Neuropilin-1 deficiency in vascular smooth muscle cells is associated with hereditary hemorrhagic telangiectasia arteriovenous malformations

Abstract

Patients with hereditary hemorrhagic telangiectasia (HHT) have arteriovenous malformations (AVMs) with genetic mutations involving the activin-A receptor like type 1 (ACVRL1 or ALK1) and endoglin (ENG). Recent studies have shown that Neuropilin-1 (NRP-1) inhibits ALK1. We investigated the expression of NRP-1 in livers of patients with HHT and found that there was a significant reduction in NRP-1 in perivascular smooth muscle cells (SMCs). We used Nrp1SM22KO mice (Nrp1 was ablated in SMCs) and found hemorrhage, increased immune cell infiltration with a decrease in SMCs, and pericyte lining in lungs and liver in adult mice. Histologic examination revealed lung arteriovenous fistulas (AVFs) with enlarged liver vessels. Evaluation of the retina vessels at P5 from Nrp1SM22KO mice demonstrated dilated capillaries with a reduction of pericytes. In inflow artery of surgical AVFs from the Nrp1SM22KO versus WT mice, there was a significant decrease in Tgfb1, Eng, and Alk1 expression and phosphorylated SMAD1/5/8 (pSMAD1/5/8), with an increase in apoptosis. TGF-β1-stimulated aortic SMCs from Nrp1SM22KO versus WT mice have decreased pSMAD1/5/8 and increased apoptosis. Coimmunoprecipitation experiments revealed that NRP-1 interacts with ALK1 and ENG in SMCs. In summary, NRP-1 deletion in SMCs leads to reduced ALK1, ENG, and pSMAD1/5/8 signaling and reduced cell death associated with AVM formation.

Keywords: Apoptosis; Cardiovascular disease; Genetic diseases; Vascular Biology.

Figure 1. NRP-1 expression in liver sections from patients with HHT and controls.

(A) Coimmunostaining for NRP-1 (green), α-SMA staining (red), and DAPI-stained (blue) nucleus in representative images. The part of the images enclosed with boxes in the left panel were digitally enlarged and shown as right panel. Arrows indicate cells that are positive for both α-SMA and NRP-1. Multiple vacuoles with minimal α-SMA⁺ cells can be found in HHT patient liver sections (lower panel). All images were captured at 10× magnification using a Zeiss Axio imager M2 equipped with an Axiocam 503 camera. Scale bar: 50 μm. (B) The overlapped green and red stain intensity index was determined using Fiji-ImageJ software with JACoP plugin. Data are shown as mean ± SEM of n = 6. Nonparametric Mann-Whitney U test was performed. *P < 0.05.

Figure 2. Direct connection of artery to vein in the lung and dilated vessels in lung and liver of Nrp1^SM22KO mice.

(A) Lung tissue sections were stained for H&E, which shows heterogenic thickness of the vessel (open and closed arrows) in Nrp1^SM22KO animals. (B) Verhoeff–Van Gieson–stained sections from Nrp1^fl/fl (WT) and Nrp1^fl/fl/SM22α^Cre+ (Nrp1^SM22KO) mice were used to distinguish the artery (closed arrow) and vein (open arrow). Ar, artery; V, vein; AW, airway in lungs. Scale bar: 50 μm. (C–F) vascular cross-sectional area was determined in CD31-stained lung (C and D) liver tissue sections (E and F). All images were captured at 10× magnification using a Zeiss Axio imager M2 equipped with an Axiocam 503 camera and a motorized stage. Scale bar: 500 μm. The area enclosed with the dotted line box was digitally enlarged to show as inset panel. The solid black arrows indicate CD31⁺ cells lining the blood vessels. Vessels were identified manually and shaded in green, as shown in right panels. The average vessel area in lung (C) and liver (E) was measured using Fiji-ImageJ software, and the average cross-sectional area in lung (D) and in liver (F) were shown as bar graphs. Data are shown as mean ± SEM of n = 5. There is a significant increase in average cross-sectional area of vessels in lung (D) and liver (E) from Nrp1^SM22KO mice compared with WT animals. Nonparametric Mann-Whitney U test was performed. *P < 0.05.

