Antidicer RNAse activity of monocyte chemotactic protein-induced protein-1 is critical for inducing angiogenesis

Abstract

Inflammatory angiogenesis involves the induction of a novel gene ZC3H12A encoding monocyte chemoattractant protein-1 (MCP-1)-induced protein-1 (MCPIP1) that has deubiquitinase and antidicer RNAse activities. If and how these enzymatic activities of MCPIP1 mediate the biological functions of MCPIP1 are unknown. Present studies with human umbilical vein endothelial cells suggest that MCPIP-induced angiogenesis is mediated via hypoxia-inducible factor (HIF-1α), vascular endothelial growth factor (VEGF), and silent information regulator (SIRT-1) induction that results in the inhibition of angiogenesis inhibitor thrombospondin-1. MCPIP1 expression inhibited the production of the antiangiogenic microRNA (miR)-20b and -34a that repress the translation of HIF-1α and SIRT-1, respectively. The RNase-dead MCPIP mutant D141N not only did not induce angiogenesis but also failed to inhibit the production of miR-20b and -34a suggesting that the antidicer RNase activity of MCPIP1 is involved in MCPIP-mediated angiogenesis. Mimetics of miR-20b and -34a inhibited MCPIP1-induced angiogenesis confirming that MCPIP1 suppresses the biogenesis of miR-20b and -34a. Furthermore, our results indicate that MCPIP expression induces nuclear translocation of HIF-1α. We show that under hypoxia angiogenesis is mediated via induction of MCPIP1 and under normoxia, in vitro, MCPIP deubiquitinates ubiquitinated HIF-1α and the stabilized HIF-1α enters the nucleus to promote the transcription of its target genes, cyclooxygenase-2 and VEGF, suggesting that the deubiquitinase activity of MCPIP may also promote angiogenesis. The present results show for the first time that the antidicer RNase activity of MCPIP1 is critical in mediating a biological function of MCPIP, namely angiogenesis.

angiogenesis is a tightly controlled process involving proper balance between the levels of pro- and antiangiogenic factors (9). Dysregulation of these processes is involved in inflammatory diseases (34). A recent study suggests that inflammatory cytokines TNF-α, IL-1β, IL-8, and monocyte chemoattractant protein-1 (MCP-1) mediate angiogenesis via the induction of ZC3H12A gene encoding MCP-1-induced protein-1 (MCPIP1; Ref. 38), originally identified as a protein induced by MCP-1 treatment of human monocytes (63). MCPIP1 is the first member of a novel CCCH-type zinc finger protein family (26), and we refer to it as MCPIP in this article. MCPIP has been shown to mediate several biological functions such as angiogenesis (33, 38), adipogenesis (60), osteoclastogenesis (53), and hyperglycemia-induced death of cardiomyocytes (61). MCPIP was reported to have deubiquitinase activity (17, 25) and RNase activity (29, 32, 43). If and how the dual enzymatic activities of MCPIP are involved in mediating any of its biological functions remain unknown.

HIF-1α, which is known to be involved in angiogenesis, is a key transcription factor that is activated under hypoxic conditions (1, 51). It plays important roles in many biological processes such as embryonic development and in pathophysiological processes involving ischemia (64). HIF-1 is a crucial regulator that induces genes assisting in cellular processes such as oxygen transport, glucose metabolism, angiogenesis, and cell survival (16). HIF-1 is a heterodimeric protein complex consisting of hypoxia-inducible subunit HIF-1α and constitutively expressed HIF-1β subunit. Under normoxic conditions, HIF-1α is an unstable protein with a half-life of ∼5 min and is under stringent negative regulation by multiple mechanisms. HIF-1α is hydroxylated in an oxygen-dependent manner by prolyl hydroxylase domain (PHD) enzymes at proline residues in its oxygen-dependent degradation domain (15). Upon HIF-1α hydroxylation, von Hippel-Lindau protein, an E3 ligase, binds to it resulting in the ubiquitination of HIF-1α and its degradation by the ubiquitin-proteosome pathway (7, 24). Under hypoxic conditions, PHD can no longer hydroxylate HIF-1α resulting in its stabilization and consequent entry into the nucleus to form a complex with HIF-1β subunit. In the nucleus the dimer can bind to the hypoxia response element (HRE; RCGTG) on the promoters of its target genes. Cyclooxygenase-2 (COX-2) and VEGF are HIF-1α target genes. COX-2 is an inducible isoform of the COX family of enzymes that are involved in the production of biological mediators of inflammation, prostanoids generated from arachidonic acid (42, 44). Induction of COX-2 is influenced by proinflammatory stimuli and has been implicated in pathologies involving inflammatory angiogenesis, such as cancer (50, 54). VEGF is a well-established proangiogenic factor (12). MCPIP is known to cause elevation of HIF-1α levels during MCPIP-induced angiogenesis (33). The molecular mechanism by which MCPIP causes stabilization of HIF-1α is unknown. It is unknown whether MCPIP-induced inflammatory angiogenesis could be mediated via stabilization of HIF-1α by removal of the ubiquitin moieties linked to HIF-1α by the MCPIP deubiquitinase activity that was reported to negatively regulate NF-κB activation (25). Moreover, if MCPIP inhibition of NF-κB activation is involved in mediating angiogenesis has not been tested. Our study aims to decipher whether MCPIP deubiquitinates ubiquitinated HIF-1α and if NF-κB, a key proinflammatory transcription factor, is inhibited by MCPIP expression both with the potential to promote angiogenesis.

