Redox regulation of lung endothelial PERK, unfolded protein response (UPR) and proliferation via NOX1: Targeted inhibition as a potential therapy for PAH

Abstract

Aims: Reactive oxygen species (ROS) play an important role in the pathogenesis of pulmonary arterial hypertension (PAH) and NADPH oxidases (NOXs) as sources of ROS are implicated in the development of the disease. We previously showed that NOX isozyme 1 (NOX1)-derived ROS contributes to pulmonary vascular endothelial cell (EC) proliferation in response to PAH triggers in vitro. However, whether and how NOX1 is involved in PAH in vivo have not been explored nor has NOX1 been examined as a viable and effective therapeutic disease target.

Methods and results: Herein, infusion of mice exposed to Sugen/hypoxia (10 % O₂) with a specific NOX1 inhibitor, NOXA1ds, delivered via osmotic minipumps (i.p.), significantly suppressed pathological changes in hemodynamic parameters characteristic of PAH. Furthermore, lungs of human patients with idiopathic PAH (iPAH) and exploratory RNA-seq analysis of hypoxic human pulmonary ECs, in which NOX1 was suppressed, were probed. The findings showed a clear indication of NOX1 in the promotion of both protein disulfide isomerase (PDI) and the unfolded protein response (UPR; in particular, the PERK arm of the pathway including eIF2α and ATF4) leading to proliferation. In aggregate, these results are consistent with a causal role for NOX1 in the development of mouse and human PAH and reveal a novel and mechanistic pathway by which NOX1 activates the UPR response during EC proliferation.

Conclusion: NOX1 promotes phenotypic changes in ECs that are pivotal to proliferation and PAH through activation of the UPR. Taken together, our results are consistent with selective inhibition of NOX1 as a novel modality for attenuating PAH.

Keywords: Hypoxia; NADPH oxidases; NOX inhibitors; NOXA1ds; Pulmonary arterial hypertension; Unfolded protein response.

Graphical abstract

Fig. 1

Selective inhibition of NOX1 ameliorates right ventricle hemodynamic parameters in experimental model of pulmonary arterial hypertension. (A–N) C57BL/6J (WT) mice were intraperitoneally (i.p.) implanted with osmotic pumps delivering 20 mg/kg/day of either scrambled sequence (Scr, control) or specific NOX1 inhibitor (NOXA1ds). Mice were exposed to either hypoxia (Hx, 10 % O₂) or normoxia (Nx, 21 % O₂) for 3 wks. Hx mice received an injection of Sugen 5416 (20 mg/kg/day, IP) at days 0, 7, 14 of Hx exposure whereas Nx mice received vehicle (A). (B–H) Values for right ventricle (RV) hemodynamic parameters: representative pressure-volume (P/V) loops (B) and individual value plots for the RV max pressure (RVMP) (C), mean pulmonary artery pressure (mPAP) (D), pulmonary vascular resistance (PVR Woods) (E), maximal rate of RV pressure rise during systole (contraction, dP/dt max) (F), maximal rate of RV pressure decline during diastole (relaxation, dP/dt min) (G), and stroke work (H). Values for the left ventricle (LV) hemodynamic parameters (I–N): representative pressure-volume (P/V) loops (I), and individual value plots for the LV max pressure (LVMP) (J), contractile index (LVCI) (K), maximal rate of LV pressure rise during systole (contraction, dP/dt max) (L), maximal rate of RV pressure decline during diastole (relaxation, dP/dt min) (M) and aortic mean pressure (AOMP) (N). All data were analyzed using a two-way ANOVA followed by a posthoc Tukey's test (C–H, J–N: n = 7–9 mice; ns = non-significant; ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001∗ vs. Nx/Scr; #p < 0.05, ##p < 0.01 vs. Su/Hx).

