Aspirin-induced histone acetylation in endothelial cells enhances synthesis of the secreted isoform of netrin-1 thus inhibiting monocyte vascular infiltration

Abstract

Background and purpose: There are conflicting data regarding whether netrin-1 retards or accelerates atherosclerosis progression, as it can lead either to monocyte repulsion from or retention within plaques depending on its cellular source. We investigated the effect of aspirin, which is widely used in cardiovascular prophylaxis, on the synthesis of different isoforms of netrin-1 by endothelial cells under pro-inflammatory conditions, and defined the net effect of aspirin-dependent systemic modulation of netrin-1 on atherosclerosis progression.

Experimental approach: Netrin-1 synthesis was studied in vitro using human endothelial cells stimulated with TNF-α, with or without aspirin treatment. In vivo experiments were conducted in ApoE(-/-) mice fed with a high-fat diet (HFD), receiving either aspirin or clopidogrel.

Key results: TNF-α-induced NF-κB activation up-regulated the nuclear isoform of netrin-1, while simultaneously reducing secreted netrin-1. Down-regulation of the secreted isoform compromised the chemorepellent action of the endothelium against monocyte chemotaxis. Aspirin counteracted TNF-α-mediated effects on netrin-1 synthesis by endothelial cells through COX-dependent inhibition of NF-κB and concomitant histone hyperacetylation. Administration of aspirin to ApoE(-/-) mice on HFD increased blood and arterial wall levels of netrin-1 independently of its effects on platelets, accompanied by reduced plaque size and content of monocytes/macrophages, compared with untreated or clopidogrel-treated mice. In vivo blockade of netrin-1 enhanced monocyte plaque infiltration in aspirin-treated ApoE(-/-) mice.

Conclusions and implications: Aspirin counteracts down-regulation of secreted netrin-1 induced by pro-inflammatory stimuli in endothelial cells. The aspirin-dependent increase of netrin-1 in ApoE(-/-) mice exerts anti-atherogenic effects by preventing arterial accumulation of monocytes.

Figure 1

Cellular distribution of netrin-1 in HUVECs. Corrected total (A) and nuclear (B) cell fluorescence of netrin-1 [corrected total cell fluorescence (CTCF) and corrected nuclear cell fluorescence (CNCF), respectively] in HUVECs following different treatments as indicated. The ratio of CTCF to CNCF is also displayed (C). Panel (D) shows representative micrographs (20× magnification fields) of netrin-1 staining (in red) in non-permeabilized (left) and permeabilized (right) HUVECs under the different experimental conditions as specified (scale bars = 10 μm). Netrin-1 fluorescent images were merged with corresponding fluorescent nuclear staining with Hoechst (in blue). Pink colour within the nuclei in permeabilized cells derives from the overlay of netrin-1 signal (red) with nuclear staining (blue), and is indicative of nuclear netrin-1 localization. Concentration of the full-length/secreted isoform of netrin-1 was measured in cell supernatants and is displayed in (E). (F) and (G) report the levels of PGE₂ and TxB₂, respectively, in HUVEC supernatant following different treatments. n = 3–7. ASA: aspirin, 0.5 mM; SC-560, 30 nM; Indom: indomethacin, 100 μM; NS-398, 10 μM; SA: salicylic acid, 0.5 mM; TSA, 400 nM; TNF-α, 10 ng·mL^–1. *P < 0.05 versus CTR (control HUVECs not treated with TNF-α).