Silent information regulator (SIRT-1) enhances the angiogenic potential of endothelial cells by deacetylating forkhead box O (FoxO), a negative regulator of angiogenesis (4). It is a member of the sirtuins family of nicotinamide adenine dinucleotide (NAD⁺)-dependent histone deacetylases that regulate several biological processes including cell survival, metabolism, longevity, inflammation, and tumorigenesis (36, 58). SIRT-1 regulates cellular differentiation by deacetylating p53, a tumor suppressor resulting in the inhibition of p53 transcription. A target gene of p53 is thrombospondin (TSP)-1, an inhibitor of angiogenesis (5, 10, 21, 57). Whether MCPIP-mediated angiogenesis involves SIRT-1 or TSP-1 is unknown.

MicroRNA (miR)s play a vital role in regulating inflammation (13) and in modulating the levels of HIF-1 α and SIRT-1. miR-20b binds to the 3′-UTR of HIF-1α and thus inhibits translation of HIF-1α. Inhibition of miR-20b production increased the levels of HIF-1α, thus suggesting its antiangiogenic role (22). miR-34a is antiangiogenic and SIRT-1 is one of its targets (62). It was reported that miR-34a inhibits SIRT-1 translation by binding to the 3′-UTR of SIRT-1 mRNA (57). MCPIP can cleave the terminal loops of precursor miRs, and this antidicer activity can suppress miR biogenesis (47). Whether this antidicer RNase activity of MCPIP plays a role in any biological function of MCPIP such as inflammatory angiogenesis is not known.

Here we report that that the antidicer RNase activity of MCPIP inhibits the production of antiangiogenic miRs, miR-20b and miR-34a, that target HIF-1α and SIRT-1 respectively, thus promoting angiogenic differentiation of HUVECs. MCPIP is shown to deubiquitinate HIF-1α and stabilize HIF-1α leading to induction of VEGF and COX-2. Thus, deubiquitinase activity of MCPIP may also promote angiogenesis. Our studies with the MCPIP mutant D141N that has intact deubiquitinase activity but has lost RNase activity show that the antidicer RNase activity of MCPIP is essential for promoting angiogenesis. We for the first time demonstrate that the role of antidicer RNase activity is critical in induction of inflammatory angiogenesis by MCPIP.

MATERIALS AND METHODS

Cell culture conditions.

The human umbilical vein endothelial cells (HUVECs; CC-2519; Lonza) were cultured in endothelial cell basal medium (EBM; CC-3124; Lonza) according to manufacturer's protocol. HUVECs used were between passages 4–8. All cells were maintained at 37°C in 5% CO₂. Experiments under hypoxic conditions (6) were performed in the chamber with 1% O₂ at 37°C and 5% CO₂.

Plasmid construction.

The human wild-type MCPIP (accession no: AY920403) was subcloned into the pCMV-MAT-FLAG vector (Sigma-Aldrich). MAT-FLAG sequence and HISx8 tag were added at the 5′- and the 3′-end, respectively. The D141N mutation was produced using the QuickChange Lightning Site-Directed Mutagenesis Kit (Stratagene), according to the manufacturer's directions using pCMV-MAT-FLAG wild-type-MCPIP as template DNA and D141N mutagenic primers (sense: 5′-AGA-CCA-GTG-GTC-ATC-AAC-GGG-AGC-AAC-GTG-GCC-3′; and antisense: 5′-GGC-CAC-GTT-GCT-CCC-GTT-GAT-GAC-CAC-TGG-TCT-3′). The Qiagen Plasmid Maxi Kit was used to prepare the pCMV-MAT-FLAG MCPIP-WT and D141N expression vectors.

Transfection procedure.

HUVECs were transfected with vectors expressing MCPIP or empty vector using Lipofectamine and PLUS Reagents (11668; 11514, Life Technologies, NY). The transfection efficiency was 60–70% and was determined by the immunoblotting with antibody against FLAG (1:500; Sigma).

Treatment/transfection of HUVECS.

HUVECs were treated with following chemical inhibitors: p38 MAPK inhibitor, SB 203580 (20 μM), 3 h before transfection with expression construct for MCPIP or empty vector. HUVECs were transfected for 6 h with 100 nmol/l of a chemically synthesized small interfering (si)RNA targeted for the HIF-1α or SIRT-1 or COX-2 with 100 nmol/l nonspecific siRNA (Santa Cruz Biotechnology,) using Lipofectamine and PLUS Reagents (Life Technologies) according to the manufacturer's protocol before transfection with MCPIP-MAT or empty vector.

Real-time PCR.