Fig. 2

Selective inhibition of NOX1 attenuates pulmonary vascular remodeling and fibrosis following chronic hypoxic exposure. (A–D) Mice were treated as in experiments described in Fig. 1. (A/B) Formalin-fixed paraffin embedded (FFPE) lung sections were stained with hematoxylin & eosin (H&E) for the assessment of histomorphology. As a measure of pulmonary vascular remodeling (PVR), medial cross-sectional area (CSA) plus neointima CSA, or total vessel wall thickening (% change) was measured in at least five visible fields containing at least one randomly selected vessel (40–80 μm diameter) per lung section (A/B). FFPE lung sections were stained with Picrosirius Red (PSR) for the assessment of pulmonary fibrosis (C/D). Lung vessel collagen deposition and, thus, fibrosis was determined using color-based thresholding. The total area of lung vessel fibrosis was calculated as a percentage of the total vessel cross-sectional area and was denoted as a change in “Lung vessel Fibrosis (PSR%)” from Nx control. For each section, 5 visible fields were captured with at least one randomly selected lung vessel (40–80 μm diameter) (C/D). Representative figures are shown wherein ∗ indicates a bar length of 50 μm while # indicates a bar length of 200 μm. All data were analyzed by two-way ANOVA followed by Tukey's test (A–D: n = 4mice: ∗p < 0.05, ∗∗∗∗p < 0.0001; vs. Nx/Scr; #p < 0.05, ###p < 0.001 vs. Su/Hx).

Fig. 3

Hypoxia upregulates and activates canonical NOX1 oxidase in lung endothelium. (A–L) Mice were treated as described in Fig. 1. Western blot analyses were performed on lung (A/B) tissue lysates using monoclonal HIF1α antibody. β-actin served as loading control. Representative blots are shown. (C/D) Freshly explanted and homogenized lung tissues were used for the detection of H₂O₂ by Amplex Red (C) and coumarin boronate (D) assays. (E–G) Formalin-fixed, paraffin-embedded lung tissue sections were incubated with antibodies against 8-hydroxy-2-deoxyguanosine (8-OHdG) (E/F), NOX1 (E/G), or CD31 (E) and visualized with a secondary antibody conjugated with Alexa Fluor 488, 560–590, or 647. Nuclei were visualized with DAPI. Fluorescence intensity was measured in five visible fields with at least one randomly selected lung vessel (40–80 μm diameter) per lung section. EC NOX1 and EC 8-OHdG expression were quantified as the average mean intensity of NOX1-or 8-OHdG-associated fluorescence, respectively, where colocalized with CD31-associated fluorescence (E–G). Representative images are shown wherein ∗ indicates a bar length of 50 μm. (H–L) Western blot of lung tissue lysates show NOX1 (H/I), NOXO1 (H/J), NOXA1 (H/K), and p22^phox(H/L). β-actin served as loading control. NOXA1ds = N1ds. Representative blots are shown. All data were analyzed by two-way ANOVA followed by Tukey's test (A-D, H–L: n = 7–9 mice, E–G: n = 4 mice; ns = non-significant; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 vs. Nx/Scr; #p < 0.05, ##p < 0.01, ####p < 0.0001 vs. Su/Hx).

Fig. 4

RNA sequencing shows NOX1 interacts with hypoxia treatment on genes involved in the unfolded protein response (A–G). Human pulmonary artery endothelial cells (HPAECs) were either silenced for NOX1 (siNOX1) using RNAi or were transfected with scrambled RNA (Scr) and exposed to hypoxia (Hx, 1 % oxygen) or normoxia (Nx, 21 % oxygen) for 24 h. Thereafter, total RNA was isolated, and RNA-Seq transcriptomics profiling was performed. (A) The top 2000 most variable genes, ranked by standard deviations, were used for principal component analysis (PCA). (B) Differential gene analysis (DEG) identified a large number of genes significantly (adjusted p-value [padj]< 0.05) altered by hypoxia treatment alone (2019 up, 1609 down, Hx/Scr vs Nx/Scr). (C) The differentially expressed genes (DEGs, adjusted p-value < 0.05) between Hx/Scr and Nx/Scr were used in an enrichment analysis with Gene Ontology biological processes, and the top 10 pathways ranked by false detection rate (FDR) were included in the dot plot. (D) The chord diagram shows the expression levels of genes (top 100 ranked by padj) involved in these pathways. The color bar indicates gene expression levels in log2 fold changes (Log2FC). (C) and (D) share the same color annotations and order for the pathways. (E) Genes showing significant interaction effects (p-value < 0.05) of Hx and NOX1 silencing were highlighted in the volcano plot (1214 up, 1517 down). The top 10 significantly enriched Gene Ontology biological processes and the corresponding gene expressions were shown in a dot plot (F) and a chord diagram (G), which share the same pathway order and color annotations. The color scale for interaction in (G) represents the log2FC of Hx/siNOX1 over the additive effects of Hx/Scr and Nx/siNOX1. (A–G: n = 3 biological replicates of cells)

