Association between Type I interferon and depletion and dysfunction of endothelial progenitor cells in C57BL/6 mice deficient in both apolipoprotein E and Fas ligand

Abstract

Patients with systemic lupus erythematosus (SLE) have a tremendously increased risk for cardiovascular disease (CVD), which could not be accounted in entirety by traditional Framingham risk factors. To study whether the accelerated atherosclerosis in SLE patients is mediated by type I interferon (IFN-I) through the regulation of endothelial progenitor cells (EPCs), we created a line of C57BL/6 mice with deficiency in both apolipoprotein E (ApoE-/-) and fas ligand (FasL-/-, gld.). As expected, the resultant gld. ApoE-/- mice exhibited both aggravated lupus-like disease and atherosclerosis under normal diet. Increased expression of IFN-I-stimulated genes (ISGs) was closely associated with depletion and dysfunction of EPCs, as well as with accelerated atherosclerotic lesion in gld. ApoE-/- mice. While only IFN-α instead of other interventions, including tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), IRS423 and IRS661, impaired EPC function in vitro. Mechanistically, activation or inhibition of the TLR7/9 signaling could modulate EPC number and function in vivo. Decreased proliferation rate and increased apoptotic rate of EPCs induced by IFN-α might contribute to the results to a certain extent. Thus, our data suggest that excessive expression of IFN-I through the activation of TLR7/9 signaling may induce accelerated atherosclerosis in lupus through the depletion or dysfunction of EPCs, suggesting that targeting IFN-I might have potential therapeutic effects on both lupus disease and premature atherosclerosis in SLE patients.

Keywords: Endothelial progenitor cells; Lupus erythematosus; Systemic; Type I interferon; gld. ApoE−/− mice.

Fig. 1.

Gld. ApoE−/− mice exhibited both lupus-like features and atherosclerosis under normal chow diet. (A) Higher splenic index in gld. ApoE−/− mice (105.88 ± 15.01 vs 125.25 ± 21.08 vs 106.25 ± 10.14 vs 150.46 ± 15.31). (B) Elevated serum levels of auto-antibodies including ANA (23.95 ± 4.13 vs 48.91 ± 8.86 vs 97.66 ± 13.03 vs 158.31 ± 11.28), anti-dsDNA (148.93 ± 10.74 vs 163.47 ± 14.87 vs 214.50 ± 4.63 vs 269.61 ± 20.40) and total IgG (248.97 ± 8.29 vs 281.89 ± 3.96 vs 361.06 ± 15.79 vs 417.73 ± 13.98) in gld. ApoE−/− mice. (C) Increased serum levels of BUN (6.00 ± 1.00 vs 8.00 ± 1.22 vs 11.00 ± 1.87 vs 16.00 ± 3.67) and urinary protein (114.18 ± 11.21 vs 151.08 ± 15.02 vs 298.38 9.42 vs 361.48 12.97) in gld. ApoE−/− mice. (D) The glomerulus of gld. ApoE−/− mice was enlarged with hyper cellularity and mesangial expansion (H & E staining, magnification, ×400). (E) A greater amount of IgG was found in kidney sections from gld. ApoE−/− mice (magnification, ×400). (F) Elevated serum levels of LDL cholesterol (0.66 ± 0.08 vs 1.23 ± 0.08 vs 4.71 ± 0.51 vs 5.86 ± 0.82) in gld. ApoE−/− mice. (G) Atherosclerotic lesion area (0.10 ± 0.10 vs 0.20 ± 0.10 vs 1.40 ± 1.00 vs 13.20 ± 2.50) was increased in gld. ApoE−/− mice, which was macroscopically visible by oil red O staining. Values were shown as means ± SEM. n = 7 mice per group in (A), (B), (F) and (G), and n = 6 mice per group in (D) and (E) n = 5 in each group for BUN and n = 9 for proteinuria. *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 2.