Figure 2

NF-κB activation and gene induction of truncated (nuclear) and full-length (secreted) isoforms of netrin-1. (A) Representative micrographs (20× magnification fields) of p65 staining under the different experimental conditions as indicated. p65 fluorescent images (in green) were merged with corresponding fluorescent nuclear staining with Hoechst (in blue). Yellow colour within the nuclei derives from the overlay of p65 signal (green) with nuclear staining (blue), and is indicative of nuclear p65 localization. p65 nuclear translocation was taken as an index of NF-κB activation and is reported in (B) as percentage of nuclei expressing p65 under the different experimental conditions. Panels (C) and (D) show accumulated data for quantitative RT-PCR for netrin-1 (NTN-1) using primers for (C) total netrin-1 (comprising both the full-length and truncated isoforms) or (D) the full-length/secreted isoform specifically, both normalized to the housekeeping gene GAPDH. The ratio of the two isoforms was obtained by normalizing the full-length isoform of netrin-1 to total netrin-1 (E). Data are reported as fold changes compared with control (untreated cells, dotted blue line). n = 3–5. Panels (F) and (G) show representative micrographs (20× magnification fields) of netrin-1 staining (in red) in permeabilized (F) and non-permeabilized (G) HUVECs either untreated (control) or TNF-α treated, either in the presence or absence of the NF-κB inhibitor Bay 11-7085 (BAY), as specified. Netrin-1 fluorescent images were merged with corresponding fluorescent nuclear staining with Hoechst (in blue). Pink colour within the nuclei in permeabilized cells (F) derives from the overlay of netrin-1 signal (red) with nuclear staining (blue), and is indicative of nuclear netrin-1 localization. In permeabilized cells (F), corrected total and nuclear cell fluorescence of netrin-1 (CTCF and CNCF, respectively) following different treatments was calculated and reported in the corresponding graphs. The ratio of CTCF to CNCF is also displayed. Secreted netrin-1 was measured by elisa in cell supernatants and results are reported in (G). n = 3. Scale bars = 10 μm. ASA: aspirin, 0.5 mM; SC-560, 30 nM; Indom: indomethacin, 100 μM; NS-398, 10 μM; TSA, 400 nM; TNF-α, 10 ng·mL⁻¹; Bay 11-7085, 5 μM. *P < 0.05 versus CTR (control HUVECs not treated with TNF-α).

Figure 3

Histone 3 acetylation and netrin-1 expression in HUVECs. Western blotting (A, B) and immunofluorescence (C) studies showing degree of histone acetylation in HUVECs in response to different treatments as specified. In the Western blots, densitometric analysis of acetylated H3 (acH3) was normalized to total H3. In immunofluorescence experiments, the corrected nuclear cell fluorescence (CNCF) was calculated (C). Panels (D)–(H) show the effect of aspirin (ASA) on HAT and HDAC activities in HUVECs. HAT (D) and HDAC (E) activities, and HDAC/HAT ratio (F), were measured in nuclear extracts isolated from HUVECs following treatments as specified. Nuclear extracts were isolated from HUVECs not stimulated with TNF-α, and HAT (G) and HDAC (H) activities were tested in the presence of either ASA or salicylic acid. n = 3. ASA, 0.5 mM; SC-560, 30 nM; SA: salicylic acid, 0.5 mM; Indom: indomethacin, 100 μM; TSA, 400 nM; TNF-α, 10 ng·mL⁻¹; Bay 11-7085 (BAY), 5 μM. *P < 0.05 versus CTR (control HUVECs not treated with TNF-α).

Figure 4

Effect of aspirin (ASA) on migration of human CD14⁺CD16⁻ monocytes. Migration of human CD14⁺CD16⁻ monocytes towards a monolayer of HAoECs (A–C) or towards an CCL2 (100 ng·mL⁻¹) gradient (D–F) is shown under different experimental conditions, as indicated in the graphs. When tested against an HAoEC monolayer, the endothelial cells, but not monocytes, were pretreated with ASA, and migration was conducted in the presence or absence of Unc5b-Fc. When tested towards a CCL2 gradient, monocytes were pretreated with ASA and migration was performed either in the presence or absence of Unc5b-Fc. (A) and (D) show cell count under light microscopy in five microscopic fields (20× magnification) on the lower surface of the membranes of the Boyden chambers, following removal of non-migrated cells on the upper surface. (B) and (E) show count, as determined by flow cytometry, of monocytes (CD14⁺ cells, gated) which have migrated to the lower well of the Boyden chamber system, with representative dot plots also displayed (C and F). n = 3. CTR: TNF-α-stimulated HAoECs (C) or untreated monocytes (F); ASA: ASA-treated HAoECs (C) or monocytes (F); ASA + Unc5b-Fc: following ASA pretreatment of either HAoECs (C) or monocytes (F). Unc5b-Fc was added during monocyte migration. Histograms show count of migrated monocytes in CTR (grey), ASA (blue) and ASA + Unc5b-Fc (red). *P < 0.05 versus ASA; †P < 0.01 versus netrin-1.