Figure 3. Augmented immune cell infiltration in lung and liver of Nrp1^SM22KO mice.

(A and B) H&E staining of lung (A) and liver (B) tissue sections from Nrp1^fl/fl (WT) and Nrp1^fl/fl/SM22α^Cre+ (Nrp1^SM22KO) mice. All images were captured at 10× magnification using a Zeiss Axio imager M2 equipped with an Axiocam 503 camera and a motorized stage. Arrows indicate infiltrated cells in the interstitial compartments of lung and liver tissue. The areas enclosed by boxes in representative images are digitally enlarged and shown as separate panels. (C) Coimmunostaining for CD68 (green) and isolectin-B4 (red)staining and DAPI-stained (blue) nuclei. The yellow arrows point to cells positive for both CD68 and isolectin-B4, and white arrows point to isolectin positive endothelium. (D) Fluorescence intensity quantification of CD68⁺ green stain in C. (E, G, and I) Chromogen stain for CD68 in liver (E) and CD45 in lung (G) and liver (I). The black arrows point to brown cells positive for CD68 in liver (E) and CD45 in lung (G) and liver (I). Scale bar: 50 μm. The brown stain intensity was measured using Zen Pro image analysis software (Zeiss). (F, H, and J) The percentage of stained area over total tissue area was presented as stain index of CD68 in liver (F) and CD45 index in lung (H) and liver (J). Data are shown as mean ± SEM of n = 5. Nonparametric Mann-Whitney U test was performed. *P < 0.05.

Figure 4. Nrp1^SM22KO mice have decreased smooth muscle cells and pericytes in adult lungs.

Coimmunostaining for (A–C) α-SMA (green), NRP-1(red), and nuclei (blue) stained with DAPI. (D) Chromogen staining for NG2⁺ pericyte lining in lung vasculature was performed. White arrows point to α-SMA⁺ lung vasculature in A and black arrows in E point to NG2⁺ brown pericytes. All images were captured at 10× magnification using a Zeiss Axio imager M2 equipped with an Axiocam 503 camera and a motorized stage. Scale bar: 50 μm. Fluorescence intensity quantification of red NRP-1⁺ (B) and green α-SMA⁺ (C) present in A. (E) NG2⁺ brown chromogen stain intensity in D was measured using Zen Pro image analysis software (Zeiss) and presented as bar graphs. There is a minimum green α-SMA and brown NG2 staining in the lung alveoli of Nrp1^SM22KO mice compared with WT mice. Data are shown as mean ± SEM of n = 4 or 5 animals. Nonparametric Mann-Whitney U test was performed. *P < 0.05.

Figure 5. Smooth muscle cell Nrp1 deletion results in increase vascular diameter and density with a decrease in pericyte lining in the retinal vasculature.

(A) P5 retinas from Nrp1^fl/fl (WT) and Nrp1^fl/fl/SM22α^Cre+ (Nrp1^SM22KO) mouse pups were stained for red isolectin-B4 (IsoB4) and NG2 for pericytes. The areas enclosed by boxes in representative images are digitally enlarged and shown as separate panels. Scale bar: 50 μm. (B) Vascular diameter was measured using Zen Pro software. Data are shown as mean diameter of artery (a), vein (v), and capillaries (cp). (C and D) Fluorescence intensity quantification of red IsoB4⁺ vasculature and green NG2⁺ (D) cells was performed using NIH ImageJ software. All images were captured at 10× objective and digitally enlarged to show as separate panels. Data are shown as mean ± SEM of n = 4–7 retinas. Nonparametric Mann-Whitney U test was performed. *P < 0.05.

