Aspartyl-(asparaginyl) β-Hydroxylase, Hypoxia-Inducible Factor-1α and Notch Cross-Talk in Regulating Neuronal Motility

Aspartyl-(Asparaginyl)-β-Hydroxylase (AAH ) promotes cell motility by hydroxylating Notch. Insulin and insulin-like growth factor, type 1 (IGF-I) stimulate AAH through Erk MAP K and phosphoinositol-3-kinase-Akt (PI3K-Akt). However, hypoxia/oxidative stress may also regulate AAH . Hypoxia-inducible factor-1alpha (HIF-1α) regulates cell migration, signals through Notch, and is regulated by hypoxia/oxidative stress, insulin/IGF signaling and factor inhibiting HIF-1α (FIH) hydroxylation. To examine cross-talk between HIF-1α and AAH , we measured AAH , Notch-1, Jagged-1, FIH, HIF-1α, HIF-1β and the hairy and enhancer of split 1 (HE S-1) transcription factor expression and directional motility in primitive neuroectodermal tumor 2 (PNET2) human neuronal cells that were exposed to H2O2 or transfected with short interfering RNA duplexes (siRNA) targeting AAH , Notch-1 or HIF-1α. We found that: (1) AAH , HIF-1α and neuronal migration were stimulated by H2O2; (2) si-HIF-1α reduced AAH expression and cell motility; (3) si-AAH inhibited Notch and cell migration, but not HIF-1α and (4) si-Notch-1 increased FIH and inhibited HIF-1α. These findings suggest that AAH and HIF-1α crosstalk within a hydroxylation-regulated signaling pathway that may be transiently driven by oxidative stress and chronically regulated by insulin/IGF signaling.

with increased AAH or HIF-1α expression, whereas 22 μM H 2 O 2 was within the optimum concentration range for stimulating these proteins. Using the ATP Luminescence Motility and Invasion (ALMI) assay, 28 we observed that treatment with H 2 O 2 significantly altered cell motility (F = 10.9, 3 df, p = 0.004), and that cells treated with 22 μM H 2 O 2 had significantly higher mean directional motility indices relative to cells that had been treated with vehicle (p < 0.01) or 13.2 μM (p < 0.05) H 2 O 2 (Fig.  1F). In addition to total cell motility, the ALMI assay measures the percentages of motile adherent and motile non-adherent cells, and therefore provides information on cell adhesion. 28 The same experiments demonstrated that treatment with 22 μM H 2 O 2 mainly increased the percentages of motile-adherent cells (p < 0.01), indicating that the increased motility was not caused by loss of adhesion (Fig. 1F).
Inhibition of AAH or HIF-1α expression impairs neuronal motility. PNET2 cells transfected with short interfering RNA duplexes targeting AAH (si-AAH) or HIF-1α (si-HIF-1α) had significantly lower mean total motility indices relative to si-Scr (negative control) transfected cells ( Fig. 2A). The major inhibitory effects of si-AAH and si-HIF-1α on motility were related to the motile-adherent populations (Fig. 2B), whereas the percentages of motile non-adherent cells remained relatively unchanged compared with si-Scr transfected control cells (Fig. 2C).
Cellular ELISA studies revealed that si-AAH transfection significantly increased the mean levels of HIF-1α, HIF-1β and Jagged-1 and decreased FIH, β-actin, AAH and Humbug immunoreactivities ( Table 2). In contrast, overexpression of AAH in cells transfected with recombinant plasmid containing full-length AAH cDNA significantly increased Notch-1, but decreased HIF-1β protein ( Table 3). Cells transfected with si-Notch had significantly increased levels of HIF-1α and HIF-1β protein expression and reduced levels of Notch-1, Jagged-1, AAH and Humbug (a catalytically inactive homolog of AAH) 7,11,29 immunoreactivity (Table 4). Finally, transfection with si-HIF-1α significantly reduced HIF-1α immunoreactivity while significantly increasing HIF-1β, FIH, β-Actin, Notch-1, Jagged-1 and AAH immunoreactivity ( Table 5). These results are consistent is composed of two subunits: HIF-1β, which is constitutively expressed, and HIF-1α, which is tightly regulated by oxygendependent prolyl hydroxylases. 13 Under normoxic conditions, HIF-1α's transcriptional activity is negatively regulated by the asparaginyl hydroxylase, factor inhibiting HIF (FIH). 14,15 FIH mediates this effect by hydroxylating HIF-1α, thereby enabling HIF-1α's interaction with the von Hippel-Lindau protein complex, resulting in its ubiquitination and degradation via the proteasomal pathway. 16 However, low oxygen tension is rate-limiting for prolyl hydroxylase activity. Consequently, under hypoxic conditions, HIF-1α becomes stabilized due to reduced ubiquitination and proteosomal degradation. 