Figure 1.
Hypoxia reduces S100A1 gene expression in Endothelial cells via its -3′UTR.
A) Primary human microvascular endothelial cells (HMVECs) were subjected to chemical hypoxia by treatment with CoCl2 (250 µmol/L) for 24 h before extract preparation and immunoblot analysis. Expression levels of S100A1 were quantitated using a LICOR Odyssey near infrared scanner and normalized to those of β-actin. The experiment was done 4 times, each in duplicate. *, P<0.05 vs normoxic. B) EA.hy926 ECs were transfected with the S100A1–3′UTR luciferase reporter and subjected to chemical (CoCl2 or Desferroxamine (Des) or gas hypoxia (1% O2), for 24 h. Luciferase expression was measured as described in the Methods section and presented as relative light units (RLU) normalized to untreated cells. *, P<0.02 vs untreated normoxic. The experiment was performed 4 times, each in triplicate. Expression of a luciferase reporter linked to a control 3′UTR (from SV40 T antigen) was not changed (not shown). C) Expression of MiR-138 was measured by qPCR in EA.hy926 ECs or primary HMVECs treated with 1% O2 or CoCl2 (250 µmol/L) or Desferroxamine (Des, 100 µmol/L) for 24 h. Levels of the small nuclear RNA U6 were assessed in parallel and used to normalize expression. The experiment was done 4 times, each in triplicate. *, P<0.01 vs untreated. U6 expression did not change during hypoxia.
Figure 2.
MiR-138 decreases S100A1 gene expression in Endothelial Cells.
A) EA.hy926 ECs were co-transfected with the S100A1–3′UTR luciferase reporter gene and either a MiR-138 mimic or scrambled control mimic. Luciferase activity in cell lysates was measured 24 h later and is reported as relative light units (RLU) normalized to untreated cells. *, P<0.01 vs ctr. mimic. The experiment was done 3 times, each in triplicate. Expression of a luciferase reporter linked to the control 3′UTR was not changed by the mimic treatment (not shown). B) Primary HMVEC were transfected with the MiR-138 mimic for 1 or 3 days or control mimic for 3 days. Cell extracts were immunoblotted for S100A1 or β-actin (as loading control). Transfection with the control mimic did not change S100A1 levels compared to untransfected HMVECs (not shown). A representative immunoblot is shown. The experiment was performed 3 times. Expression of S100A1 was normalized to β-actin. *, P<0.05 vs ctr. Mimic. C) EA.hy926 ECs were transfected with either the wild-type (WT) S100A1–3′UTR luciferase reporter gene or a S100A1–3′UTR with deletion of the 22 nucleotide putative MiR-138 target site (ΔMiR138). 24 h later cells were subjected to chemical hypoxia (250 µmol/L CoCl2). Luciferase activity in cell lysates was reported as relative light units (RLU) normalized to WT untreated cells. *, P<0.01 vs untreated. The experiment was done 3 times, each in triplicate.
Figure 3.
Hypoxia does not increase MiR-138 in skeletal muscle cells.
Murine C2C12 skeletal myoblasts (left panels) or differentiated myotubes (right panels) were subjected to hypoxia for 24 h. Protein extracts were immunoblotted for S100A1 protein expression (upper panels) and expression of MiR-138 by qPCR (lower panels). Expression of Myosin Heavy Chain (MHC) was used to verify differentiation. Experiment was performed 3 times, each in duplicate. *, P<0.05 vs normoxic.
Figure 4.
Primary human vascular smooth muscle cells do not change S100A1 nor MiR-138 levels during hypoxia.
A) Primary human microvascular smooth muscle cells (HVSMCs, obtained from the ATCC) were subjected to hypoxia for 24 h. Protein extracts were immunoblotted for Hif1-α (to verify induction of hypoxia), β-actin (to verify equal loading), and S100A1. B) Expression of MiR-138 was by qPCR. Experiment was performed 3 times, each in duplicate.
Figure 5.
Specific inhibition of MiR-138 prevents the hypoxia-induced loss of S100A1 in ECs.
A) Primary HMVECs were transfected with the antagomir-138 (or control antagomir) and subjected to chemical hypoxia (CoCl2, 250 µmol/L) for 24 h before extract preparation and immunoblot analysis. S100A1 expression was normalized to that of β-actin. The experiment was done 3 times, each in duplicate. *, P<0.05 vs control antagomir. B) EA.hy926 ECs were transfected with an antimir-138 (or control) and subjected to either gas hypoxia (1% O2) or chemical hypoxia (CoCl2, 250 µmol/L) for 24 h. Luciferase activity in cell lysates was reported as relative light units (RLU) normalized to untreated cells. The experiment was done 3 times, each in triplicate. *, P<0.02 vs antimir-138 treated.
Figure 6.
MiR-138 levels are increased in ischemic muscle tissue.