Figure 6. Flow-induced Nrp1 upregulation in the AVF in the inflow arteries.

(A) NRP-1 gene expression was assessed by qPCR in the AVF inflow arteries (GA) and contralateral carotid arteries (CA) at 3 days after AVF creation. There is a significant increase in NRP-1 gene expression in GA compared with CA from WT mice. However, there was no significant difference in NRP-1 expression in GA compared with CA in Nrp1^SM22KO mice. Immunostaining was performed to assess NRP-1 and α-SMA levels in GA and CA at 14 days after arteriovenous fistula (AVF) creation. (B and D) Representative images of NRP-1 (B) and α-SMA (D) in GA and CA from Nrp1^fl/fl (WT) and Nrp1^fl/fl/SM22α^Cre+ (Nrp1^SM22KO) mice. The dotted line in B indicates the media and adventitia layers in the vessel wall. All images were captured using 10× magnification. Scale bar: 50 μm. The arrows point to brown cells positive for NRP-1 (B) and α-SMA (D) staining in the vessel wall. (B, C, and E) The intensity of brown stain positive NRP-1 (C) index in α-SMA (B and E) index in C were quantified as described and presented as mean ± SEM of n = 5–7 animals. A 2-way ANOVA with multiple comparison was performed. *P < 0.05.

Figure 7. Nrp1 deletion in smooth muscle cell reduces gene expression of Eng, Alk1, and Tgfb1 but not Tnfa in the AVF inflow arteries.

Gene expression was assessed by qPCR in AVF inflow artery (GA) and contralateral carotid artery (CA) at 3 days after arteriovenous fistula (AVF) creation. (A–C) There was a significant increase in Eng (A), Alk1 (B), and Tgf-β1 (C) expression in GA compared with CA from Nrp1^fl/fl (WT) mice but not in Nrp1^fl/fl/SM22α^Cre+ (Nrp1^SM22KO) mice. (D–F) There was a significant increase in Bmp9 (D), Tnfa (E), and SMAD8/9 (F) expression in GA compared with CA in both WT and Nrp1^SM22KO mice. (G and H) However, there was no significant difference in (G) SMAD6 and (H) SMAD7 in CA compared with GA, regardless of Nrp1 deletion. The data were normalized to the gene expression in the CA of WT mouse and expressed as mean fold change ± SEM in GA and CA from WT and Nrp1^SM22KO of n = 6 animals. Two-way ANOVA was performed. *P < 0.05

Figure 8. Smooth muscle cell Nrp1 deletion attenuates flow induced ENG, ALK1, and TGF-β1 but not TNF-α levels in AVF inflow arteries.

Immunostaining was performed to assess ENG, ALK1, TGF-β1, and TNF-α levels in the AVF inflow arteries (GA) and contralateral carotid arteries (CA) at 14 days after arteriovenous fistula (AVF) creation. (A, C, E, and G) Representative images of (A) ENG, (C) ALK1, (E) TGF-β1, and (G) TNF-α in GA and CA from NRP1^fl/fl (WT) and Nrp1^fl/fl/SM22α^Cre+ (Nrp1^SM22KO) mice. All images were captured using 10× magnification. Scale bar: 50 μm. The arrows show the brown positive stain of ENG (A), ALK1 (C), TGF-β1 (E), and TNF-α (G). (B, D, F, and H) The intensity of brown stain positive for ENG (B), ALK1 (D), TGF-β1 (F), and TNF-α (H) index were quantified as described and presented as mean ± SEM of n = 5–7 animals. Two-way ANOVA was performed. *P < 0.05.

Figure 9. Nrp1 deletion in smooth muscle cells results in a reduction of TGF-β1–pSMAD1/5/8 signaling.

