Endothelial PHD2 deficiency induces apoptosis resistance and inflammation via AKT activation and AIP1 loss independent of HIF2α

Abstract

In hypoxic and pseudohypoxic rodent models of pulmonary hypertension (PH), hypoxia-inducible factor (HIF) inhibition attenuates disease initiation. However, HIF activation alone, due to genetic alterations or use of inhibitors of prolyl hydroxylase domain (PHD) enzymes, has not been definitively shown to cause PH in humans, indicating the involvement of other mechanisms. Given the association between endothelial cell dysfunction and PH, the effects of pseudohypoxia and its underlying pathways were investigated in primary human lung endothelial cells. PHD2 silencing or inhibition, while activating HIF2α, induced apoptosis-resistance and IFN/STAT activation in endothelial cells, independent of HIF signaling. Mechanistically, PHD2 deficiency activated AKT and ERK, inhibited JNK, and reduced AIP1 (ASK1-interacting protein 1), all independent of HIF2α. Like PHD2, AIP1 silencing affected these same kinase pathways and produced a similar dysfunctional endothelial cell phenotype, which was partially reversed by AKT inhibition. Consistent with these in vitro findings, AIP1 protein levels in lung endothelial cells were decreased in Tie2-Cre/Phd2 knockout mice compared with wild-type controls. Lung vascular endothelial cells from patients with pulmonary arterial hypertension (PAH) showed IFN/STAT activation. Lung tissue from both SU5416/hypoxia PAH rats and patients with PAH all showed AKT activation and dysregulated AIP1 expression. In conclusion, PHD2 deficiency in lung vascular endothelial cells drives an apoptosis-resistant and inflammatory phenotype, mediated by AKT activation and AIP1 loss independent of HIF signaling. Targeting these pathways, including PHD2, AKT, and AIP1, holds the potential for developing new treatments for endothelial dysfunction in PH.NEW & NOTEWORTHY HIF activation alone does not conclusively lead to human PH, suggesting that HIF-independent signaling may also contribute to hypoxia-induced PH. This study demonstrated that PHD2 silencing-induced pseudohypoxia in human lung endothelial cells suppresses apoptosis and activates STAT, effects that persist despite HIF2α inhibition or knockdown and are attributed to AKT and ERK activation, JNK inhibition, and AIP1 loss. These findings align with observations in lung endothelial cells and tissues from PAH rodent models and patients.

Keywords: AIP1; AKT; PHD2; STAT; apoptosis.

Graphical abstract

Figure 1.

PHD2 silencing or inhibition induces glycolytic gene expression and apoptosis resistance in human lung microvascular endothelial cells. A: mRNA levels in cells transfected with control (siCTRL) or PHD2 (siPHD2) siRNA for 48 h were measured by qPCR (n = 5). Data shown as geometric means ± geometric SD. B–D: protein levels were measured by Western blotting at 72 h posttransfection (n = 5). Mitochondrial superoxide and mass were analyzed by flow cytometry at the same time point (n = 5). E and F: following 48 h transfection, cells were incubated in either complete medium (CM) or serum- and growth factor-free medium (CM-S/GF) for 24 h. Apoptosis was assessed by caspase 3/7 assay and AV/PI staining (n = 5). G: HIF2α protein levels in cells cultured in CM were evaluated by Western blotting after 24 h treatment with DMOG, a PHD2 inhibitor (n = 6). H: caspase 3/7 activity was measured in CM-S/GF treated cells following 24 h DMOG treatment (n = 5). Western blot protein levels were normalized to β-actin. Data are presented as means ± SD. *P < 0.05, **P ≤ 0.01, and ***P ≤ 0.001. Paired t test (A–D), two-way ANOVA with post hoc t test (E and F), linear regression (G and H). AV/PI, annexin v/propidium iodide; DMOG, dimethyloxalylglycine; qPCR, quantitative real-time PCR.

Figure 2.

