HCAECs, obtained from PromoCell GmbH (Heidelberg, Germany), were cultured in the endothelial cell growth medium containing 10% of fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin. The cells were grown in a humidified atmosphere containing 5% of CO₂ at 37°C.

For hypoxic stimulation, a humidified temperature-controlled incubator (ProOX model 110; BioSpherix, Ltd., Redfield, NY, USA) was used. HCAECs were cultured under various hypoxic conditions: (i) 2.5% O₂, 5% CO₂, and 92.5% N₂; (ii) 5% O₂, 5% CO₂, and 90% N₂; and (iii) 10% O₂, 5% CO₂, and 85% N₂ for 1, 2, 4, or 6 h or longer. A lower oxygen concentration (1%) was attempted; however, the cultured HCAECs had difficulty surviving such severe hypoxia. As a control, HCAECs were cultured at 5% CO₂ in air.

This analysis was performed as previously described [21]. Monoclonal rat anti-human chemerin and anti-ERK antibodies (Santa Cruz Biotechnology, Paso Robles, CA) were used. Equal protein loading of the samples was verified by incubating with a monoclonal anti-α-tubulin antibody (Sigma, St Louis, MO). Proteins were incubated with specific antibodies (1:200 dilution) for 1 h at room temperature followed by incubation with a 1:10000 dilution of horseradish peroxidase-conjugated polyclonal anti-goat IgG antibody for 1 h at room temperature. Signals were visualized by chemiluminescence detection. Densitometry was used for quantification of all western blot data.

Total RNA was isolated from HCAECs using the single-step acid guanidinium thiocyanate/phenol/chloroform extraction method according to the protocol provided by the manufacturer. Real-time PCR was performed as previously described [21]. The chemerin primers were 5’-GGAAGAAACCCGAGTGCAAA-3’ and 5’-ACCAACCGGCCCAGAACT-3’.

HCAECs were transfected with 800 ng of ERK, short interfering RNA (siRNA), chemerin siRNA or HIF-1alpha, which target a specific 19–21 nt sequence to knock down gene expression (Dharmacon Inc., Lafayette, CO). The ERK sense and antisense siRNA sequences were 5’-GACCGGAUGUUAACCUUUA-3’ and 5’-UAAAGGUUAACAUCCGGUC-3’, respectively. Chemerin sense and antisense siRNA sequences were 5’-GCAUCAAACUGGGCUCUGA-3’ and 5’-UCAGAGCCCAGUUUGAUGC-3’, respectively. HIF-1alpha sense of siRNA sequence and antisense of siRNA sequence 5'-UGAGAGAAAUGCUUACACA-3' and 5'-UGUGUAAGCAUUUCUCUCA-3', respectively. (Santa Cruz Biotechnology)

As a negative control, scramble siRNA and green fluorescent protein siRNAs were used (Dharmacon Inc., Lafayette, CO). HCAECs were transfected with the siRNA oligonucleotides using the Effectene Transfection Reagent Kit (Qiagen Inc., Valencia, CA, USA). After incubation at 37°C for 24 h, HCAECs were subjected to hypoxia and then western blot analysis.

The EMSA was performed as previously described [22]. Briefly, a Bio-Rad protein assay kit was used to determined nuclear protein concentrations in HCAECs. Consensus and control oligonucleotides (Santa Cruz Biotechnology) were labeled by polynucleotide kinase to incorporate [γ-³²P]ATP. The consensus oligonucleotide sequence for transcription factor SP1 was consensus 5’-ATTCGATCGGGGCGGGGCGAGC-3’. The SP1 mutant oligonucleotide sequence was 5’- ATTCGATCGGTTCGGGGCGAGC-3’. Controls were set up in each group with mutant oligonucleotides or cold oligonucleotides to compete with the radiolabeled oligos.