Fig. 5

NOX1-derived ROS induce the unfolded protein response in murine lungs following hypoxic exposure. (A–H) Lung lysates of mice (treated as in experiments described for Fig. 1, Fig. 2, Fig. 3) were evaluated by Western blot using antibodies against: phospho- and total PERK (A/D), phospho- and total eIF2α (B/E), ATF4 (C/F), PDI (C/G), and BiP (C/H). β-actin or total protein levels (t-PERK or t-eIF2α) served as loading control or for activity comparison to total protein amounts. NOXA1ds = N1ds. Representative blots are shown. All data were analyzed using a two-way ANOVA followed by Tukey's test (A–H: n = 7–9 mice); ∗p < 0.05, ∗∗p < 0.01 vs. Scr/Nx; #p < 0.05, ###p < 0.001 vs. Su/Hx). N1ds = NOXA1ds.

Fig. 6

NOX1-derived ROS induce the unfolded protein response in HPAECs in vitro. (A–J) Human pulmonary artery endothelial cells (HPAECs) were treated with NOX1 inhibitor (NOXA1ds, N1ds) or control peptide (Scr) and exposed to hypoxia (Hx, 1 % oxygen) or normoxia (Nx, 21 % oxygen) for 24 h. Western blot analyses were performed on cell lysates using antibodies against: HIF1α (A/D), phospho- and total PERK (B/E), phospho- and total eIF2α (C/F), ATF4 (G/H), PDI (G/I), and BiP (G/J). β-actin or total protein levels (t-PERK/t-eIF2α) served as loading control or for activity comparison to total protein amounts. Representative blots are shown. (K/L) HPAECs were seeded in 6-well plates and either silenced for ATF4 using RNAi or transfected with Scr RNA. Thereafter at 100 % confluency, a void was made across the well and demarcated and HPAECS were exposed to either Hx or Nx for 24 h (K/L). Alternatively, HPAECs treated with either with a selective inhibitor of eIF2α dephosphorylation (ISRIB, 200 nM) or vehicle control (DMSO, Ctr) exposed to either Hx or Nx for 24 h (N/O). Photos show cellular front line after 24 h compared with starting cellular front (black line) (K/L, N/O). Representative images are shown wherein ∗ indicates a bar length of 100 μm (K/L). Data are shown as relative gap size reduction (wound closure) after 24 h (L/O) Alternatively, HPAECs were silenced for ATF4 (M) or inhibited with eIF2α (P) (as above) and exposed to Hx or Nx for 24 h. Thereafter, HPAECs were assessed for proliferation via crystal violet assay (M/P). All data were analyzed with two-way ANOVA followed by Tukey's test (A–P: n = 7–10 biological replicates of cells; ns = non-significant; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 vs. Nx/Scr; #p < 0.05, ##p < 0.01, ###p < 0.001, ####p < 0.0001 vs. Hx/Scr).

Fig. 7

Human iPAH lungs exhibit overexpression of NOX1 and connection with the unfolded protein response. (A/B, E-N) Western blot of iPAH or CTR lungs was conducted using antibodies against: HIF1α (A/B), NOX1 (E/F), phospho- and total PERK (G/H), phospho- and total eIF2α (I/J), ATF4 (K/L), PDI (K/M), and BiP (K/N). β-actin or total protein levels (t-PERK or t-eIF2α) served as control and representative blots are shown. (C/D) Amplex Red (C) and coumarin boronate (D) assays were employed for H₂O₂ detection. All data were analyzed with a two-way ANOVA followed by Tukey's test (A–N: n = 3 biological replicates, human lung samples; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, vs. Ctr. For Amplex Red measurement of H₂O₂ comparisons p = 0.0657 vs. Ctr; for BiP comparisons p = 0.0891 vs. Ctr.).