Neither HIF2α inhibition nor silencing reverses apoptosis resistance induced by PHD2 silencing in human lung microvascular endothelial cells. Cells were transfected with siRNAs (siCTRL, siPHD2, siHIF2α, and siHIF1β) or their combinations for 48 h in complete media (CM). A and B: following 24 h incubation in serum- and growth factor-free media (CM-S/GF) with PT2567 (HIF2α inhibitor) or vehicle, apoptosis was assessed by caspase 3/7 and AV/PI staining (n = 5). C: cells were further incubated for 24 h in CM with PT2567 or vehicle for Western blotting (n = 5). D: Western blotting for cells 48 h posttransfection (n = 4). E and F: following 24 h incubation in CM-S/GF, apoptosis was assessed by caspase 3/7 and AV/PI staining (n = 5 or 6). G: Western blotting for cells 48 h posttransfection (n = 5). H and I: cells were incubated and tested for apoptosis as in (E) and (F), respectively (n = 6). Western blot protein levels were normalized to β-actin. Data are presented as means ± SD. **P ≤ 0.01 and ***P ≤ 0.001. Two-way ANOVA with post hoc t test. AV/PI, annexin v/propidium iodide.

Figure 3.

PHD2 silencing, independent of HIF2α, activates AKT and ERK, inhibits JNK, and activates STAT1/3 in human lung microvascular endothelial cells. A–C: cells were transfected with control (siCTRL) or PHD2 (siPHD2) siRNA for 48 h. Activation of AKT (pAKT-S473), ERK (pERK-T202/Y204), and JNK (pJNK-T183/Y185) were assessed by Western blotting (n = 5). D–H: cells were transfected as in (A) and incubated for another 24 h with HIF2α inhibitor PT2567 (10 µM) or vehicle control. Activation of AKT, ERK, JNK, STAT1 (pSTAT1-Y701), and STAT3 (pSTAT3-Y705 were assessed (n = 5 or 6). I and J: cells were transfected with siCTRL, siPHD2, siHIF2α (HIF2α siRNA), or their combination for 48 h and incubated for another 24 h. Activation of STAT1 and STAT3 were assessed (n = 6). Western blot protein levels were normalized to β-actin. Data are presented as means ± SD. **P ≤ 0.01 and ***P ≤ 0.001. Paired t test (A–C), two-way ANOVA with post hoc t test (D–J).

Figure 4.

PHD2 loss reduces AIP1 protein expression and AIP1 silencing activates AKT and ERK, inhibits JNK and induces apoptosis resistance in human lung microvascular endothelial cells. A–D: cells were transfected with control (siCTRL), PHD2 (siPHD2), AIP1 (siAIP1) siRNA, or their combination for 48 h. AIP1 protein levels and activation of AKT (pAKT-S473), ERK (pERK-T202/Y204), and JNK (pJNK-T183/Y185) were assessed by Western blotting (n = 5). E and F: cells were transfected as in (A) and incubated for an additional 24 h in a serum- and growth factor-free medium and then assessed for apoptosis by AV/PI staining and caspase-3/7 assay (n = 4 or 5). Western blot protein levels were normalized to β-actin. Data are presented as means ± SD. *P < 0.05, **P ≤ 0.01, and ***P ≤ 0.001. Two-way ANOVA with post hoc t test. AV/PI, annexin v/propidium iodide.

Figure 5.

AKT inhibition inhibits STAT activation and reverses apoptosis resistance in PHD2- or AIP1-silenced human lung microvascular endothelial cells. Cells were transfected with control (siCTRL), PHD2 (siPHD2), AIP1 (siAIP1) siRNA, or their combination. A and B: 48 h posttransfection, cells were further incubated for 24 h in a complete medium with specific PI3Kδ/AKT inhibitor leniolisib (5 µM) or vehicle control. Activation of STAT1 (pSTAT1-Y701), STAT3 (pSTAT3-Y705), and AKT (pAKT-S473) was assessed by Western blotting (n = 5). C: 48 h posttransfection, cells were further incubated for 24 h in a serum- and growth factor-free medium with leniolisib or vehicle and assessed for apoptosis by AV/PI staining (n = 4). D: 72 h posttransfection, cell culture supernatants were collected for measuring secreted CXCL10 levels by ELISA (n = 6). E: lung microvascular endothelial cells isolated from patients with APAH and failed donor (FD) controls were cultured in a complete medium and assessed activation of pSTAT1-Y701 and pSTAT3-Y705 (n = 5). Western blot protein levels were normalized to β-actin. Data are presented as means ± SD. #P = 0.07; *P < 0.05; **P ≤ 0.01. Two-way ANOVA with post hoc t test (A–D), unpaired t test (E). APAH, associated pulmonary arterial hypertension; AV/PI, annexin v/propidium iodide.