The -1400 to -900 region of the rat chemerin promoter was used to generate a plasmid construct. Briefly, chemerin genomic DNA was amplified with the forward primer, 5’-AGCCTCGGAGAAAGCCTGGCT-3’ and reverse primer, 5’-TTGAGGATTAAATGAAGAAACATGAA-3’. The amplified genomic product was digested with M1uI and Bg1II restriction enzymes. The DNA fragment was then ligated into the pGL3-basic luciferase plasmid vector (Promega) digested with the same enzymes. The chemerin promoter contains a conserved SP1-binding site between positions ‒1137 and ‒1136. For the mutant construct, the SP1-binding site was mutated using a mutagenesis kit (Mission Biotech, Taipei, Taiwan). The mutations were confirmed by DNA sequencing. Chemerin-encoding plasmids were transfected into HCAECs using a low pressure-accelerated gene gun (Bioware Technologies, Taipei, Taiwan). The experimental plasmid at 2 μg and control plasmid (pGL4-Renilla luciferase) at 0.02 μg were cotransfected using the gene gun for each, and then the culture medium was replaced with the normal culture medium. After exposure of the cells to hypoxia for 1 h, HCAEC extracts were prepared using the dual luciferase reporter assay system (Promega) and analyzed for dual luciferase activity on a luminometer (Turner Designs).

The proliferation of HCAECs was quantified using [³H]thymidine incorporation. The cells were seeded in ViewPlates (Packard Instrument, Meriden, CT) at a density of 5 × 10³/well in a serum-free medium. Thymidine uptake was analyzed by addition of 500 nCi/mL [³H]thymidine (Perkin Elmer, Boston, MA) with incubation for 2–6 h with or without hypoxia. The cells were washed twice with PBS. Nonspecific uptake was studied in the presence of 10 μM cytochalasin B and subtracted from the determined values. MicroScint-20 (50 μL) was added, and the plate was read on a TopCount (Packard Instrument).

The migration activity of HCAECs was determined using the growth factor-reduced Matrigel invasion system (Becton Dickinson) as previously described.[19] HCAECs (5 × 10⁴ cells) were seeded on top of the ECMatrix gel (Chemicon International, Inc., Temecula, CA) and incubated at 37°C for 4 h during hypoxia with addition of chemerin, TNF-alpha, PD98059, an anti-TNF-alpha antibody, chemerin siRNA, or scrambled siRNA. Three different phase-contrast microscopic high-power fields per well were photographed. The migratory HCAECs were stained with Liu stain (BaSO, New Taipei City, Taiwan) and counted. The observer was blind to the experiment.

Capillary-like network formation was tested on cultured HCAECs as previously described [19]. A 24-well culture plate was coated with Matrigel (250 μL; BD Biosciences, MA), which was allowed to solidify (37°C, 1 h). HCAECs were cultured in the plate at hypoxic condition (2.5% O₂) for 4 h at 37°C. Using a phase contrast microscope, the capillary-like network formation was examined, and its branching points were quantified.

All results were expressed as mean ± SD. Statistical significance was evaluated using the ANOVA test followed by Tukey–Kramer multiple-comparison test (GraphPad Software Inc., San Diego, CA, USA). Differences with P < 0.05 were considered statistically significant.

Different degrees of hypoxia (2.5%, 5%, and 10% O₂) were used to test the expression of chemerin in cultured HCAECs (Fig 1A and 1B). Chemerin protein expression was notably increased when HCAECs were subjected to 2.5% O₂ hypoxia for 4 h as compared to other conditions. Thus, these culture conditions were used in subsequent analyses. At 2.5% O₂ hypoxia, the chemerin level increased gradually and reached a peak after 4 h (Fig 1C and 1D). Chemerin mRNA expression also reached its maximal level after 2 h and decreased gradually (Fig 1E). The pH levels showed no significant differences among the different levels of hypoxia.

To identify the pathways involved in hypoxia-induced chemerin expression, various inhibitors, including inhibitors of p38 MAPK (SB203580), JNK (SP600125), and ERK kinase (PD98059) were used. As shown in Fig 2A and 2B, western blot analysis revealed that the hypoxia-induced chemerin expression was significantly inhibited by the ERK kinase inhibitor (PD98059). This effect was confirmed using an ERK siRNA. The JNK (SP600125) and p38 MAP kinase (SB203580) inhibitors did not attenuate the chemerin protein expression induced by hypoxia (Fig 2A and 2B). These findings indicated that ERK pathways but not p38 MAPK and JNK pathways were responsible for the induction of chemerin expression in hypoxic HCAECs.