Total RNA was isolated from HUVECs by using Trizol reagent (Invitrogen). cDNA was synthesized utilizing 1 μg of total RNA (DNase-treated) as previously described (33). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as internal controls for transcript analysis. Results presented are of three independent experiments, each measured in triplicate. The sequences of the primers used for real-time analysis are stated as follows: HIF-1α: forward-5′-CTTTTACCATGCCCCAGAT-3′; HIF-1α: reverse-5′-CATTGACCATATCACTATCCACA-3′; VEGF: forward-5′-CGAGGCAGCTTGAGTTAA-3′; VEGF: reverse-5′-GCGTGGTTTCTGTATCGATC-3′; SIRT-1: forward-5′-CCCTCAAAGTAAGACCAGTAGC-3′; SIRT-1: reverse-5′-CACAGTCTCCAAGAAGCTCTAC-3′; TSP-1: forward-5′-CTCCCCTATGCTATCACAACG-3′; TSP-1: reverse-5′-AGGAACTGTGGCATTGGAG-3′; COX-2: forward-5′-CTATGGCTACAAAAGCTGGG-3′; and COX-2: reverse-5′-CCACAATCTCATTTGAATCAGG-3′.

miR analysis.

miRs were isolated from HUVECs after treatments using Trizol method. Primers (5S rRNA: 203906; U6 snRNA: 203907; hsa-miR-20b: 204755; and hsa-miR-34a: 204318) for miR analysis were purchased from Exiqon and were used as per manufacturer's recommendations. Mimics of miR-20b (MC10975) and miR-34a (MC11030) and negative control (4464058) were ordered from Life Technologies. All real-time PCR reactions were performed using the 7500 real-time PCR system (Applied Biosystems). The amplification steps consisted of denaturation for 10 min at 95°C, followed by 40 cycles of denaturation at 95°C for 10 s and then annealing at 60°C for 1 min using the SYBR green master mix (203450; Exiqon). U6 and 5s were used as endogenous controls, and fold changes were calculated for each gene. Each RNA sample assay was run in triplicate, and the assay was repeated for three times.

In vitro capillary-like tube formation assays.

Matrigel assay was performed as described previously (33). Briefly, HUVECs were transfected with expression vectors for MCPIP or its mutant, D141N or empty vector for 24 h before being trypsinized and seeded onto the surface of the Matrigel according to the manufacturer's protocol, followed by incubation in at 37°C in 5% CO₂ for 24 h. Tube formation was quantified using photographs captured by phase-contrast microscope.

Purification of MCPIP.

Human embryonic kidney (HEK)293 cells were transfected with pCMV-MAT-FLAG -MCPIP expression vector for 48 h. Transfected cells were lysed at 4°C in Cell-Lytic M lysis buffer (C2978; Sigma-Aldrich, St. Louis MO) supplemented with a protease and phosphatase inhibitor cocktails (Sigma-Aldrich). Upon centrifugation, the cleared lysate was loaded onto a column containing Ni-NTA agarose beads (Qiagen) that was previously prepared as per manufacturer's protocol. The column was washed several times with wash buffer (50 mM Tri·HCl pH 7.5 and 150 mM NaCl) along with increasing concentrations of imidazole (10 mM and 50 mM). MCPIP protein was eluted using 500 mM imidazole.

Preparation of ubiquitinated HIF-1α substrate.

HEK cells were transfected with HA-HIF-1α-pcDNA3 expression vector (cat. no. 18949; Addgene). Hypoxic conditions were induced by CoCl₂ (200 μM) and proteasome inhibition by MG132 (10 μM) for 24 h. The lysates were subjected to immunoprecipitation with HA-coated beads to yield HIF-1α-HA as the substrate. In vitro ubiquitination of the substrate was performed by Ubiquitin-Protein Conjugation Kit, (cat. no. K-960; Boston Biochem, Boston, MA) to yield ubiquitinated-HIF-1α substrate. Ub-HIF-1α was incubated with purified MCPIP enzyme with or without ubiquitin aldehyde, a deubiquitinase inhibitor, or the MCPIP mutant D141N. Hydrolysis of the Ub-HIF-1α substrate was observed by immunoblotting with ubiquitin antibody.

Deubiquitinase assay.

Experiments to determine ubiquitin hydrolysis were performed in triplicates by 1) Ub-AFC assay: purified MCPIP or D141N protein (1 μg) was incubated with 1 μM Ub-AFC (cat. no. U-551; Boston Biochem) in buffer containing 50 mM Tris·HCl pH 7.0, 10 mM DTT, and 150 mM NaCl in a final volume of 200 μl. Assays were performed at 37°C for 4 h. The fluorescence signals were detected at excitation of 400 nm and emission of 505 nm in a time-dependent manner. 2) High Molecular Weight K-63 linked PolyUbiquitin (1 μg; cat. no. UC 316; Boston Biochem) was incubated in buffer containing 50 mM Tris·HCl pH 7.0, 10 mM DTT, and 150 mM NaCl in a final volume of 200 μl with purified MCPIP or D141N protein (1 μg) at 37°C for 1 h. The reaction mixture was used for immunoblot analysis with anti-ubiquitin antibodies.

p38 MAPK activity assay.

Assay to determine activity was performed by the p38 MAPK Activity Assay Kit (cat. no. CS0250; Sigma) as per manufacturer's recommendations. Briefly HUVECs were transfected with MCPIP expression vector with or without prior treatment of p38 inhibitor SB 203580. The whole cell lysate was immune precipitated with p38 antibody, and the elute was incubated with ATF2 substrate at 30°C. After 30 min, the reaction mixture was run on SDS-PAGE and immunoblotted against phospho-ATF2 antibody.

Immunoblot analysis.