Figure 5

Aspirin (ASA) increases systemic and vascular expression of netrin-1 in ApoE^−/− mice. (A) elisa of netrin-1 in plasma, from ApoE^−/− mice on 8 week HFD, either untreated (8 weeks; n = 8) or treated with ASA (5 mg·kg⁻¹·day⁻¹; n = 8) or clopidogrel (Clop; 25 mg·kg⁻¹·day⁻¹; n = 8). Also shown in (A) are plasma netrin-1 levels in ApoE^−/− mice after 4 weeks of HFD (4 weeks; n = 4) for comparison. (B–C) The micrographs show immunofluorescence staining for netrin-1 (red), the endothelial marker CD31 (green) and nuclei by DAPI (blue) of aortas from ApoE^−/− mice after 8 week HFD either untreated (left) or ASA treated (right, ASA 5 mg·kg⁻¹·day⁻¹); magnification = 20× (scale bars = 50 μm). Bottom panels show merging of the three stainings: the white arrows indicate the luminal localization of netrin-1 (red) within the arterial wall and its co-localization with the endothelial cell marker CD31 (green), which is evident in the ASA-treated group but not in either clopidogrel-treated or untreated mice. The graph reports quantification of mean fluorescence intensity for netrin-1 and CD31 in the different groups as specified, and it is reported as fold change versus control (untreated) animals. (D) and (E) report HDAC and HAT activities measured in nuclear extracts isolated from the aortic arterial wall of ApoE^−/− mice at the end of the 8 week HFD period, either untreated (8 weeks) or treated with ASA (5 mg·kg⁻¹·day⁻¹) or clopidogrel (25 mg·kg⁻¹·day⁻¹). Also shown in (F) is the HDAC/HAT ratio in the different groups. Data for (A) are shown as median ± interquartile ranges. *P < 0.05 versus 8 weeks.

Figure 6

Effect of aspirin (ASA) and clopidogrel on monocyte/macrophage plaque content. (A, B) Representative pictures of Oil Red O(ORO) (A) and CD68 (B) staining of cryosections of the brachiocephalic artery (BA) of ApoE^−/− mice at the end of 8 week HFD in clopidogrel (25 mg·kg⁻¹·day⁻¹) or ASA-treated animals (ASA: 5 mg·kg⁻¹·day⁻¹) as indicated (magnification 20× and 40× on the left and right sides of each panel respectively). Counterstaining was with haematoxylin. Black arrows indicate the atherosclerotic lesion detected on histological section. Total and ORO-positive area of plaques under the different experimental conditions is reported in (C) (scale bars = 150 μm lower power; 50 μm higher power). (D) Total myeloid cell count [F4/80⁺ CD115⁺ Lin- (B220, CD90, Ly6G, NK1.1)], number of monocytes [F4/80^low CD115⁺ Lin- (B220, CD90, Ly6G, NK1.1)] and macrophages [F4/80^high CD115⁺ Lin- (B220, CD90, Ly6G, NK1.1)] within the BA in untreated animals (8 weeks; n = 6) and those treated with either clopidogrel (25 mg·kg⁻¹·day⁻¹; n = 6) or ASA (5 mg·kg⁻¹·day⁻¹; n = 5). Values are indicated as cell number/mg tissue. Relative prevalence of monocytes (F4/80^low cells, dark grey) and macrophages (F4/80^low cells, light grey) is reported in the pie charts and expressed as percentage of total myeloid cells [F4/80⁺ CD115⁺ Lin- (B220, CD90, Ly6G, NK1.1)]. (E) Representative flow cytometry plots showing gating strategy for monocyte/macrophage characterization in the BA of each group of animals. Briefly, total myeloid cells were identified within the Lin⁻ gate as CD115⁺ F4/80⁺ cells. Within this population, monocytes (F4/80^low; left quadrants in the dot plots indicating Ly6C vs. F4/80) and macrophages (F4/80^high and Ly6C^low cells; lower right quadrant in the dot plots indicating Ly6C vs. F4/80) were differentiated on the basis of intensity of fluorescence for F4/80 staining. Total monocytic population was further distinguished into Ly6C^high and Ly6C^low cells (bottom plots). The percentage shown in the dot plots refers to the frequency of gated cells within the parent population. In the contour plots, the relative prevalence of either monocytes (Lin⁻ CD115⁺ F4/80^low) and macrophages (Lin⁻ CD115⁺ F4/80^low Ly6C^low) or Ly6C^high and Ly6C^low monocytes is shown. (F) Ly6C^high and Ly6C^low monocyte count within the BA in the different animal groups as indicated. Values are indicated as cell number/mg tissue. Pie charts show accumulated data (8 weeks: n = 4; clopidogrel: n = 4; ASA: n = 5) reporting percentage of Ly6C^high (dark grey) and Ly6C^low (light grey) cells over total monocytes [F4/80^low CD115⁺ Lin- (B220, CD90, Ly6G, NK1.1)] measured in the BA of each group of animals. *P < 0.05 versus 8 weeks.

