Targeting Ferroptosis as a Novel Approach to Alleviate Aortic Dissection

Abstract

A variety of programmed cell death types have been shown to participate in the loss of smooth muscle cells (SMCs) during the development of aortic dissection (AD), but it is still largely unclear whether ferroptosis is involved in the development of AD. In the present study, we found that the expression of key ferroptosis regulatory proteins, solute carrier family 7 member 11 (SLC7A11), ferroptosis suppressor protein 1 (FSP1) and glutathione peroxidase 4 (GPX4) were downregulated in aortas of Stanford type A AD (TAAD) patients, and liproxstatin-1, a specific inhibitor of ferroptosis, obviously abolished the β-aminopropionitrile (BAPN)-induced development and rupture of AD in mice. Furthermore, the expression of methyltransferase-like 3 (METTL3), a major methyltransferase of RNA m⁶A, was remarkably upregulated in the aortas of TAAD patients, and the protein levels of METTL3 were negatively correlated with SLC7A11 and FSP1 levels in human aortas. Overexpression of METTL3 in human aortic SMCs (HASMCs) inhibited, while METTL3 knockdown promoted SLC7A11 and FSP1 expression. More importantly, overexpression of METTL3 facilitated imidazole ketone erastin- and cystine deprivation-induced ferroptosis, while knockdown of METTL3 repressed ferroptosis of HASMCs. Overexpression of either SLC7A11 or FSP1 largely abrogated the effect of METTL3 on HASMC ferroptosis. Therefore, we have revealed that ferroptosis is a critical cause of AD in both humans and mice and that METTL3 promotes ferroptosis of HASMCs by inhibiting the expression of SLC7A11 and FSP1. Thus, targeting ferroptosis or m⁶A RNA methylation is a potential novel strategy for the treatment of AD.

Keywords: Aortic dissection; FSP1/AIFM2; Ferroptosis; Liproxstatin-1; METTL3; SLC7A11.

Figure 1

Ferroptosis was involved in the development of AD in human. (A) Immunohistochemical staining of TFR and HMOX1 in aortas of patients with non-AD and TAAD. (B and C) Quantitative of the average optical density of TFR (B) (n=29 non-AD and n=59 TAAD) and HMOX1 (C) (n=28 non-AD and n=58 TAAD) in (A). (D) Representative western blots of SLC7A11, FSP1 and GPX4 in the aortas of non-AD and TAAD patients (n=31 non-AD and n=65 TAAD). (E-G) Quantitative results of SLC7A11 (E), FSP1 (F), and GPX4 (G) expression in the aorta. β-Actin served as the loading control.

Figure 2

Liproxstatin-1, an inhibitor of ferroptosis, largely abolished BAPN-induced AD development. (A) Flow chart of the animal experiments. (B) Representative images of excised aortas of the indicated groups (n=12 per group). (C) The overall incidences of AD were significantly lower in mice challenged with BAPN+liproxstatin-1 (AAD occurred in 7 (4 ruptured) out of 12 mice) than in mice challenged with BAPN alone (AAD occurred in 11 (8 ruptured) out of 12 mice) (n=12 per group). (D) Representative hematoxylin and eosin (HE) staining and elastin Verhoeff-van Gieson (EVG) staining of aortas (n=5 control; n=12 BAPN; n=4 BAPN+Liproxstatin-1). (E) Quantification of aortic elastic fiber fragmentation based on EVG staining in the indicated groups (n=5 control; n=12 BAPN; n=4 BAPN+Liproxstatin-1). (F) The mean aortic diameters of various aortic segments were measured based on ultrasonography images; Asc indicates ascending aorta; Arch indicates aortic arch; Desc indicates descending aorta (n=3 control; n=6 BAPN; n=7 BAPN+Liproxstatin-1). (G-J) Representative immunohistochemical staining of Hmox1, Tfr and 4-HNE (G); and their average optical density in aortic tissues of mice treated with or without BAPN or liproxstatin-1 (H-J) (n=5-6 control; n=3-7 BAPN; n=4-6 BAPN+Liproxstatin-1). (K) The serum ferrous iron levels in the indicated groups (n=9 control; n=6 BAPN; n=10 BAPN+Liproxstatin-1). *p<0.05 vs control; #p<0.05 vs BAPN.

Figure 3

METTL3 was remarkably upregulated in the aortas of TAAD patients. (A) Representative hematoxylin and eosin (HE) staining and elastin Verhoeff-van Gieson (EVG) staining of aortic tissues of non-AD (n=27) and TAAD (n=59) patients. *p<0.05 vs non-AD. (B) Quantification of aortic elastic fiber fragmentation in non-AD and TAAD patients according to the EVG staining in (A). (C and D). Representative immunohistochemical staining of METTL3 in an aortic tissue microarray including tissues from 30 non-AD and 60 TAAD patients (C). Percentages of METTL3 positive cells in the aortas of non-AD (n=30) and TAAD (n=60) patients (D). (E and F) The protein level of METTL3 in the aorta was evaluated by using western blot in 31 non-AD and 65 TAAD patients; (E) Representative western blots of METTL3; (F) Quantitative results for the METTL3 protein levels in (E). β-Actin served as the loading control. (G). Correlation of SLC7A11, FSP1, GPX4 and METTL3 protein expression in the aortas of human with or without TAAD (n=31 non-AD and n=65 TAAD patients).