HUVECs from different experimental conditions were lysed with Cell Lytic lysis Buffer (Sigma). Protein samples (50 μg) were subjected to SDS-PAGE using 10% polyacrylamide or 4–20% (NB10–420; NuSep) Tris·HCl gels and transferred using standard protocols. Immunoblot analysis was performed using the primary antibodies from Santa Cruz Biotechnology: anti-mouse GAPDH (1:1,000; cat. no. 47724); anti-rabbit HIF-1α (1:500; cat. no. 10790); anti-mouse SIRT-1 (1:500; cat. no. 74504); anti-mouse TSP-1 (1:500; cat. no. 74538) and horseradish-conjugated mouse (1:5,000; cat. no. 2005); and rabbit antibodies (1:5,000; cat. no. 2317) and ubiquitin (1:1,000; cat. no. VU 101; Life sensors). Immunoreactive proteins were analyzed using ECL kit.

Statistical analysis.

All experiments were repeated three times. The error bars are represented as ± SE. An asterisk indicates a significant difference compared with the control as indicated in each experiment. P value of <0.05 was considered significant and was determined by Student's t-test.

RESULTS

Hypoxia-induced angiogenesis is mediated via MCPIP.

Hypoxia-induced angiogenesis is augmented by cytokine production (2). MCPIP was reported to mediate angiogenesis induced by inflammatory cytokines (38). To determine if MCPIP mediates hypoxia-induced angiogenesis, HUVECs were transfected with siRNA specific for MCPIP or nonspecific scrambled control before hypoxic (1% oxygen) or normoxic (21% oxygen) incubation for 6 h. siRNA specific for MCPIP, but not scrambled siRNA, significantly inhibited hypoxia-induced production of HIF-1α and angiogenesis thus strongly suggesting that hypoxia-induced angiogenesis is mediated via MCPIP (Fig. 1, A–D).

Fig. 1.

Hypoxia-induced angiogenic differentiation is mediated via monocyte chemoattractant protein-1 (MCP1–1)-induced protein-1 (MCPIP1). Human umbilical vein endothelial cells (HUVECs) were transfected with small interfering (si)MCPIP or siScramble (Scra) for 4 h before induction with 1% oxygen (hypoxia) for 6 h. Controls were kept under 21% oxygen (normoxia). After 6 h, transcript levels (A) and protein (B) levels were evaluated for hypoxia-inducible factor (HIF-1α) after being standardized to endogenous control GAPDH; *P < 0.05. Under normoxia, HIF-1α protein levels were undetectable. After 4 h of transfection as in (A and B) cells were trypsinized and placed on Matrigel before induction with 1% oxygen (hypoxia) or 21% oxygen (normoxia) for 6 h. C: phase-contrast photomicrographs of the tube formation is represented. D: quantification of phase-contrast photomicrographs of the tube formation is represented; one-way ANOVA was calculated between siScra and siMCPIP under hypoxia; *P < 0.02. Bonferroni corrected post hoc analyses suggest that tube formation under hypoxic conditions was significantly inhibited on MCPIP knockdown.

MCPIP expression results in the HIF-1α localization in the nucleus and induction of its target genes, COX-2 and VEGF.

Forced expression of MCPIP induces HIF-1α (33). To determine whether MCPIP expression resulted in HIF-1α localization in the nuclei, HUVECs were transfected with MCPIP expression vector or empty vector for 24 h. Immunocytochemistry using antibody against HIF-1α was performed. DAPI was used for counterstaining the nuclei. Fluorescence microscopic images showed that MCPIP expression resulted in nuclear localization of HIF-1α compared with the empty vector control (Fig. 2A). Transfection with MCPIP expression vector resulted in elevated levels of MCPIP and HIF-1α (Fig. 2, B and C). Expression levels of HIF-target genes, COX-2 and VEGF, were also higher in cells expressing MCPIP. Furthermore, specific knockdown of HIF-1α inhibited MCPIP-induction of VEGF and COX-2 (Fig. 2D). These results suggest that MCPIP expression promotes entry of HIF-1α into the nucleus resulting in induction of VEGF and COX-2 production.

Fig. 2.

Forced expression of MCPIP resulted in the nuclear entry of HIF-1 and induction of cyclooxygenase-2 (COX-2) and VEGF. HUVECs were transfected with MCPIP expression vector or empty vector or siMCPIP for 24 h. A: cells were fixed and immunocytochemistry was performed using antibody against HIF-1α. Nuclei were counterstained with DAPI. Images were merged. Inset: nuclei at ×40. Total cell lysate was evaluated for MCPIP (B) and HIF-1α protein (C) levels; *P < 0.05. D: HUVECs were treated with siRNA specific for HIF-1α or scrambled siRNA as a control 3 h before being transfected with MCPIP. After 24 h, RNA was isolated for transcript analysis by real-time PCR to detect COX-2 and VEGF expression; *P < 0.05.