(A) Aortic smooth muscle cells were isolated from Nrp1^fl/fl (WT) or Nrp1^fl/fl/SM22α^Cre+ (Nrp1^SM22KO) mice. After overnight serum starvation, cells were stimulated with 10 ng/mL TGF-β1 for 30 minutes, and pSMAD1/5/8 levels were assessed by Western blot. Representative Western blot shows TGF-β1 stimulation (TGF-β1) increased pSMAD1/5/8 compared with unstimulated control (UNS) cells from WT mice but not in cells from Nrp1^SM22KO mice with no change in pSMADs 2 and 3. (B and C) Densitometric analysis of NRP-1 (B) and pSMAD1/5/8 (C) were performed using NIH ImageJ software. Data are shown as mean ± SEM of n = 3. (D–G) Immunostaining analysis of AVF inflow arteries (GA) and contralateral control arteries (CA) at 14 days after AVF creation. All images were captured using 10× magnification. Scale bar: 50 μm. Arrows indicate the brown stained nuclei of pSMAD1/5/8 (D), pSMAD2 (F), and pSMAD3 (G). The intensity of brown stain pSMAD1/5/8 (E), pSMAD2 (H), and pSMAD3 (I) was measured as described and presented as mean ± SEM of n = 5–8 animals. Two-way ANOVA was performed. *P < 0.05.

Figure 10. Smooth muscle cell Nrp1 deletion causes an increase in apoptosis of smooth muscle cells.

Aortic smooth muscle cells (SMCs) isolated from Nrp1^fl/fl (WT) or Nrp1^fl/fl/SM22α^Cre+ (Nrp1^SM22KO) mice. (A) Caspase-3/7 activity was assessed in SMCs as described to assess apoptosis of SMCs in vitro. (B) Representative images of TUNEL staining on tissue sections of the inflow artery (GA) and contralateral carotid artery (CA) at 14 days after AVF creation. All images were captured using 10× magnification. Scale bar: 20 μm. Arrows show brown staining TUNEL⁺ cells. (C) The intensity of the brown TUNEL stain was measured as described. After overnight serum starvation, SMCs were stimulated with 10 ng/mL TGF-β1 for 24 hours. (D) Proliferation was assessed in vitro as described. There was 18% increase in proliferation (P = 0.01) of SMCs isolated from Nrp1^SM22KO mice compared with cells isolated from WT sex-matched littermates. (E) Representative images of Ki-67 staining on tissue sections of GA from AVF and CA at 14 days after AVF creation. All images were captured using 10× magnification. Scale bar: 20 μm. Arrows show brown staining of KI-67⁺ cells. (F) The intensity of the brown Ki-67 stain was measured as described. Data are shown as mean ± SEM of n = 8 in vitro experiments or n = 6 animals. Nonparametric Mann-Whitney U test (A and D) or 2-way ANOVA (C and F) was performed. *P < 0.05.

Figure 11. Biochemical association of NRP-1 with ALK1 and ENG.

HEK293T cells (Ctr) were transfected with NRP-1 and FLAG-tagged constructs of ALK1 (ALK1), Endoglin (ENG), and ALK1 + ENG. (A and B) Cell lysates were subjected to immunoprecipitation with anti–NRP-1 antibody (IP: NRP-1) and probed with anti-FLAG Western blot (Blot: FLAG) (A) and FLAG (IP:FLAG) and probed with anti–NRP-1 (Blot: NRP-1) Western blot (B). Both ENG and ALK1 were detected in NRP-1 immunoprecipitate (A) and NRP-1 detected in immunoprecipitates of both FLAG-ALK1 and FLAG-ENG (B). (C) ALK1 and ENG were also detected in NRP-1 immunoprecipitate from cell lysates of TGF-β1–stimulated aortic SMCs. TGF-β1 stimulation to the SMCs did not affect the biochemical association of NRP-1 with ALK1 or ENG. (D) Total protein levels of NRP-1, ALK1, and ENG in SMCs not changed with TGF-β1 stimulation. All experiments were repeated 3 times, and representative immunoblots are shown.