively activated form of G␣ q (G␣ q -Q209L) induced the accumulation of HIF-1␣ protein (Fig.  4A, lane 3) in HEK293 cells. Expression of G␣ 12 -Q229L or G␣ 13 -Q226L also induced the accumulation of HIF-1␣ (lanes 4 and 5). M1-AchR-mediated HIF-1␣ accumulation was not affected by treatment with pertussis toxin, which blocks interaction of G␣ i GTPase with receptors (Fig. 4B). Stimulation of the endogenous ␤-adrenergic receptor with the ␤-adrenergic agonist isoproterenol did not induce HIF-1␣ protein accumulation, suggesting that G␣ s -mediated signaling did not contribute to HIF-1␣ accumulation (Fig. 4C) (35). Expression of a constitutively activated form of G␣ q -, G␣ 12 -, G␣ 13 -, or G␤␥-induced HRE-dependent reporter gene expression in HEK293 cells (Fig. 4D).
To further investigate whether mAchR signaling affected HIF-1␣ protein half-life, mAchR-expressing HEK293 cells were treated with CCH, IGF-1, or DFX for 4 h to induce HIF-1␣ expression, CHX was added to block ongoing protein synthesis, and cell lysates were prepared for immunoblot assay (Fig. 5B).
In the presence of CHX, the half-life of HIF-1␣ was more than 60 min in DFX-treated cells, ϳ30 min in CCH-treated cells, and less than 15 min in IGF-1-treated cells. These results indicate that M1-AchR stimulation induced accumulation of HIF-1␣ by increasing protein half-life.
To analyze the rate of HIF-1␣ synthesis, serum-starved FIG. 7. Effects of muscarinic signaling on HIF-1␣ transactivation domain function. Constructs encoding the DNA-binding domain (amino acids 1-147) of the yeast transcription factor GAL4 fused to the indicated amino acids of HIF-1␣ were analyzed for their ability to transactivate reporter gene pG5E1bLuc, which contains five GAL4binding sites. HEK293 cells were co-transfected with muscarinic receptors (200 ng), pTK-RL (10 ng), GAL4E1bLuc (150 ng), and GAL4-HIF-1␣ fusion protein expression plasmids (200 ng). Cells were exposed to CCH (100 M) for 16 h and harvested. The ratio of firefly:Renilla luciferase activity was determined and normalized to the value obtained from untreated cells transfected with plasmid encoding GAL4-(1-147) to obtain the relative luciferase activity (RLA).
HEK293 cells were pretreated with CCH or IGF-1 for 30 min and then pulse-labeled with [ 35 S]Met-Cys for 40 min, followed by immunoprecipitation of HIF-1␣ (Fig. 5C). In contrast to control serum-starved cells, 35 S-labeled HIF-1␣ was clearly increased in CCH-treated cells as well as IGF-1-treated cells (Fig. 5C, upper). Expression of HIF-1␣ mRNA was not affected during treatment (lower). Thus, both the cycloheximide and metabolic labeling experiments provide evidence for increased synthesis of HIF-1␣ in response to M1-mediated signal.
Effect of Kinase Inhibitors on AchR-mediated HIF-1 Activation-To investigate further the molecular mechanisms whereby muscarinic receptors activate HIF-1, M1-AchR-expressing HEK293 cells were stimulated by CCH under treatment with LY294002, PD98059, SB203580, GF109203X, or genistein which are selective pharmacologic inhibitors of PI3K, MEK, p38 MAPK, PKC, and tyrosine kinase activity, respectively. As shown in Fig. 8A Next, we investigated the effect of the inhibitors on HREdependent gene expression using the HRE-dependent p2.1 reporter (Fig. 8B). Genistein or PD98059 almost completely suppressed the gene expression by CCH. GF109203X had a partial inhibitory effect. In contrast, treatment with LY294002 or SB203580 did not have any inhibitory effect. The transcriptional activity of HIF-1␣ was examined using pGAL4/HIF-1␣-(531-826) and pG5E1bLuc (Fig. 8C). PD98059 or GF109203X almost completely suppressed HIF-1␣ TAD activation induced by CCH. In contrast, treatment with genistein, LY294002, or SB203580 did not block TAD function induced by CCH.