Since p38MAPK activation was reported to be involved in mediating VEGF-induced angiogenesis (31, 55), we sought to determine if MCPIP induces the activation of p38 MAPK by transfecting HUVECs with MCPIP expression vector. After immunoprecipitation of total p38 from the whole cell lysates, the kinase activity of p38 was checked on its model substrate, ATF2. The reaction mixture was immunoblotted with phospho-ATF2 antibody. Results showed that cells expressing MCPIP had increased ATF2 phosphorylation compared with the cells transfected with empty vector. SB 203580, a p38 MAPK inhibitor blocked phosphorylation of ATF2 (Fig. 3A). Our results suggest that MCPIP expression induces p38 MAPK activation. Furthermore, to determine if MCPIP- induced angiogenic differentiation involves the activation of p38 MAPK, HUVECs were treated with the p38 MAPK inhibitor SB 203580 before transfection with MCPIP expression vector. Our results indicate that MCPIP-induced tube formation was drastically reduced by p38 MAPK inhibition (Fig. 3B). Our data thus suggest that angiogenesis MCPIP-induced angiogenesis is mediated via induction of p38 MAPK activation, as expected from the involvement of VEGF induced by MCPIP via HIF-1α.

Fig. 3.

MCPIP induced angiogenesis via p38 MAPK activation. HUVECs were treated with or without SB 203580 (20 μM) before transfection with MCPIP expression vector. After 24 h of transfection, cells were trypsinized and the whole cell lysate was immunoprecipitated with beads coated with p38 antibody (A). After elution, p38 MAPK was incubated with substrate, ATF2. Phosphorylation of ATF2 was used as a measure to determine p38MAPK activity placed on Matrigel for 24 h (B). Phase-contrast photomicrographs of the tube formation are represented. Quantification of phase-contrast photomicrographs of the tube formation; *P < 0.002.

SIRT-1 mediates MCPIP-induced angiogenic differentiation.

Studies have shown that loss of SIRT-1 function blocks angiogenesis (36), thus suggesting its proangiogenic role. To determine if SIRT-1 is induced by expression of MCPIP, real-time PCR and immunoblot analysis were performed to examine the effect of MCPIP expression on the levels of SIRT-1 (Fig. 4, A and B). Our results show that SIRT-1 levels were significantly elevated in cells expressing MCPIP, thus suggesting that MCPIP induces SIRT-1. To determine if MCPIP-induced endothelial differentiation is mediated via SIRT-1, HUVECs were transfected with siRNA specific for SIRT-1 or siScramble before transfection with MCPIP expression vector. Knockdown of SIRT-1 inhibited MCPIP-induced tube formation (Fig. 4, C and D). Our results suggest that SIRT-1 induction is involved in the MCPIP-induced angiogenesis.

Fig. 4.

MCPIP-induced angiogenesis is mediated via silent information regulator (SIRT-1) induction. HUVECs were transfected with empty vector (MAT) or MCPIP (MAT-MCPIP). After 24 h, transcript levels (A) and protein levels (B) were evaluated for SIRT-1; *P < 0.05. C: cells were trypsinized and placed on Matrigel. D: quantification of phase-contrast photomicrographs of the tube formation is represented; P < 0.005.

MCPIP expression reduces the levels of antiangiogenic factors, NF-κB target genes TSP-1 and angiogenic inhibitor VEGI.

TSP-1 is a well-known inhibitor of angiogenesis (40). To determine if MCPIP expression has an effect on TSP-1 levels, transcript and protein analysis were performed. Expression of MCPIP resulted in lower levels of the angiogenesis inhibitor TSP-1 compared with the empty vector (Fig. 5, A and B). Furthermore, knockdown of SIRT-1 gene resulted in higher levels of TSP-1 in cells expressing MCPIP (Fig. 5, C and D). This result suggests that reduction in TSP-1 level caused by MCPIP expression is mediated via SIRT-1. Furthermore, since it was reported that MCPIP negatively regulated NF-κB activation (25), we sought to determine if this mechanism would be of importance in promoting MCPIP-induced angiogenesis. Our data show that expression of MCPIP resulted in a reduction in nuclear levels of p65 compared with cells transfected with empty vector (Fig. 5E). Our results also show that MCPIP expression caused significant decrease in the levels of antiangiogenic VEGI (Fig. 5F) and TSP-1 whose production is known to require NF-κB activation (48). Thus our data suggest that inhibition of NF-κB activation by MCPIP would contribute to the angiogenic activity of MCPIP by reducing the level of antiangiogenic VEGI and TSP-1.

Fig. 5.

MCPIP overexpression reduced levels of antiangiogenic factors, thrombospondin-1 (TSP-1) and the angiogenic inhibitor VEGI. HUVECs were transfected with MCPIP expression vector or empty vector for 24 h. After 24 h, transcript levels (A) were evaluated for TSP-1 (*P < 0.02), and protein levels (B) were assayed by immunoblot analysis with antibody against TSP-1 with GAPDH as a control. HUVECs were transfected with siRNA specific for SIRT-1 or scrambled (Scr) siRNA for 3 h before transfecting with MCPIP expression vector After 24 h, transcript levels (C) were evaluated for TSP-1 (*P < 0.02) and protein levels (D) were assayed by immunoblot analysis with antibody against TSP-1 with GAPDH as a control (*P < 0.05). E: HUVECs were transfected with for MCPIP expression vector or empty vector was used as a control. p65 nuclear protein levels were measured by immunoblot analysis. Histone was used to check for purity of the nuclear extract and as a control for densitometric analysis, and transcript levels (F) of VEGI were analyzed by RT-PCR; *P < 0.05.

Antidicer RNase activity of MCPIP suppresses the levels of miRs modulating HIF-1α and SIRT-1 expression.