Figure 6.

Transcriptome analysis of human lung microvascular endothelial cells silenced for PHD2, AIP1, and both using microarrays. A: Venn diagram of 1,236 genes differentially regulated by silencing PHD2 (siPHD2) or AIP1 (siAIP1), 928 synergistically regulated, and their overlaps. B: positive correlation between LogFC (fold change) of genes regulated by siPHD2 and siAIP1 (853 and 681 genes, respectively). C and D: positive correlation between LogFC of synergistically regulated genes and genes regulated by siPHD2 or siAIP1. E and F: top canonical pathways enriched by ingenuity pathway analysis (IPA, P ≤ 0.01) for genes upregulated by siPHD2 and siAIP1 (E) and genes synergistically upregulated (F). G–L: gene set enrichment analysis (GSEA) revealed significant activation of IFNα and inflammation, and suppression of E2F and MYC targets by silencing of PHD2, AIP1, and both (siCTRL vs. siPHD2, siAIP1, and siBoth; FDR < 0.05). This was accompanied by the induction of anti-apoptotic genes (BIRC3, TAP1, and PAK1) and repression of proapoptotic genes (BCL2L11, FAS, and BTG2). FDR, false discovery rate.

Figure 7.

Diminished PHD2, dysregulated AIP1, and activated STAT and AKT in lungs of pulmonary arterial hypertension (PAH) animal models and in lung endothelial cells (ECs) of patients with PAH. A–C: PAH was induced in rats using the SU5416/hypoxia protocol (SuHx). Protein levels of PHD2, AIP1, pSTAT3 (Y705), STAT3, pAKT (S473), and AKT in lung tissue were measured at week 3 and week 8 compared with normoxic controls (CTRL). Western blot analyses were normalized to β-actin (n = 6–13). D: mice with conditional Phd2 knockout [Egln1^Tie2(Phd2 CKO)] developed PAH. Immunohistochemistry revealed reduced AIP1 expression in the lung vessels of knockout mice compared with wild-type controls (WT). Representative images are shown, and vessels (red arrowheads) and bronchioles (Br) indicated (n = 4). Scale bar, 100 µm. E–G: lung sections from idiopathic PAH (IPAH) patients and failed donor (FD) controls were costained for von Willebrand factor (vWF, green), PHD2, pAKT (S473) or AIP1 (red), and counterstained with Hoechst 33342 (blue). Representative immunofluorescence images are shown (n = 5). Yellow staining identifies the presence of PHD2, pAKT, and AIP1 in lung vascular endothelial cells (ECs). Scale bar: 25 µm. All data are presented as means ± SD. *P < 0.05, **P ≤ 0.01, and ***P ≤ 0.001. Unpaired t test.

Figure 8.

PHD2 silencing increases SMURF1 while SMURF1 knockdown increases AIP1 and reverses PHD2 silencing-induced apoptosis resistance in human lung microvascular endothelial cells. Cells were transfected with control (siCTRL), PHD2 (siPHD2), SMURF1 (siSMURF1) siRNA, or both for 48 h. A and B: whole cell lysates were collected for Western blotting to assess SMURF1, AIP1, and SMAD1 expression. Protein levels were normalized to β-actin, and representative blots were shown. C: cells were further incubated for 24 h in a serum- and growth factor-free medium to measure caspase activities using the Caspase-Glo3/7 Assay. Data are presented as means ± SD. *P < 0.05 and **P ≤ 0.01 (n = 5 for all). Two-way ANOVA with post hoc t test (A and C), paired t test (B). D: schematic of HIF-independent mechanisms associated with PHD2 loss in pulmonary vascular endothelial cells. The resulting signal transduction (red arrows) and phosphorylation (red circles) events include AKT activation, AIP1 loss, ERK activation, and JNK inhibition, leading to apoptosis resistance and inflammation.