17 In addition, HIF-1α gene expression is stimulated by insulin, IGF-1 and IGF-2. 18 Transcriptionally activated HIF-1α binds to hypoxia-responsive elements (HRE) in the promoter or enhancer regions of hypoxiainducible genes such as insulin-like growth factor, type 2 (IGF-2), erythropoietin and vascular endothelial growth factor. 13,19 Previous studies linked activation of HIF-1 signaling to increased cell motility in both malignant neoplastic cells, [20][21][22] and cellular constituents required for skin wound healing. 23 Moreover, other studies showed that Notch signaling is sensitive to oxygen tension 24 and can be activated by hypoxia. [25][26][27] Finally, Gustafsson et al. 25 demonstrated that HIF-1α could interact with Notch's intracellular domain and induce Notch-mediated downstream responses. These findings drew our attention because both AAH and HIF-1 belong to hydroxylation signaling networks and mediate their effects through Notch. Moreover, in exploratory studies, we found that AAH expression was stimulated by oxidative stress, and that mild oxidative stress increased cell motility, suggesting potential cross-talk between HIF-1α and AAH signaling pathways. We now characterize the inter-relationships among AAH, HIF-1α, Notch and oxidative stress with respect to motility in human central nervous system (CNS) derived neuronal cells.

Results
Oxidative stress stimulates AAH and HIF-1α protein expression. PNET2 human CNS-derived primitive neuroectodermal tumor 2 (PNET2) neuronal cells were seeded into 96-well plates and treated with 0-45 μM H 2 O 2 for 20 h to examine the effects of oxidative stress on AAH, HIF-1α, HIF-1β and FIH expression using a cellular enzyme-linked immunosorbant assay (ELISA). Results were normalized to β-Actin (control) immunoreactivity measured in parallel reactions. All four proteins exhibited H 2 O 2 dose-dependent shifts in immunoreactivity, generally at concentrations between 9 and 27 μM (Figs. 1A-D). Area under the curve calculations were used to compare the magnitude of H 2 O 2induced increases in protein expression by one-way ANOVA with the post-hoc Tukey test. Those analyses demonstrated significantly greater H 2 O 2 stimulated levels of HIF-1α and AAH relative to HIF-1β and FIH immunoreactivity (p < 0.0001; Fig. 1E).
Oxidative stress promotes cell motility. We next examined the effects of oxidative stress on PNET2 cell directional motility using a more limited range of H 2 O 2 treatment, i.e., 13.2 or 22 μM. The 13.2 μM of H 2 O 2 dose was below the level associated with recent findings demonstrating that hypoxia and HIF-1α can potentiate Notch signaling. 30,31

Discussion
This study was designed to investigate the role of cross-talk between AAH and HIF-1α as a means of regulating cell motility. First, we demonstrated that AAH and HIF-1α expression and directional motility in PNET2 CNS-derived neuronal cells were stimulated by mild oxidative stress induced by low dose H 2 O 2 treatments. Then, we showed that si-RNA inhibition of either AAH or HIF-1α significantly impaired directional motility, particularly with regard to adherent cells. These results indicate that oxidative stress regulates expression of both AAH and HIF-1α, and that both molecules play key roles in regulating cell motility. Since both AAH and HIF-1α are also regulated by insulin/ IGF stimulation, 4,18,32 dual signaling pathways and mechanisms modulate cell motility. It would seem that while insulin/IGF regulatory mechanisms are important for effectuating long-term changes in cellular responses, including at the level of transcription, 33 the role of redox regulation and signaling may be to modulate short-term responses to environmental cues such as those produced by acute injury. Whether AAH and/or Notch have protective roles in the context of oxidative stress, as previously demonstrated for DJ-1, which is induced in astrocytes in response to ischemic injury, 34 or H 2 O 2 -removing enzymes, such as catalase, which is induced in hypercoagulable states that cause ischemic injury, 35 cannot be determined from the data at hand. However, since si-RNA inhibition of AAH and HIF-1α impaired motility and not cell viability, it is unlikely that either molecule mediates anti-stress responses at the low levels of oxidative stress produced in our experiments.