To explore the potential involvement of the enzymatic activities of MCPIP in its promotion of angiogenesis, a MCPIP mutant, D141N, that is known to have lost the RNase activity (29, 32, 43), was used for the in vitro Matrigel assay. Results showed that cells expressing D141N mutant showed significantly reduced tube formation compared with the cells expressing wild-type MCPIP (Fig. 6) suggesting the importance of RNase activity of MCPIP in angiogenic differentiation.

Fig. 6.

Loss of angiogenic potential by MCPIP-mutant D141N. HUVECs were transfected with MCPIP-wild-type (Wt) and MCPIP-mutant-D141N expression vectors. After 24 h, cells were trypsinized and placed on Matrigel. Quantification of phase-contrast photomicrographs of the tube formation is represented; *P < 0.005.

To test for the role of the antidicer activity of MCPIP in promoting angiogenesis, HUVECs were transfected with expression vector for MCPIP or its D141N mutant, and production of miR that could be involved in the regulation of angiogenesis was examined. miR-20b has been reported to reduce HIF-1α protein levels. When RT-PCR of the miRs isolated from cells transfected with expression vectors for MCPIP or D141N mutant or empty vector was performed, expression of MCPIP was found to cause reduction in the level of miR-20b. This reduction was, however, not seen with the RNase-dead mutant D141N (Fig. 7A). Furthermore, transcript level of miR-34a, an miR known to suppress SIRT-1 levels (37), was measured. Results showed that expression of MCPIP significantly reduced the levels of miR-34a, but the RNase-dead mutant D141N failed to inhibit miR-34a production (Fig. 7B). Our results also show that levels of miRs 20b induced by inflammatory agents such as IL-1β or TNF-α are MCPIP dependent (Fig. 7, C and D).

Fig. 7.

MCPIP inhibits the production of antiangiogenic microRNA (miR)-20b and miR-34a and their mimetics inhibit MCPIP-induced angiogenesis. HUVECs were transfected with empty vector or MCPIP or the RNase-dead mutant D141N. After 24 h, levels of miR-20b (A) or miR-34a (B) were measured by RT-PCR. C and D: HUVECs were treated with TNF-α (100 ng/ml) or IL-1β (10 ng/ml) for 2 h before being transfected with siScramble or siMCPIP for 24 h. Levels of miR-20b were measured by quantitative PCR; *P < 0.01; #P < 0.5. E: HUVECs were transfected with miR-negative control (NC) miR-20b mimic or miR-34a mimic for 3 h before transfecting with MCPIP expression vectors. After 24 h, cells were trypsinized and placed on Matrigel. Quantification of phase-contrast photomicrographs of the tube formation is represented; *P < 0.005.

If MCPIP-induced angiogenesis involves inhibition of biogenesis of these antiangiogenic miRs, mimetics of their miRs should inhibit MCPIP-induced angiogenesis. To test this possibility, HUVECs were transfected with mimetics of miR-20b or miR-34a or negative control before transfection with MCPIP expression construct. The mimetics of miR-20b or miR-34a could thus inhibit their targets, HIF-1α and SIRT-1, respectively. Our results showed (Fig. 7E) that mimetics for miR-20b or miR-34a inhibited MCPIP-induced tube formation. These results strongly suggest that antidicer RNase activity of MCPIP represses the levels of these antiangiogenic miRs and thus promotes angiogenesis.

D141N mutant was reported to have lost deubiquitinase activity when tested with octaubiquitin substrate (25). To test whether the loss of ubiquitinase activity might also be involved in the loss of angiogenic activity of D141N mutant, we tested for its deubiquitinase activity against physiologically relevant substrates. We tested whether MCPIP can deubiquitinate ubiquitinated HIF-1α in vitro. HEK cells were transfected with HA-HIF-1α expression vector under conditions described in materials and methods. HA-HIF-1α isolated by immunoprecipitation was ubiquitinated in vitro. Ubiquitinated HIF-1α substrate was incubated with purified MCPIP in the presence or absence of ubiquitin aldehyde, a deubiquitinase inhibitor. The mixture was immunoblotted with ubiquitin antibody. Results demonstrate that MCPIP deubiquitinates the ubiquitinated HIF-1α substrate and the deubiquitinase inhibitor ubiquitin aldehyde prevented the hydrolysis (Fig. 8A). This deubiquitinase activity would explain the promotion of the nuclear transport of HIF-1α seen in Fig. 2. Purified D141N showed deubiquitinase activity similar to the wild-type MCPIP also when assayed with a model substrate, Ub-AFC (Fig. 8B). We tested whether the mutant could hydrolyze high molecular weight polyubiquitin (Poly Ub) that has been reported to play an important role in NF-κB activation (56). Our results showed that both MCPIP and its mutant D141N were able to hydrolyze poly Ub (Fig. 8C) suggesting that the MCPIP mutant D141N has deubiquitinase activity on substrates relevant to the biological functions of MCPIP. Our findings thus suggest that the loss of antidicer RNase activity in D141N mutant is probably the reason for its inability to induce angiogenesis.

Fig. 8.

