Cytokine-Mediated Degradation of the Transcription Factor ERG Impacts the Pulmonary Vascular Response to Systemic Inflammatory Challenge

Abstract

Background: During infectious diseases, proinflammatory cytokines transiently destabilize interactions between adjacent vascular endothelial cells (ECs) to facilitate the passage of immune molecules and cells into tissues. However, in the lung, the resulting vascular hyperpermeability can lead to organ dysfunction. Previous work identified the transcription factor ERG (erythroblast transformation-specific-related gene) as a master regulator of endothelial homeostasis. Here we investigate whether the sensitivity of pulmonary blood vessels to cytokine-induced destabilization is due to organotypic mechanisms affecting the ability of endothelial ERG to protect lung ECs from inflammatory injury.

Methods: Cytokine-dependent ubiquitination and proteasomal degradation of ERG were analyzed in cultured HUVECs (human umbilical vein ECs). Systemic administration of TNFα (tumor necrosis factor alpha) or the bacterial cell wall component lipopolysaccharide was used to cause a widespread inflammatory challenge in mice; ERG protein levels were assessed by immunoprecipitation, immunoblot, and immunofluorescence. Murine Erg deletion was genetically induced in ECs (Erg^fl/fl;Cdh5[PAC]-Cre^ERT2), and multiple organs were analyzed by histology, immunostaining, and electron microscopy.

Results: In vitro, TNFα promoted the ubiquitination and degradation of ERG in HUVECs, which was blocked by the proteasomal inhibitor MG132. In vivo, systemic administration of TNFα or lipopolysaccharide resulted in a rapid and substantial degradation of ERG within lung ECs but not ECs of the retina, heart, liver, or kidney. Pulmonary ERG was also downregulated in a murine model of influenza infection. Erg^fl/fl;Cdh5(PAC)-Cre^ERT2 mice spontaneously recapitulated aspects of inflammatory challenges, including lung-predominant vascular hyperpermeability, immune cell recruitment, and fibrosis. These phenotypes were associated with a lung-specific decrease in the expression of Tek-a gene target of ERG previously implicated in maintaining pulmonary vascular stability during inflammation.

Conclusions: Collectively, our data highlight a unique role for ERG in pulmonary vascular function. We propose that cytokine-induced ERG degradation and subsequent transcriptional changes in lung ECs play critical roles in the destabilization of pulmonary blood vessels during infectious diseases.

Keywords: capillary permeability; endothelial cells; inflammation; lung; proteolysis; transcription factors.

Figure 1.

The proinflammatory cytokine TNFα (tumor necrosis factor alpha) promotes ERG (erythroblast transformation-specific–related gene) ubiquitination and proteolytic degradation in vitro. A, TNFα (10 ng/mL) was applied directly to cultured human umbilical vein endothelial cells (HUVECs), and ERG expression was assessed by immunoblot at the indicated times. B, TNFα (10 ng/mL for 6 h)-induced ERG downregulation in HUVECs was assessed by immunoblot (n=3) following a 30-minute pretreatment with MG132 (10 μM). C, Following a similar TNFα/MG132 treatment, ERG was immunoprecipitated from HUVECs, and samples were immunoblotted for ERG and ubiquitin (Ub). Longer exposure of the ERG immunoblot revealed the presence of high-molecular-weight, ubiquitinated ERG isoforms. D, Alignment of human, mouse, rat, and pig ERG sequences highlighting 6 conserved lysine residues (K) predicted as sites of ubiquitination. E and F, Mutation of all putative sites to arginine (Erg^K67-282R) reduced the presence of ubiquitinated ERG isoforms (E) and prolonged ERG half-life (F) when constructs were expressed in HeLa (Henrietta Lacks; human epithelial) cells (n=3). P values were determined by a nonparametric Kruskal-Wallis test followed by a Dunn multiple comparison between the indicated groups. ns indicates nonsignificant.

Figure 2.