Given the overlapping mechanisms of gene regulation, their roles in cell motility, and the fact that AAH is a hydroxylase enzyme while HIF-1α is regulated by FIH, which is also a hydroxylase enzyme, it was of interest to explore potential functional connections between AAH and HIF-1α. The main approach used was to inhibit gene expression through transient transfection of PNET2 cells with siRNA duplexes and examine the effects on AAH, HIF-1α and related signaling molecules by qRT-PCR analysis and ELISA. The findings that both si-AAH and si-HIF-1α inhibited AAH mRNA, while si-HIF-1α inhibited HIF-1α and si-AAH did not, place HIF-1α upstream of AAH in terms of gene regulation and functionally connect these genes at the level of transcription. At the protein level however, the effects were mixed in that si-RNA suppression of AAH caused parallel shifts in expression of Humbug, a truncated AAH-related protein, 7,8 but had either no effect, or it significantly increased HIF-1α and/or HIF-1β expression. Transfection with si-HIF-1α did not suppress AAH or Humbug protein expression. While the explanation for this discrepancy is not clear, conceivably other interconnecting pathways may permit AAH protein stabilization under normoxic conditions, including trophic factor stimulation (insulin/IGF in medium). In this regard, it is noteworthy that trophic factors stimulate Akt and inhibit glycogen synthase kinase 3β (GSK-3β). Since Figure 2. Inhibition of hIF-1α or aah impairs directional motility. pNeT2 cells were transfected with siRNa targeting no specific sequences (siScr), hIF-1α (sihIF-1α) or aah (siaah) using the amaxa electroporation system (see Methods). 24 hours later, directional motility was measured using the aLMI assay, which enables one to quantify the percentages of non-motile, motile-adherent and motile non-adherent cells. The total percentage of motile cells was calculated from the sum percentages of motile adherent plus motile non-adherent cells. The graphs depict the mean ± S.e.M percentages of (a) total motile, (B) motile adherent and (C) motile non-adherent cells after 30 min incubation in blind-well Boyen chambers. 2% fetal bovine serum was supplied in the lower chamber as a trophic factor. Inter-group statistical comparisons were made using one-way aNOVa with the post hoc Tukey-Kramer significance test. Significant p-values are indicated over the bars.
Previous studies demonstrated that AAH mediates its effects on cell motility by interacting with and hydroxylating Notch and Jagged, 8 and that a downstream target of Notch signaling is HES-1. 38 Since Notch-1 stimulates HES-1 transcription, 39 the reductions in HES-1 mRNA associated with si-RNA inhibition of AAH and HIF-1α suggest that Notch transcriptional activity is regulated by both AAH and HIF-1α. As demonstrated herein and in previous reports, overexpression of AAH increases Notch-1 protein levels. 11 In addition, AAH overexpression stimulates Notch's translocation to the nucleus where it regulates gene expression. 11 Once in the nucleus, Notch-1 serves as a transcription factor for other genes involved in motility. However, since si-AAH had no significant effect on Notch's mRNA levels, AAH's regulation of Notch is most likely GSK-3β phosphorylates and destabilizes both AAH 3,36 and HIF-1α, 37 si-RNA inhibition of HIF-1α may not necessarily inhibit AAH protein in the context of trophic factor inhibition of GSK-3β.
Under normoxic conditions, FIH hydroxylates HIF-1α, signaling it to undergo degradation. Under hypoxic conditions, FIH's hydroxylase is inactivated, permitting HIF-1α to enter the nucleus where it serves as a transcription factor. Our results demonstrate that, in addition to hypoxia, mild oxidative stress (induced with H 2 O 2 ) preserves HIF-1α protein. Thus, we propose that mild oxidative stress stimulates cell motility and regulates AAH protein expression by inhibiting FIH hydroxylation of HIF-1α, allowing HIF-1α to enter the nucleus and serve as a transcription factor for AAH. to interact with the intracellular domain of Notch-1, 25 or regulate AAH mRNA expression. The reduced expression of FIH mRNA effectuated by si-AAH or si-HIF-1α transfection could represent a feedback mechanism to negatively regulate HIF-1α and eventually AAH expression, thereby halting cell motility.