Gld. ApoE−/− mice showed decreased EPC number and impaired EPC function. (A) Gating bone marrow EPCs with isotype controls or Sca-1⁺ and CD309⁺ antibodies after deletion of lineage-positive cells. (B) Decreased percentage of peripheral blood (3.21 ± 0.75 vs 2.89 ± 0.58 vs 2.15 ± 0.51 vs 1.03 ± 0.41) and bone marrow (9.94 ± 1.24 vs 5.39 ± 1.34 vs 3.34 ± 0.44 vs 2.56 ± 0.73) EPCs (Sca-1⁺ and CD309⁺ of lin-cells) in gld. ApoE−/− mice compared with age matched gld. and apoE−/− mice. (C) DII-ac-LDL and FITC-UEA-1 staining showed that EPC colony-forming units (87.50 ± 4.00 vs 83.20 ± 4.80 vs 65.60 ± 8.80 vs 45.20 ± 8.10) were decreased in gld. ApoE−/− mice. (D) Reduced EPC capacity to form the cavity structure (17.50 ± 2.40 vs 14.30 ± 1.50 vs 12.50 ± 1.20 vs 7.00 ± 2.30) in gld. ApoE−/− mice compared to age matched gld. and apoE−/− mice. (E) The adherent function of EPCs (86.00 ± 8.30 vs 73.00 ± 12.00 vs 65.00 ± 8.00 vs 50.00 ± 6.30) was impaired in gld. ApoE−/− mice compared to other groups. Values were shown as means ± SEM. n = 6 mice per group. *p < 0.05, **p < 0.01, ***p < 0.001, ns = no significant difference.

Fig. 3.

Association between EPCs depletion and ISGs over-expression. (A) Percentages of peripheral blood and bone marrow EPCs in gld. ApoE−/− mice were negatively associated with atherosclerotic lesion area. (B) Elevated mRNA expression of the ISGs, including IRF7 (1.01 ± 0.11 vs 1.77 ± 0.28 vs 1.45 ± .04 vs 3.77 ± 0.38), MX1 (1.20 ± 0.03 vs 3.50 ± 1.04 vs 1.51 ± 0.06 vs 7.50 ± 2.04), MX2 (1.13 ± 0.18 vs 2.12 ± 0.33 vs 2.01 ± 0.26 vs 3.12 ± 0.93), OAS1 (1.10 ± 0.17 vs 3.56 ± 0.47 vs 2.30 ± 0.38 vs 6.57 ± 1.17), OAS2 (1.03 ± 0.03 vs 2.77 ± 0.32 vs 1.44 ± 0.06 vs 5.77 ± 1.82), IFIT-2 (1.14 ± 0.11 vs 2.12 ± 0.56 vs 1.21 ± 0.32 vs 6.81 ± 1.30) and IFN-γ (1.04 ± 0.02 vs 2.46 ± 0.56 vs 1.33 ± 0.51 vs 8.34 ± 1.34), in bone marrow EPCs from gld. ApoE−/− mice. (C) and (D) IFN score was positively associated with levels of anti-dsDNA antibody, proteinuria (C) and LDL cholesterol (D) in gld. ApoE−/− mice. (E) and (F) IFN score was negatively associated with percentages of peripheral blood and bone marrow EPCs, and positively associated with atherosclerotic lesion area in gld. ApoE−/− mice. n = 6 mice per group. *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 4.

IFN-α but not IL-1β or TNF-α affects EPC function. (A) IFN-α (16.10 ± 4.60), but not IL-1β (42.20 ± 6.10), TNF-α (46.10 ± 7.10), IRS661 (45.20 ± 5.00) or IRS423 (39.70 ± 6.80), impaired the capacity of EPCs to uptake DIL-ac-LDL, compared to PBS group (51.30 ± 10.00). (B) The adherent function of EPCs from gld. ApoE−/− mice was significantly impaired in the presence of recombinant IFN-α (40.10 ± 5.80) but not IL-1β (54.30 ± 4.00), TNF-α (50.30 ± 8.50), IRS661 (56.10 ± 6.20), IRS423 (54.20 ± 7.80), or PBS group (56.80 ± 6.00). (C) Reduced capacity of EPCs to form the cavity structure after IFN-α (2.10 ± 0.0) stimulation, but not IL-1β (5.7 ± 1.00), TNF-α (5.5 ± 0.80), IRS661 (6.20 ± 1.20), IRS423 (5.60 ± 1.00), or PBS group (6.70 ± 1.50). n = 6 mice per group. *p < 0.05, **p < 0.01, ***p < 0.001, ns = no significant difference.