Proinflammatory stimuli promote lung-specific ERG (erythroblast transformation-specific–related gene) protein downregulation. A and B, ERG expression was assessed by immunoblot (A) and quantified (n=4–5; B) in the lung and heart at 3 hours after intravenous injection of lipopolysaccharide (LPS; 1 mg/kg body weight [BW]), TNFα (tumor necrosis factor alpha), IL (interleukin)-1α, IL-1β, or IL-18 (50 µg/kg BW for each cytokine). C and D, WT (wild-type) mice were administered intraperitoneal injections of LPS (4 mg/kg BW) or vehicle (0.9% NaCl). Erg transcript expression was quantified by qPCR (real-time quantitative reverse transcription polymerase chain reaction) in the indicated organs at 2 hours after LPS exposure (n=3; C), and ERG protein expression was quantified by immunoblot at 8 hours after LPS exposure (n=4–6; D). E, Expression of ERG (gray) in CD31⁺ (cluster of differentiation 31; green) endothelial cells (white arrows) within the lung was assessed by immunostaining tissue sections collected 8 hours after LPS exposure. DAPI (4',6-diamidino-2-phenylindole; blue) was used as a nuclear counterstain. P values were determined by a 1-way ANOVA followed by a Dunnett comparison of each treatment to the vehicle-treated condition (B), unpaired t tests with Welch correction comparing vehicle- and LPS-treated samples within each organ (C), or unpaired t tests comparing vehicle- and LPS-treated samples within each organ (D).

Figure 3.

Proinflammatory cytokines promote lipopolysaccharide (LPS)-induced ERG (erythroblast transformation-specific–related gene) degradation in the lung. A through C, ERG expression was assessed by immunoblotting lung lysates at 4 hours after intraperitoneal administration of LPS (4 mg/kg body weight [BW]) in mice following global deletion of Tnfa (Tnfa^−/−; n=5–6; A), global deletion of Il1r1 (Il1r1^Gko; n=4–5; B), or endothelial cell–specific deletion of Il1r1 (Il1r1^ECko; n=5; C). D and E, Wild-type mice were given an intraperitoneal injection of LPS (10 mg/kg BW). At 0.5 to 24 hours after LPS challenge, ERG protein and transcript expression were quantified in lung (D) and heart (E) tissues by immunoblot and qPCR (real-time quantitative reverse transcription polymerase chain reaction), respectively (n=3–4). F, LPS (4 mg/kg BW)-induced ERG downregulation in the lung was assessed by immunoblot following a 3-hour pretreatment with vehicle or MG132 (10 mg/kg BW; n=3–6). P values were determined by a 2-way ANOVA followed by a Sidak multiple comparison test; data that failed a Spearman equal variance test were log transformed before analysis (B, C, and F). ns indicates nonsignificant.

Figure 4.

ERG (erythroblast transformation-specific–related gene) is downregulated in the lung during pulmonary influenza infection. A, ERG expression was quantified by immunoblot using murine lung tissue collected 2 and 6 days after low (800 egg infective dose [EID]) and high (1050 EID) doses of influenza were administered via the trachea (n=3). B, Expression of ERG (gray) in endothelial cells (ECs; CD31 [cluster of differentiation 31]; green) was assessed by immunostaining lung tissue sections collected from control or influenza-infected mice. Reduced ERG expression within pulmonary capillary ECs (but not within ECs of larger caliber vessels) of influenza-infected mice is visualized by a surface intensity representation of the ERG channel (right); DAPI (4',6-diamidino-2-phenylindole; blue) was used as a nuclear counterstain. C, Immunostaining for the influenza HA (hemagglutinin) protein (red) was used to identify uninfected (No. 1) and infected (No. 2) regions of a lung section. Reduced expression of ERG (gray) was observed in infected regions relative to uninfected regions of the same lung section. P values were determined by a 1-way ANOVA followed by a Dunnett multiple comparison test.

Figure 5.