We summarize our proposed scheme for extrinsic regulation and cross-talk between the AAH-Notch-Jagged-HES and FIH-HIF-1α hydroxylase signaling pathways in Figure 5. In brief, insulin and IGF regulate AAH and HIF-1α protein expression and function through posttranslational mechanisms including phosphorylation and attendant inhibition of GSK-3β activity. In addition, insulin and IGF stimulate AAH and HIF-1α gene expression, increasing their mRNA levels. Oxidative stress and hypoxia activate HIF-1α signaling by inhibiting FIH. This results in HIF-1α-mediated increases in AAH mRNA. Attendant increases in AAH protein expression lead to increased interactions between AAH and Notch/Jagged. AAH and HIF-1α both increase Notch signaling and cell motility. AAH functions by interacting with and hydroxylating Notch and Jagged. The cleaved N-terminal fragment of Notch translocates to the nucleus where it functions as a transcription factor and regulates target genes such as HES-1. HIF-1α potentiates Notch signaling via the mastermind-like protein 1 (MAML1) co-activator, with attendant stimulation of Notch target genes including HES-1 and HEY-1. 30 Increased Notch signaling through HES enhances expression of hypoxia responsive elements and hypoxia-inducible genes. However, attendant increased expression of FIH could serve as a negative feed-back mechanism for HIF-1α-AAH-Notch signaling. Notch activated signaling increases cell motility in part by altering expression of cell adhesion molecules. 30,40,41 We hypothesize that AAH and HIF-1α cross-talk within a hydroxylation-regulated signaling pathway that is transiently driven by fluctuations in oxidative stress, while more sustained stimulation of motility is mediated by signaling through insulin/IGF cascades. Therefore, therapeutic measures to prevent or limit invasion and metastatic spread of neuroblastic tumor cells will likely require inhibition of both redox-and mediated by post-translational mechanisms. Downstream Notch-regulated target genes that mediate cell motility include E-Cadherin, 30 tenascin C, 40 and other genes that regulate cell adhesion. 41 Jagged is a ligand for Notch, and its binding to Notch is needed for Notch cleavage and its release from the membrane for translocation to the nucleus. 42, 43 The finding that both si-AAH and si-HIF decreased Jagged-1 expression suggests an additional mechanism by which AAH and HIF-1α regulate Notch signaling. The si-HIF-1α inhibition of HES-1 mRNA could be explained by HIF-1α's ability Enzyme linked immunosorbent assay (ELISA). Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer containing protease and phosphatase inhibitors. 49,50 Protein concentrations were determined using the bicinchoninic acid (BCA) assay (Pierce, Rockford, IL). We performed direct binding ELISAs to measure AAH, HIF-1α, HIF-1β, FIH, Notch-1, Jagged-1 and β-Actin immunoreactivity. Samples containing 50 ng protein diluted in Tris-buffered saline, pH growth factor-mediated mechanisms. Prevention of neuroblastic tumor metastasis will likely require inhibition of growth factor and redox-mediated mechanisms.

Methods
Cell culture. Human CNS-derived Primitive Neuroectodermal Tumor 2 (PNET2) cells 44 were maintained as previously described. 3,32 We examined the effects of oxidative stress on HIF-1α, AAH and FIH expression by treating 96-well microcultures with 0-45 μM H 2 O 2 for 20 h and measuring immunoreactivity by a cellular enzyme-linked immunosorbent assay (ELISA). 45 The protocol for the H 2 O 2 exposures was based on previous studies demonstrating that CNS neuronal cells exhibit oxidative injury, but remain viable with altered gene expression, 24 to 48 hours after treatment with up to 45 μM H 2 O 2 . 46,47 Applying a more limited dose range of H 2 O 2 (0, 13.2 and 22 μM), we examined the effects of mild oxidative stress on directional motility. We assessed cross-talk among the AAH, FIH, HIF-1α and Notch signaling pathways in cells transfected with commercially prepared (Dharmacon, Inc., Chicago, IL) small interfering RNA duplexes (si-RNA) that targeted the AAH (ASPH NM_004318), HIF-1α (NM_001530, NM_181054) or Notch-1 (NM_0176617) genes. Finally, we examined the consequences of AAH overexpression in cells transfected with recombinant plasmid carrying the full-length AAH cDNA (pAAH) in which gene expression was under the control of a CMV promoter. 1 Control cells were transfected with recombinant plasmid carrying the green fluorescent protein gene (pGFP). Cells were transfected in suspension using the Amaxa "v" nucleofector cell line reagents and the Amaxa nucleofector apparatus (Amaxa, Inc., Gaithersburg, MD) according to the manufacturer's protocol. With this approach, we consistently achieved 75-90% transfection efficiencies as determined by GFP labeling of co-transfected cells. 16-24 hours later, cells were used to examine protein and mRNA expression.
Quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) analysis. We used qRT-PCR to measure mRNA expression. 11,48,49 In brief, cells were lysed in Qiazol reagent (Qiagen Inc., Valencia, CA) and total RNA was isolated using the EZ1 RNA universal tissue kit and the BIO Robot EZ1 (Qiagen, Inc.). RNA was reverse transcribed using random oligodeoxynucleotide primers and the AMV First Strand cDNA synthesis kit (Roche Diagnostics Corporation, Indianapolis, IN). The resulting cDNA templates were used in qPCR amplification reactions with gene specific primer pairs ( Table 1). 48 Primers were designed using MacVector 10 software (MacVector, Inc., Cary, NC) and their target specificity was verified using NCBI-BLAST (Basic Local Alignment Search Tool). The amplified signals from triplicate reactions were detected and analyzed using the Mastercycler ep realplex instrument and software (Eppendorf AG, Hamburg, Germany). Relative mRNA abundance was calculated from the ng ratios of specific mRNA to 18S rRNA measured in the same samples. Inter-group statistical comparisons were made using the calculated mRNA/18S ratios. proposed scheme for extrinsic regulation and cross-talk between the aah-Notch-Jagged-heS and FIh-hIF-1α hydroxylase signaling pathways. Large shaded oval toward the bottom of the diagram represents the nucleus, large rectangle represents the cytoplasm and outside the rectangle, extrinsic stimuli and cellular responses are indicated. Small shaded ovals represent proteins, black shaded rectangles represent genes and dash-lined rectangles depict post-transcriptional regulators of aah and hIF-1α signaling. pluses correspond to positive stimulatory effects and minus signs represent inhibitory effects. Insulin and IGF regulate aah and hIF-1α through inhibition of GSK-3β and by increasing their mRNa levels. Oxidative stress and hypoxia activate hIF-1α via inhibition of FIh, resulting in increased aah gene expression. Both aah and hIF-1α increase Notch signaling. aah interacts with and hydroxylates Notch and Jagged, resulting in nuclear translocation of Notch and increased expression of Notch-regulated genes, e.g., heS-1. hIF-1α potentiates Notch signaling and stimulation of Notch target genes. 30 Consequences include increased expression of hypoxia responsive elements and hypoxia-inducible genes that promote cell motility and alter cell adhesion.

Microtiter immunocytochemical ELISA (MICE) assay.
The MICE assay is a cellular ELISA that was used to measure the effects of oxidative stress on AAH, HIF-1α, FIH and β-Actin immunoreactivity directly in fixed cultured cells (96-well plates). 45 The main modification of the original protocol was that immunoreactivity was measured with the Amplex Red fluorophore (Ex 579/Em 595) (Molecular Probes, Eugene, OR) instead of a colorimetric reagent. Cell density was assessed by subsequently staining the cells with Hoechst H33342 (Molecular Probes, Eugene, OR) and measuring fluorescence (Ex360 nm/Em460 nm) in a Spectramax M5 microplate reader (Molecular Dynamics, Inc., Sunnyvale, CA). The calculated ratios of fluorescence immunoreactivity to H33342 were used for inter-group comparisons. At least eight replicate cultures were analyzed in each experiment.
Directional motility assay. Directional motility was measured using the ATP Luminescence-Based Motility-Invasion (ALMI) assay. 28 Briefly, culture medium containing 2% FCS was placed 7.4 (TBS) were adsorbed to the bottom flat surfaces of 96-well polystyrene plates (Nunc, Rochester, NY) overnight at 4°C. 36 Non-specific binding sites were blocked by a 3-hour room temperature incubation with 300 μl/well of TBS + 0.05% Tween 20 + 3% BSA. Samples were then incubated with 0.1-0.5 μg/ml primary antibody for 1 h at 37°C. Immunoreactivity was detected with horseradish peroxidase (HRP)-conjugated secondary antibody and Amplex Red soluble fluorophore (Molecular Probes, Eugene, OR). 36,50 Fluorescence was measured (Ex 530/Em 590) in a SpectraMax M5 microplate reader (Molecular Devices Corp., Sunnyvale, CA). Parallel negative control assays had primary, secondary or both antibodies omitted. Between steps, reactions were rinsed 3 times with TBS + 0.05% Tween 20 using a Nunc ELISA plate washer.  in the lower chambers (Neuro Probe, Gaithersburg, MD) and 8-micrometer pore diameter polycarbonate filters divided the upper and lower chambers. 100,000 viable (Trypan Blue exclusion) PNET2 cells were seeded into the upper chambers and cell migration was allowed to proceed for 30 minutes at 37°C in a CO 2 incubator. Cells collected from the upper chambers (non-motile), under surfaces of the filters (motile adherent) and bottoms of the wells (motile non-adherent) were quantified using ATPLite reagent (Perkin-Elmer, Waltham, MA) because ATP luminescence is linearly correlated with cell density. 28 Luminescence was measured in a TopCount Machine (Perkin-Elmer). The percentages of non-motile, motile adherent, motile non-adherent cells in 8 replicate assays were calculated and used for statistical analysis. Because this assay separately quantifies motile-adherent and motile-non-adherent sub-populations, it provides information about cell motility and adhesion.