Deubiquitinase activity of RNase active MCPIP mutant, D141N. A: ubiquitinated HIF-1α substrate of MCPIP was incubated with purified MCPIP (1 μg) or ubiquitin (Ub) aldehyde for at 37°C. After 1 h the mixture was immunoblotted against ubiquitin antibody. B: Ub-AFC (1 μM) was incubated with MCPIP1 or its mutant, D141N (1 μg) at 37°C. Ub-AFC without any protein was used as a negative control. Fluorescence was measured at excitation at 400 nm and emission at 505 nm in a time-dependent manner. C: high molecular weight polyubiquitin (1 μg) was incubated with purified MCPIP (1 μg) or MCPIP mutant D141N (1 μg) at 37°C for 1 h. The reaction mixture was immunoblotted with ubiquitin antibody.

DISCUSSION

Chronic inflammation plays a major role in several diseases such as cancer, cardiovascular diseases, obesity and is marked by elevated levels of proinflammatory cytokines, including TNF-α, IL-1β, IL-8, and MCP-1. Inflammatory cytokines are known to induce a novel zinc-finger protein, MCPIP encoded by the ZC3H12A gene (13, 32, 63). A recent study has shown that MCPIP mediates angiogenesis induced by inflammatory cytokines (38). Results from the present study provide new insights into the possible underlying mechanisms that mediate MCPIP-induced angiogenesis. We show that MCPIP promotes the expression of proangiogenic molecules and inhibits the synthesis of antiangiogenic molecules, thus tilting the balance towards promotion of angiogenesis. The antidicer RNase activity of MCPIP elevates the levels of HIF-1α by inhibiting the synthesis of miR-20b. MCPIP expression also induces SIRT-1 expression, which results in the inhibition of antiangiogenic TSP-1 production. Our results further show that the antidicer RNase activity of MCPIP results in the inhibition of synthesis of miR-34a that is known to suppress SIRT-1 levels. Moreover, miR-34a production is also known to be suppressed by inhibition of NF-κB activation as has been reported to occur on MCPIP expression (25). Experimental results presented with the MCPIP mutant D141N that has lost the angiogenic potential and RNase activity but has intact deubiquitinase activity, reveal for the first time that the antidicer RNase activity of MCPIP is critical in mediating MCPIP-induced angiogenesis. Deubiquitinase activity of MCPIP on ubiquitinated HIF-1α demonstrated here helps to explain the MCPIP-induced stabilization of HIF-1α and its nuclear translocation. Thus deubiquitinase activity of MCPIP probably helps to promote angiogenesis. Experimental data also suggest that inhibition of NF-κB activation by MCPIP contributes to its angiogenic activity by lowering the levels of antiangiogenic VEGI.

Sirtuins are a family of nicotinamide adenine dinucleotide (NAD⁺)-dependent histone deacetylases (HDAC) that regulate gene expression (20, 45). SIRT-1 plays a vital role in regulating cellular differentiation by transcriptional repression of several transcriptional regulators, including forkhead box type O transcription factors (FOXO) and the tumor suppressor protein p53 (4, 36). It was reported that inhibition or knockdown of SIRT-1 expression in both zebra fish and mice resulted in impairment of vasculature development suggesting that SIRT-1 mediates angiogenic signaling (11). Studies have shown that hyperacetylation of p53 results in its stabilization and results in onset of apoptosis. Conversely, p53 deacetylation by induction or overexpression of SIRT-1 would reduce p53 activity and promote cell survival (52). SIRT-1 downregulates the stability of p53 (27). p53 promotes the transcription of TSP-1, an inhibitor of angiogenesis (8). Our findings suggest that SIRT-1 modulates MCPIP-induced angiogenic differentiation. This induction of SIRT-1 resulted in the repression of the inhibitor of angiogenesis TSP-1. Furthermore, it was reported that SIRT-1 deacetylates lysine 310 on RelA/p65 protein in the NF-κB complex thus inhibiting the transactivation capacity of the NF-κB complex (59) and thus preventing NF-κB activation in HUVEC as demonstrated by our data. Also, since there are multiple NF-κB binding sites on the promoter of TSP-1, inhibition of NF-κB activation may also reduce TSP-1 levels thus adding a new dimension to the mechanism by which SIRT-1 regulates MCPIP-induced angiogenesis.

We found that D141N mutant of MCPIP could not induce angiogenesis even though we demonstrate that it has deubiquitinase activity. Therefore, the antidicer RNase activity that had been shown to be lost in this mutant (47) appeared likely to be critical in promoting angiogenesis. A recent study suggested that MCPIP cleaves the terminal loops of precursor miRs, thus antagonizing dicer activity to inhibit miR biogenesis (47). There is mounting evidence suggesting that alterations in expression of miRs are involved in several major diseases such as cardiovascular or neurodegenerative diseases and cancer (19). Since miRs are known to regulate many aspects of angiogenesis, it was essential to determine if the RNase activity of MCPIP would modulate MCPIP-induced angiogenic differentiation. The present findings suggest that MCPIP expression results in the suppression of the levels of miRs miR-20b and miR-34a that are known to bind to the 3′-UTR of HIF-1α (22) and SIRT-1 (49, 57), respectively. Interestingly, miR-34a transcription was reported to be induced by NF-κB activation as the promoter of miR-34a has NF-κB binding sites (23). Thus the reduction in miR-34a levels caused by MCPIP expression could be due to inhibition of NF-κB activation possibly by induction of SIRT-1 and/or via its own deubiquitinase activity or by the antidicer RNase activity of MCPIP repressing the miR levels. That the antidicer RNase activity of MCPIP is involved in the induction of angiogenesis by its ability to inhibit biogenesis of miR-20b and miR-34a was supported by the finding that MCPIP expression inhibited the production of the miRs and their mimetics inhibited MCPIP-induced angiogenesis. This conclusion is further strengthened by our finding that transfection with expression vector for RNase-dead mutant D141N (47) did not suppress the production of miR-20b and failed to induce angiogenic differentiation.