A) Gastrocnemius muscle biopsy specimens from patients with CLI and non-ischemic control [3] were analyzed for expression levels of MiR-138 and the housekeeping small nucleolar RNAs snoRD44 and snoRD47 by qPCR. Expression levels are presented as fold CLI/normal. n = 4; *, P<0.05 vs snoRD44 or 47, whose expression levels in CLI samples were not significantly different from normal. B) Gastrocnemius muscle biopsy specimens from mice post femoral artery resection (FAR) and non-ischemic contralateral control were analyzed for expression levels of MiR-138 and the U6 small nuclear housekeeping RNA by qPCR at times indicated. Expression levels are presented as fold FAR/normal. n = 4; * P<0.05 vs U6, whose expression levels in FAR samples were not significantly different from normal.
Figure 7.
Hif-1α mediates the reduction of S100A1-3′UTR reporter gene expression.
A) EA.hy926 (upper panel) or primary HMVEC (lower panel) cells were exposed to the prolyl-hydroxylase-2 inhibitor IOX2 (10 µmol/L) for 24 h to induce Hif1-α, prior to extract preparation. A representative immunoblot is shown to verify Hif1-α induction. The experiment was done 3 times with similar results. β-actin was used to control for protein loading. B) Expression levels of MiR-138 were assessed by qPCR in extracts prepared from EA.hy926 ECs subjected to 24 h treatment with 10 µmol/L IOX2. C) EA.hy926 ECs were co-transfected with the S100A1–3′UTR luciferase reporter gene and either the antimir-138 or scramble control (Dharmacon). 24 h later cells were incubated with IOX2 to induce Hif1-α stabilization. *, P<0.02 vs untreated. For both B, C, the experiment was done 3 times, each in triplicate. D) EA.hy926 ECs were transfected with the S100A1-3′UTR reporter gene and co-transfected with siRNA against Hif1-α, or control scramble siRNA. Cells were then subjected to chemical hypoxia with 250 µmol/L CoCl2 for 24 h before luciferase activity was assessed. *, P<0.02 vs normoxic, P<0.05 vs siRNA Hif1-α. Experiment was performed 3 times, each in triplicate.
Figure 8.
MiR-138 compromises EC Matrigel-induced capillary formation by inhibiting S100A1.
A) Primary HMVEC were transfected with the MiR-138 Mimic or scramble control Mimic. 48 h later cells were infected (MOI = 17) with either control Adenovirus or Adenovirus expressing S100A1. 24 h later cells were seeded onto Matrigel matrix. Images of EC tube formation were taken 24 h later and digitized using Image J (Original pictures of EC tube formation are included as Figure S3). B) Cell extracts of HMVEC treated in parallel to those in (A) were immunoblotted for S100A1, total or p-Thr 495 eNOS, or β-actin (as loading control). Representative images are shown. The experiment was done 3 times, each in duplicate. Expression levels of S100A1 and pT-495 eNOS were normalized to β-actin and total eNOS, respectively. The experiment was done 3 times, each in duplicate. *, P<0.05 vs untreated.
Figure 9.
MiR-138 compromises VEGF-stimulated NO production by inhibiting S100A1.
Primary HMVEC cells treated in parallel to those in Figure 8 were starved for 24 h in medium supplemented with 0.2% FBS before being treated with 50 ng/ml VEGF. Supernatants were collected 24 h later and analyzed for nitrate/nitrite levels. The experiment was done 3 times, in duplicate. *, P<0.01 vs no Mimic, #, P<0.01 vs control Adenovirus. Re-expression of S100A1 reverses the MiR-138 induced EC dysfunction.
Figure 10.
Proposed Scheme of S100A1 regulation by MiR-138.
(1) Under normal oxygen tension, the transcription factor Hif1-α is continuously hydroxylated by the action of cellular prolyl-hydroxylases (PHDs) in a reaction that requires Fe2+ and O2 as co-factors. (2) Hydroxylation of Hif1-α promotes binding of the Von Hippel-Lindau (VHL)-E3 ubiquitin ligase complex, promoting poly-ubiquitination and (3) degradation via the 26S proteasome complex. The action of PHDs are inhibited directly by IOX2 and Cobalt and indirectly by iron chelators, such as Desferroxamine (Des) as well as low oxygen levels. (4) Under low oxygen tension the Hif1-α protein becomes stabilized in the nucleus and promotes transcription of the pro-angiogenic vascular endothelial growth factor (VEGF) gene. VEGF promotes activation of eNOS by signaling through VEGFR2, promoting phosphorylation of the stimulatory Ser-1177 site. Increased eNOS activity raises nitric oxide (NO) production, which inhibits PHDs, further promoting Hif1-α stabilization in a positive feed-back loop. (5) To maintain cellular homeostasis, stabilization of Hif1-α also promotes increased production of MiR-138, (6) which binds to the 3′UTR of the S100A1 mRNA, leading to drastically reduced S100A1 levels and reduction of eNOS activity by promoting phosphorylation of the inhibitory Thr-495 site (7), in a counterbalancing negative feed-back loop. Endothelial dysfunction develops when these carefully balanced multiple feedback loops become dysregulated allowing for prolonged MiR-138 expression with consequent loss of S100A1 and reduced eNOS activity.