Tek is transcriptionally regulated by ERG (erythroblast transformation-specific–related gene). A, TEK and ERG expression in human umbilical vein endothelial cells (HUVECs) was assessed by qPCR (real-time quantitative reverse transcription polymerase chain reaction) following siRNA-mediated ERG knockdown (n=3). B and C, ERG chromatin immunoprecipitation (ChIP) sequencing (B) identified occupancy of ERG at the TEK transcriptional start site (TSS) and an intragenic region that colocalizes with H3K27Ac, a marker of active enhancers, in HUVECs. ERG binding to the TSS, an enhancer region, was confirmed by ChIP-qPCR (C; n=4). D, Tek transcript expression was assessed by qPCR in the lung, heart, liver, and kidney of Erg^iECko and control littermates (n=3). E and F, TIE2 (tunica interna endothelial cell kinase 2) protein expression in Erg^iECko and control littermates was assessed by immunoblot (E) and quantified (n=4) in the lung, heart, liver, and kidney (F). G, Expression of TIE2 (gray) in CD31⁺ (cluster of differentiation 31; green) endothelial cells in the lungs of control and Erg^iECko mice was assessed by immunofluorescence. DAPI (4'6-diamidino-2-phenylindole; blue) was used as a nuclear counterstain. H and I, Tek expression was assessed by qPCR in WT (wild-type) murine lung (H) and heart (I) tissues collected 6 to 72 hours after lipopolysaccharide (LPS; 4 mg/kg body weight [BW]) challenge (n=5–7). A, D, and F, P values were determined by unpaired t tests comparing NS and ERG siRNA-treated groups (A) or between control and Erg^iECko animals within each organ (D and F). C, H, and I, P values were determined by 1-way ANOVA followed by Dunnett multiple comparison tests.

Figure 6.

Vascular permeability, adhesion molecule expression, and immune cell recruitment occur in the lungs of Erg^iECko mice. A, Evans blue dye leakage in the lung, heart, liver, and kidney from control and Erg^iECko mice was spectrophotometrically quantified and normalized to dry tissue weight (n=4–6). B and C, Intravenously injected FITC (fluorescein isothiocyanate)-Dextran (green; molecular weight [MW], 150 000 kDa) could be visualized outside of pulmonary blood vessels (CD31 [cluster of differentiation 31]; red; white arrows) in Erg^iECko mice but not in littermate controls (B). However, extravasation of FITC-Dextran in the heart, liver, and kidney following endothelial Erg (erythroblast transformation-specific–related gene) deletion was less apparent (C) ERG (gray); DAPI (4'6-diamidino-2-phenylindole; blue). D through G, Elevated expression of VCAM1 (vascular cell adhesion molecule 1), but not ICAM1 (intercellular adhesion molecule 1), in the lungs of Erg^iECko mice was detected by immunoblot (D) and quantified (E–F; n=4). VCAM1 upregulation, particularly in the capillary bed of Erg^iECko lungs, was confirmed by immunofluorescence imaging (G). H through J, Greater numbers of CD45⁺ (red) immune cells and MPO⁺ (myeloperoxidase; green) neutrophils were observed in the lungs of Erg^iECko mice relative to their control littermates (H), which was quantified by normalization against nuclei (DAPI; blue; I and J). P values were determined by unpaired t tests comparing control and Erg^iECko experimental groups. ns indicates nonsignificant.

Figure 7.

Deletion of endothelial Erg (erythroblast transformation-specific–related gene) leads to organotypic fibrosis and disruption of vascular structures within the lung. A, Masson trichrome staining was used to assess collagen (blue) deposition within the lung, liver, heart, and kidney of Erg^iECko and control mice. B and C, Transmission electron microscopy was used to assess microvascular structures within the lung (B) and kidney (C) from Erg^iECko and control mice. In the lung, deletion of endothelial Erg led to endothelial cell (EC) hypertrophy and basement membrane delamination (red asterisks) resulting in a separation from pneumocytes (P) and the enlargement of the membrane separating alveolar air sacs (AS) and vascular lumens (L).