An imbalance between the levels of oxygen supply and its demand is critical in the development of inflammatory diseases such as diabetic retinopathy, psoriasis, and tumorigenesis (3). Cellular adaptations under hypoxia are modulated by the induction of proinflammatory cytokines and HIF-1α (18). HIF-1α is a key regulator and a transcription factor that mediates an array of cellular pathways such as angiogenesis by promoting the transcription of several target genes including COX-2 and VEGF (14, 35, 41). Under normoxia, HIF-1α is tightly regulated by O₂-dependent prolyl hydroxylation that aids in polyubiquitination by E3 ubiquitin ligase, pVHL, leading to the degradation of HIF-1α by the proteosomal pathway (30). Under hypoxic conditions, however, the hydroxylation of prolyl is inhibited thus resulting in accumulation and increased activity of HIF-1α. State of ubiquitination is an important biochemical modification, which regulate a wide range of cell biological processes (46). Deubiquitination, a mechanism of reversing ubiquitination adds, another important modulatory modification in regulating cellular functions. VDU2, a pVHL-interacting deubiquitinating enzyme 2, has been known to deubiquitinate and stabilize HIF-1α, thus preventing HIF-1α from proteosomal degradation (24). MCPIP is known to have deubiquitinase activity (25). Furthermore, angiogenic differentiation induced by MCPIP was reported to be mediated via HIF-1α induction (33). However, the mechanism/s underlying HIF-1α induction by MCPIP was unknown. Our demonstration that MCPIP can deubiquitinate ubiquitinated HIF-1α suggests that MCPIP would stabilize HIF-1α via its deubiquitinase activity. The stabilized HIF-1α would enter the nuclei and promote the transcription of its target genes VEGF and COX-2, important players in angiogenesis. Thus MCPIP-induced deubiquitination of ubiquitinated HIF-1α is a probable mechanism by which MCPIP helps to promote angiogenesis. Moreover, it was reported that the deubiquitinase activity of MCPIP negatively regulates NF-κB activation (25). Our finding showed reduced nuclear levels of p65 subunit of NF-κB on MCPIP expression suggesting a probable mechanism by which MCPIP may promote angiogenesis. The results presented here show that the antidicer activity of MCPIP is critical for its angiogenic activity and that the deubiquitinase activity also probably contributes to the angiogenic activity. However, in the absence of an MCPIP mutant, this is deubiquitinase dead but RNase active, and it is difficult to assess the relative importance of the deubiquitinase activity in the promotion of angiogenesis by MCPIP. Taken together, the findings of the present study delineate the probable molecular mechanisms by which MCPIP mediates inflammatory angiogenesis (Fig. 9). In summary, MCPIP mediates angiogenic differentiation by promoting the synthesis of proangiogenic VEGF and COX-2, and by suppressing the production of antiangiogenic miRs, miR-20b and miR-34a, TSP-1, and VEGI, thus shifting the balance towards angiogenesis. This is the first demonstration of the involvement of antidicer RNase activity of MCPIP in any of its biological functions. Furthermore, our findings that TNF-α/IL-1β-induced inhibition of miR-20b production is MCPIP-dependent also suggests a possible mechanism in angiogenesis associated with inflammation as had been reported earlier (38). Since microRNA biogenesis (28) has been shown to be involved in regulating several human pathologies such as impaired wound healing, cancer, and heart disease (39), findings from our study reveal potential targets that may contribute to the development of novel therapeutic strategies.

Fig. 9.

Schematic representation of the mechanisms involved in MCPIP-induced differentiation. MCPIP mediates angiogenic differentiation by inducing HIF-1α and SIRT-1 levels and by inhibiting NF-κB activation via following mechanisms: deubiquitination of ubiquitinated HIF-1α thus stabilizing HIF-1α (A) and suppressing the levels of miR-20b and miR-34a (B) thus augmenting the levels of HIF-1α and SIRT-1. Stabilized HIF-1α enters the nucleus for promoting the transcription of VEGF and COX-2. Increased levels of SIRT-1 repress levels of the angiogenesis inhibitor, TSP-1, thus promoting MCPIP-induced angiogenesis. SIRT-1 is also reported to inhibit NF-κB that in turn results in reduced levels of TSP-1 and miR-34a levels (C) MCPIP negatively modulates NF-κB activation resulting in lower levels of the angiogenic inhibitor VEGI. Darker lines indicate that antidicer activity of MCPIP is critical in mediating angiogenesis.

GRANTS

This work was supported in part by National Heart, Lung, and Blood Institute Grant HL-69458.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: A.R. and P.E.K. conception and design of research; A.R., M.Z., and Y.S. performed experiments; A.R. analyzed data; A.R. interpreted results of experiments; A.R. prepared figures; A.R. drafted manuscript; A.R. and P.E.K. edited and revised manuscript; P.E.K. approved final version of manuscript.