TNF-alpha expression was induced by hypoxia at 2.5% O₂ after 1 h and then decreased gradually (S1 Fig). In addition, pretreatment with an anti-TNF-alpha antibody significantly attenuated the expression of chemerin protein induced by hypoxia, and addition of TNF-alpha increased the expression of chemerin in HCAECs during normoxia (Fig 2A and 2B). These data indicate that TNF-alpha mediates the expression of chemerin by hypoxia.

We also tested the role of HIF-1alpha in HCAECs under hypoxia condition. HIF-1alpha expression was induced by hypoxia at 2.5% after 1 h and then decreased gradually (S2A and S2B Fig). However, addition of HIF-1alpha siRNA did not affect the expression of chemerin protein induced by hypoxia (S2C and S2D Fig).

Because the ERK inhibitor reduced the chemerin protein expression most significantly, we focused on the ERK pathway to study the chemerin protein expression in hypoxic HCAECs. Hypoxia at 2.5% O₂ for 1 to 6 h significantly increased the phosphorylation of ERK (Fig 3A and 3B). The ERK kinase inhibitor (PD98059) and ERK siRNA effectively blocked phosphorylation of the ERK protein induced by hypoxia. The scrambled siRNA did not affect the phosphorylation of ERK induced by hypoxia.

Under the conditions of 2.5% O₂ hypoxia, EMSA showed increased DNA-protein binding activity of SP1 in chemerin promoter (Fig 4A). Addition of PD98059 and ERK siRNA before hypoxia abolished this binding activity. Exogenous addition of TNF-alpha significantly increased the DNA-protein binding activity. This finding suggested that hypoxia in HCAECs may increase the binding of the chemerin promoter with the SP1 transcription factor. Hypoxia-induced ERK phosphorylation may increase chemerin promoter–SP1 binding and chemerin expression.

To study whether the chemerin expression induced by hypoxia is regulated at the transcriptional level, we cloned the promoter region of chemerin and contrasted a luciferase reporter plasmid (pGL3-Luc). The chemerin promoter construct contained SP1-, Stat3-, AP-1, NF-kB, and HIF-1alpha-binding sites. The mutant chemerin promoter has a mutation of SP1-binding site (Fig 4B). The luciferase reporter assay showed that hypoxia at 1 h increased transcriptional activity of SP1 on the chemerin promoter in HCAECs (Fig 4C). Mutation of SP1 binding site in the chemerin promoter inhibited this effect of hypoxia. Addition of PD98059 and the anti-TNF-alpha antibody suppressed the transcriptional activity of the chemerin promoter in hypoxic HCAECs. These results suggest that the SP1 binding site in the chemerin promoter is essential for the transcriptional regulation induced by hypoxia and that hypoxia regulates chemerin promoter via TNF-alpha and ERK pathways.

To test the effects of hypoxia on the function of HCAECs, we examined HCAECs proliferation, tube formation, and migratory activity. In addition, chemerin siRNA (S3 Fig) was designed to test the role of chemerin in HCAECs. Proliferation of HCAECs was evaluated by measuring ³H-thymidine incorporation into the cells. Results revealed an increase in HCAEC proliferation after 2 and 4 h of hypoxia (Fig 5). Pretreatment with chemerin siRNA, PD98059, and the anti-TNF-alpha antibody attenuated the proliferation induced by hypoxia. Exogenous addition of chemerin and TNF-alpha during normoxia also increased proliferation of HCAECs.

Hypoxia at 2.5% O₂ significantly increased HCAEC migration (S4 Fig) and capillary-tube formation (by counting the branching point) (Fig 6) at 4 h as compared with the control group. Addition of PD98059, the anti-TNF-alpha antibody, or chemerin siRNA attenuated the enhancement of HCAEC migration and capillary tube formation induced by 4 h of hypoxia. The control siRNA did not inhibit the enhancement of migration and capillary tube formation induced by hypoxia. Addition of chemerin and TNF-alpha increased the HCAEC migration and capillary-tube formation during normoxia.