Fig. 5.

IFN-I aggravated lupus-like disease in gld. ApoE−/− mice. (A) Elevated mRNA expression of the ISGs, including IRF7 (1.00 ± 0.10 vs 2.77 ± 0.28 vs 0.32 ± 0.04), MX1 (1.32 ± 0.02 vs 7.50 ± 0.04 vs 0.51 ± 0.06), OAS1 (1.11 ± 0.07 vs 6.57 ± 1.17 vs 0.31 ± 0.08), OAS2 (1.02 ± 0.02 vs 5.77 ± 0.82 vs 0.44 ± 0.06), IFIT-2 (1.10 ± 0.10 vs 5.12 ± 0.90 vs 0.20 ± 0.02) and IFN-γ (1.05 0.05 vs 7.47 ± 1.45 vs 0.33 ± 0.10), in gld. ApoE−/− mice was restored after IRS661 treatment. Meanwhile, IRS423 led to an opposite effect. (B) Smaller glomerulus with hypocellularity and fewer mesangial expansion in gld. ApoE−/− mice was observed after IRS661 treatment (H & E staining, magnification, ×400), while IRS423 led to a large number of inflammatory cell infiltration in kidney sections. (C) IRS661 treatment led to lower IgG deposition in kidney sections from gld. ApoE−/− mice, while IRS423 increased IgG deposition. (D) Serum levels of auto-antibodies including ANA (137.04 ± 18.20 vs 283.72 ± 53.20 vs 113.49 ± 15.78), anti-dsDNA (265.68 ± 19.69 vs 600.21 ± 80.45 vs 228.75 ± 11.42) and total IgG (398.18 ± 10.97 vs 434.29 ± 12.40 vs 310.83 ± 15.60) in gld. ApoE−/− mice elevated after IRS423 treatment. n = 7 mice per group. (E) Alterations of serum BUN (20.00 ± 4.74 vs 22.00 ± 7.68 vs 6.00 ± 1.00) and urinary proteinuria levels (360.78 ± 19.69 vs 484.13 ± 40.15 vs 273.46 ± 25.33) in gld. ApoE−/− mice after IRS661 or IRS423 treatment. For proteinuria test, n = 9 mice per group, and for BUN test, n = 5 mice per group. (F) Alterations of spleen size in gld. ApoE−/− mice (151.36 ± 18.03 vs 216.28 ± 36.13 vs 130.41 ± 11.01). IRS423: TLR7/9 agonists; IRS661: TLR7/9 antagonists. n = 6 mice per group in (A), (B), (C) and (F), and n = 7 mice per group in (D) n = 5 in each group for BUN and n = 9 for proteinuria.*p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 6.

IFN-I contributed to EPC dysfunction and atherosclerosis in gld. ApoE−/− mice. (A) IRS661 normalized the percentages of peripheral blood (1.29 ± 0.18 vs 0.64 ± 0.08 vs 1.98 ± 0.21) and bone marrow EPCs (2.80 ± 0.80 vs 1.23 ± 0.41 vs 3.89 ± 0.63) in gld. ApoE−/− mice, while IRS423 led to an opposite effect. (B) IRS661 enhanced the capacity of EPCs to uptake DII-ac-LDL (45.80 ± 5.00 vs 25.40 ± 5.00 vs 69.00 ± 7.00). (C) The capacity of EPCs to form the cavity structure (6.70 ± 1.20 vs 2.00 ± 1.00 vs 12.50 ± 1.50) in gld. ApoE−/− mice was restored after IRS661 treatment. (D) The impaired adherent function of EPCs (48.50 ± 5.30 vs 31.00 ± 3.00 vs 68.00 ± 5.00) from gld. ApoE−/− mice was significantly overturned after IRS661 treatment. (E) Atherosclerotic lesion area (11.10 ± 2.10 vs 20.20 ± 7.10 vs 6.20 ± 3.50) in gld. ApoE−/− mice was significantly aggravated after IRS423 treatment, while IRS661 treatment markedly improved the lesion. n = 6 mice per group. *p < 0.05, **p < 0.01, ***p < 0.001.