Figure 7

Effect of aspirin (ASA) on blood monocytosis and monocyte trafficking into atherosclerotic lesions. (A) and (B) display monocyte count (SSC^low B220⁻ CD115⁺ F4/80⁺ cells) and number of both Ly6C^high and Ly6C^low subsets in the peripheral blood of ApoE^−/− mice on HFD at different time points as indicated, either with or without concomitant treatment with ASA (5 mg·kg⁻¹·day⁻¹) or clopidogrel (25 mg·kg⁻¹·day⁻¹). n = 4–8. (C) and (D) show the prevalence of total monocytes (reported as percentage over SSC^low B220⁻ mononuclear cells) in the blood of the different animal groups as indicated, and the prevalence of circulating Ly6C^high (light grey) and Ly6C^low cells (dark grey) is reported in the pie charts in (F) (calculated as percentage of total monocytes). (E) Degree of platelet activation as measured by percentage of P-selectin-expressing platelets in the circulation under the different experimental conditions. *P < 0.05 versus 8 weeks. (G) Autoradiographic exposure of the brachiocephalic arteries from ApoE^−/− mice on HFD and ASA treatment (5 mg·kg⁻¹·day⁻¹) for 8 weeks, excised 36 h after adoptive transfer of ¹¹¹In-labelled monocytes, either injected with Un5cb-Fc antibody (left) or isotype control (right) 1 h before receiving labelled monocytes. Exposure of tissue plates was for 12 h. Dark spots in the grey-scale images (left) and red spots in the colour images (right), as indicated by the black arrows, indicate cell accumulation and localization within the brachiocephalic artery. The corresponding signal was quantified and reported as relative recruitment of cells (labelled monocytes injected into animals receiving antibody isotype control scaled to 100%). n = 3 per group. *P < 0.05 versus isotype control. (H) Representative flow cytometry dot plots indicating the gating strategy and analysis for blood monocyte characterization in the different animal groups as indicated. Briefly, B220 expression was evaluated within the mononuclear blood population only. B220⁺ cells were excluded and further analysis was performed on the population of B220⁻ cells (prevalence of these within the mononuclear cell population is indicated). Within the SSC^low B220⁻ mononuclear cell gate, the expression of CD115 and F4/80 was studied to identify CD115⁺F4/80⁺ events representative of the total monocytic population. Expression of Ly6C within the SSC^low B220⁻ CD115⁺ F4/80⁺ cells was carried out to calculate the percentage and number of Ly6C^high and Ly6C^low monocytes (upper and lower right quadrants, respectively, in the bottom plots).