Figure 4

METTL3 suppressed SLC7A11 and FSP1 expression via promoting their mRNA degradation. (A) Representative immunohistochemical staining of Mettl3, Slc7a11, Fsp1 and Gpx4 in aortic tissues of mice treated with or without BAPN (n=6 control, n=3 BAPN). (B and C) The protein levels of METTL3 (Flag), SLC7A11, SLC3A2, FSP1, and GPX4 were detected by using western blot analysis in HASMCs infected with Lenti-Flag or Lenti-METTL3-Flag (B) or infected with Lenti-pLKO or Lenti-shMETTL3 (C) (n=4 per group). #1 and #2 indicate two different target sequences for knockdown of METTL3. The quantitative results are displayed. β-Actin served as the loading control. *p<0.05 vs Lenti-Flag or Lenti-pLKO control. (D and E) The fraction of FSP1 (D) or SLC7A11 (E) mRNA remaining in HASMCs with METTL3 overexpression or not after treated with 5 µg/mL Actinomycin D (ACTD) for indicated times (n=4 per group).

Figure 5

METTL3 facilitated cystine deprivation- and IKE-induced ferroptosis of HASMCs. (A and B) Content of ferrous ions (Fe²⁺) in HASMCs with METTL3 overexpression (A) or knockdown (B) treated with cystine deprivation or IKE (n=3 per group). (C-F) Cell viability was evaluated by using CCK-8 kit in HASMCs infected with Lenti-Flag, Lenti-METTL3-Flag, Lenti-pLKO or Lenti-shMETTL3 and treated with or without cystine deprivation or IKE (n=4 per group). (G and H) Cellular injury was measured by using an LDH kit in HASMCs infected with Lenti-Flag, Lenti-METTL3-Flag, Lenti-pLKO or Lenti-shMETTL3 under cystine deprivation or IKE treatment (n=4 per group). (I-L) The representative western blots of TFR, HMOX1, and PTGS2, and their quantitative results in HASMCs with METTL3 overexpression (I and J) or knockdown (K and L) (n=4 per group). β-Actin served as the loading control. *p<0.05 vs Lenti-Flag or Lenti-pLKO control; # p<0.05 vs Lenti-Flag or Lenti-pLKO with cystine deprivation; &p<0.05 vs Lenti-Flag or Lenti-pLKO with IKE treatment.

Figure 6

METTL3 promoted cystine deprivation- and IKE-induced lipid peroxidation in HASMCs. (A and B) The level of lipid ROS (oxidized BODIPY-C11 (green)/non-oxidized BODIPY-C11 (red) ratio) detected by using BODIPY-C11 kit in HASMCs with METTL3 overexpression (A) or knockdown (B) which treated with cystine deprivation or IKE (n=4 per group). (C and D) Lipid peroxidation was evaluated by using an MDA assay kit in HASMCs with the indicated treatments (n=3 per group). (E-H) Immunofluorescence staining of 4-HNE in HASMCs with METTL3 overexpression (E) or knockdown (G) which treated with cystine deprivation or IKE, and quantitative results are displayed in (F) and (H) (n=3 per group). *p<0.05 vs Lenti-Flag or Lenti-pLKO control; # p<0.05 vs Lenti-Flag or Lenti-pLKO with cystine deprivation; &p<0.05 vs Lenti-Flag or Lenti-pLKO with IKE treatment.

Figure 7

SLC7A11 and FSP1 overexpression largely abrogated the effects of METTL3 on ferroptosis in HASMCs. (A and B) The protein levels of SLC7A11, SLC3A2, FSP1, and GPX4 were detected by using western blot analysis in HASMCs infected with Lenti-Flag or Lenti-METTL3-Flag (A) or with Lenti-pLKO or Lenti-shMETTL3 (B) (n=4 per group). β-Actin served as the loading control. (C) Cell viability was evaluated by using CCK-8 kit in HASMCs with the indicated treatments (n=4 per group). (D) Cellular injury was measured by using an LDH kit in HASMCs with the indicated treatments (n=4 per group). (E) Lipid peroxidation was evaluated by using an MDA assay kit in HASMCs with the indicated treatments (n=3 per group). (F-I) Immunofluorescence staining of 4-HNE in HASMCs with the indicated treatments (F and H), and quantitative results are displayed in (G) and (I) (n=4 per group). *p<0.05 vs Lenti-Flag or Lenti-pLKO (A and B), or vs Lenti-METTL3 (C-I).

Figure 8

Schematic summary. Our results demonstrate that METTL3 is upregulated in the aortas of TAAD patients, and METTL3 overexpression facilitates ferroptosis of HASMCs by promoting the mRNA degradation of SLC7A11 and FSP1, then reducing their protein levels. Furthermore, ferroptosis is activated during the development of AD, and inhibition of ferroptosis by liproxstatin-1 largely abrogates BAPN-induced AAD in mice. These results suggest that inhibition of METTL3 or ferroptosis is an effective intervention strategy for AD.