Previously, we demonstrated that small G protein Ras and Rac1 are involved in hypoxia-induced HIF-1 activation process (29). We next examined involvement of Ras and Rac1 in the CCH-induced HIF-1 activation (Fig. 8D). Expression of dominant negative forms of Ras (Ras N17 ) or Rac1 (Rac1 N17 ) inhibited CCH-induced HRE-dependent gene expression in M1-AchR-expressing HEK293 cells. Expression of a dominant negative form of MEK1 or MEK5 also inhibited CCH-induced gene expression (Fig. 8E), which is consistent with the observed effects of PD98059 in Fig. 8B. DISCUSSION Molecular cloning studies have revealed the existence of five distinct muscarinic receptor subtypes referred to as M1-M5 (33). The M1-M5 receptors are members of the GPCR superfamily. Although even numbered receptors (M2 and M4) are selectively coupled to G proteins of the G␣ i /G␣ o family, the odd-numbered receptors (M1, M3, and M5) are preferentially linked to G␣ q /G␣ 11 proteins (16,17). Our results demonstrate that muscarinic acetylcholine receptor-mediated signals induce HIF-1 activation under non-hypoxic conditions in a receptor subtype-specific manner. Only odd numbered receptors induced HIF-1␣ protein expression and transcriptional activity in a receptor-ligand-dependent manner in HEK293 cells. Expression of a constitutively activated form of G␣ q , G␣ 12 , G␣ 13 , or G␤␥ is sufficient to induce accumulation of HIF-1␣ and HIF-1-mediated transcription (Fig. 4D). The selective effect of M1-or M3-versus M2-or M4-AchR signaling does not appear to be related to the level of receptor expression based on the ERK activation (Fig. 1A). HIF-1␣ protein expression level is determined by the balance between protein synthesis and degradation (12,13). CCH stimulation of M1-expressing HEK293 cells increased the half-life of HIF-1␣ protein compared with that in IGF-1-treated cells although the effect was less than that induced by DFX, which inhibits HIF-1␣ prolyl hydroxylases (Fig. 5B). CCH also induced stabilization of HIF-1␣-(548 -603)-␤-gal fusion protein in HEK293 cells expressing M1-AchR (Fig. 5A). Moreover, lysate of M1-expressing HEK293 cells stimulated by CCH was less effective than lysate from unstimulated cells in promoting the interaction between HIF-1␣ and VHL. Together, these results suggest that M1-AchR signaling regulates HIF-1␣ hydroxylation, ubiquitination, and/or proteasomal degradation. However, the effect of mAchR signaling on HIF-1␣ accumulation is not explained fully by the stabilization mechanism, because the effect on the half-life of HIF-1␣ is much less that that induced by DFX (Fig. 5B). We have reported that certain growth factors such as IGF-1 (12,13), prostaglandin E 2 (14), and the nitric oxide donor NOC18 increase the rate of HIF-1␣ protein synthesis rather than increasing the stability of HIF-1␣. The pulse-labeling data (Fig. 5C) indicate that mAchR stimulation also increases the rate of HIF-1␣ synthesis. The dual effects on both synthesis and stability account for the high levels of HIF-1␣ induced by mAchR signaling.
Src family kinases are also activated by G␣ q /G␣ 11 -coupled receptors via the proline-rich tyrosine kinase 2 (PKY2) (36). v-Src and Ras V12 signaling in RCC4 and 786 -0 cells has been shown to stabilize HIF-1␣ by inhibiting hydroxylation of Pro-564 (37). The effects of dominant negative forms of Ras and Rac1 (Fig. 8D) suggest that similar mechanisms may be involved in AchR-mediated HIF-1 activation. We demonstrate that the activation of HIF-1 in response to muscarinic signaling is suppressed by the PKC inhibitor GF109203X or the tyrosine kinase inhibitor genistein. (Fig. 8). M1 and M3 muscarinic receptors stimulate phospholipase C ␤ , which hydrolyzes phosphatidylinositol 4,5-bisphosphate to generate the second messengers diacylglycerol and inositol 1,4,5-triphosphate (34) causing PKC activation. In fact, genistein or GF109203X inhibits M1-AchR-mediated HIF-1␣ accumulation. Taken together, the pathway may lead to HIF-1␣ accumulation (Fig. 9). Regulation of HIF-1 activity involves changes in both the protein expression and transcriptional activity of HIF-1␣ (8,20,29). Our data analyzing transactivation mediated by Gal4-HIF-1␣-TAD fusion proteins demonstrate that muscarinic receptor signaling also induces HIF-1␣ TAD activity under nonhypoxic conditions. A major determinant of TAD function is the interaction between HIF-1␣ and the coactivators p300/ CBP, which regulated by O 2 -dependent hydroxylation of Asn-803 by FIH-1 (7,20). TAD activity is also regulated by a MAPK-dependent mechanism (29,38) that enhances recruitment of p300/CBP (39). The MEK inhibitor PD98059, which is a selective pharmacologic inhibitor of MEK1 and MEK5, blocked muscarinic signal-induced HIF-1, suggesting a link between muscarinic signaling, MEK/ERK, and HIF-1␣. The data presented in Fig. 8E suggest that MEK1 and MEK5 cooperatively play a critical role in this process. The signal transduction pathway leading from muscarinic receptor activation to HIF-1␣ TAD activation is also sensitive to the PKC inhibitor GF109203X.
The M1-M4 receptors are widely expressed throughout the central and peripheral nervous systems (34). The M1 receptor is found in greatest abundance in the cortex and hippocampus where it constitutes 40 -50% of the total mAchR. Activation of HIF-1 may provide a mechanism to increase glucose uptake and/or perfusion in response to increased neuronal activity via increased transcription of genes encoding glucose transporters and angiogenic factors such as VEGF. In PC12 pheochromocytoma cells, HIF-1 serves as a neuroprotective factor against amyloid ␤ peptide, which is involved in the pathogenesis of Alzheimer's disease (40). Notably, the M1 receptor is coupled with processing the amyloid precursor protein (41). Finally, mice with a deletion of the HIF-1 binding site within the HRE of the Vegf gene develop motor neuron degeneration similar to amyotrophic lateral sclerosis (42), indicating an important role for HIF-1 in VEGFmediated motor neuron survival. Further studies are required to determine whether HIF-1 activation by muscarinic receptor signaling may play an important role in regulating cell survival in neurodegenerative disorders.